CCE2007 Proceedings

CIVIL ENGINEERING
C. Ionescu, F. Paulet-Crainiceanu, R. Scinteie, A. Nicuta

editors
EDITURA SOCIETATII ACADEMICE "MATEI - TEIU BOTEZ"

Iasi, 2007
C i v i l E n g i n e e r i n g
C. Ionescu, F. Păuleţ-Crăiniceanu, R. Scînteie, A. Nicuţă

Editors
EDITURA SOCIETĂŢII ACADEMICE "MATEI-TEIU BOTEZ"

Iaşi, 2007
Proceedings of the Fifth International Symposium
“Computational Civil Engineering 2007”
Iaşi, România, May 25, 2007
C. Ionescu, F. Păuleţ-Crăiniceanu, R. Scînteie, A. Nicuţă
Editors
Descrierea CIP a Bibliotecii Naţionale a României
Civil engineering / ed.: C. Ionescu, F. Păuleţ-Crăiniceanu,

R. Scînteie, A. Nicuţă. - Iaşi : Editura Societăţii
Academice "Matei - Teiu Botez", 2007
Bibliogr.
ISBN 978-973-8955-14-1
I. Ionescu, Constantin (ed.)

II. Păuleţ-Crăiniceanu, Fideliu (ed.)
III. Scînteie, Rodian (ed.)
IV. Nicuţă, Ana (ed.)
624
Computerized processing: F. Păuleţ-Crăiniceanu, R. Scînteie

All rights reserved, © Societatea Academică "Matei-Teiu Botez" Iaşi, România, 2007
“Computational Civil Engineering 2007”, International Symposium
Table of Contents
1. Păuleţ-Crăiniceanu, F., Ionescu, C., Scinteie, R., Nicuta, A.M.

Civil Engineering, a domain based on computation 7
2. Apeltauer, T., Holcner, P., Kysely, M., Macur, J.
Non-linear effects in traffic simulations 25
3. Kaspars, B., Korjakins, A.
Influence of deformation at a heritage building support on stability of groined masonry
arch 39
4. Jacek, B.
Application of MS Excel® Solver in Solving Structural and Management
Problems 55
5. Boštík Jiří
Underground structure in discontinuous rock mass 66
6. Dósa Adam, Valentin-Vasile Ungureanu
Discrete model for the stability of continuous welded rail 75
7. Ene Petre
Soil-Structure Interaction Effects of Damped Structures 85
8. Gherman Liviu, Petru Mihai, Nicolae Florea, Constantin Gavriloaia, Ioan Paul Vodă
Numerical Modeling of Fiber Reinforced Concrete (FRC). Fiber orientation
angle α = 00, fiber length lf = 50mm 95
9. Gilewski Wojciech
A new look into finite element templates 103
10. Iancovici Mihail, Georgiana Ionică
Evaluation of the inelastic demand of structures subjected to multiple ground motions113
11. Irimia Roxana-Adina, Vlăsceanu Alina Nicoleta
Databases-business support for the Civil Engineering 124
12. Boţi Nicolae, Irina Lungu, Ioan Boţi
Foundation solutions for a marketing center in Cluj 134
13. Kopenetz Ludovic G., Ferdinand-Zsongor Gobesz
The Expertise of Offshore Structures 140
14. Kotek Petr, Jordán Filip
Scenario and impact of global warming in building energy
performance simulation 148
15. Lungu Irina, Nicolae Boţi, Anghel Stanciu
Settlement prediction of a pile group based on design charts 159
16. Mesároš Peter, František Mesároš
The importance of controlling subsystem in information systems for building industry165
17. Musil Roman
Analysis of Water Consumption in Residential Building 172
18. Radinschi Irina, Brindusa Ciobanu
Improving Civil Engineering Physics Teaching-Learning with Mathematica 5.1 180
3
19. Radinschi Irina, Brindusa Ciobanu, Mircea Daniel Frunza
Physics with Maple 9.5 191
Physics Studies – Computational Methods, a Strong Connection 201
Computer Algebra System for Energy-Momentum Localization 211
22. Siwiec Przemysław, Sebastian Toś
The state of stresses in RC beams with installation holes located in compressed zone 220
23. Slonovschi Andrei, Ion Antonescu, Liviu Prună
Computer Graphical Communication an Efficient Tool for Understanding the Design of
Wooden Trusses 231
24. Smutny Jaroslav, Lubos Pazdera
Analysis of dynamic parameters of rail fastening by Rihaczek transformation 246
25. Smutny Jaroslav, Lubos Pazdera
The use of time frequency transformations in testing structural elements 256
26. Stanciu Anghel, Oana Colţ, Irina Lungu
Design of reinforced soil works – Textomur structures – based on the computer program
Cartage 269
27. Voiculescu Dragos, Daniela Preda
Computer program for Pipe Section Column Sizing 277
28. Cornelia Victoria Anghel
Computational method and application using Newton Algorithm 282
29. Libuše Beckerová, Gergely Bölckei and Jiří Brožovský
Partial substitution of Cement in Concrete by Finely Ground Brick Body 287
30. Constantin Amariei, Iulian Gabriel Mihai
Minimum weight buildings design using inequalities method 295
31. Brindusa Ciobanu, Ion Silisteanu, Irina Radinschi
Scientific Perspectives for Future Research Work in Fundamental Properties of Short-
Lived Radionuclides from Decay and from in-Beam Studies 307
32. Gabriel Oprisan, Nicolae Taranu, Vlad Munteanu
Experimental and numerical analysis of compressed concrete elements confined with
FRP composites 316
33. Florin Ţepeş Onea
Rayleigh models coefficients 326
34. Magda Brosteanu, Teodor Brosteanu
NDT Devices as Long-Term Vision in Assessing of Old Masonry Buildings by
Probabilistic Approach 334
35. Magda Brosteanu, Teodor Brosteanu
IT Approach in Heat Transmission Visualizing through Opaque Walls 346
36. Adrian Prodescu, Daniel Bîtcă
Comparative Study on the Results of Analytical and Experimental Analysis of a Steel
Taintor (Radial) Gate 354
37. Jiří Brožovský, Alois Materna and Ivan Kološ
Contribution to non-linear constitutive modelling of masonry structures in 2D 362
38. David Horak, Martin Zlamal, Petr Danek
Behavior of newly developed FRP reinforcement in structures under various load
schemes 368
4
39. Josef Fink, Lubomir Ondris
New Shear Connectors for Composite Girders – Experiences with ABAQUS Push-Out
Test Simulations 379
40. Ludovic G. Kopenetz, Ferdinand-Zsongor Gobesz
The Structural Expertise of Steel Cables 396
41. Mária Kozlovská, Michal Danko
Computer aided building diary 410
42. Ioan Moga, Ligia Moga
Determination of the heat flows, of the solar and luminous characteristics of the glazing
and of the devices for the solar protection. Automaton calculus program “SOLAR” 425
43. Radomir Sokolar
Effect of moulding humidity on the properties of dry pressed ceramic tiles 437
44. Jerzy Szołomicki
Application of damage model for analysis of masonry structures 442
45. Constantin Ionescu
A New Form of the Active Moments Method 448
46. Elena Carmen Teleman, Elena Axinte
Laboratory analysis in B.L.W.T. – priorities in insuring a safety curtain walling design465
47. Grzegorz Dmochowski, Przemysław Siwiec and Piotr Berkowski
FEM 3D analysis of RC frame foundations of rotary cement kiln 477
48. L. De Doncker, P. Troch, R. Verhoeven, K. Buis, P. Meire
Flood routing using ‘Femme’. 487
49. Przemysław Siwiec, Sebastian Toś
Simulation of cooling tower collapse on the basis of non-linear concrete model and FEM
analysis 498
50. Irene Daprà, Giambattista Scarpi
Unsteady axial Poiseuille flow of a Bingham fluid in an annulus 508
51. Tomáš Fojtík, Jiří Brožovský
Formation of lime efflorescence on concrete and concrete products 521
52. Monica Gheorghiu, Mirela Chelcea
The use of computer means in the representation of the ruled surfaces 530
53. Horaţiu Mociran, Eugen Panţel
Numerical studies on the seismic performance of three structural systems 544
54. Doina Stefan, Violeta-Elena Chitan
Identification Methods of the Nonlinear Systems Subjected to Seismic Actions 554
55. GRAITEC ADVANCE software
GRAITEC – CAD and Analysis Software for the Construction Industry 562
56. Jaroslav Smutny, Lubos Pazdera
The use of the bispectrum for analysis dynamic parameters of rail fastening 568
57. Petr Kalvoda, Jiri Vondrak, Miroslava Sucha, Radovan Machotka
Subpixel Measurement in Close-Range Photogrammetry 576
58. Mircea Balcu, Ştefan Marian Lazăr
The Generalized Mohr-Coulomb (GMC) Yield Criterion and some implications on
characterization of pavement materials 585
59. Alina Nicuţă, Constantin Ionescu
Quality costs in Bridge Engineering 592
5
60. Lenka Nevrivova
Investigation of mechanical properties of castable with respect to their application as
ceramic anchor elements 600
61. Valentin-Vasile Ungureanu, Marius Comanici
Sensitivity study of a model for the stability analysis of continuous welded rail 608
62. Valentin-Vasile Ungureanu, Adam Dósa
Parametrical study of the effect of the torsional resistance of the fastenings on the
stability of continuous welded rail 616
63. Călin Moga, Ştefan I. Guţiu, Gabriel Urian
Concrete-filled steel tubular columns. Shear connection design 625
64. Petru Moga, Ştefan Guţiu, Cătălin Moga
Composite columns with circular section. Shear transfer mechanism 631
65. Ioana Entuc, Mihai Budescu, Ioan Ciongradi, Gabriel Oprisan
The meshing influence upon the design model of reinforced concrete chimneys
structures 639
66. Dan Diaconu-Şotropa, Ioan Păvăloi
Temperature Distribution Computing System for 2D domains 647
67. Jiří Bydžovský, Amos Dofka, Zdeněk Šnirch
Possibility increase cooling tower service life in terms of long-term monitoring 656
68. P. Tragakis, D. Stoica, S. Majewski, M. Voiculescu
Some general considerations about the behaviour and retrofitting solutions for the
existing buildings with masonry structures 664
69. V. Warnotte, D. Stoica, S. Majewski, M. Voiculescu
State of the art in the pounding mitigation techniques 681
70. Dragos Stan Caciula and Nicolae Ungureanu
The calculation of a cable network using the tangent stiffness matrix 697
71. Constantin Ionescu
City of Iassy’s monumental constructions 710
72. Ludovic Kopenetz, Alexandru Cătărig, Titu Hodişan
The design of large sized floors 734
73. Alexandru Cătărig, Ludovic Kopenetz, Titu Hodişan
Structural analysis of sandwich type composites 742
74. Jiří Brožovský, Tomáš Fojtík, Amos Dufka
Resistance of Concrete with Fine-Ground Granulated Slag Admixture to Sulphates and
Acidic Aggressive Media 750
75. Mihai Coveianu, Olimpia Marchelov, Daniel Bîtcă
Comparative Analysis of Some Methods for Dynamic Behavior and the Structural Design
of Multistory Structures 760
76. Oana Mihaela Stan
Wind Dynamic Response of a Suspension Pedestrian Bridge 769
77. Ion Anghel
The modeling, simulation and numerical analysis of a fire in a construction with light
structure 779
78. Laura Fraifeld, Ramona Agachi, Horst Werkle, Franz Zahn, Gabriela Atanasiu
Capacity based earthquake design of RC shear wall according to EC8 790 7
6
Civil Engineering, a domain based on computation

Fideliu Păuleţ-Crăiniceanu1, Rodian Scînteie2, Constantin Ionescu1 and
Alina Mihaela Nicuţă1
1
"Gh. Asachi" Technical University of Iaşi, România
2
CERT-CESTRIN Bucharest, Romania
Summary
In Civil Engineering, the computation is at the base of almost all activity. An

international symposium as Computational Civil Engineering 2007, [1], is a good
opportunity to present researches, technical and scientific work in the field and
represents a show on the importance computers and computations play in this
domain.
This paper tries to withdraw the main ideas, approaches and achievements
reflected in the proceedings of the above mentioned symposium.
A nice tradition regarding computation in Civil Engineering is already established.
The diversity of work together with the diversity of countries, universities and other
institutions implied in Computational Civil Engineering symposiums lead to the
conclusion that the field is strong and needed.
KEYWORDS: Civil Engineering, computers, research, education, computation in

engineering and science.
1. INTRODUCTION
In Civil Engineering, as in other engineering fields, information processing is vital.

This processing must be done quickly and reliable.
As computational means are progressing exponentially, the engineering is also
progressing. Civil Engineering is now in the middle of this progress. It could be stated
that computational means and instrumentation are ahead of their full use in Civil
Engineering.
Every year since 2003 in Iasi, Romania, the international symposium Computational
Civil Engineering is held. The proceedings of the symposiums show that the
computational/instrumental “revolution” is underway.
During the Computational Civil Engineering 2007 symposium complex applications
were presented [1]. The general evolution (compared to the previous years) is obvious.
The mixture between civil engineers and computer/automation engineers (sometimes in
8 F. Păuleţ-Crăiniceanu, R. Scînteie, C. Ionescu, A.M. Nicuţă
the same person) is becoming essential, though the inertia (specific to construction
domain) is still powerful.
This paper is doing a review of the achievements/problems specified by the authors
participating at the above mentioned symposium.
2. COMPUTER-BASED CIVIL ENGINEERING
Tomas Apeltauer and a team from Brno University of Technology [2] present their
results of the research on traffic behavior. Traffic jams become a challenge all over the
world and huge losses in terms of time and resources are caused by this phenomenon.
In order to mitigate the effects one must understand the intimate processes and overall
behavior. It is therefore necessary to use most advanced tools available. For the
simulation a program was created using Java programming language, due to its easy
accessibility, simplicity and portability on internet. Each vehicle is represented by an
object with its own parameters and the whole system of vehicles is represented by the
so called binded list. In such way it is possible to dynamically change number of
vehicles in the system. The paper presents the nonlinear effect in traffic simulation.
The analysis uses real data collected with different methods like aerial photography and
inductive loops. The traffic was studied at both microscopic and macroscopic using
cellular automata and analog models. Authors work also with the concept of chaos in
traffic flow analysis.
The influence of deformation at a heritage building support on stability of groined
masonry arch is studied by Kaspars Bondars and Aleksandrs Korjakins with
applications at Riga Cathedral [3]. The settlements monitoring of cathedral’s supports
and numerical modeling to define settlement inequality were used as initial information
to compare computer modeled deformed structure shape with existing cracked masonry
structure. Geodesic monitoring by optical tools and mechanical tensometers have been
used for determination of a crack expansion.
Wind, solar radiation and temperature loading give no big influence on computational
results but can be important. Two-step computational method highlights deformed
shape influence on thrust line position in cross section of arch shell. For further study a
long term monitoring program was set since 2005 using SOFO optical tensometer tool.
Jacek Boroń [4] is interested in solving structural and management problems with tools
offered by MS Excel® Solver. Spreadsheet programs are very popular software and
many engineers are using them. The article is offering a model about using their
resources to solve technical and managerial issues. This tool is able to solve small-
scale linear programming (LP), smooth nonlinear programming (NLP), and mixed
integer programming (MIP) problems. Examples are drafted both from structural
design and economic calculation.
Calculation and analysis of the position of underground structures in discontinuous
rock mass are presented by Jiří Boštík in his articles [5]. Three typical alternatives were
“Computational Civil Engineering 2007”, International Symposium 9
considered: the underground structure in rock mass with one-sided dip of

discontinuities, the underground structure situated in synclinal bend and the
underground structure situated in anticlinal bend. Mathematical models were used for
the analysis by Finite Element Method using Plaxis software. Mathematical models
were planar – plane strain. Maximum vertical displacement is calculated for each case.
Adam Dósa and Valentin-Vasile Ungureanu present in their paper [6] a discrete model
which was developed for the buckling analysis of continuous welded rail subjected to
temperature load. The model is based on a nonlinear analysis in total lagrangean
formulation. The structure consists of beam elements and lateral, longitudinal and
torsional spring elements. The source of nonlinearity is due to the geometric
nonlinearity of the rail high axial forces and also to the nonlinearity of material type for
the lateral and longitudinal resistance of the ballast and the torsional resistance of the
fasteners. The track model is encoded into a special purpose program which allows a
parametric study of the influence of vehicle loading, the stiffness properties of the
structure and of the geometric imperfections on the track stability. Both theoretical and
practical experiments data are presented.
In his article Petre Ene [7] deals with the effects of soil-structure interactions in damper
structures. The classical approach is to investigating the seismic behavior of structures
with seismic dampers by neglecting the effects of foundation-structure interaction. The
paper presents a rigorous time-domain procedure to address the interaction effects of
structures equipped with fluid viscous dampers and foundations with an unbounded
medium. Quantitative results show that, during earthquakes, there are significant
differences between a system with and without radiation damping. For greater
accuracy, radiation damping
should be properly taken into account. More-over, the efficiency of fluid viscous
dampers in reducing seismic disturbance of a structure is very dependent on the
flexibility of the foundation.
A new approach on fiber reinforced concrete is presented in the study Gherman et al.
[8]. Numerical modeling is applied, using a finite element method (FEM) program, to
study the behavior of linear structural elements, made from steel fiber reinforced
concrete (SFRC). The results stress out the favorable influence of the fibers on the
bearing capacity and especially on the ductility of the loaded members. Lack of design
standards and practice codes on this material is bringing this field of study under
researchers’ consideration. The results are largely illustrated with tables and pictures,
outputs from the analysis software.
A new look on finite element templates is presented by Wojciech Gilewski [9]. Beams,
plates and shells are widely considered in engineering applications. However the
corresponding discretization procedures are not yet sufficiently reliable. The
formulation aims to satisfy several conditions: ellipticity, consistency and the inf-sup
condition. A template, an algebraic form of element matrices, which contains free
parameters is generated. Setting the parameters to specific values produces element
instances. Two templates are analyzed: Bernoulli and Timoshenko beam. The number
of free parameters is discussed by a general method.
The article of Mihail Iancovici and Georgiana Ionică [10] presents an evaluation of the
inelastic demand of structures subjected to multiple ground motions. The paper
investigates the effect of repeated Vrancea strong ground motions on the behavior
factors and the related parameters that accounts for cyclic structural deterioration due
to inelastic response. A large number of integrated analyses, nonlinear response
analyses and energy balance-based analyses were carried out and estimates were made
on the behavior factors for inelastic SDOF systems controlled by flexure with stiffness
degradation. It is shown that multiple ground motion of Vrancea type for Bucharest,
may lead to an important increase of force and drift demand of structures that usually is
not taken into account by the present seismic design codes.
Roxana-Adina Irimia and Alina Nicoleta Vlăsceanu present in their paper [11] a view
on the use of databases-business support for the civil engineering. The actual
development trend regarding the Databases Management Systems must focus on: using
intensively the integration of the Information Systems; obtaining open applications;
ensuring interoperability and communication for the Information Systems; focusing on
the global performances; Databases Management Systems must become platforms for
the databases. Spatial databases (which have resulted by combining the technology of
data bases with geographical systems and assisted design) and the multimedia
databases (which store and process classical data-text and graphics and also multimedia
data-images, audio and video) are strongly recommended for civil engineering use.
Nicolae Boţi et al. [12] present foundation solutions for a marketing center in Cluj. In
order to accommodate the site conditions three foundation solutions has been
discussed: open caissons, embedded 0.3m into the good foundation soil; drilled short
piles or precast driven piles inserted into the good foundation soil at lest 15d;
foundation beam networks onto an improved soil by a granular material cushion. Based
on the analysis performed on the foundation solutions, the foundation system is
recommended to consist of flexible continuous footing for the external shear walls and
a foundation beam network of the columns inside, all performed on a compacted
ballast cushion.
The expert analysis and assessment of the offshore structures is the subject of the paper
of Ludovic G. Kopenetz and Ferdinand-Zsongor Gobesz [13]. The structural expertise
must cover, beside the vastness of interfering random variables, all-important aspects
in order to avoid the pollution of the pelagic environment. Load evaluation, structural
analysis aspects and problems concerning fatigue test are presented.
Petr Kotek and Filip Jordán [14] are interested by scenario and impact of global
warming in building energy performance simulation. This paper is focused in impact of
different outside temperature and sun radiation during the year and mainly in point of
view: “what if in future”. The climate is lunatic every year and we can see and feel that
the temperature is slowly rising all over the world. Nowadays the question of global
warming is more and more discussed. Two case studies were chosen for different
scenarios of climate change First and second cases are one real administration building
in Czech Republic, Prague. One is performed without thermal insulation on the wall
construction and second with and their appropriate sensitivity in point of view of
climate change is shown.
Lungu Irina, Nicolae Boţi and Anghel Stanciu [15] analyze the possibility to predict
settlement of a pile group based on design charts. The article starts from the strong
intention to modify the actual capacity based design to a settlement based design when
foundation systems on piles are concerned. Two main issues are treated: settlement
prediction of a pile group based on the single pile behavior; and combination of
boundary element with finite element approach.
Design charts proved to be useful tools to set comparative values at work in order to
select the foundation elements in the best combination to face even the most
pessimistic hypothesis of soil-foundation interaction. A successful and economic
design is reached only when the interaction effects are taken into account.
The importance of controlling subsystem in information systems for building industry
is illustrated in the article of Peter Mesároš and František Mesároš [16]. The paper
presents the necessity, possibility, assumptions and advantages of implementation of
the controlling subsystem. In this context authors advise to institutionalize the position
of the controller who, on the basis of his knowledge and practical experience, could
analyze the detected deviations between the plan and reality. This could assure the
uninterrupted operation of information flows, i.e. providing objective information for
decision-making on the basis of real processes taking place in the construction.
Roman Musil is interested in the analysis of water consumption in residential building
areas [17]. Water supply is essential for residential areas. Water need must be assessed
and predict. Using monitoring systems and computation tools the author is evaluating
the daily water need, but also minimal and maximal hourly consumption. Results are
used for better planning of the water supply and better use of accumulate reservoirs and
automatic pumping stations.
Computational instruments are not only used for industrial purpose but also as teaching
tools. Irina Radinschi and Brindusa Ciobanu [18] are using Mathematica 5.1 to
improving civil engineering physics teaching-learning process. Some examples for
studying physical concepts and phenomena and for plotting a general class of physics
phenomena are presented. The software provides a good environment for computation,
a high-level programming language, text, graphics, and animation. Further, embedded
tools allow carrying out length calculations and evaluating the components of some
tensor and pseudo-tensor quantities.
Radinschi Irina, Brindusa Ciobanu and Mircea Daniel Frunza [19] are presenting the
possibility to study Physics with Maple 9.5. The paper sketches how computer algebra
systems like Maple 9.5 can be used to helping the students connecting physical
concepts and phenomena with powerful mathematical formalisms and graphical
representations. The experience in this area demonstrated that computational physics,
together with traditional theoretical and experimental physics, is a reliable tool which
helps students to explore a wide variety of phenomena and give them a deeper
understanding of these topics. Static charts and animation might be used to demonstrate
physical behavior like wave propagation or kinetic and potential energy. The main goal
is to allow students the ability to learn physics, to perform length calculations, to make
graphical representations and introduce animations, and also to solve practical

problems.
Irina Radinschi and Brindusa Ciobanu [20] uncover a strong connection between the
computational methods and study of physics. Platforms as Mathematica and Maple
have advantages as flexibility and speed, and more advanced graphical facilities.
Further, the computational methods allow a rapid learning of physics and they can
successfully replace the traditional methods.
Radinschi Irina and Brindusa Ciobanu [21] develop a computer algebra system for
energy-momentum localization. Maple 9.5 with the GrTensorII package is used to
perform complex calculations. Procedures are used for evaluating the energy and
momentum components of the Einstein, Landau-Lifshitz and Møller energy-
momentum complexes. Some of the results are presented, and the metric under
consideration describes a magnetic stringy solution. GrTensorII yields the energy and
momentum components which are involved in the expression for energy and
momentum densities. The expressions for energy in the Einstein, Landau-Lifshitz and
Møller prescriptions are plotted.
The state of stresses in RC beams with installation holes located in compressed zone is
studied by Przemysław Siwiec and Sebastian Toś [22]. The paper presents the
influence of locating holes in the compressed area of the reinforced concrete beam. In
order to research the actual behavior of the elements a numerical model based on Finite
Element Method was created. Two independent analyses were performed in which
linear and nonlinear constitutive relations for concrete and for steel have been used.
The results acquired have been compared and discussed. The nonlinear computer
model proved to be highly reliable for the analyses of cracked RC elements.
Computer graphics become efficient tools in understanding the design of wooden
trusses. Andrei Slonovschi, Ion Antonescu, and Liviu Prună [23] use 3D models to
design study and verify wood structures. Use of computer graphical communication
and of CADD programs (like AutoCAD) together with classical graphical disciplines
(Descriptive Geometry, Technical Drawing) is essential in forming and developing of
spatial vision and the ability of reading triple orthographic projection drawings, tools
useful for future design and execution engineers.
The paper of Jaroslav Smutny and Lubos Pazdera [24] analyze the dynamic parameters
of rail fastening. For evaluation of response signals obtained by rail fastening analysis
a new method using time and frequency related transformations has been developed. In
the paper the laboratory measurements and dynamic parameter analyses of flexible
fastening of Vossloh SKL14 type were described. Short Time Fourier Transformation
and Rihaczek Transformation were used for time-localization of the occurrence of
frequency elements of stationary and non-stationary signals. The method can be used
for designing new rail fastening systems and their parts.
Smutny Jaroslav and Lubos Pazdera present the use time frequency transformations in
testing structural elements. The paper [25] deals with the non-destructive testing of
structural elements by means of the acoustic response using the jointed time and
frequency transforms: short time Fourier and wavelet transforms. These methods make
possible to localize the beginning and the end of frequency components contained in
the measured signal and they enable to perfectly analyze the spectrum of the non-
stationary noise. This way, this mathematical procedure enables to distinguish a good
specimen from a defective one.
Based on the measurement and analyses made, it is possible to state that the experiment
checked the possibilities of using the given methodology for the detection of structural
defects of the measured products and materials from the homogeneity and the cracks
formation points of view.
Reinforced soil represents an alternative solution to consolidate earth massifs and
perform retaining structures included in the present transportation infrastructure.
Anghel Stanciu, Oana Colţ and Irina Lungu [26], based on the computer program
Cartage, present a method to design of reinforced soil works. The computer program
can evaluate the stress within the reinforcement, the displacement of the soil massive,
and the safety coefficients at anchorage as well, based on the stability analysis of the
massive in case of cylindrical failure surfaces going through the toe of the reinforced
soil structure. The method may significantly decrease the design time.
Voiculescu Dragos and Daniela Preda developed computer software for pipe section
column sizing. The paper [27] presents the Excel program which has three main parts:
the first one is the data input, where the proposed size of the column is set; the second
one is the load section, where the maximum loads are input from the structural analysis
program, such as SAP or ETABS; the third part gives the results of the structural
checks on the proposed section
In [28], Anghel is presenting an implementation of Newton’s algorithm (tangent
method) in solving nonlinear equations. A C++ program is shown together with an
example. Conditions of availability and discussions on the method are extensively
treated.
Libuše Beckerová, Gergely Bölckei and Jiří Brožovský observe [29] that the use of the
brick chipping from brick factories could be very useful in composition of concrete.
This solution is a way to recycle a part of the waste and to have a more environment
friendly construction industry. The finely ground brick body shows pozzuolana activity
and, tehrefore, it can replace in part even the cement part in the concrete. As a
consequence, the price of concrete could be lowered. The authors are showing an
analysis of the chemical composition of the compounds in the brick body. Also,
experiments are presented in details (compositions, parameters, methods). Conclusions
show very good behavior of the concrete characteristics (compression strength,
bending strength, depth of the pressure water ingress, dynamic modulus of elasticity).
The method of inequalities method is proposed by Amariei and Mihai [30] for
minimum weight building design. Ways and hypotheses to compute optimization of the
weight for structures is firstly analyzed. Then the proposed inequalities method is
presented. Static equilibrium equations, plastic yielding conditions and the weight
function are discussed. An extended numerical example is probing the method.
Ciobanu, Silisteanu and Radinschi [31] are presenting a project focussed on

fundamental properties of short-lived radionuclides from decay and from in-beam
studies. A first goal is shown to be the formulation of a reaction theory for cluster
decay in which the clustering and reaction amplitudes will be represented by means of
resonance formulas. A brief outlook of the experimental results and theoretical ideas
which define the field is presented together with quantities estimations for the
resonance decay widths and relative intensities. A direct numerical integration using
step-by-step methods on computer is envisioned.
In their paper, “Experimental and numerical analysis of compressed concrete elements
confined with FRP composites”, [32], Gabriel Oprisan, Nicolae Taranu and Vlad
Munteanu show a method proposed for increasing the strength and ductility of concrete
subjected to compressive loading. A detailed testing methodology and an experimental
program are developed. Compositions used for fiber reinforced polymer (FRP) resin
together with the properties of glass/carbon fibers are shown. Experimental stages were
noticed. Finite element models for the specimen used in experiments were analyzed.
Experimental and analytical results have shown very good agreement, proofing the
chosen solutions.
Coefficients for Rayleigh models of damping are investigated by Florin Ţepeş Onea in
[33]. Damping coefficients are obtained as a linear combination between the mass and
stiffness matrices, using a Cauchy sequence. The approach leads to solving the
problem of the influence of superior modes of vibration on the energy dissipation in the
backlash of a dam. FEM 2D and 3D models were used. Analyzing the results it was
observed that, for two modes of vibration, the effect of the mass matrix increase and
the effect of stiffness matrix decrease in the same direction with the increase of the
second frequency taken into account.
In [34], the authors, Magda Brosteanu and Teodor Brosteanu, refer to the non-
destructive testing devices in assessing of old masonry buildings. They focus on flat-
jacks technique to determine the state of stress in masonry. Also, they are performing
health monitoring using IR thermal-graphic method. The obtained data is used for
strengthening structural design of masonry members and thermal renovation design.
Photos on instrumentation and buildings are presented.
The same authors write a paper titled, “IT Approach in Heat Transmission Visualizing
through Opaque Walls”, [35]. Heat flow image and density, surface temperature
distribution, isothermal lines are the problem solved through computational processing
of infrared (IR) photographs. Mathematical models are presented and analyzed.
Applications using Finite Element Methods for determination of the thermal resistance
for different typical wa;; structures are shown. The effectiveness of the IT means is
underlined.
A “Comparative Study on the Results of Analytical and Experimental Analysis of a
Steel Taintor (Radial) Gate” is performed by Prodescu and Bîtcă in [36]. They intended
to test the stress and strain state of a hydraulic gate undertaking a pressure of 20m of
water. For this, a 1:4 scale model was built. Also the model was described and
analyzed through FEM. The comparison between the two cases has shown good
accuracy. Therefore, the models were validated and the required adjustments for design
of the original (to be build) gate were withdrawn.
Jiří Brožovský, Alois Materna and Ivan Kološ are contributing to the solving of the
non-linear constitutive modelling of masonry structures in 2D, [37]. They apply
different laws for bricks and mortar. In the case of mortar, the 2D failure criteria is the
Kupfer's criteria. The Bazant's crack-band model is employed for avoiding the use of
re-refining the FEM mesh. For bricks, an approach similar to non-local material
models is proposed. It guarantees that crack is detected. However, the authors stress
that the brick model must be improved. A FEM model of a small wall and the
corresponding numerical analysis is shown and discussed. Results were the expected
ones.
An extended experimental and numerical work is done in [38]. David Horak, Martin
Zlamal, and Petr Danek propose newly developed FRP reinforcements and see their
effects under various loads (static and dynamic). The reinforcement is used on
longitudinal and (in some cases) on transversal direction in concrete beams. Also,
application of FRP for masonry strengthening was in the views of the authors. Data
experimentally obtained was used in improving the analytical model for the reinforced
structures (FE software based on fracture mechanics of quasi-brittle materials). All
results are very encouraging and the paper shows a high quality experimental and
analytical achievement.
Josef Fink and Lubomir Ondris propose new shear connectors for composite girders
that are subject to patenting at the Technical University of Vienna, Institute of Steel
Constructions. The authors are analyzing the new connectors in experimental work as
in FEM models. Experiences with ABAQUS push-out test simulation are largely
described. Importance of parameters in FEM analysis (mesh, precision, procedure,
careful visual check of graphical results etc) is commented. Validation of numerical
FEM model parameters was achieved and future connectors’ experiments could be,
therefor, optimized.
“Structural Expertise of Steel Cables” is the title of a paper [40] presented by Ludovic
G. Kopenetz and Ferdinand-Zsongor Gobesz. Taking into account the importance of
cable structures, an exhaustive consideration of the main parts an expertise should be
composed from is outlined. In situ and laboratory tests together with analytical work
for determination of fatigue or corrosion are reviewed. The authors discuss the main
topics to be observed in an expertise: steel quality and brands, type of bearing with
cables, the effects of loads (wind, self-weight, thermal loads, dynamic loads, etc). A
diagram of the site’s investigation method based on digital cameras (currently under
patenting) is proposed for cables. Typical patterns of wire rope degradation and failure
are also shown. A computer program (using FEM) is developed in order to help in
cable analysis.
In order to have a better building management, Mária Kozlovská and Michal Danko are
proposing a “Computer aided building diary” [41]. They have developed a database
application that covers all the legal and technical problems involved in building
management. The structure of the application in graphical format is shown. Main
screens are also presented and discussed. Advantages for building management of this
new tool are revealed.
Ioan Moga and Ligia Moga are concerned about the introduction of the legislation in
“determination of the heat flows, of the solar and luminous characteristics of the
glazing and of the devices for the solar protection” [42]. They develop a computer
program called “SOLAR”. Considerations on: glazing system and the solar protection
devices; solar and optical characteristics of the glazing; and heat transfer are
elaborated. Then a validation of the computer program is performed according to the
European (and national) standards.
“Effect of moulding humidity on the properties of dry pressed ceramic tiles” is the
focus of the paper [43] presented by Radomir Sokolar. Use of fly ashes in ceramic
(green or fired) products is in the views of the study. Taking into account a large
interval of moisture for the pressing granulate, and for the two types of ceramic bodies,
the properties of the products were experimentally determined. Also capillarity of fired
bodied in dependence on pressing moisture of pressing granulate is presented. Tables
and graphics illustrate the parametrical studies.
In his paper, [44], Jerzy Szołomicki is applying a damage model for analysis of
masonry structures. The model is based on the finite elements method to simulate the
ultimate response and the mode of failure. It takes into account the nonlinear analysis
of masonry structures (effect of stiffness degradation and the problem of the finite
element mesh). After the mathematical concept (of damage accumulation functions) is
developed the author is defining the function of the damage criterion. Also global
damage indices are developed. They provide accurate data on the state of any
component of a damaged structure and the importance to the entire structure.
“A new form of the Active Moments Method” is proposed by Constantin Ionescu in
[45]. Based on the work of former famous Romanian Professors Anton Sesan, Nicolae
Orlovschi and Gh. Em. Filipescu, the author is improving the method by generalizing
it. Equilibrium equations are written together with a description of the Active Moment
Method. Then the proposed improvements are theoretically demonstrated.
In [46], Elena Carmen Teleman, Elena Axinte are dealing with priorities in insuring
curtain walling’s safety design. The authors are showing the influence of climate
change on Romanian wind action for buildings. Because of new high speed winds, the
effects on the glass-covered buildings rise questions on safety. Design criteria, studies
on safety (accuracy in the determination of realistic values of wind pressure action;
consideration of wind-induced vibrations in design of light envelopes) and laboratory
studies of buildings aerodynamics are the main parts of the paper.
Grzegorz Dmochowski, Przemysław Siwiec and Piotr Berkowski, [47], are performing
a “FEM 3D analysis of RC frame foundations of rotary cement kiln”. Firstly, a
description of the damaged foundations for cement tubular oven are presented. Those
foundations have been designed using 2D methods. Authors show the way they have
been taken into account the damage to the concrete and steel (passive or active). They
also present the high temperature load conditions. For the FEM model osoparametric 3-
D finite elements with eight nodes have been used. Results of the analysis revealed that
the allowable stress was exceeded in some points and/or because of initial in-plane
design damages/cracks are explained. Measures to strengthen the structure are
proposed.
Using ‘Femme’ (a open computer program that simulates the environment and
developed by NIOO-Netherlands Institute of Ecology) L. De Doncker, P. Troch, R.
Verhoeven, K. Buis, and P. Meire, are researching the way to integrate into that
software the flood routing [48]. The complexity for the description and propagation of
waves in rivers is solved through different wave models (based on the kinematic,
parabolic and Saint-venant equations). As an immediate application, the case of river
Aa in Belgium is studied. The importance of the research on environment and
prediction of floods is underlined.
In order to explain observed failure and prevent eventual future failures, Przemysław
Siwiec and Sebastian Toś are conducting a “Simulation of cooling tower collapse on
the basis of non-linear concrete model and FEM analysis”, [49]. They firstly describe
the sequence of the collapse of the cooling tower. The collapse (from 1987) occurred in
a calm day and without any previous signs, therefore practically unexplained. A
computer FEM detailed model (geometry and materials’ properties) is shown and
material and geometric non-linear analyses under various loads are carried out. The
computer simulations lead to the explanation: because of strong winds that occurred
only few days before the collapse together with material degradation were the cause of
that failure phenomenon.
Irene Daprà and Giambattista Scarpi study the “Unsteady axial Poiseuille flow of a
Bingham fluid in an annulus”, [50]. They propose a new model based on the error
function to regularize the constitutive equation for Bingham fluid (as bentonite,
montmorillonite, slurries), fluids which present a yield stress. Steady and unsteady
flow (start-up flow, cessation of flow, pulsatile flow) laws are presented. Numerical
analyses were conducted and solution has been carried out via an implicit finite
difference method.
The problem of lime efflorescence on concrete is a hard task and studied from long
time ago. A paper on “Formation of lime efflorescence on concrete and concrete
products”, [51], is presented by Tomáš Fojtík and Jiří Brožovský. Firstly a description
of the conditions the phenomenon occurs is performed. Then primary and secondary
lime efflorescence notions are explained. The 15 conditions that could lead to
efflorescence occurring and the mechanism involved are shown. Measure that can be
taken for avoiding and measures to remove the efflorescence are also presented.
Various experiments for combating the lime efflorescence were underwent and detailed
by the authors.
Monica Gheorghiu and Mirela Chelcea are showing how “The use of computer means
in representation of the ruled surfaces”, [52], is very useful in teaching Descriptive
Geometry to university students. Very difficult to imagine 3D representations become
easy to produce and easy to understand. The use of animations also helps the teaching
process. Examples of paraboloids, conoids, cylindroids, helicoidal surfaces and
hyperboloids illustrate the teaching/learning proposed procedure.
Horaţiu Mociran and Eugen Panţel are concerned about the problem of limiting
earthquakes’ effects on structures, [53]. They propose a numerical study on seismic
performance of three different structural systems: simple framed structure, base
isolation and viscous damped structure. The loads are three scaled artificial earthquakes
of Vrancea type that matches on P100-1/2006 code response spectrum. Base shear,
interstory drift, floor accelerations and some performance indexes show that viscously
damped frame was best performing.
The topic of [54] is the “Identification Methods of the Nonlinear Systems Subjected to
Seismic Actions”. The two authors, Doina Stefan and Violeta-Elena Chitan, present an
extended overview on the development of stochastic methods in the field of
identification. Techniques for linearization of stochastic equation of motion are firstly
explained for single degree of freedom systems and then for multi-degree of freedom
systems. The case of the non-linear oscillator model is also treated.
An interesting presentation of software is shown by GRAITEC, [55], a company that is
involved in producing computing solutions for Civil Engineering. Structural analysis
with FEM for concrete and steel structures is resolved. Also advanced concrete and
steel design is remarkable. The demonstrations of the software are exemplifying and
recommend the qualities of the proposed integrated design software environment.
In [56] Jaroslav Smutny and Lubos Pazdera are proposing the use of the bispectrum for
analysis of dynamic parameters of rail fastening. The main idea is concentrating around
the Higher Order Statistics (HOS) that can give more insights of nonlinearities and
characteristics of a system. The HOS is a extension of second-order characteristics
such as the auto-correlation function and power spectrum to higher orders in signal and
data processing. In the theoretical background, the authors define some terms as
moments, cumulants, and polyspectra specific to HOS. An example of application is on
the laboratory measurements and the analyses of the dynamic parameters of a rail
fastening specimen. The authors show that using only a Fourier analysis it is almost
impossible to detect the loosened bolts situation. However, using the bicoherence
spectrum, the defective situation is clearly detected.
Petr Kalvoda, Jiri Vondrak, Miroslava Sucha, and Radovan Machotka are concerned
about “Subpixel Measurement in Close-Range Photogrammetry”, [57]. At the
beginning, a theoretical background on image processing and identification techniques
is presented. Matlab with its Image Processing Toolbox is used for the needed
operations in close-range Photogrammetry. Tests on some standard ideal cases were
carried out.
In [58], Mircea Balcu and Ştefan Marian Lazăr propose a “generalized Mohr-Coulomb
(GMC) Yield Criterion” and study the “implications on characterization of pavement
materials”. The requirements for modeling elasto-plastic material deformations and the
Novozhilov’s Approach for Mohr-Coulomb Yield Criterion are shown. Then, a
generalization of the Mohr-Coulomb Yield Criterion is developed. An application for
asphalt mixtures is producing different (better) results using the exposed procedure.
A interdisciplinary research is proposed by Alina Nicuţă and Constantin Ionescu, [59].
They study the costs that must be involved in order to achieve quality goals in Bridge
Engineering. The authors observe that quality costs are the costs associated with
preventing, finding, and correcting defective work. These costs can be significantly
reduced if careful design is performed. A general approach for evaluation of quality
costs is presented. Quality system for Bridge Engineering is explained and the main
factors influencing this system are analyzed.
An “investigation of mechanical properties of castable with respect to their application
as ceramic anchor elements” is developed by Lenka Nevrivova in [60]. The necessity
of the study is shown. Preparation of samples and the laboratory tests for analyzed
materials and elements are developed. The tests on samples burnt at different
temperature included: permanent length change of testing bodies burnt under high
temperatures; volume mass and apparent porosity; compressive strength; tensile
strength in bending; bearing capacity in heat and resistance against sudden temperature
changes. Extensive graphical show of results is presented and deeply commented.
Valentin-Vasile Ungureanu and Marius Comanic are dealing with the problem of
stability analysis of continuous welded rail, [61]. They observe that new European
regulations referring to the mentioned analysis must be implemented in Romania.
Therefore they develop new software and compare it to existing known software. The
authors conduct many parametric and comparative studies that certify good agreements
in results.
A “parametrical study of the effect of the torsional resistance of the fastenings on the
stability of continuous welded rail “ is performed by Valentin-Vasile Ungureanu,
Adam Dósa in [62]. The way to experimentally determine the torsional resistance of
fastenings is shown together with related characteristics diagrams. Linear and non-
linear FEM models were used for comparisons and for showing the influence of the
torsional resistance on the rail stability in the presence or in the absence of moving
loads. The influence is experimentally and analytically proved to be relevant.
In their paper, [63], “Concrete-filled steel tubular columns. Shear connection design”,
Călin Moga, Ştefan I. Guţiu and Gabriel Urian propose firstly a way to calculate the
shear stress appearing between the filling concrete and the tubular columns. The
number of needed mechanical shear connectors between the two materials is also
determined. Then a numerical application is done and results are commented.
Petru Moga, Ştefan I. Guţiu and Cătălin Moga follow a similar approach with that
shown above but, this time, referring to circular columns. They analyze the shear
transfer mechanism specific to this type of composite (steel and concrete) sections of
columns and also they determine the needed number of connectors between the two
materials. A numerical example is illustrating the procedure.
A study on “the meshing influence upon the design model of reinforced concrete
chimneys structures” is carried out by Ioana Entuc, Mihai Budescu, Ioan Ciongradi and
Gabriel Oprisan, [65]. The authors consider the case of a chimney located in Portugal.
Two main finite element models (with beam elements and with solid elements) are
used for analytical study. Results were compared with the experimental ones. For the
analytical research, the number of elements has been modified in order to notice the
differences in the dynamic characteristics. Also plate elements have been used.
Another parameter taken into account was the general shape of the chimney (conical or
cylindrical). Plate elements were the worst in use while the brick elements were
appropriate for use. Also, beam elements with lengths between one fiftieth and one
tenth from the total height of the chimney were found proper to use.
Dan Diaconu-Şotropa and Ioan Păvăloi present in their paper [65] the way of technical
accomplishment of objective meant to analyze stationary heat transfer at discretizeable
two-sized structures with quadrilateral type finite elements. They focused on the study
of the domain of structures deformation capability, conductive heat transfer and
thermal deformation capacity using finite elements method (MEF). The main purpose
is to develop numerical experiment with applications in extreme temperatures area.
The long-term monitoring of records from long-term cooling towers construction is
considered by Jiří Bydžovský, Amos Dofka and Zdeněk Šnirch in [67] as necessary to
define time of service life and dependability of cooling tower as reinforced
construction. More exactly, the paper deal’s with the influence of repair action on
service life and degradation processes at construction.
In their paper [68] P.Tragakis et all take into consideration the similarities between
architectural époques or structural styles, especially because of the imposed
gravitational rules, by different Design Codes, with similar content. In particular they
refer to the buildings with masonry structures considered in the idea to optimize the
possible retrofitting solutions. In order to approximate assessment of the response of
unreinforced masonry buildings to seismic loads were used incremental-iterative
procedures. The authors propose a constitutive model for the resisting elements able to
account for the variation in lateral stiffness induced by cracking, and their sort of
failure.
A research on the “pounding mitigation techniques” [69] was developed by
V.Warnotte et. all. The research is on the impact loads introduced by pounding that
have to be superimposed on those caused by the ground acceleration itself. In order to
avoid pounding induced collapse of buildings have been introduced several methods.
Dragos Stan Caciula and Nicolae Ungureanu present in their paper [70] “the
calculation of a cable network using the tangent stiffness matrix”. Their research is
based on the finite element method which can analyze all sorts of structure elements.
More particular the paper follows the method and the equation for the finite nonlinear
elements of a cable network.
In his paper [71] Constantin Ionescu makes an incursion through the “monumental
constructions” of town of Iasi. The main sightseeing is referred from an historical point
of view together with connected images.
The problem of large sized floors is being discussed by Ludovic Kopenetz et. all in
their paper [72]. They consider that it needs a special attention due to the deformation
of the structural concrete and the function of the building. The paper presents aspects
of the structural analysis that the structural designers have to take into consideration.
Similar with the previous presented paper the authors, Alexandru Cătărig, Ludovic
Kopenetz, Titu Hodişan make an analysis of the sandwich type composites in [73].
They consider that the structural analysis must take into consideration the general and
local phenomena occurring both in the elements and at the level of the whole, in order
to obtain the standard structural security.
In their research [74] Jiří Brožovský et. all have monitored the concrete resistance to
sulphates and to acidic aggressive media. The methodology used for the research in
order to investigate concrete resistance to corrosion, was the preparation of test
specimens conformable to the ČSN 73 1340.
“A comparative analysis of dynamic behavior and design of multistory structures” [75]
was made by Mihai Coveianu et. all. The paper compared different calculation methods
for the linear dynamic analysis of high multistory steel structures. For exemplification
was used a structure which is up to be build in Victoria Place, Bucharest.
Oana Mihaela Stan with the help of LUSAS (FEM) software program made a research
on “wind dynamic response of a suspension pedestrian bridge” [76]. In order to obtain
the dynamic response of the structure were performed two analyses: linear analysis and
a geometric nonlinear analysis.
In his article Anghel Ion deals with “the modeling, simulation and numerical analysis
of a fire in a construction with light structure” [77]. The numerical simulation was
created by using the high level software Fire Dynamics Simulator (FDS). The results
obtained were concretized by temperature measurements in different points on a
burning structure.
There was made an analysis of the seismic response of reinforced concrete building
using displacement based design model [78] by Laura Fraifeld, Ramona Agachi et all
by considering a Romanian residential building as a case study. The paper presents the
comparison between two methods of analysis of the shear wall nonlinear behavior.
There were used two software programs ADINA and ATENA from which use were
drawn several conclusions.
3. CONCLUSIONS
The Computational Civil Engineering symposiums held every year since 2003 in Iasi,
Romania, show that civil engineers are united through their need for computation and
also by the need to exchange their findings.
The progress in Civil Engineering is strongly connected to the implementation of
advanced computational means and to the fast spread of the knowledge to the
education/research/production sites.
Computational Civil Engineering symposiums series are becoming, no doubt, a vector
of progress especially in the Eastern part of Europe with nice perspectives for growing.
References
1. Ionescu, C., Păuleţ-Crăiniceanu, F., Scinteie, R., Nicuta, A. (editors), Civil Engineering,
Proceedings of the International Symposium “Computational Civil Engineering 2007”, Editura
Societăţii Academice "Matei-Teiu Botez", Iaşi, 2007, ISBN 978-973-8955-14.
2. Apeltauer Tomas, Petr Holcner, Martin Kysely and Jiri Macur: Non-linear effects in traffic
simulations, idem, pp.25-38;
3. Bondars Kaspars, Aleksandrs Korjakins: Influence of deformation at a heritage building support
on stability of groined masonry arch, idem, pp.39-54;
4. Boroń Jacek: Application of MS Excel® Solver in Solving Structural and Management Problems,
idem, pp.55-65;
5. Boštík Jiří: Underground structure in discontinuous rock mass, idem, pp.66-74;
6. Dósa Adam, Valentin-Vasile Ungureanu: Discrete model for the stability of continuous welded
rail, idem, pp.75-84;
7. Ene Petre: Soil-Structure Interaction Effects of Damped Structures, idem, pp.85-94;
8. Gherman Liviu, Petru Mihai, Nicolae Florea, Constantin Gavriloaia, Ioan Paul Vodă: Numerical
Modeling of Fiber Reinforced Concrete (FRC). Fiber orientation angle α = 00, fiber length lf =
50mm, idem, pp.95-102;
9. Gilewski Wojciech: A new look into finite element templates, idem, pp.103-112;
10. Iancovici Mihail, Georgiana Ionică: Evaluation of the inelastic demand of structures subjected to
multiple ground motions, idem, pp.113-123;
11. Irimia Roxana-Adina, Vlăsceanu Alina Nicoleta: Databases-business support for the Civil
Engineering, idem, pp.124-133;
12. Boţi Nicolae, Irina Lungu, Ioan Boţi: Foundation solutions for a marketing center in Cluj, idem,
pp.134-139;
13. Kopenetz Ludovic G., Ferdinand-Zsongor Gobesz: The Expertise of Offshore Structures, idem,
pp.140-147;
14. Kotek Petr, Jordán Filip: Scenario and impact of global warming in building energy performance
simulation, idem, pp.148-158;
15. Lungu Irina, Nicolae Boţi, Anghel Stanciu: Settlement prediction of a pile group based on design
charts, idem, pp.159-164;
16. Mesároš Peter, František Mesároš: The importance of controlling subsystem in information
systems for building industry, idem, pp.165-171;
17. Musil Roman: Analysis of Water Consumption in Residential Building, idem, pp.172-179;
18. Radinschi Irina, Brindusa Ciobanu: Improving Civil Engineering Physics Teaching-Learning with
Mathematica 5.1, idem, pp.180-190;
19. Radinschi Irina, Brindusa Ciobanu, Mircea Daniel Frunza: Physics with Maple 9.5, idem, pp.191-
200;
20. Radinschi Irina, Brindusa Ciobanu: Physics Studies – Computational Methods, a Strong
Connection, idem, pp.201-210;
21. Radinschi Irina, Brindusa Ciobanu: Computer Algebra System for Energy-Momentum
Localization, idem, pp.211-219;
22. Siwiec Przemysław, Sebastian Toś: The state of stresses in RC beams with installation holes
located in compressed zone, idem, pp.220-230;
23. Slonovschi Andrei, Ion Antonescu, Liviu Prună: Computer Graphical Communication an
Efficient Tool for Understanding the Design of Wooden Trusses, idem, pp.231-245;
24. Smutny Jaroslav, Lubos Pazdera: Analysis of dynamic parameters of rail fastening by Rihaczek
transformation, idem, pp.246-255;
25. Smutny Jaroslav, Lubos Pazdera: The use of time frequency transformations in testing structural
elements, idem, pp.256-268;
26. Stanciu Anghel, Oana Colţ, Irina Lungu: Design of reinforced soil works – Textomur structures –
based on the computer program Cartage, idem, pp.269-276;
27. Voiculescu Dragos, Daniela Preda: Computer program for Pipe Section Column Sizing, idem,
pp.277-281;
28. Cornelia Victoria Anghel: Computational method and application using Newton Algorithm, idem,
pp.282-286.
29. Libuše Beckerová, Gergely Bölckei and Jiří Brožovský: Partial substitution of Cement in
Concrete by Finely Ground Brick Body, idem, pp.287-294.
30. Constantin Amariei, Iulian Gabriel Mihai: Minimum weight buildings design using inequalities
method, idem, pp.295-306.
31. Brindusa Ciobanu, Ion Silisteanu, Irina Radinschi: Scientific Perspectives for Future Research
Work in Fundamental Properties of Short-Lived Radionuclides from Decay and from in-Beam
Studies, idem, pp.307-315.
32. Gabriel Oprisan, Nicolae Taranu, Vlad Munteanu: Experimental and numerical analysis of
compressed concrete elements confined with FRP composites, idem, pp.316-325.
33. Florin Ţepeş Onea: Rayleigh models coefficients, idem, pp.326-333.
34. Magda Brosteanu, Teodor Brosteanu: NDT Devices as Long-Term Vision in Assessing of Old
Masonry Buildings by Probabilistic Approach, idem, pp.334-345.
35. Magda Brosteanu, Teodor Brosteanu: IT Approach in Heat Transmission Visualizing through
Opaque Walls, idem, pp.346-353.
36. Adrian Prodescu, Daniel Bîtcă: Comparative Study on the Results of Analytical and Experimental
Analysis of a Steel Taintor (Radial) Gate, idem, pp.354-361.
37. Jiří Brožovský, Alois Materna and Ivan Kološ: Contribution to non-linear constitutive modelling
of masonry structures in 2D, idem, pp.362-367.
38. David Horak, Martin Zlamal, Petr Danek: Behavior of newly developed FRP reinforcement in
structures under various load schemes, idem, pp.368-378.
39. Josef Fink, Lubomir Ondris: New Shear Connectors for Composite Girders – Experiences with
ABAQUS Push-Out Test Simulations, idem, pp.379-395.
40. Ludovic G. Kopenetz, Ferdinand-Zsongor Gobesz: The Structural Expertise of Steel Cables,
idem, pp.396-409.
41. Mária Kozlovská, Michal Danko: Computer aided building diary, idem, pp.410-424.
42. Ioan Moga, Ligia Moga: Determination of the heat flows, of the solar and luminous
characteristics of the glazing and of the devices for the solar protection. Automaton calculus
program “SOLAR”, idem, pp.425-436.
43. Radomir Sokolar: Effect of moulding humidity on the properties of dry pressed ceramic tiles,
idem, pp.437-441.
44. Jerzy Szołomicki: Application of damage model for analysis of masonry structures, idem, pp.442-
447.
45. Constantin Ionescu: A New Form of the Active Moments Method, idem, pp.448-464.
46. Elena Carmen Teleman, Elena Axinte: Laboratory analysis in B.L.W.T. – priorities in insuring a
safety curtain walling design, idem, pp.465-476.
47. Grzegorz Dmochowski, Przemysław Siwiec and Piotr Berkowski: FEM 3D analysis of RC frame
foundations of rotary cement kiln, idem, pp.477-486.
48. L. De Doncker, P. Troch, R. Verhoeven, K. Buis, P. Meire: Flood routing using ‘Femme’, idem,
pp.487-497.
49. Przemysław Siwiec, Sebastian Toś: Simulation of cooling tower collapse on the basis of non-
linear concrete model and FEM analysis, idem, pp.498-507.
50. Irene Daprà, Giambattista Scarpi: Unsteady axial Poiseuille flow of a Bingham fluid in an
annulus, idem, pp.508-520.
51. Tomáš Fojtík, Jiří Brožovský: Formation of lime efflorescence on concrete and concrete products,
idem, pp.521-529.
52. Monica Gheorghiu and Mirela Chelcea: The use of computer means in the representation of the
ruled surfaces, idem, pp.530-543.
53. Horaţiu Mociran, Eugen Panţel: Numerical studies on the seismic performance of three structural
systems, idem, pp.544-553.
54. Doina Stefan, Violeta-Elena Chitan: Identification Methods of the Nonlinear Systems Subjected
to Seismic Actions, idem, pp.554-561.
55. GRAITEC ADVANCE software: GRAITEC – CAD and Analysis Software for the Construction
Industry, idem, pp.562-567.
56. Jaroslav Smutny, Lubos Pazdera: The use of the bispectrum for analysis dynamic parameters of
rail fastening, idem, pp.568-575.
57. Petr Kalvoda, Jiri Vondrak, Miroslava Sucha, Radovan Machotka: Subpixel Measurement in
Close-Range Photogrammetry, idem, pp.576-584.
58. Mircea Balcu, Ştefan Marian Lazăr: The Generalised Mohr-Coulomb (GMC) Yield Criterion and
some implications on characterisation of pavement materials, idem, pp.585-591.
59. Alina Nicuţă, Constantin Ionescu: Quality costs in Bridge Engineering, idem, pp.592-599.
60. Lenka Nevrivova: Investigation of mechanical properties of castable with respect to their
application as ceramic anchor elements, idem, pp.600-607.
61. Valentin-Vasile Ungureanu, Marius Comanici: Sensitivity study of a model for the stability
analysis of continuous welded rail, idem, pp.608-615.
62. Valentin-Vasile Ungureanu, Adam Dósa: Parametrical study of the effect of the torsional
resistance of the fastenings on the stability of continuous welded rail, idem, pp.616-624.
63. Călin Moga, Ştefan I. Guţiu, Gabriel Urian: Concrete-filled steel tubular columns. Shear
connection design, idem, pp.625-630.
64. Petru Moga, Ştefan Guţiu, Cătălin Moga: Composite columns with circular section. Shear transfer
mechanism, idem, pp.631-638.
65. Ioana Entuc, Mihai Budescu, Ioan Ciongradi, Gabriel Oprisan: The meshing influence upon the
design model of reinforced concrete chimneys structures, idem, pp.639-646.
66. Dan Diaconu-Şotropa, Ioan Păvăloi: Temperature Distribution Computing System for 2D
domains, idem, pp.647-655.
67. Jiří Bydžovský, Amos Dofka, Zdeněk Šnirch: Possibility increase cooling tower service life in
terms of long-term monitoring, idem, pp.656-663.
68. P. Tragakis, D. Stoica, S. Majewski, M. Voiculescu: Some general considerations about the
behaviour and retrofitting solutions for the existing buildings with masonry structures, idem,
pp.664-680.
69. V. Warnotte, D. Stoica, S. Majewski, M. Voiculescu: State of the art in the pounding mitigation
techniques, idem, pp.681-696.
70. Dragos Stan Caciula and Nicolae Ungureanu: The calculation of a cable network using the
tangent stiffness matrix, idem, pp.697-709.
71. Constantin Ionescu: City of Iassy’s monumental constructions, idem, pp.710-733.
72. Ludovic Kopenetz, Alexandru Cătărig, Titu Hodişan: The design of large sized floors, idem,
pp.734-741.
73. Alexandru Cătărig, Ludovic Kopenetz, Titu Hodişan: Structural analysis of sandwich type
composites, idem, pp.742-749.
74. Jiří Brožovský, Tomáš Fojtík, Amos Dufka:Resistance of Concrete with Fine-Ground Granulated
Slag Admixture to Sulphates and Acidic Aggressive Media, idem, pp.750-759.
75. Mihai Coveianu, Olimpia Marchelov, Daniel Bîtcă: Comparative Analysis of Some Methods for
Dynamic Behavior and the Structural Design of Multistory Structures, idem, pp.760-768.
76. Oana Mihaela Stan: Wind Dynamic Response of a Suspension Pedestrian Bridge, idem, pp.769-
778.
77. Ion Anghel: The modeling, simulation and numerical analysis of a fire in a construction with light
structure, idem, pp.779-789.
78. Laura Fraifeld, Ramona Agachi, Horst Werkle, Franz Zahn, Gabriela Atanasiu: Capacity based
earthquake design of RC shear wall according to EC8, idem, pp.790-800.
Non-linear effects in traffic simulations

Tomas Apeltauer, Petr Holcner, Martin Kysely and Jiri Macur
Faculty of Civil Engineering, Brno University of Technology, Brno, CZ-602 00, Czech Republic
Summary
Growing requirements on transport of passengers and goods have a principle

impact on life and economy. It is not need not intricately needed to search accurate
formulation of economic consequences, on the contrary would be enough to quote,
for example, United States Secretary of Transportation Norman Mineta, according
to which local drivers lose in traffic jams almost 4 billions hours and 2 billions
gallons of fuel each year. Upward trend is perceptible especially on highways,
where some out of traffic jams affects two-thirds drivers every day, which is double
compared to situation twenty years ago.
In Los Angeles at that time transportation problem lasted 4.5 hours every day,
about two decade later in the same town transport breaks down for 7 hours every
day. It is evident it is not always possible extend capacity carriage according to
their users needs and is necessary to search alternative manners. One of these is
transition to description of traffic flow via nonlinear dynamics, which explains the
behavior of very complicated systems.
If we explain properties of the traffic flow, we can try to describe emergency of
traffic jams and search manners, how those situation precede at least in part.
KEYWORDS: non-linear dynamics; traffic flow; congestions; traffic modeling;

simulations
1. TRAFFIC CHAOS AS A TECHNICAL TERM
Nonlinear dynamics as a tool for understanding chaos emergency already gets

through wide spectra branch. Presently it is perceptible that we cannot apply the
clean deterministic look even on very simple system, whereas the meaning of
nonlinear system features grows with their complexity. One of such dynamic
systems is the traffic flow. A group of German physicists headed by Dirk Helbing,
Boris Kerner and Michael Schreckenberg published in top class magazines like
Physical Review Letters, Journal of Physics and Nature results of computer
simulations of the vehicle movement. They noticed following interesting matter:
while using equations for a gas molecule movement, supplemented by some
specific characteristics of the human driver (among others the effort to avoid
26 T. Apeltauer, P. Holcner, M. Kysely, J. Macur
collision), they discovered special effects which truly copies real characteristics of
traffic. For example in places of bottlenecks traffic started compression, whereas
this perturbation expands as a density wave against the direction of traffic flow. It
is the exact analogy of known retardation and accumulation out in fronts before
bottlenecks – as soon as vehicles near bottleneck slow down, other vehicles will
slow down too, which causes the wave of “stop-and-go” spreading upstream.
At all the biggest surprise was the following result: the traffic jam can emerge in
certain circumstances spontaneously, without bottlenecks, traffic accident or other
at first sight obvious causes. Traffic flow is able to flow freely, and nevertheless
suddenly will change to the slowly ridden flow. Under certain condition the
“negligible” fluctuation in the vehicle velocity or headway can lead to the system
collapse, which last for long hours after primary impulse.
In spite of it is relatively cheerless inquest, it is absolutely in agreement with
results of mathematical models of many physical and biological systems. All these
studied systems have the common emergency of phenomenon known from popular
literature as the “chaos”. Simply: in every complex system with many parts which
influences each other a weak fluctuations can lead to big consequences. But we
cannot predict the emergency of such turnover. Scientists these phenomena
describe like non-linear – seemingly unimportant change in one characteristic
might have unreasonably extensive incidence on entire system. Non-linear
characteristics were detected in weather, biological systems or chemistry. And
during recent decades it is also mentioned in transport.
2. REAL TRAFFIC DYNAMICS OBSERVING
Every correct physical theory requires agreement with experimental data. Good
theory had to reproduce known system behavior as also predict its future. It will be
shown further that selection of the right method for data collection on road is a
very idyllic business with comparison to the construction of the right theory.
2.1. Measuring of the traffic variables
Probably the oldest and at the same time most complicated technology of traffic
detection is aerial photography, by the help of whose we can obtain total view of
the situation on tracked section. If we need to obtain concrete traffic variable such
as traffic flow velocity, we have to use digital picture analysis. Problems of visual
analysis generally belongs to the most complicated tasks, it is very time consuming
and not balanced with corresponding result accuracy. Its usage in current traffic
research is therefore rather border. But some advantage aerial photograph has
further: emergency of traffic jam and its progress is apparent at the first sight.
Figure 1. The traffic jam behavior, derived from aerial photography [17].
The second way can be taken with floating vehicle, which works as a movable
detector. But the biggest data sets are now collected by detectors placed on
concrete fixed place on the motorway. Most widespread is the induction loop
placed closely below the road surface. Incoming vehicle turns its inductance and
creates a weak signal.
From a simple induction loop we can obtain especially information about vehicles
that passed tracked place during definite time, which matches the definition of
traffic intensity. Today double induction loops ale placed in the most cases, which
make possible to measure in addition vehicle lengths and its velocity. Combination
of these values finally makes it possible to determine traffic density like second
crucial traffic variable. Correlation between traffic intensity and its density is a
very basic tool in the traffic research.
2.2. Measurement with induction loops
We can obtain a good notion about possibilities of detection via induction loops
from A5-Nord highway nearby German Frankfurt.
Figure 2. Location of induction loops on the part of A5-Nord highway [5].
Described section of highway has on the whole four intersections which join it on
next communications. On that section there are deployed about thirty double
induction loops with labels from D1 to D30. Each set of loops includes three
detectors for every lane, special cases are slip and trunk roads.
Figure 3. Data collected on A5-Nord highway. We can see one minute averages of traffic
densities and intensities from lanes on the place od detector D6. Measurement was on 9th
October 1992 between 7.00 – 12.00 [5].
We can see results on selected detectors on fig. 3. Three graphs a) to c) shows the
dependence of traffic flow intensity on traffic density for every of the three lanes.
How we will see further, in all cases this values correspond to free traffic flow and
at first sight there are perceptible expressive differences in maximal measured
values. Here we have to remark that the on German motorways there is a big
asymmetry in vehicle types among lanes. Trucks here usually do not move in the
left lane, only use the right one in some cases the middle one for overtaking.
2.3. Free flow and congestions
As mentioned above, there are three basic variables which characterize traffic flow
in specific time and place: traffic density (number of vehicles in one kilometer),
velocity (in most cases km.h-1) and traffic intensity (number of vehicles which
passed specific point in some period of time, one hour generally). It holds in steady
traffic flow the traffic intensity is a product of its velocity and density.
Unfortunately both these variables are not mutually independent and that is why it
is necessary to determine its correlation.
We can start from obvious and empirically tested facts: maximal velocity in traffic
flow corresponds to the density near zero, when each one vehicle in not influenced
by another one. Then we denote that case as free traffic flow. From this we can
define congestion as a state supplemental to the free flow, which contains some
other traffic situations. One of these states is the critical value of density, when the
speed of traffic flow drops to zero and the flow moves via small jumps or it stops
definitely (headways are too small for safe and useful constant speed). We can see
that between the free flow and this critical value the velocity must go down
(usually supposed monotonously and continuously). Determination of this
dependency is very basic and also very complicated task during investigation of
traffic flow.
Figure 4. Basic and very simple traffic model according to Greenshields, which comes
from strictly linear velocity-density dependency. Maximal traffic intensity is reached in
middle velocity for this model.
2.3. Fundamental diagrams
The basic instrument for description of steady flow and its global properties is the
fundamental diagram. Its most common type describes dependency of the traffic
intensity on the traffic density. In is not obvious from diagram itself that the key
correlation between density and velocity is uncertain and efforts to determine it
precisely were not successful yet. Nevertheless we can some derive some basic
characteristic also from empirically gained average values of traffic quantities. We
can use for demonstration data from A5-Nord again (fig. 5).
Contrary from previous entry are now showed all measured values which clearly
apportioned into two independent groups. Left part graph appertain to already
discussed free flow, the right pat to congestions. The waveform corresponds to
reality, when the traffic intensity is able to growth almost linear as far as to the
certain critical value. Then be enough so ever minor perturbation to switch free
stream into congestion. Traffic density nevertheless doesn't need to necessarily
growth and may even temporarily tail off. We can see again the phenomenon of
sudden emergency of traffic jams at the same time key feature of complicated
systems.
Figure 5. Fundamental diagram of traffic flow, gained from one minute averages traffic
density and intensity. Data was collected on A5-Nord on 23. March 2001 between 9.00 -
15.00 via detector D20, it is an average from all three lanes [5].
3. TRAFFIC FLOW SIMULATIONS
It is possible to see dynamic properties of traffic also through the numerical

simulation. We have to create a model, which truly describes vehicle behavior and
at the same time won't contain unbearable quantity of parameters. As well it is
necessary to choose a good simulation method which generates results in a good
time. We will present methods at first.
Till this time plentifully used access are cellular automata. Its principle is in
dividing the motorway to the finite count of blocks and each block in each
computation step contains or does not contain just one vehicle. That is why are
cellular automata extreme discreet and then only crudest approximation of the real
situation which is balanced by high simulation speed and low hardware exigencies.
Analog models as an opposite access have slow entrance only in last years. In
contrast to cellular automata it works with vehicles, which are equipped with
moving equations. Evolution of this model looks as moving of bubbles in narrow
pipe for example. The model behavior is more realistic but also much more time-
consuming for computational time.
Cellular automata were used for investigation of traffic flow at the beginning of
wider focus to this problem. The main reason was relatively low level of computers
and quite good results of this model. Especially the first reason is not actual in last
years.
Together with choosing of good method we have to find a good model. This
situation is much more transparent, because two main groups of models have
difference in the basic principle.
We can consider the traffic flow as a complicated system with separated vehicles.
Each vehicle is defined by its features and the main one is dependency between
acceleration and the situation around the vehicle. Some used algorithms try to
Describe observed properties of a real traffic flow (dynamics, physiological and
psychological ability of a driver). These models are known as microscopic.
Macroscopic access assume with traffic flow as a continuum defined by the
common presumptions. These can be derived by inference from border conditions,
measured values in real flow may be based on analogies with physical phenomena,
for example from fluxion liquids or gas.
It is often used the third combine approach, when the characteristics macroscopic
features are derived from integration of microscopic transport models and
presumptions about fluctuations.
The main advantage of the macroscopic access is the ability to describe evolution
of traffic flow by one or more differential equations. Is however debatable, if is this
presumption of possibilities this unification lawful – analytical complication and
venture of this access appears to be also heavy cost behind aesthetics of classical
theory.
Users of microscopic models on the other hand can get round of the danger of
complicated equations. But it is not all, they early snarl up on crossroad of
necessary numeral definition of this model pivotal parameters. About the
uncertainty generally witnesses for example historical survey of 23 models defined
by “safe distance from previous vehicles in dependency of its speed” and
“responsing time” [1]. Response time is between 0 – 2 s, its arithmetic mean is 0.78
s. The maximal value of traffic intensity is 1050 – 4800 vehicles per hour, with the
realistic average 2050 vehicles per hour. The vehicle speed, at which is achieved
full intensity, is between 10.6 – 55.2 kmh-1, in some cases even growth to infinity
(virtually it is necessary to consider about technical or legislative limit). The
average (without infinite values) is 29.0 km.h-1. Arithmetic mean itself do not give
any help – rather only underscore frustrating motley of models, from which only
little is unusable and unfortunately till this time any model is not implicitly good.
Perhaps that is just it at projection and scheduling communication meanwhile
dominates experiential access, which does not consider behavior inside traffic flow.
3.1. Car following models
Vehicle-following models, exactly Car Following Models (CFM) are investigated

nearly for a half century. Determination of the dependency of vehicle acceleration
on traffic conditions is in this case fundamental – in the simplest case just on the
condition of the previous vehicle (the vehicle driving before). In spite of the fact
that this task may appear at the first sight trivial, we did not manage to find a
generally respected model which would reflect the reality and would be at the same
time sufficiently compact, transparent and especially reasoned until today.
Thus certain general demands were accepted which would meet this proposal
without the apparent internal connection of the model with reality. Among these
“outer” demands belong especially:
• Non-collision character of simulations performed in the whole spectrum of
possible parameters and initial conditions.
• Physically reasonable values of the vehicle velocities and accelerations in
course of simulation.
• Asymmetrical character of the model – acceleration different from
deceleration (usually a stronger deceleration is admissible, e.g. in case of a
threatening collision).
• Emergency of global conditions corresponding to the real observation – non
linear character of the model (waves “stop and go”, spontaneous emergency
of congestions in case of above critical densities, hysteresis of traffic flow
intensity in case of above critical and under critical density etc.).
One of the most often studied models in last time, which partly satisfies the
requirements mentioned above, is the so called Intelligent Driver Model (IDM)
which was used for experiments of simulation. This model may be considered to be
an interpolation of two members for acceleration and deceleration with different
shapes in both parts.
Figure 6. Vehicle velocity and acceleration according to IDM model. Start and stop before
an obstacle at the distance 3000 m were used. This kind of test is a good indicator of the
model relevance.
Before the simulation we can check for example the first outer demand via the
simple experiment, when we test the ability of vehicle to stop ahead the block (fig.
6).
3.2. Model simulations
For the simulation itself we have created a program environment in Java

programming language due its easy accessibility, simplicity and portability in
internet. Each vehicle is represented by an object with its own parameters and the
whole system of vehicles is represented by the so called binded list. In such way it
is possible to dynamically change number of vehicles in the system.
For a traffic flow simulation it is advantageous to use the cyclic border conditions
which are commonly used for a systems of interacting particles – in this case it
means to put vehicles on a circular road with sufficiently large radius. In our case
we have chosen the radius 1 km, the total road length is more than 6 km and so we
can suppose the system behavior is not rather deformed by correlation between the
first and the last vehicle.
3.2.1. System visualization
Direct system animation, when a movement of each vehicle is animated by an

object on the circular road, is suitable just for program tuning – it does not bring a
good idea about actions in the flow and more over it decelerates the calculation.
The representation of time dependent development of density field is more
interesting. We unwrap the road in abscissa, where we record each vehicle position
by one point. After the selected time interval we similarly indicate a further line of
the density field. After sufficiently long interval there arises a model giving a good
idea about the development of a vehicle density on the whole road (fig. 7).
Figure 7. Development of vehicle density. In the beginning vehicles do not move and are
homogeneously dislocated in the first half of the road length.
In the mentioned visualization of the density field development there is the distinct
quick fixation to the stable state with three synchronized congestions moving
upstream. The system moves to stable state with arbitrary initial vehicle
distribution. Total number of vehicles in the system is 200.
Figure 8. Screen shot from the animation of density distribution in stabilized condition. Full
ellipse represents the average density level, which remains constant. The first picture shows
the system state after one hour of simulated time, the second one shows the system after
seven hours.
Another type of visualization of system development could be the local density

animation (fig. 8). The calculation speed evidently suffers from the animation;
however it gives a good idea about mutual interaction of density fluctuations
during the stabilization of system and their motion in stabilized condition, which is
always upstream.
We can make an interesting conclusion from the simulation: In spite of the fact that
congestions in a flow can be unevenly distributed, they have the same form and
size which is obviously an expression of vehicles with identical properties.
3.2.2. Obtained results
The main question could be the shape of fundamental diagram. In simulations we

used mean values of traffic density and intensity gained from all vehicles on whole
road. The final shape of fundamental diagram is too “smooth” and does not show
real transient processes in the traffic flow (fig. 9).
Figure 9. Dependency of simulated traffic intensity on average traffic density.
But we do not have so complex data during the real physical measuring. That is
why the real fundamental diagram has so different shape. The main difference in
not only in clear reasons: physical measuring do not work with stable traffic flow
and vehicles do not have the identical properties. The basic reason is finite time
interval, when we measure traffic intensity and density. With the sufficiently long
time interval we can obtain similar results in modified simulation experiments,
which reflect physical measuring.
It is possible to choose reference point on the simulated road. Each vehicle in each
step tests, if it crossed over this point. In positive case the global counter
increments with one. By this way we can simply “measure” the traffic intensity in
the selected point for the selected time interval (corresponding number of
computational steps).
The same trivial situation is during measuring of corresponding traffic density.
This quantity is in real situation derived from measured vehicle speeds and time
gaps. Each simulated vehicle has the value of distance from previous vehicle for
computing of acceleration.
This value is then possible to use for calculation corresponding local mean density
for a period of measuring interval. Then we have to collect to the other counter
headways of vehicles, which goes through the reference point.
It is obvious that for traffic densities, when we have stable traffic flow without
congestions, we will obtain same values for fundamental diagram as in averaging
in the whole system. Totally different situation occurs when we have traffic flow
with congestions. Then some few measurements will not give same values, but set
of different values on discreet levels of traffic intensities. Dividing into groups is
caused by counting only integer over of the vehicle on the bounds of time interval.
By this way we will get fundamental diagram which looks more realistic despite of
same vehicle properties and stable flow (fig. 10).
Figure 10. During profile measuring the fundamental diagram looks more realistic as the
method of its construction.
From the figure it is apparent that in linear part of diagram is profile measuring
concurrent with measuring on the whole system. With emergency of congestions
data “blurred” and distances of these levels depends on the period of measuring.
We can anticipate that for sufficiently long measuring interval the measured values
will converge to values of the whole system.
3.2.3. Non-linear phenomena
The basic simulation experiment was performed on a system with a maximum

number of 400 vehicles, i.e. with density of 63.7 vehicles/km. The flow
stabilization was indicated by a stabilization of root-mean-square deviation of
distances from an average value. The stabilization was considered as a difference
between individual cycles (the passage of reference vehicle through the start)
smaller than 10-5. After each stabilization one vehicle was removed from the
system while maintaining the existing states of other vehicles. The gap thus
generated was quickly filled, though the existing imbalance was sufficient for the
system to jump over into near state with various numbers of congestions. For
comparison there an experiment was performed – after each change of the vehicle
number, a new initial situation was set (vehicles homogenously distributed along
the road without move).
Figure 11. Comparison of dependency of velocities congestion on average density at

various strategies of setting initial conditions. Bold line corresponds to the setting of the
homogenous traffic flow, thin line to the taking away of the vehicle from the flow without
any change of other vehicles states.
From fig. 11 we can see that the disturbance created by taking away the vehicle
will be sufficient for the system to overlap into the congestion sooner than it is
created by rounding up errors and discretization. In another words in the interval 48
– 53 vehicles/km at least two stable situations exist side by side – homogeneous

distribution of vehicles and balanced congestions, while the final situation depends
on the initial conditions. At densities about 45 vehicles per km there are stable
situations with various numbers of congestions. During the stabilization from a
homogeneous initial situation the system converges alternatively to five or six
congestions distributed along the road; with a strategy of taking away vehicles the
system stays in a stable state of four congestions. For a lower density the state with
six congestions become unstable and extinguishes. A detailed investigation of
stable areas of this model is difficult with regard to the fact that space of initial
conditions is created among others by the density function.
Together with the state overlapping, occurrence of higher number of stable
situations is a typical property of non-linearities.
4. CONCLUSIONS
The outlined microscopic models of traffic flow are only part of a wide spectrum of
further models being used. Many of them are not sufficiently theoretically
described because they are parts of commercial products, and therefore subject to
commercial confidence.
At present a coexistence of models is admitted with some of them describing traffic
flow better at higher densities and others at lower densities or other modes. Traffic
engineers still wait for a model which would not contain large number of
parameters with the necessity of calibration and which would at the same time
characterize various traffic modes. Nevertheless it is possible to make a series of
interesting conclusions even from the existing models which may be verified in
practice – this is what we tried to do in this paper.
References
1. Bernard, M., Notes Notes on the Design Concepts for Transport Infrastructure: Past and Future,
4th Swiss Transport Research Conference, Monte Verita, Ascona, 2004.
2. Transportation Research Board, Highway Capacity Manual 2000, National Academy of
Sciences, USA, 2000.
3. Lighthill, M. J., Whitham G. B., Proceedings of the Royal Society A 229, 1955
4. Helbing, D., Hennecke, A., Shvetsov, V., Treiber, M., Micro- and Macrosimulation of Freeway
Traffic, Physical Review Letters 82, 2000
5. Kerner, B. S., The Physics of Traffic, Springer, New York, 2004
6. Schreckenberg, M., Schadschneider, A., Cellular automaton models and traffic flow, Journal of
Physics A: Mathematical and General 26 L679-L683, 1993
7. Schreckenberg, M., Schadschneider, A., Car-oriented mean-field theory for traffic flow models,
Journal of Physics A: Mathematical and General 30 L69-L75, 1997
8. Helbing, D., Traffic and Related self-driven many particle systems, Reviews of Modern Physics
73 1067-1141, 2001
Influence of deformation at a heritage building support on

stability of groined masonry arch
Kaspars Bondars1, Aleksandrs Korjakins1
1Faculty of Building and Civil Engineering, Riga Technical University, Riga, Kalku str. 1 LV-1658,
Latvia
Summary
Masonry arch stability criteria definition has been developed by numerical

modeling of cracked arch shell structure of Riga cathedral. Settlement of supports
induced strains in masonry arch shells which lead to stress redistribution and
crack forming in masonry shell structures is serious problem for heritage
buildings. Deformation criteria should be developed for safety exploitation of
heritage masonry structures. The main task of the present research is - developing
deformation criteria for evaluating of masonry shell construction stress-strain
state. Three-dimensional scanning of arch structure by means of Leica 3D laser
scanners were used to determine geometric model of masonry arch shell.
Previously accomplished monitoring data was used and a new monitoring program
was developed to define unequal settlements of supports. The settlements
monitoring of Riga cathedral supports and numerical modeling to define settlement
inequality were used as initial information to compare computer modeled deformed
structure shape with existing cracked masonry structure. Geodesic monitoring by
optical tools and mechanical tensometers have been used for determination of a
crack expansion since 2005. Twenty one Optical Fiber Sensors of SOFO type are
used now to measure deformations to establish safe exploitation limits. Multilevel
analysis of masonry shell has been fulfilled by finite element software and applied
for developing safe exploitation criteria.
KEYWORDS: structural masonry, masonry fracture, groined arch, heritage
building safety, numerical modeling.
1. INTRODUCTION
Since past masonry structures were built by time-honored method of trial and error.
The one of the first written rules and building code was “de Architectura” in ten
volumes written by Marcus Vitruvius Pollio, roman architect approximately at 1st
century. Medieval time masonry unions use time approved methods of masonry
structure cross sections, shell geometry, proportions and forms. Masonry code in
medieval Europe was distributed by Christianity influence as sacral building forms,
proportions and architectural styles. An objective of investigation is deformed
40 K. Bondars, A. Korjakins
structure of Riga cathedral in Latvia capital city shown in figure 1. 13th century
built masonry Gothic arch shells which are built from solid clay bricks are one of
the first masonry shell covered structures in Latvia.
Figure 1. North front of Riga Cathedral
Cross arch and star arch shell where commonly used as type of heritage building
shell. Long building process and changes in finale view of cathedral was usual by-
effect in building process centuries ago. Riga cathedral had long erection process
with stops and design changes. Few steps of erection and enlargement by changing
city demands took place also. Fire took place in 1547 when tower and nearest roof
parts were lost. A cross and star type masonry arches were partly rebuilt after
Swedish / Russian war in 16th century. Medieval European Hansa union cities were
usually located near river deltas. The supporting piles and subsoil situation
therefore in Riga cathedral are survived today and found in not satisfied conditions.
Weak subsoil under building foundations is common situation in Old Riga city. For
lot of buildings wooden piled foundations are used to transfer loads to dense sandy
layer through loose sand located under footings. Wooden pile life time practically
is not limited under ground water minimum level but pile condition after eight
centuries of building life time were found in unconditional situation. Seismic risk
in Latvia region is low but support deformations proceeds in most of Old Riga
piled supported buildings.
Structural weakness or overloading, dynamic vibrations, settlement, and in-plane
and out-of-plane deformations can cause failure of masonry structures. To prevent
the accidental situation in heritage buildings safety criteria must be specified
determining deformation limit between masonry arch shell interacting parts.
Surveillance des Ouvrages par Fibres (SOFO) optical fiber sensors will be mounted
on arch shell cracks for deformation monitoring. Cathedral arch internal surface
laser scanning were made by tools of three-dimensional 3D Leica laser scanner for
geometrical description of structure. Whole building thee-dimensional model is
available at modern software computing. The plastic analyze approach is
developed to define safety exploitation of deformed structure. Proposed

methodology on existing heritage buildings reduce necessary time for first
approximation of deformed structure. Numerical modeling of masonry structures
by modern computing tools demand additional material parameters such as
elasticity, stiffness, compression, tensile and shear resistance of stone material and
lime mortar, friction angle and cracking energy. Cracked masonry shell part
interaction also must be specified by friction because of safe exploitation
possibility after crack forming. Thrust line in non-cracked arks show center of
compression force in cross section and for medieval buildings almost is located on
center of cross section predicted by geometric proportions carefully improved in
mason unions. Based on “middle third rule” later computing method also is in safe
exploitation zone. Uncracked masonry arch shells can be easily computed using
modern GEM software’s and plastic material approach. After support deformations
and crack forming in shell structures thrust line change position in cross section
and internal forces relocates significantly. Therefore the main task of present
research is safety exploitation criteria definition for deformed masonry arch shell
of Riga cathedral.
2. MASONRY COMPUTING EVOLUTION
The main problems in the structural analysis of historic buildings are following:
1. Lack of data about geometric dimensions,
2. Material properties of the inner parts those are not visible exteriorly of the
structural members that are huge in cross-sectional dimensions,
3. Difficulties in identifying the characteristics of construction materials,
4. Excessive cost of detailed laboratory analyses,
5. Variety of the data due to construction techniques and natural material
utilization,
6. Altering material properties even along the same structural member due to
long-lasting construction process,
7. Uncertainties in construction process and steps,
8. Indefiniteness of general stability and strength continuity due to the existing
damage on the structure,
9. Inability in the application of modern construction materials, structural
analysis and loading conditions.
However modern computing technique mainly based on FEM software gives
possibility to detail analyze of heritage buildings. The historical review is
necessary to understand the computing evolution and background of modern
computing technologies.
2.1. Geometrical approach
Time honored geometry of arch structures was approved by geometric methods till
17th century. Arch and cross arch is a statically indeterminate structure therefore
the first calculation methods were developed to simplify design methodology. In
17th century Hook’s ca. 1670 proposed hanging flexible line thrust line definition
method was brilliant idea for very first arch geometry approach. Adequately weight
distributed hanged line give ideal tension line which in inverse form gives ideal
compression line or trench line as brilliant solution of statically indeterminate
structure. That type of solution is useful for vertically loaded arch structure but
doesn’t include horizontal loading. In Riga cathedral three arch horizontal
interacting forms the horizontal loading eliminate this methodology usage for that
type of buildings. Developed by Moseley (1833) “new Theorem in Statics” where
the Principle of Least Pressure was proposed describes structures widely. Later
Moseley (1843) specialized this approach to arches. Methodology, proposed by
Villarceau (1853) reduces the statically indeterminate level of arch by inserting the
three hinges to statically determinate system, was mostly useful analytic method till
development of elastic analysis method. Towards the end of the 17th century, the
studies of Galileo on strength analysis brought the medieval structural theory to the
end. In those times, the scientific deal was mainly in the analysis of string-
resembling the arch form- under vertical forces. In 1826, Navier put forward that
the buildings should stand by means of calculations of the stresses on its structural
members rather than the application of some geometric rules.
2.2. Elastic analysis
In end of 19th century elastic analysis was developed. In fact, until ca. 1880,
engineers divided arches into "elastic", made of wood or wrought iron, and "rigid",
made of masonry. Poncelet (1852) was conscious of the problem and in his
historical review of arch theory suggested to apply the elastic theory to masonry
arches in order to obtain a unique solution. Already in the 1860=s some elastic
analysis of masonry arches were made. Winlker (1879) make a discussion of
elastic material approach to masonry structure. However, he added a discussion on
the "Störungen" that can affect the position of the line of thrust. Their main origins
were: the deformation of the centering during construction, the yield of the
buttresses under the thrust and the effect of changes of temperature. All these
perturbations would produce some cracking of the arch and Winkler was well
aware this would affect notably the position of the line of thrust, which could be
very different from the calculated. Winlker suggested that some means of
controlling the position of the line of thrust by inserting internal hinges during
construction. Castigliano (1879) applied his theory of elastic systems also for
masonry bridges. Elasticity theory gives admissible results till elasticity limit of
material. In heritage building supporting structures compressive stress level is not
high and elastic material behavior is applicable. In early stage of elastic analysis
develop for masonry structures engineers showed a certain resistance because of
material anisotropic and irregularity. Elastic material approach in masonry shell
systems uncertain results is giving therefore computing approach development
seams necessary. The historical development review of first computational
methods was done by Heyman [1,2].
2.3. Kinematic approach and arch collapse mechanism
Hinge forming by loading, thermal variations or geometric changes is most of arch

collapse reasons. Material plasticity in hinges in self-weight loaded masonry shell
system is main reason of safety limit exceeding initiated by support deformation.
Support deformation forms the hinges in arch masonry structure as interaction
contact by sliding surface. Safe exploitation criteria therefore must be defined. The
kinematic or 'mechanism' method was first introduced by Heyman [3]. Heyman
assumptions for arch safety analyze method were infinite compressive strength and
friction resistance and zero tensile strength. He pointed out that plastic limit
analysis can apply well to the case of masonry gravity structures, such as piers and
arch bridges. In his book, for the single span arch, Heyman assumed the arch will
collapse when four hinges form. Furthermore, he assumed that the arch has infinite
compressive strength and sliding failures cannot occur. The hinges can alternately
form at the intrados or extrados of the arch, and at failure, the thrust line must pass
through the hinges. Therefore, it is possible to determine the magnitude of the
applied load which will cause the arch to collapse. By Heyman, [4,5] the kinematic
approach takes into account a collapse mechanism activated by an adequate
number of plastic hinges. As main criteria of safe exploitation definition was
Heyman proposed methodology. When the thrust line in a cross section is adjacent
to the ring of the arch, a hinge is opened in that point. According to the upper
bound theorem from the theory of plasticity, the maximum load corresponding to
some collapse mechanism is greater or equal to the maximum load corresponding
to the real collapse mechanism. This theorem implies that when the thrust line is
adjacent to the ring of the arch in four points then the arch is not safe. Friction is
high enough between stones and sliding failure cannot occur. The masonry has an
infinite compressive strength. More complicate collapse mechanism has been
considered for twin-span models by Hughes [6]. The pioneer in using modern
computational methods to determine the collapse load of block structures was
Livesley [7], who attempted to solve the problem as a linear programming problem
using the lower bound formulation of limit analysis. In his paper Livesley showed
that the adoption of associative friction leads to an incorrect collapse mechanism.
More importantly it may also give an overestimate of the true collapse load. When
using linear programming to solve limit analysis problems, flow will always occur
normal to the specified surface (i.e. according to the so-called ‘normality rule’).
3. MASONRY MATERIAL APPROACH
In present research are proposed two step computational approaches. A complete

and detailed review of micro and macro material approach may be found in
Lourenco [8] for mechanical aspects at both scales. As first stage of presented
computational methodology is deformed structure modeling by elastic macro
modeling approach to understand whole building situation and decrease computing
amount. In second stage of analysis deformed structure situation must be
investigated defining cracked masonry part interaction. Therefore micro modeling
material approach is most convenient.
3.1. Macro modeling approach
Homogenizations of material included in most of building codes give simplify

rules for material property description and usage in calculations. In previously done
research other heritage brick building was investigated to find masonry material
elasticity properties and work out material definitions according Latvian building
code. The homogenization of material means generalization of stone and mortar
properties for elastic material property definition adopted for computational usage
in elastic stage. Hooke law defined in 1676 long lasting linear elastic material
approach was used for analytic calculations. As main parameter for elastic material
approach elastic modulus was defined. Elasticity dependence from brick and
masonry class defined from cube resistances of brick and mortar material shown in
figure 2. Not cracked cross groined arch loading capacity computational analyze
for 1912 year built building using defined lower elasticity modulus in previously
done investigations earlier.
Compression elasticity modulus E, [MPa] for solid clay brick

masonry with layer thickness 75 and 100 mm
1600 1455
Compression elasticity modulus
1364
1400
1182 1227
1200 1091 1136
1000 1000 1000 1000 M75
E, [MPa]
909 909 909

800 864 818 M100
818 818
773 773 M125
600 682 682
682 636 M150
636 545
400 591 591 M200
443
500 409
200 318
341 273
0 307 250 207
205 159
91
M200
143
M150
159 127 73
M100
64
M75
95
M50
55
M200
M25
M150
45
M10
M125
M4
Mortar class
M2
M100
M"0"
M75
Brick class
Figure 2. Latvian building code methodology defined mean elasticity modulus for elastic
masonry approach
To understand the heritage masonry there are 3 point brick bending tests done.
Shown in figure 3 bending test results highlight main problem of heritage masonry
material definition – load bearing resistance result dispersion induced by home
manufacturing process variability.
Figure 3. Three point brick bending test results
3 point bending test shows different tensile resistance by crash force. Test also
shows equal elasticity for all specimens and brittle material approach of stone
material. Structural analysis in elastic stage using FEM linear masonry approach is
not acceptable for close to compression resistance calculations. In that cause
plasticity theory must be used. Not only brick manufacturing significantly affect
whole masonry material properties but also various stone material usage, block size
and form variations, mortar component variations and joint thickness variations.
Our aim is to give possibility to understand and analyze masonry arch, in fact, any
combination of them, i.e., a masonry building. The 3D surface scanning with
20mm precision is so detailed that the internal structure can be described easily.
Behind plaster layers forming the regular forms and levels of masonry wall surface
a most irregular internal structure is found. Homogeneity, isotropy, uniform
mechanical properties, etc., all the common assumptions of modern conventional
structural analysis cannot be made in this case without violence to the most basic
common sense. Taken from Riga Technical University building material laboratory
test report Nr 65-2000, masonry compression resistance, 3 point bending test given
tensile resistance was used for definition of masonry property. Used in
computational software material behavior is approximate by material property
variations but can give very close deformed stage structural understanding. Used in
FEM software Staad Pro material properties are described as: Prime modulus of
elasticity (E0 = 3600 MPa), average modulus of elasticity (Emean = 818 MPa),
Yong’s modulus (G = 1200 MPa), poisson’s ratio (0.2), thermal linear expansion
coefficient (αt = 0.000005 1/degree), density (ρ = 18 kN/m3), brick compression
resistance (R1 = 4.5 MPa), mortar compression resistance (R2 = 2.5 MPa), masonry
compression resistance (R = 1.1 MPa), masonry centric tensile resistance (Rt = 0.05
MPa), masonry shear resistance for head joint (Rsq = 0.11 MPa), masonry tensile
resistance in bending for head joint (Rtb=0.08 MPa), resistance to main tensile
stresses (Rtw = 0.08 MPa), coefficient of creep effect (v = 2.2).
τ red ≤ Rsq + 0.8 ∗ µ ∗ σ 0
(1)
Defined in equation 1 by Latvian building code methodology limiting tangential
stresses, where µ = 0.7 friction coefficient by joint mortar; σ0 – compression
stresses average value from lightest loading with reduction coefficient 0.9.
Tangential limiting value in second stage of calculation is defined as crack forming
value for shear forces in macro modeling material approach. Macro modeling
approach is simplifying material description and may normally be used in analytic
FEM structural computing by Rikards [9]. The macro modeling of masonry as a
composite is latest developed material approach by Rots [10]. Lourenço and Rots
[11] macro-modeling technique, specifically formulated for the analysis of masonry
constructions, is based on lumping all inelastic phenomena to the joints by means
of a composite interface model. This model, stemming from plasticity,
comprehends three different failure mechanisms, namely, a straight tension cut-off
for mode I failure, the Coulomb friction model for mode II failure as well as an
elliptical cap for compression and combined shear-compression failure. As in the
previous case, Lourenço and Rots model requires the values of the initial axial and
shear stiffnesses KN and KT as input data.
3.2. Micro modeling approach
Micro modeling of masonry material is precise stone material interaction

description possibility but also significantly increase computing time and amount.
In early stage of micro modeling dry joint approach was used. In latest
developments the block and the mortar in the joints are represented by continuum
models, whilst the interface unit-mortar is represented by discontinuous elements.
The Young model, the Poisson coefficient and the inelastic properties of the units
and the mortar are taken into account. Micro modeling is widely investigated and
described by Lourenco and Rots [12] and find mortar material and stone material
plasticity, load bearing tensile and compression resistances depending from joint
position, friction angle a.c. Micro modeling material approach must be used in
second step of computational investigation to analyze support settlement deformed
structure situation. For those purposes additional masonry material properties must
by involved.
4. MONITORING DATA COLLECTION
After World War Two Riga cathedral was closed as church and reorganized to
concert hall. Soviet government put serious amount of funding to uplift and repair
the structure, installed new ventilation, heating and fire safety systems. Latest
funding in those systems renovation, crack monitoring and geological survey make
a possibility to install long term monitoring system and make the present research.
4.1. Cracked arch photo fixation
Since 1959 cracks has been survived in Riga cathedral. Photo fixation made by
heritage building protection institutions show lack of cracks in masonry shells. This
is useful information to understand the new crack forming and prolongation of
existing in long term monitoring.
4.2. Arch masonry building sequence and layer orientation fixation
From 1960 to 1961 each arch has been sketched and described by J. Stukmanis. In
very careful way each brick and block sizes, form, orientation and position in
masonry structure. Shown in figure 4 arch orientation, rib positions, brick sizes and
forms were fixed in his work. Also description of damages and cracks for each arch
was written totally on 68 pages and 30 sketches done. Collected in State Cultural
Monument Archive materials about structure investigations and reconstruction
process give we better understanding of existing situation.
Figure 4. Careful fixation of all cross arch shells was done by J. Stukmanis
Fixation of damages after World War Two is very useful information in nowadays
to understand changes since that time. Crack prolongation, widening and new
cracking fixation of masonry structure therefore can be done by present crack
monitoring. After crack fixation plastering of arch surface was accomplished and
cracks hided.
4.3. Cathedral scanning by 3D Leica laser scanners
Building scanning gives the possibility of virtual three dimensional presentations in

tourist web sites. This very detailed scanning by three dimensional Leica laser
scanner give precise geometric surface defining for computational analyze of
structure. Made by Kalinka and Reiniks, specialists from geodesic company Merko
laser scanning in 2006 used for FEM computational analyze. Precision of 20mm on
internal surfaces and less detailed for outside is today used for structures. Increase
out coming virtual model precision theoretically gives whole structure computing
possibility. Future development of computational hardware will provide us with
powerful method of masonry structure analyze.
Figure 5. From full 3D model made plan and arch marking in plan made by Stukmanis
Figure 5 shows building plan section of model highlights geometrical uniformity of

structure in comparing with Slavietis, Seglins and Drugis in 1959 made hand tool
measurements. Scanned Gothic arch geometry also is preferable in comparing with
Erdmanis 1963 proportionality findings shown in figure 6. Proportion findings are
simplifying method of geometric approach to use the hand calculations of thrust
line. Geometrical inequalities lead to serious difference between existing situation
and real situation in cross section.
Figure 6. Proportionality findings by Erdmanis in 1963
By means of modern computational FEM software there is a possibility to analyze

the whole building. We find the laser scanning as very fast geometric modeling for
FEM software.
4.4. Microclimate monitoring
Staff managed microclimate monitoring for internal temperature and humidity was
performed from January 2002 to June 2004. Temperature oscillation gives
significant influence on masonry deformations and crack movement. In developed
now SOFO type based monitoring temperature external and internal measure is
included.
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
2003.01.02
2003.01.16
2003.01.30
2003.02.13
2003.02.27
2003.03.13
2003.03.27
2003.04.10
2003.04.24
2003.05.08
2003.05.22
2003.06.05
2003.06.19
2003.07.03
2003.07.17
2003.07.31
2003.08.14
2003.08.28
2003.09.11
2003.09.25
2003.10.09
2003.10.23
2003.11.06
2003.11.20
2003.12.04
2003.12.18
Temp, A RH, A TempLim RHLim
Figure 7. 2003 year internal temperature and humidity monitoring by staff
At figure 7 shown staff made monitoring for temperature and humidity for year
2003. Line Temp,A shows temperature variation during 2003 year period. RH,A
line shows humidity variations along 2003 year period. Temp,Lim and RH,Lim
lines show lower limiting values favorable to fungous forming. Temperature
initiated deformations simulated by computer modeling show significant influence
on crack forming, prolongation and movements. Cyclic temperature loading in
eight hundred years of cathedral life time affect not only internal comfort but also
stress situation in cross arch system seriously changing thrust line position. Solar
radiation effect on cathedral external surface can be defined as load on FEM
elements also.
4.5. Support condition survey
Support condition and subsoil situation survey in last two years made by CM GIB
Geotechnical Company presented by Celmins and Markvarts show weak soil layer
presence under footings. Started in 2005 geotechnical survey is more detailed than
in former Soviet period ever done. Cross section of piled supported column footing
is shown in figure 8. Marked as 7’’’D low density sand layer around the wooden
pile footing is main Riga cathedral part unequal deformation reason.
Figure 8 Column footing section in building middle span
Also an unsuited condition wooden piling was found. Support deformations along
time counted as main reason of crack forming in masonry arch shell system. This is
object of present research to fix the safe exploitation limits. Extreme support
deformation of masonry shell system change force lines and part interacting
stresses in shell section. Cracking changed thrust line position in cross section is
safe situation definition criteria. Present deformed situation and LBS Konsultants
make support deformation simulation by Plaxis in 2005 show approximate unequal
support deformations in 16 cm by various parts of cathedral. Future support
deformations predict crack prolongation, plastic hinge forming and unequal support
deformation possibility.
To understand ground water flow and level changes Riga Department ordered
ground water control monitoring from Balt-Ost-Geo Company. The ground water
table monitoring during 2006 showed approximately 0.4m wooden pile coverage
and less. In situation when wooden pile caps are not covered with ground water
table wood structure degradation is possible. Solutions must be found to prevent air
exposure possibility.
4.6. SOFO deformation monitoring system development
Surveillance des Ouvrages par Fibres (SOFO) optical fiber sensors monitoring
system received from Smartec and would be mounted in few following months to
change mechanical tensometer crack monitoring to a long term. Optical sensor

mounting cracks to provide masonry part movement in five years is decided to
control safe exploitation. According with mechanical tensometer monitoring
reports presented from 2005 crack oscillation varies 1-2mm every year depending
on season. The main reason of optical tensometer usage is computational control of
measurement data, long lasting life period and minimized side factor’s influence on
measurement’s quality.
Figure 9 SOFO principles of optic tensometer
The SOFO measuring system gives precision of 0.02mm therefore defining high
level of monitoring. Used as reading unit Smartec Bee allows control of 24 optical
sensor units and communicates with registration PC trough telephone line.
Additional thermo sensors for external and internal temperature control are
included in measured program. Internal memory of reading unit gives possibility of
data downloading by reasonable schedule. Battery support of reading unit gives
possibility of non-stop monitoring. Support deformation caused crack widening is
the way of safe criteria exploitation definition.
5. STABILITY CALCULATION AND SAFETY CRITERIA

DEFINITION
As discussed before critical thrush line position must be found on deformed

structure of masonry arch shell structure. Laser three dimensional scanning
geometrical models were used to define geometric forms of Gothic arch system.
The latest FEM software usage in heritage building masonry calculations in two
stages is developed. In first stage shown in figure 10 linear elastic material
approach is used. Support deformation loading is applied by Staad Pro software. In

second stage of analyze deformed structure shape and plastic material approach is
used to define stability by thrust line position in section.
Figure 10 First stages analyzed with Staad Pro elastic material approach using
The second stage of deformed structure includes the fixed and measured by
monitoring support settlements, cracked part interaction and subsoil elasticity.
Plaxis computing tool is developed for taking into account masonry plasticity and
stiffness. Each arch part is stabile and safe till thrust line locate in curved shell
cross section. Defined by Heyman four hinge forming as safety criteria is also used
to find the support deformation limits by numerical computing.
6. CONCLUSIONS
The present research is a small part of huge monitoring and investigation amount
done in Riga cathedral. All previously collected investigations data and
calculations increase the cathedral structure detailed understanding. Our view about
masonry material computing technologies and possible material assumptions
received from available papers are used to prepare FEM model and computational
strategy choice. Laser scanning based surface geometric data transfer to FEM
computational software significantly reduce modeling time. Improving laser
scanned precision and computational devices will give the powerful tool for
structural analyze of heritage buildings. Detailed investigations of material

properties, exploitation conditions and deformed situation must be taken into
account. None methodology describing masonry material is possible to include the
side effects of building process. It is necessary to point out the side effect of
building process: significant dispersion of material properties; geometric and
material variations; shell structure deformed conditions and various material
interactions. Wind, solar radiation and temperature loading give no big influence
on computational results but can be important. Two step computational method
highlight deformed shape influence on thrust line position in cross section of arch
shell. The monitoring improvement by long term SOFO optical tensometer tool is
the way for detailed structure investigation and safe exploitation system establish
of Riga cathedral. Methodology developed with curved arch surfaces laser
scanning in reason to use data as geometrical model for FEM computing software
is our aim of future improvement.
Acknowledgements
Thanks to Riga cathedral staff for huge amount of former Soviet time investigation
and monitoring data, which is included in this paper. Involved investigator’s permit
give us big acknowledge to understand the cathedral building. Thanks to the Riga
cathedral representatives understanding in the latest technology usage on arch long
term monitoring advantage. Thanks a lot for complex monitoring development and
funding in heritage building to keep it in good health.
References
1. Heyman J., The Stone Skeleton: Structural Engineering of Masonry Architecture, Cambridge
University Press, Cambridge, 1990 pp 45.
2. Heyman J., The science of structural engineering, Imperial College Press, 1999.
3. Heyman, J. The equilibrium of shell structures. Oxford: Clarendon Press. 1997
4. Heyman, J., “The masonry arch”, Ellis Horwood Limited, UK. 1982.
5. Heyman J., ‘The safety of masonry arches’, Int. Journ. Mech. Sci., 11 (4), 1969, pp. 363-385.
6. Hughes, T.G., Analysis and assessment of twin-span masonry arch bridges, Proc. Instn. Civ.
Engrs., 110, 1995, 373-382.
7. Livesley, R. K. Limit analysis of structures formed from rigid blocks. Int. Journ. for Num. Meth.
in Eng. 12, 1978, p. 1853-1871.
8. Lourenço, P. B. Computational Strategies for Masonry Structures. PhD thesis, Delft University of
Technology, 1996
9. Rikards, R., and Cate A., Finite element method, Riga Technical University, Riga, pp 78, 2002
10. Rots, J.G. - Numerical simulation of cracking in structural masonry, Heron, 36(2), 1991, p. 49-63.
11. Lourenço, P.B., de Borst, R., Rots, J.G., A plane stress softening plasticity model for orthotropic
materials, Int. J. Numer. Meth. Engng., 40, 1997, p. 4033-4057
12. Lourenço, P.B., Rots, J.G., A multi-surface interface model for the analysis of masonry structures,
J. Engrg. Mech., ASCE, 123(7), 1997, p. 660-668
Application of MS Excel® Solver

in Solving Structural and Management Problems
Jacek Boroń
Building Engineering Institute, Wroclaw University of Technology, Wroclaw, 50-377, Poland
Summary
Solvers, or optimizers, we call software tools that help us as users to find the best
way to allocate scarce resources. The resources may be everything in limited
supply. The "best" or optimal solution may mean maximizing profits, minimizing
costs, or achieving the best possible quality. An almost infinite variety of problems
can be tackled this way, but here are some typical examples: finance and
investment, manufacturing, distribution and networks.
To use the solver you have to build a model that specifies: decision variables,
constraints and objective. The Solver will find values for the decision variables that
satisfy the constraints while optimizing (maximizing or minimizing) the objective.
Spreadsheets such as Microsoft® (MS) Excel provide a convenient way to build
such a model.
Anyone who has used a spreadsheet is already familiar with the process: Cells on
a worksheet can hold numbers, labels, or formulas that calculate new values, such
as the objective of an optimization. Constraints are simply limits (specified with
<=, = or >= relations) on formula cells. And the decision variables are simply
input cells containing numbers. Whether the Solver can find a globally optimal
solution, a locally optimal solution, or a good solution depends on the nature of the
mathematical relationship between the variables and the objective function and
constraints (and the solution algorithm used)[1].
In this paper it will be shown how we can use the Excel-Solver (the product
developed by Frontline Systems, Inc. for MS Excel) to find the best (optimal)
solutions in some structural problems, as well tasks connected with management of
an enterprise. The illustrative examples will be these, which I always use for my
didactics at Building Engineering Institute. Numerical results for two of sample
problems demonstrate how versatile (multipurpose) and useful tool was used.
KEYWORDS: excel-solver, modeling, optimizers, structural optimization

problems, managing optimization problems
56 J. Boroń
1. INTRODUCTION
The Solver being part of Microsoft Excel (developed by Frontline Systems for
Microsoft) is one of the widely used optimization tools. It is able to solve small-
scale linear programming (LP), smooth nonlinear programming (NLP), and mixed
integer programming (MIP) problems. An upgraded Premium Solver Platform
(PSP) is a tool that extends functionality and speed of the MSE Solver to handle
industrial-scale problems of over tens of thousands variables and constraints. PSP
is also challenging global optimization problems using multi-start or clustering
methods and non-smooth problems using methods based on genetic and
evolutionary algorithms (or tabu search).
Table 1. The Characteristic of the Enhanced Excel Solvers [1].

Remark for the Table 1: For integer problems B&B refers to branch and bound and
P&P refers to preprocessing and probing. For nonlinear problems “GRG” refers to
the generalized reduced gradient method and “SQP” refers to sequential quadratic
programming.
2. USING SPREADSHEETS
2.1. Spreadsheets vs. Individual Programming
Spreadsheets such as Microsoft Excel provide a convenient way to build such a

model. Anyone who has used a spreadsheet is already familiar with the process:
Cells on a worksheet can hold numbers, labels, or formulas that calculate new
values (such as the objective of an optimization).
Constraints are simply limits (specified with <=, = or >= relations) on formula
cells.
The decision variables are simply input cells containing numbers. Frontline's
Premium Solver products provide powerful tools for solving, or optimizing, such
model.
Another way to build an optimization model is to write code in a programming
language such as Visual Basic or C/C++. Instead of spreadsheet cells, variables or
arrays in the program hold the decision variables and calculate the constraints and
objective.
2.2. How do we define a model?
Every optimization model consists of:

-decision (design) variables and parameters,
-objective function,
-constraints.
2.2.1. Decision variables
They usually measure the amounts of resources, such as money, to be allocated to

some purpose, or the level of some activity, such as the number of products to be
manufactured, the number of pounds or gallons of a chemical to be blended, etc. In
engineering problems decision variables are dimension of the structures or cross-
sections.
58 J. Boroń
2.2.2. Objective function
After defining the decision variables, the next step is to define the objective, which
is normally some function that depends on the variables. We'd be finished at this
point, if the model did not require any constraints (but in engineering problems it is
rather impossible). For example, in a curve-fitting application, the objective is to
minimize the sum of squared differences between each actual data value, or
observation, and the corresponding predicted value. This sum has a minimum
value of zero, which occurs only when the actual and predicted values are all
identical. If we asked a solver to minimize this objective function, we would not
need any constraints.
In most models, however, constraints play a key role in determining what values
can be assumed by the decision variables, and what sort of objective value can be
attained.
2.2.3. Constraints
Constraints reflect real-world limits on production capacity, market demand,

available funds, and so on. To define a constraint, you first compute a value based
on the decision variables. Then you place a limit (<=, = or >=) on this computed
value.
General Constraints. For example, if A1:A5 contains the percentage of funds to
be invested in each of 5 stocks, you might use B1 to calculate =SUM(A1:A5), and
then define a constraint B1 = 1 to say that the percentages allocated must sum up to
100%.
Bounds on Variables. Of course, you can also place a limit directly on a decision
variable, such as A1 <= 100. Upper and lower bounds on the variables are
efficiently handled by most optimizers and are very useful in many problems.
Policy Constraints. Some constraints are determined by policies that you or your
organization may set. For example, in an investment portfolio optimization, you
might have a limit on the maximum percentage of funds to be invested in any one
stock, or one industry group.
Physical Constraints. Many constraints are determined by the physical nature of
the problem. For example, if your decision variables measure the number of
products of different types that you plan to manufacture, producing a negative
number of products would make no sense. This type of non-negativity constraint is
very common. Although it may be obvious to you, constraints such as A1 >= 0
must be stated explicitly, because the solver has no other way to know that
negative values are disallowed.
As another example of a physically determined constraint, suppose you are
modeling product shipments in and out of a warehouse over time. You'll probably
need a balance constraint, which specifies that, in each time period, the beginning
inventory plus the products received minus the products shipped out equals the
ending inventory (and hence the beginning inventory for the next period).
Integer Constraints. Advanced optimization software also allows you to specify
constraints that require decision variables to assume only integer (whole number)
values at the solution. If you are scheduling a fleet of trucks, for example, a
solution that called for a fraction of a truck to travel a certain route would not be
useful. Integer constraints normally can be applied only to decision variables, not
to quantities calculated from them. A particularly useful type of integer constraint
specifies that a variable must have an integer value with a lower bound of 0, and
upper bound of 1. This forces the variable to be either 0 or 1 (nothing in between)
at the final solution. Hence, it can be used to model "yes/no" decisions. For
example, you might use a 0-1 or binary integer variable to represent a decision on
whether to lease a new machine. Your model might then calculate a fixed lease
cost per month, but also a lower cost per item processed with the machine, if it is
used. A solver can help determine whether leasing the machine will yield higher or
lower profits.
2.2.4. Interpreting solutions
A solution (set of values for the decision variables) for which all of the constraints
in the Solver model are satisfied is called a feasible solution. Most solution
algorithms first try to find a feasible solution, and then try to improve it by finding
another feasible solution that increases the value of the objective function (when
maximizing, or decreases it when minimizing). An optimal solution is a feasible
solution where the objective function reaches a maximum (or minimum) value.
A globally optimal solution is one where there are no other feasible solutions with
better objective function values. A locally optimal solution is one where there are
no other feasible solutions "in the vicinity" with better objective function values
(we can picture this as a point at the top of a "peak" or at the bottom of a "valley"
which may be formed by the objective function and/or the constraints). The Solver
is designed to find optimal solutions (ideally the global optimum) but this is not
always possible. In many cases, though, we may be happy to find a good solution
(one that is better than the solution we are using now). Whether the Solver can find
a globally optimal solution, a locally optimal solution, or a good solution depends
on the nature of the mathematical relationship between the variables and the
objective function and constraints (and the solution algorithm used).
2.2.5. What makes a model hard to solve?
Solver models can be easy or hard to solve. "Hard" models may require a great deal
of CPU time and random-access memory (RAM) to solve (if they can be solved at
60 J. Boroń
all). The good news is that, with today's very fast PCs and advanced optimization
software from Frontline Systems, a very broad range of models can be solved.
Three major factors interact to determine how difficult it will be to find an optimal
solution to a solver model:
-the mathematical relationships between the objective and constraints, and the
decision variables
-the size of the model (number of decision variables and constraints) and its
sparsity
-the use of integer variables - memory and solution time may rise exponentially as
you add more integer variables
2.3. Using the Solver dialogs
Assuming that we have the standard Excel Solver or a Premium Solver product
installed, the next step is to create a worksheet where the formulas for the objective
function and the constraints are calculated. In the worksheet below, we have
reserved cells C15, D15, E15 and F15 to hold our decision variables x1, x2, x3 and
x4: amount of advertisments through TV, radio, direct mail and newspapers.
Fig.1. Example 1: Advertising
To add the constraints, we click on the Add button, select cells C16:F16 in the Cell
Reference edit box (the left hand side), and select cells C11:F11 in the Constraint
edit box (the right hand side); the default relation <= is OK (Fig.2.)
Fig.2. Adding the constraints in the edit box (polish version of MS Office)
We choose the Add button again (either from the dialog above, or from the main
Solver Parameters dialog) to define the non-negativity constraint on the decision
variables.
When we've completely entered the problem, the Solver Parameters dialog looks
like this (Fig.3.) This dialog appears for the Standard Solver Platform.
Fig.3. The Solver Parameters Dialog Box (polish version of MS Office)
2.4. Finding and using the solution
To find the solution, we should click on the Solve button. After a moment, the
Solver returns the optimal solution: $9000 in cell C15, $3000 in D15, $0 in E15
and $1875 in cell F15. This means that we should advertise 18 times in TV, 15
times in radio and newspaper, to make a total, minimal cost $13875 (shown at cell
G15). We shouldn’t invest in ads through direct mail. The message "Solver found a
solution" appears in the Solver Results dialog, as shown here (Fig.4.). We now
click on "Answer" in the Reports list box to produce an Answer Report, and click
OK to keep the optimal solution values in cells C15:F15.
62 J. Boroń
Fig.4. Solver results dialog box
After a moment, the Solver creates another worksheet containing an Answer

Report, like the one below, and inserts it to the left of the problem worksheet in the
Excel workbook. (Fig.5.)
Fig.5. Answer report worksheet
3. ILLUSTRATIVE EXAMPLES
3.1. Design (structural) problem – optimization of the welded cross-section
Our target is to optimize a double-T, bisymmetric cross-section of a welded, steel

beam, connected with a stiff plate made of armed concrete (Fig.6.). We allow only
one loading condition: bending moment M. We consider 4 design variables t, h, b

and g. Objective function Fc going to be minimized:
Fc = (0,125 · Re + s) · (h · t + 2 · b · g) (1)
The worksheet prepared to solve the problem is presented below (Fig. 6.).
STEEL BEAM OPTIMIZATION

Input data:
M= 5000 kNm (bending moment)
l0 = 4,000 m. (beam span)
steel class: St3S
Strenght constraints (simple bending, by Polish Standard PN-90/B-03200)
M <= MR * φL , where MR = αp * W z * f d ,
therefore W z >= Mz/(αp *φL* f d )= 0,022148
Considering conection of our beam with stiff plate gives: φL = 1 and αp = 1,05.
Design (geometric) constraints:
t >= 0,008
0< h <= 0,85
g>0
6*g<= b <= 20*g ◄prevents stretching out of the belts during welding
SOLVER
cost F c = 1663,082 61688,97
design b= 0,387 1,000 starting
variables g= 0,850 1,000 values
h= 0,064 1,000 ◄
t= 0,008 0,100
Re = 2,35E+05 2,25E+05 2,15E+05 kPa ◄plasticity border

4
J(b,g,h,t) = 0,011 m
3
W z(b,g,h,t) = 0,022 m
s= 0,7
fd = 215 205 195 MPa ◄strenght of the steel
Hp = 1,05
Bending moment M(x)
6000,0
M(x) values in kNm
5000,0
4000,0
Tutaj wstaw wykres momentów zginających belki
3000,0
2000,0
1000,0 M
0,0
0
4
8
2
6
0
4
8
2
6
0
0,
0,
0,
1,
1,
2,
2,
2,
3,
3,
4,
x coordinate (from 0,0 do 4,0) m
Fig.6. Optimization of the welded steel beam cross-section (4 design variables)

64 J. Boroń
3.2. Management of an enterprise problem - purchasing & transportation
Problem. A concrete manufacturer has 4 different plants that all require a certain
amount of cement. There are 5 different companies (cement factories) where the
cement can be bought. Where should the concrete manufacturer buy the cement
and how much should it buy, to minimize its' cost and shipping?
Solution. The variables are the amounts of cement to be bought from each
company for each plant. On worksheet Purchasing & Transportation Model these
are given the name Amounts_of_cement_to_buy: cells C15:G18 (see Fig.7.). The
constraints are simple and straightforward:
Amounts_to_buy (C15:G18) >= 0 via the Assume Non-Negative option (2)
Total_amounts_to_buy (H15:H18) >= Demand (I15:I18) (3)
Total_sold(C19:G19) <= Supply(C20:G20) (4)
The objective is to minimize the cost. This is defined as Total_cost on the
worksheet (cell H24).
Fig.7. Purchasing & Transportation Model (concrete manufacturing)
Even though this model is very simple, it is one of the most used models in the
industry. It routinely saves many companies thousands or even millions a year.
4. CONCLUSIONS
The optimizer developed by Frontline Systems, Inc. under simple name “Solver” is
multipurpose and friendly-in-use tool also for engineering and management
problems. Therefore I decided to include it successfully in my didactic program
beginning from the third semester at Faculty of Civil Engineering.
However in real world optimization problems can be both simple and “cheap” or
complicated and “expensive”, so the types of mathematical relationships in a model
(for example, linear or nonlinear, and especially convex or non-convex) determine
how hard it is to solve, and the confidence we can have that the solution is truly
optimal.
These relationships also have a direct bearing on the maximum size of models that
can be realistically solved. A few advanced solvers can break down a problem into
linear, smooth nonlinear and non-smooth parts and apply the most appropriate
method to each part, but in general, we should try to keep the mathematical
relationships in a model as simple (i.e. close to linear) as possible [2].
References
1. Fylstra, D. et al., Design and Use of the Microsoft Excel Solver, Copyright 1998, Institute for
Operation Research and the Management Sciences, 0092-2102/98/2805/0029.
2. Nenov, I., Fylstra, D., Interval Methods for Accelerated Global Search in the Microsoft Excel
Solver, Reliable Computing 9: 143-159, 2003, Copyright 2003, Kluver Academic Publishers.
Underground structure in discontinuous rock mass

Jiří Boštík
Institute of Geotechnics, Brno University of Technology, Brno, 602 00, Czech republic
Summary
The results of parametric analysis on influence of discontinuities orientation on the

response of rock mass during driving an underground structure in the direction of
discontinuity planes are presented in the contribution. Three typical alternatives
were considered: the underground structure in rock mass with one-sided dip of
discontinuities, the underground structure situated in synclinal bend and the
underground structure situated in anticlinal bend. In all three cases, the variability
of discontinuities dip angle in the range of 0 - 90 o was considered.
Mathematical models were used for the analysis by Finite Element Method using
software Plaxis. Mathematical models were planar – plane strain. Rock Jointed
model was used to simulate mechanical behaviour of the rock environment. The
underground structure is driven in full -face with no lining.
All the analysed alternatives are compared to one another considering the vertical
displacement at the top of the underground structure and maximum vertical
surface displacement. From the results presented, it can be concluded that from the
point of view of vertical displacement at the top of the underground structure,
situating the underground structure in the axis of the synclinal bend, if the dip of
discontinuities is not very steep, is the least favourable. On the other hand, taking
into account the maximum vertical surface displacement, situating the
underground structure, in relation to the discontinuities orientation, seems to be of
no importance.
KEYWORDS: underground structure, discontinuities, mathematical simulation.
1. INTRODUCTION
The rock environment, in which the underground structures are build and which is
an inseparable part of them, is usually weakened by several discontinuities. These
discontinuities are created either during the rock formation itself (e.g. stratification
of sedimentary rocks) or in rock which have already been formed (e.g. joints, fault
planes).
In case of sedimentary rocks, the stratification, which is developed by lithological

changes of deposited material, is the most significant structural property.
Occurrence of the discontinuities in rock mass described above has an impact to its
behaviour during driving of underground structures. For instance, in [1] the authors
presented that in case of the stratified rocks their placing, thickness and dip of the
layers and strength of the rock need to be considered. The angle between tunnel
axis and layers direction, which are generally of a various aslope, contorted or
folded, is a very important feature as well.
Driving of a tunnel in the direction of layers [1] is less convenient, especially when
the rock is non-homogenous. The tunnel can be subjected to one-sided pressures
when the layers are skewed (Fig. 1a). When the tunnel is situated in the axis of
anticline (Fig. 1c), the dip of layers need to be taken into account. Such a situating
the tunnel is very inconvenient in case the layers are steeply skewed as the crown
of the tunnel was loaded by full weight of the loose layers. Situating the tunnel in
synclinal bend (Fig. 1b) is also considered to be very inconvenient – the rock
would pressure on the tunnel from either side and huge water inflow would occur if
permeable layers were presented. Situating the tunnel path in the direction of the
vertical or steeply skewed layers is very unsuitable due to large weight of
overburden affecting the crown of the tunnel.
a b c
Figure 1. Situating tunnel in stratified rock in the direction of the layers [1]
2. PARAMETRIC STUDY
In order to analyse an influence of discontinuity planes orientation on the response

of rock mass during driving an underground structure, the following model
example was considered. This was solved by Finite Element Method using
software Plaxis [2].
68 J. Boštík
2.1. Geometric arrangement
A linear underground structure (collector) of horse-shoe shape, 3.3 m wide, 4.05 m

high and sectional area of 12.2 m2 (Fig. 2) is considered. Top of the underground
structure is 13.2 m under the horizontal surface, i.e. in depth that equals four times
width of the underground structure.
Figure 2. Schematic arrangement of underground structure
2.2. Geology conditions
Rock environment is created by one kind of stratified rock. It is assumed, that

single layers are of the same width (D), which is relatively small as compared to
transverse dimensions of the underground structure (B), as schematically illustrated
in Fig. 3. Further, it is supposed the discontinuities are not filled.
Figure 3. Discontinuities – 1, Underground structure – 2
Mechanical behaviour of rock environment in the mathematic model is

approximated by the Jointed Rock model. This model is an anisotropic elastic
perfectly – plastic model, especially meant to simulate the behaviour of stratified
and jointed rock layers and its more detailed description can be found in [2].
Within the analysis, one set of discontinuities (stratification) is used and elastic
behaviour of rock environment is fully isotropic, i.e.
E 2 = E1 and ν 2 = ν 1 , (1)
1
G2 = E1 (1 + ν 1 ) , (2)
2
where E1 ,ν 1 stand for Young’s modulus and Poisson’s ratio of the rock as a
continuum respectively, E 2 ,ν 2 , G2 stand for Young’s modulus, Poisson’s ratio
and Shear modulus in the discontinuity plane direction respectively.
In the discontinuity plane direction it is assumed that shear stresses are limited
according to Coulomb’s criterion. Parameters of strength in the direction of
stratification are as follows: cohesion (c), friction angle (ϕ), dilatancy angle (ψ),
tensile strength (T - Strength).
Throughout the calculation, the material parameters from Table 1 were used.
Table 1. Material parameters used

γ E1 ν1 E2 ν2
[kN/m3] [kPa] [-] [kPa] [-]
23.7 6x104 0.33 6x104 0.33
G2 c ϕ ψ T-Strength
[kPa] [kPa] [o] [o] [kPa]
2.3x104 20 24 0 0
Three basic alternatives according to Fig. 4 are solved. In the first one (J), one-
sided dip of discontinuities is considered. In the second one (S), the underground
structure is driven in the symmetrical synclinal bend. In the last one (A) the
underground structure is situated in symmetrical anticlinal bend. In all the three
cases above, the various orientation of the discontinuities, determined by dip angle
α1, is involved. Value of α1 is chosen from the range of 15 - 75 o with 15 o step.
2.3. Technology
From the point of view of driving method, driving of the underground structure in
full - face with no use of lining is considered.
70 J. Boštík
Figure 4. Considered alternatives: J – one-sided dip of discontinuities, A – symmetrical

anticlinal bend, S – symmetrical synclinal bend
2.4. Mathematical model
The planar mathematical model (plane strain) is formed by Finite Element Method
using the software Plaxis [2]. The analysed area is of a rectangle shape of the size
of 35 x 27.7 m. The mesh with 1 336 elements, 2 753 nodes and 4 008 stress points
(Fig. 5) was created by discretisation using six-node triangular element.
Figure 5. Finite element discretisation
The top boundary of the mathematical model representing the surface is freely
sliding, while the other model boundaries are tied with standard boundary
conditions: the vertical boundaries restricted on slide in horizontal plane, the
bottom boundary restricted on slide in both horizontal and vertical directions.
Solving the mathematical models was performed in the two calculation phases
illustrated in Fig. 6:
1. Calculating the primary state of stress caused, in this case, only by self-
weight of rock mass γ. Value of horizontal and vertical effective stresses ratio
(coefficient of lateral earth pressure) K0 was 0.593.
2. Carrying out the excavation of the underground structure.
For all the three alternatives – J, S and A, five models with different discontinuities
orientation (see section 2.2) were created. Furthermore, cases where set of the
discontinuities is horizontally and vertically oriented were solved as well. In total,
seventeen models were created.
Figure 6. Calculation phases
3. RESULTS
Vertical displacement at the top of the underground structure (named as a point I

further in the text) and maximum vertical surface displacement were selected in
order to compare the results of the solved alternatives and their variations. Relation
between these values and discontinuities orientation (dip angle) for the analysed
cases are graphically illustrated in Fig. 7 and Fig. 8. In these diagrams, the vertical
displacement between calculated values for the specific discontinuities orientation
are approximated by bisector.
Case J
In the case J (one-sided dip of discontinuities), the maximum vertical displacement
at the top of the underground structure was found for dip angle 45o and equals
24.0 mm. If the dip angle increases or decreases the value of vertical displacement
at the top of the underground structure decreases. The lowest value (19.5 mm) of
the vertical displacement in the point I was verified for the case of horizontally
oriented discontinuities (α1 = 0o). This value practically corresponds to the value
obtained in the case of vertically oriented discontinuities (α1 = 90o).
72 J. Boštík
-15
0 15 30 45 60 75 90
vertical displacement - point I [mm] -20
-25
-30
-35
-40
dip angle [o]
J A S
Figure 7. Vertical displacement at the top of the underground structure
-5
0 15 30 45 60 75 90
max. vertical surface displacement [mm]
-6
-7
-8
-9
-10
-11
dip angle [o]
J A S
Figure 8. Maximum vertical surface displacement
Similar relation can also be found for maximum value of the vertical surface
displacement. The highest value of 10.3 mm is achieved for α1 = 60o (almost the
same as 10.2 mm for α1 = 45o). Minimum value of the vertical surface
displacement of 5.9 mm was verified for α1 = 0o and in fact corresponds to the
value of 6.4 mm calculated for α1 = 90o.
Case A
In the case A, the maximum vertical displacement at the top of the underground
structure was found for dip angle 60o and equals 32.5 mm. The dependency on α1
is of almost the same as in the case J except the displacements are bigger and the
maximum value is found for steeper discontinuities dip as presented above.
Variations with α1 = 15, 30 and 75o when the calculated displacements are slightly
smaller; more specifically 1.4, 0.3 and 1.6 mm respectively, are an exception.
Maximum vertical surface displacement reaches the biggest value of 11 mm for α1
= 45o. For α1 = 60o, maximum vertical surface displacement of 10.3 mm is the
same as in the case J. Nevertheless, this does not apply to vertical displacement of
a point at the top of the underground structure, which is for α1 = 60o significantly
bigger than in the case J. The difference equals 8.8 mm, which is approximately
37 %.
Case S
In the case of situating the underground structure in synclinal bend for all
alternatives considered, the maximum vertical displacement at top of the structure
(40.0 mm) is achieved for α1 = 15o. Concurently, for dip angle from the range of
30o – 75o, vertical displacement of the point I does not significantly vary, similarly
to the case J.
The maximum vertical surface displacement (11.0 mm) was found for α1 = 60o as
in the case J. The change in relation to the discontinuities orientation is basically of
simillar nature as in the earlier cases.
a b c
Figure
9. Total displacements around the underground structure: a – case J (α1 = 45o), b–
case A (α1 = 60o), c – case S (α1 = 15o)
A field of displacements in the very vicinity of shape of the underground structure

is illustrated by arrows in Fig. 9. Only those alternatives of analysed cases, where
the biggest vertical displacement at the top of the underground structure was
calculated, are presented. Here, the deformation of the underground structure shape
for one-sided dip of discontinuities is asymmetric and the biggest displacement is
approximately localised in the area of vault support in the direction transverse to
the direction of discontinuities (Fig. 9a). In the case A, the biggest displacements
74 J. Boštík
occur at the top of the underground structure and at the sides of the underground
structure closely below its vault support (Fig. 9b). In the case S, the biggest
displacement was found at the top of the underground structure (Fig. 9c).
4. CONCLUSION
From the results of the parametric study, it can be concluded that from the point of
view of vertical deformation at the top of the top heading, driving of the
underground structure in the axis of synclinal bend, if the dip of discontinuities is
not very steep (α1 = 15o), is the least favourable. Further, for the case of situating
the underground structure in the axis of anticlinal bend, the calculated maximum
vertical displacement at the top of the underground structure is 19 % less than in
the previous case. From this perspective, the most favourable is driving of the
underground structure in case of one-sided askew aslope layers, where maximum
value of vertical displacement of the top of the top heading is 60 % of the value
obtained for the underground structure in synclinal bend.
From the perspective of the maximum vertical surface displacement, situating the
underground structure, in relation to the discontinuities orientation, seems to be of
no importance. This observed quantity is practically of the same value for all the
cases analysed.
It also results from the analysis, that maximum values of vertical displacement of
the top of the underground structure and vertical surface displacement are not
reached for the same discontinuities orientation, which is significant especially in
the case of synclinal bend.
Along to the all above it needs to be considered, that the alternatives were analysed
only for certain selected values of dip angle α1 in the presented analysis. Therefore,
the maximum values of the monitored displacements and corresponding values of
dip angle do not have to be the values of absolute maximum for the case itself.
Acknowledgements
The contribution was processed within the grant project GA ČR 103/07/P323 and
the research project of MSM0021630519.
References
1. Záruba, Q., Mencl, V., Engineering geology, Publisher ČSAV, Prague, 1954. (in Czech)
2. Plaxis 2D – Version 8, Edited by R. B. J. Brinkgreve, A.A. Balkema Publishers, 2002.
Discrete model for the stability of continuous welded rail

Adam Dósa, Valentin-Vasile Ungureanu
Department of civil engineering, “TRANSILVANIA” UNIVERSITY, Braşov, Romania
Summary
In this paper a discrete model is developed for the buckling analysis of continuous
welded rail subjected to temperature load. The model is based on a nonlinear
analysis in total lagrangean formulation. The structure consists of beam elements
and lateral, longitudinal and torsional spring elements. The source of nonlinearity
is due to the geometric nonlinearity of the rail high axial forces and also to the
nonlinearity of material type for the lateral and longitudinal resistance of the
ballast and the torsional resistance of the fasteners. The use of a displacement
control algorithm leads the analysis beyond the critical point and permits a more
realistic computation of the structural safety. The track model is encoded into a
special purpose program which allows a parametric study of the influence of
vehicle loading, the stiffness properties of the structure and of the geometric
imperfections on the track stability.
The validity of the present model is verified through a series of comparative
analyses with other author’s results.
KEYWORDS: Continuous welded rail, Non-linear stability analysis, Temperature

loading, Structural safety.
1. INTRODUCTION
The first computational models of the buckling of the continuous welded rail
(CWR) were developed at the beginning of the 1930 years. These models can take
into account the main parameters which control the stability of CWR like the
horizontal and vertical stiffness of the rail, the longitudinal and transversal
resistance of the rail, the torsional resistance of the fasteners, the stresses induced
by the vehicle and temperature loading, the geometry and the misalignment of the
rail. In the SCFJ model presented in this paper the structure consists of beam
elements and lateral, longitudinal and torsional spring elements. The beam
elements are modeling the rail and have geometric nonlinear characteristics due to
high compressive thermal stresses. The spring elements are describing the material
nonlinear behavior of the ballast and the fasteners.
76 A. Dósa, V. Ungureanu
2. DEVELOPMENT OF THE TRACK MODEL
2.1. The longitudinal ballast behavior
In the SCFJ model the longitudinal resistance of the ballast is introduced by spring
elements having the, linear or bilinear displacement-force curves given in figure 1.
In the case of vehicle loading, the bilinear curve is corrected [6] by the equation (1)
taking into account the vertical force Q on each sleeper.
2
U vc = U v + Q ⋅ tan φL , U vc ≥ U v (1)
3
In the above equation U v is the reference value of the longitudinal resistance
(without vehicle loading), U vc is the corrected value of this resistance and φL is the
angle of the longitudinal friction between the sleeper and the ballast.
Figure 1. Longitudinal resistance versus longitudinal displacement of the ballast
2.2. The transversal ballast behavior
The transversal resistance of the ballast is introduced by spring elements having the
displacement-force curves given in figure 2. In both cases the elasto-plastic model
includes softening. This kind of ballast behavior has been measured for
consolidated ballast. In the case of vehicle loading, the bilinear curve is corrected
[6] by the equation (1) taking into account the vertical force Q on each sleeper
using equations (2), (3) or (4).
2
Vvc = Vv + Q ⋅ tan φT , Vvc ≥ Vv (2)
3
Vrc = Vr ⋅Vvc / Vv (3)
v − vv
−2
vr − vv
for tipV=1 : V = Vrc + ( Vvc − Vrc )⋅ 2 (if v > vv ) (4)
In the above ecuations Vv is the reference value of the peak transversal resistance
(without vehicle loading), Vvc is the corrected value of this resistance, φT is the
angle of the transversal friction between the sleeper and the ballast, Vr is the
reference value of the rezidual transversal resistance (without vehicle loading), and
Vrc is the corrected value of this resistance. In the case of exponential softening the
difference Vvc − Vrc is half at the middle of vvc − vrc interval.
Figure 2. Transversal resistance versus transversal displacement of the ballast
2.3. The torsional stiffness of the fasteners
The resistance of the fasteners is introduced by torsional springs having the linear
or tri-linear behavior shown in figure 3. In the case of loaded rail this behavior also
can be corrected taking into account the vertical force acting on each sleeper.
Figure 3. The torsional stiffness of the fasteners

2.4. The geometrical and physical characteristics of the rail
The rail is modeled by beam elements having area of the cross section A, second
order moment about the vertical and horizontal axes Iz and Iy respectively. The
Young modulus and the thermal expansion coefficient of the material are E and
respectively α. In the model the misalignment of the rail can be described by two
types of curves: a complete or a half cosine wave having the total length λ and the
amplitude δ (figure 6). The length of the model is an input of the program. At the
end of the model special infinite boundary elements are introduced -equivalent
with the theoretical infinite rail [6]. This elements lead to the reduction of the
length of the model and hence the computational effort. Further reduction can be
obtained by using the symmetric half structure.
3. THE NUMERICAL ALGORITHM
Since in a simplified manner, the horizontal and vertical behavior are considered
decoupled, the numerical algorithm has two phases.
3.1. The computational model for vertical loadings
This model is linear elastic consisting of a beam on elastic springs. The nodes of
the structure are considered at the sleepers. Each node has two degrees of freedom:
the vertical translation w and the rotation θy. The system of equations of
equilibrium is:
Ka = F . (5)
where:
K is the stiffness matrix of the structure and results by assembling the stiffness
matrices k of the beams and the vertical stiffness of the fasteners.
a is the displacement vector of the nodes of the structure.
F is the vector of forces at the nodes of the structure, which (in this case) results by
assembling the vectors f0 of the forces on the beams.
The stiffness matrix k(4x4) of a beam is given by the equation:
k = BT k d B . (6)
Here B(2 x 4) is a transformation vector, which links the vector of displacements of
the beam and the reduced vector of displacements of the beam. The reduced vector
of displacements does not contain the rigid body displacements.
 θ yid  1 L 1 − 1 L 0
a del =  d  = B ⋅ a el =   (wi θ yi wi +1 θ yi +1 )
T
(7)
θ yi +1  1 L 0 − 1 L 1
Figure 4. The displacements of the beam in the vertical plane
kd(2 x 2) is the reduced stiffness matrix of the beam.

EI y 4 2
kd = (8)
L 2 4
If the beam is loaded, the vector f0 of equivalent forces in the nodes is given by
equations (9).
M i0 = P ⋅ b 2 a L2 Ti 0 = ( P ⋅ b + M i + M i +1 ) L
M i0+1 = − P ⋅ a 2 b L2 Ti 0+1 = ( P ⋅ a − M i − M i +1 ) L
f0 = ( Ti 0 M i0 Ti 0+1 M i0+1 )T (9)
Figure 5. Equivalent nodal forces of the beam
The stiffness matrices and the load vectors of the beams are assembled by the
relation (10).
K ind ,ind = K ind ,ind + k , Find = Find + f0 . (10)
Here ind is the vector of the indices of the displacements of the current beam.
The stiffness of the sleepers is assembled with the help of the equation (11).
K jnd , jnd = K jnd , jnd + Rz L (11)
In the above equation jnd is the set of indices of vertical displacements of the nodes
jnd=1, 3, ..., 2nnd-1. The constraints of the structure are introduced by setting to
zero the displacements of the supports. The free displacements of the nodes result
by solving the system of linear equations:
aid = ( K id ,id )−1 Fid . (12)
In equation (12) id is the set of the free displacements of the structure.

Using the vertical displacements, the vertical force on each sleeper can be
computed by the equation (13)
Q = − wRz L + Gsleeper . (13)
The transversal, longitudinal and torsional resistances are corrected taking into
account the forces Q on each sleeper using equations (1) to (4).
3.2. The computational model in the horizontal plane
The model is a straight or curved beam on elastic supports with misalignments

(figure 6). The nodes of the structure are considered at the sleepers. At each node
are introduced longitudinal, transversal and rotational spring elements which are
modeling the sleepers. The infinite boundary elements at the ends of the model
have equivalent characteristics (Young modulus and thermal expansion coefficient)
in order to replace the theoretical infinite rail [6]. The loading of the model is an
increase of the temperature in the rail. The characteristics of the beams and of the
springs correspond to the two rails of the track panel. A node has three degrees of
freedom: two linear displacements in the horizontal plane, u and v and the rotation
θz around the vertical axis. In the analysis of the structure the goal is to obtain the
displacement-temperature curve. The problem is solved by a displacement control
based incremental process. The behavior of the system is determined as a sequence
of increments of state parameters (forces and displacements). In the current
increment j characterized by the small control displacement δvcj, the nonlinear
behavior of the system can be approximated by a linear relation between the
successive increments of the state parameters:
a j +1 = a j + δa j , δF j = K jδa j . (14)
In the above equation aj is the displacement vector in the current configuration, δaj
is the increment of the displacements, δFj is the incremental load vector and Kj is
the incremental (tangent) stiffness matrix of the structure.
Figure 6. The model for horizontal displacements
By using equations (14), the following incremental scheme results:

δa j = ( K j )−1δF j , a j +1 = a j + δa j . (15)
In this paper an improved scheme is used, known as Heun’s or midpoint rule:

a j +1 2 = a j + 1 2 ( K j )−1δF j ,
K j +1 2 = K( a j +1 2 ),
(16)
δF j +1 2 = δF( a j +1 2 ),
a j +1 = a j + 1 2 ( K j +1 2 )−1δF j +1 2 .
The incremental load vector δFj is not computed. The incremental displacement δaj
is the result of a yet unknown increment of the temperature produced by a known
increment δvcj of the control displacement. For simplicity, in the next equations
indices j of the current configuration are dropped. The displacement control
consists of loading the system with displacement increments δvc in a specific node.
As a rule in this paper: the control displacement is taken as the maximum
transversal displacement of the node on the symmetry axe of the structure. The
phases of the computation are the following:
- It is adopted a base system with the control displacement fixed at zero.
- This base system is loaded with two load cases:
i) Load 1 is a temperature increase δT =1, which produces displacements δa(1) at
the free nodes and reaction R(1) in the artificial support.
ii) Load 2 is a displacement δvc of the artificial support, which produces
displacements δa(2) at the free nodes and reaction R(2) in the artificial support.
Figure 7. The determination of the temperature and displacement increments
The base system and the initial system are identical if the total reaction in the
artificial support is zero. R= R(1)δT + R(2)=0. This equation yields the unknown
variation δT of temperature and the incremental displacements δa of the free nodes.
δT = − R ( 2 ) / R ( 1 ) , (17)
δa = δa( 1 ) ⋅ δT + δa( 2 ) . (18)

The tangent stiffness matrix Kj in the j increment depends on the parameters of the
system in the current step and results by assembling the stiffness matrices kt(6 x 6) of
the beams and of the springs which model the sleepers.
k t = EA / L ⋅ r T r + BT ( k d + k Gd )B + N j / L ⋅ z T z . (19)
In this equation r=(-cosβ -sinβ 0 cosβ sinβ 0 ),
z=( sinβ -cosβ 0 -sinβ cosβ 0 ),
 0 0 1 0 0 0 1  z  EI z  4 2 N j L  4 − 1
B= −  , kd =  2 4 , k Gd = ,
0 0 0 0 0 1  L  z  L   30  − 1 4 
and Nj is the axial force in the beam in the j-th incremental step: Nj=EA ∆ Lj/ Lj,
∆L j = L j − L0 , L0 = ( x i0+1 − x i0 )2 , L j = ( x ij+1 − x ij )2 .
Figure 8. The axial deformation of the beam
Matrices k d and k Gd are the material and geometric stiffness matrices respectively.
They are expressed with the reduced set of displacements which produce
deformations and they are not containing the rigid body displacements of the beam.
This reduced form of the stiffness matrices needs less computational effort and
speeds up significantly the computation. Equation (19) introduces the non-linear
effect of the axial force Nj. The complete tangent stiffness matrix in the updated
lagrangean formulation used here has two more terms corresponding to the
variation of the length of the beam in bending and to the effect of the shear force
[1], [2], [3], [4], [5]. Since in the current cases the structure is divided in a
sufficient number of beams, the errors are very small, when neglecting these two
terms. In a study using the complete tangent stiffness matrix and equation (19) the
differences between the resulting limit temperatures were only at the fifth digit.
4. NUMERICAL EXPERIMENTS
The validity of the present model is verified through a series of comparative

analyses with other author’s results. The numerical example presented here
corresponds with one given in [1]. The track length is L=24.359/2 m corresponding
to 21 sleepers on the symmetric half of the structure. The curvature radius is R=400
m. The horizontal misalignment is characterized by a half wave cosine with a
length λ=9.144 m and an amplitude δ=0.0381 m. The rails have the characteristics
of two AREA 136 rails. The vertical stiffness of the ballast elements is Rz=68900
kN/m per meter of track. The longitudinal stiffness is Rx=1378 kN/m per meter of
track. The torsional stiffness of the fasteners is 111.250 kNm/rad per meter of
track. Laterally the ballast is modeled by the tri-linear constitutive behavior given
in figure 2. The reference values of the lateral peak resistance and residual
resistance are Vv=17.508 and Vr=9.630 kN per meter of track. These values are
corrected with the vertical forces resulting from the vehicle loading. The lateral
displacement at the peak value is vv=0.00635 m and at the limit value is vr=0.0381
m. The model is vertically loaded by a vehicle with two bogies represented by four
vertical loads of 293 kN each. The centre spacing between the bogies is 12.85 m.
The spacing between the axles in a bogie is 1.78 m. The centre of the misalignment
is located in the middle between the bogies. The track is loaded by a temperature
increase from zero to a maximal value corresponding to the buckling of the rail.
The lateral displacement of the middle node versus the resulting temperature
increase is shown in figure 9.
50
Tmax
Temperature increase (deg. Celsius)
45
40
35
30 Tmin
25
20
15
10
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Lateral displacement (m)
Figure 9. Lateral displacement versus temperature increase
The curve in the figure 9 is characterized by two points: Tmax - the maximum
increase of temperature for which the buckling certainly starts, and Tmin - the
minimum increase of temperature which occurs in the post-buckling domain. The
values computed by the SCFJ model - Tmax=49.5 0C and Tmin=33.3 0C - are in a
good agreement with those given in [1].
References
1. Bănuţ, V. Calculul neliniar al structurilor, Editura tehnică, Bucureşti, 1981.
2. Crisfield, M.A. Non-linear Finite Element Analysis of Solids and Structures, Wiley, 1991.
3. Dósa,A., Litră E. Elemente de bară încovoiată cu precizie îmbunătăţită pentru calculul neliniar şi
de stabilitate , Revista Construcţiilor, 2006.
4. Dósa,A., Popa L. High order beam elements for the stability and non-linear analysis of frame
structures, “Computational Civil Engineering 2006”, Iaşi, România.
5. Felippa C.A., Nonlinear Finite Element Methods,
www.colorado.edu/engineering/CAS/courses.d/NFEM.d.
6. Van, M.A., Stability of Continuous Welded Rail Track, Delft University Press, 1997.
Iaşi, Romania, May 25, 2007
Soil-Structure Interaction Effects of Damped Structures

Petre Ene
Technical University of Construction, Bucharest, Romania
Summary
In the past, much effort has been devoted to investigating the seismic behavior of
structures with seismic dampers by neglecting the effects of foundation-structure
interaction. In order to ensure the safety of structures, it is necessary to apprehend
the interactive behavior of damped structures and unbounded foundations under
intense earthquake ground motions. This paper presents a rigorous time-domain
procedure to address the interaction effects of structures equipped with fluid
viscous dampers and foundations with an unbounded medium. Quantitative results
show that, during earthquakes, there are significant differences between a system
with and without radiation damping. For greater accuracy, radiation damping
should be properly taken into account. More-over, the efficiency of fluid viscous
dampers in reducing seismic disturbance of a structure is very dependent on the
flexibility of the foundation.
1. INTRODUCTION
Seismic dampers have been recently used to enhance seismic performance for
some tall buildings in California (Pong, 1999). For example, an eight story
concrete building located on the campus of the California State University at Los
Angeles was retrofitted by adding viscous dampers (Elhassan et al., 1996). An
eleven story steel building, the new national headquarters for the Money Store
located in Sacramento, California, was one of the first buildings in the United
States that use seismic dampers to limit story drift during earthquakes (Miyamoto
and Scholl, 1998). However, many structural analyses and designs ignore the
behavior of soil-structure interaction due to the complexity and uncertainty of
modeling techniques. In the last twenty years, some advances have been made in
this area. Most notable is the development of procedures based on boundary-
integral or boundary-element methods for unbounded media. Nonetheless, a major
need still exists for alternative approaches, particularly for procedures that could be
implemented within the context of finite-element analyses. These procedures
would be more familiar to structural analyses, and would allow greater flexibility
in terms of the geometrical and material characterization of the unbounded
medium.
In this paper, a new time-domain procedure, based on the finite element method,
the semi-analytical solution and the infinitesimal finite-element cell method, is
86 P. Ene
presented to address the interactive behavior of structures equipped with fluid

viscous dampers and foundation with an unbounded region (Fig.1a). To yield more
accurate results, a time domain procedure has been developed to simulate the
radiation features of an infinite domain.
Fig. 1a Ten-Story Building Equipped with Fluid Viscous Dampers and Foundation
Fig. 1b Finite Element Mesh for Building Equipped with Fluid Viscous Damper
and Foundation
2. TIME-DOMAIN SUBSTRUCTURE METHOD

The substructure method is adopted for the study. The entire system is divided into
three substructures including the building, foundation and fluid dampers.
Furthermore, the soil medium is also divided into two regions: the near and the far
field of the soil. The far field of the soil extends to infinity.
Equations of Motion of Building and Near Field of Foundation
As shown in Figs. 1a and 1b, the structure, fluid damper and the near field of the
soil can be descretized by the finite element method. The equation of motion of the
structure and the near field of the foundation is given as:
M U&& ( t ) + C U& (t ) + KU (t ) = − M B u&& g (t ) + Q (t ) (1)
M, C, K = symmetrical mass, damping, and stiffness matrices, respectively, for the

building and the near field of the foundation;
U ( t ) , U& ( t ) , U&& ( t ) = vectors of nodal displacements, velocities, and accelerations
relative to the ground, respectively;
= ground motion;
B = displacement transformation matrix;
And Q (t) = vector of dynamic forces resulting from the far field of the foundation.
Analytical Model for Fluid Viscous Damper
The fluid damper exhibits viscoelastic fluid behavior over a large frequency. The
simplest model to simulate the mechanical behavior of the fluid viscous damper is
the Maxwell model (Bird et al., 1987) given as:
P ( t ) + λ P& ( t ) = C 0U& ( t )
(2)
Where “t” is the relaxation time, is the damping constant at zero frequency, P (t)
is the force acting on the fluid damper, is the damper position velocity.
Another finite element formulation model for fluid dampers was adopted in this
study (Pong et al., 1994). Using the virtual work principle, the equivalent modal
forces F (t) can be obtained as:
(3)
Where the matrix is the added damping result from fluid dampers:
(4)
System Equations for Far Field of Foundation
As shown in Fig. 1a, the foundation medium is divided into two regions: the near
and far fields of the foundation. The far field of the foundation extends to infinity.
Adopting the similarity theorem, as shown in Fig. 2, Song and Wolf developed the
consistent infinitesimal finite-element cell method accounting for the radiation
damping for the unbounded medium in the time domain (Song and Wolf, 1995;
1996).
88 P. Ene
Fig.2 Fundamental Concept of Consistent infinitesimal Finite-Element Cell

Method
The governing equation of the far field of foundation can be expressed as:
(5)
Where the interaction is force at the interface of the near and far fields of the
unbounded medium and is the acceleration unit-impulse response matrix in
the time domain. The relation of the vector of dynamic forces from the far field
to those acting on the near field is given as:
(6)
As shown in Fig.3, owing to similarity, the coordinates of the nodes on the exterior
boundary can be expressed by those of the nodes on the interior boundary and the
dimensionless cell width
(7)
(8)
The nodes on the interior boundary are identified by subscripts 1, 2, 3 and the
nodes on the exterior boundary are identified by subscripts 4, 5, 6, respectively.
The cell width can be expressed as and
(9)
(10)
Fig.3 Two-dimensional Finite-element Cell and Parent Element on Descretized

Interface
Note that the subscript i and e portray the interior boundary and the exterior
boundary, respectively. denote the coordinates of the nodes on the
exterior boundary. indicate the coordinates of the nodes on the interior
boundary. The shape function matrix of the finite-element is:
(11)
, (12)
Where , N = the shape function vector in the direction.

Therefore, the stiffness matrix of the finite element equals:
, (13)
The Jacobean matrix of a particular element equals:
(14)
Where (15)
The determinant of the Jacobean matrix is given as:
(16)
where is denoted as (17)
The strain-nodal displacement matrix is

90 P. Ene
(18)
, (19)
And (20)
, (21)
the shape functions at nodal point k. The final form of the consistent
infinitesimal finite-element cell equation in the time domain is given as:
(22)
Where (23)
(24)
(25)
(26)
Fig.4a Comparison of Relative Displacements between Points A and B while

Elastic Modulus of Near Field is and Elastic Modulus of Near Field is
(27)
(28)
mass density. After determining from eq. (22), the acceleration unit-
impulse response matrix may be written as:
(29)
The entire history of time is divided into n equal intervals . It is assumed that
matrix and the acceleration is piece-wise constant over each time step:
, (30)
With the aid of eq. (30), of eq. (5) can be expressed as:
(31)
Where (32)
(33)
Time-Domain Finite Element Method for Structure Equipped with Fluid Viscous
Dampers and Unbounded Media
Substituting eq. (6) and eq. (31) into eq. (1) results in
(34)
92 P. Ene
Fig.4b Comparison of Shear Forces at Point B while Elastic Modulus of Far Field
is and Elastic Modulus of Near Field is
is an added-load matrix representing the previous time effect. Applying the new
mark method to eq.(34), at the Nth time step, we have
(35)
Where (36)
(37)
Where
.
Application
A ten story steel building frame was used for this application. The building span
width is 9,20m. The typical story height is 3,70m, except the first story height is
4,60m. Point A is designated at the roof, while Point B is designated at the base.
The material used for the building has Poisson‘s ratio and elastic modulus equal to
0,3 and 200 GPa, respectively. The weight for each floor is 4460,60N/m. The
radius of the near field of the soil below the building is around the 249m. The
Poisson’s ratio of the stratum is 0,30 and the weight density is 1,80KN/m3.
It was assumed that the floor were in their own plane. The damping constant C0 of
fluid viscous damper is 525,59KNs/m and the whole system was subjected to 1940
El Centro earthquake ground motion. Fig. 4a shows the shear force responses at
point B when the building equipped with fluid viscous dampers during
earthquakes. In these figures, the bold-solid line is the seismic response of the
building on rigid foundation (without soil) and the dotted line is the response of the
building on the flexible foundation with radiation damping during earthquakes.

With reference to Fig. 4a and 4b, they indicate that the seismic responses of the
building have remarkable differences on different foundations.
Fig. 5a Comparison of Max. Relative Displacements between Points A and B while

Elastic Modulus of Far Field is and Different Modulus of Near Field
are Applied
Fig. 5a Comparison of Max. Shear Forces at Point B while Elastic Modulus of Far
Field is and Different Modulus of Near Field are Applied
When the elastic modules of the near field of foundation are different, the
comparison of the maximum relative displacement responses at points A and B of
the building equipped with fluid viscous dampers on the unbounded soil are shown
in Fig. 5a. In addition, Fig. 5b shows the comparisons of the maximum shear force
responses at point B when the building equipped with fluid viscous dampers on the
unbounded soil. These two figures show that significant differences exist between
the responses of the building on different foundations during earthquakes. They
also show that the whole system becomes more flexible due to the interaction
effects of the foundation and the effectiveness of the fluid viscous damper in
reducing seismic responses of the building is considerably downgraded.
94 P. Ene
Conclusions
Seismic dampers have been recognized as an effective tool in improving seismic
resistance for both new and existing buildings. Recently, they have been used
extensively to achieve enhanced structural performance for the seismic
rehabilitation of tall buildings in California. Nevertheless, soil-structure interaction
is a very complicated behavior and it requires time-consuming effort in structural
analyses and designs. To ensure the safety of a structure, it necessitates considering
the interaction behavior of the damped structure under intense ground motions. It is
prudent to consider the effects of radiation damping and the flexibility of the soil
media when designing tall buildings with seismic dampers since taller buildings
tend to be much more flexible than shorter buildings.
This study demonstrates differences between a damped structure with and without
consideration of the soil media. The structural behavior of a building can be quite
different if the soil-structure interaction is ignored. Quantitative results show that
remarkable differences exist in the seismic behavior of the system with and without
a flexible foundation. The role of radiation damping becomes extremely important
during an earthquake when supplemental damping is added to the structure.
Therefore, the design of a tall structure with supplemental damping in an
earthquake one area should consider the effects of soil-structure interaction to
ensure optimal seismic performance.
References
1. Elhassan, R.M., Arminak, A., 1996 – Design and Analysis of a Seismic Retrofit of a Tall
Concrete Building using Supplemental Viscous Dampers;
2. Miyamoto, H.K., Scholl, R., 1998 – Modern Steel Construction;
3. Pong, W.S., 1999 – Performance - Based Design Procedures for Buildings with supplemental
Dampers;
4. Pong, W.S., Tsai, C.S., 1994 – Seismic Study of Building Frames with added Energy-absorbing
Devices;
5. Song, C., Wolf, J.P., 1995 – Consistent Infinitesimal Finite-Element Cell Method: Out of Plane
Motion;
6. Song, C., Wolf, J.P., 1996 – Consistent Infinitesimal Finite-Element Cell Method: Three-
Dimensional Vector Wave Equation.
Numerical Modeling of Fiber Reinforced Concrete (FRC).

Fiber orientation angle α = 00, fiber length lf = 50mm
Liviu Gherman1, Petru Mihai2, Nicolae Florea3, Constantin Gavriloaia4, and
Ioan Paul Vodă5
1
Faculty of Civil Engineering, BMTO, Tehnical University “Gh. Asachi”, Iasi, Romania
2
Faculty of Civil Engineering, BMTO,Tehnical University “Gh. Asachi”, Iasi, Romania
3
4
5
Summary
The paper presents a new approach on FRC study, which has similar
characteristics to a multi-phase composite material. Numerical modeling is
applied, using a finite element method (FEM) program, to study the behavior of
linear structural elements, made from steel fiber reinforced concrete (SFRC).
The results stress out the favorable influence of the fibers on the bearing capacity
and especially on the ductility of the loaded members. Lack of design standards
and practice codes on this material is bringing this field of study under
researchers’ consideration.
KEYWORDS: numerical modeling, steel fiber reinforced concrete, influence of the

steel fibers.
96 L.Gherman, P.Mihai, N.Florea, C-tin.Gavriloaia, I.P.Vodă
1. INTRODUCTION
Enhancement of SFRC properties is more evident for hardened concrete than for
fresh concrete. It is influenced by the fiber characteristics: length and geometry,
reinforcement percentage, orientation and adherence to the concrete matrix.
For this new type of material, the workability is decreasing, the concrete having
low to very low plasticity. If a high reinforced percentage is used, or fibers having
length higher than the critical one, the workability decreases dramatically, the
fibers having the tendency to bundle. This shortcoming can be overcome with
supplementary execution costs.
Studying the influence of the fibers on the concrete behavior, as well as the wide
possibilities for using this material makes it necessary that it is studied using
numerical modeling.
2. NUMERICAL MODELING USING FEM
2.1. Used Finite Element Types
Any loaded body with a given shape must be analyzed, in order to establish the
proper finite elements to be used for meshing. Usually, the choice of the finite
element is determined by the geometry of the body and by the number of
independent spatial coordinates necessary to describe the system. Some types of
finite elements, one-dimensional or bi-dimensional, are presented in Figures 1, 2.
When the geometry, properties and other parameters of the material (stresses,
displacements, pressure, temperature) can be described using only one spatial
coordinate, the one-dimensional finite element can be used (Figure 1). Even though
this element has a transversal cross-section, it can be idealized as a line.
In some cases, the cross-sectional area of this element may vary along its length.
node node node node
1 2 1 2
Figure 1 – One-dimensional finite element
When the geometry and details of the studied problem can be expressed function of
two independent spatial coordinates, bi-dimensional finite elements are used
(Figure 2), the basic type having triangular shape.
3
3 3
1 2 1 4 2 1 4 5 2
Figura 2 – Bi-dimensional triangular finite element
2.2. Finite Element Analysis Steps
Most of the finite element analysis steps are performed successfully by the FEM
program, so their succession is highly simplified:
• The preparatory stage, consisting in the precise meshing of the element, using
a CAD program, after which it was exported to a FEM program.
• Definition of the physical and mechanical characteristics of the materials, and
of the plain stress analysis manner in the nonlinear domain.
• Analysis and interpretation of the results.
2.2.1. The meshing was performed in two variants.

The corresponding free body diagrams are also presented (Figure 3, 4).
a.) plain concrete beam, where triangular finite elements were used,
b.) SFRC beam, where triangular elements were used for the concrete and one-
dimensional finite elements were used for the steel fibers. The fiber orientation
angle was considered α = 00.
1.)
2.)
3.)
Figure 3 – Meshing
0
3.1 – fiber orientation angle α = 0 , 3.2 – plain concrete beam, 3.3 – SFRC beam
2.2.2. Definition of the physical and mechanical characteristics of the materials

and of the plain stress analysis manner in the nonlinear domain
The studied model has the cross-sectional area A = 108cm2 (12x9cm) and the
length 150cm, being simply supported. The distance between the supports is
120cm. It is acted by two concentrated forces, 40cm apart from each other and
from the supports, so that on the middle third the bending moment is constant
(Figure 4). The concrete class is C20/25, with Young modulus E = 3000N/mm2.
The disperse fibers (the fiber orientation angle being considered α=00), have the
diameter Φ =1,05mm, the cross-sectional area A ≈ 1mm2, Young modulus E =
210000N/mm2, tensile strength Rti = 200N/mm2, and length l f = 50mm .
P P
+
Q
-
M
+
Figure 4 – Free body diagram of the beam
2.2.3. Analysis and interpretation of the results.
After applying the loading steps, there were obtained the load-displacements
curves presented in Figures 5.1 and 5.2, and the stress maps corresponding to the
loading steps (Figures 6 and 7).
P [KN]
P [KN]
δ [mm] δ [mm]
1.) Plain concrete beam 2.) SFRC beam
Figure 5 – Load-displacement curve (P-δ) for:

5.1. – the plain concrete beam, 5.2. – the SFRC beam
-1,4
-1,1
-0,86
-0,57
-0,29
Y
0
0,29
0,57 Z X
0,86
1,1
Max 1.365 at Node 3014
Min -1.502 at Node 627
-2,8
-2,3
-1,9
-1,4
-0,93
-0,47 Y
0
0,47 Z X
0,93
1,4
Min -2.893 at Node 627
-4,6
-4
-3,3
-2,6
-2
Y
-1,3
-0,66
0 Z X
0,66
1,3
Min -4.865 at Node 3095
LOAD CASE = 10
Increment 10 Load Factor = 4.05000
RESULTS FILE = 1
STRESS
CONTOURS OF SX
-5,8
-5
-4,1
-3,3
-2,5
-1,7 Y
-0,83
0 Z X
0,83
1,7
Min -6.511 at Node 3095
LOAD CASE = 11
Increment 11 Load Factor = 4.12500
RESULTS FILE = 1
STRESS
CONTOURS OF SX
-7
-6,1
-5,2
-4,4
-3,5
Y
-2,6
-1,7
-0,87 Z X
0
0,87
Min -6.999 at Node 3091
Figure 7 – Stresses maps for the plain concrete beam.

STRESS
CONTOURS OF S
0
0,0842
0,168
0,253
0,337
0,421
0,505
0,589
0,674
0,758
Y 0,842
0,926
1,01
Z X
1,09
1,18
1,26
Max 1.347 at Nod
Min 0.9148E-04 a
STRESS
CONTOURS OF S
0
0,288
0,576
0,864
1,15
1,44
1,73
2,02
2,3
2,59
Y 2,88
3,17
3,46
Z X
3,74
4,03
4,32
Max 4.609 at Nod
Min 0.7766E-03 a
STRESS
CONTOURS OF S
0
0,521
1,04
1,56
2,08
2,6
3,12
3,65
4,17
4,69
Y 5,21
5,73
6,25
Z X
6,77
7,29
7,81
Max 8.334 at Nod
Min 0.1030E-02 a
STRESS
CONTOURS OF S
0
0,767
1,53
2,3
3,07
3,84
4,6
5,37
6,14
6,91
Y 7,67
8,44
9,21
Z X
9,98
10,7
11,5
Max 12.28 at Nod
Min 0.1278E-02 a
STRESS
CONTOURS OF S
0
0,243
0,485
0,728
0,97
1,21
1,46
1,7
1,94
2,18
Y 2,43
2,67
2,91
Z X
3,15
3,4
3,64
Max 3.882 at Nod
Min 0.7169E-03 a
Figure 8 – Stresses maps for the SFRC beam

There are presented the load values for each loading steps and the corresponding
displacements, for the plain concrete beam (Table 1) and for the SFRC beam,
respectively (Table 2). The displacement decrease, for each loading step, for the
SFRC beam compared to the plain concrete one, is presented in Table 1, column 3.
Notes and comments on the behavior of the SFRC beam under loading are
presented in Table 3.
Table 1
Loading step Load Displacement Displacements decrease for the Comments
magnitude [mm] SFRC beam, compared to the
[KN] plain concrete beam [mm]
0 1 2 3 4
1 1,5 0,119 0,002
2 1,8 0,143 0,003
3 2,1 0,169 0,004
4 2,4 0,2 0,005
5 2,7 0,241 0,008
6 3 0,296 0,014
7 3,3 0,369 0,024
8 3,6 0,483 0,062
9 3,9 0,639 0,105
10 4,05 0,729 0,056
11 4,125 0,825 0,074 Beam failure
Table 2
Loading step Load magnitude [KN] Displacement Comments
[mm]
0 1 2 3
1 1,5 0,117
2 1,8 0,14
3 2,1 0,165
4 2,4 0,195
5 2,7 0,233
6 3 0,282
7 3,3 0,345
8 3,6 0,42
9 3,9 0,534
10 4,2 0,673
11 4,35 0,751 The first fissure occurs
12 4,5 0,86 The fissure opens
13 4,8 1,231 In-between steps 12 and 13,
the fibers take over the tensile
stresses
14 5,1 1,371 In-between steps 14 and 38,
15 5,175 1,48 the concrete and the fibers
16 5,3 1,6 work together to take over the
tensile stresses
38 6,598 4,127
Table 3
Bearing capacity increase for Displacements of the SFRC
the SFRC beam, compared to beam, compared to the plain
the plain concrete beam concrete beam
Bearing capacity increase until 5,45% The displacement of the SFRC
the first fissure occurs (4,35 compared to 4,125) beam is 9.1% lower
(0,751 compared to 0,825)
concrete failure (4,8 compared to 4,125) beam is 67,02% higher
fibers yielding (5,1 compared to 4,125) beam is 166,18% higher
fibers failure (6,598 compared to 4,125) beam is 500,24% higher
3. CONCLUSIONS
Compared to the plain concrete, the SFRC behavior, on elastic and elasto-plastic
domains, is characterized by:
1. elastic domain – an increase of the bearing capacity:
- until the first fissure, of 5.45%, and the displacement is 9.1% lower;
- until the concrete failure, of 16.36%, and the deformation is 67.02 higher,
leading to the increase of the area under the force-displacement curve.
2. plastic domain - an increase of the bearing capacity:
- until the fibers yielding, of 23.64%, and the displacement is 166.18% higher;
- until the fibers strain hardening and failure, of 23.64%, and the deformation is
500.24 higher.
References
1. Chiaia, B., Fantilli, A.P., Valini, P., Kalamaras, G. – Fiber reinforced concrete for massive
structures the case of craviale tunnel, FIB, Italy, Napoli, 2006.
2. Falkner, H. and Henke, V. Steel fibre reinforced concrete, from research to standards. Concrete
Structures – Annual Technical Journal of the Hungarian Group of fib, 2005; 6: 39-46. Proceedings of
the 2nd Congress Session 2 June 5-8, 2006 – Naples, Italy Design and construction
3. Rilem TC 162-TDF, σ − ε design method – final recommendation. Materials and structures, 36,
560-567, 2003,
4. Barla, G., Barpi, F., Bertolino, C. and Chiaia, B. A note on the design of fibre-reinforced shotcrete
linings for underground support. Computational Modeling of Concrete Structures EURO-C 2003,
Eds. Bicanic, N., de Borst, R., Mang, H. and Meschke, G., A.A. Balkema Publishers, Lisse, 2003:
627-634.
5. Balaguru, P.N. and Shah, S.P. Fiber - Reinforced Cement Composites. McGraw-Hill Inc, New
York, 1992.
6. Cuteanu, E., Marinov, R., - Metoda elementelor finite în proiectarea structurilor, Ed. Facla,
Timişoara, 1980.
A new look into finite element templates.
Wojciech Gilewski1
1
Institute of Structural Mechanics, Warsaw University of Technology, Warsaw, Poland
Summary
The present paper is dedicated to evaluation and new method of construction the
finite element templates. A template is an algebraic form of element matrices,
which contains free parameters. Setting the parameters to specific values produces
element instances. Two templates are analyzed: Bernoulli and Timoshenko beam.
The number of free parameters is discussed by a general method.
KEYWORDS: finite element template, consistency, ellipticity
1. INTRODUCTION
Beams, plates and shells are widely considered in engineering applications.

However the corresponding discretization procedures are not yet sufficiently
reliable. It is difficult to obtain an element that is optimal. In a formulation we
should aim to satisfy: ellipticity, consistency and the inf-sup condition.
Ellipticity ensures that the finite element model is solvable and physically means
there are no spurious zero energy modes. This condition can easily be verified by
studying the zero eigenvalues and corresponding eigenvectors of the stiffness
matrix of a single unsupported finite element. Consistency is related to the
convergence. The finite element solution must converge to the solution of a
mathematical problem the element size h is close to zero. The bilinear forms used
in the finite element discretization must approach the exact bilinear forms of the
mathematical model as h approaches zero. The inf-sup condition ensures optimal
convergence in bending-dominated problems and is not a subject of this paper.
One of the interesting concepts are finite element templates proposed by Felippa
[1,2]. The objective of this paper is to evaluate finite element templates, proposed
by Felippa, with the use of the energy-difference criterion (to check if the template
is consistent) and the spectral analysis (to check if the template is elliptical). Two
beam templates are discussed: for Bernoulli theory and for Timoshenko theory. A
new, general method for development of finite element templates is proposed.
104 W. Gilewski
2. TEMPLATES
A finite element template is an algebraic form for element matrices, which contains
free parameters (Felippa [1,2]). Setting those parameters to specific values
produces element instances. The set of free parameters is called the template
signature. Borrowing the terminology from biogenetics, the signature may be
viewed as an “element DNA” that uniquely characterizes it as an individual entity.
Elements derived by different techniques thet share the same signature are called
clones. The template should fulfill the following conditions:
• consistency (the individual element test is passed),
• stability (correct rank and nonnegativity conditions),
• parametrization (free parameters)
• invariance (the element is observer invariant).
The element stiffness matrix derived through the template approach is based on the
fundamental decomposition
K = K b (α i ) + K h ( β j ) ,
where K b and K h are the basis and higher-order stiffness matrices respectively.
α i , β j are free parameters. These two matrices play different and complementary
roles. The basic stiffness K b takes care of consistency and element-type-mixing.
The higher order stiffness K h is a stabilization term that provides the correct rank
and may be adjusted for accuracy.
3. BERNOULLI BEAM TEMPLATE
Let us consider a finite element template for Bernoulli beam [1] of two-noded
element of the length L with natural d.o.f. q e = {wi , φ i , wk , φ k } .
0 0 0 0  4 2L −4 2L 
 
0 − 1  2
− 2 L L2 
EJ 0 1 EJ  2 L L
K Template = Kb + Kh = +β 3
L 0 0 0 0 L − 4 − 2 L − 2L
Bernoulli
4
   
0 − 1 − 2 L L2 
2
0 1  2L L
For β = 3 the well known beam element with Herimite’s polynomial shape
functions. The template depends on one free parameter β . The energy-difference
procedure (described in details in [3,5], and used for evaluation of beam and plate
finite elements in [4,5]) can be used to check if the template satisfies the
consistency condition.
The strain energy density of a beam is expected to be:

2
~  d 2w 
2 E s = EJ  2  .
 dx 
The equivalent density of element strain energy is a quadratic form
~ 1
2 E sES −Template = q Te K Template
Bernoulli q e .
L
The nodal displacements can be expressed by average displacements and its
derivatives in the midpoint of the element with the use of Taylor series expansion
2 3
∆w L 1 ∆2 w L 1 ∆ w
3
L
wi = w( x) − ( x) + ( x )  − ( x)  + ...,
∆x 2 2 (∆x) 2
 2  6 (∆x)
3
2
2 3
∆w L 1 ∆2 w L 1 ∆3 w L
wk = w( x) + ( x) + ( x)  + ( x)  + ...,
∆x 2 2 (∆x) 2
2 6 (∆x) 3
2
2 4
∆w ∆2 w L 1 ∆3 w L 1 ∆4 w L
φi = ( x) − ( x ) + ( x )  − ( x)  + ...,
∆x (∆x) 2
2 2 (∆x) 3
2 6 (∆x) 4
2
2 4
∆w ∆2 w L 1 ∆3 w L 1 ∆4 w L
φk = ( x) − ( x ) + ( x )  − ( x)  + ...
∆x (∆x) 2
2 2 (∆x) 3
2 6 (∆x) 4
2
After collecting the expressions with respect to L we have
~ ES −Template  ∆2 w 
2
 β  ∆3 w 
2
1 ∆2 w ∆4 w 
2Es = EJ  2  + L2 EJ    +  + 0( L4 ).
 ∆x   36  (∆x) ∆ ∆
3 2 4
 12 x x 
In the limit case L → 0 the following relation is valid

~ ~
lim
L→0
E sES −Template = E s .
This is a proof that the element template satisfies the consistency requirement for
any β . The basis stiffness matrix K b is responsible for the first term of the strain
energy. Formally the β parameter is free from zero to infinity.
106 W. Gilewski
The element template should also elliptical. To check this condition it is necessary
to calculate the eigenvalues and eigenvectors of the template. Two zero
eigenvalues related to rigid body motions are expected. The two other eigenvalues
are to be positive and related to the deformed element. The results are the
following
EJ EJ
λ1 = 0, λ 2 = 0, λ 3 = 2 , λ 4 = 10 β ,
L L
w 1 = {− L, 1, 0, 1 },
w 2 = { 1, 0, 1, 0 },
w 3 = { 0, −1, 0, 1 },
w 1 = { 2 L, 1, −2 L, 1 }.
It is seen that the element is elliptical if β > 0 . If β → 0 the fourth eigenvalue

tends to be zero with the eigenvector that describes deformed element.
Since the development of the element template is rather complicated, a couple of

questions arises after reading the texts of Felippa [1,2]:
How many free parameters exist in the template ?
Are the free parameters exist in both matrices (basis and higher order) ? etc.
Let us examine some more general matrices to answer the questions.
0 0 0 0   4b11 2 Lb12 − 4b13 2 Lb14 

0 a 
0 − a 24  EJ  2 Lb12 L2 b22 − 2 Lb23 L2 b24 
EJ 
K Extended = 22
+
L 0 0  L3 − 4b13 − 2 Lb23 − 2 Lb34 
Bernoulli
0 0 4b33
   
0 − a 24 0 a 44   2 Lb14 L2 b24 − 2 Lb34 L2 b144 
The basis matrix is responsible for the first term of the strain energy, so it
depends only on rotations. There are 3 formally independent parameters in
this matrix. The second matrix formally depends on 10 parameters.
Following the procedure described above one can receive the following
strain energy density of the element:
~ 4
2 E sES − Extended = 4 EJ (b11 − 2b13 + b33 )w 2 +
L
4 ∆w
+ 3 EJ (− b11 + b12 + b14 − b23 + b33 − b34 )w +
L ∆x
2
1  ∆w 
+ EJ (b11 − 2b12 + 2b13 − 2b14 + b22 − 2b23 + 2b24 + b33 − 2b34 + b44 )  +
 ∆x 
2
L
1 ∆2 w
+ EJ (b − 2b − 2b + 2b + 2b + b − 2b )w +
∆x 2
11 12 13 14 23 33 34
L2
2
1  ∆w 
+ 2 EJ (a 22 − 2a 24 + a 44 )  +
L  ∆x 
1 ∆w ∆2 w
+ EJ (− b11 + 3b12 − b14 − 2b22 + b23 + b33 − 3b34 + 2b44 ) +
2L ∆x ∆x 2
1 ∆3 w
+ EJ (− b11 + 3b12 + 3b14 − 3b23 + b33 − 3b34 )w 3 +
2L ∆x
1 ∆w ∆ w
2
+ EJ (− a 22 + a 44 ) +
L ∆x ∆x 2
(b11 − 4b12 + 2b13 − 4b14 + 3b22 − 4b23 + 6b24 + b33 − 4b34 + 3b44 ) ∆w ∆ w3 +
3
EJ
+
12 ∆x ∆x
(− b12 + b23 + b14 − b34 )w ∆ w4 +
4
EJ
+
12 ∆x
2
(b11 − 4b12 − 2b13 + 4b14 + 4b22 + 4b23 − 8b24 + b33 − 4b34 + 4b44 ) ∆ w2  +
2
EJ
+
16  ∆x 
+ 0(L )
To fulfill the consistency condition the following equations are to be satisfied:
 1 − 2 1 a 22  0
− 1 0 1 a  = 0
   24    .
 1 2 1 a 44  4
108 W. Gilewski
1 0 0 −2 0 1 0 0 0  b11   0 
− 1 b   
 1 0 0 −1 1 1 − 1  2  12   0 
1 −2 1 2 −2 1 −2 − 2 0 b22   − 1 
     
1 −2 0 −2 2 1 2 0 − 2 b13   0 
− 1 3 −2 0 1 1 − 1 0 − 3 b23  = − 2b44
     
− 1 3 0 0 −3 1 3 0 − 3 b33   0 
1 −4 3 2 − 4 1 − 4 6 − 4 b   − 3
   14   
0 −1 0 0 1 0 1 0 − 1 b24   0 
     
1 −4 4 − 2 4 1 4 − 8 − 4 b34  − 4
There are 3 equations for 3 parameters “a”. We have

a 22 = a 24 = a 44 = 1 .
It is seen that the basis matrix of Bernoulli beam template is independent on the
free parameters.
There are 9 equations for 10 parameters “b”. Thus
b44 = β , b11 = b12 = b22 = b13 = b23 = b33 = b14 = b24 = b34 = b44 = β .
3. TIMOSHENKO BEAM TEMPLATE
Let us consider a two-noded element of the length L with natural d.o.f.

q e = {wi , φ i , w k , φ k } . Timoshenko beam template proposed by Felippa [2] is more
complex than for Bernoulli beam and depends on 3 parameters:
K Template
Bernoulli
= Kb + Kh =
0 0 0 0  4 2L −4 2L 
   − 2 L ψL2 
αEJ 0 1 0 − 1 EJ  2 L ψL
2
= + 3β 3 .
L 0 0 0 0 L − 4 − 2 L 4 − 2L
   
0 − 1  2 L ψL − 2 L ψL2 
2
0 1
Following a similar procedure like for the Bernoulli beam we have:

• Density of strain energy
2 2
~  dφ   dw 
2 E s = EJ   + H φ − 
 dx   dx 
• Nodal parameters
2 3
∆w L 1 ∆2 w L 1 ∆ w
3
L
wi = w( x) − ( x) + ( x )  − ( x)  + ...,
∆x 2 2 (∆x) 2
 2  6 (∆x)
3
2
2 3
∆w L 1 ∆2 w L 1 ∆3 w L
wk = w( x) + ( x) + ( x )  + ( x)  + ...,
∆x 2 2 (∆x) 2
2 6 (∆x) 3
2
2 4
∆φ L 1 ∆2φ L 1 ∆3φ  L
φ i = φ ( x) − ( x) + ( x )  − ( x)  + ...,
∆x 2 2 (∆x) 2
2 6 (∆x) 3
2
2 4
∆φ L 1 ∆2φ L 1 ∆3φ L
φ k = φ ( x) + ( x) + ( x )  + ( x)  + ... .
∆x 2 2 (∆x) 2
2 6 (∆x) 3
2
• Density of the template strain energy
2 2
~  ∆φ  12 β EJ  ∆w 
2 E sTemplate = αEJ   + φ −  +
 ∆x  ∆x 
2
L 
∆2φ  ∆w  ∆3 w  ∆w 
+ β EJ 3 2 ψφ −  − β EJ 3 φ −  + 0( L )
2
∆x  ∆x  ∆x 3  ∆x 
To fulfill the condition of consistency it is necessary to take α = 1 and β in the

following form
HL2
β= (1 + ...)
12 EJ
with any ψ .
110 W. Gilewski
Let us propose more general form of the stiffness decomposition:
0 0 0 0   4b11 2 Lb12 − 4b13 2 Lb14 

0 a 
0 − a 24  H  2 Lb12 L2 b22 − 2 Lb23 L2 b24 
EJ 
K Extended = 22
+
 0  L − 4b13 − 2 Lb23 − 2 Lb34 
Bernoulli
L 0 0 0 4b33
   
0 − a 24 0 a 44   2 Lb14 L2 b24 − 2 Lb34 L2 b144 
• Density of the extended template strain energy
~ φ2 φ ∆φ
2 E sExtended = EJ (a 22 − 2a 24 + a 44 ) 2 + EJ (− a 22 + a 44 ) +
L L ∆x
2
(a 22 + 2a 24 + a 44 ) ∆φ  + H (b11 − 6b13 + 3b33 ) w2 +

2
EJ
+
4  ∆x  L
φw w ∆w
+ H (3b12 − 3b23 + 3b14 − 3b34 ) + H (− b11 + 3b33 ) +
L L ∆x
H ∆φ
+ (3b22 + 6b24 + 3b44 )φ 2 + H (− 3b12 + 3b23 + 3b14 − 3b34 )w +
4 ∆x
H ∆w H
+ (− 3b12 − 3b23 − 3b14 − 3b34 )φ + (b11 + 6b13 + 3b33 ) + 0( L)
2 ∆x 4
To fulfill the consistency requirement 3 equations for 3 coefficients “a” and 7

equations for 10 coefficients “b” are to be satisfied:
 1 − 2 1 a 22  0
− 1 0 1 a  = 0
   24   
 1 2 1 a 44  4
This matrix equation is equivalent to the conditions received for the Bernoulli
beam template in the previous chapter.
1 0 0 − 6 0 3 0   b11   0 0 0 0
0  
 3 0 0 − 3 0 3  b12   0 3 0  0
 
− 1 0 0 0 0 3 0  b22   0 0 0 b24  0
      b  +  
0 0 3 0 0 0 0  b13  = − 6 0 − 3  34  4
 0 −3 0 0 3 0 3  b23   0 3 0 b44  0
       
 0 − 3 0 0 − 3 0 − 3 b33   0 3 0  2
1
 0 0 6 0 3 0  b34   0 0 0   4
 
One can receive a 22 = a 24 = a 44 = 1 .
There are 3 independent constants in the group “b”. If we put
b 24 = C 1 , b 34 = C 2 , b 44 = C 3 the other coefficients are the following:
1
b11 = 1, b12 = −(1 + C 2 ), b22 = 1 (4 − 6C1 − 3C 3 ),
3 3
1 1 1
b13 = , b23 = − (1 + C 2 ), b33 = , b34 = C 2 .
3 3 3
The derivation above is a proof that there are 3 independent coefficients in the
Timoshenko beam template.
4. CONCLUSIONS
The finite element templates, proposed by Felippa [1,2], are algebraic forms that
contains some free parameters. The finite element templates for Bernoulli beam
and Timoshenko beam are examined in the paper, from the point of view of the
consistency condition. A new way of template construction is proposed. It is
confirmed that there is 1 free parameter for Bernoulli beam template and 3 free
parameters for Timoshenko beam template. The new method of creation the
template can be extended for 2D and 3D problems.
112 W. Gilewski
References
1. Felippa C.A., A template tutorial I: panels, families, clones, winners and losers.
Report No. CU-CAS-03-03, University of Colorado, 2003
2. Felippa C.A., The amusing history of shear flexible beam elements.
Report No. CU-CAS-05-01, University of Colorado, 2005

3. Gilewski W., Correctness of plate bending finite element with physical shape
functions, Finite Element News, 3, 1993, pp.29-34
4. Gilewski W., Evaluation of finite elements.
16th Int.Conf. on Computer Methods in Mechanics CMM-2005, Garstecki A., Mochancki B.,
Szczygiol N. Eds., Częstochowa 2005
5. Gilewski W., On the Criteria for Evaluation of Finite Elements – From Timoshenko Beam to
Hencky-Bolle Plate (in Polish), OW Politechniki Warszawskiej, Warsaw 2005

Evaluation of the inelastic demand of structures subjected to

multiple ground motions
Mihail Iancovici1,2, Georgiana Ionică1
1
Department of Mechanics, Statics and Dynamics of Structures, Technical University of Civil
Engineering (UTCB), Bucharest, 020396, Romania
2
National Center for Seismic Risk Reduction (NCSRR), Bucharest, 021652, Romania
Summary
In the current seismic design format, the key issue in establishing realistic seismic
loads is the behavior factor. It accounts for all the dissipative mechanisms that a
structural system may develop under a strong ground motion, however not clearly
enough stated yet. It corresponds to the performance level associated to the
ultimate limit state (i.e. life safety), related to a 100 years mean return interval of
earthquake ground motion with a prescribed peak acceleration of ground.
The paper investigates the effect of repeated Vrancea strong ground motions on the
behavior factors and the related parameters that accounts for cyclic structural
deterioration due to inelastic response. A large number of integrated analyses,
nonlinear response analyses and energy balance-based analyses were carried out
and estimates were made on the behavior factors for inelastic SDOF systems
controlled by flexure with stiffness degradation. The correlation between behavior
factors and damage level are investigated, using the Bozorgnia and Bertero
(2001), improved damage index. It is shown that multiple ground motion of
Vrancea type for Bucharest, may lead to an important increase of force and drift
demand of structures that usually is not taken into account.
KEYWORDS: multiple earthquake ground motions, behavior factor, hysteretic
energy, damage index, artificial accelerogram
114 M.Iancovici, G.Ionică
1. INTRODUCTION
Romanian territory and neighboring countries are repeatedly exposed to medium to

high intensity earthquake ground motions generated from the same source, located
in Vrancea region. Bucharest is one of the most exposed cities to damage to
buildings and human losses as well. It is therefore obviously needed to explore the
effects of repeated Vrancea strong ground motion and the related implications that
may improve the seismic design of new buildings and the evaluation procedures of
the existing ones. In the current Romanian seismic design format, based on strength
principles, the key issue in establishing realistic seismic loads that account for the
actual inelastic response, is the force reduction factor/behavior factor, namely q.
Basically it accounts for all the dissipative mechanisms that a structural system
may develop under a strong ground motion, however not clearly enough stated yet.
It is recognized however that the complexity of inelastic behavior phenomena
cannot be reproduced through a single parameter that is intended to fully describe
the actual structural response. q factor is primarily related to the structural inelastic
response (ductility and cumulative effects of repeated cycles of inelastic
deformations) and contains some of the ground motions properties.
In most of the seismic design codes and Romanian as well, q factor primarily
addresses to the selected structural type and includes the effect of inelastic behavior
and the over-strength effect. It does not directly account for the influence of strong
motion duration nor for the hysteretic behavior of the structural elements. It
corresponds to the ultimate limit state performance level (i.e. life safety), related to
a 100 years mean return interval of a prescribed peak ground acceleration (PGA).
For repeated earthquake ground motions however, there is no clear evidence on
how this important factor might be interpreted and used in analyses.
The purpose of the paper is to study the effect of repeated strong ground motions of
Vrancea type, on the behavior factors of buildings located in Bucharest; we study
the variability of q factor and related parameters on structural vibration period and
ductility, and on ground motion parameters as well.
2. BEHAVIOR FACTORS FOR SINGLE INPUT GROUND

MOTIONS; CYCLING LOADING EFFECT
Currently it is usual to estimate the actual force demand by dividing the base shear
force that corresponds to a fully elastic response by the behavior factor.
Early studies revealed the fact that the equal displacement assumption and the
equal energy assumption provide a fairly good estimation of the force reduction
factors at long and short periods, respectively. These developments accounts for
ductility properties only. A study by Newmark and Hall (1973) using 10 ground
motions records of 1940 El Centro earthquake, proposed reduction factors that
include the effect of both, ground motion and structural properties. Nassar and
Krawinkler (1991), Miranda and Bertero (1994), Watanabe and Kawashima (2002)
have been conducted studies on the force reduction factors that give fairly good
estimates on q factors especially for routine buildings. Ordan et al., 1998 and
Arroyo et al., 2003 found that the value of q strongly depends on ductility and
natural vibration period, and is significantly influenced by the soil type.
Inelastic behavior depends on many parameters, associated to the excitation and the
structural system. In order to uniformly grasp the effect of multiple earthquake
ground motions, and for the sake of clarity we used spectral representations of
SDOF systems response, having 5% damping ratio, and bilinear restoring force
characteristics, with stiffness degradation and 10% post-yielding stiffness ratio; all
systems have equal displacement ductility. For flexible structures however, SDOF
models are expected to reproduce with some degree of inaccuracy the actual
response. For routine buildings we expect however realistic estimates.
By definition, the behavior factor is given by
Fel (ξ el , T )
q= (1)
Fy ( µ , ξ nl , T )
where Fel and Fy are the maximum linear and nonlinear base shear force
respectively, µ is the displacement ductility factor, ξel and ξnl are the damping ratios
in the linear and nonlinear behavior range respectively, and T is the vibration
period of the model. By simplicity, usually ξnl is taken same as ξel.
We first compared the formulations of Newmark and Hall (1973) and Miranda and
Bertero (1994), with q factors obtained from analysis for low and high
displacement ductility systems, using the NS component accelerograms recorded
during 1977, 1986 and 1990 at the INCERC Bucharest station (fig.1).
3 µ=2 10 µ=6
9 New mark&Hall
2.5 Miranda&Bertero
8 analysis
7
2
6
q,Rµ
5
q,Rµ
1.5
4
1 3
New mark&Hall
Miranda&Bertero 2
0.5
analysis 1
0 0
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
3 10
New mark&Hall
µ=2 9 µ=6 Miranda&Bertero
2.5 8 analysis
7
2 6
q,R µ
5
q,R µ
1.5
4
1 3
New mark&Hall 2
0.5 Miranda&Bertero
1
analysis
0
0
0 1 2 3 4 5
0 1 2 3 4 5
T, s T,s
3.5 µ=2 9 µ=6

New mark&Hall
8
3 Miranda&Bertero
7 analysis
2.5
6
2 5
q ,R µ
q,R µ
1.5 4
3
1 New mark&Hall
2
Miranda&Bertero
0.5 1
analysis
0 0
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 1. q factors for VN77NS, VN86NS, VN90NS records (INCERC Bucharest station)
From the above plots can be observed that q tends to one as T approaches zero.
Proposed relationships fairly estimate the analysis results, especially for short-
medium vibration period and for low-medium displacement ductility; for flexible
structures with high ductility, the results are grossly overestimated, especially in
the case of 1977 strong ground motion. On the other hand, a large variability of q
in terms of T and µ can be observed. The variability of spectral response was more
detailed investigated, using a number of 15 accelerograms of 1986 Vrancea
earthquake, recorded in Bucharest (fig.2).
0.6 0.5
duct2 duct3 duct4
0.5 duct5 duct6

0.4
0.4
0.3
COVSA
COVq
0.3
0.2
0.2
duct1 duct2 duct3 0.1
0.1 duct4 duct5 duct6
0 0
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 2. COV of SA and q factors for 1986 EQGM (15 records, Bucharest)
The results are showing a larger variability corresponding to high ductility in small
vibration period region. As for the q factor, the scattering is much more
pronounced; the variability is lower for low ductility structures.
From the definition, one of the major disadvantage of q factor is that does not
account for the effect of hysteretic demand, as a powerful damage indicator (Uang
et al., 1990; Iancovici, 2005). The mass normalized hysteretic energy is given by
t
EH (t ) = (1 − β k )ω 2 z (τ ) x& (τ )dτ ∫
0
(2)
where, βk is the ratio of pre- and post-yielding stiffness, ω is the natural circular
frequency and z(t) is the nonlinear (hysteretic) displacement. For the sake of
investigating the patterns of q factors and hysteretic energies, we plotted their mean
plus one standard deviation values on the same graph (fig.3).
4
µ=2 0.12 18
µ=6 0.1
q q EH
3.5 EH 16 0.09
0.1
14 0.08
3
0.07
0.08 12
EH(m^2/s^2)
EH(m^2/s^2)
2.5 0.06
10
2 0.06 0.05
q
q
8
1.5 0.04
0.04 6
0.03
1 4 0.02
0.02
0.5 2 0.01
0 0 0 0
0.1 0.8 1.5 2.2 2.9 3.6 4.3 5 0.1 0.8 1.5 2.2 2.9 3.6 4.3 5
T,s T,s
Figure 3. q factor and hysteretic energy spectral representations (15 records, Bucharest)
As suggested, the variation pattern differs considerable; generally q factor could

not correctly reproduce the hysteretic energy distribution over the whole vibration
periods range. The coefficients of variation for the input energy that the structure
will receive and the hysteretic energy that the structure will absorb are plotted in
fig. 4.
0.8 0.9
0.7 0.8
0.6 0.7
0.6
0.5
COV EH
COVEI
0.5
0.4
0.4
0.3
0.3
0.2
duct2 duct3 duct4 0.2 duct2 duct3 duct4
0.1 duct5 duct6 0.1 duct5 duct6
0 0
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 4 . COV of input and hysteretic energies 1986 EQGM (15 records, Bucharest)
A high coefficient of variation can be observed for both parameters. For short
natural periods range the variation is higher for low ductility; for flexible systems,
the ductility effect on COV nearly vanishes.
3. MULTIPLE GROUND MOTIONS EFFECT ON STRUCTURAL

RESPONSE
The multiple input ground motion effect was introduced by a set of two and three
accelerograms respectively. The effect of longer duration motions was removed
from analyses by considering 40 seconds relaxation time intervals between
excitations. We chose for our purpose again the NS component accelerograms
recorded at the same site, INCERC Bucharest station during 1977, 1986 and 1990
earthquakes and the generated pulses are shown below.
2
ga1,m/s2
-2
0 5 10 15 20 25 30 35 40 45
t,s
2
ga2,m/s2
-2
0 20 40 60 80 100 120 140
t,s
2
ga3,m/s2
-2
0 50 100 150 200 250
t,s
Figure 5. Multiple input motions of 1977, 1986 and 1990 EQGM, NS components
(INCERC Bucharest station)
The spectral representation corresponding to low and high displacement ductility

show that there is no sensitivity on q factors for the considered multiple input
motions.
µ=2 8 µ=6
3. 5
7
3
6
2.5
5
2
4
q
q
1.5
3
1 sm1 sm1
2
sm12 sm12
0.5 sm123 1 sm123
0 0
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 6. q factors for 1977, 1986 and 1990 EQGM, NS components (INCERC Bucharest
station)
Taking the VN77NS record as reference and representing the ratios of the
corresponding hysteretic energies (fig. 7), it can be observed that the variation is
pronounced for SM12 and has almost doubled values for SM123, especially in the
case of low-medium vibration periods. This fact is not reproduced by the behavior
factors ratios.
1.6
duct2 duct3 duct4
duct5 duct6
1.4
EH2/EH1
1.2
1
0 1 2 3 4 5
T,s
2.2
duct2 duct3 duct4
duct5 duct6
2
1.8
EH3/EH1
1.6
1.4
1.2
1
0 1 2 3 4 5
T,s
Figure 7. Hysteretic energy ratios for 1977, 1986 and 1990 EQGM, NS components
(INCERC Bucharest station)
It is therefore desirable to relate the behavior factors with damage. Earthquake

structural damage may be expressed as a contribution of excessive inelastic
deformation and cyclic reversal loading effect. For the purpose of studying the
correlation that might exist between behavior factors and damage, we used an
improved damage model developed by Bozorgnia and Bertero, 2001. This
addresses to inelastic SDOF systems and tends to eliminate the disadvantages of
the well known Park and Ang damage index (Williams et al., 1995; Mehanny et
al., 2000).
By definition, the Bozorgnia and Bertero damage index is given by
µ − µe EH
DI = (1 − α 1 ) + α1 (3)
µ mon − 1 EH mon
where µ is the displacement ductility, µe is the ratio of maximum elastic portion of

deformation to yield displacement, µmon is the monotonic displacement ductility
capacity, EHmon is the hysteretic energy capacity under monotonic load and 0<α1
<1.0 is a constant. The associated damage index ratios are represented in fig. 8.
1.30 1.70
duct 2 duct3 duct4 duct 2 duct3 duct4
duct5 duct6 duct5 duct6
1.25 1.60
1.50
1.20
1.40
DI3/DI1
DI2/DI1
1.15
1.30
1.10
1.20
1.05 1.10
1.00 1.00
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 8. Damage index ratios for 1977, 1986 and 1990 EQGM, NS components (INCERC
Bucharest station)
The spectral distribution of damage shows attenuation with the vibration period,
with small influence of ductility. However the increase in damage ratio is up to
25% for SM12 and up to 60% for SM123.
There are a very limited number of records that can serve to our purpose, able to
drive the structure up to the desired performance level.
We chose then to simulate a large number of artificial strong ground motion
accelerograms which realistically aim to reproduce a prescribed structural response
and have similar amplitude, frequency content and duration as the real ones,
recorded in Bucharest on soft soil conditions (fig. 9).
Artificial ground motions were simulated using the procedure described by
Gasparini et al. 1976), who’s response spectra match the elastic acceleration
response spectra, given in the Romanian seismic design code P100-2006.
Sokolov and Bonjer (2006) proposed a procedure to estimate the ground motion
parameters based on site-dependant Fourier spectra, obtained by attenuation
relationships. Based on this procedure, estimates of peak ground accelerations, for
1940 event in Bucharest, were made as 0.24g. Comparisons with recorded
accelerograms during 1977, 1986 and 1990 events were proven the availability of
the proposed technique.
Based on these estimates, we generated three sets of 20 accelerograms compatible
with the elastic response acceleration spectra for Bucharest, having peak ground
accelerations of 0.24g (first and last) and 0.20g (middle). Since the problem of
generating artificial ground motions has many issues, our simulation intends to
reproduce the possible succession of events that occurred and may occur in the
future, primarily focusing on the structural response and ground motion

characteristics from engineering point of view (fig. 9)
0.8 Bucharest 1.4 Bucharest

PGA=0.24g PGA = 0.24g
0.7 1.2
0.6
1
norm P S D (1/Hz )
0.5
0.8
SA(T)/g
Bucharest
0.4
PGA=0.20g
0.6
0.3
0.2 0.4
0.1 0.2
0 0
0 1 2 3 4 0 1 2 3 4 5
T,s f, Hz
Figure 9. Code spectra compatible accelerograms and corresponding normalized PSD

functions
Typical simulated accelerograms are presented below (fig.10).
5
2
ga1,m/s
-5
0 5 10 15 20 25 30 35 40
5
t,s
2
ga2,m/s
-5
0 20 40 60 80 100 12 0
5
t,s
2
ga3,m/s
-5
0 20 40 60 80 1 00 12 0 14 0 1 60 1 80 20 0
t,s
Figure 10. Typical code spectra compatible simulated accelerograms
By analyses, means plus one standard deviation values of q factors were obtained
and are represented in fig.11.
4 µ=2 14 µ=6
3.5 12
3
10
2.5
8
2
q
6
1.5
sm1 4 sm1
1
sm12 sm12
0.5 sm123
2
sm123
0 0
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 11. q factors (means plus one standard deviation), artificial accelerograms
The results are showing a general decreasing trend of the q factor; a slight decrease
of q factor corresponding to SM12 and a more pronounced decrease in the case of
SM123, especially in the case of flexible structures.
1.1 1.1
1.0
q2/q1
q3/q1
1.0 0.9
0.8
duct2 duct3 duct4 duct2 duct3 duct4
duct5 duct6 duct5 duct6
0.9 0.7
0 1 2 3 4 5 0 1 2 3 4 5
T,s T,s
Figure 12. q factors ratios
Similarly, we computed the ratios of means plus a standard deviation of the

behavior factors, corresponding to SM12 and SM123, taking as reference those
corresponding to SM1 (fig.12).
One can be observed that generally, in the case of two input motions, the decrease
of behavior factor goes up to 6% and up to 21%, in the case of three input motions,
for flexible structures having low ductility. In the range of low-rise buildings, the
variation is rather uniform and goes up to 2-3% for two input motions and up to
12% for three input motions. Consequently the decrease of q will generate an
increase in the force and drift demands, that the structure should be able to supply.
Slight increases of q can be observed in the case of high-rise buildings with low
ductility.
4. CONCLUSIONS
The behavior factors cannot reproduce the effect of hysteretic energy demand. We
have shown that for a typical multiple strong motions recorded at the same site,
INCERC Bucharest station during 1977, 1986 and 1990 Vrancea earthquakes, there
is no variation on behaviour factors. However, in the same time, the hysteretic
energy demand goes up to 40% and 80% respectively. Same tendency was
observed in the case of associated damage indices.
For a large number of artificially generated accelerograms, the effect of repeated
ground motions on q factor might be considerable, especially in the case of three
input motions, for the flexible structures having low ductility. Consequently, the
base shear demand would increase considerably. This could be the case of some
reinforced concrete buildings in Bucharest, designed in the pre-code or low code
era, with low inelastic deformation capacity and low hysteretic energy
performance.
These are however estimates that might be reliable for routine buildings. For
flexible buildings however changes are expected due to the higher modes effect
and the effect of over-strength.
Amadio et al., 2002 suggested that the reduction on q factor is even larger than for
SDOF systems, in the case of steel moment resisting frames. A seismic design
procedure that does not take into account the cumulative inelastic deformation
demand that a structure will likely undergo during severe ground motion could lead
to unreliable performance.
References
1. Amadio, C., Fragiacomo, M., Rajgelj, S. , The effects of repeated earthquake ground motions on
the non-linear response of SDOF systems, Earthquake Engineering and Structural Dynamics,
32:291-308, 2003
2. Gasparini, D.A., Vanmarcke, E.H., Simulated Earthquake Motions Compatible With Prescribed
Response Spectra, Department of Civil Engineering, Research Report R76-4, Masschusetts
Institute of Technology, Cambridge, Massachusetts, 1976
3. Iancovici, M. , Evaluarea performantei structurale a cladirilor de beton armat, Ph.D. thesis,
Technical University of Civil Engineering Bucharest (in Romanian), 2005
4. Miranda, E. Bertero, V., Evaluation of strength reduction factors for earthquake resistant design,
Earthquake Spectra, 10(2), 357-379, 1994
5. Nassar, A. A., Krawinkler, H., Seismic demands for SDOF and MDOF systems, Report No. 95,
The John A. Blume Earthquake Engineering Center, Stanford University, California, USA, 1991
6. Newmark, N. M., Hall, W. J., Seismic design criteria for nuclear reactor facilities, Report No. 46,
Building Practices for Disaster Mitigation, National Bureau of Standards, U.S. Department of
Commerce, 209-236, 1973
7. Sokolov, V., Bonjer, K-P., Modeling of distribution of ground motion parameters during strong
Vrancea (Romania) earthquakes, 1st ECCES, Geneva, 2006
8. Terán-Gilmore, A., Jirsa J., The concept of cumulative ductility strength spectra and its use within
performance-based seismic design, ISET Journal of Earthquake Technology, Paper No. 446, Vol.
41, No. 1, pp. 183-200, 2004
9. Uang, C.M. and Bertero, V.V., Evaluation of Seismic Energy in Structures, Earthquake
Engineering and Structural Dynamics, Vol. 19, pp. 77-90, 1990
10. Watanabe, G., Kawashima, K., An Evaluation of the Force Reduction Factor in the Force-Based
Seismic Design, Tokyo Institute of Technology, O-Okayama, Meguro, Tokyo, Japan, 152-8552,
2002
11. Romanian Seismic Design Code P100/2006, Ministry of Transport, Construction and Tourism,
Romania, 2006
Databases- business support for the Civil Engineering

Irimia Roxana-Adina1, Vlăsceanu Alina Nicoleta1
2
Department of Computer Science, Christian University”Dimitrie Cantemir”, Bucharest, Romania
Summary
The current changes in modern society and the evolution of the market, defined by
the generalization of the Internet-Intranet-Extranet technologies, produce
significant changes in the management of large databases and new ways of
exploring the huge potential of information provided by Internet interactions must
be found. Companies that need to develop predictive models using business
expertise use the concepts of Data Mining and Web mining into business
operations to improve decision making. In addition, Data mining software allows
them to find relationships and patterns in raw data and use it in an automated
decision supports system; they also can extract knowledge from real-time and
historical data and predict outcomes of future situations.
The actual development trend regarding the Databases Management Systems must
focus on:
-using intensively the integration of the Information Systems;
-obtaining open applications;
-ensuring interoperability and communication for the Information Systems;
-focusing on the global performances;
- Databases Management Systems must become platforms for the databases.
The technology of data bases have interfered with others Computer Science-based
technologies, generating new types of hybrids DBMS, such as evolved (advanced)
data bases.
Related with the Civil Engineering field, it can be mentioned the spatial databases
(which have resulted by combining the technology of data bases with geographical
systems and assisted design) and the multimedia databases (which store and
process classical data-text and graphics and also multimedia data-images, audio
and video).
KEYWORDS: spatial databases, multimedia databases Data Mining, Internet, digital

economy.
1. INTRODUCTION
One of the factors that stimulated the development of database technology was the
need to guarantee the consistency of the stored data. On the other hand, the demand
to process more complex data with more semantic content have also lead to a better
understanding of the properties of the data and motivated the evolution of the Data
Base Management Systems (DBMS) toward the inclusion of facilities to handle
such properties.
The handling of data properties within the context of a database engine faces two
main difficulties: 1) Properties are hard to identify, especially those obvious in the
actual world; 2) Data property presentation varies during the process of design of
the software.
The proper requirements elicitation of a software product is a key factor in the
success of the whole software development process. However, these requirements
are not easy to deal with. They have different intrinsic natures and they may appear
showing different faces. In many cases, some requirements are totally or partially
hidden in the information collected by the software developers.
Pragmatic issues such as the degree of adhesion of a given DBMS or others related
with performance may introduce additional mutations. The understanding of the
evolution from actual world data properties to database world integrity restrictions
is another key factor in the success of the software artifact during its life cycle.
This understanding must be accompanied with a well defined traceability
mechanism, which must clearly identify the actual world data property behind
every database integrity restriction. Traceability is also a key factor especially
during software maintenance activities.
Not every data property must be modeled, but each property must be looked at
carefully to see if modeling is needed in the context of the scope and in the
objective of the software artifact. A more analytical approach may order the data
properties by taking into account their importance.
2. EVOLUTION OF DATA PROPERTIES
Figure 1 shows all possible mappings from data properties in the actual world to
database integrity restrictions. It should be noticed that there are neither arrows
from Business Rules to Object Restrictions, nor from Attribute Properties to
General Rules. These mappings are extremely hard to find. Also, some other
mappings shown in Figure 1 occur very seldom, while others are common.
126 Irimia R., Vlăsceanu A.
Figure 1. Mappings data properties into integrity restrictions
Attribute Properties. Database Engines have a set of built-in data types whose main
purpose is to deal with the Attribute Properties of the data to be stored. These data
types are useful and have been used for decades, helping users to take care of their
data processing needs.
Relationship Properties. The connection among different attributes is the source of
most of the data processing richness and problems. These connections have a scope
larger than Attribute Properties since they involve several attributes, at least two,
usually belonging to different objects or entities.
Business Rules. Every organization restrains behavior in some way. This is
strongly associated to constraints that define which data may (or may not) be
actualized. A Business Rule is a declarative sentence usually describing a correct
state of a piece of data.
3. METHODS TO ORGANIZE THE DATA AND DBMS

(DATA BASE MANAGEMENT SYSTEMS)
As mentioned above the need to guarantee the integrity of the stored data was and
still is one of the key factors driving the evolution of the Data Base Management
Systems (DBMS), creating different Database paradigms. The following
paragraphs give a brief summary about how most popular paradigms see the
Database integrity problem.
Relational Databases. A rich theory for constraints has emerged in relational
databases, mainly based on a fundamental class of constraints called dependencies.
Its main motivation is to incorporate more semantics into the relational model. The
term relational model has thus come to refer to the broad class of database models
that have relations as the data structure and that incorporate some or all of the
query capabilities, update capabilities, and integrity constraints. Many ideas of
other database paradigms already in the scenario (and in their own evolution
process) have been borrowed by researchers and vendors, and introduced in current
relational products. This fact has created types of incomplete and extended
relational engines, which started to be known as post relational databases. They are
incomplete because not all the prescribed issues for the relational model are
satisfied, and extended because they go further in some other areas such as active
characteristics.
Active Databases. Supporting reactive behavior implies that a database
management system may be viewed from a production rule system perspective.
These production rules are well known today, in database terminology, as active
rules or simply triggers. It is a form of computation, which is motivated by the
occurrence of some event, typically a database operation, executing a reaction to
that stimulus. Active rules may pose queries to the database to collect information
about events and database objects and decide whether events require an action.
Then they may execute actions, normally any sequence of database operations.
Reactive behavior is seen as an interesting and practical way to check for
satisfaction and enforcement of integrity constraints. Nevertheless, integrity
constraint maintenance, materialized view maintenance (especially useful in the
warehousing area) and implementation of Business Rules are not the only areas of
application of data repositories with reactive behavior. Other interesting
applications areas are replication of data for audit purpose, data sampling,
workflow processing, scheduling, real-time applications and many others. In fact,
practically all products offered today in the marketplace support complex reactive
behavior on the client side.
Spatial Databases. Even though there is a very active research area interested in
the design of robust and efficient spatial databases, the inability of current GIS
regarding the implementation and management of spatial integrity constraints is
still evident. This is due to the fact that within the scope of geographic applications,
special problems come up due to the location aspects of data. A modification in a
spatial database may cause simultaneous updates in a large number of records in
multiple files, making it hard to manage the entire environment. A very
sophisticated control is required to avoid redundancy and loss of integrity.
Object-Relational Databases. Object-oriented literature typically uses the term
"relationship" to mean, specifically, relationships supported by foreign keys in a
relational system. Various levels of support for referential integrity have been
implemented in those systems. As in the relational paradigm, triggers are very
effective in supporting data integrity in a database, especially for dealing with
those restrictions that cannot be expressed declaratively. ORDBMSs demand a
system of triggers even more flexible than the relational one.
Object-Oriented Databases. Given that in an OO environment, objects are
collected into classes and relationships are established at the class level. A
relationship between classes denotes a set of relationships between the objects of

the respective classes. Relationships may have attributes. A model is always a
compromise to achieve the right amount of expressive power while keeping
simplicity and clarity. Domain Restrictions, keys, and referential integrity
constraints may all be used straight-forwardly in OODBMSs. Other kinds of
constraints are peculiar to OODBMSs, for instance: constraints of the migration of
objects between classes (roles); exclusivity constraints between classes; constraints
on the definitions of subclasses; and existence dependencies. The last ones are the
key to semantic integrity checking since they allow the designers to track and solve
inconsistencies in an object-oriented conceptual schema.
Distributed Databases. A distributed database managing system (DDBMS) is a
database management system (DBMS) that supports characteristics of a distributed
database, that is, the possibility of handling information contained in multiple
locations, and preserving data integrity at the same time. A DDB differs from a
centralized DB mainly in that data are placed in a number of locations instead of
being located in only one site. This characteristic of DDBs causes the control of
data integrity to be more complex than for centralized environments. Each
transaction can involve more than one location, and it is hard to keep an execution
order of the instructions of the transaction to preserve data integrity.
The strength of the SQL standards induces one to think that in the future the object
relational paradigm will continue to be dominant in the field of DBMS.
Improvements to check-and-trigger mechanisms are strongly needed to integrate
data properties in a more friendly way with the data itself. Development platforms
where data properties become a concern of the engine and the data integrity is fully
factorized will form application programs leading to an evolution and the
maintenance of less traumatic DBMS-based software artifacts. Naturally, this
process will require very active participation on the part of researchers and DBMS
producers.
4. SPATIAL DATABASES AND GIS-USEFULL

INSTRUMENTS FOR CIVIL ENGINEERING
A spatial database is a database that is optimized to store and query data related to
objects in space, including points, lines and polygons. While typical databases can
understand various numeric and character types of data, additional functionality
needs to be added for databases to process spatial data types. These are typically
called geometry or feature. The Open Geospatial Consortium created the Simple
Features specification and sets standards for adding spatial functionality to
database systems. Database systems use indexes to quickly look up values and the
way that most databases index data is not optimal for spatial queries. Instead,
spatial databases use a spatial index to speed up database operations. In addition to

typical SQL queries such as SELECT statements, spatial databases can perform a
wide variety of spatial operation. The following query types and many more are
supported by the Open Geospatial Consortium:
• Spatial Measurements: Finds the distance between points, polygon area, etc.
• Spatial Functions: Modify existing features to create new ones, for example
by providing a buffer around them, intersecting features, etc.
• Spatial Predicates: Allows true/false queries such as 'is there a residence
located within a mile of the area we are planning to build the landfill'
• Constructor Functions: Creates new features with an SQL query specifying
the vertices (points of nodes) which can make up lines.
• Observer Functions: Queries which return specific information about a feature
such as the location of the center of a circle.
In the category of Spatial Database Systems, it can be mentioned Oracle Spatial,
PostgreSQL DBMS (which uses the spatial extension PostGIS to implement the
standardized data type geometry and corresponding functions) and MySQL DBMS
(which implements the data type geometry plus some spatial functions that haven't
been implemented according to the OpenGIS specifications. Functions that test
spatial relationships are limited to working with master bounding rectangles rather
than the actual geometries. MySQL versions earlier than 5.0.16 only supported
spatial data in MyISAM tables. As of MySQL 5.0.16, InnoDB, NDB, BDB, and
ARCHIVE also support spatial features).
A geographic information system (GIS) is a system for capturing, storing,
analyzing and managing data and associated attributes which are spatially
referenced to the earth. In the strictest sense, it is a computer system capable of
integrating, storing, editing, analyzing, sharing, and displaying geographically-
referenced information. In a more generic sense, GIS is a tool that allows users to
create interactive queries (user created searches), analyze the spatial information,
edit data, maps, and present the results of all these operations. Geographic
information science is the science underlying the applications and systems, taught
in degree programs at many universities. Geographic information system
technology can be used for scientific investigations, resource management, asset
management, Environmental Impact Assessment, Urban planning, cartography,
criminology, history, sales, marketing, and route planning. For example, a GIS
might allow emergency planners to easily calculate emergency response times in
the event of a natural disaster, a GIS might be used to find wetlands that need
protection from pollution, or a GIS can be used by a company to find new potential
customers similar to the ones they already have and project sales due to expanding
into that market.
Civil Engineering projects involve the management, analysis and integration of
large amounts of geographic information to ensure success. This can include a
wide range of information such as detailed design drawings originating from CAD
solutions, detailed mapping, air photography, geological investigations, population

information, traffic flows and environmental models.
With increasingly complex projects, the interaction of often international
multidisciplinary functions requires the introduction of effective and efficient data
management and sharing solutions. GIS provides the civil engineer with tools for
creating, managing, analysing and visualising all types of geographic information.
GIS is now being used by Civil Engineers in every application domain and
throughout every aspect of the enterprise, be it on the desktop, in the field, or in
collaboration with others.
The Geographical Information Systems provides a rich set of databases and spatial
tools needed to manage civil information for design, modeling and maintenance.
Data is taken from many different sources and file formats and integrated into a
single application platform to support complex civil engineering workflows.
A geodatabase is a spatially enabled relational database that has the ability to
represent various types of geographic data, manage attributes associated with that
data, maintain spatial relationships (topology and networks), and keep track of
user-defined thematic layers and datasets. GIS typically supports three different
“views” of geographic information: the geodatabase view, the geovisualization
view, and the geoprocessing view.
Spatial relationships, such as topologies and networks, are also crucial parts of a
GIS database. Topology is employed to manage common boundaries between
features, define and enforce data integrity rules, and support topological queries
and navigation
Networks are used to describe a connected graph of GIS objects. The graph, or
network, can be used to model transportation pathways, pipelines, utilities,
hydrology, and many other network based applications.
5. MULTIMEDIA DATA BASES AND OPEN

COMUNICATIONS IN THE INTERNET ENVIRONMENT
Nowadays, we are witnessing an explosive growth in use of multiple media forms

(voice, data, images and video etc.) in varied application areas including
entertainment, communication, collaborative work, electronic commerce, and
Civil Engineering. The increasing computing power, integrated with multimedia
and telecommunication technologies, is bringing into reality the dream of real time,
virtually face-to-face interaction with collaborators sitting far away from us. In the
process of realizing the technological ambitions, it’s necessary to address a number
of technology, management and design issues and to be familiar with exciting
current applications. It is impossible to track the magnitude and breadth of changes
that the multimedia and communication technology is bringing daily to us in many

different ways throughout the world. The major multimedia networking areas
includes: Development and management of real time distributed multimedia
applications; Audio/video applications and streaming issues; Protocols and
technologies for building Internet multimedia applications; QOS (Quality of
Service) frameworks and implementation; Collaborative applications; Multimedia
synchronization in distributed environment; Multicasting technology and
applications; Use of mobile multimedia over wireless network.
The Internet has evolved from a provider of the simple TCP/IP best-effort service
to an emerging integrated service Internet. This development provides tremendous
opportunities for building real-time multimedia applications over Internet, although
they are often large and complex. The CORBA Media Streaming Framework and
Java Media Framework are the two emerging environments for implementing
Internet multimedia. They provide applications with a set of APIs to hide the
underlying details of Internet real-time protocols. The Internet multimedia
applications developed using these two frameworks are expected to appear in the
near future.
In addition to HTML (Hyper Text Markup Language) pages and web sites, a
number of new and more sophisticated technologies have been developed for
creating and developing Web pages:
• Dynamic HTML introduces movement and the ability to react to a user’s
actions to Web pages;
• JavaScript, Java, and ActiveX are technologies that provide a way of
performing complex tasks within a Web page and can even function as full-
featured programs;
• CGI scripts is a technology used on some Web servers to provide server-based
features such as visitor counters and discussion groups;
• XML (Extended Markup Language) provides a way of giving detailed content
information about a Web page, allowing for more meaningful searching and
information gathering.
In the last few years the rapidly growing Internet has pushed new multimedia
applications in the field of communication and Civil Engineering. The next step in
the information age is the mobile access to multimedia applications: everything,
everywhere any time, using data transmission for mobile multimedia applications
in wireless cellular networks.
Multimedia services are offered by service providers in the field of: broadcast (TV
and audio channels, editors); information on demand (video, audio, weather,
documents); communication (voice and video telephony); commerce (banking,
electronic commerce, publicity); and industry (collaborate work, VPN).
Multimedia has the intrinsic goal of improving the effectiveness of computer-to-
human interaction and, ultimately, of human-to-human communication. While the
human-computer interaction community looks predominantly at the application

layer and the telecommunications community at the lower end of the ISO/OSI
stack, little work has been published in bridging the gap between these two
communities. Indeed, the human element is often neglected in quality of service
negotiation protocols. Not only does this have a negative and undesirable impact
on the user's experience of multimedia, it also discards the potential for more
economical resource allocation strategies. With the proliferation of ubiquitous
multimedia in predominantly bandwidth-constrained environments, more research
is needed towards integrating and mapping perceptual considerations across the
protocol stack and building truly end-to-end multimedia communication solutions.
6. CONCLUSIONS
The new vision over the knowledge based society increases the importance of
knowledge, information and communication, presented in a multidimensional
form, all these creating the nucleus of humanity’s progress. The digital revolution
based on these new values has transformed the business and governmental
environment, revolutionizing the educational processes and generating new ways
of memorizing, storing, and sharing information and data to a global level, causing
an irreversible transformation in the social and individual level.
The new knowledge-determined society is inclusive, which means that any person,
without any kind of discrimination, has the right to create, receive and use
information and knowledge, no matter what media it being used and without border
restrictions. The new form of social organization is individually-centered, focusing
on citizens and communities as creators and beneficiaries of large- scaled
dissemination of information. This implies a restructuring of the social
organization in the sense of increasing social equity and democracy, mainly by
increasing the transparency offered by the wide access to information, resulting in
improving the quality of life. The new perspective, the openness of
communication, also implies new forms of solidarity and cooperation, together
with an increase of the degree of responsibility in using and spreading information.
Civil Engineering companies are recognizing the significant productivity gains to
be obtained through the introduction of project intranets/extranets for the
management, and publishing of shared project information. Geographic
information is at the heart of all engineering projects and is by many different
departments or organizations during any successful project. Geographic
information collected during the planning stage is then necessary for detailed
design, construction and maintenance phases of any project. The creation of a
geoportal to share centralized data allows all project team members to access the
relevant geographic information as and when required.
References
1. Jorge H. Doorn J., Database Technologies, Idea Group Inc, Argentina, 2005.
2. Shashi Shekhar and Sanjay Chawl, Spatial Databases, Prentice Hall, 2003.
3. Philippe Rigaux, Michel Scholl and Agnes Voisard, Spatial Databases - With Application to GIS,
2000
4. Mehdi Khosrow-Pour, Encyclopedia of Information Science and Technology, Idea Group
Publishing, 2005
5. Syed Mahbubur, RahmanIdea, Multimedia Networking: Technology, Management and
Applications, Group Publishing, 2002
6. Bruce Kogut, The Global Internet Economy, The MIT Press, 2003
7. Alan Simon and Steven Shaffer, Data Warehousing and Business Intelligence for e-Commerce,
Morgan Kaufmann Publishers, 2001.
8. Mahesh Raisinghani, Business Intelligence in the Digital Economy: Opportunities, Limitations,
and Risks, IGI Publishing, 2004.
9. IBM Framework for e-business Technology, Solution, and Design Overview, IBM Redbooks,
2001.
Foundation solutions for a marketing center in Cluj

Nicolae Boţi1, Irina Lungu2 and Ioan Boţi3
1
Department of Roads and Foundations, Technical University, Iaşi, 700050, Romania
2
3
Department of Geotechnical Eng., Technical University of Civil Eng., Bucureşti, 020396, Romania
Summary
The construction of the marketing center is set on the left bank of the river Somes,
in difficult site conditions. The underground water table increases from
approximately 2.4-3.5m deep to 1.0-1.8m from the ground surface.
The lithological succession consists of clayey soils with various thicknesses and a
significant swelling and contraction potential continued by silty sands, sand and
gravels up to 10m deep. The cohesive soil layers display a reduced consistency
index, less than 0.5, as for the fat clay, the void ratio may result in values up to 1.5.
In order to accommodate the site conditions three foundation solutions has been
discussed: open caissons, embedded 0.3m into the good foundation soil; drilled
short piles or precast driven piles inserted into the good foundation soil at lest 15d;
foundation beam networks onto an improved soil by a granular material cushion.
The technology to perform the granular compacted layer is set in order to allow
the heavy machines to enter the excavated area that transformed into a swamp. In
this respect, the principle of the reversed filter method is required, by laying coarse
materials at the bottom and finer materials on top.
The thickness of the granular cushion is set to prevent a deformation limit state at
the footing level as well as at the cushion base level. The effective pressure should
be less than the plastic pressure developed within the granular material first and
second, into the natural foundation soil.
The design of the cushion concluded that a 2m thickness will comply with the
restrictions set to prevent reaching a deformation limit state under the foundation
system.
The solution proved to be efficient in case of poor soil conditions such as
unconsolidated earth fillings, mud and loose sands, a non-uniform soil layering
which is specific to swampy areas, to prevent the frost sensitivity and to ensure the
presence of a drainage layer.
KEYWORDS: active clays, soil improvement, granular compacted cushion,

deformation limit state, equivalent layer method.
1. GEOTECHNICAL SURVEY
1.1. Geological and geomorphological data
The construction site is situated at the South border of the Somes’s Plateau limited
at North-East by Transilvania Field, on the left side terrace of the river Somes. The
geological eras are considered a necessary base for this survey due to the fact that a
texture is a result of the deposit conditions and the successions of the events. At
high depths there is a salt formation which, in its ascension generated the folding of
the strata into an anticline with West-East direction. Marls are found above the salt
deposits, which are visible on the Desmir hills. The quaternary is represented by
the alluvial deposits of the second terrace of the river Somes. With ages, the other
streams created a swamp in the area of interception with the river Somes.
1.2. Hydrogeological data
Within the three boreholes performed at the construction site, water is present in
sand and gravel strata, at depths between 2.4-3.5m and the level increased from
3.5m to 1.0m (F1), from 3.1m to 2.1m (F2) and from 2.4m to 1.8m (F3) from the
ground surface. Water is strongly sulphurous, with very low carbonide, very low
salts content.
The fluctuating underground water level influences the geotechnical category
together with the foundation solution: the excavation goes below the water table
and consequently gravitational drainage works are required, without any risks of
damaging the surrounding buildings.
1.3. Soil profile
Based on the performed boreholes and borehole logs, the soil profile consists of the
following layers:
• clayey soil, brown-black, 0.2m thickness (F1, F3) and 1.2m (F2);
• fat clay, black, with high plasticity, variable thickness 2.3m (F1), 1.2m (F2),
0.5m (F3) – γ = 16.41-17.83 kN/m3, Ic = 0.41-0.77, e = 1.06-1.49;
• grey clayey sand, with metal insertions, 1m thickness in F1 borehole only – γ
= 18.6 kN/m3, Ic = 0.33, e = 0.78;
• yellowish grey gravel, only in the F2 borehole, 3.1m thickness – γ = 20.36-
21.37 kN/m3;
• fat marly clays with fine gravel, 0.6m thickness in F3 borehole – γ = 21.74
kN/m3, Ic = 0.8, e = 0.84;
• brownish silty sands, 1.2m thickness – γ = 19 kN/m3, e = 0.59, Φ = 370;
136 N. Boţi, I. Lungu, I. Boţi
• fine gravel with sand, 3.5m (F1), 1.5m (F2), 2.0m (F3) thickness – γ =
22.8kN/m3;
• brownish gravel, 1m thickness in F1 borehole only – γ= 20 kN/m3;
• coarse gravel and sand, 2m thickness in F1 borehole only – γ = 22.8 kN/m3;
• yellowish-brown sand, below 10m depth all over the place – γ = 18.11 kN/m3,
e = 0.77
2. FOUNDATION SOLUTIONS
According to the site conditions previously presented the following foundation

solutions were considered appropriate:
• open caissons,
• short piles – drilled or precast driven,
• foundation beam network onto a granular material compacted cushion.
2.1. Open caissons
Caissons ensure the crossing of the soft soils, preferable used in river beds, since
the soil removal from the inside is easily done by a ship scraper and the developed
friction force is low. It is considered an efficient solution where piling is difficult
due to the presence of large boulders or other obstacles. It is important to maintain
the verticality of the caisson advance into the ground. By using hydro-mechanically
excavation under the water table in sands or silty sands, a hydraulic fracture of the
ground may develop immediately under the cutting edge of the caisson.
2.2. Short piles
The limited use of a pile foundation according to this site conditions can be
explained by the difficulties during piling – drilling or driving – and also by the
fact that such short piles result with a good behaviour during service when
designed to carry loads up to 250-300 kN with an optimum pile interdistance of 1.5
– 2.0m. A subevaluation of the pile bearing capacity can happen when using tables
for the point resistance and side friction in case of coarse sands, gravels, and high
consistency clays.
2.3. Foundation beam network onto a granular material compacted cushion
As a soil improvement method of the initial soft soil, a granular compacted cushion
is considered appropriate, the layering process beginning over an initial layer made
of crushed stones and boulders that is stable enough to ensure the access of the
heavy machines to start the compaction process.
The composition of the ballast cushion is dictated by: the nature and composition
of the soil, the groundwater table, and the construction design parameters [1]. In
case the cushion is built under the water table, on permeable soils the reverse filter
principle has to be applied to avoid hydraulic fracture of the soil underneath, by
layering a coarser material at the bottom and finer material on top.
3. DESIGN PRINCIPLES FOR THE GRANULAR COMPACTED

CUSHION
The foundation soil design subjected to a shallow foundation was performed using
the conventional pressure according to the Romanian STAS 3300/2-1985 of
350kPa. In figure 1, the cushion dimensions are presented in order to support a
shallow foundation: the thickness hp and the enlargement (the width of the guard
area) c with respect to the foundation edges, with a total of Bp under the footing.
Figure 1. The load transfer from the footing to the cushion and then to the foundation soil
The minimum thickness of a granular compacted cushion is considered to be 1.00m

and in case the cushion acts as a reverse filter, the cushion is designed over the
entire building and with a minimum thickness of 1.50m. In this hypothesis of
cushion’s pre-dimensioning, the pressures verification at the base of the cushion is
given by:
p0 = γ ⋅ D f + γ p ⋅ hp + α ⋅ p ≤ pconv (1)
138 N. Boţi, I. Lungu, I. Boţi
where: γ is the soil unit weight of the soil filling above the ground level, γp is the
unit weight of the granular material compacted into the cushion, Df the foundation
depth, hp the cushion’s thickness, p the pressure deliver from the footing, α the
distribution coefficient for the vertical normal stress along the footing axis.
The final verification of the cushion is made based on the isobar curves, with the
restriction that the cushion shall be set between the isobars of 0.2p of the vertical
normal stress σz and 0.3p of the horizontal normal stress σy, as well as for the shear
stress τ (figure 2) [2].
Figure 2. Isobar curves for a continuous flexible footing with the width b
The method of the equivalent stratum is applied by using Odemark’s relationship in

order to asses the distribution of the stress along the thickness of the cushion based
on the pressure delivered from the footing to the cushion (figure 3) [2], [3].
E1 ⋅ γ 2
hequivalent = α ⋅ h1 ⋅ n (2)
E2 ⋅ γ 1
where: γ1, γ2 are the unit weight of the two layers involved, α a coefficient
experimentally set, E1 and E2 are the corresponding moduli of the materials within
the two layers, h1 the thickness of the layer for which the equivalent thickness is
established. For the present case γ 1 ≅ γ 2 , α = 1 and n = 2.5 . The granular
material is considered to be a ballast with the following characteristics:
γ d = 20.5 − 21.5kN / m3 , wopt = 6 − 8% and a modulus E1 = 400daN / cm 2 . The
thickness of the cushion is set initially to 1.00m, a foundation depth of 1.50m and a
deformation modulus of the soil underneath E2 = 100daN / cm 2 .
Figure 3. The method of the equivalent stratum
The bearing capacity of the foundation soil is set as pconv = 350kPa, the pressure
delivered by the footing p = 250kPa, the distribution coefficient 0.55 and
consequently, the pressure at the base of the cushion p 0 = 137.50kPa. The resulted
equivalent thickness is set as hequivalent = 1.50m. This case is similar with the ones
mentioned in literature, where the ration between deformation modulus of the soil
is smaller 1.8-2.5 the cushion modulus, 150-200daN/cm2, the ground can be
considered homogeneous.
3. CONCLUSIONS
Based on the analysis performed on the foundation solutions, the foundation

system is recommended to consist of flexible continuous footing for the external
shear walls and a foundation beam network of the columns inside, all performed on
a compacted ballast cushion. The thickness is finally set at 2.0m to accommodate
poor soil conditions such as unconsolidated earth fillings, mud and loose sands, a
non-uniform soil layering which is specific to swampy areas, and to ensure the
presence of a drainage layer.
References
1. Mihu, A., Silion, T., Ciubotaru, V., Boţi, N., Muşat, V., Grecu, V., Beche, V., Sîrbu, G., Folosirea
pernelor de balast la fundarea unor construcţii de clădiri pe pământuri puternic compresibile şi
saturate cu apă, Revista nr. 1 Construcţii, INCERC Bucureşti, 1976. (in Romanian)
2. Boti, N., Silion, T., Raileanu, P., Cijevschi, Maria, Musat, V., Grecu, V., Stanciu, A., Nicuta,
Ana, Criterii de dimensionare a pernelor de pamant, Sesiunea Ştiinţifică Jubiliară „40 de ani de
Învăţământ Superior de Construcţii la Iaşi”, 1981. (in Romanian)
3. Pop, V., Popa, A., Maniu, I., Popa, A., Fundaţii de suprafaţă pe pernă de balast, Rotaprint IP Cluj
Napoca, 1978. (in Romanian)
The Expertise of Offshore Structures

Ludovic G. Kopenetz, Ferdinand-Zsongor Gobesz
Department of Structural Mechanics, Faculty of Civil Engineering and Building Services, Technical
University of Cluj-Napoca, Cluj-Napoca, 400020, Romania
Summary
The problem of energy in tight liaison with ecology is representing one of the
greatest problems of humanity. Considering an exponential broaden of marine
production in the following years, the international organizations have imposed the
adoption of some highly strict internal and international prescriptions in order to
avoid environmental accidents (pollution). In this context, the structural safety of
the offshore-type complex plants and equipments is quite actually.
The structural expertise must cover, beside the vastness of interfering random
variables, all-important aspects in order to avoid the pollution of the pelagic
environment.
The present paper wishes to bring a humble contribution to the clarification of the
aspects which are producing sensitive alterations of structural safety (quality of
material, global and local stability, fatigue, corrosion and erosion etc.), using
advanced and lifetime oriented analysis concepts.
KEYWORDS: offshore structure analysis safety stability assessment
1. INTRODUCTION
The behavior in situ of the steel offshore structures (figure 1) and the ecological
aspects are tightly bonded to structural safety. In this way, the structural expertise
must answer to the problems of structural safety [1], [2], [3]. At its turn, the
concept of structural safety is bundled to the design, construction and operation of
offshore structures, considering the extent of interfering random variables (wind
waves, earthquake, construction faults, corrosion, fatigue etc.).
The international standards concerning the conservancy of the planetary ocean
(covering approximately 70% of the Earth’s surface) are very severe [4]. In this
context the structural expertise of the offshore structures which are located in the
area of the coastal shelf of the Black Sea has a peculiar importance, considering the
fact that these structures were conceived and erected at the level of knowledge of
the years 1970-1980 [5].
Figure 1. Offshore structure types.
It is known that during the years of 1990 a series of such structures collapsed
worldwide, such as the Sleipner platform made from reinforced concrete) in the
North Sea (figure 2, 3), an offshore structure made from steel in the area of Brazil’s
territorial waters etc.
During construction Perspective view Section views

Figure 2. Illustrations of the Sleipner platform.
The following factors underlay the occurrence of such accidents:

- misconceived structural analysis aspects (even using the most advanced
computing programs and supercomputers);
- corrosion due to severe operating conditions in marine environment;
- imperfections during construction and assembling;
- great number of parameters which were hard to control;
142 L. G. Kopenetz, F. Zs. Gobesz
- post critical structural behavior;

- mismatches between lab-test results and the real situation.
Figure 3.
In this paper the authors square up structural analysis aspects which are required
for structural expertise in order to assess the level of safety of these kinds of
structures in situ.
2. LOAD EVALUATION ASPECTS [6], [7], [8]
The permanent and non-persistent actions will be considered both by their lasting
time and their time variation.
The evaluation aspects will relay upon a sternly hydro mechanical analysis (wave
theory, hydrostatics and hydrodynamics). In case of floating offshore structures,
the Archimedean modification of pushing forces will be considered
Model experiments and the simulation of some regular loadings on such structures
have a distinctly relevance:
- sea rolling waves generated by intermittent wind having 150 ÷ 175 km/hour
traveling speed,
- seismic waves (tsunami).
3. STRUCTURAL ANALYSIS ASPECTS AND PROBLEMS

CONCERNING FATIGUE TEST [9], [10], [11], [12], [17]
Stresses in offshore structures are highly dependent from global stiffness (jackets
coupling manner, cable inclination and tension etc.).
A dissimilar effect on stresses is coming from the vibration of these structures [13].
The effects of vibrations are originated from the frequency of the acceleration as
well as from the reversing speed of the acceleration. These vibration effects are
acting not only on the bearing structure, but also on the operating crew (there are
persons to whom 0.01·g acceleration at a 1Hz frequency causes sensitivity).
The stress ranges are determined upon an elastic analysis of the structure from
loadings caused by waves, wind etc. [14], [15], [16]. The action of waves has a
dynamic character, similar to seismic action, but having a quasi-permanent
persistence. These loads are determined from the hydrodynamic analysis
considering the FROUDE-KRYLOV components, the incidental hydrodynamic
mass etc.
The coupling between stress status and corrosion is well known. There is a
slackness of the corrosion in case of compressive stress, and an accelerated
corrosion in case of tensile stress.
In the same way there is a connection between fatigue loading and corrosion speed.
The methodologies used for fatigue examination are based on the principle of the
linear cumulating of damages (MINER), based on the prescriptions from:
• DNV – Det Norske Veritas (Norway);
• GL – the German Lloyd (Germany);
• ECCS – the European Convention for Constructional Steelwork,
• API – the American Petroleum Institute (USA).
In order to trail in site these kinds of structures, the authors are advising the
permanent monitoring of the essential joints of the structure and the comparison of
the behavior with a synthetically WÖHLER curve.
At the point when the loading spectrum is approaching to the curve, intervention
can be instantly made.
This procedure allows displaying the contribution of corrosion on the level of the
stress spectrum. The special tensomatic markers which are able to withstand
marine environment are already on the market at a reasonable cost, the problem
being resumed to the mounting and to the procurement of proper software.
The quantification or numerical computation of the stress spectrums can be done
through the RESERVOIR method.
4. CONCLUSIONS
The recent collapses of some offshore platforms, although these were made using
highly advanced technology, shows that structural expertise and examination works
are indispensable for the existing platforms.
The emergence of any damage and the modification of the threshold for structural
collapse impose a reasonable limitation of hazards. Lowering the risk to a certain
level is done usually with uncommon financial exertion. Considering risk
limitation actions, the expertise and eventually monitoring of these structures have
an incommensurably lower cost than other measures, adopted usually on
probabilistic bases.
Uncontrolled movements due to wind, ice loading and

combination of wind and ice
OFFSHORE MOVEMENTS Collapsing movements due to earthquakes
Destructive movements due to waves

Local instability
internal
COLLAPSE TYPES ON Explosions
external
OFFSHORE STRUCTURES Earthquake
Fire
The large changes in configuration (geometric
nonlinearity)
NONLINEARITIES OF
The drag loading and the dependency of the load on the
OFFSHORE STRUCTURES
position and orientation of the structure
The nonlinear load-deformation behavior
Figure 4.
In case of steel offshore structures, the actions which are causing sensitive
modifications of the structural safety can be grouped in the following classes
(figure 4):
- problems concerning base material quality (cold-flow, different brittle fracture
types, lamellar fracture);
- problems concerning global and local stability (curling of thin walls and
plates etc.);
- fatigue and repeated stress;
- creeping;
- problems concerning deformations induced by static and dynamic loads;
- corrosion and erosion problems.
The actual prescriptions need improvements in the way of extending the computing
concepts related to the so called “structural” and “specific imperfections”.
The fabrication and maintenance costs of the offshore structures made from steel
are huge, due to the aspects mentioned in this paper. In such conditions o possible
way and with spectacular results appears to be the using of reinforced concrete by
the accomplishment of the offshore constructions. Alternative structures in this
direction are presented in figures 5, 6 and 7, having the decisive advantage of
corrosion withstanding, relatively simple erection ashore, simple anchoring in
comparison with the special ones from steel structures.
Figure 5.
Figure 6.
Figure 7.
The tackling of fatigue problems is advisable to be done through permanent

monitoring and by comparing the stress spectrum with a synthetically WÖHLER
curve.
Structural engineers dealing with such structures are advised to accept only those
results which are confirmed by a simplified, but not shallow structural analysis.
References
1. Clauss, G., Lehmann, E., Östergaard, C., Offshore Structures, Springer Verlag, London-Berlin-
New York, 1992.
2. Cătărig, A., Kopenetz, L., Alexa, P., Analysis Problems of Tubular Offshore Structures,
Proceedings of the Seventh International Symposium on Tubular Structures, A. A. Balkema,
Rotterdam, Brookfield, 1996.
3. Kopenetz, L., Cătărig, A., Practical Structural Dynamics of Marine Cables, Ovidius University
Annals, Series: Civil Engineering, vol. 1, No. 6, 2004.
4. Kopenetz, L., Cătărig, A., Comportarea “in situ” a structurilor din oţel de tip Offshore, Revista
Construcţiilor, Nr. 1, 2005.
5. CECM–TC6, Recommandations pour la vérification à la fatigue des structures en acier, CTICM
Centre technique et industriel pour la construction métallique, Saint-Rémy-lès Chevreuse, France,
1987.
6. Brebbia, C. A., Walker, S., Dynamic Analysis of Offshore Structures, Nekines-Butterworths,
London-Boston, 1981.
7. Kinsman, B., Wind Waves, their Generation and Propagation on the Ocean Surface, Prentice-
Hall, 1965.
8. Hauptmanns, U., Werner, W., Engineering Risks, Springer Verlag, Berlin-London-New York,
1991.
9. Haibach, E., Betriebsfestigkeit, VDI-Verlag, Düsseldorf, 1989.
10. EUROCODE-3, Design of Steel Structures, 1993.
11. STAS 10108/0-1978, Design of Steel Structures (in romanian), 1978.
12. Dalban, C., et alii, Construcţii metalice, EDP, Bucureşti, 1983.
13. Kopenetz, L., Cătărig, A., Probleme de analiză structurală pentru consolidarea podurilor,
Simpozion Reabilitarea Drumurilor şi Podusilor, Cluj-Napoca, 2000.
14. British Standard–BS5400, Part 10, Steel, Concrete and Composite Bridges, Code of Practice for
Fatigue.
15. RP2A, Recommended Practice for Planning, Designing and Constructing Fixed Offshore
Platforms, Washington, American Petroleum Institute, 1987-1989.
16. API-RP2A, LRFD, its consequences for adoption to North Sea Offshore design practice,
Advanced Mechanics & Engineering Ltd, 1991.
17. EUROCODE-1, Basis of Design and Actions on Structures, 1991.
Scenario and impact of global warming in building energy

performance simulation
Petr Kotek1,2, Jordán1
1
Department of Microenvironmental and Building Services Engineering,
CTU Technical University in Prague, Czech Republic
2
contact: kotek.petr@centrum.cz
Summary
Building performance simulation (BPS) tools are more and more widely used in
practice and developers are trying to facilitate and anticipate the settings of input
parameter for ordinary users. BPS tools are used to determine the energy
consumption, indoor environment and to diagnose heating and cooling load for
relevant HVAC systems in the building under investigation. Every simulation
model is still under a big uncertainty in input setup data in case of using BPS tool
before the building is built. An uncertainty and sensitivity analysis (UASA) was
performed in our latest research by using MonteCarlo simulation (with Latin
Hypercube sampling). The aim of this research was to estimate the most influential
input parameter for building envelop and other boundary condition.
From UASA was assumed the most influential input parameters such as outside
temperature and sun radiation (OT, SR) which we (i.e designer) can not change or
optimize. And moreover, these outside climate data are different every year. It
means that if we use some reference climate database we will put a big uncertainty
in simulation model and outputs are with a big different uncertainty every year.
This paper is focused in impact of different OT, SR during the year and mainly in
point of view: “WHAT IF IN FUTURE”. The climate is lunatic every year and we
can see and feel that the temperature is slowly rising all over the world. Nowadays
the question of global warming is more and more discussed. According
Intergovernmental panel on climate change (IPCC) we performed a study how the
energy consumption will change in predicted years according to IPCC scenario.
Two case studies were chosen for different scenarios of climate change First and
second cases are one real administration building in Czech Republic, Prague. One
is performed without thermal insulation on the wall construction and second with
and is shown their appropriate sensitivity in point of view of climate change
KEYWORDS: climate change, IPCC scenario, sensitivity analysis, Monte Carlo

simulation, Latin Hypercube sampling
1. INTRODUCTION
Uncertainty and sensitivity analysis by using MonteCarlo (MCA) simulation is a

powerful method for estimation the most influential parameters and for clarifying
how big the uncertainty in results is. The simple crude MCA simulation technique
is shown on Figure 1. Inputs Xi can be for example thickness, density and
conductivity of thermal insulation, transmittance of transparent construction,
emissivity of opaque envelope or some inputs for HVAC components etc. [2]
Monte Carlo Analysis (MCA)
input x1
probability density function (pdf)
results with their

uncertainty
output value
new input file
for simulation
thermal
with all n
model
input x2 inputs
changed
time
repeat for all simulations from 1 to E

input xn
Figure 1. basic principle of MonteCarlo (MCA) simulation.
Many of them can be future optimize by using some optimization algorithm but
some of these like outside dry bulb temperature and sun radiation could not. These
not “predictable” inputs can bring huge uncertainty in results and they can influent
heating and cooling load for HVAC equipment and change heating and cooling
energy consumption every year. Whence it follows that if we use BPS tool for
assessment energy consumption we have to present range of values where the real
energy consumption per year probably will be. It means, that if we use some “old”
climate database with average temperature for 50 or 30 years, we have to analyze
these values, compare with measured values for 5 years backwards for example
and performed new ones.
From Figure 2. we can see the difference between 30 years observation in Prague
and actual measure temperature. From the chart we can also see, that the long term
observation is little bit “colder” than nowadays temperature. From this we can say,
that the distribution of the temperature is not normal during the year but more Beta
distribution (with α=2 , β=5). It is obvious that if we use the “red” values for the
building performance simulation we will have different output values instead of
using the “blue” one.
150 P. Kotek, F. Jordán
Figure 2. difference between long term observation and actual temperature in Prague
From the Figure 2. we can feel about, that the climate is wormer and wormer than
long term observation. Also this winter 2007 a lot of people save money for paying
heating bills because the temperature was so high.
And moreover in 2003, 2004, 2005, 2006 was the annual average temperature
higher than observation among years 1971-2000 in Czech Republic (CR). Nobody
knows if our planer will rapidly warming-up but all over the world we can see, that
rising temperature changing a lot of thing just now.
In 1988 United Nations Organization (OSN) established Intergovernmental Panel
on Climate Change (IPCC) by World Meteorological Organization (WMO) and by
the United Nations Environment Programme (UNEP). IPCC reports are widely
cited in almost any debate related to climate change.
The IPCC published a first assessment report in 1990, a supplementary report in
1992, a second assessment report (SAR) in 1995, and a third assessment report
(TAR) in 2001. A fourth assessment report (AR4) is currently under way.
The IPCC has made a series of reports and scenarios related to climate change and
these it can be find out in TAR. Statements of the IPCC or information from the
TAR are often used as a reference showing a scientific consensus on the subject of
global warming. An increasing body of observations gives a collective picture of a
warming world and other changes in the climate system (The global average
surface temperature has increased over the 20th century by about 0.6°C)
The TAR estimate for the climate sensitivity is 1.5 to 4.5 °C over the period 1990
to 2100. Each scenario then has a range of possible outcomes associated with it.
The most optimistic outcome assumes an aggressive campaign to reduce CO2
emissions, while the most pessimistic is a "business as usual" scenario. The more
realistic scenarios fall in between. [6]
Figure 3. The 140 years observation of global temperature (IPCC)
Figure 4. The 1000 years climate model observation of global temperature (IPCC)
According the IPCC long term observation (Figure 3., Figure 4.) we can see, that
the Earth is really warming-up but maybe this is the normal “behavior”of our
planet and it happened many times in the past. But maybe is it true and every single
human on our planet making our planet warmer by driving car for example. This
paper is based on work of 2500 scientist all over the world in 120 countries and we
used the IPCC outcomes from their climate model as an input for our investigation.
Investigation in point of view: How the heating and cooling energy consumptions
will change over the period 1990 to 2100.
2. IPCC SCENARIOS
From TAR outcomes we used the scenarios

• Global mean temperature change (ºC) (Figure 5a.)
• Radiative solar forcing (W/m2) (Figure 5b.)
a)
b)
Figure 5. The IPCC scenarios a) temperature b) solar radiation
For creating new climate database it can be used crude MCA (via analysis in
Figure 2. and prediction in Figure 5). But the pure MonteCarlo simulation cannot
be applied for time-consuming problems such as dynamic BPS tool, as it requires a
large number of simulations. A small number of simulations can be used for the
acceptable accuracy of statistical characteristics of response using the stratified
Latin Hypercube sampling technique LHS [1]. For our aim is enough to create
climate database based on scenario B1 (lower bound) and A1Fl (upper bound) with
uncertainty shown in Figure 2. The IPCC report also includes one inconsiderable
fact. Fact, that annual temperature is rising and also every year occur more days
with extreme temperature in winter and summer. This fact can be seen on Figure 6.
Figure 6. Globe mean temperature change and variance in future
3. PROGRESSION
For Czech Republic exist two *.epw climate databases – for Prague and for Ostrava
city (see Figure 7.). These *.epw files can be download from EnergyPlus website
free of charge. [8] and used for every king of BPS tool (ESP-r, Trnsys, DOE-2,
Design Builder, IES<VE>, etc).
Figure 7. Czech Republic and marked two cities with available climate database
Sometimes these two climate databases are used for the whole Czech Republic and
hereby bring big uncertainty in results. But this has not been considered in this
paper. The case study (administration building) is in the capital city Prague and we
will modify the downloaded climate file CZE_Prague_IWEC.epw with Weather
Statistics and Conversions (WSC) tool developed by EnergyPlus. [8]
The statistical values for both climate databases are shown in Table 1.
Table 1. Statistical report of CZE_Prague_IWEC.epw

GPS / elevation 50°44' N, 14°29' E/ 364 m
Annual mean temperature 7.9 °C
Days in heating period 245
Average temperature during heating period 4.09 °C
min / max temperature -15.2 / 30.5 °C
Hours with temperature > 0°C /year 7078 hours (295 days)
Hours with temperature ≤ 0°C /year 1682 hours (70 days)
By using EnergyPlus WSC tool the files were converted to the *.csv and exported
to the MS Excel → modified for two scenarios (step 10years) and uncertainty made
by SimLab [7]. These new weather files (12 for each scenario) were converted
back to the *.epw file with WCS tool and used for simulations.
Table 2. Statistical report of CZE_Prague_IWEC.epw in 2100 according IPCC

scenario B1 in 2100 scenario A1Fl in 2100
Annual mean temperature 10.4 °C 12.7 °C
Days in heating period 204 180
Average temperature during 3.86 °C 4.22 °C
heating period
min / max temperature -16.3 / 36.7 °C -17.8 / 42.6 °C
Hours with temperature > 0°C 7393 hours (308 days) 7596 hours (317 days)
Hours with temperature ≤ 0°C 1367 hours (57 days) 1164 hours (48 days)
0.06
0.05 epw_Prague_1995
0.04
IPCC_scenario A1Fl_2100
0.03
0.02
0.01
0
-30 -20 -10 0 7.9 10 12.7 20 30 40 50
Figure 8. Globe mean temperature change and variance according IPCC for Prague
For this prototype we used IES <VE> [5] which incorporates ApacheSim, a
dynamic thermal simulation tool based on first-principles mathematical modeling
of building heat transfer processes. It has been tested using ASHREA Standard 140
[4] and qualifies as a Dynamic Model in the CIBSE system of model classification
[3].
4. CASE STUDIES - ADMINISTRATION BUILDING
For the real administration building in Prague was performed two possible
situations.
• Adm.case A without thermal insulation like many old building in CR
• Adm.case B with 8cm insulation on the wall.
Another construction and setting were used the same for both cases.
Figure 9. Photo of administration building and energy model in IES <VE>
The material prosperities for both cases can be found in Appendix 1.

Table 3. Basic proportions and settings of the building

Width x length x hign 12.4 x 15.3 x 8.1 m
Volume 1537 m3
Area of glazing 18 %
Set point for heating - day / night, weekend 20ºC / 15ºC
Set point for cooling - day / night, weekend 24ºC / 32ºC
Working hours 6 - 18 hours
For these cases (adm.case A, adm.case B) were made new 48 “IPCC weather files”
and for each one made 20 “uncertainty files” according probability distribution.
The automatic simulation were performed during approximately 4 hours on Intel
Core2 CPU T5500, 1.66 GHz, 1GB RAM.
5. RESULTS
Results shown in Figure 10 represents both cases in point of view energy load for
heating (red curve) and cooling (blue curve) and annual energy consumption.
30 60
25 50
20 40
Adm.case A
15 30
10 20
5 10
0 0
1995
2015
2035
2055
2075
2095
1995
2015
2035
2055
2075
2095
30 60
25 50
20 40
Adm.case B
15 30
10 20
5 10
0 0
1995
2015
2035
2055
2075
2095
1995
2015
2035
2055
2075
2095
Energy demands (kW) Energy consumption per year (MWh)

Heating Cooling
Fiure 10. Impact of global warming according scenario IPCC - the uncertainty in results on
the basic shown in Figure 2 represents the black boundary.
From results in Figure 10 can be seen, that if the temperature and solar radiative
forcing rise up, than heating load will stay almost the same. The annual heating
will little bit fall down because of warmer winter. Obviously the huge impact will
be in cooling demand and consumption. If we compare the results for non insulated
and insulated building it is nice to see, that cooling demand will rise up from now
in 2100 about 70%±10% (without insulation) and 40%±9% (with insulation) and
consumption will be higher about 218%±21% (without ins.) and only 82%±12%
(with ins.) This big difference is because of easier accumulation in the brick wall
without insulation.
6. CONCLUSIONS
The aim of this paper is to show how big the impact of global warming can be in
total annual energy consumption per year. The fact, that we will need more power
for cooling bring together with higher temperature also higher pollution from
cooling sources and the ozone hole will be bigger and bigger because of
greenhouse gases. It is not obvious, that thermal insulation can also cut down the
cooling consumption and it is time to do it for every single building not only
because of reducing heating demand.
Acknowledgements
The authors thank for support CTU research aim No. CEZ MSM 6840770003
References
1. McKay M.D., Conover W.J. and Beckman R.J., A Comparison of Three Methods for Selecting
Values of Input Variables in the Analysis of Output from a Computer Code, Technometrics Vol.
21: 239-245., 1979.
2. Kotek, P., Jordán, F., Kabele, K., Hensen, J., Technique of uncertainty and sensitivity analysis for
thermal and HVAC calculations, IBPSA 2007, Beijing, Sep/2007. (not published yet)
3. Crawley, B.D., Hand, W.J., Contrasting the capabilities of building energy performance
simulation programs, United States Department of Energy, University of Strathclyde, Wisconsin,
2005.
4. ANSI/ASHRAE Standard 140-2001, Standard Method of Test for the Evaluation of Building
Energy Analysis Computer Programs.
5. IES <Virtual Environment> 5.1. Building performance simulation tool,
http://www.iesve.com/content/ , last visited May 2007
6. IPCC-The Scientific Basis, Published by the press syndicate of the university of cambridge for the
Intergovernmental Panel on Climate Change (IPCC), www.cambridge.org ISBN 0521 01495 6,
2001
7. SIMLAB Version 2.2. 2004. Simulation Environment for uncertainty and sensitivity analysis,
developed by the Joint Research Centre of the European commission. http://simlab.jrc.cec.eu.int
8. EnergyPlus website: http://gundog.lbl.gov/
Appendix 1 - Thermal prosperities of administration building

Adm.case A - without thermal insulation
Wall Construction (outside to inside)
EN-ISO
Conductivity Thickness Capacity Density
Layer U-Value
[W.m-1.K-1] [m] [J.kg-1.K-1] [kg.m-3]
[W.m-2.K-1]
Outside plaster 0,190 0,0125 1000 800
Brickwork 0,800 0,450 900 1700 1,21
Inside plaster 0,350 0,010 1000 1000
Adm.case B - with thermal inslutaliton
Wall Construction (outside to inside)
EN-ISO
Conductivity Thickness Capacity Density
Layer U-Value
[W.m-1.K-1] [m] [J.kg-1.K-1] [kg.m-3]
[W.m-2.K-1]
Outside plaster 0,190 0,0125 1000 800
Polystyrene 0,044 0,080 1270 20
0,38
Brickwork 0,800 0,450 900 1700
Inside plaster 0,350 0,010 1000 1000
EN-ISO
Other construction U-Value
[W.m-2.K-1]
Roof 0,28
Ground Floor 0,9
Floor 0,84
Window 1,7
Glass door 1,7
Settlement prediction of a pile group based on design charts

Irina Lungu1, Nicolae Boţi2 and Anghel Stanciu3
1
2
3
Summary
The rehabilitation of transportation infrastructure is an on-going process that

involves an increased complexity of bridges and retaining structures with
corresponding complex requirements to comply with during service.
When the necessary bearing capacity is not provided by means of a shallow
foundation, a deep foundation, i.e. a piles group is considered appropriate for both
bearing and deformation requests.
Pile foundations represent current foundation solutions for these constructions,
given the poor soil conditions often encountered on the construction site. Their
design is based on the soil limit states – both deformation and bearing capacity.
Settlement restrictions that are included with the deformation limit state require an
accurate evaluation of the settlement prediction regarding the pile group.
Depending on the purpose of using piles for the foundation system, one of the
following three cases will be the most suitable in a practical design: (a) the piles
are supposed to carry the full load; (b) the piles will act as settlement reducers for
the raft which is a less important bearing element of the foundation system; the
piles are required to have a factor of point and shaft capacity safety somewhat
greater than unity; (c) the piles are provided as settlement reducers but they will
operate at a shaft factor of safety of unity, and piles of different length or stiffness
are used . The raft is the principal bearing element of the foundation.
Sometimes, simple settlement calculations seem to be adequate for practical
purposes provided that the appropriate in-situ soil properties have been carried
out with engineering judgment. The errors introduced by the simple methods are
usually small compared with those that can occur during sampling and testing.
The laboratory and field tests on single piles and pile groups as well as a careful
monitoring activity of buildings with pile foundation have known a tremendous
evolution, implementing all this in data base systems used in the computer field and
advanced numerical methods for engineering purposes.
KEYWORDS: pile group, settlement, design charts, pile-soil interaction, single

pile, Finite Element Method, Boundary Element Method.
160 I. Lungu, N. Boţi, A. Stanciu
1. INTRODUCTION
There is a strong intention to modify the actual capacity based design to a

settlement based design when foundation systems on piles are concerned
[1],[2],[5]. In pursuing this issue there are the following arguments:
• bearing capacity of the single pile is very sensitive to the pile installation
procedure while its stiffness is less affected;
• bearing capacity of an individual pile in a pile group may significantly vary
when piles have different dimensions and spacings, or when due to soil
variation, the installation procedure may introduce some differences among
them, while pile group behavior under loading intend to make this differences
less significant;
• the connecting raft may deliver a significant part of the load to the subsoil by
direct contact and in the same time, its influence on the foundation settlements
may be less important [5],[39];
• when piles act as settlement reducers the most important factor that decide on
the necessary number of piles is the differential settlement.
2. SETTLEMENT PREDICTION OF A PILE GROUP BASED ON

THE SINGLE PILE BEHAVIOR
Considering that the single pile rigidity was established by load tests on single
piles, the stiffness of a pile group may be quantified through the use of an
efficiency, η, the inverse of the group settlement ratio, Rs, defined as:
1 kp
η= = (1)
Rs nk1
with k1 - pile-head stiffness of a single pile; kp - stiffness of the pile group (in terms
of average settlement).
To a first approximation, the group efficiency was taken as a simple power law of
the number of piles in the group, as: η ≈ n − e , therefore the group stiffness
becomes: k p ≈ n1− e k1 [3].
The exponent, e, generally lies in the range of 0.3-0.5 for primarily friction piles,
rising to 0.6 or higher for end bearing piles. Fleming presented a set of design
charts (figure 1) in terms of a base value ,e1, and four correction factors c1 to c4, as
indicated in the equation:
e = e1 (l d ) ⋅ c1 ( E p / G ) ⋅ c2 ( s / d ) ⋅ c3 ( ρ ) ⋅ c4 (ν ) (2)
Figure 1. Design chart to evaluate the pile efficiency
The simple use of efficiency factors does not provide any indication of the
increased proportion of load carried by the pile bases in a group of piles - and thus
the greater importance of underlying soil strata compared with the case of a single
pile [5].
Poulos [2] pointed out that since the base response is generally more non-linear
than the shaft response, this may lead to lower group efficiency than using linear
elasticity. But ignoring non-linear effects will overestimate the interaction between
piles, which will compensate for overestimation of the base stiffness.
3. COMBINATION OF BOUNDARY ELEMENT WITH FINITE

ELEMENT APPROACH
Hain and Lee [4] developed a method for pile raft settlement prediction which
included a finite element approach for the raft and a boundary element approach
using Mindlin’s solution for the pile group interaction (so called raft on a pile
reinforced continuum).
The model considers a variable stiffness raft of any geometry supported by a
random group of identical piles and a soil mass idealized as an elastic continuum.
Consideration is given for a deep homogeneous mass, a modified linear elastic
analysis for a layer of finite thickness, a failure of individual piles at load less than
the total group capacity by applying an excess load cut-off procedure.
Four types of interactions are included in this approach and represented in figure 2.
Figure 2. Pile-soil-raft interactions – basic representations

Imposing the compatibility and equilibrium conditions between the soil-pile group
system and the raft a set of stiffness equations can be formed. Two assumptions are
involved:
• only vertical forces are transmitted from the raft to the pile heads;
• each pile occupies the whole of the “constant pressure” area around a
particular node.
These assumptions can not be appropriate to be applied for the analysis of piled
isolated footings where lateral forces and overturning moments are the major
proportion of the applied loads.
The pile compressibility is also considered and as a conclusion, the author stated
that the percentage of load suported by the pile group diminishes as the pile
compressibility increases.
A decrease in the length of the piles had the similar effect on the load distribution
in the foundation system. The final results with all the influences assessed at the
begining are presented in figure 3 as design charts.
Figure 3. Design charts with parameters influencing the pile group behavior under vertical
loads
The method is considered the first attempt in considering a sound interaction in a

piled raft foundation with the limitations of a linear elastic behavior of the soil, a
regular pile arrangement: the raft and piles with the same length and geometry [5].
4. CONCLUSIONS
The stiffness of a single pile becomes more important in design practice not only
by its value but for learning the motivation of a suggested procedure in the
available literature and thus to be able to make a sound engineering judgement
each time a new case comes up.
Design charts proved to be useful tools to set comparative values at work in order
to select the foundation elements in the best combination to face even the most
pessimistic hypothesis of soil-foundation interaction. A successful and economic
design is reached only when the interaction effects are taken into account.
References
1. Van Impe, W.F., Developments in pile design, DFI Conference, Stressa, 1991
2. Poulos, H.G., Foundation economy via piled-raft systems, Piletalk International, K. Lumpur,
Malayesia, pp 97-106, 1991
3. Fleming, W.G.K., A new method for single pile settlement prediction and analysis,
Géotechnique, vol 42, no 3, pp 411-425, 1992
4. Hain, S.J. and Lee, I.K., The analysis of flexible raft-pile systems, Géotechnique, vol 28, no 1,
pp 65-83, 1978
5. Lungu, I., Teză de doctorat, Contribuţii privind conlucrarea între terenului de fundare, fundaţie
şi structura de rezistenţă, Iaşi, 1997
The importance of controlling subsystem in information systems

for building industry
Peter Mesároš1, František Mesároš2
1
Department of Management Sciences, Faculty of Business Economics, University of Economics,
Košice, 04130, Slovakia
2
Research Institute of Building Informatics Ltd., Košice, 04001, Slovakia
Summary
The paper presents various aspects and opportunities of using information systems
and information technologies in building industry considering their impact on
competitiveness. Especially we focus on new subsystem of controlling and its
integration in corporate information system. It is a contribution to the research
project Nr. 2330256/07 of University of Economics in Bratislava, named
Knowledge Management as Assumption of Successful Development of Slovak
Enterprises.
KEYWORDS: information systems, information technologies, controlling system,
building industry
1. INTRODUCTION
An increasingly rapid change and development in the business environment of

building firms have been noted in recent years. The growth of domestic as well as
foreign competition creates a constant pressure on intensive exploitation of the
latest information technology, innovation of products and increase in the quality of
construction performance. This appears to be the way of acquiring a competitive
advantage and achieving the building firm prosperity. Important factors of
prosperity are mainly composed of:
• profit which is not formed at the expense of the future gains,
• short-term liquidity and long-term capacity of funding,
• economy of the whole firm with an accent on economization and quality of
the construction performance,
• effectiveness in using relevant information and information technologies for
its processing in order to achieve competitiveness.
Managers in the building firms, even despite the fact that various programme
systems have already been introduced in the corporate information systems, report
on the lack of relevant current and exact information related to the corresponding
dates and suitable form of decision making in an efficient and effective way. The
current corporate information systems in the building industry are not capable of
166 P. Mesároš, F. Mesároš
providing the information that is needed by managers. In connection with this, it is

particularly the information required for:
• comparing the plan and reality as to the ongoing conduct of the building i.e.
the information necessary for the ongoing management of profit and costs of
the construction,
• making the deviations analysis as compared to the planned state,
• rapid conduct of calculation with providing the data for the needs of further
measures to eliminate the deviations from the plan.
The current information systems do not assure a rapid indication of problems and
analysis of their causes with the proposal for their elimination (Čarnický, 1999).
Managers can find this information in a huge amount of data concerned with the
plan, accounting, marketing, various reports, etc. They select only certain
information, make their own analysis in order to make the optimum decisions.
However, this activity is rather time-consuming and not so effective, as a result of
this process – making a relevant directing decision – appears to be frequently rather
subjective and intuitive. Similar associations are also mentioned by various other
authors (M. Kozlovská, 2002). In recent years a large business has grown up
around the subject of information systems and information technologies and its
implementation in business practice of all industries of national economies. ICT
companies and management consultants have developed methods and techniques for
identifying and agreeing investment in information systems that support business
strategy and create new strategic options and build the competitiveness. Information
technology has changed the nature of competition and has created new
opportunities for using information services. Using a technology to reduce the cost
of building company activities or to add value to existing products and services can
provide a competitive advantage.
2. INFORMATION SYSTEMS FOR BUILDING INDUSTRY
2.1. ERP systems and its exploitation in building companies
According to Wikipedia definition, Enterprise Resource Planning systems (ERPs)

integrate all data and processes of an organization into a unified system. A typical
ERP system will use multiple components of computer software and hardware to
achieve the integration. A key ingredient of most ERP systems is the use of a
unified database to store data for the various system modules.
Enterprise Resource Planning is a term originally derived from manufacturing
resource planning (MRP II) that followed material requirements planning (MRP).
MRP evolved into ERP when routings became major part of the software
architecture and a company's capacity planning activity also became a part of the
standard software activity. ERP systems typically handle the manufacturing,
logistics, distribution, inventory, shipping, invoicing, and accounting for a
company. Enterprise Resource Planning or ERP software can aid in the control of
many business activities, like sales, marketing, delivery, billing, production,
inventory management, quality management, and human resource management.
ERPs are cross-functional and enterprise wide. All functional departments that are
involved in operations or production are integrated in one system. In addition to
manufacturing, warehousing, logistics, and Information Technology, this would
include accounting, human resources, marketing, and strategic management.
Implementing ERP software is typically not an internal skill, so even smaller
projects are more cost effective if specialist ERP implementation consultants are
employed. The length of time to implement an ERP system depends on the size of
the business, the scope of the change and willingness of the customer to take
ownership for the project.
2.2. Controlling subsystem – new component in ERPs for building industry

Special applications
In our view, controlling presents a specific form of work with information, its
gathering, processing, evaluating and providing for the needs of managerial
performance and decision-making. From the point of view of time, two main
directions have been formed, namely the strategic and operational controlling. The
strategic controlling is aimed at direction of a long-term detecting of the business
company potentials. Externally, it analyzes the threats and opportunities. The
operational controlling directs its activities within the given profit potential. Its aim
is the optimisation of subject, time and value parameters of the corporate activities
aimed at achieving the prosperity and profit in the process of building production
conduct.
2.2.1. Assumptions of Controlling Subsystem Implementation
In the controlling subsystem the existing databases are used that have been formed
within the calculation-planning subsystem (ex ante information) and the database
of personal and economic information subsystem (ex post information).
A simplified connection of the given subsystems and their groups is illustrated in
Fig.1 and Fig. 2.
The data are taken from the group of tasks Production calculation that presents the
proposed economic and capacity expression of technical, technological, material
and organizational performance of the object construction or just a part of it.
A whole range of data can be found here, namely the limit ones, the norm/standard/
and planned data for the whole construction, an object, or a part of it. Continuously
once a month, these data are taken from the group of tasks Production billing for
the concretely undertaken volume in the given construction, object, part, or an
order.
Subsystem
OPERATIONAL
CONTROLLING
Calculation-planning Subsystem of social and

Subsystem economic information
1. standarbasis
budget 1. wages
2. production calculation 2. human resources
3. production billing 3. accounting
transportation 4. transportation
4. budgeted indexes 5. material and technical
5. time plan assurance
Figure 1. Operational controlling as a subsystem of the information system
The production billing uses the same standard basis as the production calculation.
The production billing (PB) is calculated as a multiplication of the data of the real
vector of production (Q sk) and the data from the standard basis (SB).
PB = Qsk ∗ SB (1)
The production billing can be processed from:
• organizational points of view for orders, or objects when the order includes
more objects (constructions, centres, centre of costs, division, plant),
• the point of view of time in a month, exceptionally even a shorter time period
(summarized quarterly, annually, for the whole order).
The area of drawing the building costs: The tool for performing this activity is
accounting. Even smaller enterprises should have an independent area of the so-
called (managerial) accounting of costs. The accounting of costs enables to use
such procedures and methods of accounting that are not allowed for financial
accounting. The objective of this procedure is to enable to define the respective
volume of variable costs and to find the so-called contribution profit (a contribution
to cover the fixed costs from the profit production) at every calculation unit, centre,
etc.
The essential difference between adding the costs to the financial accounting and
accounting of costs consists in the fact that:
• financial accounting is managed by the accounting classes for the needs of
making a record on the profit and losses. The so-called balance sheet and
therefore the main criterion here, is a strict evidence of documents;
• accounting of costs is more governed by the principle of causality. The basic
kinds are classified with an accent and in dependence on certain decisions so
that they could directly enable the respective management measures.
The data on the reality are mainly taken from the group of tasks of intracorporate
accounting in the structure defined within the calculation-planning subsystem in
connection with the organizational structure of the given building firm. For
example:
EC - economic centre
CC - costs centre
EA - economic activity
CU - calculation unit/entity (construction – object)
Therefore, it is inevitable to create a code of numbers of the actual suppliers in the
structure - a plant, EC, CC, EA.
CONTROLLING AND ITS IMPLEMENTATION
Determination Detection of Analysis of Proposal of

of quantities deviations deviations measures
Plan of costs Comparing plan Reasons of Recommendatios

– reality deviations for management
Figure 2. Content of activities and procedural steps in the controlling subsystem
Within the controlling subsystem even the links to the material and technical
supply for the purposes of continuous comparing the planned material need with
the real material use, not only in the financial but also in the material expressing,
are defined as well.
When selecting the followed indexes the individual building firms themselves
determine the type of indexes, as well as the extent and degree of details. However,
their extent is given by the principle of effectiveness of the information system. It

is required that the costs for their provision are lower than the effects provided by
the detected deviations and accepted decisions.
2.2.2. Advantages of implementing and exploiting the controlling subsystem
The advantages arise from technically calculated and justified norms of

manufacturing and technical character of unit costs elaborated by the Department
of manufacturing and technical preparation of production. Thus, controlling
• contributes to the prosperity of a building firm,
• creates presumptions for acquiring a higher quality of building work
• help achieve rapid outcomes of costs drawing in the course of construction
(once a month at the minimum), i.e. not after the completion of construction,
• serves to common check-up of drawing the individual costs for the
construction,
• enables to carry out a detailed analysis of deviations, provides the possibility
of analysing the causes of their origin and points to the divisions responsible
for their origin,
• has a psychological significance since it upgrades the motivation of workers
towards savings (as this method enables a preventive check-up of costs
drawing, it also motivates the workers towards higher responsibility for their
own economic behaviour),
• creates pressure towards the perfect arrangement of organizational relations
among divisions, sticking to the technical, technological and organizational
conditions of the construction performance,
• enables to evaluate the effectiveness of production divisions activities as well
as the activities of divisions of pre-production and production preparation
(Trávnik, 1998), e.g. incomplete project documentation of the construction
with a poor quality, on the basis of which the technical preparation of the
construction is elaborated. The production calculation may thus cause rather
great deviations from the norms that are not due to production but are
generally ascribed to production,
• enables to analyse continuously the objectivity of understanding the right
overheads for the respective construction job,
• provides important data for the analysis of the secondary budgeted costs that
are often incorrectly accounted within the construction overhead costs.
In the context of the above said it is highly recommended to set up a position of the
controller who, on the basis of his knowledge and practical experience, could
analyse the detected deviations between the plan and reality. He thus assures the
working of information flows, i.e. provides objective information for decision-
making on the basis of real processes taking place in the construction.
Within his work capacity, the controller assures the following tasks, in particular:
• prepares supporting documents for planning and decision-making,
• processes the methodological materials, regulations for processing the
normative basis of the enterprise, production calculations, budgets,
production billing, budgeted indexes, supporting papers for a follow-up of
production cost indexes in the intracorporate accounting,
• follows the legal acts in the field of pricing, accounting, calculations,
statistics, taxes, etc., and tries to implement them into the corporate
information system as soon as possible,
• provides methodological counselling in the given areas for the enterprise
workers.
The setting of the controller tasks provides only a framework of his activities.
Therefore, it is rather necessary to take into account the specific features of every
building firm as well.
3. CONCLUSIONS
Application of the controlling subsystem appears to be a relatively simple and

transparently conceptual fact. It assures the complex approach and presents an
effective form of acquiring, processing and interpreting the information on the
costs and profit, which support decision-making and formulating the strategy by
managers. It gives presumptions for increasing the efficiency of a building firm and
its prosperity according to the objective indexes. It enables to act successfully in
the competitive environment as well as to respond to constantly increasing
requirements of investors to the quality of building production.
References
1. Čarnický, Š., Why the top management workers need executive information systems? Acta
Oeconomica Cassoviensia N3, Faculty of Business Economics Košice of the University of
Economics in Bratislava, 1999.
2. Kozlovská, M., Management of purchasing, Eurostav, vol. 8, 2002.
3. Mesároš, F., Mesároš, P., Information Systems Strategy in Building in Industry, Economics and
Business Management, vol. 1, 2003.
4. Trávnik, I., et al., Directing the building work value, Slovak Technical University in Bratislava,
Bratislava, 1998.
Analysis of Water Consumption in Residential Building

Roman Musil
CTU in Prague, Faculty of Civil Engineering, Department of Microenvironmental and Building
Services Engineering, Thákurova 7, 169 34 Prague 6,e-mail: roman.musil@fsv.cvut.cz
Abstract
This paper consider about measurement consumption of cold water in residential

buildings. In the year 2005 ran over extensive monitoring cold water consumption
in 62 residential building in Pilsen implement by waterworks Pilsen with
cooperation department of microenvironmental and building services systems in
Prague. The paper considers about analysis this measurement and comparsion
with Czech norm and prescriptions. There will be evaluate daily and hourly
maximal coefficient cold water take-off and daily and hourly cold water
consumption profiles for residential buildings.
Keywords
monitoring, residential building, water consumption
1. INTRODUCTION
Need and consumption of cold water in residential building is magnitude which is

not initial only for dimension supply water systems but it shows life style and user
consumer habits. This value is evolving and changing with society development.
Impact on user behaviour is in relation prices development. The second aspect is
technical building facilities development in hot water preparation which enables
using flexible facilities reacting on variable water needs.
2. WATER NEED
Water need is supposed take-off while consumption is the real take-off water
amount for certain time period. Water consumption assessment is input value
which is covered with sufficient equipment capacity for water supplying. Problem
is when consumption is function of time and we have to observe it in different time
scale. Therefore water need assessment isn’t possible simplify to one value but
time period has to be written by more parameters which are:
• water need during the chosen period (specific water need)

• water take-off during time period distribution
Basic statement is specific water need qp. This value is given in dependence on
kind of operation and activity in designed building. It is water need per basic
consumption unit (person, bed, eat,…) and it is in liter/(unit*day). Probably water
losses from source to consumer are included in specific water need. Probably water
losses are supposed about 20%. State specific water need for residential buildings
according to valid Czech public notice number 428/2001 sb. are in table 1.
Table 1. Specific water need [m3/person/year] according to valid Czech notice 428/2001 for
residential buildings.
3. WATER NEED ASSESSMENT
We calculate with subsequent values for assessment of water need whichever

object:
1) Daily average water need
QP = q . n [ l/d] [1]
Where:
q – specific water need (liter/unit/day)
n – number of unit
3.1 Daily maximum water need
Facilities for water take–off from source, capacity of water and water pipelines for
supplying water to water–supply tank are designed on maximal water need.
Qm = QP . kd [ l/d] [2]
Where:
174 R. Musil
Qp – average daily water need

kd – daily inequality coefficient
Table 2. Take-off daily inequality coefficient [ČSN 736701]
3.2 Hourly maximum water need
We dimension facilities for supply water to consumers on this value. Running of

vary hourly need is use for dimension of pumping device which supply water
directly to consumers (for example pumping station for increasing water pressure).
Hourly water need is one value from component for assessment of water-supply
tank value.
Qh, max = Qm . kh, max . z [ l/hod] [3]
3.3. Hourly minimum water need
Qh, min = Qm . kh, min . z [ l/hod] [4]

Where: Qm – daily maximum water need [l/day]
kh, max – hourly maximum inequality coefficient
kh, min – hourly minimum inequality coefficient
z – take-off water period – residential building a hotels z = 24 hours
– administrative buildings z = 10 hours
Balance values Qm and Qh serve to accumulate reservoir and automatic pumping

station designing which are components of internal water-supply. Determination of
daily and hourly inequality coefficient is very complex for residential and civil
buildings. These coefficients depend on various factors. Water need during day is
unequal and depends on building operation character, on yearly season,
temperature change, flat and house furniture of building services water systems, on
social and age of user groups and their behavior. Water take-off varies during week
too.
In 2005 year went ahead wide monitoring water consumption in 62 residential

buildings in Pilsen practiced by the waterworks Pilsen in cooperation with
Department of Microenvironmental and Building Services Systems, CTU in
Prague. Monitoring on cold water was done in 3 week period by water meter fixed
on building foot with hourly period saving measured values. These values are
primarily determined for processing stochastic loads profiles and mathematical
urban space modelling. Final values don´t include hot water which is supplied to
building by centralized heating system. Evaluate measurement results are in next
graphs.
Cold Water Consumption
6.00
5.00
4.00
litre/perso
3.00
2.00
1.00
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
time (hour)
Monday Tuesday Wednesday Thursday

Friday Saturday Sunday average
Figure 1. Daily cold water consumption during week
Table 3. Comparsion of hourly inequality coefficient in several days with average value
176 R. Musil
Cold Water Consumption
8.00
7.00
6.00
litre/perso
5.00
4.00
3.00
2.00
1.00
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
time (hour)
Monday Tuesday Wednesday Thursday

Friday Saturday Sunday average
Figure 2. Maximal cold water consumption during week – envelope curve [2].
Daily Water Consumption

120.00
Cold Water (l per person and d
100.00
80.00
60.00
40.00
20.00
0.00
ay
ay
ay
y
y
ay
ay
da
da
sd
sd
rd
nd
id
es
on
tu
ne
Fr
ur
Su
Tu
M
Sa
Th
ed
W
Time (h)
Average daily water consumption

Maximum daily water consumption-envelope spline
Figure 3. Total daily cold water consumption

Average Cold Water Consumption Regarding to Building

Occupancy
Average consumption (litre /person/
90
80
70
60
50
40
30
20
10
0
ay
ay
ay
y
y
ay
ay
da
da
sd
sd
rd
nd
id
es
on
tu
ne
Fr
ur
Su
Tu
M
Sa
Th
ed
W
40 - 50 inhabitants 50 - 75 inhabitants
100 - 300 inhabitants 300 - 400 inhabitants
400 - 500 inhabitants above 500 inhabitants
average
Figure 4. Total daily cold water consumption divided according to number of occupant in
the building
Average hourly take-off inequality coefficient for residential buildings
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
ay
ay
y
ay
ay
ay
da
da
sd
sd
rd
nd
id
on
es
ne
tu
ur
Fr
Su
M
Tu
Sa
Th
ed
W
kmax for 40-50 inhabitants kmin for 40-50 inhabitants

kmax for 50-75 inhabitants kmin for 50-75 inhabitants
kmin for 75-100 inhabitants kmax for 100-300 inhabitants
kmin for 400-500 inhabitants kmax for more than 500 inhabitants
kmin for more than 500 inhabitants
Figure 5. Hourly inequality coefficient according to flat occupancy

178 R. Musil
4. CONCLUSION
From measurement results result that total yearly water consumption from all 62
measurement is 56 liter/person/day (Fig. 3 and Fig 4.). Average water consumption
in same object type is for us important at total water need evaluation and is very
depended on concrete measured residential buildings, its inhabitants and technical
equipment. For building services systems designing average value doesn’t have
predicative value. As well as in other technical areas for example structural
mechanics for beam calculation we have several load states which are partly
covered and results is maximum load envelope curve. It is similar with water
consumption, in this case we have 62 load states (done measurement) where
envelope curve copy maximal take-off values in separate hours (Fig. 2). Average
maximal take-off is 90 liter/person/day. From equation [2] is coefficient inequality
1,6 which in comparable with Czech norm value ( for 1000 inhabitant is kd=1,5).
From Fig. 3 is possible to see that average daily water consumption during week
has minimum difference. Differences are in time period of take-off (Fig. 1 and
Fig.2) where is possible to see extreme in noon and in the evening. Greatest take-
off peak is in Saturday (noon) and in Sunday (noon and evening). In Saturday
evening is surprisingly lowest from the whole week. Measurement results shows
also number of inhabitants influence on average cold water consumption in object
(Fig. 4). Average cold water consumption in object with lower number of
inhabitant is bigger than by bigger occupancy of building. The next important
indicator is hourly inequality coefficient (Fig.5) which is in most of groups
expressively other than ČSN 736701 values, how is shown in consequent table.
Table 4. Comparsion of hourly inequality coefficient in several days with average value
Values were calculated from measured dates always for every day in week and
were averaged for compare with Czech norm values. Table shows that values given
by calculation from measured cold water consumption in residential buildings are
extensively different. Residential building with smaller number of inhabitants has
bigger differences (as far as about 55%) and other way round with more people in
object is difference smaller. Smaller daily inequality coefficient values mean lower
hourly take-off water differences which is influenced by fact that was measured
only cold water and there are no taken in peak hot water take-offs. In table 4, we
can see percentage comparison calculated khmax for individual days with average
hourly maximal coefficient during week. Table shows, that weekly average value
of hourly inequality coefficient is with regard to small daily value differences

sufficient.
Acknowledgement
This paper would not be created without excellent cooperation and technical
facilities of Waterworks Pilsen, a.s which ensured measurement and date
collection. Paper was created with support of investigative intention CEZ MSM
6840770003.
References
1. Kadlecová M., Notices to new Czech norm of hot water design, Topenářství Instalace 3/99
2. Kabele K. and collective, Energy and Environmental building systems, CTU 2005
3. Czech notice of Agriculture Ministry 428/2001
4. Vrána J. Water and Sewerage in Home and Flat, Grada publishing 2005
5. Musil V.and collective, Technical Buildings Facilities I
Improving Civil Engineering Physics Teaching-Learning with

Mathematica 5.1
Irina Radinschi1, Brindusa Ciobanu2
1
Department of Physics, “Gh. Asachi” Technical University, Iasi, 700050, Romania
2
Summary
In this paper we present some important examples of Mathematica 5.1 techniques

for studying physical concepts and phenomena and for plotting a general class of
physics phenomena. We point out that a computer algebra system like Mathematica
is very useful for physics studies. We have a long-term experience in using this
program in physics teaching and learning, as in approached fields as mathematical
physics and computational physics.
Mathematica 5.1 is described through some examples of plotting waves
interference, constructive interference and destructive interference. These
examples highlight the power and versatility of Mathematica and indicate its
application to a much wider range of problems in computational physics.
Mathematica 5.1 assists the students for solving differential equations,
differentiating, integrating, making sums, finding roots, plotting and producing
animations. We emphasize the graphical capability of Mathematica and use its
power to go beyond finding solutions and performing length calculations and bring
the studies alive with animations, and other graphical tools. This is possible
because Mathematica 5.1 provides a good environment for computation, a high-
level programming language, text, graphics, and animation. Further, GrTensorM
which is embedded in Mathematica is a package which allows us carrying out
length calculations and evaluating the components of some tensorial and
pseudotensorial quantities.
For these reasons we implement Mathematica 5.1 in our course of physics, for
exploring, teaching, and applying powerful mathematical methods in physics. This
computer algebra system allows the students to develop their ability to learn
physics, to solve difficult problems with a symbolic computing engine, and to
review the basic mathematics. We introduce Mathematica program into our
courses for students who major in technical fields because is a reliable tool for
teaching and learning physics in a modern way.
KEYWORDS: Mathematica 5.1, engineering physics teaching-learning,computer

algebra system, physics, physical concepts and phenomena.
1. INTRODUCTION
It is well-known that nowadays, physics studies are not allowed without using
adequate software. Mathematica is a powerful mathematical software system for
teachers and students. In the recent years, Mathematica has changed the way
physics is taught and teached. Mathematica 5.1 gives the possibility to visualize
and display physics concepts and phenomena, to perform length calculations and to
generate numerical and graphical solutions to physics problems. It also provides
modern techniques and for plotting 2D and 3D having many options for graph
drawing.
Our experience in computational physics [1]-[6], [10]-[14] and particularly in using
Mathematica 5.1, make us to decide that a better way for engineering physics
teaching-learning is to incorporate this computer algebra system in our course of
physics. Mathematica 5.1 is a very powerful and useful general purpose program
[7]-[8] which can find solutions to algebraic equations, it can do calculus, it can
evaluate equations numerically and it can plot and generate animations and sounds.
Mathematica 5.1 also covers essential physical situations and phenomena. This
allows the students leaving the length calculations to Mathematica and pay more
attention to the physics studies. Mathematica has also a powerful online help built
in, which can be used even by the new users [7]-[8].
Mathematica is a reliable tool of choice, in engineering analysis and modeling, and
in technical education in universities. Mathematica 5.1 combines powerful
computing software with a convenient user interface. Mathematica's features
include symbolic and high-performance numeric computation, 2D and 3D data
visualization, broad programming capabilities, and one-step creation of web
documents [9].
2. PRESENTATION OF MATHEMATICA 5.1 APPLICATIONS
2.1. Applications of Mathematica 5.1 in physics
Mathematica 5.1 includes advanced programming methods from modern computer

science as well as adding a host of new ideas of its own [9].
Mathematica 5.1 applications in physics studies allow the students to test their
understanding of the concepts and phenomena, and applying them to many
physical situations. It also allows solving rapidly physics problems without
performing length calculations, and plotting and producing animations. We show
how Mathematica 5.1 will be applied to take the students beyond the limitations of
182 I. Radinschi, B. Ciobanu
traditional instruction. We present some Mathematica 5.1 applications in physics

which are implemented in our courses.
2.2.1. Mathematica 5.1 commands for studying the wave interference. Plotting 2D
interference.
The wave interference is an interesting topic, and with the graphical and animation
capability of Mathematica, we can improve our physics courses. Once the students
are able to handle the equations which describe the wave interference, they also
discover results on their own by varying the amplitudes, the values of pulsation,
initial phase and time, and generating constructive and destructive interference.
Mathematica 5.1 is also a reliable tool for graph drawing. It has many options for
the color of the curves and for the color of the background. These can be set by
using the option Graphics`Colors` and choosing from there the colors. The built-in
Mathematica color directives RGBColor, Hue, and CMYKColor handle the most
common systems. This package gives color directives using other standard
systems. For axes we use the option AxesLabel and we plot the physical quantities
on Z-axis, Y-axis and X-axis.
The Mathematica 5.1 commands for plotting 2D the wave interference are given
below. We give some examples for constructive and destructive interference.
-Graphics-
ln[1]:=Plot[{0.2*Sin[Pi/4*t+Pi/3],-0.2*Sin[Pi/4*t+Pi/3]},
{t,1,10},PlotStyle→{{Hue[0.1]},{Hue[0.2]}},Background→RGBColor[0.94174`,
1.`,1.`],AxesLabel→{t, Ψ}]

0.2
0.1
t
2 4 6 8 10
-0.1
-0.2
Figure1. Plotting two waves
Out[1]= -Graphics-
ln[2]:=Plot[{0.2*Sin[Pi/4*t+Pi/3],0.4*Sin[Pi/4*t+Pi/7],0.2*Sin[Pi/4*t+Pi/3]+0.4*
Sin[Pi/4*t+Pi/7]},{t,1,10},PlotStyle→{{Hue[0.1]},{Hue[0.2],
{Hue[0.3]}},Background→RGBColor[0.94174`,1.`,1.`], AxesLabel→{t, Ψ}]

0.6
0.4
0.2
t
2 4 6 8 10
-0.2
-0.4
-0.6
Figure 2. Plotting 2D interference
Out[2]= -Graphics-
In this case, the resultant wave function is evaluated using the command
ln[3]:= TrigExpand[0.2*Sin[Pi/4*t+Pi/3]+0.4*Sin[Pi/4*t+Pi/7]]
Out[3]=0.346759*Cos[Pi/4*t]+0.460388*Sin[Pi/4*t]
πt πt
Out[3]= 0.346759 Cos[ ] + 0.460388 Sin[ ]
4 4
ln[4]:=Plot[{0.2*Sin[Pi/4*t],0.4*Sin[Pi/4*t],0.2*Sin[Pi/4*t]+0.4*Sin[Pi/4*t]},{t,0.
1,50},PlotStyle→{{Hue[0.1]},{Hue[0.2]},{Hue[0.7]}}, Background→RGBColor
[0.94174`,1.`,1.`],AxesLabel→{t,ψ}]

0.6
0.4
0.2
t
10 20 30 40 50
-0.2
-0.4
-0.6
Figure 3. Plotting constructive interference

Out[4]= -Graphics-
The resultant wave function is computed with the command
ln[5]:= TrigExpand[0.2*Sin[Pi/4*t]+0.4*Sin[Pi/4*t]]
Out[5]=0.6*Sin[Pi/4*t]
πt
Out[5]= 0.6 Sin[ ]
4
ln[6]:=Plot[{0.2*Sin[Pi/4*t+Pi/3],-0.2*Sin[Pi/4*t+Pi/3], 0.2*Sin[Pi/4*t+Pi/3]-
0.2*Sin[Pi/4*t+Pi/3]}, {t,1,10}, PlotStyle→{{Hue[0.1]}, {Hue[0.2]}},
Background→RGBColor[0.94174`,1.`,1.`], AxesLabel→{t, Ψ}]

0.2
0.1
t
2 4 6 8 10
-0.1
-0.2
Figure 4. Plotting destructive interference
Out[6]= -Graphics-
ln[7]:=Plot[{0.2*Sin[Pi/4*t],0.2*Sin[Pi/4*t+Pi],0.2*Sin[Pi/4*t]+0.2*Sin[Pi/4*t+Pi
]},{t,0.1,50},PlotStyle→{{Hue[0.1]},{Hue[0.2]},{Hue[0.7]}},
Background→RGBColor [0.94174`,1.`,1.`],AxesLabel→{t, Ψ }]

0.2
0.1
t
10 20 30 40 50
-0.1
-0.2
Figure 5. Plotting destructive interference

Out[7]= -Graphics-
2.2.2. Mathematica 5.1 commands for plotting 3D wave interference
We give an example of plotting 3D the interference. The Mathematica 5.1

commands for plotting 3D waves interference are presented.
ln[8]:=p1=Plot3D[0.2*Sin[ω*t],{t,1,10},{ω,Pi/7,Pi/4},Background→RGBColor
[0.94174`,1.`,1.`],AxesLabel→{t,ω,Ψ}]
0.2
0.1

0
-0.1 0.7
-0.2
2 0.6 
4
6
t 0.5
8
10
Figure 6. Plotting 3D wave function
Out[8]= -SurfaceGraphics-
ln[9]:=p2=Plot3D[0.2*Sin[ω*t+Pi],{t,1,10},{ω,Pi/7,Pi/4},Background→RGBColo
r [0.94174`,1.`,1.`],AxesLabel→{t,ω,Ψ}]
0.2
0.1

0
-0.1 0.7
-0.2
2 0.6 
4
6
t 0.5
8
10

ln[10]:= -SurfaceGraphics-
Show[{p1,p2},Axes→True]
0.2
0.1

0
-0.1 0.7
-0.2
0 0.6 
2.5
5
t 0.5
7.5
10
Figure 8. Plotting 3D destructive waves interference
The sum of two wave functions can also be plotted using the commands given
below.
ln[11]:=p3=Plot3D[0.2*Sin[ω*t]+0.2*Sin[ω*t+Pi],{t,1,10},{ω,Pi/7,Pi/4},Backgro
und→RGBColor [0.94174`,1.`,1.`],AxesLabel→{t,ω,Ψ}]
1 10 - 1 6
 0
-1  10 - 1 6 0.7
2 0.6 
4
6 0.5
t
8
10
Figure 9. 3D graph for destructive waves interference

The resultant wave function for destructive waves interference is computed using
the command
ln[12]:= TrigExpand[0.2*Sin[ω*t]+0.4*Sin[ω*t]]
Out[11]=0*Sin[ω*t]
For plotting constructive interference we use the next commands.
ln[13] =p3=Plot3D[0.2*Sin[ω*t+Pi/3],{t,1,10},{ω,-
Pi/4,Pi/4},Background→RGBColor [0.94174`,1.`,1.`],AxesLabel→{t,ω,Ψ}]
0.2
0.1

0
0.5
-0.1
-0.2
0 
2
4
6 -0.5
t
8
10
ln[14] =p4=Plot3D[0.2*Sin[ω*t-Pi/4],{t,1,10},{ω,-
0.2
0.1

0
0.5
-0.1
-0.2
0 
2
4
6 -0.5
t
8
10
The resultant wave function is computed using the command
ln[15] =TrigExpand[Sin[ω*t+Pi/3]+0.2*Sin[ω*t-Pi/4]]
Out[15]=0.0317837*Cos[ω*t]+0.241421*Sin[ω*t]
ln[16] Plot3D[0.0317837*Cos[ω*t]+0.241421*Sin[ω*t],{t,1,10},{ω,-
0.2

0
0.5
-0.2
0 
2
4
6 -0.5
t
8
10
Figure 12. Plotting 3D constructive waves interference
Plot3D includes a setting for the PlotPoints and one for the Mesh option. For the
graph from the Figure 12 we use these options and we obtain
ln[17] Plot3D[0.0317837*Cos[ω*t]+0.241421*Sin[ω*t],{t,1,10},{ω,-
Pi/4,Pi/4}, Background→RGBColor [0.94174`,1.`,1.`], PlotPoints→40,
Mesh→False, AxesLabel→{t,ω,Ψ}]
0.4
0.2

0
0.5
-0.2
-0.4
0 
2
4
6 -0.5
t
8
10
Figure 13. Plotting 3D constructive waves interference
3. CONCLUSIONS
The latest version Mathematica 5.1 raises the level in mathematical software,
providing more than 50 new functions, toolkits, and performance improvements in
a powerful program [15]-[16]. The Mathematica 5.1 version adds a host of new
capabilities, especially for working with large-scale, diverse types of data. New
innovative algorithms are introduced to deliver unmatched performance for all
steps in the data handling process, importing, analyzing, manipulating, or plotting.
We have introduced Mathematica 5.1 as a teaching and learning tool in our course
of physics for civil engineering. Mathematica 5.1 is a powerful program and has
many advantages. It provides a good environment for computation, a high-level
programming language, text, graphics, and animation. We assume that the students
are already familiar with Mathematica, but we also offer a basic training in this
computer algebra system for beginers.
The cases of Plot and Plot3D wave interferences was illustrated with some
examples and commands for plotting 2D and 3D waves interference, and also
options like PlotStyle, Background→RGBColor, AxesLabel, PlotPoints, and Mesh

have been added for improving the graph drawing.
Our main goals are to implement Mathematica 5.1 and Maple 9.5 programs to
assist students in physics learning and for solving problems. We want our students
to make progresses and we also want them to make these progresses rapidly. On
the other hand, we’ll give them the possibility to choose between Mathematica 5.1
and Maple 9.5, or to work with both programs.
References
1. Ciobanu B., Radinschi I., One Computational Algorithm for Physics Modeling, Proceedings of 5th
International Conference on Electromechanical and Power Systems, SIELMEN, Chisinau, Rep.
Moldova, October 6-8, 2005, pp. 222-225.
2. Radinschi I., Ciobanu B., Testing Physics, Junimea Publishing House, Iasi, Romania, 2006.
3. Radinschi I., Ciobanu B., Physics for Engineers, Junimea Publishing House, Iasi, Romania, 2006.
4. Ciobanu B., Radinschi I., A Sequential Programming Algorithm – a Tool for Study the Standing
Waves, Proceedings of International Symposium IEEI 2006, Iasi, Romania, Bulletin of IP Iasi,
Vol. LII(LVI), Fasc.5, pp. 129-134.
5. Radinschi I., Scripcariu L., Ciobanu B., Frunza M. D., Online Teaching-while-Quizzing Test,
Proceedings of the 2nd National Conference on Applied Physics, June 9-10, Galati, Romania,
2006, in press Romanian Journal of Physics
6. Ciobanu B., Radinschi I., Teaching and Learning Physics for Engineers by Aid Computers,
Proceedings of International Symposium IEEI 2006, Iasi, Romania, Bulletin of IP Iasi, Vol.
LII(LVI), Fasc.5, pp. 135-140.
7. Mathematica for Scientists and Engineers: Using Mathematica to do Science, by Richard Gass
(Prentice Hall).
8. www.wolfram.com.
9. http://www.pugh.co.uk.
10. Radinschi I., Frunza M. D., Ciobanu B., Online Virtual Model for Testing the Knowledge,
Proceedings of INTED 2007, March 7-9, Valencia, Spain.
11. Radinschi I., Ciobanu B., Implementation of Computational Methods in Physics Learning,
Proceedings of 4th International Symposium Computational Civil Engineering 2006, Iasi,
Romania, 2006, pp. 251-257.
12. Frunza M. D., Radinschi I., A Virtual Model for Testing Mathematics, Physics and Chemistry,
Proceedings of International Symposium for Design and Technology of Electronic Packages,
SIITME, 2006, p. 154, Iasi, Romania.
13. Ciobanu B., Radinschi I., The Electrical Properties of sample Investigation by Computational
Programs, Proceedings of 5th International Conference on Electromechanical and Power Systems,
SIELMEN, October 6-8, 2005-Chisinau, Rep. Moldova, 2. p. 218-221.
14. Ciobanu B., Radinschi I., One computational program for predictions of properties of the yarns,
Proceedings of International Symposium Present and Perspective in Textile Engineering,
November 10-12, 2005, Iaşi, Romania, p. 563-569.
15. www.gtpcc.org
16. www.pcmag.com; www.usc-b2b.com; www.cambridge.org; www.uc.edu.
Physics with Maple 9.5

Irina Radinschi1, Brindusa Ciobanu1, Mircea Daniel Frunza2
1
2
Faculty of Electronics and Communications, “Gh. Asachi” Technical University, Iasi, Romania
Summary
In this paper we sketch how computer algebra systems like Maple 9.5 can be used
to helping the students connecting physical concepts and phenomena with powerful
mathematical formalisms and graphical representations. We demonstrate that a
computer algebra system like Maple is very useful for physics studies. The
experience in this area demonstrated that computational physics, together with
traditional theoretical and experimental physics, is a reliable tool which helps
students to explore a wide variety of phenomena and give them a deeper
understanding of these topics.
Maple 9.5 provides a good environment for computation, a high-level
programming language, text, graphics, and animation. Maple 9.5 is also capable
of, and used for, high-level computation by students and this making it an ideal
computer assisted teaching and learning tool. It is also a comprehensive
environment for exploring, teaching, and applying mathematical methods in
physics. Using Maple the students will be unconstrained by traditional
mathematical and physical limitations, and they will also be capable of a better
understanding of computational methods. Maple 9.5 for physics helps the students
to develop their ability to learn physics, to solve difficult problems with a symbolic
computing engine, and to review the basic mathematics. Maple 9.5 assists the
students for solving differential equations, differentiating, integrating, making
sums, finding roots, plotting and producing animations. In this work are presented
some applications of Maple 9.5 in physics, which help the students to bridge the
gap between mathematical formalism and physical phenomena.
Step by step we want to introduce Maple program into our courses for those who
plan to major in technical fields. The main features of our course are the use of
computer program like Maple for teaching and learning physics, and for solving
physics problems. The students can also use the computers outside class for further
instruction and problem solving. In this way we want to take our students beyond
the limitations of traditional instruction and make them to become familiar with
high-level computer algebra systems.
KEYWORDS: Maple 9.5, computer algebra system, physics, physical concepts

and phenomena.
192 I. Radinschi, B. Ciobanu, M.D. Frunza
1. INTRODUCTION
Analyses of current developments in physics teaching and learning have shown the
importance of implementation of high-level computer algebra systems like Maple
[7] in our courses. Maple is a mathematical application package that supports
symbolic and numeric computation and graphics. Maple 9.5 helps students
connecting physical concepts and phenomena with powerful mathematical
formalisms and graphical representations. The extensive mathematical assistance,
symbolic manipulations, computational power and graphical abilities of programs
like Maple [11]-[14] can greatly help students to explore physical topics and
experiment with ideas without performing cumbersome calculations.
We have a long-term experience in computational physics [1]-[6], [8]-[10] and
particularly in using Maple 9.5 program, with applications in many fields of
physics. Our course will be designed to incorporate the use of Maple for solving
differential equations, differentiating, integrating, making sums, finding roots,
plotting and producing animations.
Students use Maple V software to achieve a better conceptual understanding of the
material of the course while still gaining a good knowledge of the methods of
problem-solving. Further, this computational method allows a rapid learning of
physics and is also a powerful tool for solving problems.
2. PRESENTATION OF MAPLE 9.5 APPLICATIONS
2.1. Applications of Maple 9.5 in physics
Maple 9.5 applications in physics gives students the opportunity to show what they
have learned by testing their understanding of the concepts and phenomena, and
applying them to real situations. Further, it allows solving rapidly physics problems
without performing length calculations, plotting and producing animations. We
present some Maple 9.5 applications in physics which are implemented in our
courses.
2.2.1. Evaluation of speed and displacement
> ode1 := m*diff(x(t),t)=4*t^2/3;

4 2
ode1:= m ( d x(t))÷= t
dt 3
> ans1 := dsolve(ode1);
4 t3
ans1 := x(t) = + _C1
9m
and
> diff(4*t^3/(9*m),t);
4 t2
3m
2.2.2. Plotting the speed and displacement
The commands of Maple 9.5 for plotting 2d, for producing animations and for
plotting 3d for the speed are:
a) plotting 2d
>with(plots);
>plot(4*t^2/3,t=0...10);
Figure 1. Graph 2d of the speed
b) producing animations
>with(plots);
>animate(plot,[A*(4*t^2/3),t=0...10], A=-5...5);
Figure 2. Animation 2d graph of the speed
c) plotting 3d
>with(plots);
>plot3d(4*t^2/(3*m),m=0.1...1,t=0...10);
Figure 3. Plot 3d of the speed

The commands of Maple 9.5 for plotting 2d, for producing animations and for
plotting 3d (in this case we consider that the integrating constant is C1=0) for the
displacement are
a) plotting 2d
>with(plots);
>plot(4*t^3/9,t=0...10);
Figure 4. Plot 2d of the displacement
b) producing animations
>with(plots);
>animate(plot,[A*(4*t^3/9),t=0...10], A=-5...5);
Figure 5. Animation 2d graph of the displacement
c) plotting 3d
>with(plots);
>plot3d(4*t^3/(9*m),m=0.1...1,t=0...10);
Figure 6. Plot 3d of the displacement

2.2.3. Plotting the wave propagation
There are considered three wave functions. The commands of Maple 9.5 for
plotting the waves propagation in 2d and 3d cases are given below.
a) plotting 2d
>with(plots);
>f:=4*sin(Pi/4*t+Pi/7);
>h:=2*cos(Pi/4*t+Pi/3) ;
>g:=0.7*sin(2*Pi/3*t+Pi/7);
> plot({f(t),g(t),h(t)},t=0...10);
Figure 7. Plot 2d of the wave propagation
b) plotting 3d
>with(plots);
>f:=4*sin(omega*t+Pi/7);
>g:=0.7*sin(omega*Pi/3*t+Pi/7);
>h:=2*cos(Pi/4*t+Pi/3) ;
>plot3d({f,g,h},omega=0...Pi/4,t=0...10);
Figure 8. Plot 3d of the wave propagation
2.2.4. Plotting the kinetic and potential energy
The commands of Maple 9.5 for plotting the kinetic energy and the potential
energy (with respect of SI units) are given bellow
>with(plots);
>plot3d({m*v^2/2,m=0.1...1,v=0...10);
Figure 9. Plot 3d of the kinetic energy
>plot3d({k*x^2/2,k=0.1...10,x=1...100);
Figure 10. Plot 3d of the potential energy
We gave some examples which demonstrate the usefulness of Maple 9.5 for
teaching and learning physics.
Maple 9.5 can also be used as a powerful tool for solving physics problems. After
the students will become familiar to most of Maple commands they will be capable
to solve more complicated problems.
3. CONCLUSIONS
This paper points out some advantages of using Maple [7], [11]-[14] in studying
physics. We presented some examples from our course to demonstrate the
usefulness of this high-level computer algebra system. We used Maple 9.5 for
evaluating the speed and displacement in a particular case, for plotting 2d and 3d of
the speed and displacement, and for producing animations. The program was also
used for plotting the wave propagation of three waves in 2d and 3d cases, and for
plotting the kinetic energy and the potential energy in 3d case.
The main goals are to allow students the ability to learn physics, to perform length
calculations, to make graphical representations and introduce animations, and also
to solve physics problems using Maple 9.5. The applications will be introduced
step by step, firstly students have to learn how to handle with derivatives and
integrals, solving equations and making graphs. At a higher level, they will be able
to solve challenging and physically realistic problems.
References
1. Ciobanu B., Radinschi, I., One Computational Algorithm for Physics Modeling, Proceedings of
5th International Conference on Electromechanical and Power Systems, SIELMEN, Chisinau,
Rep. Moldova, October 6-8, 2005, pp. 222-225.
Vol. LII(LVI), Fasc.5, pp. 129-134
7. http://grtensor.org; http://www.maplesoft.com/maplebooks.html.
Romania, 2006, pp. 251-257.
11. http://www.wfu.edu/physics/cel/maple.html
12. http://www.maplesoft.com/applications/
13. www-teaching.physics.ox.ac.uk
14. www.adsabs.harvard.edu
Physics Studies – Computational Methods, a Strong Connection

1
Department of Physics, “Gh. Asachi” Technical University, Iasi, 700050, Romania,
radinschi@yahoo.com
2
bciobanu2003@yahoo.com
Summary
The aim of this paper is to point out the importance of the connection between
theoretical physics and computational methods. At national and international level,
considerable efforts have been made for implementing the computational methods
in the study of physics. This work is focused on the use of Mathematica and Maple
programs for physics studies. Because our applications are part of the algebraic
computations in physics, new calculation programs in Mathematica and Maple are
conceived. We also used methods of animated graphs for illustrating the physics
phenomena. The calculations are performed with Mathematica and Maple
programs. In some particular cases of our applications, these programs have
attached the GrTensor packages. Many graphs were also made using the
Mathematica and Maple programs. This goal will be reached easily establishing
good collaborations with specialists in the field from our country and from abroad.
On the other hands, undergraduate physics implies to work with students and in
this way they have the possibility to enlarge the circle of collaborations and gather
a lot of information in the field and in adjacent areas. Platforms as Mathematica
and Maple have some advantages as flexibility and speed, and more advanced
graphical facilities. GrTensorII which is embedded in Maple is a package which
allows us carrying out length calculations and evaluating the components of some
tensorial and pseudotensorial quantities. This is because it has the facility for
defining new tensors. Furthermore, it allows the simplication of large terms. The
duration of the calculations is shorter compared to that which used other software
such as REDUCE. GrTensorM runs under Mathematica. GrTensorM has a large
number of predefined objects, with automatic index generation and facilitates the
creation of new objects with a compact intuitive interactive definition facility. An
important step is the development of the existing methods, and in this light we
intend to enlarge the areas of study by using the Mathematica and Maple programs
to more classes of applications. In this way a powerful connection between physics
studies and computational methods is established.
KEYWORDS: physics studies, computational methods, Mathematica, Maple,

GrTensorII, GrTensorM.
1. INTRODUCTION
The connection between the physics studies and the computational methods is one
of the most important issues at national and international level. The mathematical
methods implemented in our physics course represent the main way of instruction
and provide the skills needed for advanced studies in physics. Mathematica and
Maple have the potential to strongly impact the way problem solving methods are
taught in undergraduate physics and the curriculum of such courses. In order to
perform length calculations, evaluate the components of some tensorial and
pseudotensorial quantities and make graphs we use powerful programs like
Mathematica and Maple. In physics we want to use more animated graphs for
illustrating the phenomena. Nowadays, physics studies are not allowed without
using adequate software. Because the use of good software is very important, this
is the reason why specialists in the field from our country and from abroad have
been consulted. Further, special attention was paid for involving our students in
this interactive study: physical phenomena-computational methods.
We consider that this work can be extended to many physical situations and a lot of
information in the field can be achieved. Because the study of physics requires in
many cases making graphs and carrying out calculations which can be performed
using computer algebra methods, we strongly recommend programs like
Mathematica and Maple.
We have a long-term experience in computational physics [1]-[3], [5]-[7] and
particularly in using these programs, with applications in many fields of physics.
Our course will be designed to incorporate the use of the Mathematica and Maple
programs as the primary means of teaching mathematical methods for solving
physics problems. Further, we want to understand how this course impacts the
students’ understanding of the course material, and to what extent Mathematica and
Maple can be reliable problem solving tool for students at a similar level in their
study of physics. This is an important point, because one of the main problems that
would arise is how will be the course instruction affected by the use of these two
computer programs and what are the strengths and weaknesses of Mathematica and
Maple as teaching and learning tools.
These studies imply to work with students, and in this way they have the possibility
to enlarge the circle of collaborations and gather a lot of information in the field
and in adjacent areas. Further, the computational methods allow a rapid learning of
physics and our viewpoint is they can succesfully replace the traditional methods.
We hope that our students will feel confident and they will be able to handle with
the Mathematica and Maple programs, in order to easily and rapidly solve and
explore most problems they encounter in their physics studies, and enrich their
knowledge.
2. CONNECTION BETWEEN PHYSICS STUDIES AND

COMPUTATIONAL METHODS
2.1. Presentation of Mathematica and Maple programs
As we pointed out in Introduction, the study of physics requires adequate software.

In recent years computer algebra systems have become increasingly powerful tools
in teaching physics. The extensive mathematical assistance, symbolic
manipulations, computational power and graphical abilities of programs like
Mathematica and Maple [4] can greatly help students to explore physical topics and
experiment with ideas without performing cumbersome calculations. We consider
that we are able to obtain reliable results in teaching and learning physics, and this
is due to the good connection between the physics studies and the methods of
computer calculation and graphic simulation. The applications which are obtained
for many physical situations also allow to propose some useful suggestions for
future studies. Concerning our courses and seminars we want to perform the
calculations with Mathematica and Maple. Many graphs will also be made using
the Mathematica and Maple programs. Special attention will be paid to animated
graphs for illustrating the physics phenomena. In the whole process of teaching and
learning physics we will take into account that many of our students are novice
users of these programs, and will choose the adequate level of applications.
Platforms as Mathematica and Maple have some advantages as flexibility and
speed, and more advanced graphical facilities. GrTensorII is a package which
allows calculating the components of tensorial and pseudotensorial quantities, and
is embedded in Maple. This is because it has the facility for defining new tensors.
Furthermore, it allows the simplification of large terms. GrTensorM runs under
Mathematica. GrTensorM has a large number of predefined objects, with automatic
index generation and facilitates the creation of new objects with a compact intuitive
interactive definition facility. We have a long-term experience in using these
programs in physics teaching and learning, as in approached fields as mathematical
physics and computational physics.
2.2. Advantages in using Mathematica and Maple for teaching and learning
physics
We sketch the advantages of these powerful programs in teaching and learning

physics. It is obviously that Mathematica and Maple program have to satisfy some
requirements which indicated them as powerful tools. These requirements are:
1) to be a good mathematical tool for solving equations, carrying out length
calculations, and writing tensorial and pseudotensorial quantities,
2) Mathematica and Maple computer algebra systems have to allow the students to
easily and quickly change the mathematical models and the parameters and then
generate new answers,
3) giving the possibility to be applied for a large number of applications and
yielding meaningful results,
4) facilitating to make predictions and analogies between different physical
situation,
5) giving the possibility to make graphs and animated graphs in order to illustrate
the physical phenomena,
6) having the entire documentation electronically searchable.
These features make Mathematica and Maple ideal for use in upper-level physics
courses for students completing assignments.
3. APPLICATIONS OF MATHEMATICA AND MAPLE
3.1. Maple 9.5 applications
We present some examples from a typical calculus course. Further, some graphs
are inserted. The topics that we have chosen are solving differential equations,
evaluating the first and second derivative of some physical quantities and making
graphs and animated graphs. The following examples, solutions and graphs
demonstrate such cases.
3.1.1. Maple commands for simple harmonic oscillations and damped harmonic
oscillations
Firstly, we consider the simple harmonic oscillations and damped harmonic

oscillations and we give the Maple sequences for dsolve – solve ordinary
differential equations (ODEs). The first example is using the command
dsolve(ode1) which yields the roots of ode1. In the second case we use the
command dsolve(ode2) which yields the roots of ode2.
> ode1 := diff(x(t),t,t)+omega^2*x(t);
ode1 := (diff(x(t),(t, 2)))+omega^2*x(t)
ans1 := x(t) = _C1 sin(w t) + _C2 cos(w t)
and
> ode2 := diff(x(t),t,t)+2*gamma*diff(x(t),t)+omega^2*x(t);
ode2 := (diff(x(t),(t, 2)))+2*gamma*(diff(x(t), t))+omega^2*x(t)
ans2 := x(t) = _C1*exp((-gamma+(gamma^2-omega^2)^(1/2))*t)+_C2*exp((-
gamma-(gamma^2-omega^2)^(1/2))*t)
3.1.2. Maple commands for evaluating the first and second derivative of
displacement for the simple harmonic oscillations
Secondly, we give some examples of evaluating the first and second derivative of
displacement for the simple harmonic oscillations. This will give the velocity and
acceleration for the simple harmonic oscillations.
> diff(A*sin(omega*t+phi),t);
A cos(w t + f ) w
> diff(A*sin(omega*t+phi),t,t);
2
-A sin(w t + f ) w
3.1.3. Maple commands for plotting 2d and 3d for simple harmonic oscillations
Thirdly, we make the graphs for simple harmonic oscillations. We make a 2 –

dimensional graph and a 3 – dimensional graph.
For the 2 – dimensional case we plot the expression for the displacement with the
amplitude A=2, the pulsation ω=Pi/4 and the initial phase φ=Pi/3 and t=0…10, and
we obtain the graph from the Figure 1.
Figure1. Simple harmonic oscillations, the 2d graph

> with(plots);
> plot(2*sin(Pi/4*t+Pi/3),t=0...10);
For the 3 – dimensional case we plot the expression for the displacement with the
amplitude A=0.1…1, the pulsation ω=Pi/4, the initial phase φ=Pi/3, and t=0…10.
We give the sequences bellow and in the Figure 2 we plot the displacement for
different values of the amplitude and time.
> plot3d(A*sin(Pi/4*t+Pi/3),A=0.1...1,t=0...10);
Figure 2. Simple harmonic oscillations, the 3d graph
For the graphs the selection animate or animate3d can be selected.
3.2. Mathematica 5.1 applications
We give some examples connected to the topics of solving differential equations,

evaluating the first and second derivative of some physical quantities and making
graphs and animated graphs.
3.2.1. Mathematica commands for simple harmonic oscillations and damped

harmonic oscillations
We present the simple harmonic oscillations and damped harmonic oscillations and
we give the Mathematica sequences for DSolve. We obtain
ln[1] DSolve[x''[t] == -ω^2 x[t], x, t]

Out[1] {{x→Function[{t},C[1] Cos[ω t]+C[2] Sin[ω t]]}}
and
ln[2] DSolve[x''[t] == 2 δ x'[t] + ω^2 x[t], x, t]
Out[2] {{x→Function[{t}, e ( −δ − δ 2 −ω 2 ) t
C[1] + e ( −δ + δ 2 −ω 2 ) t
C[2]
3.2.2. Mathematica commands for evaluating the speed and acceleration for the
simple harmonic oscillations
In this case the first and second derivatives of displacement for simple harmonic
oscillations are computed. They will give the velocity and acceleration for the
simple harmonic oscillations.
ln[1] D[A*Sin[ω*t+φ],t]
Out[1]A ω Cos[ω*t+φ]
ln[2] D[A*ω*Cos[ω*t+φ]]
Out[3] -A ω^2 Sin[ω*t+φ]
3.2.3. Mathematica commands for plotting 2d and 3d for simple harmonic

oscillations
c) We plot the displacement for the simple harmonic oscillations. We make a 2-

dimensional graph and a 3-dimensional graph, and use the same values for the
amplitude, pulsation, initial phase and time parameters as in the case of Maple
program.
For the 2-dimensional case we obtain the graph from the Figure 3.
ln[1] Plot[2*Sin[Pi/4*t+Pi/3],{t,1,10}, PlotStyleHue[.1], Background→RGBColor

[0.94174`,1.`,1.`], AxesLabel→{t, Ψ }]
Out[1]-Graphics-

2
t
2 4 6 8 10
-1
-2
There are many options for the color of the curve and for the color of the
background. These can be set by using the option Graphics`Colors` and choosing
from there the colors. The built-in Mathematica color directives RGBColor, Hue,
and CMYKColor handle the most common systems. This package gives color
directives using other standard systems. For axes we use the option AxesLabel and
we plot the wave function Ψ on Y-axis against time t on X-axis.
The 3-dimensional case is illustrated in Figure 4. We plot the expression for the
wave function using the same option for RGBColor as in the 2d case. For axes we
use the option AxesLabel and we plot the wave function Ψ on Z-axis, against
amplitude A on X-axis and time t on Y-axis.
ln[2] Plot3D[A*Sin[Pi/4*t+Pi/3], {A,0.1,1}, {t,1,10}, Background→RGBColor

[0.94174`,1.`,1.`], AxesLabel→{A,t,Ψ}]
Out[2] -SurfaceGraphics-
1
0.5
 10
0
-0.5 8
-1
6
0.2 t
0.4 4
0.6
A 2
0.8
1

ln[3] Show[f,Mesh→False]
Out[3] -SurfaceGraphics-
1
0.5 10
0
-0.5 8
-1
6
0.2
0.4 4
0.6
0.8 2
1
Further, the SelectionAnimate or Sound Generation can be selected. Mathematica

is a powerful computational tool which provides a large spectrum of choices for
SelectionAnimate and for Sound Generation. It also provides simulations of
relevant concepts.
In some particular cases of our applications, Mathematica and Maple programs
have attached the GrTensor packages. These are useful for evaluating the
components of some tensorial and pseudotensorial quantities. Important
calculations can be carried out with GrTensorII and GrTensorM for studies
concerning the general relativity. These packages allow evaluating the determinant
of a metric, the covariant and contravariant components of the metric tensor, the
components of the Einstein and Ricci tensor, and also to define new tensors and
pseudotensors.
We also strongly encouraged the students to use either Mathematica or Maple to
work on the homework problems.
4. CONCLUSIONS
In this paper, we examined the connection between physics studies and Computer
Algebra Systems. Teaching and learning undergraduate physics using Mathematica
5.1 and Maple 9.5 is described through examples. Both Mathematica 5.1 and
Maple 9.5 are very useful for learning, teaching, and carrying out research in
physics.
We presented some examples from our course to demonstrate the usefulness of

these computer programs. We have chosen the cases of simple harmonic
oscillations and damped harmonic oscillations. Mathematica and Maple are used
for solving the equation of motion in these two cases, for evaluating the velocity
and acceleration in the case of simple harmonic oscillations and for plotting the
displacement in both cases. We work with Mathematica 5.1 and Maple 9.5. For
evaluating tensorial and pseudotensorial quantities the GrTensorII and GrTensorM
packages are attached to Mathematica and Maple programs, respectively.
We will implement Mathematica and Maple as teaching and learning tools in our
course of physics for civil engineering taking into account the advantages of these
programs.
References
1. Ciobanu B., Radinschi I., One Computational Algorithm for Physics Modeling, Proceedings of 5th
International Conference on Electromechanical and Power Systems, SIELMEN, Chisinau, Rep.
Moldova, October 6-8, 2005, pp. 222-225.
Romania, 2006, pp. 251-257.
Vol. LII(LVI), Fasc.5, pp. 129-134.
Computer Algebra System for Energy-Momentum Localization

1
radinschi@yahoo.com
2
bciobanu2003@yahoo.com
Summary
The aim of this paper is to sketch how computer algebra systems like Maple 9.5
with the GrTensorII package is used for performing calculations for energy-
momentum localization. This computer algebra platform Maple and GrTensorII
allows defining new objects like tensorial and pseudotensorial quantities, and for
this purpose there are created procedures with GrTensorII 1.50 version. As the
software Maple allows the use of complicated graphical methods, they were used
to make the graphical representations. Maple has some advantages as flexibility
and speed, and also advanced graphical facilities as animation. GrTensorII which
is embedded in Maple is a package which allows us carrying out length
calculations and evaluating the components of some tensorial and pseudotensorial
quantities. This is because it has the facility for defining new tensors and
pseudotensors. Furthermore, it allows the simplification of large terms and gives
the possibility to evaluate the energy and momentum density components for a
given metric. The duration of the calculations is shorter compared to that which
used other software such as REDUCE. At international level, considerable efforts
have been done to find a generally accepted expression for the energy-momentum
density. This includes working with superenergy tensors, energy-momentum
complexes, quasi-local expressions and tele-parallel quantities. Concerning the
development of the existing theories and methods, we mention the enlargement of
the areas by applying the energy-momentum complexes and their tele-parallel
versions in the case of different 2, 3 and 3+1 dimensional space-times. Special
attention has been paid to the connections between the results and furthermore to
the connections with the tele-parallel theory of gravitation. In this context, we
present some programs elaborated in GrTensorII which have been used for
evaluating the energy and momentum components for the Einstein, Landau-Lifshitz
and Møller energy-momentum complexes and make an investigation of the
connection between physics studies and computational methods. We point out
several important results that the energy-momentum complexes have yield for
different geometries and some scientific perspective for future research in this
area. Further, special attention was also paid for involving our master students in
this interactive study: physical phenomena-computational methods.
KEYWORDS: computer algebra system, Maple, GrTensorII, energy-momentum
localization.
1. INTRODUCTION
It is well-known that a computer algebra system like Maple is very useful for
physics studies, and GrTensorII package [1] can be used for performing length
calculations and evaluating the components of some tensorial and pseudotensorial
quantities. Further, GrTensorII is a reliable tool for evaluating the energy and
momentum components of the energy-momentum complexes of Einstein [2],
Landau-Lifshitz [3] and Møller [4] and gives a deeper understanding of these
prescriptions.
The different attempts at constructing an energy-momentum density don't give a
generally accepted expression. The “center” of general relativity is one of the most
interesting and challenging ideas in modern science, the one that gravity is the
geometry of curved four-dimensional space-time. In conclusion, gravity is
geometry. For localizing the energy there have been used many definitions,
including superenergy-tensors, the energy-momentum complexes, the quasi-local
quantities, and the tele-parallel theory of gravitation. The energy-momentum
complexes definitions include the Einstein [2], Landau-Lifshitz [3], and Møller [4]
prescriptions. These prescriptions have been criticized because of their drawback,
they are coordinate dependent. Except the Møller energy-momentum complex
which allows evaluating the energy distribution in any coordinate system, the other
prescriptions give meaningful results only if the calculations are restricted to quasi-
Cartesian coordinates. There are doubts that these pseudotensorial definitions yield
good results for a given space-time. In this light, some remarks are needed: 1) the
energy-momentum complexes (ELL) are non-covariant, coordinate dependent
expressions and yield acceptable results only the calculations are carried out in
quasi-Cartesian coordinates, 2) only the Møller energy-momentum complex
enables one to calculate the energy in any coordinate system and 3) this agrees with
the equivalence principle which states that gravity cannot be detected at a point.
Chang, Nester and Chen [5] showed that the energy-momentum complexes are
actually quasi-local and legitimate expressions for the energy-momentum. They
concluded that there exist a direct relationship between energy-momentum
complexes and quasi-local expressions because every energy-momentum
complexes is associated with a legitimate Hamiltonian boundary term. In the last
decades the issue of the energy-momentum localization by applying the energy-
momentum complexes was re-opened and many interesting results have been
obtained [6], which demonstrate that these prescriptions are powerful concepts for
energy-momentum localization. All these considerations point out the significance
of these prescriptions and stress the usefulness of Maple program and GrTensorII
package which allow calculating the energy and momentum components for a
given space-time, and also making graphical representations.
2. GRTENSORII PACKAGE AND ENERGY-MOMENTUM

COMPLEXES
2.1. Energy-momentum complexes
2.2.1. General comments
As we pointed out in Introduction, the lack of the pseudotensorial prescriptions is

connected to the dependence of the energy-momentum complexes on the
coordinate systems. Only the Møller prescription allows evaluating the energy and
momentum in any coordinate system. The other energy-momentum complexes of
Einstein and Landau-Lifshitz yield reliable results if the calculations are carried out
in quasi-Cartesian coordinates. In this context, another important problem appears.
According to recent studies, although the energy-momentum complexes behave
under general coordinate transformations like non-tensorial objects, the local
conservation laws obtained by them hold in all coordinate systems. The Einstein
and Landau-Lifshitz energy-momentum complexes give the same energy
distribution as the Penrose definition for a general non-static spherically symmetric
metric of the Kerr-Schild class [7]. In the case of the most general non-static
spherically symmetric metric these definitions disagree concerning the expression
for the energy distribution [7].
2.2.2. Maple and GrTensorII platform for the Einstein, Landau-Lifshitz and
Møller prescriptions
The GrTensorII programs for the Einstein, Landau-Lifshitz and Møller energy-
momentum complexes are given below. Further, examples for a class of solutions
belonging to the string theory are presented [8]-[10].
GrTensorII for the Einstein energy-momentum complex:
>readlib(grii):
>grtensor();
> grdef(`H{k^m^l^n}:=(-detg)*(g{^k^m}*g{^l^n}-g{^l^m}*g{^k^n})`):
Created definition for H(dn,up,up,up)
> grdef(`H1{i^k^l}:=1/sqrt(-detg)*g{i m}*H{k^m^l^n,n}`):
Created a definition for H(dn,up,up,up,pdn)
Created definition for H1(dn,up,up)
GrTensorII for the Landau-Lifshitz energy-momentum complex:

>readlib(grii):
>grtensor();
> grdef(`S3{^m^j^n^k}:=-detg*(g{^m^n}*g{^j^k}-g{^m^k}*g{^j^n})`):
Created definition for S3(up,up,up,up)
> grdef(`S4{^m^n^k}:=S3{^m^j^n^k,j}`):
Created a definition for S3(up,up,up,up,pdn)
Created definition for S4(up,up,up)
GrTensorII for the Møller energy-momentum complex:
> grdef(`M{i^k^l}:=sqrt(-detg)*(g{i m,n}-g{i n,m})*g{^k^n}*g{^l^m}`):
Created definition for M(dn,up,up)
The first two commands in these procedures are for starting the GrTensorII
package. The command grdef is used for defining the required energy-momentum
complexes.
The Einstein, Landau-Lifshitz and Møller energy-momentum complexes are used
for evaluating the energy and momentum components for the space-time which
describes a magnetic stringy solution. The metric is used in the initial form for the
Møller prescription and is transformed in quasi-Cartesian coordinates for the
Einstein and Landau-Lifshitz prescriptions. The following results are obtained:
> qload(kar1);
Calculating ds for kar1 ... Done. (0.000000 sec.)
> grcalc(M(dn,up,up)):
Calculating g(dn,dn,pdn) for kar1 ... Done. (1.000000 sec.)
Calculating detg for kar1 ... Done. (0.000000 sec.)
Calculating g(up,up) for kar1 ... Done. (0.000000 sec.)
Calculating detg for kar1 ... Done. (0.000000 sec.)
Calculating M(dn,up,up) for kar1 ... Done. (0.000000 sec.)
> grdisplay(M(dn,up,up));
> qload(kar);
Calculating ds for kar ... Done. (0.000000 sec.)
> grcalc(H1(dn,up,up)):
Calculating g(dn,dn,pdn) for kar ... Done. (1.000000 sec.)
Calculating detg for kar ... Done. (0.000000 sec.)
Calculating g(up,up) for kar ... Done. (0.000000 sec.)
Calculating H1(dn,up,up) for kar ... Done. (0.000000 sec.)
> grdisplay(H1(dn,up,up));
> qload(kar);
Calculating ds for kar ... Done. (0.000000 sec.)
> grcalc(S3(up,up,up,up,pdn),S4(up,up,up)):
Calculating g(dn,dn,pdn) for kar ... Done. (1.000000 sec.)
Calculating g(up,up) for kar ... Done. (0.000000 sec.)
Calculating S3(up,up,up,up,pdn) for kar ... Done. (0.000000 sec.)
Calculating S4(up,up,up) for kar ... Done. (0.000000 sec.)
> grdisplay(S4(up,up,up));
The command qload in these procedures is for loading the metric, grcalc is
designated for calculating the energy and momentum components for the
aforementioned energy-momentum complexes, and the command grdisplay is used
for displaying the results. The energy components of the Einstein, Landau-Lifshitz
and Møller energy-momentum complexes are involved in calculating the
expression for energy in the case of the aforementioned magnetic stringy solution.
For plotting the expressions for energy obtained in the Einstein, Landau-Lifshitz
and Møller prescriptions on the same graph the next Maple 9.5 commands are
used. We denote with f, g and h the expressions for energy in the Landau-Lifshitz,
Einstein and Møller prescriptions, respectively. In all expressions we consider the
case M=1.
> with(plots);
> f:=r/2*(r*Q^2+2*r-2*Q^2)/((r-2)*(r-Q^2));
> g:=1/2*(r*Q^2+2*r-2*Q^2)/(r-Q^2);
> h:=r/2*(2-Q^2)/(r-Q^2);
> plot3d({f,g,h},r=0...100,Q=0.1...1);
Figure 1. Energy on Z-axis plotted against r on X-axis and Q on Y-axis

3. CONCLUSIONS
In the last decade the problem of the energy-momentum complexes has been re-
opened and acceptable results have been obtained [6] for different space-times.
Many authors used different energy-momentum prescriptions to investigate the
energy distributions for a given geometry. They demonstrated that in many cases
several energy-momentum complexes yield the same and acceptable result. Chang,
Nester and Chen [5] showed that the energy-momentum complexes are actually
quasi-local and legitimate expressions for the energy-momentum. Their idea
supports the energy-momentum complexes and the role which these are playing in
energy-momentum localization. Furthermore, important studies have been done
about the new idea of quasi-local approach for energy-momentum complexes [11]-
[12] and a large class of new pseudotensors connected to the positivity in small
regions have been constructed [13]. In this light, the quasi-local quantities are
associated with a closed 2-surface (L. B. Szabados, [12] and
http://relativity.livingreviews.org/Articles/lrr-2004-4/). The Hamiltonian boundary
term determines the quasi-local quantities for finite regions and the special quasi-
local energy-momentum boundary term expressions correspond each of them to a
physically distinct and geometrically clear boundary condition [14].
Maple and GrTensorII procedures are used for evaluating the energy and
momentum components of the Einstein, Landau-Lifshitz and Møller energy-
momentum complexes. Some of our results are presented, and the metric under
consideration describes a magnetic stringy solution. GrTensorII yields the energy
and momentum components which are involved in the expression for energy and
momentum densities. The expressions for energy in the Einstein, Landau-Lifshitz
and Møller prescriptions are plotted on the same graph using Maple 9.5 commands
for 3d graphs. The computer algebra platform Maple 9.5 and GrTensorII package is
used for performing the calculations and graphical representations, and in this way
a powerful connection between physics studies and computational methods is
established. The young master students play an important role in this activity and
their tasks are clearly defined. They are mostly needed to clarify the problems
related to theoretical physics and computational and mathematics demands. Their
contributions are meant to be strictly related to their dissertation papers. They also
have the possibility to enlarge the circle of collaborations abroad and gather a lot of
information in the field and the adjacent topics.
References
2. Einstein, A., Preuss. Akad. Wiss. Berlin 47, 778, 1915; Addendum-ibid. 47, 799, 1915; Trautman,
A. in Gravitation: an Introduction to Current Research, ed. L. Witten, Wiley, New York, p. 169,
1962.
3. Landau, L. D., Lifshitz, E. M. The Classical Theory of Fields, Pergamon Press, p. 280, 1987.
4. Møller, C. Ann. Phys. (NY) 4, 347, 1958.

5. Chang, Chia-Chen, Nester, J. M., Chen, Chiang-Mei, Phys. Rev. Lett. 83, 1897, 1999.
6. Aguirregabiria, J. M., Chamorro, A., Virbhadra, K. S., Gen. Rel. Grav. 28, 1393, 1996; Vagenas,
E. C., Mod. Phys. Lett. A21, 1947, 2006; Sharif M., Fatima, Tasnim, Astrophys. Space Sci. 302,
217, 2006; Gad, Ragab M., Gen. Rel. Grav. 38, 417, 2006; Grammenos, Th., Mod. Phys. Lett.
A20, 1741, 2005; Radinschi, I., Mod. Phys. Lett. A15, Nos. 11&12, 803, 2000; Radinschi I.,
Grammenos, Th., Int. J. Mod. Phys. A21, 2853, 2006; Aydogdu, O., Salti, M., Astrophys. Space
Sci. 302, 61, 2006.
7. Virbhadra, K. S., Phys. Rev. D60, 104041, 1999.
8. Radinschi, I,. Rom. J. Phys. Vol.50, Number 1-2, 57, 2005.
9. Radinschi I., Ciobanu B., International Conference Fifty Years of Romanian Astrophysics,
Bucharest, September 27-30, AIP, 2006
10. Radinschi, I., Yang, I-Ching, On the Energy of Stringy Black Holes, New Developments in String
Theory Research, ed. Susan A. Grece, Nova Science Publishers, Inc New York, U.S.A., p. 17,
2006.
11. Chang, Chia-Chen, Nester, J. M., Class. Quant. Grav. 16, 1279, 1999.
12. Chang, Chia-Chen, Nester, J. M., Chen, Grav. Cosmol. 6, 257, 2000; Szabados, L. B., Living.
Rev. Relativity 7, 4, 2004.
13. So, Lau Loi, Nester, James M., Chen, Hsin, gr-qc/0605150, to appear in Proceedings of the 7th
International Conference on Gravitation and Astrophysics; Deser, S., Franklin, J.S., Seminara, D.,
Class. Quant. Grav. 16, 2815, 1999.
14. Nester, James M., Class. Quant. Grav. 21, S261, 2004; Chen, Chiang-Mei, Nester, J. M., Tung,
Roh-Suan, Phys. Rev. D72, 104020, 2005.
The state of stresses in RC beams with installation holes located

in compressed zone
Przemysław Siwiec1, Sebastian Toś1
1
Institute of Building Engineering, Wroclaw University of Technology
Summary
This paper presents a question of influence of installation holes located in the

compressed zone of the reinforced concrete beam on functioning. In order to
research the actual functioning of the element a numerical model based on Finite
Element Method has been created.
Two independent analyses were performed in which linear and nonlinear
constitutive relations for concrete and for steel have been used. The results
acquired have been compared and discussed.
KEYWORDS: RC beam, nonlinear FEM analysis, nonlinear material, numerical
model, numerical analysis.
1. INTRODUCTION
Designing process of reinforced concrete structures (RCS) is based on current state

of knowledge. Codes of Practice (CP) in many countries are based on this
knowledge. It can be considered that if an element is designed according to
requirements of the CP and it fulfills requirements of border states (in both phases:
service and mount) as well as construction requirements, there is certain probability
that it will be safe and will properly carry adequate loads.
The problem occurs when there is need to specify the state of safety of reinforced
concrete element not designed in accordance with requirements of the CP and
especially when it doesn’t fulfill construction requirements. Then it is necessary to
perform a complete analysis of stresses and strains according to rules of mechanics.
And here a problem arises: according to rules of a linear analysis it is not possible
to specify correctly the state of stresses in reinforced concrete elements in the II
(2nd) phase. So there is a need to make the analysis based on physical and
geometrical nonlinearities in reinforced concrete.
2. NONLINEAR MODEL
Fundamental physical stress-strains relations based on assumption of linear elastic

relation come from the Hook’s law. The structure deformations are not
proportional to the loads and this irregularity in material of a structure. This kind of
material is called plastic.
From the point of view of theory of elasticity, material turns to undetermined state
after exceeding limit value of stresses (criterion of plasticity) defined by
constitutive relationships between stresses and strains. There are many criteria of
plasticity in complex state of stresses [1,2,3]. The most popular are: Trasca and
Von Mises (M-H-V) yield criterion for describing materials such as steel and
aluminum; and Coulomb and Druckner-Prager yield criterion for describing such
materials as concrete, rocks, soils and sands. In literature, for describing plasticity
in reinforced concrete the most popular are Von Mises criterion (M–H–V) for steel
and Druckner-Prager criterion for concrete under compression. The most
divergences among researchers of the problem have been caused by concrete under
tension. The solutions used [4.5.6.7] are mainly connected with adopted calculation
techniques. In this paper the Finite Element Method as a basic analytical tool has
been used. Within a framework of this method the smeared crack approach for
modeling concrete under tension has been used.
Basic assumption of theory of plasticity in FEM formulation is decomposition of
strains to elastic and plastic parts:
ε = εe +ε p (1)
Such a model is called elasto-plastic. The most important consequence of adopting
this material model relation between total stresses for certain time t and total strains
for time t, is additional parameter – a function of stresses and strain history. This
function is taken into account by internal parameter of state , governed by a
specific evolution law. Properties of elasto-plastic material are determined by
assumptions:
• the elastic relation between total stresses and strains is given by:
σ = Dε e (2)
• the yield condition determines the state of stresses when the plastic flow is
initiated. This condition is determined by vector function of stresses and by
the parameter of state M, according to relationship:
f (σ ,κ ) = 0 (3)
• the flow rule specifies the plastic strain rate vector as a function of the state
of stresses in relationship:
222 F. Lastname, S. Author, T. Author
n ∂g j
ε& p = ∑ λ& j (4)
j =1 ∂σ
for n functions of plastic potential gj which can also be considered as a function of
the stress vector and state parameter gj(U,M).
The plastic multiplier λ& j is determined from Kuhn-Tucker conditions:
 f ≤0
 &
 λj ≥ 0 (5)
λ& f = 0
 j
• the hardening hypothesis describes evolution of the internal state parameter M.
Generally, changes of M parameter are given by a function of the stresses
vector and the plastic strain rate vector, i.e.:
κ& = h(σ , ε& p ) (6)
Fig. 1 – The stress-strain relation in crack.
The smeared crack approach puts forward two concepts: smeared cracks with fixed
direction which is determined when the crack is initiated (constant direction of the
crack during the whole iteration) and the rotated cracks (the crack direction
changes during iteration according to redistribution of main stresses). Transient
concept is multi-directional crack model which makes an assumption that a new
crack arises when angle between direction of vector of actual main stresses and
direction normal to the old crack exceeds certain value but the old one is still
active.
The fundamental feature of smeared crack approach is taken from decomposition
of the total strain into elastic strain and a crack strain as:
ε = ε e + ε cr (7)
During increasing of loading, the new cracks simultaneously occur and everyone of
them makes new local co-ordinate system n-t. For the crack strain for crack i which
is given by:
cr
[
ei = ε cr nn ,i , γ cr nt ,i ] T
(8)
total strain vector in crack is given by:
[ cr
e cr = e1 , e2 ,K, en
cr cr
] T
(9)
Basic cracking parameters of the smeared crack concepts are: tensile strength of
concrete which determines when crack appears, fracture energy which determine
how much of energy is dissipated when a unit of crack appears (the area under UG
curve in crack in fig. 1, shear retention relation in crack, the tension softening
relations as a relation between crack stress and crack strain (shape of descending
part of relations in fig. 1).
3. DESCRIPTION OF PROBLEM.
During the analysis of large precast concrete structure [8] it has been noticed that
almost every beam has from 4 to 8 holes located in compressed zone. Through
these holes go installation pipes.
Fig. 2 – Shown Precast concrete beams with installations holes.
Detailed analysis of the project has shown that no analysis of influence of these
holes on load-carrying capacity and operating conditions had ever been carried out.
The analysis has been made for 11,54 m long precast concrete beam (0,40x1,025).
It had 8 holes (see fig. 2) 3,60 m from the end of the beam 4 installation holes
(15cm in diameter), have been made. They have been placed 20cm from the upper
edge (i.e. the compressed zone).
Fig. 3 – Draft of the beam. Installation holes displayed.
Fig. 4 – Scheme of nodes arrangement around the hole.
4. DESCRIPTION OF NUMERICAL MODEL.
The Finite Element Model has been designed with 4-node plain stress elements
(fig. 4 and 5). Reinforcement has been modeled as embedded bars in structural
elements – reinforcement strains have been computed on the basis of the
displacement field of the mother element. This approach guarantees full connection
of finite element and reinforcement (without any bond-slip). In numerical model
both linear and nonlinear constitutive principles for concrete and steel have been
used. The computer model has been performed only for a support zone 4 m long. It
has been assumed that material was homogeneous – cross-sections only along
neutral axis have been examined.
5. CALCULATIONS AND RESULTS.

For numerical calculation the actual load from the design project (216,78kN/m) has
been taken into account. This load yields the bending moment at 3568,89kN/m.
In the figures 6 and 7 the linear static analysis results have been presented.
Fig. 5 – Scheme of finite elements arrangement.
Fig. 6 – Beam deformations draft – linear static analysis.
It can be noticed that elastic relations in concrete don’t allow to analyze properly
the stress distribution in concrete. According to this analysis the tensile stresses
reach 11,71 MPa, which is not true. Therefore the correctness of stress distribution
and its values around the hole could be questioned.
Fig. 7 – Stress distribution around the hole – linear static analysis.
In order to eliminate this problem the nonlinear analysis has been performed.
Fig. 8 – Element deformation for Lf= 2,496 (damage of the beam) – nonlinear analysis.
The numerical analysis has been carried out with use of Newton-Rawston iterative
algorithm for solving a nonlinear system of equations. For iteration process the arc-
length spherical method for prediction of load increment and energy norm criterion
for convergence criterion of equilibrium iteration have been used. During
numerical analysis load has been progressively increased in order to fulfill
equation:
Lf ⋅K × v = F (10)
Where:
- Lf – scalar nonlinear load coefficient,
- K – global stiffness matrix,
- v – general displacements vector,
- F – general loads vector.
Response of the structure has been presented as a relation between selected degree
of freedom in function of nonlinear load coefficient (Lf) which allows to determine
critical values of loads at which structure fails (see fig. 10).
Structure damage mechanism can be observed through the research (albo
investigation) of model deformation during the nonlinear analysis (fig. 8). The
breaking of the beam close to the hole can be noticed very clearly. This has been
caused by (the lack of) elastic recovery of other parts of researched element (no
flexion).
Crack development has presented the state of stresses in the structure. In fig.9 areas
of active cracks have been displayed.
An important fact is appearing of skew cracks (in the support zone) for coefficient
of deformation Lf equal to 1,0 which has been observed in real structure as well.
During damage phase almost the whole tensile zone of the beam has been cracked.
Fig. 9 – Development of cracking zone for different values of nonlinear loads coefficient
Lf.
Fig. 10 – Relation between displacement of node 1 and nonlinear load coefficient Lf.
Through nonlinear analysis and stress distribution it is possible to determine crack

development and neutral axis position of cross-section of RC beam (fig. 12). As
contrasted with the nonlinear analysis, the linear analysis of RC structure has
proved almost totally useless (fig. 11).
Fig. 11 – Stress distribution around the hole received from nonlinear analysis.
Fig. 12 – Stress distribution from linear and nonlinear analysis for various loads
coefficients. Left – linear analysis, Right – nonlinear analysis.
6. CONCLUSIONS
In result of the analysis of the computer model an image of stresses around the hole
has been received (fig. 7 and fig.11, 12). It can be seen very clearly there occurred
a disorder in stress distribution around the hole. It is significant that stress
increment at the edge of the hole reaches up to 70% of material strength
(compression strength of concrete). However this disorder decays and about 10 cm
from the hole edge stresses reach regular values (similar to those of elements
without holes).
The analysis of a nonlinear loads coefficient has shown that the beam fails at large
values of this coefficient (about 2,5 times of service load) which leaves large safety
margin for the structure. Therefore it can be stated that the influence of the hole is
not significant for decreasing of the beam safety.
Additionally the nonlinear computer model has been proved highly reliable for the
analyses of cracked RC elements.
References
1 Blaauwendraad J.: Numerical Methods in Structural Mechanics – Delft University of Techniology –
1993
2 De Borst R.: Computational Methods in Non-linear Solid Mechanics – Delft University of
Technology, seria B20, Delft – 1993
3 Feenstra P. H., De Borst R., Rots J. G.: The application of various crack models to F.E. analysis of
reinforced concrete panels - Nonlinear Engineering Computation, ed. N. Bićanić, D.R.J. Owen,
1991
4 Grootenboer J. J., Leijten S. F. C. H., Blaauwenraad J.: Concrete mechanics, Part C, Numerical
models for reinforced concrete structures in plain stress – Heron vol. 26, Delft – 1981
5 Rots J. G., Blaauwenraad J.: Crack model for concrete: Discrete or smeard? Fixed, multi-linear or
rotatin? – Heron 34, Delft – 1989
6 Rots J. G., Nauta P., Kusters G. M. A., Blaauwenraad J.: Smeared crack approach and fracture
localization in concrete – Heron, vol. 30, 1985
7 Siwiec P., Strouven P., Moczko A.: Study of steel bar corrosion by experimental and analytical
simulation – DIANA Computational Mechanics ’94, ed. Kusters M. A., Hendriks A. N., Kluwer
Academic Publishers, Dordrecht – 1994
8 Szechiński M, Kamiński Kamiński.: Research on reinforced Concrete structure after the damage,
Reports of Institute of Buiding Ingeneering of the Wrocław University of Technology, Wrocław.
Poland. 2001.
Computer Graphical Communication an Efficient Tool for

Understanding the Design of Wooden Trusses
Andrei Slonovschi1, Ion Antonescu1 and Liviu Prună1
1
Department of Descriptive Geometry and Technical Drawing,
Technical University “Gh. Asachi” Iasi, 6600, Romania
Summary
In design or workmanship the civil engineers (designers, chiefs of building yards,

etc.) often encounter moments when they need to create or read technical drawing
plans.
The skill of plan's reading is firstly due to the capacity of engineer to create mental
images. This ability differs from one individual to another allowing it's exploiting
and training in a manner that allow him to communicate with the other engineers,
making him fully understood, by learning and training the concepts of Descriptive
Geometry, Technical Drawing and Computer Graphics.
The routine of mentally converting space representations in plane drawings and
reverse, by using an engineering common language or, in other words, to
communicate through graphical approach engineering data, is achieved in Civil
Engineering Faculty in the firs year of study.
Practical experience in teaching disciplines regarding graphical communication
shows, as expected, that exclusive using of didactic charts using bi-dimensional
representations isn’t sometimes enough for students to understand the three-
dimensional shape of complex construction elements.
Taking account of these factors three-dimensional didactic models for drawings
with a high level of understanding difficulty were proposed and developed.
KEYWORDS: Computer Graphical Communication, Descriptive Geometry,

Technical Drawing, Computer Graphics, 3D Model, AutoCAD, Triple
Orthographic Projection, Axonometric Isometric Projection.
232 A. Slonovschi, I. Antonescu, L. Prună
1. INTRODUCTION
In design and execution activity, a fundamental role in creation and understanding

of technical drawing blueprints has the ability of person to create mental images,
without which it be practically impossible to read and relate the views of a drawing
for understanding the shape of existing or virtual parts composing the future
buildings.
The bulk of first year students from technical faculties, hence the ones from Civil
Engineering Faculties, don't have a trained ability for creating mental images, in
other words, to see with their mind's eyes the three-dimensional shape of a part
staring from it's 2D graphical representations and haven't any knowledge about
existing conventions or standards referring to graphical communication.
This skill together with data related to graphical communication language is
acquired in the firs year of study at Descriptive Geometry and Technical Drawing
disciplines.
Introduction of computer graphics and CAD programs (e.g. AutoCAD) facilitate
the acquiring of this skill by their capacity to create virtual images suggesting the
3D shape of bodies.
Let see further how computer graphics together with descriptive geometry and
technical drawing help in developing of future engineers' 3D space vision.
2. THE NECESSITY OF USING COMPUTER GRAPHICS IN

TRAINING OF ABILITY TO CREATE MENTAL VIEWS
At Technical Drawing discipline in workshop lessons the students have to create

five blueprints referring to:
- Geometrical constructions;
- Wooden structures (wooden roof truss);
- Reinforced concrete steel structures (concrete steel beam);
- Metal structures (metal truss);
- Dwelling buildings representation (plan of current floor and façade).
These aim equally the training of capability to see with their mind's eyes and the
gradual spreading of data referring to Technical Drawing, namely to the
engineering graphical communication norms.
Because some students have a reduced capacity of creating mental images, and
others with such an ability more developed don't have trained it at all, the blueprint
which rises frequently understanding problems is the wooden truss referring to a
portion of a roof, it's length being of 4 meters and it's span of 12 meters (fig. 1).
Figure 1. The Wooden Roof Truss
As noticed from figure 1 on the blueprint are drawn the front view, the plan (view
from above), the profile (section view from the left) and the isometric
representation of a ridge beam assembly knot detail of a wooden roof truss.
In the classical presentation variant of the theme referring to the wooden structures
(with the help of presentation plans) the composing parts of the truss (rafters,
purlins, ridge beam, props wall plates, tongs, struts, bases, etc.) are presented to the
students together with the way of vertical loads transmission (rafters → purlins →
props → bases→ ceiling → etc.).
From practical experience it is noticed that the students have major ambiguities in
deciphering the 3D shape of the truss (the way of positioning the props, rafters,
tongs, struts, bases, etc.). So, they partially understand the representing way of the
front view but they do not understand the depicting of view from above, left view
and of the ridge beam assembly knot detail. These facts demonstrate that there are
major problems on the road of materializing the mental image of the body
orthographical represented on the plan.
Starting from the above mentioned problems we have proposed that, by using
computer graphics, to create a three-dimensional solid model of the wooden truss
which to significant contributes to the smoothing of the above mentioned road.
This model was developed with the support of AutoCAD program.
3. REALISING OF THE THREE-DIMENSIONAL MODEL OF

THE WOODEN TRUSS
In the frame of the lesson dedicated teaching the "wooden truss" theme, parallel
with the classical exposure, with the help of the plans containing it's orthographic
plane representation, the presentation of the 3D model is made. This 3D model was
realised at the 1:1 scale according to the dimensions from the 2D representation.
For a relatively rapid learning of the four representations from the didactic plan
(front view, plan view, left view and isometric projection of the ridge beam
assembly knot) the screen of AutoCAD program was split into four view-ports thus
(fig. 2):
- front view in the upper left corner;
- view from above in the lower left corner;
- view from the left in the upper right corner;
- 3D model in the lower right corner.
If working is done in any view-port, the changes made are automatically updated in
the others three view-ports.
The vision of 3D model take place on stages the model being structured using the
AutoCAD layers technique.
Front Left side

view view
View
from
3D Model
above
Figure 2. Projection disposal on view-ports
In the first step, taking in account the walls’ thickness and the distances between
them, the slices from the walls supporting the cappings and the beams of the
ceiling are depicted (fig. 3).
The Brick walls
Figure 3. The Brick walls
As the upper and lower level dimensions of all walls are equal the picture
appearing in the “left view” view-port can create confusion at first sight because a
single wall is seen, the other ones being covered by the visible wall. This confusion
is eliminated by the views appearing in the other three view-ports.
In the second stage the ceiling and its beams and cappings are shown. In the “view
from above” view-port the walls, the beams and the cappings are not seen because
the ceiling is placed above these elements. By the other hand in the “3D model”
view-port some portions of these elements are seen because of the different
location of the observer (fig. 4).
The ceiling
Figure 4. The Ceiling
In the third step the bases are positioned. These have a parallelepiped shape with
32 x 32 x 5 cm sizes and are placed on A, B, D şi E axes at 3 meters distances one
axe from another along the building.
The only view-ports in which all bases are seen are the “view from above” and “3D
Model”. In “left side view” the bases are not seen because they are covered by the
wall plates. In “front view” only the bases from the first row are seen, the others
are overlapping in this view with these (fig. 5).
The bases
Figure 5. Bases position

In the fourth step the props located on A, B, D and E axes are depicted. The props
from C axe are drawn after the drafting of the rafters and the ridge beam.
The props are structural parts of the roof taking the vertical loads from the purlins
and transmitting them to the ceiling’s girders. The adopted dimensions for the
props’ sections are 12 x 12 cm.
The props from A, B, D and E axes are centred positioned on bases (fig. 6). In fig.
7 the detail of a prop leaning on a base is depicted.
The props
Figure 6. Placement of Props from A, B, D and E axes
Prop Concrete steel

ceiling
Base
Figure 7. Detail of a prop leaning on a base

The only view-ports in which all props are seen are the “view from above” and
“3D Model”.
In “left side view” view-port the props from one half of the roof opening are seen,
the others overlapping in view with these. In the “front view” only the props of the
first rank are seen, the others being placed on rear of the ones from the first row
(fig. 6).
In the fifth stage the purlins and the wall plates are positioned parallel with the
building’s edges. The purlins and the wall plates are structural part of the roof that
take the vertical loads from the rafters and transmit them to the props and the
capping of ceiling. The purlins lean on the props. The adopted section for purlins
and wall plates is 12 x 15 cm.
Next the current purlins and the wall plates are represented (fig. 8). The view-ports
in which all elements are seen are the “view from above” and “3D Model”. In “left
side view” view-port only two purlins and one wall plate from a total of four
purlins and two wall plates are visible, the other elements overlapping with the
visible purlins and wall plates.
Wall plate
Purlin
Prop
Base
Brick wall
The ceiling
Figure 8. Current Purlins and the Wall Plates
In the sixth stage the rafters are drawn. The rafters are structural elements of the
roof that transmit the vertical outside loads to the purlins and wall plates and giving
shape to the roof.
As noticed in fig. 1 the roof is coupled thus for its construction on the opening will
be pairs of two rafters positioned along the building. The axial distance between
rafters is 1 meter.
The adopted sizes for rafters’ sections are 75 x 12 cm. As these must be positioned
on purlins on edge (the smaller size of the section) results that in the front view the
section’s dimension is 12 cm and in the left and upper views the dimension is 75
cm (fig. 9).
Figure 9. Knot detail in triple orthographic projection
For representing the rafters, the level assembly dimensions of these and the cutting-
out of the rafters in order to lean them on the purlins and wall plates have to be
taken in account. The cut-out is realised with a maximum deep of one third of the
greater dimension of the section (fig. 10 and 11). Knowing this information and
taking in account the upper and lower level positioning dimensions all the roof’s
rafters are represented.
Figure 10. Detail of cut-out in a rafter

Rafter
Cut-out
zone
Purlin
Figure 11. Exploded detail of purlin – rafter joint
The view-ports in which all rafters are seen are the “view from above” and “3D
Model”. In the “front view” only the front rafters are visible, the others being
positioned behind them, and in the view-port “left side view” only the rafters
positioned on one half of the roof opening are seen, the rafters positioned on the
other half of the opening being behind them (fig. 12).
Rafter
Figure 12. Rafters positioning

In the seventh step the ridge beam is positioned. The same conditions presented in
case of mounting the purlins are taken in account for representing the ridge beam
(fig. 13 ... 15).
Figure 13. Detail of upper rafters joint
Cut-out
zone Rafters
Ridge
beam
Figure 14. Exploded detail of upper knot

Ridge
beam
Figure 15. Positioning of Ridge Beam

In the eighth stage the props from C axe are depicted. The props are assembled
under the ridge beam at a distance between axes of 3 meters in the same transversal
planes as the existing props (fig. 16 and 17).
Prop
Figure 16. Placement of Props from C axe
Ridge
beam
Rafter
Prop
Figure 17. Detail of Ridge Beam mounting
In the ninth step the tongs are represented. The tongs are stiffness elements of the
roof truss. These are positioned in pairs under the purlins, containing between them
the rafters and the props (fig. 18).
Tongs
Figure 18. Placement of Tongs
The view-ports in which all tongs are seen are view from above” and “3D Model”.
In the “front view” only the front tongs are visible, the other tongs overlapping in
this view with them. In the “left side view” the ends of the tongs from one half of
the truss opening are seen.
In the overlapping zone between props and tongs, the props are cut out for tongs
crossing (fig. 19).
Rafter
Purlin
Prop Tongs
Cut-out
zone
Figure 19. Exploded detail of prop’s cut-outs for tongs mounting
In the last step the struts are positioned. The struts are stiffness elements of the
truss also (fig. 20).
Struts
Figure 20. Positioning of Struts
The view-ports in which all struts are seen are view from above” and “3D Model”.
In the other two view-ports only the struts from the first rank are visible.
Using the AutoCAD program specific commands for space vision of the model
(3DORBIT, VPOINT, ZOOM, PAN) some positions of the roof truss’s 3D model
are presented to the students for a better understanding of its structure and shape.
Also with the help of the 3D model the students better understand the way of
structuring and depicting in axonometric isometric projection of the ridge knot (fig.
21 and 22).
Rafter
Tongs
Ridge
beam
Prop
Figure 21. Detail of Ridge Knot
Rafter
Rafter
Ridge
beam
Tongs
Cut-out
zone
Prop
Figure 22. Exploded detail of Ridge Knot
3. CONCLUSIONS
Using of computer graphical communication and of CADD programs (like

AutoCAD) together with classical graphical disciplines (Descriptive Geometry,
Technical Drawing) is essential in forming and developing of spatial vision and the
ability of reading triple orthographic projection drawings, tools useful for future
design and execution engineers.
References
1. Stănilă, A., Hîncu, G., Caiet de desen tehnic, Editura Societăţii Academice “Matei-Teiu Botez”,
Iaşi, 2006. (in Romanian)
2. Gheorghiu, M., Chelcea, M., Indrumator pentru desen tehnic de constructii, Editura Matrix Rom,
Bucureşti, 2001. (in Romanian)
3. Harrington, D., Burchard, B., Pitzer, D., Inside AutoCAD 2002, New Riders Publishing, 2002.
4. Simion, I., AutoCad 2006 pentru ingineri, Editura Teora, Bucureşti, 2006. (in Romanian)
Analysis of dynamic parameters of rail fastening

by Rihaczek transformation
Jaroslav Smutny1, Lubos Pazdera2
1
Department of Railway Constructions and Structures, Faculty of Civil Engineering,
University of Technology Brno, Brno, 602 00, Czech Republic
2
Department of Physics, Faculty of Civil Engineering,
Summary
For evaluation of response signals obtained by rail fastening analysis a new

method using time and frequency related transformations has been developed. In
the paper the laboratory measurements and dynamic parameter analyses of flexible
fastening of Vossloh SKL14 type have been described. The method can also be used
for designing new rail fastening systems and their parts.
KEYWORDS: Rail fastening, dynamic test, time frequency transform
1. INTRODUCTION
The basis for the selection and comparison of new components of rail
superstructure are also theoretical analysis (simulation) and static and dynamic
tests carried out in the laboratory and in the field (directly on the railway) [9]. It is
necessary to mention that theoretical analysis of application of mathematical
simulation is often based on idealized assumptions. Hence, when the real
conditions on the railway or tramway superstructure are encountered, the model
may be inaccurate.
For testing the railway superstructure construction, different methods and different
criterions were applied. Dynamic testing [4] often uses the method of exciting the
structure by mechanical shock. Exciting by shock is useful for the setting up a
given set of frequencies as the shock, according to the theory, stimulates all
frequencies, mainly resonant. Mechanical shock is often stimulated by a special
hammer, which has an incorporated power sensor in radial direction to the railhead.
The response is measured by accelerometer sensors at different points of the rail
structure (rail foot, clip plate, clamp, sleeper etc). This method makes it possible to
record frequency components in the range 1 Hz to 10 kHz. Recorded data are often
recalculated and presented in the form of the frequency transfer function. This
shows important frequency components (mainly resonant) that include information

about dynamic properties of particular parts of tested structure.
Often in the experimental investigation of dynamic properties [5] of rail fastening,
the measurement and calculation of the transfer function is called accelerance
(inverse function of dynamic weight). It is for that reason that the acceleration
transducer is the most suitable electromechanical measurement device for the
measuring of oscillation. Accelerance is defined by the relation
S aF ( f )
H aF ( f ) = (1)
S FF ( f )
where SaF(f) is cross spectrum of response and entry signal, SFF is auto-spectrum of
entry signal. From the relation (1) it can be seen that measured acceleration is
standardised for power measured during the shock.
In spite of the advantages, it is not possible to localise the time behaviour of
frequency components included in the signal. Therefore for the evaluation of
response signals when analysing the rails fastenings, the authors supplemented the
methods of the measurements by utilising progressive processes of signal analysis,
i.e. by utilising time frequency transformations.
One possible procedure to analyse time occurrence of frequency components of
transfer and non-stationary signals, is the use of the so-called time frequency
proceedings (transformations). These can be distributed according to two basic
groups [3]:
▪ linear (including mainly Short Time Fourier Transformation, Wavelet
Transformation etc.)
▪ non-linear (including mainly quadratic Cohen Transformations, Affine and
Hyperbolic Transformation, eventually further special proceedings)
2. THEORY OF TIME FREQUENCY ANALYSIS
Given a time series, x(t), it can readily be seen how the “energy” of the signal is
distributed in time. By performing a Fourier transform to obtain the spectrum, X
(ω), it can also be seen how the “energy” of the signal is distributed in frequency.
For a stationary signal, there is usually no need to go beyond the time or frequency
domains. However, most real world signals have characteristics that change over
time, and the individual domains of time and frequency do not provide a means for
extracting this information. The general goal of this contribution is to demonstrate
some lesser-known methods for creating functions that represent the energy of the
signal simultaneously in time and frequency.
248 J. Smutny, L. Pazdera
Example of linear time-frequency distributions is the Short Time Fourier

Transformation. The main idea of the Short Time Fourier Transform (STFT) is to
split a non-stationary signal into fractions within which stationary assumptions
apply and to carry out a Fourier transform on each of these fractions. The STFT is
defined by equation [3, 4]
∞
STFT (τ , f ) = ∫ [ x(t ) ⋅ g ∗ (t − τ )] ⋅ e − j⋅2⋅π ⋅ f ⋅t ⋅ dt (2)
−∞
where ‘*’ denotes the complex conjugate, g is the short time window, x(t) is the
signal, τ is the time location parameter, f is frequency and t is time. In the two
dimensional time-frequency joint representation, the vertical stripes of the complex
valued STFT coefficients STFT(τ, f) correspond to the Fourier spectra of the
windowed signal with the window shifted to given timesτ. The main disadvantage
of linear time-frequency transform is that the time frequency resolution is limited
to the Heisenberg bound. This is due to the imposition of local time window g(t). If
this window is more resolved in time, the frequency resolution suffers because the
effective width of its Fourier transform G(f) increases, and vice-versa.
Quadratic (non-linear) methods present the second fundamental class of time
frequency distributions. Quadratic methods are based upon estimating an
instantaneous power (or energy) spectrum using a bilinear operation on the signal
x(t) itself. The class of all quadratic time-frequency distributions to time shifts and
frequency-shift is called Cohen’s class. Similarly, the class of all quadratic time-
frequency distributions covariant to time-shift and scales is called the Affine class.
The intersection of these two classes contains time-frequency distributions, like the
Wigner-Ville distribution, that are covariant to all operators.
Cohen [1] generalised the definition of the time frequency distributions in such a
way as to include a wide variety of different distributions. These different
distributions can be represented in several ways. Cohen’s class definition like the
Fourier Transformation, with respect toτ, of the generalised local correlation
function is most common. With a two-dimensional kernel, the bilinear time
frequency distribution of the Cohen´s class is defined according to equation [2]:
 τ  τ
C x (t , f ) = ∫∫∫ e − j⋅2π ⋅θ ⋅t ′− j⋅2π ⋅ f ⋅τ ⋅+ j⋅2π ⋅θ ⋅t ⋅ψ (θ ,τ ) ⋅ x t +  ⋅ x *  t −  ⋅ dθ ⋅ dt ⋅ dτ (3)
 2  2
where x is the signal, t (t’) is the time, τ is the time location parameter, ω is
angular frequency, θ is shift frequency parameter, ψ(θ, τ) is called the kernel of the
time frequency distribution. A distribution Cx (t, f) from Cohen’s class can be
interpreted as the two-dimensional Fourier Transformation of a weighted version of
the ambiguity function of the signal
C x (t , f ) = ∫∫ Ax (θ ,τ ) ⋅ψ (θ ,τ ) ⋅ e − j⋅2π ⋅ f ⋅τ ⋅ e − j⋅2π ⋅θ ⋅t ⋅ dτ ⋅ dθ (4)
where Ax(θ, τ) is the ambiguity function of the signal x(t), given by equation:
 τ  τ
A(θ ,τ ) = ∫ x t +  ⋅ x *  t −  ⋅ e − j⋅θ ⋅t ⋅ dt (5)
 2  2
We note that all integrals run from -∞ to ∞. The weighted function ψ(θ, τ) is called
the kernel. It determines the specific properties of the distribution. The product
Ax(θ, τ)⋅ψ(θ, τ) is known as the characteristic function. Since the ambiguity
function is a bilinear function of the signal, it exhibits cross components, which, if
allowed to pass into time frequency distribution, can reduce auto-component
resolution, obscure the true signal feature, and make interpretation of the
distribution difficult. Therefore, the kernel is often selected to weight the ambiguity
function such that the auto-components, which are centred at the origin of the (θ, τ)
ambiguity plane, are passed, while the cross-components, which are located away
from origin, are suppressed. This means that the suppression of cross-components
might be understood as the frequency response of a two-dimensional low-pass
filter.
When a low pass kernel is employed, there is a trade-off between cross-
components suppression and auto-component concentration. Generally, as the
band-pass region of the kernel is made smaller, the amount of cross-component
suppression increases, but at the expense of auto-component concentration.
There is definition of the kernel for Rihaczek Transformation in equation 6
j ⋅θ ⋅τ
ψ (θ ,τ ) = e
2
. (6)
Equation 4 can also be rewritten into the following form [5]
∞ ∞
C x (t , f ) = ∫ ∫ Π(τ − t ,θ − f ) ⋅ WVT (τ ,θ ) ⋅ dτ ⋅ dθ (7)
−∞ −∞
where
∞ ∞
Π (t , f ) = ∫ ∫ψ (θ ,τ ) ⋅e
− j ⋅ 2 ⋅π ⋅( f ⋅τ −θ ⋅t )
⋅ dt ⋅ dω (8)
−∞ −∞
is the two-dimensional Fourier transform of the kernel ψ and WVT presents

Wigner-Ville transform. Cohen’s class has a simple interpretation as a smoothed
Wigner-Ville distribution [5].
3. ANALYSIS OF DYNAMIC PARAMETERS
The model used for laboratory measurements and analysis of dynamic parameters
of a sample of rail fastening is presented below. The rail grid model was
constructed of concrete sleepers B 91, on which there were fastened rails of
construction shape UIC 60 by flexible fastening Vossloh SKL14.
For the testing of the dynamic properties of the sample, the method of measuring
the response to mechanical shock was used. Mechanical shock was stimulated by a
special hammer in the radial direction on the railhead. A part of this hammer is a
force detector.
The response was measured by accelerometers at different points of the rail
structure, on the rail foot and sleepers (10 cm from fastening). Figure 1 show the
location of detectors. From the response time signals frequency transfer functions
(accelerance) were calculated in order to obtain standardised responses [5].
Figure 1 General view of the working place
Signals from measurements on the rail and sleepers were used for the presentation
of particular analyses in this contribution. The measuring system consisted of a
Brüel and Kjaer PULSE modular analyses for recording the vibration parameters
together with B&k cubic acceleration detector and a B&k shock stimulation
hammer (Figure 1).
The accelerometers were fastened to the measured construction by means of bee

wax. The results were recorded digitally.
The analysis of the response to mechanical shock was evaluated by means of the
following methods and parameters [5]:
• Time records of the duration of impulse response function (in principle
standardized acceleration value)
• Frequency analysis with the use of frequency response function (according to
equation 1)
• Time-frequency method of spectral analysis (for the transfer from time to
time- frequency domain, the algorithm of Short Time Fourier Transformation
and Rihaczek Transformation was used)
Time histories of the impulse response function, recorded by accelerometers,

located on the rail foot, are depicted on the upper graph of Figure 2. The maximum
positive value of acceleration of 300 m⋅s-2 is reached 1 ms from the observed
beginning. The maximum negative value of acceleration of -300 m⋅s-2 is reached 2
ms from the beginning. Damping of the signal from the acceleration 300 m⋅s-2 to
the acceleration lower than 30 m⋅s-2 took 15 ms.
In the left graph of Figure 2 is depicted the amplitude spectrum of this frequency
response function calculated according to equation 1. In the graph, six important
frequencies (0.2 kHz, 0.7 kHz, 1.9 kHz, 2.4 kHz, 3.3 kHz and 3.7 kHz), are visible.
The important values are taken as those which have the damping up to 20 dB from
the maximum value of amplitude spectrum.
Time frequency amplitude spectrum estimated by application of Short Time
Fourier Transformation to the impulse response function is depicted in the middle
graph in Figure 2. As shown on this graph, the time history of important frequency
components essentially differ.
Frequency component 1.9 kHz reaches the highest values for a relatively long time
(compared to other frequency components). It appears in the signal nearly in its full
history, i.e. approximately 40 ms by damping up to 40 dB. The second most
important component is the frequency 3.3 kHz. This appears in signal up to the
time of 20 ms from the above. Other notable frequencies 2.4 kHz and 3.7 kHz are
in the signal for the time of 5 ms up to 15 ms.
Similar conclusions are visible from the middle graph of Figure 3, which present
the analysis of impulse response function on the rail foot by the use of Rihaczek
Transformation. This transform belong to the category of non-linear time
frequency proceedings from the Cohen class.
Figure 2 Accelerometric detector located on the rail foot, time frequency analysis by
Short Time Fourier Transformation
Figure 3 Accelerometric detector located on the rail foot, time frequency analysis by
Rihaczek Transformation
Signals (impulse response function) taken by a second transducer, located on the

concrete sleeper, have different character. From the time record (see upper graph of
Figure 4) it is apparent that the maximum impulse response function amplitude
acquires lower frequency values as a result of the influence of the transformation of
the signal through the fastening of rail, clip plate, sleeper to the accelerometer and
reaches values of 50 m⋅s-2. These values were reached 2 ms from the first rise time
from “amplitude pack". Values of acceleration are considerably lower than those
by the transducer located on the rail foot which was located nearer to the source of
mechanical impulse.
In the left graph of Figure 4 is depicted the amplitude spectrum of frequency
response function. The form of spectrum considerably differs from the
characteristics measured by the first transducer located on the rail foot. The most
important components appear in the lower frequencies from the transducer located
on the rail foot: in the interval of 0.2 kHz up to 2 kHz, there are also more in
number.
Similar conclusions are given by the middle graph of Figure 4 which presents the
time frequency amplitude spectrum estimated by the application of the Short Time
Fourier Transformation. From this graph it can be seen that time occurrence of
significant components included in signal is considerably shorter (the longest is
approx. 20 ms from the imaginary beginning) than it is from the signal from
transducer located on the rail foot.
Figure 4 Accelerometric detector located on concrete sleeper, time frequency analysis by

Short Time Fourier Transformation
Figure 5 Accelerometric detector located on concrete sleeper, time frequency analysis by

Rihaczek Transformation
Similar conclusions apply to the middle graphs of Figure 5 which present the
analysis of signals from transducers located on the sleeper by the use of Rihaczek
Transformation. The significant frequency components which are calculated by the
Rihaczek Transformation (Figure 6) are frequencies of 0.2 kHz, 0.7 kHz, 1.9 kHz,
2.7 kHz, 3.2 kHz, 3.4 kHz and 3.7 kHz. The most significant spectrum component
is the frequency component 0.2 kHz which appears within this spectrum for a
relatively long time in relation to the activity of other components.
On the whole, it is possible to state from the middle graphs in Figure 2 to Figure 5
that in contrast to linear methods whose ability to resolve the frequency elements in
the time region is limited by certain window functions, quadratic methods can
achieve this objective. Higher distinguishing makes more favourable localization of
significant frequency components in time possible. The quality of time and
frequency achieved by measuring the signal response to mechanical shock and
applying by these transformations is a good choice.
4. CONCLUSIONS
Based on measurements and analyses, it is possible to state that the methods

presented above are very good for the measurement of dynamic parameters of rail
fastenings. The use of these methods enables the testing of new types of rail
fastenings and different types of rail washers under rails and the opportunity to
optimise the geometric location of damping elements on rail etc. From the
mathematical means of signal analysis it is possible to utilise both Short Time
Fourier Transformation and Rihaczek Transformation for time-localisation of the
occurrence of frequency elements of stationary and non-stationary signals.
Based on the experience acquired, it is of great advantage for the analysis of real
signals to utilise the properly selected time and frequency sections. This procedure
seems to be more suitable than the spatial arrangement. It is possible for more
precise localisation of time records to separate significant frequency components or
to depict all important frequencies. Analysis of signals, acquired by measurement
and analysis of response to mechanical shock gives new, more detailed insights to
transition characteristics of railway and tramway structures. Hence, it grants
valuable knowledge for a thorough analysis of these constructions, which can be
important for consequent optimisation of construction and operational conditions.
Also the fact that by time frequency proceedings analysis of dynamic load of
railway and tramway constructions provides real data for consequent formulation
of mathematical models. From this point of view, both linear and non-linear time
frequency transformations are applicable. These methods give a fast and accurate
localisation of frequency components included in the measured signal. It is possible
to apply the described method successfully not only on samples of several
constructions of railway and tramway superstructure but also directly in the field
on real tracks.
Acknowledgements
This research has been supported by the research project 103/07/0183

("The investigation of dynamic effects due to the rail transport by the method of
quadratic time and frequency invariant transformations ") and by research project
MSM 0021630519 (“Progressive reliable and durable load-bearing structural
constructions”)
References
1. Cohen L., Time-frequency distributions, Proc. IEEE, 1989, vol. 77 no. 7, pp. 941-981
2. O'Neill J.C., Quartic Functions for Time-Frequency Analysis with Applications to Signal
Adaptive Kernel Design, SPIE - Advanced Signal Processing Algorithms, 1997
3. Poularikas A. D.: The Transform and Applications Handbook, IEEE Press, 1996
4. Melcer, J., Kucharova, D. : Mechanical properties of rubber pads under static and dynamic load,
proceedings of International Conference on Materials Science and Engineering,
BRAMAT 2003, Romania, Brasov, 3/2003, University of Brasov, 2003, pp. 18-314.
5. Smutný J., Pazdera L.: New techniques in analysis of dynamic parameters rail fastening, InSight,
The Journal of The British Institute of Non-Destructive Testing, Vol 46, No 10, October, 2004,
pp. 612-615, ISSN 13542575
The use of time frequency transformations

in testing structural elements
1
2
Summary
The paper deals with the non-destructive testing of structural elements by means of
the acoustic response using the jointed time and frequency transforms. These
methods make it possible to localize the beginning and the end of frequency
components contained in the measured signal and in this way, they enable us to
analyze perfectly the spectrum of the non-stationary noise. In this way, this
mathematical procedure enables us to distinguish a good specimen from a
defective one.
KEYWORDS: Time Frequency Transform, Wavelet Transform, Acoustic

Response Analysis, Ambiguity Function
1. INTRODUCTION
Many times we have already been able to convince ourselves that utilising the
experience and knowledge having their roots even in the far past have brought
surprisingly good results. One of such experience is the knowledge that noise
resulting from the shock applied to a structure with cracks (disturbances) is
significantly different from the same subject without cracks. This phenomenon has
been known for a long time. As early as in the Middle Ages this phenomenon was
used to detect cracks in ceramic pots after their firing. The phenomenon mentioned
makes it possible to detect cracks in metallic materials as well. Its very old
application in the railway transport is generally known. However, in the
development and application of the methods used to detect defects in structural
elements and materials, this phenomenon has often been neglected. The absence of
the advanced measuring techniques and appropriate mathematical instruments
necessary for the evaluation of measured signals were the main reasons.
There are only the methods of the time-frequency analysis in connection with the
classic spectral analysis that make a thorough analysis of the signals measured with
good possibilities of classifying and identifying possible defects. The noise excited
by force impulse to the tested construction or element is a very interesting

phenomenon since it contains a number of mutually independent and well
recognisable symptoms according to which it is possible to differentiate materials
with cracks from those without cracks [8, 9].
These symptoms are particularly included multi-spectra of the measured noise
signals. The composition of the spectrum of noise of each material is given by
several characteristics with their own side elements with respect to time and
frequency. The spectrum of the measured noise is changed for the duration of the
process. The symptoms of cracks are as follows:
• Changes of the amplitudes of particular characteristic frequency components
• Frequency shifts of characteristic frequency components
• Appearance of new frequency components
• Presence of modified spectrum in comparison with a good product
2. THEORY OF TIME FREQUENCY ANALYSIS
Information relating to any technical or physical occurrence is represented in the

signal by time changes of immediate values or physical phenomena described. The
direct evaluation of the time-amplitude representation often appears not to be easy.
That is why there is a practice of the signal transform from the time domain to
some different ones. In some cases, some important pieces of information from the
frequency area may be obtained. There are a lot of various transforms applied for
transition between time, frequency and jointed time frequency domain. The best
known method is the Fourier transform and some of its modifications.
The Fourier transform and some of its modifications are the techniques which are
especially suitable for processing stationary signals. These methods can analyse
transient and non-stationary signals as well as in the cases when we are interested
only in the frequency components contained in the whole time behaviour of signal.
However, these do not provide us with information on the occurrence of important
frequency components in the time flow.
One of the possible way how to analyse the time occurrence relating to the
frequency components of transient and non-stationary signals is the use of the so-
called time-frequency analysis method.
The goal of this paper is to demonstrate some less-known methods for creating
functions representing the energy of the signal simultaneously in time and
frequency domain.
The Short Time Fourier and Wavelet Transforms represent the examples of the
linear time-frequency distributions. The main idea of the Short Time Fourier
Transform (STFT) is to split a non-stationary signal into segments where the signal
is considered to be stationary. The Fourier transform on each of these segments is

computed. The STFT is defined by equation [3]
∞
STFT (τ , f ) = ∫ [ x(t ) ⋅ g ∗ (t − τ )] ⋅ e − j⋅2⋅π ⋅ f ⋅t ⋅ dt (1)
−∞
where ‘*’ denotes the complex conjugate, g(t) is a short time window, x(t) is
a signal, τ is a time location parameter, f is a frequency and t is a time. In the
two dimensional time-frequency joint representation, the vertical slice of the
complex valued STFT coefficients STFT(τ, f) correspond to the Fourier
spectra of the windowed signal with the window shifted by time τ. The time
frequency resolution is limited to the Heisenberg principle, which is the
main disadvantage of the linear time-frequency transform. The signal
component cannot be presented as a point in the time frequency space. Only
its position inside the ∆t⋅∆f rectangle region may be determined. This is due
to the imposition of local time window g(t). If the width of the window is
increased, frequency resolution improves but time resolution becomes poor
and vice-versa.
The Wavelet transform (WT) is a new mathematical tool developed mainly since
the middle of the 1980s. It is efficient for the local analysis of non-stationary and
fast developing transient signals. Similarly to the STFT, the Wigner distribution,
ambiguity function and the Wavelet transform map the signal to the time-scale
(frequency) joint presentation. The temporal aspect of the signal is preserved. The
WT provides a multi-resolution analysis with a dilated window. The high
frequency analysis is made using a narrow window and the lower frequency
analysis is made using a wide window. The Wavelet analysis is similar to the
Fourier analysis because it breaks a signal down into its constituent parts for the
analysis. Whereas the Fourier transform decomposes the signal into a set of sine
waves of different frequencies, the Wavelet transform decomposes the signal into
its “wavelets”, scaled and shifted versions of the “parent Wavelet”. The Wavelet
transform allows us an outstanding localisation in both the time domain via
translation of the parental wavelet and in the scale (frequency) domain via
dilatation. The translation and dilatation operations applied to the parent wavelet
are performed to calculate the wavelet coefficients representing the correlation
between the wavelet and the localised section of the signal. The wavelet
coefficients are calculated for each wavelet segment, giving a time-scale function
relating to the wavelets correlation to the signal [6, 7].
∞
1  t −τ 
WT (τ , s ) = ⋅ ∫ x(t ) ⋅ψ ∗   ⋅ dt (2)
s −∞  s 
where ‘*’ denotes the complex conjugate, x(t) is a signal, t is a time, τ is a

translation factor, s is a scale factor (frequency) and ψ (x)is a wavelet. This
transform is called Continuous Wavelet Transform because the analysing wavelet
can be used at any scale, and its position can also be shifted continuously over the
entire time domain of the signal being analysed.
As the parent wavelet (which is obviously the most suitable especially for transient
processes) the Morlet Wavelet and the Mexican Hat Wavelet are often used. The
parent Morlet Wavelet is preferred for the analysis of the vibration measurement.
The Morlet’s basic wavelet function is a multiplication of the Fourier basis with a
Gaussian window according to equation (3)
t2
−
− j ⋅ω 0 ⋅t
ψ =e ⋅e 2
(3)
Its real part is a Cosine-Gaussian function and the imaginary part is a Sine-
Gaussian function. The Wavelet transform is of a particular interest for the analysis
of non-stationary and transient signals. The Wavelet transform provides an
alternative to the classic Short Time Fourier Transform or the Gabor transform, and
it is more efficient than the Short Time Fourier Transform.
The Wigner-Ville transform [2, 4] is another alternative method for a Short-Time
Fourier Transform and the Wavelet Transform for processing both the stationary
and non-stationary signals. The Wigner distribution was proposed by Wigner in
1932 for the region of quantum physics, and about 15 years later, this was modified
for the region of the signal analysis by Ville. The Wigner-Ville transform is
defined for the time frequency region by relation
∞
 τ  τ
WVTx (t , f ) = ∫ x t + 2  ⋅ x  t − 2  ⋅ e
* ( − j ⋅ 2⋅π ⋅ f ⋅τ )
⋅ dτ (4)
−∞
where ‘*’ is a complex conjunction, t is a time, τ is a shift along the time axis, x(t)
is a time representation of the signal and WVTx(t,f) is a jointed time and frequency
representation of the signal. In contrast to the linear time frequency transforms in
which the resolution is limited by the window function, the Wigner-Ville spectrum
offers us an excellent resolution both in the frequency and time domain. The
calculation is not limited by the Heisenberg principle of uncertainty which is its
important characteristic since it is a more general transformation that does not
utilise the weighing function.
Quadratic (non-linear) methods represent the second fundamental class of time
frequency distributions. The quadratic methods are based upon estimating an
instantaneous power (or energy) spectrum using a bilinear operation on the signal
x(t) itself. The class of all quadratic time-frequency distributions to time-shift and
frequency-shift is called the Cohen’s class. Similarly, the class of all quadratic
time-frequency distributions covariant to time-shift and scale is called the Affine
class. The intersection of these two classes contains time-frequency distributions

like the Wigner-Ville distribution that are covariant to all operators [2]. Cohen [1]
describes the non-linear time frequency transformation (especially shift-invariant
class) that can be derived from the Wigner distribution.
Cohen [1] generalised the definition of the time frequency distributions in such a
way which includes a wide variety of different distributions. These different
distributions can be represented in several ways. The Cohen’s class definition like
the Fourier Transform, with respect toτ, of the generalised local correlation
function, is the most common. With a two-dimensional kernel, the bilinear time-
frequency distribution of the Cohen´s class is defined according to equation [1, 10]
 τ  τ
C x (t , f ) = ∫∫∫ e − j ⋅2π ⋅θ ⋅t ′ − j ⋅2π ⋅ f ⋅τ ⋅+ j ⋅2π ⋅θ ⋅t ⋅ψ (θ ,τ ) ⋅ x t +  ⋅ x*  t −  ⋅ dθ ⋅ dt ⋅ dτ (5)
 2  2
where x(t) is a signal, t is a time, τ is a time location parameter, ω is an angular
frequency, θ is a shift frequency parameter, a function ψ(θ, τ) is called the kernel
of the time frequency distribution. Distribution Cx (t,ω, ψ) from the Cohen’s class
can be interpreted as the two-dimensional Fourier Transform of a weighted version
of the ambiguity function of signal
C x (t , f ) = ∫∫ Ax (θ ,τ ) ⋅ψ (θ ,τ ) ⋅ e − j ⋅2π ⋅ f ⋅τ ⋅ e − j ⋅2π ⋅θ ⋅t ⋅ dτ ⋅ dθ (6)
where Ax(θ, τ) is the ambiguity function of the signal x(t), given by equation:
 τ  τ
A(θ ,τ ) = ∫ x t +  ⋅ x*  t −  ⋅ e − j ⋅θ ⋅t ⋅ dt (7)
 2  2
Note that all integrals run from -∞ to ∞. The weighted function ψ(θ, τ) is called the
kernel. It determines the specific properties of the distribution. The product Ax(θ,
τ)⋅ψ(θ, τ) is known as the characteristic function. Since function Ax (θ ,τ )
represents a bilinear operation for processing the signal, the contributions from the
so-called cross components are exhibited during its calculation, which
consequently deteriorates the differentiation of the given transformation. This
effect may be limited by a suitable choice of the so-called kernel function. Then the
kernel function unambiguously determines the properties of a given transformation.
Therefore, the kernel is often selected to weight the ambiguity function such that
the auto-elements that are centred at the origin of the (θ, τ) ambiguity plane are
passed, while the cross-elements that are located away from origin are suppressed.
This means that the suppression of cross-elements might be understood as the
frequency response of a two-dimensional low-pass filter.
When a low-pass kernel is employed, there is a trade-off between the cross-
elements suppression and the auto-element concentration. Generally, as the band-
pass region of the kernel is made smaller, the amount of the cross-element
suppression increases, however, at the expense of the auto-element concentration.
Table 1 presents definitions of the kernels for various time frequency distributions
[5].
Table 1 Definitions of the kernels for various time frequency distributions

Distribution Kernel function ψ(θ, τ)
j ⋅θ ⋅τ
Rihaczek ψ (θ ,τ ) = e
2
 θ ⋅τ 
Margenau-Hill ψ (θ ,τ ) = cos 
 2 
j ⋅θ ⋅ τ
Page
ψ (θ ,τ ) = e 2
Equation 6 can also be rewritten into the following form [5, 10]
∞ ∞
C x (t , f ) = ∫ ∫ Π(τ − t ,θ − f ) ⋅ WVT (τ ,θ ) ⋅ dτ ⋅ dθ (8)
−∞ −∞
where
∞ ∞
Π (t , f ) = ∫ ∫ψ (θ ,τ ) ⋅e
− j ⋅ 2 ⋅π ⋅( f ⋅τ −θ ⋅t )
⋅ dt ⋅ dω (9)
−∞ −∞
is the two-dimensional Fourier transform of the kernel ψ and WVT presents the
Wigner-Ville transform. Cohen’s class has a simple interpretation as a smoothed
Wigner-Ville distribution.
3. ANALYSIS OF DYNAMIC PARAMETERS
The text to follow pursues the determination of the quality of structural elements
(in this case the “hurdis” brick) by the method of acoustic response analysis. The
basis for the methodology designed by the author is the analysis of the response of
the noise signal to a mechanical impulse, particularly by time-frequency
procedures. The measured specimen was placed on a special device fixed to
supports (Figure 1). The mechanical shock was excited by a special pendulum with
a defined choice of the shock intensity. The measuring device (made by the
Bruel & Kjaer Company) consisted the PULSE 3360C signal analyser, a
microphone and the measuring software. The sampling frequency was 12 kHz. The
electric signal detected and digitally recorded resulted from the measurement, and
it was adequate to the amplitude of the acoustic pressure in the place of the
microphone location.
After completing the analysis, the check measurement and calculations, the
following methods and parameters were used to the analysis of the response to the
mechanical shock:
• diagram time history of the acoustic pressure
• frequency analysis with help of behaviour of the power spectral density (the
algorithm of the fast Fourier transform was used)
• linear jointed time and frequency methods of the spectral analysis
(the algorithms of the Wavelet transform were used)
• non-linear jointed time and frequency methods of the spectral analysis
(the algorithms of the Rihaczek transform were used)
The analyzed figures are composed of three diagrams. The upper diagram shows
the time history of the acoustic pressure. In the left diagram, the amplitude
spectrum of the acoustic response calculated by the direct employing of the Fourier
transform on measured signal is shown. The middle diagram shows the 3D view of
jointed time and the frequency spectrum of the amplitude spectrum of the acoustic
pressure. The spectra in particular diagrams in Figures 2 to 5 were gradually
calculated by means of the Wavelet transform or more precisely by the Rihaczek
transforms.
The values of the acoustic pressure in the decibel scale are depicted in the middle
diagrams in different colours. It should be said that the maximum value is in black
colour. Figures 2 to 5 present three groups of diagrams of the signals measured by
the microphone using the impact-echo method used to a defective product and for
the defect-free one.
Comparing the time records of a good and a defective specimen, it may be
considered that the signal coming from the good product has a lower damping, i.e.
the specimen sounds for a longer time.
Figure 1 General view of the working place
Then the frequency characteristic shows that in a defective specimen the distinctive
frequency components have “moved” in the direction of the lower frequency
values and that the spectrum is wider compared with defect-free specimens. The
comparison of the course of the time-frequency transform clearly shows the
different characters of the spectra of the defective specimens and the defect-free
specimens.
The text to follow will only deal with the diagram of the time-frequency
characteristics obtained by means of different mathematical procedures.
The Wavelet transform will be considered as a basic transform. The time-frequency
curve for the defect-free product (Figure 2) clearly shows two frequency
components on frequencies 2.5 kHz and 3.3 kHz. The component on frequency
2.5 kHz is damped by 20 dB approximately in 30 ms, while the frequency
component 3.3 kHz is damped in 50 ms. These two frequency components
compared with other frequency components are remarkable in their spectra. The
change of the time-frequency spectrum of a defective specimen (Figure 3) can
easily be distinguished. The distinctive frequency region, approximately from
700 Hz to 3.2 kHz contains more than two important frequency components – e.g.
it is possible to choose the values on 700 Hz, 1.1 kHz, 1.8 kHz, 2.2 kHz, 2.5 kHz,
2.8 kHz and 3.2 kHz – their values are less than one order lower than the value of
the maximum component. The time periods for damping of these frequency
components by 20 dB are usually shorter than the periods for selected components
in defect-free specimens. These periods reach the values of damping approximately

between 15 ms and 30 ms.
Let us mention that the Wavelet transform is one of the basic and also fast
procedures for the time-frequency analysis of signals. However, the accuracy and
appropriateness of this method depends upon the choice of the window function.
The application of the method requires a certain experience gained for the
“rational” definition of input parameters and also in the interpretation of its
spectrum.
Therefore it is often more advantageous to use non-linear time frequency
transformations for this analysis of the transient signals. The characteristic feature
of non-linear transformations rests in their resultant resolution in time and
frequency which is not limited by the Heisenberg uncertainty principle. This fact
includes a high resolution capability in the time frequency level which
demonstrates itself by the “accurate” localisation of important frequency
components in time.
The Rihaczek transforms applied to both signals (defective ones in Figure 4 and
defect-free ones in Figure 5) show a similar jointed time-frequency spectrum of the
standardised acoustic pressure level as in case of the Wavelet transform. The time-
frequency spectrum (the Rihaczek transform) shows local frequency maximums of
particular frequency components in time more distinctively (precisely) than it is
with the spectrum calculated by the Wavelet transform. This is given especially by
the type of transform where mainly the kernel and the local auto-correlation
function participate in the phenomenon presented.
The time-frequency spectrum calculated by the Rihaczek transform shows sharper
courses of particular distinguished frequency components ("slighter lines") than it
is with the spectra calculated by the Wavelet transform.
As apparent from Figures 2 to 5, the whole process of response may be roughly
divided into three stages. In the first stage, a very fast growth of the amplitudes of
the key frequency components to maximum values is apparent. In the second stage,
the descent and extinguishing of higher frequency components occur. However, in
both phases of the response the contents of the frequency cross-sections remain
approximately the same. In the third stage, the fading of the response occurs. This
is usually characterised by the existence of the lowest frequency component.
If some material cracks occur, the speed of the growth and the declination of
amplitudes of the spectrum are higher than for a defect-free material. It should,
however, be stated that the speed of the growth and declination of amplitudes of
particular important elements are not the same. The frequency or the time-
frequency spectrum of defective materials is generally much wider, and the so-
called clusters of important frequency components occur there. This phenomenon
is properly interpreted by particular diagrams in Figures 2 to 5.
Figure 2 Defect-free product - time behaviour signal from microphone,

frequency spectrum and time - frequency analysis by Wavelet transform
Figure 3 Defective product - time behaviour signal from the microphone,

frequency spectrum and time - frequency analysis by the Wavelet transform
Figure 4 Defect-free product - time behaviour signal from the microphone,

frequency spectrum and time - frequency analysis by the Rihaczek transform
Figure 5 Defective product - time behaviour signal from the microphone,

frequency spectrum and time - frequency analysis by the Rihaczek transform
4. CONCLUSIONS
Based on the measurement and analyses made, it is possible to state that the
experiment checked the possibilities of using the given methodology for the
detection of structural defects of the measured products and materials from the
homogeneity and the cracks formation points of view. The analysed parameters
enable us to distinguish a good product from a defective one. The given
methodology may be successfully applied to all ceramic and concrete products. It
can also be established that modern means of signal analysis, especially the time-
frequency transforms highly contributed to a quality processing of the
measurement. These methods provide us with a time localisation of frequency
components contained in the measured signal. In this way, these methods offer new
possibilities in the experimental analyses of structural elements and materials. This
possibility is included in both the linear and the non-linear time-frequency
procedures. By comparing the obtained results we can clearly observe that when
analysing the responses to the mechanical shock (heavily damped time signals) the
procedures, namely those including into the class of the Cohen time-frequency
transforms, show a very good time localisation of significant frequency
components contained in the analysed signal.
In some particularly important cases the depicting of results of the computes and
analyses by the time and frequency sections may be completed. Thus, these provide
us with a profound support in the analysis of the
time-frequency results. This procedure often appears as more suitable than e.g. a
separate space arrangement. This makes it possible to locate very precisely the time
behaviour of particular important frequency components or to show all important
frequency components contained in the spectrum in a given time.
In conclusion it may be mentioned that the given methodology can also be
incorporated in the process of the half-automated quality control of products under
in the production line. When using the above methodology together with the
methods of qualitative analysis or the artificial intelligence (fuzzy and rough sets,
neural networks, genetic algorithms etc.), then the process of evaluating the quality
of products can fully be automated.
Acknowledgements

and by research project MSM 0021630519 (“Progressive reliable and durable
load-bearing structural constructions”)
References
1. L. Cohen: Time-frequency distributions - a review, Proc. IEEE, Vol. 77 No. 7, 941-981, 1989
2. Wahl T. J.: Bolton J. S., The Application of the Wigner Distribution to the Identification of
Structure-Borne Noise Components, Journal of Sound and Vibration, pp. 101-122, 1993, ISBN
0022-460X/93/100101
3. Poularikas A. D.: The Transform and Applications Handbook, IEEE Press, 1996
4. J.C. O'Neill: Quartic Functions for Time-Frequency Analysis with Applications to Signal
Adaptive Kernel Design, SPIE - Advanced Signal Processing Algorithms, 1997
5. Hammond J. K., White P. R.: The Analysis of Non-Stationary Signals Using Time-Frequency
Methods, Journal of Sound and Vibration, pp. 419-447, 1996,
and ISBN 0022-460 X/93/100101
6. Marasek C., Piotrkowski R., Serrano E., Ruzzante J. E.: Monitoring of the tool condition with
acoustic emission signal analysis using wavelet packets, Insight - Non-Destructive Testing and
Condition Monitoring Vol. 44 No. 12, December 2002, ISSN 13542575
7. Chang Y. F.: Wavelet deconvolution before scanning in ultrasonic non-destructive testing,
Insight - Non-Destructive Testing and Condition Monitoring Vol. 44 No. 11, November 2002,
ISSN 13542575
8. Young S. Cho: Non-destructive testing of high strength concrete using spectral analysis of
surface Wales, NDT & E International, Volume 36, Issue 4, June 2003, pp. 229-235, ISSN 0963-
8695
9. Allen G. Davis : The non-destructive impulse response test in North America: 1985–2001
NDT & E International, Volume 36, Issue 4, 2003, pp. 185-193, ISSN 0963-8695
10. Smutný J., Pazdera L.: New techniques in analysis of dynamic parameters rail fastening, InSight,
The Journal of The British Institute of Non-Destructive Testing, Vol. 46, No 10, October, 2004,
pp. 612-615, ISSN 13542575
Design of reinforced soil works – Textomur structures – based on

the computer program Cartage
Anghel Stanciu1, Oana Colţ 2 and Irina Lungu3
1
Department of Roads and Foundations, Technical University, Iasi, 700050, Romania
2
3
Summary
Reinforced soil represents an alternative solution to consolidate earth massifs and

perform retaining structures included in the present transportation infrastructure.
One of the advantages consists in applying this type of structure in poor foundation
soil conditions, due to the fact that the induced pressures are small and uniformly
distributed. Construction costs may also be decreased by using the local material
from the construction site as filling material within the reinforced soil work.
Finally, the resulted platform may also be support for additional construction
works.
Regarding the construction method, the variety of the reinforcements (steel,
geosynthetics, geogrids) or of the facing elements (precast concrete blocks, steel
structures, reversed geogrids or geosynthetics) allows various performing
procedures to develop when building reinforced soil structures.
Textomur system is included in the category of reinforced soil with geosynthetics,
developed mainly in France. The main difference between this system and other
reinforcing techniques consists in building up the facing element, by using
prefabricated steel elements.
The design of the reinforced soil works in Textomur system is implemented via a
computer program – Cartage. This program was initiated in 1985 LCPC
(Laboratoire Central du Ponts et Chaussée), France and it is based on the
analytical calculation method developed also at LCPC, considering the limit
equilibrium.
By running this computer program stresses within the reinforcements can be
evaluated, the displacements of the soil massive, and the safety coefficients at
anchorage as well, based on the stability analysis of the massive in case of
cylindrical failure surfaces going through the toe of the reinforced soil structure.
KEYWORDS: reinforced soil, retaining structures, geosynthetics, Textomur,

Cartage.
270 A. Stanciu, O. Colţ, I. Lungu
1. INTRODUCTION
The first modern system of soil reinforcing was initiated and developed in the mid
‘60s by the French engineer Henri Vidal, with the name Terre Armeé (reinforced
soil), using steel reinforcements. Beginning with 1970, geosynthetics have been
created as an alternative of the steel reinforcements. The acceptance of
geosynthetics in reinforced soil construction has been triggered by a number of
factors, including aesthetics, reliability, cost, simple construction techniques, good
seismic performance, and the ability to tolerate large deformations without
structural distress.
In Romania, the first attempts of performing retaining walls by reinforced soil were
made in 1973...1974, when such a construction work was designed and built over
50m length. After 1990 many retaining walls have been performed as soil
reinforced with geotextiles and geogrids within road rehabilitation works.
Textomur system is included in the category of reinforced soil with geosynthetics,
developed mainly in France. The main difference between this system and other
reinforcing techniques consists in building up the facing element, by using
prefabricated steel elements.
2. TEXTOMUR REINFORCING SYSTEM
Textomur is essentially a reinforcing technique of soil using geogrids or

geotextiles. There are many variants of the Textomur system and the name
corresponds to Textomur system with mineral, vegetal, or neutral facing elements.
The facing elements of Textomur have a length of 4.75m the steel within has a
diameter of 6 and 8mm. For the Textomur system with mineral facing the
formworks are galvanized after welding and bending at the appropriate angle. The
height of one layer is 60-65cm, being variable depending on the requirements of
the project. The slope of the formwork may vary between 40° to 90°. The
reinforcement length is established for each project but an initial 0.7 of the total
height is recommended.
2.1. Advantages
• Local material from the construction site may be used as filling material.
• Flexibility – curves, edges, terraces of variable length may be performed.
• Long lasting life – the life span is estimated for approximately 120 years.
• Easy to built, without a foundation.
• Construction may be performed without additional bracing.
2.2. Textomur with mineral facing
Textomur with mineral facing includes galvanized welded steel facing elements
behind which large boulders of 30-100cm are manually placed over a width of
30cm and along the entire height of the facing (figure 1). The boulders are
preferred to be with plan surfaces and consequently the arrangement will display
less voids.
Figure 1. Cross section of a Textomur system with mineral facing
This type of Textomur is best applied for retaining structures (figure 2), the ensure
slope stability, for significant height, reinforced soil works in steps can be
performed, with an aesthetic role as well.
Figure 2. Retaining wall of Textomur with vertical mineral facing

3. COMPUTER PROGRAM - CARTAGE
The computer program Cartage is based on the LCPC calculation method of the
reinforced soil [1], [2], considering the limit equilibrium of the massive. Slope
stability is analyzed based on circular failure surfaces through the toe of the
massive. The program computes the stresses within reinforcements, considering
each failure surface. The safety coefficient for internal stability is set as 1.5. The
maximum stress in each reinforcement resulted from the program is considered the
service stress for the corresponding failure surface. The calculation continues until
the safety requirements are met, altering the reinforcing scheme or the
reinforcement type.
3.1. Calculation method of L.C.P.C.
The method is based on Rankine’s theory to compute the active earth pressure on
the retaining element (figure 3). The soil layer horizontally placed between the i-1
and i reinforcements is considered to establish the tension force within the
reinforcement [3], [4]. Regarding the main stress as the vertical one ( σ v = γ ⋅ H ),
the tension force in the i reinforcement would counteract the active earth pressure
acting on the facing of height ∆H .
H
σv=γ H
σh=Κa σv i-1
i
Τ=Κa σv ∆H
∆H
Figure 3. Principle of L.C.P.C. method
Consequently the tension force is given by:

1
Ti = γ ⋅ H ⋅ K a ⋅ ∆H + ⋅ K a ⋅ γ ⋅ (∆H )
2
(1)
2
but H = i ⋅ ∆H and thus the final relationship is the following:
 1
Ti = 1 +  ⋅ K a ⋅ γ ⋅ ∆H 2 (2)
 2
3.2. Design hypotheses
The stability analysis is performed considering the simultaneous action of several

pessimistic factors the highest height, a horizontal platform at the top, the service
load as close as possible towards the facing, and ignoring the presence of the
material in front of the reinforced structure toe.
A case study is presented as a retaining wall on the road RN 94, Embrun, France.
The less favourable case is adopted in the design profile (figure 4), height of 5,40m
corresponding to 9 levels of Textomur formwork, with an inclination of 270 [5].
The presence of soil over the first formwork height was disregarded.
Figure 4. Design profile for Cartage, considering one potential failure surface
The geotechnical parameters of the local material are: unit weight - γ = 21kN/m3;
internal friction angle - φ = 35o; cohesion - c = 5kPa.
For filling the space between reinforcements, the local material is considered
cohesionless.
Initially, 5 layers of geotextile were considered, with strength in tension of
100kN/m, with 0.60m interdistance, 2 layers with 200kN/m strength at 0,60m
interdistance and 3 layers of 200kN/m strength at 0,30m interdistance.
The input data for Cartage are represented by: geometry of the massive;
characteristics of the filling soil, foundation soil and soil behind the retaining
structure; characteristics of the reinforcements; coordinates of the failure circles;
displacement value at the massive top; the presence of underground water.
The input data are presented in figure 5, a verification of their accuracy being
easily done.
Figure 5. Format of the input data for the computer program Cartage
3.3. Results
By running the program the output data are presented in a table consisting of the
extreme values for the safety coefficients according to the failure surfaces
considered, for the displacement, coefficient of anchorage and maximum stress.
Based on the required values by the designed norms, the necessity of altering the
reinforcing scheme is analyzed.
The alteration of the reinforcing solution is made according to the difference
between the imposed and obtained values and not in the least, based on the design
experience of the engineer.
Within the presented case study, the final values are set in a table for the first 12
failure circles as in figure 6 [5].
Figure 6. Final values as output data from the Cartage program

LEGEND:
• Values between parantheses ( , ) - number of the massive and reinforcement
with the smallest value of the coefficient of anchorage (ANCR), respectively
the maximum stress (TENR) for each failure circle.
• DELTA – displacement value ;
• F0 – value of the safety coefficient with reinforcing the massive ;
• FSOL – value of the safety coefficient after reinforcing the massive.
4. CONCLUSIONS
Using computer programs to design retaining structures made of reinforced soil

decreases significantly the design time. In order to perform an efficient design it is
required to master correctly the behaviour of the reinforced soil generally, and
especially the response of such structures subjected to various external loads.
References
1. Silion T., P. Răileanu, A. Stanciu - Fundaţii - Pământ armat, Institutul Politehnic Iaşi, Facultatea
de Construcţii, 1980.
2. Cartage – manual de utilizare, L.C.P.C, 1985.
3. Schlosser F - La terre armée, Recherches et realisations, Bull. liaison Labo P. et.ch. 62, nov. –
dec., 1972.
4. Schlosser F., Long T.N. – Dimensionnement de murs en terre armee, Session de formation
permanente, E.N.P.C., 1974.
5. Proiect tehnic şi detalii de execuţie - „RN 94 - Déviation d’Embrun", Geonove, Franţa.
Computer program for Pipe Section Column Sizing

Dragos Voiculescu1, Daniela Preda1
1
Steel Structures, Technical University of Civil Enginering, Bucharest, Romania
Summary
The paper presents the Excel Program for Pipe Section Column Sizing. The
program has three main parts: the first one is the data input, where the proposed
size of the column is set; the second one is the load section, where the maximum
loads are input from the structural analysis program, such as SAP or ETABS; the
third part gives the results of the structural checks on the proposed section.
The program is interactive, meaning that if one of the checks is wrong, it can be
corrected by changing the size of the Pipe section.
The program is very useful in applied engineering design, as it gives immediately
the final section of a specific column.
KEYWORDS: pipe section; column; sizing.
1. INTRODUCTION
The computer program was developed in Microsoft Excel in order to be used

without major problems on any machine. It has some Visual Basic features to make
it very easy to use; also it was realized in such a manner to be printed as calculation
notes in real projects.
2. PROGRAM PRESENTATION
The program has three main parts, interconnected. The first part is the data input,
where the initial geometrical characteristics of the column are given. This data is
used to compute all sectional characteristics of the pipe section, which will be used
later in the program. Also the section class is shown. (Fig.1)
Immediately after, the pipe steel grade is input. The user may choose in between
the two common steel grades for pipes. After choosing the steel grade, the factored
strength is shown. All strength checks will be reported to this strength.
278 D. Voiculescu, D. Preda
Figure 1.Input data
After this data is established, the program demands the input of the buckling length
factors (Fig.2). These ones are computed using another program, according to the
column end supports and beam characteristics. The buckling length factors will be
used to compute the buckling coefficient which will be used further.
Figure 2.Buckling data input
Then the second part of the computer program starts, the part where the column
loads are input. These loads are taken from a structural analysis program such as
SAP or ETABS. The data is input in a table, and is used in all the strength and
stability checks. (Fig.3)
Loads are given in three hypothesis: The first one is the maximum axial load with
the correspondent bending moments; the second one is the maximum z-z bending
moment with the correspondent others and the third one is the maximum y-y
bending moment with the correspondent others. All further calculations is referred
to these hypothesis.
Figure 3.Load input
The third part of the computer program starts with the strength checks for the
proposed pipe section. Here the applied formulas are written and the result is given
for each hypothesis. (Fig.4)
Figure 4.Strength checks

280 D. Voiculescu, D. Preda
Also the buckling checks are given in this part of the program (Fig.5).
Figure 5.Buckling checks

3. CONCLUSIONS
This program is very useful when fast section sizing is demanded. It may be
printed and gives the calculation sheet for the specific column.
References
1. STAS 10108/0-78.
Computational method and application using Newton Algorithm

Cornelia Victoria Anghel
Computer Science Department, “EFTIMIE MURGU” University, Resita, 320085, Romania
Summary
The article presents some aspects about Newton method and algorithm and a
numerical example. The C++ program illustrated the important steps and
advantages to apply the Newton method to solve the nonlinear equations.
KEYWORDS: Newton (tangent) method, Newton algorithm, C++ program.
1. INTRODUCTION NEWTON METHOD (TANGENT METHOD)
The Newton Method (or Tangent Method) is used to solve the un-linear equations,
and to reduce substantially the number of iterations.
If we considered x* - a separate solution of equation f(x) = 0 on [a, b] interval and
f(x) is continue and monotone on this interval, the values of function in a and b
points f(a) and f(b) have opposed signs (such as in Figure 1).
Figure 1.
We write the tangent equation in A(a, f(a)), point

y – f(a) = f ’(a)· (x - a)
intersect the Ox axis.

f (a ) f (a )
We obtain x=a− or x1 = a − .
f ' (a ) f ' (a )
From geometrical point for view (figure 1) signifies the fact that the arch/spring
curve is replaced by the tangent, and the place of the exact value solution x1 (the
abscissa/x-coordinate of the intersection point of the curve and the axis Ox) it is
taken as an approximate value x1 (the abscissa of the intersection point of the
tangent at the curve with the axis Ox).
By repeatedly applying Newton’s method we can correct the first approximation,
finding closer values to the exact one.
The approximate values result from the series:
f ( x n −1 )
x n = x n −1 −
f ' ( x n −1 )
2. THE ALGORITHM OF THE TANGENT METHOD

(NEWTON)
• We draw the tangent at the curve in the point B(b, f(b)) and we find the
intersection point of the tangent and the axis OX – (x1,0). The abscissa of this point
x1 is taken as I the solution approximation.
• Computing the function value in x1 we could deduce the tangent at the
curve in the point with the coordinates (x1, f(x1)). The tangent will intersect the axis
OX in the point with the coordinates (x2 , 0). The abscissa of this point x2 0 taken as
I solution approximation.
• Continuing this process several times, we can observe that we obtain a
value series x1, x2, x3, ... , which tend towards the equation solution x°.
3. NUMERICAL APPLICATION
We propose to solve the equation x3 – 2x – 5 = 0.

Solution: We notice that if we note f(x) = x3 – 2x – 5, then
f(1) = - 6; f(2) = -1; f(3) = 16.
From the conditions f(2) < 0 and f(3) > 0 we deduce that a solution of the equation
appears in the interval (2, 3).
284 C.V. Anghel
We consider the first approximation x1 = 2.

f (2)
We determine x 2 = 2 − .
f ' (2 )
For f(x) = x3 – 2x – 5, f’(x) = 3x2 – 2 and f(2) = -1; f’(2) = 10
1
x2 = 2 + = 2 + 0.1 = 2.1 .
10
f (2.1)
We repeat the sentence and we get: x3 = 2.1 −
f ' (2.1)
where f(2,1) = (2,1)3 - 2·(2,1) – 5 = 0,061.
f’(2.1) = 3·(2,1)2 – 2 = 3· 4,41 - 2 = 13,23 – 2 = 11,23.
0.061
It results x3 = 2.1 − = 2.1 − 0.0054 = 2.0946
11.23
f ( x3 )
Then we determine x 4 = x3 −
f ' ( x3 )
f (2.0946)
By substitution we get: x 4 = 2.0946 − = 2.09455.
f ' (2.0946)
We approximate x4 = 2,095.
The value of the function is:
f(2,095) = (2,095)3 – 2 ·(2,095) – 5 = 9,195 - 4,190 – 5 = 9,195 – 9,190 = 0,005
hence f(2,095) > 0.
From the conditions f(2) = -1 < 0 and f(2,095) > 0 we determine that the value x =
2,095 can be considered the solution of the given equation.
As an observation, we can affirm that if the graphic of the function f(x) is almost
parallel to the ax Ox, i.e. the size of the derivative f’(x) is closed of zero, then the
Newton method is not applicable, but another method is recommended, for
example the string/cord method.
4. PROGRAM C++
The program written in C++ which determines the solution of the function given in
the above mentioned example, using the algorithm of the tangent method (Newton)
is:
#include <stdio.h>
#include <conio.h>
#include <math.h>
float Func(float x)
{
return x*x*x-2*x-5;
}
int Newton(float Func(float, float *x), float *x)
{
const float eps=1e-6;
const int itmax=100;
float df, dx, f;
int it;
for(it=1; it<=itmax; it++){

f=Func(*x, &df);
dx=(fabs(df) > eps) ? -f/df : -f;
*x += dx;
if (fabs(dx) <= eps*fabs(*x)) return 0;
}
printf("maximum no. of iterations exceeded within the Newton method! \n");
return 1;
}
The following notations were used:

Func – for the user function
a, b - the limits for the search interval
*x - the found zero (output)
it - the number of iterations
itmax – the maximum number of iterations
The error number is returned:

0 – for a normal execution
1 – if the maximum number of iterations is exceeded
2 – when the interval does not contain a root.
After each partition of the interval, not only the ends of the interval are upgraded,
but also the corresponding values of the function f(ai) and f(bi).
5. CONCLUSIONS
This method helps us to efficiently obtain the solution of the problem. The small
number of iterations necessary for the determination of the equation solution from
the presented example indicates that the choice of the tangent method (Newton) in
286 C.V. Anghel
the solving of this equation was the right one. These observations can be picked out
from the analysis of the results obtained after the development of the program
implemented in the above presented programming language C++.
If the graphic of the function f(x) is almost parallel to the Ox axis, i.e. the size of
the derivative f’(x) is closet o zero, then the Newton method is not applicable, but
another method is recommended, for example the string/cord method.
References
1. Anghel C.V. – Metode numerice. Algoritmi şi programe de calcul. Ed. Orizonturi Universitare,
Timişoara 2005 (in Romanian);
2. Anghel, I., Anghel, C.V.- Algebra liniară. Programare liniară, Curs vol.1, Ed. Eftimie Murgu,
Reşiţa, 2003 (in Romanian);
3. Precup, R. ş.a – Matematici asistate de calculator. Aplicaţii. Ed. Politehnică, Timişoara, 2002 (in
Romanian);
4. Kilyeni, Şt. – Metode numerice, vol.1, Ed. Orizonturi Universitare, Timişoara, 1997 (in
Romanian).
Partial substitution of Cement in Concrete by Finely Ground

Brick Body
Libuše Beckerová, Gergely Bölckei and Jiří Brožovský
1,2,3
Faculty of Civil Engineering, Brno University of Technology, Brno, 602 00, Czech Republic
Summary
Brick chippings are the waste which can partially substitute the cement. Therefore
this waste is utilized, in light of the environment, into the useful new product - the
concrete. The finely ground brick body is characterized by certain pozzuolana
activity which enables the partially substitution of the energetically demanding
binder which is the cement. The paper indicates the observations concerning the
tests of concrete in which the cement was substituted by 10% and 20% of finely
ground brick bodies from the brick plants Kryry and Šlapanice.
KEYWORDS: brick body, concrete, additive to concrete, pozzuolana activity.
1. INTRODUCTION
The brick plants produce on the average 1 till 5 % of waste, which is recycled only
in low range and utilized as opening material. The rest is stored in deposits. The
price of this waste material rarely overreaches 5 till 10% of the cement price. It is
true that for the utilization as substitute for cement it is necessary to prepare it
(by grinding, separation etc.). However also after this preparation its price does not
overreach 20% of cement production costs. The finely ground brick body shows
pozzuolana activity so that we can hold it for the active pozzuolana admixture and
we can substitute the part of binder in cement composite by this material.
By the addition of finely ground brick body we can achieve, with comparable
concrete parameters, the decrease of production costs.
The reaction extent of the burned brick body with the hydrating cement is given on
the one hand by the composition of raw materials and on the other hand by the
burning conditions.
The brick products are on the present as a rule burned at the temperature 1000 –
1100°C. The brick body is formed by raw glass surrounded by other crystals and it
can show variable pozzuolana activity in dependence on the used earth and on the
burning temperature. The greatest portion in the resulting product forms as the rule
the glassy phase with high content of silicon and of alkalis and it contents further
288 L. Beckerová, G. Bölckei, J. Brožovský
also aluminiuum, different iron oxides, calcium and magnesium surrounded by

crystals of mullite, silicon and sometimes also cristoballite. The brick products
with high lime content can also contain considerably amounts of feldspars which
support the increase of activity.
The reaction activity is highest mostly in the temperature range between 600 –
900°C. Higher temperature causes nucleation and crystallization of new phases in
the body and the reaction activity rather decreases. The glassy phase excels during
cooling and it is in dependence on the composition more or less active [4].
The following active admixtures were applied in practice:
- Finely ground brick body from the brick plant KRYRY
- Finely ground brick body from the brick plant ŠLAPANICE
The effect of the cement substitution in concrete by finely ground brick body, on
its physico-mechanical characteristic, was tested in experimental work.
2. METHODS OF EXPERIMENTAL WORK AND THE

COMPOSITION OF CONCRETE
The experimental work is based on the properties comparison of concrete made

with different batch of finely ground brick body (substitution 10 % and 20 % of the
cement weight) with the properties of reference concrete which was made without
the use of brick body.
2.1. Input materials:
The following components were used for the production of concrete:
− cement CEM 42,5R from cement plant Mokrá
− sand fraction 0-4 mm from the locality Žabčice
− crushed aggregate fraction 4-8 mm from the locality Želešice
− crushed aggregate fraction 8-16 mm from the locality Želešice
− admixture: finely ground brick body from the brick plant KRYRY and
ŠLAPANICE
− The volume weight of these brick bodies are 2750 kg/m3
− Additive: poly-carboxyl-ether ACE 40, from the firm BASF which shows
liquefying effects, increases the initial and the final strength of concrete
- batch water – it fulfils the demands of the EN 1008.
The basic parameters of the aggregate are in table 1.

Table 1: Basic parameters of the aggregate

Žabčice Želešice Želešice
Parameter Unit
0/4 mm 4/8 mm 8/16 mm
Volume weight [kg/m3] 2630 2650 2730
Bulk weight in shed
[kg/m3] 1770 1670 1710
state
Porosity [%] 32,7 37,0 37,4
2.2. Composition of concrete
The composition of concrete is in table 2.
Table 2. Composition of concrete in 1 m3

Concrete Concrete
Reference with10% with 20%
Component
concrete of brick of brick
body body
CEM I 42,5 R Hranice cement 400 kg 360 kg 320 kg
Finely ground brick body 0 kg 40 kg 80 kg
Sand 0 to 4 mm from the Žabčice
745 kg 740 kg 739 kg
Gravel Pit
Aggregates 4 to 8 mm from the
235 kg 230 kg 229 kg
Želešice Gravel Pit
980 kg 972 kg 971 kg
Želešice Gravel Pit
Plasticizer BASF ACE 40 4 kg 4 kg 4 kg
Mixing Water 151 kg 147 kg 141 kg
Water/Cement Ratio 0,38 0,37 0,35
2.3. Observed parameters
The following parameters were observed with tested concrete:

− fresh concrete - flow table test F (EN 12350-5 Standard)
− density of fresh concrete D (EN 12350-6 Standard)
− density of hardened concrete D (EN 12390-7 Standard)
− compression strength fc,cu (EN 12390–3 Standard) in the age of concrete
1, 3, 7, and 28 days
− bending strength fcf (EN 12390–6 Standard) in the concrete age 28 days
− depth of penetration of water under pressure htv (EN 12390–8 Standard)
− dynamic modulus of elasticity Ebu (CSN 731371 Standard) in the age of
concrete 1, 3, 7, 14 a 28 days.
Marking of individual tested concrete:

REF - reference concrete (strength class C 55/67)
S10 - concrete with 10% substitution of cement by finely ground brick body
from brick plant Šlapanice
S20 - concrete with 20% substitution of cement by finely ground brick body
from brick plant Šlapanice
K10 - concrete with 10% substitution of cement by finely ground brick body
from brick plant Kryry
K20 - concrete with 10% substitution of cement by finely ground brick body
from brick plant Kryry
3. TEST RESULTS
The test results are in table 3 and for selected parameters they are
represented graphically in figures:
- Fig.1: Effect of the finely ground brick body addition on the compression
strength of concrete with different age
Fig.2: Effect of the finely ground brick body addition on the bending
strength of concrete and on the ingress of pressure water - age of concrete
28 days
- Fig.3: Effect of the finely ground brick body addition on the dynamical
elasticity modulus – for different ages of concrete.
Table 3: Results of concrete tests
Age
Properte Unit REF S10 S20 K10 K20
[days]
FRESH CONSCRETE
F -- [mm] 370 435 450 420 435
D -- [kg/m3] 2500 2480 2460 2470 2460
HARDENED CONCRETE
D 28 [kg/m3] 2520 2490 2480 2490 2480
1 39,8 30,2 20,8 28,6 20,4
3 49,8 40,4 34,9 36,5 33,6
fc,cu [MPa]
7 59,6 54,6 43,1 52,6 44,1
28 74,0 68,6 56,6 69 52,1
fcf 28 7,8 6,5 5,5 6,6 5,5
htv 28 [mm] 11 8 6 10 7
1 47,5 42,8 40,3 40,8 40,7
3 52,2 46,5 44,5 44,6 44,5
Ebu 7 [GPa] 56,7 52,8 50,3 51,1 49,8
14 57,1 53,0 51,7 52,2 51,5
28 57,4 54,6 52,0 52,5 52,5
80
70
60
50 1 day
fc, cu [MPa] 40 3 days
30
7 days
20
28 days
10
7 days
0
REF S10 1 day
S20 K10 K20
Figure 1.Compression strength in dependence on the quantity of added brick body and on
the hardening time of concrete
9 12
8
10
7
6 8
fcf [MPa]
htv [mm]
5
6
4
3 4
2
2
1
0 0
REF S10 S20 K10 K20
fcf htv
Figure 2.Effect of the quantity of added brick body on the bending strength and on the
depth of the pressure water ingress
60
50
40
Ebu [GPa] 30
20
10
0
REF S10 S20 K10 K20
1 day 3 days 7 days 14 days 28 days
Figure 3.The dynamical elasticity modulus in dependence on the quantity of added brick
body and on the hardening time of concrete.
4. CONCLUSION
Experimental works proved the usability of finely ground brick body as active
admixture into concrete. The test results showed that the utilization of this active
admixture is also adequate for the improvement of some special concrete
properties.
a) Compression strength:
The concrete with the substitution of 10 % cement by finely ground brick body
achieved after 28 days 92 - 93% of the reference concrete compression strength
values.
The concrete with the substitution of 20 % of cement by finely ground brick body
achieved after 28 days 71 – 73% of the reference concrete compression strength
values.
b) Bending strength:
The bending strength decreases with the increasing content of the brick body. The
highest strength had the reference concrete.
c) Depth of the pressure water ingress:
The depth values of pressure water ingress into the reference concrete were
comparable with the water ingress into concrete with the addition of brick body
and they are very favourable.
d) Dynamical modulus of elasticity:
The dynamical modulus of elasticity decreases with the increasing content of the
brick body. The reference concrete had the highest value.
The results of experimental works show that better values of compression strength
are achieved in the case of 10 % cement substitution by brick body – the concrete
class corresponds with the concrete class of the reference concrete. Concrete with
20 % cement substitution by brick body achieves values which are by a class
lower.
We can conclude that the utilization of finely ground brick body in concrete is
useful from the ecological and also from the economical point of view.
ACKNOWLEDGEMENTS
The work was supported by the MSM 0021630511 Plan: Progressive Building
Materials with Utilization of Secondary Raw Materials and Impact their on
Structures Durability.
References
1. Drochytka, R.et al. Progressive Building Materials with Utilization of Secondary Raw
Materials and their Impact on Structures Durability. Brno University of Technology, Final
report of the project VVZ CEZ MSM: 0021630511, Brno 2006. Brožovský, J.: Subtask 3 (in
Czech).
2. Brožovský, J. and Martinec, P. Durabbility of Concrete with Fly Ash. In Proceedings of the
2d International Conference on Concrete and Reinforced Concrete: Concrete and
Reinforced Concrete – Development Trends. NIIZHB, Moscow, Russia, Vol. 4, 2005. (in
Russian).
3. Brožovský, J., Zach, J., Brožovský, J., Jr.: Durability Of Concrete Made From Recycled
Aggregates. In Proceedings Of The International Rilem Jci Seminar Concrete Durability
And Service Life Planning: Curing, Crack Control, Performance In Harsh Environments,
Ein-Bokek, Dead Sea, Israel, 2006
4. Wild, S., Gaillius, A., Hansen, H., Pederson , L. , Szabowski, J. : Pozzolanic proprietes of a
variety of european clay bricks. Building research and information vol. 25, number 3, 1997
5. EN 12350-5 Testing fresh concrete - Part 5: Flow table test
6. EN 12350-6 Testing fresh concrete - Part 6: Density
7. EN 12350-3 Testing hardened concrete - Part 3: Compressive strength of test specimens
8. EN 12350-5 Testing hardened concrete - Part 5: Flexural strength of test specimens
9. EN 12350-7 Testing hardened concrete - Part 7: Density of Hardened Concrete
10. EN 12350-8 Testing hardened concrete - Part 8: Depth of penetration of water under
pressure
11. CSN 73 1371 Method of Ultrasonic Pulse Testing of Concrete
12. EN 1008 Mixing Water for Concrete. Specification for Sampling, Testing and Assessing the
Suitability of Water, Including Water Recovered from Processes in the Concrete Industry,
as Mixing Water for Concrete
Minimum weight buildings design using inequalities method

Constantin Amariei1, Iulian Gabriel Mihai2
1
Structural MechanicsDepartment, TU“Gh.Asachi”Iaşi, 700050, România
2
S.C.EdilConst S.A Câmpina, Prahova, 105600, România,
Summary
Based on inequalities method and on possibilities of solving by automatic

computation mathematical computation model of minimum weight steel structures
is presented.
KEY WORDS: optimum design, minimum weight, inequalities method, linear

algorithm.
1. GENERAL PROBLEMS
Optimization computation due to weight criterion can be done in many ways,

according to structure equilibrium equations which are designed function of weight
reduction condition formulation. Most important ways are presented below:
1. Establishment, function of a certain cross-section previously chosen, of all
possible solutions, determination for each of it of the allowable load and choosing
from all “sure” (for which the capable load is bigger than real load) of the one with
the smallest one. The procedure is very elaborated and represents an empirical way
for obtaining of some economic solutions.
2. The computation is based on choosing the most efficient solution from a variety
of solutions, using one of the post-elastic computation methods. Such a possibility
is given by the bending moments distribution in plastic domain, in its usual form or
in a generalized operation form with mechanical work measures. [1] Using of this
method supposes a certain experience in choosing the adjusting way of nodes, bars
and kinematic chains equilibration, respectively in choosing of some combinations
and constructive constraints which have to be taken in view.
3. Static methods, based on plastic yielding conditions (safety) corresponding to
critical cross-sections of the structure, as presented below:
− S p (i ) ≤ S i ≤ S p (i ) (1)
where: Sp(i) is the capable effort (plastic) of the critical cross-section “i”;
Si is the effective effort in the critical cross-section “i”.
296 C. Amariei, I.G. Mihai
In the case of bar structures subjected mostly in bending, relation (1) becomes:
− M p (i ) ≤ M i ≤ M p ( i ) (2)
Choosing a statically determined system and writing the bending moments Mi as:
n
− M i = M 0i + λ ⋅ M 0' i + ∑ xh ⋅ m0hi (3)
h =1
where: M 0i - bending moments on the base system, produced by permanent loads;

λ - loading coefficient;
M 0i' - bending moments in the base system by λ = 1 ;
x h - statically undetermined values;
m0hi - bending moments produced in the base system by each x h = 1 .
The limit state computation is reduced to a linear algorithm problem consisting in
increasing at maximum the loading factor λ in obtained relations by replacing the
() ( )
equations 3 and 2 . By utilizing bending moment diagrams in equilibrium
with external loads, disposed on base systems, judiciously choused, and which can
differ from a load case to another, the plastic yielding conditions become [7]:
− M p ≤ M p0 + ∑ xi ⋅ mi ≤ M p (4)
where ( for a critical cross-section “i” ):
M p - plastic moment of the cross-section;
M p0 - bending moment produced by several loads on a certain base system;
xi - proportionality coefficients (statically undetermined values);
mi - bending moments from auto equilibrium diagrams.
Using unknown factors decomposing or axes translations, are obtained resolvable
formulations by simple procedure, due to which is determined the minimum of
weight function.
4. The method based on using the elementary mechanisms combination and on the
minimum weight solution theory elaborated by J.Foulkes and B.G.Neal.
5. The method proposed by J.Heyman and W.Prager, based on the general
conditions of the limit state computation and on the minimum weight design
theories and following the static way, in a two cycles of solving, each cycle
consisting of two stages.
Generally, additionally to general hypothesis of the post elastic

computation domain, at minimum weight structures design are also taken into
consideration some supplementary hypothesis:
a) There exist an infinite variety of cross-sections (rolled steel shapes or
composed cross-sections);
b) It is known the variation law on the weight per unit length of elements (q),
due to plastic strength modulus (Wp). If there are graphically represented the
pairs of values of (q, Wp) for the usual :I” rolled steel shape, is obtained the
curve presented in figure 1 (approximated as being a continuous curve), which
can be expressed by an exponential relation as :
q = k ⋅ W pα (5)
or:
α
M 
q = k  p  = k ' ⋅ M αp (6)
 σc 
q(dan/cm
A C
3
Wp(cm )
Fig. 1
The coefficients k and α vary function of the cross-section type.

Since for a certain structure the used cross-sections don’t vary generally between
very high limits, the curve (q, Wp) can be approximated with a polygonal diagram.
For example, if the values of plastic module corresponding to points A and B from
the figure 1 have a 1:2 ratio, the error coming from the approximation of the curve
on the specified portion with a straight line, is only of 1%. Due to this assumption,
the weight per unit length of the element can be written as a linear function:
q = a +b⋅Mp (7 )
The total weight of the structure is:
Q = ∑ qi ⋅ li = ∑ (a + b ⋅ M p (i ) ) ⋅ li = a ⋅ ∑ li + b ⋅∑ M p (i ) ⋅ li
n n n n
(8)
i =1 i =1 i =1 i =1
where li represents the lengths of the bars, Mp(i) are the plastic bending moments of
the bars cross-sections, and the number „n” is the number of bars having different
values of plastic bending moments.
n
Because a ⋅ ∑ l and b are constant, results that the determination of the
i =1
i
minimum weight is reduced to determination of the minimum solution of the

equation:
n
X = ∑ M p (i ) ⋅ li (9)
i =1
called weight function.

Regarding the loads considered for minimum weight structures
computation, according to American norms, they will be considerate as follows:
• for the combination consisting of permanent and live loads, they will be
multiplied with a unic coefficient 1,7;
• for the combination consisting of permanent loads, live loads and wind action
or earthquake action, they will be multiplied with a unic coefficient 1,3.
2. COMPUTATION MODEL BASED ON INEQUALITIES

METHOD
Adaptation of the inequalities method for optimization computation
Inequalities method, usually used as determination way of loading limit factor can
be adapted for a structure weight optimization computation, representing some
important advantages, as simple and direct way of writing the constraining
relations (especially of plastic yielding conditions), and the fact that it can be taken
into consideration in the optimization computation of axial force influence, fact
that can influence a lot the conceiving and behavior of certain structures categories.
Adaptation of inequalities method for weight optimization computation requires
two important elements:
a) Taken in consideration as unknowns – in the relations that express plastic
yielding conditions – of the plastic moments and introducing in these relations
of the loading factor with imposed values by real loads acting on the structure.
b) Joining to these relations the weight function (linear or nonlinear) which has
to be optimized.
The relations which compose the mathematic model for optimum design in this
way are the following:
1. Static equilibrium equations:

a) for loaded bars ( figure 2.a ):
M s(k ) ⋅ bk + M c(k ) ⋅ l k + M d(k ) ⋅ a k = λk ⋅ a k ⋅ bk (10)
where: k=1,2,…b (b being the number of loaded bars over the length).
n
∑ λr
r =k
+ M s(k ) ()
+ M ck − M dk
() − M (jik )
ak bk
lk hij +θ1 +θ 2 +θ m+1
−θ1 −θ 2 −θ m+1
-θ1 -θ2
+θ1,2
1 2 ……... m
(a) (b)
n
∑ λr
r =k
− M (jik )
hij ()
hijk +θ
+θ
−θ −θ
1 2 ……... m
(c)
Fig. 2
b) for kinematic chains:

-unequal columns over the level ( figure 2.b ):
s  M (ji ) + M s(i ) 
∑ 

 = Hk
 (11)
i =1  hi 
where: i=1,2…s ( s being the number of the columns from a level );
k=1,2…m ( m being the number of levels of the structure );
hi=column height;
Hk=sliding load for a level.
- unequal columns over the level (figure 2.c ):
∑ (M ( ) + M ( ) ) = h
s
j
i
s
i
k ⋅ Hk (12)
i =1
hk level height.
c) for nodes:
∑ [± M ( ) ] = 0ij
k
(13)
k=1,2…n (n being the number of nodes).
2. Relations which express plastic yielding conditions:

− M p (i ) ≤ M i ≤ M p ( i ) (14)
where: Mp(i) is the plastic moment of the critical cross-section“i” (the unknowns of
the problem);
Mi is the effective bending moment in the critical cross-section “i”, i=1,2…c (c
being the number of critical cross-sections).
Totally are written a number of (e) statically equilibrium relations (e=b+m+n) and
a number of (2c) inequalities – plastic yielding conditions.
3. Weight function, which has the usual form:

p
X = ∑ li ⋅ M αp (i ) (15)
i =1
i=1,2…n ( n being the established number of different plastic bending moments of

the structure ).
In a matrix form, the relations can be written as:
- statically equilibrium relations:
[B] ⋅ {Μ} = {λ } (16)

- plastic yielding conditions:
− {M p } ≤ {Μ} ≤ {M p } (17 )
- weight function:
X = {C} ⋅ {M p } (18)
T
where:
[B] is the matrix of coefficients from the statically equilibrium relations;

{M } vector of bending moments in the critical cross-sections;
{λ} is the vector of constants from statically equilibrium;
{M } is the vector of plastic moments in critical cross-sections;
p
{C} is the vector of plastic moments coefficients from the weight equation.
Removing from the plastic yielding conditions a number of (e) bending
moments by their replacing with the values obtained in the (e) statically
equilibrium equations, function of the other (c-e) bending moments, will result (2c)
inequalities with [p+( c-e )] variables, [p necessary plastic bending moments and
(c-e) bending moments in the critical cross-sections, as:
[A] ⋅ {M } ≥ {λ} (19)
Some inequalities will be eliminated, obtaining a reduced number of constraining
conditions which, together with the weight function, compose the relations of
programming problem for minimum weight determination.
From the constraints number reducing problem, the most important are the
following:
a) Imposing – constructively taking – of some ratios between necessary plastic
moments, meaning:
Mp(i)><Mp(k) (20)
which will have as effect the decreasing of the unknowns number of the
optimization problem and also will eliminate some constraining relations referring
to plastic joints appearance possibility on each bar in nodes.
b) “Selection” of inequalities meaning eliminating the least restrictive relations
(which are satisfied including the remained inequalities).
c) Partially or totally knowing the shape of failure bending moments distribution
(based on static and loading schemes), which makes possible to write – for critical
cross-sections where is certainly known the sign of bending moment – only some
simple inequalities, instead of double ones which usually appear in plastic yielding
conditions:
− M p (i ) ≤ M i (21)
or:
M i ≤ M p (i ) (22)
In case of a linear weight function, simplex method can be used for

solving, when is necessary the transformation of constraints inequations in
equalities relations, by introducing of some compensation variables M , so the
matrix relation [19] becomes (eventually after elimination of some constraints
conditions):
[A] ⋅ {M } − [E ] ⋅ {M } = {λ} (23)

{ }
where: M is the vector of compensation variables;
[E ] is the unit matrix.
Or, can be used the extended form of the problem, by introducing of some auxiliary
variables M ' :
[A]⋅ {M }− [E ]⋅ {M }+ [E ]⋅ {M ' } = {λ} (24)

or:
[A] ⋅ {M } + [E ] ⋅ {M * } = {λ} (25)
where:
{M }= {M }− {M }
* '
(26)
in this case, is necessary to respect the negativity conditions for all three variables
categories:
M , M ,M ' ≥ 0 (27 )
and the extended weight function (the lower bound of the weight function X) is:
X * = {C} ⋅ {M p }+ {0} ⋅ {M i } + {µ } ⋅ M *
T T T
{ } (28)
{ }
where: M p is the vector of necessary plastic bending moments, which makes the
minimum weight structure;
{M i } is the vector of remain bending moments;
{M } is the vector of auxiliary compensation variables.
*
The simplex solution contains also the values of “p” necessary plastic bending
moments, and the values of remain bending moments in (c-e) critical cross-
sections; the other (e) values of bending moments are determined with statically
equilibrium equations, so it is possible a complete statically analyze for checking
the plastic yielding conditions fulfill, and also the failure mechanism establishment.
In case of some big dimensions, the problem will be solved with a linear or
nonlinear computer program.
3. COMPUTATION EXAMPLE
Determination of minimum weight solution for the frame in the figure 3.a.
4·F
2 3 4 θ 3 θ
2 4
Mp(2) δ3
e=L
e=L
θ=δ3/a
Mp(1)
Mp(1)
δ3=a·θ
1 5 1 5
a=1,5L b=1,5L a=1,5L b=1,5L
(a) δ2 (b)
4·F
2 4
e=L
θ θ=δ2/e
δ2=e·θ
1 5
(c)
Fig. 3
It is considered: F=1; L=1.

It will be noted: Mp(1)=Y1 şi Mp(2)=Y2 which are the necessary plastic bending
moments for the columns and beams.
Will be noted the bending moments in the five critical cross-sections (1,2,3,4,5) of
the frame, as: M3=Y3; M4=Y4; M5=Y5; M1=Y6; M2=Y7
The relations which compose the mathematical model of the minimum weight
frame design problem by inequalities method are the following:
1) Statically equilibrium relations:

a) bar equilibrium (figure 3.b):
M2·θ+ M3·θ+M3·θ+M4·θ=F·δ1
or: M2+M3+M3+M4=F·a
having the established notations:
Y7+Y3+Y3+Y4=F·1,5·L
or Y7+2·Y3+Y4=1,5
b) displacement equilibrium (figure 3.c ):
M1·θ+ M2·θ+M4·θ+M5·θ=4·F·δ2
or:M1+M2+M4+M5=4·F·e
having the established notations:
Y6+Y7+Y4+Y5=4·F·L
Y6+Y7+Y4+Y5=4
2) Plastic yielding conditions:

-Y1≤Y6≤Y1;-Y1≤Y7≤Y1; -Y2≤Y7≤Y2;-Y2≤Y3≤Y2;-Y2≤Y4≤Y2;-Y1≤Y4≤Y1; -
Y1≤Y5≤Y1
These relations can also be written:
Y1+Y6≥0; Y1+Y7≥0; Y2+Y7≥0; Y2+Y3≥0; Y2+Y4≥0; Y1+Y4≥0; Y1+Y5≥
respectively:
Y1-Y6≥0; Y1-Y7≥0; Y2-Y7≥0; Y2-Y3≥0; Y2-Y4≥0; Y1-Y4≥0; Y1-Y5≥0
3) Weight function:
n
X = ∑ li ⋅ M p (i )
1
so: X=2·Mp(1)+3·Mp(2)+0·M3+0·M4+0·M5+0·M1+0·M2
or: X=2·Y1+3·Y2+0·Y3+0·Y4+0·Y5+0·Y6+0·Y7
Using a usual computation program for solving linear problems, will be obtained
the following results:
-plastic bending moments values:
Y1=Mp(1)=1.625; Y2=Mp(2)=0.375
-valorile momentelor din secţiunile critice:
Y3=M3=0.375; Y4=M4=0.375; Y5=M5=1.625; Y6=M1=1.625; Y7=M2=0.375
-value of the weight function:
X=4.375
Knowing the values of plastic bending moments on the columns and beams will be
established the cross-section of bars:
a) the column:
M p (1)
M p (1) = σ c ⋅ Wcolumn so : Wcolumn =
σc
bcolumn ⋅ hcolumn
2
or: Wstalp =
4
Will be imposed: hcolumn = 1,5 ⋅ bcolumn
bcolumn ⋅ (1,5 ⋅ bcolumn )

2
M
and results : = p (1)
4 σc
4 ⋅ M p (1)
where bcolumn = 3
2,25 ⋅ σ c
4 ⋅ M p (1)
hstalp = 1,5 ⋅ 3
2,25 ⋅ σ c
b) the beam:
M p (2 )
M p (2 ) = σ c ⋅ Wbeam deci : Wbeam =
σc
bbeam ⋅ hbeam
2
or: Wbeam =
4
It will be imposed: hbeam = 1,5 ⋅ bbeam
bbeam ⋅ (1,5 ⋅ bbeam )

2
M
and results: = p (2 )
4 σc
4 ⋅ M p (2 )
where: bbeam = 3
2,25 ⋅ σ c
4 ⋅ M p (2 )
hbeam = 1,5 ⋅ 3
2,25 ⋅ σ c
References
1. Amariei C. – Calculul structurilor în domeniul plastic, I.P.Iaşi, 1974 (in Romanian)
2. Bălan Şt, Petcu V. – Calculul structurilor în domeniul plastic – Optimizări, Editura Academiei
R.S.R., Bucureşti, 1979 (in Romanian)
3. Dancea I. – Metode de optimizare, Editura Dacia, Cluj Napoca, 1976 (in Romanian)
4. Gheorghiu Al. – Concepţii moderne în calculul structurilor, Editura Tehnică, Bucureşti, 1975 (in
Romanian)
5. Mihăilă N. – Introducere în programarea liniară, Editura Didactică şi Pedagogică, Bucureşti, 1980
(in Romanian)
6. Nădejde I. – Probleme de cercetare operaţională. Programare matematică, Editura Academiei
R.S.R., Bucureşti, 1971 (in Romanian)
7. Răutu S., Chiroiu V. – În legătură cu calculul construcţiilor cu greutate minimă (in Romanian)
8. Vuc I. – Modele matematice în proiectarea construcţiilor, Editura Facla, Timişoara, 1981 (in
Romanian)
Scientific Perspectives for Future Research Work in Fundamental

Properties of Short-Lived Radionuclides from Decay and from
in-Beam Studies
Brindusa Ciobanu1, Ion Silisteanu2 and Irina Radinschi 3
1
2
“Horia Hulubei” Natl. Inst. of Physics & Nuclear Engrg., 077125, Bucharest-Magurele, Romania
3
Summary
The aim of this paper is to point out some scientific perspectives for future research
work in fundamental properties of short-lived radionuclides from decay and from
in-beam studies, arisen in the frame of our research project. The project
approaches one of the most important themes of nuclear physics. In recent years
the investigation of unstable short-lived nuclei has gained world-wide interest as
well on the experimental as on the theoretical side. These nuclei are characterized
by unique structure properties: the weak binding of the outermost nucleons and the
effects of the coupling between bound states and the particle continuum.
Particularly, the modification of the effective nuclear potential leads to the
formation of the nuclei with very diffuse nucleon densities, to the occurrence of the
nucleon skin and halo structures. Our first goal is to formulate a reaction theory
for cluster decay in which the clustering and reaction amplitudes will be
represented by means of resonance formulas. A brief outlook of the experimental
results and theoretical ideas which define the field is presented toghether with
quantitative estimations for the resonance decay widths and relative intensities.
The resonance solution of systems of coupled equations is obtained by a direct
numerical integration using step-by-step methods on computer. Thus, we obtain
most complete information on the development of typical structures from stable to
exotic losely-bound nuclei by: improving the structure models in order to describe
essential features and to obtain spectroscopic information on a number of nuclei
that can then tested against data extracted from decay and in-beam studies;
extending the range of applicability of reaction models by using accurate reaction
channels methods (e.g. via additional channels including deformation, exchange
effects, antysimmetrisation, etc); including microscopic structure information in
coupled channel reaction models and treating more carefully of "intermediate"
systems that are more or less bound or have mixed composition.
KEYWORDS: short-lived radionuclides, cluster decay, numerical integration of

coupled channel equations, computational methods.
308 B. Ciobanu, I. Silisteanu, I. Radinschi
1. INTRODUCTION
In recent years the investigation of unstable short-lived nuclei has gained world-
wide interest as well on the experimental as on the theoretical side. These nuclei
are characterized by unique structure properties: the weak binding of the outermost
nucleons and the effects of the coupling between bound states and the particle
continuum. Particularly, the modification of the effective nuclear potential leads to
the formation of the nuclei with very diffuse nucleon densities, to the occurrence of
the nucleon skin and halo structures. These phenomena will also affect collective
vibrations and rotations of unstable nuclei, particularly the electric dipole and
quadrupole excitations, and new modes of excitations and decay might arise in
nuclei near the drip line.
Apart from general structure aspects in the physics of nuclei far from stability,
there is also strong relations to the question of clustering in nuclei and to few-body
physics. Furthermore, there will be a strong coupling to the continuum both in
static and dynamic descriptions. Nuclear clustering and fine structure are two of the
most important typical structures in nuclear physics which are strongly connected
to the stability and deformation.
Superheavy elements (SHE) probe the extremes of nuclear structure with respect to
the number of nucleons that can form a bound system. Their existence and decay
properties are one of the most fundamental problems in nuclear physics [1-4].
There are data that confirm the existence of Z=111 and 112 and their connection to
lighter decay chains [1]. The first data for Z=113,114, 115 and 116 also exist [3, 4],
with suggested A=288-292, respectively, but the A values are not certain since the
connection to lighter nuclei is not known. A new element with atomic number 118
was synthesized for the first time in the 249Cf + 48Ca reaction.
Atomic and mass numbers of the isotope of element 118 were determined from the
measured excitation functions and decay characteristics of the daughter nuclei
produced in cross-bombardments [5]. Theoretical models for SHE have evaluated
from microscopic-macroscopic models [6] to fully microscopic deformed Hartree-
Fock (HF) models [7]. In addition to their importance for many-body nuclear
structure, theoretical models for the prediction of the decay properties of the SHE
are important when designing experiments since the technique used will depend on
half-life and decay mode.
In order to obtain the most complete information on these new nuclei, effects and
phenomena a use of radioactive beams in a wide range of isospin and energy is
necessary. High energy beams produced byin- flight fragmentation and post-
accelerated beams proved to be complementary with the respect to their intensity,
purity, and optical quality. When one analyses a proper reaction model, the
measurements of core or halo momentum distributions after breakup, can provide
important information on the structure of a such nucleus. The detection of gamma-
rays from outgoing fragments allows to access experimentally whether core excited
components are present in a halo state.
The experimental signatures of clustering and "fine structure" (FS) far from
stability are supported by data from selective excitation in nucleon and α-transfer
reactions, α-decay, heavy cluster decay, and by rotationally spaced energy levels,
enhanced transition strengths and intensities, and appreciable emission width for
the resonant states above the decay threshold. Recent decay and in-beam studies of
unstable nuclei are revealing an unexpected diversity and richness of shapes, and
new decay modes and typical structures which are related to very low binding
energy and the strong influence of the continuum.
To treat the decay rates for these nuclei we describe the nuclear clustering and
penetration of the particle through the potential barrier taking into account the
correspondence between different decay channels and different final states of the
daughter nucleus.
The decay process being governed by nuclear clustering and quantum tunneling
over the Coulomb barrier is strongly influenced by couplings of the relative motion
of the fragments to several nuclear intrinsic motions. We address to effects of
coupling between the intrinsic degrees of freedom and relative motion by solving
numerically the coupled channel equations, including all the relevant reaction
channels.
Our first goal is to formulate a reaction theory for cluster decay in which the
clustering and reaction amplitudes will be represented by means of resonance
formulas. A brief outlook of the experimental results and theoretical ideas which
define the field is presented together with quantities estimations for the resonance
decay widths and relative intensities. The resonance solution of systems of coupled
equations is obtained by a direct numerical integration using step-by-step methods
on computer.
Thus, we obtain most complete information on the development of typical
structures from stable to exotic loosely-bound nuclei by: improving the structure
models in order to describe essential features and obtain spectroscopic information
on a number of nuclei that can tested against data extracted from decay and in-
beam studies; extending the range of applicability of reaction models by using
accurate reaction channels methods (e.g. via additional channels including
deformation, exchange effects, anti-symmetrization, etc); including microscopic
structure information in coupled channel reaction models and treating more
carefully of "intermediate" systems that are more or less bound or have mixed
composition.
The modern facilities (EUROBALL, GAMMASPHERE, RIA, etc.) or recent
research projects (SPIRAL-2, etc.) of the important international centers in Oak
Ridge (USA), Darmstadt (Germany), Dubna (Russia), etc., are mainly devoted to
the laboratory investigation of exotic nuclear structures [6].
The very large number of publications and conferences in this field underlines its
importance and define it as an excellence research area. The research team has a
long-term experience in the research fields approached by the project, as:
- short lived nuclear production with secondary beams;
- spectroscopy of exotic nuclei by transfer and breakup reactions. One notes
experimental methods like resonant particle spectroscopy (RPS) method [8],
recoil decay tagging (RDT) method [9], cross-sections measurements in
breakup reactions and theoretical methods (highly unstable states
spectroscopy, semiclassical models for the evaluation of one-nucleon breakup
reactions at intermediate energies, calculation and design for detection
methods, RPS and RDT);
- theoretical modelling of spontaneous decay, transfer and break-up reactions
(clustering and “fine structure” at limits of stability, coupled channel
approximation for cluster decay including rotational degrees of freedom,
quantum tunneling induced by rotational coupling, multichannel resonant
decay).
In the following we will shortly present the current status and directions within the
topics related to our project.
2. RESULTS AND PERSPECTIVES
2.1. Short lived nuclei production with secondary beams
One tries to estimate the most favourable projectile - target combination to

maximize the fusion cross section at energies near the Coulomb barrier, taking into
account the deformation degrees of freedom and mass assymetry.
These thematics investigate the orientation effect by considering the role of
quadrupolar and hexadecapolar deformations for the projectile and the role of
rotational degrees of freedom for the target. In this propose one developes coupled
channel approximation for the Schrödinger equation, treating the cold fusion as a
scattering problem with appropriate boundery conditions.
Both Numerov and Runge-Kutta numerical methods will be use for the computing
of the radial wave functions. One establishes the succesfull application to the
description of fusion of 48Ca on actinide nuclei 238U and 248Cm, recently performed
at the Flerov Laboratory Dubna. It was estimated that this projectile-target
combination is one of the most promising one to reach the superheavy island of
stability [1,2,10].
2.2. Spectroscopy of decay, transfer and breakup reactions
One attempts to measure one nucleon breakup cross sections for proton and
neutron rich nuclei. One estimates breakup cross sections for weakly bound exotic
nuclei. One performes the design and simulation of complex experimental setup for
reactions with secondary beams.
Exotic nuclei (far from stability valley) play an important role in the structure
studies, as well in the modern astrophysical studies (nucleosynthesis, structure an
evolution of neutron stars, black holes, etc.)
One of the most sophisticated and succesful experimental methods has been
proposed by Romania (C. Borcea et al.) in which a measurement of total reaction
cross section and the inclusive breakup one nucleon cross section for 8B and 11Be
have been proposed for the definite establishment of the existence of halo states
[11]. In these experiments, a secondary beam of (e.g.) 11Be has been produced by
fragmentation of a primary beam of 18O at 70 MeV/nucleon on a thick target of
9
Be. Fragmentation products have been separated with the doubly achromatic
LISE separator in GANIL (France). An approximately 1000 11Be counts/sec has
been obtained in impinged on a stack of 17 Silicon detectors which served as a
degrader, target and analysing media for reaction products. In front of the telescope
a position sensitive detector x-y (position sensitive parallel plate avalanche
detector-PPAD ) has been placed to define the beam spot on the detector. The
telescope segmentation allowed identification of the reaction products as well as
the identification of reaction mechanisms such as stripping or nuclear dissociation.
The first mechanism has been identified by an increased rate of reaction events in
adiacent segments of the telescope, following an reaction event in a given segment.
Later on in a subsequent experiment at Michigan the nuclear dissociation
mechanism has been identified by a supplementary energy loss deposition by the
proton in the breakup reaction of 8B on a silicon target. The method is based on the
assumption that to a good approximation after breakup the heavy fragment will
have 10/11 or 7/8 from the incident energy. This is based on the common
understanding of halo states (i.e. low separation energy and low angular
momentum characterising the state). This small change in the energy translates into
a different range in the detector which can be measured. Backwards to the
telescope an array or 32 neutron detectors has been placed in an attempt to better
identify reaction mechanism such as nuclear dissociation.
A similar experiment has been performed recently in a Oberlin -Michigan -
Bucharest collaboration. A cocktail of proton rich nucleai has been impinged on a
silicon telescope. The experiment allowed identification of new candidates for
proton halos, weakly bound nuclei, close to the proton drip line, which shows a
very low separation energy and a wave function dominated by low angular
momenta (p-waves). This work is in progress. These exeperments have a very nice
future. New plannigs are projected at GANIL for studies of neutron rich nuclei in
the f-shell with the improvements that core states will be identified by an array of
efficient gamma detetectors [12-16].
2.3. Spontaneous decay, transfer and breakup theoretical modelling
The experimental results and theoretical ideas which define the clustering and "fine
structure" is presented together with quantitative estimations for the resonance
emission widths. We present both formal considerations, derived from a
microscopic (shell model) formulation of the reaction theory, and practical
computational programs based on coupled channel methods, with many-body
effects included in formation and reaction amplitudes, energy shifts and total
decay widths.
The main conceptual issues in reaction theory are identified and their relative
importance is assessed: these include, couplings between the relative motion of the
fragments and several nuclear collective motions, the existence and convergence of
the resonance scattering solution, account the effects of non-linear couplings to all
orders, self-consistence of the scattering potential, and the rotational excitation of
nuclei by the cluster transfer.
The cluster decay properties of heavy drip-line nuclei are considered and some
spectroscopic information on the continuum states populated in the unbound
intermediate systems in the decay channel is obtained. The limitations and
advantages of current computational approaches are addressed, with particular
regard to quantitative experimental comparisons for superheavy nuclei [1- 4], [17,
18].
2.4. Numerical computer codes for nuclear reactions with exotic nuclei
The aim is to obtain with good accuracy the eigenstates (bound and resonant) of
deformed nuclei for realistic studies of phenomena like fission, alpha decay and
proton emission.
The motion of a particle in a deformed potential is governed by the tri-dimensional
Schrödinger equation. When the potential is axially symmetric, the wavefunction
can be expanded in partial waves with angular and radial components. The radial
wavefunctions are the solutions of a set of coupled differential equations of second
order. In order to obtain the eigenstates (either bound, resonant or antiresonant) one
has to solve (numerically) the eigenvalue problem for this system subject to some
physically consistent boundary conditions. For this purpose several tasks should be
accomplished: the calculation of the solution at the boundaries of the numerical
domain (near the origin and at large distances), the propagation of the solution on
the integration interval, the determination of the eigenvalues and the construction
of the normalized wave functions. Many investigations have been devoted to this
subject. See for example [19-24], including our own experience [4,17,18,25]. In the
present project one intendes to improve the existing procedures, increasing the
accuracy and reducing the amount of computer time and memory requirements.
Thus realistic cases with important deformations and large number of equations
could be handled, allowing the study of nuclear processes in complex systems (like
exotic and heavy nuclei).
2.5. Role of charge density in cold fusion reactions
Cold fusion barriers are studied with respect to the change of the charge density
within the overlapping region. Charge evolution from separated target and
projectile up to the compound nucleus is taken into account by meaning of a
deduced transition formula which depends on geometric parameter variation
defining the shape. Macroscopic shell correction and total deformation energy for
fusion like configurations are calculated for different charge density paths.
The way geometric parameters influence on the total deformation energy is also
studied and compared with already published work [26, 27]. Influence of changes
on isotopic composition of the reaction is emphasized in final shell corrections, and
an analogy is made with isospin equilibration observed and calculated in [28].
Total barriers are computed as a function of charge density taken as free parameter,
a procedure also accounted in [29]. Minimization along this coordinate produces
variations of about 4 MeV for light nuclei and up to 8 MeV for superheavy
synthesis, for the deformation energy in the last part of the process.
4. CONCLUSIONS
The aim of this paper is to point out some scientific perspectives for future research
work in fundamental properties of short-lived radionuclides from decay and from
in-beam studies, arisen in the frame of our research project. Our first goal is to
formulate a reaction theory for cluster decay in which the clustering and reaction
amplitudes will be represented by means of resonance formulas. A brief outlook of
the experimental results and theoretical ideas which define the field is presented
toghether with quantitative estimations for the resonance decay widths and relative
intensities. The resonance solution of systems of coupled equations is obtained by a
direct numerical integration using step-by-step methods on computer.
The degree of novelty is given by the field of the proposal itself: exotic nuclear
systems are intensively investigated by different experimental and theoretical
groups in all important international centers during the last decade. Important
experimental and computing facilities were assigned to this field in this period and
many conferences and workshops were organized in order to analyse the obtained
results. The complexity is given by sophysticated experimental techniques and

theoretical methods.
The participants are well recognized scientists in several related to this field areas.
This project is a good opportunity to enlarge the international cooperation with our
foreign partners and to prepair the participation within the framework of various
international projects.
Acknowledgements
Research of authors was supported by MedC- CERES in the framework of the

grant CEX05-D10-08/03.10.2005.
References
1. Yu. Ts. Oganessian, et al., Phys. Rev. Lett, 83, 1999, pp. 3154.
2. Yu. Ts. Oganessian, et al., Phys. Rev., C70, 2004, pp. 242.
3. S. Hofmann and G. Münzenberg, Rev. Mod. Phys., 72, 2000, pp.733.
4. I. Silişteanu, W. Scheid, A. Sandulescu, Nucl. Phys. A, 679, 2001, pp. 317.
5. Yu. Ts. Oganessian, et al., Synthesis of the isotopes of elements 118 and 116 in the 249Cf and
245
Cm+ 48Ca fusion reactions, C 74, 2006, 044602.
6. G. Münzenberg, Perspectives for nuclear structure research at GSI- from halo nuclei to
superheavy elements, Proceedings of the 174. WE- Heraeus- Seminar, Il Nuovo Cimento, Vol.
110 A, N. 9- 10, 1997, pp. 1103.
7. S. G. Nilsson, I. Ragnarsson, Shapes and Shells in Nuclear Structure, Cambridge University
Press, ISBN-13 978-0-521-01966-8, ISBN-10 0-521-01966-4, 2005, pp. 50, 72, 157.
8. D. Robson, Nucl. Phys., A 204, 1973, pp. 204.
9. R. S. Simon et al., Z. Phys., A 325, 1986, pp. 197.
10. Y. S. Tsyganov, Detection of ultra rare α decays of super heavy nuclei, Nuclear Instruments
and Methods in Physics Research, A 573, 2007, pp. 161-164.
11. N. K. Skobelev, R. Anne, C. Borcea, D. Guillemaud - Mueller, S. Grevy, Z. Dlouhy, F. Carstoiu,
M. Lewitowicz, S. M. Lukyanov, A. C. Mueller, F. Negoita, Yu. E. Penionzhkevich, M. G.
Saint - Laurent, O. Sorlin, Reactions of breakup of 11Be Nuclei interacting with 28Si Nuclei,
Bull. Rus. Acad. Sci. Phys., 63, 1999, pp. 771.
12. A. De Vismes, P. Roussel-Chomaz, F. Carstoiu, Global analysis of proton nucleus cross sections,
Phys. Rev., C 62, 2000, 064612.
13. E. Sauvan, F. Carstoiu, N. A. Orr, J. C. Angelique, W. N. Catford, N. M. Clarke,
M. Mac Cormick, N. Curtis, M. Freer, S. Grevy, C. Le Brun, M. Lewitowicz, E.Liegard,
F. M. Marques, M. Mac Cormick, P. Roussel-Chomaz, M. G. Saint Laurent, One-neutron
removal reactions on neutron-rich psd-shell nuclei, Phys. Rev., C 69, 2004, 044603.
14. E. Sauvan et al., Phys. Rev., C 69, 2004.
15. L. Trache et al., Phys. Rev. Lett., 271102, 2001, pp. 87.
16. L. Trache et al, Phys. Rev., C 69, 2004, 032802(R).
17. I. Silisteanu, A. O. Silisteanu, W. Scheid, B. I. Ciobanu, Alpha Half-Time Estimates for the
Superheavy Elements, Proceedings of the Carpathian Summer School of Physics 2005,
Exotic Nuclei and Nuclear/ Particle Astrophysics, World Scientific, 2006, pp. 423.
18. I. Silisteanu, A. Sandru, A. O. Silisteanu, B. Popovici, A. Neacsu, B. I. Ciobanu, Alpha Half-Time
Estimates for the Superheavy Elements, Proceedings of the Predeal International Summer School
in Nuclear Physics, Collective Motion and Phase Transition in Nuclear Systems, World Scientific,
2006, pp. 569.

19. S. Cwiok, J. Dudek, W. Nazarewicz, J. Skalski, T. Werner, Single particle energies, wave
functions, quadrupole moments and g-factors in axially deformed Woods - Saxon potential with
applications in the two - centre - type nuclear problems, Comp.Phys.Commun., vol. 46, 1987,
pp.379.
20. E. Maglione, L. S. Ferriera, R. J. Liotta, Nucleon decay from deformed nuclei, Phys.Rev.Letters,
vol. 81, 1998, pp. 538.
21. K. Rykaczewski, J. C. Batchelder, C. R. Bingham, T. Davinson, T. N. Ginter, C. J. Gross,
R. Grzymack, M. Karny, B. D. MacDonald, J. Mas, J. W. McConnel, A. Piechaczek, R.C. Slinger,
K.S. Toth, W. W. Walters, P. J. Woods, E. F. Zganjar, B. Barmore, L.Gr. Ixaru, A.T. Kruppa, W.
Nazarewicz, M. Rizea, T. Vertse, Proton Emitters: Probing the Structure of Unbound Nilsson
Orbitals, Phys.Rev.,C, vol. 60, 1999, 011301.
22. T. Vertse, A. T. Kruppa, B. Barmore, W. Nazarewicz, L. Gr. Ixaru, M. Rizea, Proton emission
from Gamow resonance, AIP Conference Proceedings 518, Proton-Emitting Nuclei, First
International Symposium, Oak Ridge, Tennessee U.S.A., 2000, pp. 184.
23. H. Esbensen, C. Davids, Coupled-channel treatment of deformed proton emitters, Phys.Rev.,C,
vol. 63, 2000, 014315.
24. M. Rizea, PERSYS - A program for the solution near the origin of coupled channel Schrödinger
equation with singular potential, Comput. Phys. Commun., vol.143, 2002, pp. 83.
25. I. Silişteanu, W. Scheid, A. O. Silisteanu, Rom. Rep. Phys.,57, 2005, pp. 4.
26. C. Y. Wong, Phys. Rev. Lett., 31, 1973, pp. 766.
27. L. C. Vaz, J. M. Alexander, Phys. Rev., C 18, 1978, pp. 2152.
28. B-A. Li, S. J. Yennelo, Phys. Rev., C52, 1995, R1746.
29. K. Satou, H. Ikezoe, S. Mitsuoka, K. Nishio, S. C. Jeong, Phys. Rev., C 65, 2002,054602.
Experimental and numerical analysis of compressed concrete

elements confined with FRP composites
Gabriel Oprisan, Nicolae Taranu and Vlad Munteanu
Civil Engineering Department, “Gh. Asachi” Technical University, Iasi, 700054, Romania
Summary
Confinement of concrete by suitable selection of transverse internal reinforcement

or by externally bonded reinforcement results in significant increase in both the
compressive strength and the ductility of concrete subjected to compressive
loading. Traditional confinement solutions have been developed using steel hoops
or steel jackets and only recently fibre reinforced polymer composites (FRP) have
been perceived as reliable confinement solutions for concrete elements. Extensive
research projects have been performed since 1990s and theoretical and
experimental results confirm the validity of confinement with FRP composites
jackets. An experimental program has been initiated at the Faculty of Civil
Engineering, the Technical University of Iasi, to evaluate the confining effect with
glass fibre/epoxy and carbon fibre/epoxy composites. The influence of the material
type and the thickness of FRP confining jacket were the variables involved.
The results obtained have proven the effectiveness of confinement solutions based
on FRP composites. Both compressive strength and ductility of the confined
specimens have shown dramatic increase compared to unconfined concrete
specimens. The experimental set-up, the testing procedure and the main results are
presented, emphasizing the influence of the composite nature and the thickness of
the confining jacket.
KEYWORDS: confining, FRP jackets, compressive strength, ductility
1. INTRODUCTION
To establish the mechanical characteristics of unconfined concrete, cylindrical

specimen and standard cubes adequate to compression test were cast. Also in order
to determine the concrete class and the compressive strength, 9 concrete cylinders,
100 mm in diameter and 250 mm high, and 9 cubes, 100 mm in side, have been
tested in uniaxial compression under standard condition. To avoid a rapid failure of
the unconfined concrete cylinder and determine the complete stress-strain curve
including the post peak strength domain of the material it was necessary to use a
special installation for the post elastic testing of brittle materials. The arrangement
of all transducers is illustrated in figure 1, for both unconfined and confined

samples.
a. b.
Figure 1. Instrumentation of test samples with LVDTs a. unconfined concrete cylinder
specimen; b. confined concrete cylinder specimen
The compressive strength determined on plain concrete samples, after 28 days is

31.64 N/mm2 on the concrete cylinders and 32.16 N/mm2 on the concrete cubes,
and a complete stress-strain curve for unconfined concrete is illustrated in Figure 2.
2. EXPERIMENTAL PROGRAM
To avoid a rapid failure of the unconfined concrete cylinder and determine the
complete stress-strain curve including the post-yielding domain of the material it
was necessary to use a genuine installation for the post elastic testing of brittle
materials. Before the experimental testing of the specimen, the testing machine had
been calibrated and the equipment had been prepared for data acquisition, fig. 2.
Data processing was done with a developed soft based on “Test Point” program.
Rigid plate “Test Point” software
3T1
Specimen
Concrete
Data acquisition
T2 Strain indicators interface
32 channels
3T3
Load cell
Rigid plate Printer
Figure 2. Acquisition and processing experimental data
Stress-strain curve obtained on unconfined concrete cylinders (fig.3) has two

branches, a rising one associated with elastic, viscous and plastic strains, and a
318 G. Oprisan, N. Taranu, Vl. Munteanu
falling one corresponding to pseudo plastic strain and characterized by an

accelerated decrease in load and increase in strain over time. The maximum value
on the curve corresponds to the ultimate compressive strength of concrete. The
compressive strength determined on plain concrete samples, after 28 days was
31.64 N/mm2 on the concrete cylinders and 32.16 N/mm2 on the concrete cubes.
Figure 3. The complete stress-strain curves for unconfined concrete specimens
Nine specimens with 2, 3 and 4 unidirectional CFRP and GFRP layers (three of
each type) with main fibers orientated in the hoop direction have been prepared
using a wet hand lay-up technique. Prior to the application of the FRP layers the
surface of the samples has been properly prepared to provide a hard, dry and clean
surface. The main characteristics of GFRP for epoxy resins are presented in table 1
whilst the properties for glass and carbon fibers are given in table 2 and table 3.
Table 1 - Properties of epoxy resin

Appearance Component A: light/yellow to amber
Component B: pale yellow to clear
liquid
Mixing ratio A:B = 100:34,5
Density 1,16 g/cm3
Tensile strength (ASTM D 638) 7 days at +21 0C: 55N/mm2
Tensile modulus (ASTM D 638) 7 days at +21 0C: 2000N/mm2
Fracture elongation (ASTM D 638) 7 days at +21 0C: 3,2%
Confined concrete specimens with GFRP and CFRP have been tested with the
same installation used for the unconfined concrete samples. The confined
specimens have been kept under laboratory conditions kept for 7 days to enable the
complete cure of the polymeric resin after confinement.
Table 2 - Properties of glass fiber reinforcement

Fiber type E-Glass fibers
Fiber orientation 00C (unidirectional)
Construction 93% warp, 7% weft
Areal weight 920 g/m2
Fabric thickness 0,36 mm
Density of glass fibers 2,58 g/cm3
Tensile strength of fibers 2250 N/mm2
Tensile E-modulus 72400 N/mm2
Ultimate fiber strain 3,7 %
Table 3 - Properties of carbon fiber reinforcement

Fiber type Carbon fibers - high modulus
Fiber orientation 00C (unidirectional)
Construction 97% warp, 3% weft
Areal weight, [g/m2 ] 610
Fabric thickness, [mm] 0,34
Tensile strength of fibers, [N/mm2] 3900 (nominal)
3700 (minimal)
Tensile E-modulus, [N/mm2] 231000
Ultimate fiber strain, [%] 1,5
Three stages have been observed on the stress-strain curve of confined concrete
specimens:
• the first stage is linear and it corresponds to the stiffness of the unconfined
concrete specimens, fig.4;
• in the second stage the concrete specimen exhibits larger lateral strains and
the GFRP jacket creates a confining pressure on the concrete core. When the
concrete specimen is cracked, a dramatic change in curve configuration
occurs. At this stage much higher stresses and strains are attained than in the
case of unconfined concrete;
• in the third stage, the concrete specimen is thoroughly cracked and the GFRP
maintain the concrete confined specimen intact till the explosive failure of the
jacket occurs. The stress-strain curve increases by a constant angle up to the
rupture, fig.4.
A comparative set of curves for unconfined and CFRP confined samples is
illustrated in fig.5; the confined samples have been wrapped with 2, 3 and 4 layers
of CFRP composites.
The stress-strain curves of the CFRP composites have a bilinear shape with sharp
softening in the transition area around the strength of the unconfined concrete. In
the first stage the slope of the stress-strain curve is similar to that of the unconfined
concrete. In the second stage the concrete is cracked and the confinement is
activated. The stress of the confined concrete linearly increases with increasing the
CFRP strain. All experimental results confirm the similar work published in the
field of structural rehabilitation with composites [1].
Figure 4 - Stress-strain curves of compressed unconfined and confined concrete specimens

with GFRP
Figure 5 - Stress-strain curves of compressed unconfined and confined concrete specimens

with CFRP
When the test is properly carried out the maximum stress is reached at the CFRP
rupture. Some samples failed prematurely due to separation of the composite layers
at the lap joints. A step forward of the experimental program has been the cyclic
loading of the confined samples as a first stage to seismic retrofit with polymeric
composites. Some preliminary results are shown in fig.6 [2].
Figure 6 - Stress-strain curves of confined specimens subjected to cyclic load
3. NUMERICAL ANALYSIS
In order to verify the suitability of a FEM based software (LUSAS) for modelling
concrete confinement problems, a numerical analysis was performed.
A statical non-linear analysis was conducted on a model that was built under the
same dimensional conditions with the experimental program.
The concrete cylinder was modelled as a 3D solid element. In order to obtain a
radial regular meshing it was first modelled a quarter of the final concrete element.
The CFRP jacket was modelled as a sum of thick shell elements, “glued” onto the
concrete surface using the SLIDELINE option, proper for contact problems, as in
this case. The attributed thickness of the jacket was equivalent to a 3 layers
disposal of CFRP sheets.
3D solid elements were used for the concrete cylinder meshing, as long as for the
jacket, 3D surface elements were used. Although the software allows the use of
higher order finite elements, only linear interpolation between nodes option was
used in order to reduce the total number of elements and nodes for having a
reasonable solving time.
The cylinder was considered as simply supported at the top end and totally
restrained at the bottom end. The load was applied as a uniformly distributed force
at the top end of the concrete element. In order to ensure a good convergence of the
load increments, an initial value of 1 N/mm2 was used and variable load increments
(3 and 10) were used afterwards in the Solution Software Manager.
Figure 7 Support and load conditions of the concrete cylinder; radial meshing of the
concrete volume
A Ducker-Prager material model based was used for defining the material
properties for concrete as long as for the composite CFRP jacket an orthotropic
material model available in the software’s material options was used.
The mechanical properties for both concrete and composite jacket were defined in
accordance with the materials that were used in the laboratory tests, and then
attributed to the geometric elements.
From the large range of numerical and graphical results available after performing
the run of the program, representative data charts and maps were selected and
presented below. They are related to stresses, strains, displacements data needed in
order to realise a good picture of the confined concrete cylinder behaviour under
axial loading.
Figure 8. Normal stresses (unconfined Figure 9. Normal stresses (CFRP confined

concrete cylinder) concrete cylinder)
Figure 10. Longitudinal displacements Figure 11. Longitudinal displacements

(unconfined concrete cylinder) (CFRP confined concrete cylinder)
Figure 12. Transversal displacements Figure 13. Transversal displacements

(unconfined concrete cylinder) (CFRP confined concrete cylinder)
Figure 14. Longitudinal stress resultants in Figure 15. Transversal stress resultants in
the CFRP jacket the CFRP jacket
Characteristic stress-strain curves for both confined and unconfined concrete were
obtained; they are very similar to those obtained from the experimental analysis.
Figure 16. Stress-strain characteristic curves (CFRP confined / unconfined cylinder)

Figure 17. Stress-strain characteristic curves (GFRP confined / unconfined cylinder)
4. CONCLUSIONS
The use of FRP confinement significantly improves compressive strength and

ductility of concrete. It should be noted that epoxy resins adhere well to the
concrete surface, being also an excellent adhesive. Strengthening solutions with
FRP used for reinforced concrete columns prove to be very good for several
reasons. On the one hand, the strength and ductility values obtained are favourable
and on the other hand, the weight of the consolidated system and the cost related to
execution time and work are considerably reduced. Experience has shown that the
failure modes of FRP are different and they depend on a series of factors which can
influence the results obtained, such as: the characteristics of the concrete to be
confined, casting moulds and vibration methods, fibre volume fraction,
confinement techniques and accurate application by qualified workers.
References
1. Teng, J.G., Chen, J.F., Smith, S.T., Lam, L., FRP strengthened RC structures, John Wiley,
Chicester, 2002.
2. Oprisan, G. Soluţii moderne de consolidare a structurilor penrtu construcţii industriale, Teză de
doctorat - U.T. Iasi, (2002). (in Romanian)
Rayleigh models coefficients

Florin Ţepeş Onea
Universitatea “Ovidius” Constanţa
Summary
The propose of this work is to study the effect of the superior vibration modes on
the energy dissipation in the backlash between two plots of a idealized concrete
dam.
The intern energy dissipation inside the structural material is generating material
internal variations or hysterical damping. For viscous damping this energy is
frequency balanced. The Rayleigh damping could be considered as a linear
combination between structural masses matrix and stiffness matrix. The damping
matrix is obtained from the Cauchy sequence:
p −1
C = M ∑ ak ( M −1 K ) k (1)
k =0
Where the coefficients ak k=1,2..p are obtained from p simultaneous equations:

1  a0 
ξi =  + a1ωi + a2ωi 2 + ... + a p −12 p −3  (2)
2  ωi 
For p=2:
C =αM + βK (3)
Where α and β are constants that can be obtained from two dumping ratios of two
different frequencies.
This study is made for an idealized symmetric concrete dam
It was used two calculus models for the dam-foundation ensemble; a plain one and
spatial model, with simultaneous calculation.
KEYWORDS: Rayleigh, spatial mesh, excitation, dumping

1. STRUCTURAL ANALYSYS
The plane finite element mesh is made by 80 quadrilateral elements for the
foundation and 56 elements for the dam. The elasticity modulus for the dam was
pick Eb = 300000daN / cm 2 and for the foundation E f = 150000daN / cm 2 , with
the dam’s high of 30m and the slope λ1 = λ = 0,5 .
The dam is made by two plots 15m width each other separate for a backlash of
1mm. The two plots adjacent nodes, corresponding with the space, have the same
quota on x and z axis. This nodes can be connected with the help of springs in
order to model the friction.
We can notice that in the case of plane mesh the fundamental vibration mode is
flexural on upstream-downstream, the second mode is flexural too but on the high
of the dam, the others modes are of torsion.
For the spatial mesh the first vibration mode implies a symmetrical displacement
and flexural on upstream- downstream direction of the two plots, and for the
second mode a antisymmetrical displacement. To start with the 6’s vibration mode
it appear also a rotation of the two plots, imples a relative moving of the plots
surfaces in the spae between them. The propose of this work is to study the effect
of the superior vibration modes on the energy dissipation in the backlash between
the two plots.
Because both masses and stiffness matrix are orthogonal, damping matrix is
orthogonal too. From orthogonal condition we obtain:
φiT (α M + β K )φ j = 2ω (4)
Where φi , φ j are eigenvectors, ωi is circular frequency, ξi fraction of critical

dumping. The equation become:
α + βωi 2 = 2ωiξi
For determine the α and β coefficients influence, it was made a parametric study
for plane and spatial dam- foundation discrete mesh. Critical damp fraction was
taken as constant ξ = 0, 05 for whole vibration modes because of the fact that
massive structure as a concrete dam is, it is possible to obtain, after the structure
excitation (with a value lower that the seismic value), only the fraction of critical
dumping corresponding to the first vibration mode.
The calculus was made in both cases of finite elements, for the first 10
vibration modes.
328 F. Ţepeş Onea
If we coupled ω1 + ωi and solving the equation systems obtained result α

and β coefficients. So for plane discrete mash α = 1, 44, β = 1,51E − 3 for ω1 and
ω3 and α = 1.97, β = 5.66 E − 4 for ω1 and ω10 . We can observe that the effect
of the mass matrix increase and the effect of stiffness matrix decrease in the same
direction with the increase of the second frequency taken into account.
In spatial mesh case, the fundamental vibration mode is reduce
ω1 = 13, 48rad / s and for the plane mesh ω1 = 21.3rad / s . This difference
appears because in spatial mesh we take account of the torsion vibration modes
also.
For the spatial mesh and for the frequency ω1 and ω3 α=0,893 and
β = 2,5E − 3 while if use the frequency ω1 and ω9 α = 1, 063 and
β = 1.56 E − 3 .
In the case of spatial discrete mash we can notice a mass matrix influence
grow and a stiffness matrix influence diminution in the same time as the pulsation
value grow. The variation of the α and β factors is much reduce when is use the
spatial mash. For the spatial mash case we obtain a mean value α Ed = 0,99 and for
the plane mesh α med = 1.72 . All the results are presented in tables 3.1 and 3.2
Table 1 Plane mesh

Num. Ωi (rad/s)
1 21.3
2 38.59
3 44.43
4 79.44
5 87.59
6 96.43
7 109.0
8 118.8
9 121.3
10 155.1
ω1 + ωi α β
1+3 1,44 1,519E-3
1+5 1,71 9,17E-4
1+7 1,78 7,67E-4
1+10 1,97 5,66E-4
Table 2 Spatial mesh

Num. ωi
1 1
2 17,44
3 26,49
4 34,28
5 37,21
6 42,17
7 45,64
8 49,8
9 50,25
ω1 + ωi α β
1+3 0,893 2,5E-3
1+5 0,989 1,97E-3
1+7 1,041 1,689E-3
1+9 1,063 1,567E-3
The calculus was resumed for a critical dump fraction 8% in witch case we
obtain for spatial mash and the pears ω1 and ω9 the following results α=1.69 and
β=2.51E-3, results with no big difference compare with the case of critical dump
fraction of 5%.
It is obvious that only for a spatial discrete mash the obtained results are
close to reality. The influence of superior modes use in the case of spatial mash
have no significant effect on the α and β coefficients as it presented in table 2.
It is noticed that in the same time with the increase of the frequency the
mass matrix effect increase to and also the stiffness matrix effect decrease.
So, we can say that the stiffness matrix effect connected with frequency is
major.
After coefficients calculus, it was analyzed the dam response at the same
excitation with and without damping matrix effect. It was followed the effect of
using vibration modes 1-3, 1-5,1-7,1-10 in mass end stiffness matrix coefficients
on the stress and displacement response. In the table 3,4,5,6 are presented stress
and displacement values for different coefficient pairs α and β, for plane and spatial
mesh.
1.1 Plane mesh-displacements compare

(nods was select on a vertical for the coping of the weir to the foundation of the
dam)
330 F. Ţepeş Onea
Table 3
Node Without ω1 + ω3 ω1 + ω5 ω1 + ω7 ω1 + ω10
damping
161 0.728E-2 0.4200E-2 0.4184E-2 0.4178E-2 0.4104E-2

152 0.6079E-2 0.3508E-2 0.3495E-2 0.3490E-2 0.3429E-2
143 0.4149E-2 0.2407E-2 0.2399E-2 0.2396E-2 0.2876E-2
124 0.3706E-2 0.2154E-2 0.2147E-2 0.2144E-2 0.2107E-2
94 0.2404E-2 0.1406E-2 0.1401E-2 0.1399E-2 0.1375E-2
1.2 Plane mesh-stress compare (stress are presented in upstream-downstream

points)
Table 4
Without ω1+ω3 ω1 + ω5 ω1 + ω7 ω1 + ω10
damping
σ(-)daN/cm2 -18.2 -10.93 -10.92 -10.91 -10.75
σ(+)daN/cm2 14.9 7.57 7.54 7.53 7.37
1.3 Spatial mesh-displacements compare (nods was select on a vertical for the
coping o the weir to the foundation of the dam)
Table 5
damping
161 0.433E-2 0.3319E-2 0.3326E-2 0.3325E-2 0.3326E-2
152 0.3646E-2 0.2778E-2 0.2783E-2 0.2786E-2 0.278E-2
143 0.3096E-2 0.2352E-2 0.2357E-2 0.236E-2 0.236E-2
124 0.2328E-2 0.1764E-2 0.1769E-2 0.1771E-2 0.1773E-2
94 0.1593E-2 0.1205E-2 0.1209E-2 0.1212E-2 0.1213E-2
1.4 Spatial mesh – stress compare (stress are presented in upstream –

downstream points)
Table 6
damping
σ(-)daN/cm2 -12.03 -9.2 -9.2 -9.2 -9.2
σ(+)daN/cm2 7.16 5.2 5.2 5.2 5.2
It is presented in the table 7 displacement comparative values for the spatial mesh
for two critical dumping ratios of 5% and 8%.
Table 7
5% Rayleigh 8% Rayleig Diference %
323 -0.3878E-3 -0.3623E-3 -6.57
-0.8799E-2 -0.7545E-2 -14.25
-0.2340E-2 -0.2272E-2 -2.9
314 -0.2895E-3 -0.2678E-3 -7.49
-0.7398E-2 -0.6348E-2 -14.19
-0.2276E-2 -0.2210E-2 -2.89
299 -0.1890E-3 -0.1738E-3 -8.04
-0.5644E-2 -0.4849E-2 -14.08
-0.2144E-2 -0.2082E-2 -2.89
278 0.1383E-3 0.1354E-3 -2.09
-0.4047E-2 -0.3480E-2 -14.01
-0.1965E-2 -0.191E-2 -2.79
256 0.1032E-3 0.1024E-3 -0.77
-0.2943E-2 -0.2529E-2 -14.06
-0.1779E-2 -0.1731E-2 -2.69
The displacement comparative graphics are presented in figures 1 and 2 and the
stress calculus points in figure 3.
Figure 1 Displacement comparison for spatial mesh
It is also notice that as well as for plane and spatial discrete mash, if damping
matrix is used, the stress and efforts values are almost similar for all the coefficient
pairs α and β used. It was also noticed that for Rayleigh models use, only the first 3
vibration modes are required.
332 F. Ţepeş Onea
Major response differences of 14% are obtained only between 2 critical dams of
2% and 8%.
Figure 2 α (ω ) for plane mesh and spatial mesh
Figure 3 β (ω ) for plane mesh and spatial mesh

Figure 4 Modal analyze. Spatial mesh for a concrete dam. The six vibration mode
Freq=42.17rad/s; T=0.149s
References
1. Clough. R.W., Penzien J., Dynamics of Structures, New York: McGraw-Hill Book Co.1975
2. Bathe. K.J. and Wilson, E.L. Numerical Methods in Finite Element Analysis. Englewood Cliffs,
N.J.: Prentice-Hall,Inc.,1976
3. D.Stematiu, Calculul structurilor hidrotehnice prin metoda elementelor finite, Editura Tehnică
Bucureşti, 1988. (in Romanian)
NDT Devices as Long-Term Vision in Assessing of Old Masonry

Buildings by Probabilistic Approach
Magda Brosteanu1 and Teodor Brosteanu2
1
CCI Department, Faculty of Civil Engineering, Technical University of Iasi, Romania
2
CFDP Department, Faculty of Civil Engineering, Technical University of Iasi, Romania
Summary
Assessment and retrofitting of existing buildings made by masonry structure

represent a huge engineering challenge as a distinct problem versus the design of a
new building. Structural Strengthening and Thermal Renovation have to be applied
at the same time.
This article refers to Two Non-Destructive Testing Devices in Diagnostic and
Assessing under realistic conditions of Old Masonry Buildings evaluated to the
real performance. It focuses on flat-jacks technique to determine the state of stress
in masonry, and on monitoring of buildings health using IR thermal-graphic
method. Minimum Equipment for Compressive Strength Qualifying Test on
Masonry Large Specimen, and Minimum Equipment for Infrared Thermal-Graphic
Test in Quality Control of Thermal Insulation are presented.
The results deal with the Masonry Strengths of walls and columns for the
computation of Nominal Assurance Degree of Seismic Action of the existing
masonry dwellings and monumental buildings according to the Strengthening
Structural Design for Compressing Loading and Lateral Loading of damaged
buildings.
Other results deal with the Surface Temperature Distribution and Moisture
Content in Visible and Infrared Images for Quality Control of Thermal Insulation
according to the Thermal Renovation Design for Heat Losses and Yearly Thermal
Balance of damaged buildings.
The measured values are random variables. The characteristic values of random
variables as fractiles for a given probability will be computed using the
Probabilistic Approach.
KEYWORDS: monitoring, non-destructive testing (NDT), flat-jacks technique,

infrared thermal-graphics, masonry characteristic compressive strength, surface
temperature, random variables, input data in strengthening design.
1. INTRODUCTION
Many masonry structures were built over long periods of time using different
technologies and materials. There are hidden damages out of a professional control.
There are dwellings and monumental buildings made by masonry structure, placed
on seismic hazard zone that need to be retrofitted at new performance levels.
The Romanian law L10:1995 concerning the quality in constructions defines one
main component from 11 ones of quality system such as serviceability behavior
and remedies during the life span of buildings.
Investor and/or owner or users have to be able to define correctly the users’
requirements on custom mode offer and contract form of technical project for
assessment or retrofitting.
Structural Safety (strength and stability) and Thermal Protection (thermal
insulation and moisture protection for heat energy saving) are two essential
performance requirements of buildings. Assessment and retrofitting of existing
buildings made by masonry structure represent a huge engineering challenge as a
distinct problem versus a new building design. Framing of damages into potential
risk degrees has a cultural, social and economic impact.
Structural Strengthening and Thermal Renovation have to be applied at the same
time.
The existing buildings are big consumers of energy. The volume of existing
buildings is so big and it contains many masonry buildings that need an upgrading
of the insulating by thermal retrofitting that means no interstitial condensation risk
and no surface dew point temperature. There are almost 8.0 million of dwellings as
2.8-rooms conventional apartments with an amount of fuel required for one
season’s heating of 10.4-21.5 mil tones conventional fuel per year. By summing-up
almost 20% for other buildings, the Romanian annual energy consumption for the
heating of buildings is between 12.5-25.8 mil tones conventional fuel per year.
There are damaged old buildings placed on seismic hazard zones. Local remedies
were been applied from economic reasons. Few local inventories were been
performed with modest results. The existing buildings with classified damages
need an upgrading of cross walls structure in order to reduce the seismic risk for
users’ safe.
No general rules are expected. Case-studies will be performed. Few official
inventory and preliminary decisions of alternative spaces of living or working
about were been performed. This is a worried factor.
There will be a sustainable activity to restore old buildings and monuments, not
matter where. The trends in building costs in Germany (7) for new buildings and
existing buildings are presented in the Fig.1.
336 M. Brosteanu, T. Brosteanu
Fig.1. Trends in building costs in Germany
This article refers to the non-destructive testing device in assessing of old masonry
buildings which were originally built over a long time ago, and it is focuses on flat-
jacks technique to determine the state of stress in masonry, and on health
monitoring using IR thermal-graphic method, as input data for strengthening
structural design of masonry members in compression, and thermal renovation
design.
2. DIAGNOSTIC SURVEYS
Diagnostic strategy is presented in the Fig.2. Keywords of diagnostic strategy using

the probabilistic approach are the followings: sources of risk, damages, weak
points, indicators, probability of failure, limit states, monitoring, early damages
survey, random variables, input re-design data, costs etc.
Diagnostic
Survey
Weak Direct Adaptive Measuring Data

Points Monitoring Model Technique Management
Fig.2. Diagnostic Strategy
Preliminary surveys focus on:

(i) geometric survey using topographic method and photogram-metric

technique to identify the geometry of building;
(ii) crack-detection using piezoelectric arrays to identify static pattern and
probable causes of instability.
Final aim of preliminary surveys of the overall pattern of cracks is the framing of
existing buildings with classified damages into 4 potential classes of seismic risk
RS /5/. RS1 means a high risk of collapse. RS2 means a low probability of
collapse but there are major damages of bearing elements. RS3 means that the
damages of bearing elements do not disturb the structural safety but there are major
damages of nonbearing elements. RS4 means that the expected seismic answer of
the existing building and the designed seismic answer for a new building are the
same.
Note that should be kept in mind the influence factors of quantifying: the design
seismic zone, the structural pattern on height, the aseismic conformity of structure,
the nominal assurance degree to seismic action, the probable nature of collapse of
bearing elements (either a ductile nature, a semi-ductile one or a breakable one),
the reinforcement details, the building importance factor, the building life span, and
the building inventory.
Diagnostic surveys focus on:
(iii) coring survey using coring and sampling technique to provide specimens
for material testing;
(iv) X-ray survey inside the boreholes using a rotating micro-camera to provide
3-D image of wall;
(v) non-destructive testing using flat-jacks technique (6) to determine the state
of stress in masonry realistic specimen, and the strengths of masonry:
the fk compressive strength, the fvk shear strength, and the fxk tensile strength in
bending, and the fvk shear strength of mortar layer;
(vi) sonic testing using an ultrasonic pulse excitation through, to perform a
detailed survey of sound velocity distribution over;
(vii) radar-screen survey using high frequency electro-magnetic waves to
investigate the interfaces between material layers having different dielectric
constants;
(viii) health monitoring using fiber optical sensors and IR thermal-
graphics for data capture to control any change in behavior, such as temperature
distribution, moisture distribution, strains in all directions, relative movements,
inclination etc;
(ix) material-destructive testing on coring samples using axial compression and
other tests to assess the masonry strengths, and the degree of physical-
chemical damages.
The calibration is a very important step.
3. FLAT-JACKS TECHNIQUE
Characteristic values of the compressive strength for the computational design

purpose of new masonry structures are only regulated by design codes. Hence, the
behavior of old masonry structures in compression has to be investigated to provide
strength values.
Flat-jacks technique is a non-destructive testing method for quantification of
phenomenon with regard to the state of stress in masonry.
Minimum Equipment for Compressive Strength Qualifying Test on Masonry Large
Specimen consists of:
Loading Devices: Electric Drill, Core Drill, Hydraulic Pump, Hydraulic Flat-Jacks,
Hydraulic Presses, Service Gauges, and Flexible Hose;
Recorder and Dataloggers: Data Acquisition Board, Inductance Strain Gauges
Transducers, Wheastone Bridge Circuit, Graphic Recorder (Printer), Deformeters
(Comparing Devices), Micrometers, Dilatometers, Sonic Detector, and Software.
Testing Method consists of: Stress-Strain Analysis of Large Specimen in
Compression, Modulus of Elasticity, Shear and Dilatometer Tests on Masonry
Sample Area, Diagnostic Surveys, Sonic Surveys, Structural Monitoring, and
Structural Strengthening.
Fig.3. Flat-Jacks Compressive Starter Kit
Flat-Jacks Compressive Starter Kit is shown in the Fig.3. It contains: 5 flat-jacks,

10,000 psi capacity hydraulic hand pump, 1,000 psi test quality pressure gauge,
2,…,6 feet hydraulic hoses with quick connect fitting, removable extensometer of
0.0001 inch resolution, inch reference bar, marking gauge, 100 extensometer gauge
points, rugged field case, where 1 psi=10-3 ksi=10-3 kip.sqin-1=0.006895
MPa=0.006895 N.mm-2, 1 in=25.4 mm, 1 ft=12 in=0.3048 m.
Fig.4. Flat-Jacks Shear Starter Kit
Flat-Jacks Shear Starter Kit is shown in the Fig.4. It contains: 10,000 psi capacity
hydraulic hand pump, 10,000 psi test quality pressure gauge, 5,000 psi test quality
pressure gauge, 10 feet hydraulic hose, 10 tones capacity hydraulic ram with tilt
saddle, 2-3.5 inch x 2.0 inch x 0.5 inch hardened steel bearing plates, 2 inch Ф x
1.4 inch cylindrical aluminium spacer, 2 inch Ф x 4.0 inch cylindrical aluminium
spacer, rugged field case.
Fig.5. Survey to perform flat-jacks test
The survey using flat-jacks technique carried out without scaffolding is presented
in the Fig.5.
Masonry sample will be marked by two horizontal cuts into which two flat jacks
will be inserted and it will be subjected to an axial compression. The pressure of
flat-jacks will be increased up to the re-establishment of the initial degree of
deformation. It is possible to find the masonry compressive strength fk in
MPa=N.mm-2 by increasing of compression load until the first vertical crack will
arise in the brick (1), (2), (3) because the lateral strain of mortar will only meet the
effect of adhesion and friction between bricks and mortar.
Masonry sample will be marked by two vertical cuts, a brick will be removed out,
and two small horizontal-acting hydraulic flat jacks will be inserted. It is possible
to find either the shear strength of mortar layer fvk or the friction angle between
bricks and mortar by modifying the vertical and horizontal stresses.
Masonry strengths are random variables. By the probabilistic approach (9), the
characteristic value of a material strength is the probability p=0.05 of lower values
than this. The fractiles x p will be computed as follows:
xp = mx ±k σ x (1)
where,
m x is the mean of random variables;
k is the factor or standard deviation number depending on the p probability
and the type of distribution;
σ x is standard deviation.
From the field data arranged into an increasing row, data selection, graphic
probability distribution such as histogram and frequency polygon, on recommends
Henry’s line for checking of the theory-testing concordance, and the elimination of
the absurd results. If the mean line that rectifies the plotted line is very close by the
middle field points, it will be a concordance between the measured values
distribution and the theoretical distribution. A few values will be rejected, and a
few values will be kept.
The intersection of Henry’s line with the abscissa is m x point. The intersection of
Henry’s line with the 16% line is the a angle between Henry’s line and the
ordinate, where σ x is tangent of a . The intersection between Henry’s line with the
5% line means the x p fractile.
Mean value with relative frequency named first moment or math hope and the
variance named central second moment are computed using locating and spreading
of statistical evaluation from Table 1 and Table 2, using Eq.2 and Eq.3.
Fractile is computed using Eq.1 for 0.05 probability of lower value than x p .
Table 1. Data selection
Intervals Central Absolute Relative Sum Diagram of events

Value frequency on interval
Frequency
-2
R
N.mm
Ni fi=ni/n Σni Σfi Ni
%
1.46-1.50 1.48 6 0.1111 6 11 xxxxxx
1.51-1.55 1.53 8 0.1482 14 26 xxxxxxxx
1.56-1.60 1.58 9 0.1666 23 43 xxxxxxxxx
1.61-1.65 1.63 12 0.2222 35 65 xxxxxxxxxxxx
1.66-1.70 1.68 8 0.1482 43 80 xxxxxxxx
1.71-1.75 1.73 6 0.1111 49 91 xxxxxx
1.76-1.80 1.78 5 0.0926 54 100 xxxxx
7 1.63 54 1.0000 54 100 5 10
Table 2. Mean and standard deviation computation
Central Relative Multiplication Central Variance

value frequency deviation
Fi=ni/n 2
Xi fi Xi A i= σ=
2
Xi - m x f i (Xi – m x)
1.48 0.1111 0.1644 -0.1424 0.0022
1.53 0.1482 0.2267 -0.0924 0.0012
1.58 0.1666 0.2632 -0.0424 0.0002
1.63 0.2222 0.3621 0.0076 0.0001
1.68 0.1482 0.2489 0.0576 0.0005
1.73 0.1111 0.1922 0.1076 0.0013
1.78 0.0926 0.1648 0.1576 0.0023
1.63 1.0000 1.6224 0.0078
8
mx = ∑f X
1
i i = 1.6224 (2)
σx = ∑ f (X i i − m x ) = 0.0883
2
(3)
The characteristic compressive strength of masonry large specimen fk computed

as fractile of 1.4771 MPa:
x p = 1.6224 - (1.645)(0.0883)=1.4771 (4)
4. IR THERMAL-GRAPHICS
The key feature to reduce the amount of heat losses through the building envelope
to a very low value is a very high standard of thermal insulation. Hence, the
thermal behavior of masonry envelopes has to be investigated to provide thermal
resistance R -values and thermal transmittance U -values (4).
Lock in thermal-graphics is a picture-based method for visualization and
quantification of phenomenon with regard to the heat transfer.
The thermal energy caught by the IR camera from an object is focused by the
radiation detector and the electronic system converts the electrical signal produced
by detector into a video signal on monitor screen.
The total energy emitted by a body’s surface is defined by Eq.5:
∞
ET = ∫ε
λ =0
λ E λ dλ = εT σ T4 (5)
where,
ET is the total radiation energy emitted with the wavelength λ at
T temperature;
ε λ is the body’s emissive power for the radiation with the wavelength λ ;
E λ is the radiation energy emitted with the wavelength λ ;
ε T is the body’s emissive power for the radiation with the whole wavelength
spectrum;
σ = 5.672 · 10-8 W. m -2.K-4 is Stefan Boltzmann’s constant.
Minimum Equipment for Infrared Thermal-Graphic Test in Thermal Insulation
Quality Control consists of: Thermnocouples Meters and Probes, Equipment for
Recording, Stocking, and Processing, Thermal CAM P-Series (professional
infrared camera), and Thermal CAM Connect Software.
Testing Method consists of: Data Acquisition of Temperature, Temperature
Control, and Thermal Imaging System.
IR equipment can catch the body’s thermal image in real time using a receiving and
recording system (IR camera), and a processing system of the thermal images
(computer and software). IR camera consists of a scanning system, an infrared
radiation detector, an electronic system, and a microprocessor.
Working parameters are as follows: temperature range from -20°C till 500°C with
the sensibility increment of 0.1°C, working autonomy of 3 hours, and operation
temperature from -15°C till 45°C. The calibration between the surface emissive
power and the surface temperature is given.
A few results are presented (11). In the Fig.6, the visible image and IR image with
cold bridges zones are presented. In the Fig. 7, the earth moisture infiltration at the
building ground floor level (12) is presented. In the Fig.8, temperature values on a
window with the layering effect are presented as screened results.
Fig.6. Visible and infrared images with defects due to the cold bridges
Fig.7. Infrared image of the earth infiltration tracing
Fig.8. Screened results of temperature distribution

5. CONCLUSIONS
There will be a sustainable activity to restore old buildings and monuments, not
matter where it will be. In Germany, trends in building costs for new and existing
buildings confirm this.
Assessment and retrofitting of existing buildings made by masonry structure
represent a huge engineering challenge as a distinct problem versus a new building
design.
Inspection techniques via monitoring and parallel testing i.e. the diagnostic surveys
under realistic conditions are presented.
Flat-jacks technique as a non-destructive testing method is suggested for input data
capture of masonry strengths in strengthening design of damaged old buildings.
Note that for high buildings, few experts have to be rock-climbers.
Lock in thermal-graphics as a picture-based method for visualization and
quantification of phenomenon with regard to the heat transfer, but not only is
suggested for input data capture of masonry thermal resistance in thermal
renovation design of damaged old buildings. It is attractive. IR thermal-graphics is
an important instrument to stimulate the interest concerning the quality of works.
The right decision is taking the best insulation available. Note that for high
buildings, few experts have to be rock-climbers, too.
Lock in acoustics and thermal-graphics is suggested for research.
On recommends the equivalent rigid frame pattern for simplified structural design
(10), and for the best approximation to the specimen experimental behavior (8),
and for actual brick shear walls structure, as well.
The equivalent rigid frame consists of columns with the same features as posts, and
slabs with the same features as lintels. The frame columns will substitute the posts
spanning between posts centre lines. The slabs will substitute the lintels on the
entire span of columns centre lines.
References
1 xxx The EC 6-1-3:1996 “Design of Masonry Structure-General Rules for Buildings-Rules

for Reinforced Masonry”.
2 xxx CR 6:2006 Codul de proiectare pentru structuri din zidarie.
3 xxx The EN 1052-3:1996 “Methods of tests for masonry”.
4 xxx SR EN ISO 13790:2005:2004 Evaluarea performantei termice a cladirilor.
5 xxx P 100:2006:2004 Codul de proiectare seismica cap. 1”Prevederi de proiectare pentru
cladiri”, cap.3 “Prevederi pentru evaluarea si proiectarea consolidarii constructiilor
vulnerabile seismic, cap.8 “Prevederi pentru proiectarea consolidarii monumentelor
istorice si a constructiilor cu valoare arhitecturala”.
6 xxx The former ISMES spa “Building and Monument Restoration-Surveys and Control”,
Bergamo, Italy, 1996.
7 xxx Collaborative Research Center SFB 477, founded by the German Science Foundation
(DFG), “Life Cycle Assessment of Structures via Innovative Monitoring”, University
Carola-Wilhelmina-Braunschweig, Germany, http://www.sfb477.tu-bs.de
8 Hendry, A.W., Sinha, B.P., Davies, S.R. -Design of Masonry Structures, Third edition of
Load Bearing Brickwork Design, E& FN Spon Publishing House, An Imprint of Chapman
& Hall, London, Weinheim, New York, Tokyo, Melbourne, Madras, 1997, ISBN0 419-
21560-3.
9 Brosteanu, M. -Constructions. Philosophy of Design, The first edition. Cerni Publishing
House, Iasi, 1998, ISBN973-9378-13-7.
10 Brosteanu, M., Brosteanu, T. -Selection of theoretical method for lateral load analysis of
brick shear walls structures, In the Proceedings of International Symposium
“Computational Civil Engineering”, “Matei-Teiu Botez” Academic Society Publishing
House, Iasi, May 26, 2006, ISBN(10)973-7962-89-3, ISBN(13)978-973-7962-89-8.
11 Avram, C. -Investigatii IR pentru controlul calitatii protectiei termice, Satisfacerea
exigentelor de izolare termica si conservare a energiei in constructii, “Matei-Teiu Botez”
Academic Society Publishing House, Iasi, 2003, ISBN973-85882-7-8.
12 xxx The former http://www.agema.com; http://www.flir.com.
IT Approach in Heat Transmission Visualizing through

Opaque Walls
Magda Brosteanu1 and Teodor Brosteanu2
1
CCI Department, Faculty of Civil Engineering, Technical University of Iasi, Romania
2
CFDP Department, Faculty of Civil Engineering, Technical University of Iasi, Romania
Summary
This article refers to the 2-D thermal field computation in the steady-state
conditions for thermal design of the building envelope. A procedure to analyze the
heat transfer through the opaque wall is provided. The results are as follows: Heat
Flow Image and Density, Surface Temperature Distribution, and Isothermal Lines.
Computing programs are very useful instruments in the designing of adjustment to
R-value and in the solving of problems.
KEYWORDS: building, envelope, opaque wall, heat transfer, steady-state

conditions, computational analysis, surface temperature visualizing.
1. INTRODUCTION
The Romanian standard SR EN 10211-2:2001 /1/ is identical to the European

Standard EN ISO 10211-2:2001-Thermal bridges in construction-Calculation of
heat flows and surface temperatures-Part2: Linear thermal bridges.
Solving problems by 2-D analysis has required using the dedicated software and
computer aided design systems.
2. THERMAL FIELD
The thermal field represents all temperatures θ (x, y, z, τ )-values, where x, y, z are
the spatial co-ordinates of the point and τ is the time.
The thermal field is computed from the Laplace’s Eq., as follows:
∂θ
λ ∆ θ = cρ +I (1)
∂τ
where,
λ is the thermal conductivity of the material, in W.m-1.K-1;

c is the heat capacity of the material, in J. kg -1.K-1;
ρ is the density of the material, in kg.m-3;
I is the source of heat, in W.m-3.
There are many heat transfer problems from which 2-D problem in steady-state
conditions is the most useful formula for θ(x, y ) -values computation by the
Numerical and Computational Techniques:
in steady-state conditions τ =0, and in transient ones τ ≠ 0;
in linear conditions with λ ,c , ρ =constants, and non-linear ones with
λ (θ ),c(θ ), ρ(θ ) ;
1-D problem of θ(x ) , 2-D problem of θ(x, y ) , and 3-D problem of θ(x , y, z ) .
3. MATHEMATICAL MODEL
By the Finite Difference Method or Finite Element Method, the Differential

Equations with second partial derivative of θ with respect to dx2 squared at
constants (m, n) have to be changed into a system of Algebraical Linear Equations
using difference operators or finite operators:
∂ 2θ ∂ 2θ
λ + λ =0 (2)
∂x 2 ∂y 2
The digitization after x and y in elementary layers supposes the nodes disposal at
the borders, at intermediary surfaces between different materials, and in the same
material. The differential operators could be approximated with the help of two
levels schemes resulted from θ(x, y ) development in the Taylor’s series around the
node with (m, n) co-ordinates.
Finally, the formulae have resulted by the imposed, uniqueness conditions, initial
and at the limits ones:
the 1st case condition (Dirichlet’s one) in steady-state conditions has indicated that
the inside air temperature θ i and the outside air temperature θ e are fixed;
the 2nd case condition (Newman’s one) has indicated the occurrence of a
continuous density of heat flow q s at the boundaries, i.e. we have to consider
many nodes up to the breadth of the influence zone lp of the cold bridge limited by
the normal heat flow on both surfaces of the element:
∆θ
qs = (3)
R
the 3rd case condition (Fourier’s one) has indicated the heat transfer on the internal
surface and external surface of the element:
θ si − θ 2 θ 3 − θ se
α i (θi − θ si ) = λ and λ = α e (θ se − θ e ) (4)
∆y ∆y
the 4th case condition has indicated the heat transfer between different materials
deduced from the conservation Eq.:
∑q = 0 (5)
4. HANDLING THE PROGRAM
There are thermal bridges as parts or portions of a built element, given by the
technological or structural reasons, in which the insulating material is completely
absent, i.e. the insulation is bridged.
The heat flow increases across the thermal bridges and the inside surface becomes
cooler, giving the reason of “the cold bridge term”. There is also an increased risk
of condensation.
Sometimes, it is necessary to calculate the thickness of the additional insulating
material in order to produce the same specified R0-value as in the current section.
The best position of thermal insulating material is at the outside surface of building
element.
Thermal resistance of the envelope R0 in m2.K.W-1 can be found from the Eq. as:
R0 = R i + R + R e (6)
All details selected can be developed for the claimed input data.
The results of computational approach are as follows:
the image of heat flow;
the density of heat flow q s , in W.m-2;
the distribution of the outer surface temperature , θ es (x , y ) in 0C ;
the distribution of the inside surface temperature θ is (x , y ) , in 0C;
the isothermal lines θ(x , y ) =constant;
the zones of almost constant temperatures;
the lines of almost constant densities of heat flow;
the zones of almost constant densities of heat flow.
Considering a case-study /2/ of a building outer wall joint subjected to 2-D heat
flow through, in the winter time conditions, for a masonry wall with a cold bridge
due to a concrete column with and without adjustment at the outside surface using
an EPS layer, the results are shown in the Fig. below. Wall structure thermal
properties and indoor-outdoor surface resistances are taken, and the temperatures
remain steady.
Wall Joint Digitization is presented in the Fig. 1.
Fig.1. Wall Joint Digitization
The image of lines of conductive heat flow through the outer wall after the thermal
adjustment is presented in the Fig.2. There is a non-uniform heat flow through
because of the effect of lateral heat flow is significant.
Fig.2. Image of Heat Flow through

The distribution of heat flow density and surface temperature on a wall joint with
and without adjustment of R0-value is compared in the Fig. 3, 4, 5, 6, 7, and 8.
Fig.3. Variation of the Density of Heat Flow after the Adjustment
Fig.4. Zones of almost Constant Densities of Heat Flow after the Adjustment
Fig.5. Variation of the Density of Heat Flow before the Adjustment
Fig.6. Zones of almost Constant Densities of Heat Flow before of the Adjustment
Fig.7. Zones of almost Constant Temperatures after the Adjustment
After the adjustment, the isothermal lines are almost parallel to the wall external
surface and the effect of lateral heat flow is not significant.
Fig.8. Zones of almost Constant Temperature before of the Adjustment

5. CONCLUSIONS
Thermal comfort and protection requirements mean a positive state of mind due to
a friendly inside environment, and a lower cost of heating for energy conservation
resulting from the best thermal isolation available concerning the improvement of
relative humidity, condensation risk, comfort temperature, thermal resistance,
thermal transmittance, and heat losses.
2-D FEM wall pattern in the steady-state conditions is suggested for the design of
the adjustment of R0-value. After the adjustment, the effect of lateral heat flow is
not significant. Note that the best position of thermal insulating material is at the
outside surface of building element.
IT Approach is an important instrument to stimulate the taking of right decision by
heat transmission visualizing. It is attractive.
On recommends the dynamic heat transmission through the building envelope with
solar gain and material thermal storage, for research and effective design, as well.
References
1 xxx The Romanian code SR EN ISO 10211-2:2001 “Thermal bridges in construction-
Calculation of heat flows and surface temperature. Part2: Linear thermal bridges”.
2 Moss, J. Keith –Heat and Mass Transfer in Building Services Design, “E&FN Spon”
Publishing House, an imprint of Routledge, London, 1998, ISBN 0 419 22650 8.
Iaşi, România, May, 2007
Comparative Study on the Results of Analytical and

Experimental Analysis of a Steel Taintor (Radial) Gate
Adrian Prodescu and Daniel Bîtcă
Steel Structures Department, University of Civil Engineering, Bucharest, Romania
Summary
The main load of a hydraulic gate is the water pressure, especially for the
submerged gates. A good accuracy of the computation methods leads to a more
secure structure.
The paper presents a comparison between different computation methods of this
type of structures. A radial gate was analyzed. The size of the submerged outlet is
4,0 x 4,0 m and the water depth is 20,0 m
The following methods were used in the analysis:
• The finite elements method, using a mesh of the structure of the gate.
• Model studies. An 1:4 scale model was built in order to perform this study.
The model was subjected to water pressure. In several points of this model,
strain measuring devices were installed and the strains were measured at
several water pressure magnitudes.
The results of the first two computation methods are compared in some points of
the structure. With the aid of the second method, the strains were computed in
order to compare them with the strains measured on the model.
These comparative studies try to establish the accuracy of each of the methods
presented above.
KEYWORDS: radial gate, water pressure; experimental study
1. STUDY’S OBJECT
In this paper it is analyzed a depth radial gate, which have the dimensions
4,00x4,00m, located on 20,00m depth. The structure of radial gate is presented in
figure 1.
Figure 1. The structure of analyzed radial gate
The main dimensions of the 1:4 scale model are:

- the outlet dimensions Lxh = 1050x1000 mm;
- the plate radius is 1300 mm;
- the plate thickness is 2,5 mm, in order to be 1:4 by the thickness of the gate's
plate (10 mm);
356 A. Prodescu, D. Bîtcă
Fig 2 Assembling the scale model

2. THE HYDRAULIC PRESSURE TEST OF THE SCALE MODEL
In order to apply water pressure on the model, a steel tank was built such
as the model would fit on one of the tank's lateral faces. The tank was made out of
6 mm thickness sheet, which was stiffened with angle steel shapes.
Fig. 3. The scheme of the hydraulic pressure test
The sealing between the model and the tank was made using siliconic sealant with
good elastic properties, in order to allow the displacements between the tank and
the model.
The boundary conditions were made similar to the real ones, when the gate is
closed.
A vertical pipe was attached to the top face of the tank. This pipe allows the
increase of water pressure in 1 m steps, the maximum pressure being 6 m above the
tank.
The deflections of the model were measured in 6 load steps; the water pressure was
increased with 1 m in each step.
The deflections of the model in the upstream - downstream direction were
measured with the aid of some measuring devices with 1/100 mm precision. The
points in which these measuring devices were placed are shown in fig. 4 and 5. In
pts 1 and 2 (where the displacements should be 0) two measuring devices were
installed in order to correct the other measured deflections.
Fig. 4. Aspects during the test

Fig. 5. The points where the measuring devices were installed
The corrected deflections in points 3 to 7 are shown in fig. 6. The corrections were
made using the deflections measured in points 1 and 2 (which are the bearings
where, theoretically, the displacement is null)
Figura 6. The corrected deflections in points 3 to 7.

3. COMPARRISON BETWEEN THE RESULTS OF THE TEST AND

THOSE OBTAINED BY COMPUTATIONS
A finite elements model was build, having the same characteristics as the scale
model. The load steps were similar to the ones from the test.
The graphs in fig. 7, 8 and 9 shows comparisons between the displacements
obtained in the test and in the computations.
One can notice a good match between the results obtained using the two methods,
especially in the first four loading steps.
Fig 7. Displacements in point 3

In points 3 and 6, situated in areas of the structure having rather big stiffness, one
can see the results are very close, the maximum differences being of about 19 %
for point 3 and about 26% in point 6.
As a conclusion, the closeness between the results confirms the reliability
of the finite element method, if the mesh is appropriate and sufficiently detailed.
References
1 Prodescu, A. Acurateţea modelelor de calcul folosite la analiza stavilelor segment, Teză de
doctorat, Universitatea Tehnica de Construcţii, Bucureşti, 1998
2 Beleş, I; Răduică, N. Construcţii metalice şi elemente de construcţii; Stavile metalice
Institutul de Construcţii Bucureşti 1979
Contribution to non-linear constitutive modelling of masonry

structures in 2D
Jiří Brožovský1, Alois Materna2 and Ivan Kološ1
1
Dept. of structural mechanics, VSB – Technical University of Ostrava, CZ708 33, Czech Republic
2
Dept. of building structures, VSB – Technical University of Ostrava, CZ708 33, Czech Republic
Summary
This paper discusses a non-linear constitutive model for modelling of masonry in

2D. The model consists from two independent parts: the model for mortar and the
model for bricks or stones.
The reason of using two different constitutive models is our need to be able to
modell individual bricks (or stones in the case of a stone masonry) and the mortar
between the bricks.
We have decided to use a smeared crack approach for the modelling fo a mortar. It
is an approach that is widely and successfully used for a constitutive modelling of
concrete and we assume that it can be also applied to a mortar that is very similar
to a concrete.
The model of mortar is based on an equivalent one-dimensional stress-strain
relation that depends also on 2D state of stress (through the Kupfer's failure
criteria that is used to obtain the limit stresses for the one-dimensional stress-
strain relation) and on some special material properties such is fracture energy of
mortar.
Bricks are modelled in a different way, As a generally brittle, they often can't be
effectively modelly by the smeared crack approach. The damage of the brick (a
crack) usually goes through the whole brick and it can be assumed that its
occurrence depends on a stress intenzity on an area and not on the stress size in an
individual material point. Thus we have decided to use a simple approach that can
be describes as a very basic non-local material model.
After the ckack on the brick is detected the material properties are changed in a
moment. Unlike the model for a mortar, it is assumed that there is no unloading
curve for this model,
KEYWORDS: finite element method, constitutive modelling, masonry, mortar,

bricks, crack band model, smeared cracks.
1. INTRODUCTION
The static analysis of masonry can be done in several ways. The most common is
approach is a linear elastic constitutive modelling. It allows to provide a relatively
simple computational analysis and many structures can be analysed with very
simple computational models (beams and frames) and often even without need of a
computer. Also the finite element analysis can relatively easy and non-complicated.
It is often said that linear elastic modelling is enough for a design of masonry
structures because they are not allowed to work in situations when their behaviour
can be non linear (under tension loads, for example).
But there are situations when the assumption of the linear elastic behaviour is not
sufficient. For example there are often needs of analysis of already existing
buildings or of historic structures and monuments that can work in an non-optimal
mode for masonry. For these cases a lot of different non-linear constitutive models
have been developed by many authors (for example [1]). Many of these approaches
are very complex and they can offer a high level of accuracy of results if they are
properly used. But it isn’t often easy (or even possible) to get all the necessary
input data that are needed for such models. In these cases the results can be even
less precise than the traditional computing approaches. There are also a lot of not
so advanced models that cannot offer so high precision of results but they usually
require lower number of input data and may they may be easier to use.
In this paper we present a constitutive model for masonry that we are developing.
This model was designed as a compromise between needs and possibilities of
computing and access to input data (however, the proposed model still can be too
complicated in some cases).
2. CONSTITUTIVE MODEL DESCRIPTION

2.1. Overview
The constitutive model includes two independent parts: the model for mortar and
the model for bricks. These parts are based on different assumptions and they are
impemented in different ways. The mortar is implemented using the smeared crack
approach and the crack-band model. Bricks are modelled with a simple variation of
a non-local model approach. This division of the model into two parts allows us to
create a models that can include individual bricks (or stones in a case of stone
masonry that is common in historic buildings) and also locate the mortar in
positions that are correspond with the real structure.
The proposed constitutive model is developed for 2D cases (for the plane stress
case).
364 J. Brožovský, A. Materna, I. Kološ
2.2. Model for mortar
The constitutive law for mortar is controlled by an equivalent uniaxial stress-strain

relation. The simplification of the problem from 2D to 1D is not ideal but it is
relatively easy to develop and understand. To make the uniaxial relation to be more
corresponding with the real behaviour the limits of the relation (strenghts in tension
and in compression) are computed from a 2D failure criteria (the Kupfer's criteria
is used here) and depends on the actual 2D stresses.
Figure 1. Uniaxial stress-strain relation for mortar
There used approach has several disadvantages. The one of the most important is
that results depend of properties of a finite element mesh. It is obvious, because
when a “damage” is detected the properties of the material became be reduced and
they are reduced on whole area of the element (or – in our case – on a whole area
that corresponds with an integration point of a finite element).
This issue can be minimized by usage of the Bazant's crack-band model when the
properties of the stress-strain diagram (the “Ez” value in Figure 1) depends on the
size of the cracked area (value L on Figure 2) and the material properties (the
fracture energy GF). The fracture energy can be obtained from special experiments
or it can be (with limited precision) computed (Karihaloo []).
The behaviour of the material in the compression is modelled in a similar way.
Figure 2. Lenght L of the cracked area
2.3. Model for bricks
The real bricks (or stones in stone masonry) have a different behaviour. They
usually are much more brittle than the mortar. The material model for mortar can
be also used for a modelling of bricks but it usually isn't ideal for this purpose.
We have prepared a different constitutive model. We assume that behaviour of
bricks is linear elastic until the damage (crack) is detected. Usually, the crack goes
throught the full height (or width) ob brick so it is obvious that the failure condition
must respect this.
We have proposed a use an approach that is similar to non-local material models:
the stresses in the brick are controlled on an area A with dimensions that are
comparable with height of the brick. It should guarantee that crack is detected if the
stresses in the brick are big enough to be able to create a crack. Thus a brittle
demage of the brick should occur.
After the brick is damage the material properties should be changed. Now we have
adopted the elasto-plastic behaviour of a material. After the crack is detected then
the normal stiffness of the material (represented by Young's modullus) is reduced
to zero. It is not an ideal approach because it means that the cracked material still
carry the previous stresses but we selected it for the relatively ease of the
implementation. After we will test the other parts ot the alghorithm (namely the
computation of the failure condition) we will improve the behaviour of the cracked
material to be more realistic.
366 J. Brožovský, A. Materna, I. Kološ
3. ILLUSTRATIVE EXAMPLE
Model description
To show the behaviour of the model we have prepared a simple (not very realsisic)
numerical example. The geometry is shown on the Figure 3. The sizes of bricks are
270x160 mm and the width of mortar is 10 mm in all cases.
Figure 3. Illustrative example – finite element model
The Figure 4 shows the load-displacement relation that was obtained from the
analysis of the model.
The relatively sharp change of the stress-strain curve is the result of a selected
constitutive model of mortar – the bricks under the loading cracked in a moment
and then the strength of the structure became reduced. From the Figure 4 it is
obvious that the mortar has a relatively small influence of the results in this case.
The relatively high strength of a damaged material in this case was a result of a
relatively large residual stiffness of a bricks (we used a elastoplastic model with
hardening for the bricks here to help the convergence of solution).
The results show that the model features a behaviour that we have expected to
obtain.
Figure 4. Computed stress-strain relation
4. CONCLUSIONS
The article shows a basic material model for masonry. It can be improved in
several ways, namely in the area of the modelling of bricks.
Acknowledgements
The works are supported from the Czech state budget through the Czech Science
Foundation. The project registration number is GA CR 103/03/P389.
References
1. Bazant Z. P., Planas, J., Fracture and size effect in concrete and other quassibrittler materials,
CRC Press, Boca Raton, 1998.
2. Cervenka V., Constitutive models for cracked reinforced concrete, ACI Journal, vol. 82, 1985.
3. Cervenka V., Inelastic finite element analysis of of reinforced concrete panels under in-plane
load, University of Colorado, Colorado, 1970
Behavior of newly developed FRP reinforcement in structures

under various load schemes
David Horak1, Martin Zlamal1, Petr Danek2
1
Department of Concrete and Masonry Structures, Brno University of Technology, Czech Republic
2
Department of Building Testing, Brno University of Technology, Brno, Czech Republic
Summary
An own reinforcement based on glass or carbon fiber reinforced polymers in frame

a Czech ministry of industry and trade research task was developed. A set of
experiments was made for reinforcing of concrete structures with this FRP internal
reinforcement.
The developed reinforcement was used for reinforcing of several concrete
elements. These elements were exposed to different types of loading. Their
behavior was monitored to verify the functionality of new reinforcement. Based on
this results it is possible to determine required properties of reinforcement used for
every sort of reinforcing (longitudinal or shear reinforcement).
This reinforcement was also used to additionally strengthen the masonry vaults
loaded with static and dynamic loads. Obtained results are compared with
theoretic results of nonlinear numerical analysis of constructions.
KEYWORDS: Longitudinal and shear GFRP reinforcement, reinforced concrete

structures, strengthened masonry vaults.
1. INTRODUCTION
At present non-metallic reinforcement is used very frequently (because of their

resistibility) in constructions that are exposed to aggressive environment’s
influence. It makes possible to reduce costs needed for special arrangements for
protection the common reinforcement and eventually consecutive repairs.
However the price of the non-metallic reinforcements is quite high (see Fig. 1).
And because this reinforcement form the significant part of the final costs of the
cross-section price, it is very advisable to (next to economical optimizing of the
cross-section [1]) use local non-imported (i.e. probably cheaper) materials.
Figure 1. Comparison of the average prices in the Czech Republic

(common materials and developed reinforcement)
The economic aspect mentioned above isn’t in the Czech Republic so strong,
because there isn’t any native producer of this kind of reinforcement. While using
it is necessary to import the reinforcement from abroad, which makes construction
sometimes more expensive.
In terms of research in frame of the Czech Ministry of Industry and Trade,
development of “home” reinforcement based on glass and carbon fibers has begun.
It is of course necessary to check out functionality of this system – i.e. functionality
of interaction of reinforcing bar and surrounding concrete.
But not only new concrete structures are in the centre of interest. Masonry
continues to be popular because of its relative simplicity of application in the
technical practice. Indeed, for a new use of structural masonry reasonable
constructional rules are required, because conventional approach based on the
experience is unacceptable nowadays. In addition, most methods of carrying
capacity assessment and of strengthening for the existing masonry construction are
increasingly based on analyses of mathematical simulation and appropriate (linear
and nonlinear) computational models. One method of load-bearing elements
strengthening is application of additional external non-prestressed reinforcement
into chases in masonry on intrados of vaults, which will provide stiffening and
increasing of load carrying capacity of the individual load-bearing elements.
The existing and especially historical masonry structures are nowadays
considerably monitored. Many of them are in need of some retrofitting or
strengthening. In such cases the non-metallic with minimal requirements for
reinforcement cover even in aggressive environment could be the best solution.
Therefore some tests were undertaken to learn about behavior of masonry vaults
additionally strengthened with GFRP bars. These test logically followed previous
research of the additionally strengthened masonry structures.
To achieve good usable results it is necessary to provide also statistical evaluation
and theoretical backgrounds for further designs of such structures. Therefore all
data obtained from the tests are used to create and verify the numerical model of
370 D. Horak, M. Zlamal, P. Danek
FRP reinforcement material used in calculations. This model should allow to

predict as accurate as possible the behavior of concrete and masonry structures
reinforced with FRP bars.
Mathematical model is created in physically non-linear FE software based on
fracture mechanics of quasi-brittle materials. Results obtained from real tests are
used as input data for all materials. It means all material characteristics for both
concrete (strengths, modulus of elasticity, fracture energy, etc.) and non-metallic
bars (tensile strength, load-deflection diagram, modulus of elasticity). The cohesion
between reinforcement and concrete (grouting) is modeled via cohesion parameters
for each type of the surfacing.
Comparison of the real and numerical results shows very good correspondence
(some results are shown in the text below and in [2] and [3]).
1.1. Concrete reinforcement
Tests are performed in several partial fields:

• obtaining physical-mechanical characteristics of reinforcement,
• obtaining cohesion between reinforcement and concrete,
• monitoring behavior of specimens reinforced with non-metallic reinforcement
(i.e. real function of reinforcement in loaded construction).
The first two research points were completed and all the results were analysed [2].
Choice of the most suitable type of reinforcement was achieved based on obtained
results. The best cohesion with the concrete, material properties and demand factor
of the production of the reinforcement and the surface preparation were confronted.
All these parameters influence the price and the efficiency of the developed
reinforcement.
After the decision about the surfacing of the reinforcement it was necessary to
confirm the functionality of the reinforcement. Therefore several tests were
performed on the concrete specimens. GFRP bars were used as both longitudinal
and shear type of reinforcement.
1.2. Strengthening of the masonry structures
The method of additionally inserted non-prestressed reinforcement allows

additional strengthening of masonry structures without a necessity of large
intervention into vaults especially in case of external application. This system is
capable redistributing newly originated stresses from load that act on a
strengthened construction. The aim of reinforcement is to restrict the development
of existing cracks and eliminate possibly an origin of the new ones, and to improve
load-bearing capacity of vaulted masonry constructions.
From the static viewpoint, unreinforced masonry structure is unable to transfer

tensile forces that can originate on existing structure from following action:
• action of the higher imposed load against the designed one,
• action of either identical or the lower load against the designed one.
Another consequence of the retrofit reinforcement application into masonry
structures is the rigidity improvement. The effect is evident especially at the
structures cracked by the previous traffic utilisation. Nevertheless, from the
practical viewpoint this consequence could be smaller for railway bridges.
For reinforcing the masonry structures it were used two types of reinforcing
materials (shape of this reinforcement bars can be seen in Figure 2):
• commonly used steel reinforcement (Helifix),
• non-metallic reinforcement (GFRP bars).
Figure 2. Shape of Helibar and wrapped surface GFRP
As a binding (transferring) medium between reinforcement and origin masonry was

used special mortar (grouting substance). It is important to mention that it is
essential the reinforcing bars compose with grouting substance and with origin
masonry the reliable and durable system.
2. CONCRETE MEMBERS WITH LONGITUDINAL GFRP

REINFORCEMENT
These tests are related to concrete beams (dimensions 350 x 100 x 2200 mm)
reinforced only with longitudinal GFRP reinforcement (diameter 14 mm, one-side-
wrapped bars).
This test was classical four-point bending test (Figure 3). The span of the beam was
2.2 m and the loading forces were applied at 1/3 of the length of the beam. Beams
were designed to obtain failure caused by a bending moment. During the
experiment following input data were monitored – force load, deflection on several
points and strain of the reinforcing bars (monitoring units build into the
reinforcement [2]).
Three specimens with the longitudinal reinforcement were exposed to load forces.
Also three specimens without the reinforcement were loaded to provide reference
data and to make possible the comparison of effects of the reinforcement. Results
of specimens without reinforcement allowed also validating the input data (i.e.
material model of the concrete) used in FEM numerical model.
Figure 3. Deflection of the loaded beam before the collapse
All three reinforced beams collapsed because of exceeding the tensile strength of
the GFRP bars. Two of them collapsed under the load force, one beam collapsed in
the middle. Maximal average load carrying capacity of this beam improved from
total 6.11 kN (calculation presumption 6.09 kN) to 17.19 kN (calculation
presumption 16.38 kN - according to ACI 440.1R-03 without any safety factors).
The tests results demonstrated the functionality of the developed non-metallic
reinforcement.
The development of the load-deflection curve is in Figure 4. Also comparison of
the behavior between reinforced beams (specimen 1-3), non-reinforced beams
(reference specimens 1-3) and numerical model of the reinforced beam (Atena 3D
results) can be found there.
Figure 4. Deflection of the beam with longitudinal GFRP reinforcement

2. CONCRETE MEMBERS WITH LONGITUDINAL AND SHEAR

GFRP REINFORCEMENT
The non-metallic reinforcement was tested also as shear reinforcement. The

longitudinal reinforcement in these beams were the GFRP bars (diameter 14 mm,
one-side-wrapped bars – same as mentioned before). The shear reinforcement was
created from one GFRP bar (diameter 8 mm, one-side-wrapped bars) shaped into
spiral looped around all longitudinal reinforcement bars – see Figure 5.
This reinforcement had to be shaped before hardening. Thus the curing method was
changed and the curing of the already shaped and fixed reinforcement proceeded at
the room temperature. The hardening of the bar took more time, but the material
characteristics were not reduced.
Figure 5. Shear reinforcement shaped into spiral
All beams (dimensions 115 x 240 x 2100 mm) were loaded by the same way as the
beams with longitudinal reinforcement only. It means the load scheme was
classical four-point bending test with load points at 1/3 of the span. Supposed
failure mode was exceeding the shear capacity in the area near supports.
Again the test set was made from three test specimens with shear reinforcement
and from three “reference” specimens without shear reinforcement. The
reinforcement influenced positively the shear capacity of the tested beam and
confirmed its functionality. The shear capacity improved from 54.7 kN (reference
specimens without shear reinforcement) to 82.2 kN (reinforced specimens).
Figure 6. Cracked beam with GFRP shear reinforcement
Figure 7. Deflection of the Beam with GFRP Shear Reinforcement

(REF Specimens are Reference Specimens Without Shear Reinforcement)
The tests results demonstrated the functionality of the developed non-metallic

reinforcement.
3. MASONRY STRUCTURES STRENGTHENED WITH GFRP

BARS
Within experimental parts of the project three sets of masonry vaults with for
various loading types were manufactured. For the distinction of individual vaults
are used notation jKi, where „j” corresponds to series number (1-3) and „i” to the
strengthening method (1-3). The vaults were symmetrically loaded in ½ of the span
- 1.series (j=1), asymmetrically in ¼ of the span - 2.series and symmetrically in

both quarters of the span - 3.series (j=3).
Figure 8. Loading scheme of the vaults and distribution zones of the load in the vaults
Each series consists of three vaults: non-strengthened – comparative (i=1), a vault

reinforced in two chases (i=2) and a vault reinforced in three chases (i=3). The
vaults were bricked up from full burnt bricks on lime-cement mortar of the width
890 mm, span 2600 mm, deflection 750 mm and radius 1500 mm.
Into every reinforcing chases were embedded 2 bars. First experiments were
performed with reinforcement HeliBar of special helical shape of diameter 8 mm
and the second set of test specimens were reinforced with GFRP bars of diameter 6
mm (one-side wrapped). Only unsymmetrical loading in ¼ of the span was used for
testing vaults with non-metallic reinforcement (it is the case of the biggest
influence of the additional strengthening [4]).
3.1. Behavior under static load
From the comparison of the load-bearing capacity of the individual vaults in the
series results that essential growth of the load-bearing capacity was achieved
especially in the case of 1st series and 2nd series of the vaults, namely more than
eight multiple growth. This growing of carrying-capacity can be watch for both
cases of reinforcement – helical metallic and non-metallic. It was related to the
vaults stressed by either concentrated or one-sided load, at which the vaults were
loaded by the interaction of normal forces and bending moments.
In the case of 3rd series the experiments did not prove the effects of strengthening
by additionally inserted reinforcement on the vaults load-bearing capacity; no
effects of reinforcement demonstrated themselves because the vaults were mainly
compressed.
In the case of non-strengthened vaults of 1st and 2nd series the failure was acute,
main crack was opened and the vault ruptured. In the case of the strengthened
vaults of 1st and 2nd series came to the gradual opening of separate cracks until the
failure, which was accompanied by the rapture of the metallic reinforcement from
the chases. All glass reinforcing bars were in the ultimate limit state ruptured.
45
FORCE F [KN]
40
35
30
25
20
15
10
0
-50 -40 -30 -20 -10 0 10 20 30 40 50
GLASS - HELIFIX - DEFORMATION [MM]
S 1/4 without load [mm] - 2K1 - glass S 1/4 under load [mm] - 2K1 - glass
S 1/4 without load [mm] - 2K1 - helifix S 1/4 under load [mm] - 2K1 - helifix
Figure 9. Comparison of deformations on vaults loaded in ¼ of the span strengthened with

GFRP and metallic helical reinforcement
On the basis of thus obtained results from numerical studies and on the base of the
designed algorithm, it will be possible to obtain (substantiate) simple constitutive
relations for the evaluation and design of strengthening by simplified designed
methods used in the practice and to set up simple algorithms for design and
checking calculation of the masonry vaulted construction with additional
reinforcement for the practice.
3.2. Behavior under dynamic load
Dynamical tests were performed on vaults loaded asymmetrically in 1/4 of the span
and reinforced with glass reinforcement (GFRP) only. From results of first
dynamical tests it is visible increasing of load-bearing capacity of the reinforced
vaults (2K2, 2K3) compared to the non-reinforced vault (2K1) (Figure 10).
20
FORCE F [KN]
18
16
14
12
10
0
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
DEFORMATION [MM]
S 1/4 without load [mm] - 2K1 - glass S 1/4 with load [mm] - 2K1 - glass
Síla def
Figure 10. Comparison of deformations on vaults loaded in ¼ of the span strengthened with
GFRP reinforcement – dynamical test
Unfortunately the low set of tested specimen prohibited comparison with the test
data from statical experiments. The results are also influenced by the big non-
homogeneity of masonry structure. Also the fracture mode (i.e. failure of the vault
by opening of tension cracks in the bed joint) is not uniform and the position of
cracks can influence the final load bearing capacity.
Strengthened vaults can be partially compared by relation of their load-bearing

capacity. Ratio of the load-bearing capacity of the vaults with three reinforcing
chases and with two chases (2K3/2K2) at statical loading is 1,33 and at dynamical
loading is 1,29.
3. CONCLUSION
The tests showed that the developed system is functional. The reinforcement bars
can work as concrete reinforcement and they are capable to transfer the load forces
generated in the construction during the loading.
There is also very positive benefit for the strengthening of the masonry vaults. This
system can be used to repair the historical structures with minimal impact to the
structure itself (thanks to low requirements for cover – there is no need to provide
additional layers of cover materilas).
So far only short term tests were performed. To fully confirm the functionality and
safety of the newly developed reinforcing system it will be necessary to verify the
long term characteristics of the reinforcement. The main subjects of research will
be the behavior of the reinforced structures under the long term loads. Also the
resistibility of the reinforced structures in the aggressive environment has to be
verified.
Acknowledgements
This research has been proceeded with support of Czech ministry of industry and
trade (MPO) in frame research task 1H-PK2/57 – “Durable concrete structures”
and also has been prepared with the financial support of Ministry of Education,
Youth and Sports, project No. MSM0021630519 “Progressive reliable and durable
load bearing structures”.
References
1. Plsek, J., Stepanek, P., Optimisation of design of croos-section in concrete structures, Procceding
of the 4th International Conference Concrete and Concrete Structures, 2005.
2. Stepanek, P., Fojtl, J., Horak, D., Pull-out test of non-metallic FRP reinforcement for concrete
structures, Proceedings of the 4th International Specialty Conference on Fibre Reinforced
Materials, 2006, Hong Kong.
3. Horak, D., Krupa, P., Stepanek P., Prokes, J., Development of the new methods of measurement
of the strain in GFRP concrete reinforcement, Worshop NDT 2006 – Non-destructive testing in
engineering practice, 2006, Brno, Czech Republic.
4. Stepanek, P., Zlamal, M.: Additional strengthening of masonry vaults with non-prestressed
additional reinforcement, Proceeding “STRUCTURAL FAULTS + REPAIR-2006”, 2006,
Edinburgh, Scotland.
New Shear Connectors for Composite Girders

– Experiences with ABAQUS Push-Out Test Simulations
Josef Fink 1, Lubomir Ondris 2
Institute of Steel Structures, TU Vienna, Karlsplatz 13/212, A-1040 Vienna, Austria
Summary
Composite structural members consisting of steel and concrete parts utilize the
advantages and suppress the disadvantages of individual material components. The
shear loading between steel and concrete parts is transferred by shear connectors.
They have to be designed to maximize the shear load capacity and ductility and to
minimize the overall manufacturing costs.
At the Technical University of Vienna, Institute of Steel Constructions, new shapes
of shear connectors for composite steel and concrete beam structures are being
developed. The composite structure consists of steel I-beam with connector welded
lengthwise in the middle of the flange and surrounded by a concrete reinforced
slab. The first sheet-type crown shaped connector has already been accepted for
patent procedure. To verify its efficiency both numerical analyses and physical
experiments with push-out tests were carried out. For numerical analyses of
composite structures the well-established FE-program ABAQUS is used. Results
depend on many material and control parameters therefore first an extensive
parametric study was necessary /1/. Here missing information is given, originally
suppressed in /1/ because of patent protection procedure. Further numerical
experiences with different loading velocities, with ABAQUS option NLGEOM and
experiences relating to results evaluation are described. Different results of
ABAQUS versions 653 and 662 are mentioned. Necessity of a very careful
preparation of numerical simulation and of using realistic material data for
concrete is emphasized.
A comparison in /2/ shows a very good matching between physical tests and
numerical simulation.
KEYWORDS: Composite steel and concrete beam structure, welded shear

connector, push-out test, finite element modeling, ABAQUS, material model for
concrete, loading velocity, geometric nonlinearity, kinetic energy, total strain
energy, new shape of shear connector.
1. INTRODUCTION
380 J. Fink, L. Ondris
Developing of new sheet-type connectors at the Technical University of Vienna,

Institute of Steel Structures consists of physical tests with push-out specimens and
of their numerical simulation as well. A recent extensive parametric study /1/ of
options given in the known FE-program ABAQUS /3-5/ was concerned with a new
sheet-type crown shaped connector. After accepting it by European Patent Office in
Vienna for patent procedure under proposed name "VieCro", the information
originally suppressed in /1/ can here be given for completeness.
The welded steel parts of the specimen can be seen in Fig. 1. In Fig. 2 the complete
specimen with reinforced concrete slabs is ready for testing in a horizontal loading
machine. In Fig. 3 the basic parts of the FE-model utilizing twofold symmetry are
shown. Because of their very high stiffness, the head plate and the I-beam could be
removed from the FE-model and replaced by corresponding boundary conditions.
The loading was simulated by prescribing boundary condition on the lower surface
of the steel connector (welded to the I-beam) in longitudinal direction. ABAQUS
static and dynamic nonlinear procedures have been invoked, using quadratic FE-
elements C3D20R and linear FE-elements C3D8R, respectively.
In /1/ results of 123 FE-calculations are given in form of P-d diagrams, d being the
actual value of the prescribed connector displacement, P being the corresponding
total reaction force in supported right side surface in Fig. 4. In diagrams the in-
fluence of ABAQUS material parameters, of friction coefficient in frontal and la-
teral contact surfaces and of clamped or contact support while using static and dy-
namic solution procedures can be observed.
2. DIFFICULTIES WITH NUMERICAL SIMULATIONS
The difficulties are caused by highly nonlinear character of the problem and by
missing specific material parameters for concrete. If there is no P-d diagram from
physical test and/or if the material data are not reliable then it is not easy to decide
whether the calculated peak load, post-peak response and failure load are "true" i.e.
whether they correspond to the parameters chosen only or whether they are influen-
ced also by potential "internal" numerical instabilities.
In a recent EU-Project FE-packages ABAQUS, ANSYS, DIANA, GEFDYN and
LUSAS used for solution of concrete dams have been compared /6/. It has been
confirmed that the result of a complex highly nonlinear problem often depends on
nonlinear control parameters used. According to /6/, correct estimating of the
failure load using a non-convergence stopping criterion is problematic because
with a "wrong" criterion the non-convergence on relatively small loads can occur.
The quotation marks signalize some uncertainty always present while solving
complicated nonlinear problems: The result depends on the solution path, this
again depends on a set of parameters used. Every change in parameters used means
a different solution path, sometimes with surprising effects.
Because of problems described a very careful preparation of solution strategy and
mesh data and a very thorough results evaluation is necessary. Even then it is
always useful to carry out more calculations with slightly changed parameters to
check the result sensitivity.
Throughout the study /1/ emphasis has been placed on the highest possible
accuracy and reliability of the system answer rather than on efficiency of
calculations. As far as possible, for the most control and material parameters
ABAQUS default values were used uniformly throughout all calculations.
With static calculations maximum number of iterations as a general stopping
criterion was set sufficiently high (and never reached), for starting and minimum
increments very small values were set. With dynamic calculations the ABAQUS
default automatic incrementation with global stable increment estimator was used.
3. SOLUTION PROCEDURES
From ABAQUS nonlinear procedures tested in /1/ (*STATIC,RIKS,

*STATIC,STABILIZE, *DYNAMIC,EXPLICIT) the last one seems to be the
most suitable for simulation of physical push-out tests. The reasons why just an
explicit integration package like ABAQUS/Explicit should be used for solving
highly nonlinear problems are convincingly given in /7/. Practical comparison of
results obtained by particular procedures is given in /1/.
4. FE-MODEL FOR DYNAMIC ANALYSES
With the *DYNAMIC,EXPLICIT procedure of ABAQUS/Explicit package only

linear elements can be used. However, the FE-model has to take into account
material and contact nonlinearities and the geometry of the connection.
The authors prefer a rather conservative solution with some "FE-overkill". The
mesh for dynamic analyses (Fig. 4) consists of following parts: steel connector:
6474 linear 8-node hexahedrons C3D8R (3 elements crosswise on the half-
thickness of 10 mm), concrete slab: 30495 elements C3D8R, reinforcement: 525
elements T3D2, each anchor rod: 1 element T3D2. This mesh is able to register
high stress gradients occurring in interactions between sheet connector and
concrete parts near connector.
5. MATERIAL MODELS
ABAQUS material models used are completely described in /1/, thus the descrip-
tion will not be repeated here. It is to emphasize that for concrete parts a character-
ristic material curve with loss of strength after reaching the tensile strength has
been used instead of artificial material features.
For steel parts an elastic-ideal plastic model without hardening was used.
6. FRICTION IN CONTACT SURFACES
In /1/ each calculation has been carried out three times with friction coefficients
0.3, 0.5, 0.7. For simplicity, between frontal surfaces (Fig. 5) the same friction
coefficient was used as between lateral surfaces (Fig. 6). Generally, P-d curve from
physical experiment, especially in post-peak area, can better be approximated in
numerical simulation using small friction coefficient. Calculations in /8/ confirm
this conclusion and analyse the contribution of frontal and lateral friction.
7. RESULTS EVALUATION – ENERGIES AND VISUAL

INSPECTION
In dynamic calculations used for simulation of quasi-static processes an important

role plays the ratio of kinetic energy to internal energy ALLKE/ALLIE with a peak
value accompanying distortion. ABAQUS-recommendation that ALLKE/ALLIE
should not be over 0,05-0,1 has to be used with some caution. The ultimate value
can be different if the distortion is localized in a small area and the energy output in
history output request is required for the whole FE-model. This is the way how the
ratio ALLKE/ALLIE is judged usually, at least in first calculations. In all
calculations described in /1/ and here energies were output for the whole model.
It turns out that in the curve ALLKE/ALLIE-d sometimes the existence of the peak
regardless of its magnitude determines the distortion process. Usually, the peaks
are very narrow and disappear soon. Often they can not be observed because of
such a trivial reason as too big distance between output points. Also in such cases a
careful visual inspection of deformed mesh using deformation scale factor > 1 can
detect beginning of distortions or numerical difficulties since the accompanied
oversized unnatural displacements do not disappear.
8. LOADING VELOCITY
Dynamic explicit calculations in /1/ have been carried out conservatively with a
very small constant velocity of v=0.5 mm/s (prescribed connector displacement
linear increasing with time). Using higher velocities can shorten the computer
times substantially, on the other side inertial forces can arise and influence the
result. Therefore it would be useful to find the upper bound of admissible velocities
for quasi-static calculations of structures in question.
First, push-out test simulations based on an anvil connector shape (red in Fig. 7)
and utilizing two planes of symmetry have been carried out. ABAQUS concrete
material model CDP with dilation angle ψ=30o, friction coefficient f=0.5 between
interaction surfaces, clamped support surface and loading in the same way as in /1/
but using five constant loading velocities v were used. The corresponding Pmax-v
diagram (Fig. 8) was however unacceptable, contradicting expectations. By
detailed inspection of particular solutions it has been found out that, in contrary to
/1/, the pre-peak response in P-d diagrams is characterized by several irregularities
(Fig. 9). For example, upon velocity v=1 mm/s, the first peak value at d=0.5 mm
(blue in Fig. 9) is followed by a downcome caused by local numerical problems
which can clearly be seen in Fig. 7 (excessive local displacements; deformation
scale factor: 10) and on ALLKE/ALLIE-d curve in Fig. 10. Just the curve for v=2
mm/s could be accepted for estimating Pmax as a measure for inertial effects.
Differences in Pmax and irregularities are not caused by inertial forces directly but
by the extreme numerical instability of the solution: Already a negligible parameter
change (in this case the loading velocity v) effects on the solution path. Thus,
simulations based on an anvil connector shape could not be used to identify the
range of admissible velocities.
After this experience the influence of the loading velocity has been examined using
the simplest possible example. The test beam 100x100x200 mm consisting of con-
crete only (without any contact interactions and steel parts) was clamped on one
end and loaded by prescribed displacement of -25 mm on the opposite surface.
Again, linear increasing displacement was prescribed. Following constant
velocities were used: v=0 (static solution), 0.5, 5, 25, 100, 500, 1000 mm/s.
ABAQUS CDP material model and C3D8R elements were used. Because of
twofold symmetry only 1/4 of the beam has been meshed. In Fig. 11 the final de-
formed shape with Mises stress upon v=5 mm/s can be seen. The corresponding
Pmax-v diagram in Fig. 12 conforms to expectations. In Fig. 13 the pre-peak res-
ponse up to v=5 mm/s is free of irregularities (diagram parts with d>2 mm do not
contain any substantial information and are omitted). Clear visible is the force
delay upon v=100, 500 and 1000 mm/s due to the distance between the surface on
the left with applied displacement d and the clamped surface on the right with the
reaction force P. Thus, upon higher velocities the solution is ever more dominated
by inertial effects unacceptable in a quasi-static structural response.

ALLKE/ALLIE values in Fig. 14 confirm the conclusion that in this type of prob-
lems the constant loading velocity should not exceed, let's say, just to be sure, 10
mm/s.
However, also the difference between ABAQUS procedures is to be noticed. With
the procedure *STATIC (v=0) the peak load of Pmax=115.2 kN is slightly higher
than with the procedure *DYNAMIC,EXPLICIT (v=5 mm/s: Pmax=109.9 kN),
both being used with NLGEOM=YES..
At last, push-out test simulations based on a crown shaped connector (in /1/ loaded
by v=0.5 mm/s) have been recalculated using additional loading velocities v=5, 25,
100, 1000 mm/s. As in /1/, support with contact interaction, concrete material CDP
M2, dilation angle ψ=36.31, friction coefficient f=0.3 and NLGEOM=YES were
used.
Corresponding P-d curves can be seen in Fig. 15 (diagram parts with d>10 mm do
not contain any substantial information and are omitted). The P-d curve for v=0.5
mm/s was already given in Fig. 27 of /1/.
Generally, P-d curves for support with contact interaction are characterized by
oscillations, the same curves for clamped support are smooth (Fig. 28 in /1/). In
Fig. 15 only the curves v=0.5 and v=5 mm/s are acceptable. The bifurcation of P-
values after reaching peak area could not be explained yet. The reason is probably
again an extreme numerical instability of the solution process.
With v=25 mm/s and more, oscillations become dominant and influenced by
inertial effects. Also the force delay is clear visible. The ALLKE/ALLIE values in
Fig. 16 are acceptable for v=0.5 and v=5 mm/s only.
In Fig. 17 Mises stress is shown to demonstrate the influence of inertial effects.
Upon the velocity v=1000 mm/s the lower red part of the steel connector moving to
the right is already upon d=0.76 mm (ideally) plastic, whereas the support reaction
force on the specimen right side still equals zero. The yielding area in this case is
not originating around contact points with concrete as usually but due to prescribed
fast growing displacement on lower connector surface.
A general recommendation concerning the highest admissible loading velocity for
numerical simulation of push-out tests can not be given. It is not easy to distinguish
whether the difference in peak values and/or calculation break were really caused
by inertial effects or by taking a different solution path. The critical velocity value
depends on many parameters, e.g. connector shape, mesh quality, boundary
conditions used, way how the loading is applied, and so on. Therefore, not only P-d
curves should be compared but also ALLKE/ALLIE values have to be checked and
the deformed mesh should be carefully visually inspected throughout the solution.
At least two calculations with different loading velocities are to be carried out. If
the results differ substantially, the reasons have to be investigated in more detail.
9. USING NLGEOM AND DOUBLE PRECISION
This option on the data card *STEP controls the inclusion of nonlinear effects of
large displacements. With static procedures in ABAQUS/Standard default setting is
NLGEOM=NO, with the procedure *DYNAMIC,EXPLICIT in ABAQUS/Explicit
default setting is NLGEOM=YES.
In /1/ all calculations have been carried out taking into account the nonlinear effect
of large displacements. In dynamic calculations in post-peak response this has led
to very long calculations offering at last nonrealistic distorted FE-mesh by far not
corresponding to the brittle behavior of concrete especially in situations with no or
little reinforcement. E.g. the curve v=0,5 mm/s in Fig. 15 ends at d=23,6 mm, but
between d=11,8 mm (expressive ALLKE/ALLIE-peak) and d=23,6 mm
(calculation collapse) the structure in Fig. 18 (suppressed Fig. 30 in /1/) is
nonrealistic distorted and the calculation does not give any real information.
Therefore, using NLGEOM=NO can be a reasonable alternative generally, not only
if BRITTLE material model is used.
Generally, always DOUBLE PRECISION solution should be used. Upon SINGLE
PRECISION the solution process aborts prematurely and using NLGEOM=YES
does not change the situation substantially.
10. ABAQUS VERSIONS 653 AND 662
For calculations in /1/ ABAQUS version 653 was used. At TU Vienna shortly a
new computer system came into service using ABAQUS version 662. Calculations
on both systems with the same input file showed in P-d diagram considerable
differences in post-peak area. The reason could not be explained yet.
11. CONCLUSIONS
The basic question is which goals are to be fulfilled by numerical simulation of

physical push-out tests. A pure matching of the P-d curve from an existing physical
test can be achieved also by using a coarse FE-mesh, inadequate material model
"sometimes without regard to the reasonableness of the material input variables" /9/
and controversial calculation parameters. Such a simulation, of course, keeps the
simulation costs minimal. However, it proves rather the versatility of the FE-
program and the user's inventiveness than an accurate representation of the
physical behavior itself.
A consequent numerical simulation should not only match the data record from a
foregoing physical test. It also should be able to reduce the number of physical
tests in the course of selecting more effective connector shapes. This goal can be
achieved only by a detailed visual inspection of analysis results get by a reliable
FE-calculation. This has to be based on a sufficiently fine mesh (one linear finite
element crosswise the modeled steel connector half-thickness is not sufficient to
follow stress gradients), on a realistic material model (no "ideal plastic concrete" in
tension), on a tolerable loading velocity (for safe excluding of additional inertial
forces), etc. In case of doubts the influence of an unknown (e.g. material)
parameter has to be checked by changing its value and repeating the calculation.
In near future the authors will continue developing of new connector shapes
accompanied by exploring the influence of reinforcement location on properties of
specific connector shape in composite steel and concrete beam structures.
Acknowledgment
The corresponding author thanks to Wojciech Lorenc for valuable and inspiring
discussions /10/. Other research colleagues are encouraged to join and exchange
experiences, too.
References
1. Fink J., Petraschek Th., Ondris L.: Push-Out Test Parametric Simulation Study of a New Sheet-
Type Shear Connector, in: Projekte an den zentralen Applikationsservern, Berichte 2006,
Zentraler Informatikdienst (ZID) der Technischen Universität Wien, Wien 2007, pp. 131-153,
also: http://www.zid.tuwien.ac.at/projekte
2. Fink J., Petraschek Th., Ondris L.: Neuartige Schubverbinderleisten – experimentelle Untersu-
chungen und numerische Vergleichsbetrachtungen, ÖIAZ, accepted for publishing 2007
3. Hibbitt, Karlsson & Sorensen, Inc.: ABAQUS/CAE, ABAQUS/STANDARD,
ABAQUS/EXPLICIT, Dokumentation Ver. 6.5.3, http://www.abaqus.com
4. ABAQUS Verification Manual, 2.2.24
5. ABAQUS Example Problems Manual, 2.1.15
6. Jefferson A.D., Bennett T. and Hee S.C.: Fracture Mechanics Based Problems for the Analysis of
Dam Concrete, NW-IALAD-Final Technical Report-Task Group 2.4, Cardiff University (UWC),
15. March 2005
7. Zimmermann S.: Finite Elemente und ihre Anwendung auf physikalisch und geometrisch nicht-
lineare Probleme, Report TUE-BCO 01.05, Technische Universität Eindhoven, Niederlande,
März 2001
8. Iwancsics M.: Numerische Untersuchungen zu neuartigen Schubverbindern, Diplomarbeit,
Institute of Steel Structures, TU Vienna 2007
9. Evaluation of LS-DYNA Soil Material Model 147, Report No. FHWA-HRT-04-094, November
2004, Federal Highway Administration, 6300 Georgetown Pike, McLean, VA 22101-2296
10. Lorenc W., Institute of Building Engineering, Wroclaw University of Technology:
Private Communication 2006-2007
Fig. 1 Push-out specimen's welded steel parts
Fig. 3 Basic parts of the FE-model with twofold symmetry

Fig. 2 Complete push-out-specimen ready for testing
Fig. 4 FE-mesh for dynamic analyses incl. reinforcement, contact support on the right
Fig. 5 Frontal master contact surface of the steel connector
Fig. 6 Lateral master contact surface of the steel connector

Fig. 7 FE-simulation with anvil shape connector, v= 1 mm/s, d=0.625 mm
600
500
Pm ax kN
400
300
200
0 0,5 1 1,5 2
v mm/s
Fig. 8 Pmax – v curve, anvil shape connector

600
500
400
Load P /kN/
300
200
v=0,125 mm/s
0,25
0,5
100
1
2
0
0 0,5 1 1,5 2 2,5 3
Displacement d /mm/
Fig. 9 P-d curves, anvil shape connector
0,005
0,004
ALLKE / ALLIE
0,003
0,002
0,001
0
0 0,5 1 1,5 2 2,5 3
Displacement d /mm/
Fig. 10 ALLKE/ALLIE-d curve, anvil shape connector, v=1 mm/s

Fig. 11 Simple concrete beam with displacement controlled loading
600
500
400
kN
300
P m ax
200
100
0
0 100 200 300 400 500 600 700 800 900 1000
v mm/s
Fig. 12 Pmax – v curve, simple concrete beam

125
100
Load P /kN/
75
v=0 (static)
50 0,5 mm/s
5
25
25 100
500
1000
0
0 0,5 1 1,5 2
Displacement d /mm/
Fig. 13 P-d curves, simple concrete beam
0,8
ALLKE / ALLIE
0,6
0,4 v=0,5 mm/s

5
25
0,2 100
500
1000
0
0 0,5 1 1,5 2
Displacement d /mm/
Fig. 14 ALLKE/ALLIE-d curves, simple concrete beam

600
500
400
Load P /kN/
300
200 v=0,5 mm/s

5
100 25
100
1000
0
0 2 4 6 8 10
Displacement d /mm/
Fig. 15 P-d curves, crown shaped connector /1/, NLGEOM=YES
0,8
A L L K E / A L L IE
0,6
0,4
v=0,5 mm/s
v=5
0,2 v=25
v=100
v=1000
0
0 2 4 6 8 10
Displacement d /mm/
Fig. 16 ALLKE/ALLIE-d curves, crown shaped connector /1/, NLGEOM=YES

Fig. 17 Crown shaped connector /1/, NLGEOM=YES, v=1000 mm/s, d=0.76 mm
Fig. 18 Crown shaped connector /1/, NLGEOM=YES, v=0,5 mm/s, d=11,8 mm

The Structural Expertise of Steel Cables

Ludovic G. Kopenetz, Ferdinand-Zsongor Gobesz
Department of Structural Mechanics, Faculty of Civil Engineering and Building Services, Technical
University of Cluj-Napoca, Cluj-Napoca, 400020, Romania
Summary
Considering the fact, that steel cables are structural elements forming statically
determinated systems, their rupture can lead to catastrophe. In this context, the
structural assessment of steel cables represents a primary necessity, generated by
the requirement to assure security and safety in use.
Generally the main causes which induce degradation/deterioration of cables are:
fatigue and corrosion.
The effects of corrosion and fatigue are displayed usually through fiber laceration
sequent to a certain service time, after which the number of ruptures and
lacerations increases exponentially.
This paper covers some problems of structural expertise along with numerical
simulation aspects of corrosion and fatigue, as well as a methodology for the
deduction of the presumed service life of steel cables.
KEYWORDS: steel cable, structural expertise, corrosion, fatigue.
1. INTRODUCTION
The irrefutable qualities of bearing cable structures justify, from one part the
extension of their application area, from another part the remarkable effort for
world-over research concerning a better insight and knowledge about their
behavior under corrosion and fatigue, envisioning a safer and more judicious
design.
The aim of the authors through the submission of this paper is to present some
aspects of their research done by the Faculty of Civil Engineering and Building
equipments from the Technical University of Cluj-Napoca.
Since the middle of 70’ the pursuit of this kind of research was focused in the
following directions:
• Tests carried out in situ (generally considering only statical behaviour, just
infrequently considering dynamical too), including the exact quantification of
cable structure geometry based on survey techniques. Due to the fact that
those kinds of structures are located usually at considerable heights, the
applicable survey techniques have a dynamic character, recording also the
swing and vibration of the structure and the movements induced by dynamic
loads (wind, functioning equipment, traffic).
• Laboratory tests, in order to identify the structural materials through physico-
mechanical and chemical analysis.
• Statical and dynamical analysis and calculus, including the evaluation of the
service time for every structural component considering possible future loads
and actions over the operative period.
It has to be pointed out the complex and time consuming character of any credible
structural expertise in this field, considering the vastity of involved factors which
are combined with the unstationary character of the loads (wind, temperature,
vibrations from equipment and traffic etc.) [9], [10], [11].
Figure 1. The Agigea Bridge from Romania

2. CABLES AND TIE-BACKS IN MEMBRANOUS CABLE

STRUCTURES
Common materials used in cable fabrication during history were: papyrus, camel
hair, flax and hemp, until in 1834 when the first cables and ropes from steel wires
were made. This new building material soon became indispensable in many fields,
due to its special properties – high breaking strain compared to its self weight,
great flexibility and durability [17]. In the field of construction, cables were used
initially as bearing parts for suspended bridges, and much more lately for covering
large areas without intermediate holders [18].
In the last period, in highly corrosive environments, cables made from
polypropylene (specific cable weight / specific water weight = 0.91), polyester and
nylon (specific cable weight / specific water weight = 1.14) are used.
2.1. Steel qualities and brands for cables and tie-backs [19], [20]
Cables and tie-backs are made from high- and very high-grade steel, with an
average carbon content of 0.5% and a breaking strain around 60 daN/mm2.
Considerable growth of mechanical strength can be obtained through repeated
deformations applied on steel rods during the fabrication process of wires. Thus, a
cylindrical steel bar is transformed on the drawbench in wire, while its breaking
strain rises up to 120 – 200 daN/mm2. After that, the wire is subdued to a thermal
treatment and hereby the material regains its plastic properties. The wire yarns are
entwisted on a central wire, in one or more layers, composing strands. At their turn,
the strands are coiled around a central core, forming the cable.
Nowadays in Romania two types of steel are used in the fabrication of wires which
can be embodied in cables:
- carbon steel with 0.6 – 0.9% C and 0.3 – 0.7% Mn,
- thin alloyed steel, usually with manganese and silicon.
From carbon steel are wires with smooth (SBP) or marked (SBPA) surface made,
each type in two qualities (I and II). From thin alloyed steel, high strength rods
(PC90) are produced, with geometrical, chemical, mechanical and technological
properties prescribed by the STAS438/1-74 standard.
The semi-product which is used in the fabrication of strands and cables is the
proprietary carbon steel wire (through initial thermal treatment the steel is heated
up to 880 – 930 °C, followed by a quick cooling to 450 – 500 °C in a lead bath
after which the cooling continues slowly in the air) and drawn SBP type wire (the
laminated wires are forced on a drawbench through a smaller hole than the actual
diameter of the wire) without final annealing treatment.
The cables can be classified upon several criteria:
• Classification based on shape: The cables can be flat or round shaped. At their
turn, round shaped cables can be simple (made from a single strand), double
(composed from several strands) or coupled (formed by wrapping double
cables around a central core).
• Classification according to the number of strands: Cables can be made from 1,
6, 8, 18 or 36 strands.
• Classification upon the core material: The cable core can be manufactured
from vegetal, mineral, metallic or synthetic wires.
• Classification after the quality of fibers: Steel cables can be produced from
uncoated (mat) or zinc coated wires.
• Classification upon the laying of strands: The wire yarns can be coiled in a
strand towards right (Z) or left (S). At their turn, the strands can be wrapped
within a cable towards right (Z) or left (S).
2.2. Bearing structures with cables
From the point of view of structural analysis, bearing structures with cables can be
divided in the following two categories:
- isolated cables,
- cable nets and suspended structures.
Considering these structures, cables can be:
- isolated parallel or twisted wires,
- fascicles of stranded wires,
- ropes,
- thin steel rods,
- ribbons,
- chains,
arranged in one direction or in different directions (Fig. 2).
Figure 2. Different cable types
A synthetically exhibition of bearing structures with cables is presented in fig. 3.

Figure 3. A classification of bearing structures with cables.

3. EXAMINATION FACETS FOR STRUCTURAL CABLES
In case of structural cables the problems which are generating noticeable

modifications of the structural safety can be grouped in the following classes:
• Problems concerning the quality of the constitutive material (cold flow, brittle
or breaking in different manners etc.).
• Fatigue problems due to repeated stress.
• The problem of considerable displacements caused by static and dynamic
loads.
• Corrosion and erosion problems.
Bearing structures with cables are characterized by loads which are strongly
depending from the basic geometry of the structure, namely by the initial balanced
state (including also the geometrical and physical imperfections) induced by the
steady and working loads and by eventual prestressing.
The expertise and checking of these structures must be done by studying the
nonlinear geometrical (eventually physical) behaviour and all the factors upon
which relies the structural safety. As a decisive element rises in this context the
occurrence of corrosion and endurance to oligocyclic fatigue. The resistance to
oligocyclic fatigue will be studied from secondary stresses (from vibrations or from
daily thermal expansion – contraction) considering the intensity of the stress range.
3.1. Loads on structures with cables
The main loads applied on bearing structures with cables are arising from self
weight, wind pressure, temperature, working loads with connected dynamic effects
and support displacements.
3.1.1. The influence of weight
The following effects will be considered from dead load, combined also with other
loads and forces from different causes:
- active loads, containing self weight and snow build-up, ice (frost);
- inactive loads, containing the self weight of the structural element and other
permanent loads.
3.1.2. The influence of wind
This action will be considered in the context of KÁRMÁN vortices, combined

from case to case with the phenomena of galloping and fluttering.
3.1.3. The influence of temperature
The low environmental temperature must be considered in any cases. Structural

elements with lower working temperature than 0 °C will be additionally loaded
with ice (frost) through the condensation of the moisture from the atmosphere.
In case of cable sustaining tubular structures with closed ends, through the cooling
of the built-in gas or steam the inner pressure can drop enough to create vacuum
inside the tube. In such case these tubular structural parts must bear up to the
external pressure at low temperature.
3.1.4. The influence of dynamic effects
Buildings with structural cables are checked against the detrimal effects of
vibrations that can arise from several sources, such as:
- impact forces;
- resonance developing from the operation of equipments (including air
conditioning and ventilation appliances, loud musical gears, traffic etc.);
- seismic actions and wind.
3.1.5. The influence of thermal effects (expansion, contraction)
Thermal effects should be considered in combination with loads and forces from
other causes:
- thermal actions through constraints and restraints;
- effects due to different coefficients of thermal expansion in case of structures
with mixed materials (like steel and aluminium).
3.2. Case of laboratory-tests
The minimal bench tests which are carried out in a laboratory in order to identify
the compounding materials of bearing structures with cables are:
- axial extension test, at different velocities;
- repeated bending test;
- torsion test;
- chemical test of the base material;
- simulation of corrosion and fatigue.
3.3. Site investigations [21]
Examining the behavior in time (ageing property) of steel cables is a special assay,
imposed primary by the necessity to insure operational safety. In situ tests will
pursue:
• Diminution of cable section, namely loss of metallic cross sectional area

(LMA) due to corrosion, plastic deformation (afterflow) etc. [22].
• Modification of the shape (geometry).
• Broken wires, laceration and other local faults (LF) due to fatigue [22].
In principle the following methods are used for site investigations:
- visual inspection,
- electrochemical (potential, electrochemical sounds, magnetic etc.) methods,
- other nondestructive testing (microscopical examination, gravimetry, infrared
thermography, gammagraphy, radiography and radar processing) [23], [24],
[25].
For structural cables only a few of these methods give adequate results.
For visual inspection carried out in site, the authors are advising a new method:
MOV_CAM (currently under patenting), using sliding digital cameras in order to
view and process the image of the cable (fig. 4).
a
a
a-a
CAM2
Dist.
CAM1 Cable 120°
CAM2
CAM3 CAM3 CAM1
Figure 4. Visual investigation with digital cameras (currently under patenting).
LF LMA
Figure 5. Sample results from LF and LMA testing.
For LMA and LF type investigations electromagnetic processes are successfully

used lately, based on a magnetic head equipped with strong permanent magnets
and sensors with distance meter. The software enables to inspect and to display the
local faults (LF) and the sectional area diminishments (LMA) in a synthetically
manner (fig. 5).
The investigation of the occurrence of corrosion in situ has a peculiar meaning, on
one side due to the increasing intensity of the polluting agents in the environment,
on the other side due to the requirement regarding the prolongation of the
operational service time of cables.
Cables are highly sensitive to corrosion because their self constructional
embodiment enables the penetration and stagnation of moisture.
One of the most delicate problems of structural engineering is to track and to keep
under control the evolution of degradations caused by corrosion, because this
aspect entail in time changes in the physico-mechanical properties and therefore in
the strength of materials, leading also to stress redistribution in the structural
elements. That is the reason why in the following part a method for monitoring the
corrosion of steel cables and the principle of numerical simulation for this process
will be described.
The acquisition of primary data about corrosion, in the case of cables, can be done
with the method of electric resistance, which is easily applicable in situ. The
method of electric resistance is based on the principle that cable corrosion is
accompanied by cross sectional reduction. In this way, if there is no intercrystalline
corrosion, the raising of electric resistance is produced by the diminution of the
cross section.
In case of intercrystalline corrosion, the area of the cross section is not modified
significantly, but the specific resistance increases. Thereby, measuring the electric
resistance on different portions of a steel cable, considering 500 – 1000 mm long
segments, the commencement of corrosion can be promptly recorded. Monitoring
the corrosion of steel cables by the means of the electric resistance method enables
the undelayed signaling of the occurrence of corrosion, including intercrystalline
corrosion. This method is simple, safe and relatively cheap, implying low costs.
4. NUMERICAL SIMULATION OF CABLE CORROSION
The numerical simulation of corrosion can be done in order to study the

phenomena in time, using probabilistic degradation functions or data acquired
through monitoring. Some of the most frequently encountered fault types are
presented in table 1.
Table 1. Typical patterns of wire rope degradation and failure [25]

Illustration Short explanation
A “bird cage” caused by sudden release of tension and
resultant rebound of rope from overloaded condition.
These strands and wires will not return to their original
positions.
A close-up of a rope subjected to drum crushing. The
distortion of the individual wires and displacement from
their normal position is noticeable. This is usually
caused by the rope scrubbing on itself.
A wire rope jumped from a sheave. The rope itself is
deformed into a “curl” as if bent around a round shaft.
Close examination of the wires show two types of
breaks – normal tensile “cup and cone” breaks and shear
breaks which give the appearance of having been cut on
an angle with a cold chisel.
Localized wear over an equalizing sheave. The danger
of this type wear is that it is not visible during operation
of the rope. This emphasizes the need of regular
inspection of this portion of an operating rope.
A wire which has broken under tensile load in excess of
its strength. It is typically recognized by the “cup and
cone” appearance at the point of the fracture. The
necking down of the wire at the point of failure to form
the cup and cone indicates that failure occurred while
the wire retained its ductility.
An illustration of a wire which shows a fatigue break. It
is recognized by the squared off ends perpendicular to
the wire. This break was produced by a torsion machine
which is used to measure the ductility. This break is
similar to wire failures in the field caused by excessive
bending.
A wire rope which has been subjected to repeated
bending over sheaves, under normal loads. This results
in “fatigue” breaks in individual wires, these breaks
being square and usually in the crown of the strands.
An example of “fatigue” failure of a wire rope which
has been subjected to heavy loads over small sheaves.
The usual crown breaks are accompanied by breaks in
the valleys of the strands, caused by “strand nicking”
resulting from the heavy loads.
A single strand removed from a wire rope subjected to
“strand nicking”. This condition is the result of adjacent
strands rubbing against one another and is usually
caused by core failure due to continued operation of a
rope under high tensile load. The ultimate result will be
individual wire breaks in the valleys of the strands.
The SACOC (Structural Analysis of Corrosion for Cables) software package is

based on the finite element method and it proved to be very useful in the study of
the corrosion of structural cables. A schematic block diagram of this program is
presented in figure 6 [21].
The user can take advantage of several available degradation functions simulating
cable corrosion in time, or he can use quantifications from site investigations
gathered in a data base.
1. GLOBAL BEARING CAPACITY ESTIMATION

BASED ON F.E.M., IN STATIC AND DYNAMIC
RANGE
2. COMPUTATION OF ACTUAL STRESS
3. CORROSION MENU
4. CORROSION SIMULATION DATA BASE

WITH TIME FUNCTIONS 5. STRUCTURAL DEGRADATION
THROUGH CORROSION, BASED
ON TESTS IN SITU.
RESULTS FROM
NONDESTRUCTIVE TESTING
6. ALTERATION OF THE GEOMETRICAL AND

MECHANICAL PROPERTIES DUE TO
CORROSION
7. ACTUAL STRESS RESULTED FROM

CALCULATION, CONSIDERING THE EFFECT
OF CORROSION (STEP 6)
8. COMPARISON BETWEEN THE RESULTS

FROM STEP 7 AND STEP 1
9. RESOLUTION FOR INTERVENTION

YES / NO
Figure 6. The schematic block diagram of the SACOC software.

5. CABLE QUALITY CERTIFICATION
The quality of cables is guaranteed by the manufacturer through quality assurance

certificates, released for each product range. Deviations in dimensions, properties
and shape shall not exceed the prescribed limits from the relevant Romanian
technical regulations, nor those stated by the manufacturer. The list of standards for
wires and steel cables is presented in table 2.
Table 2. The list of Romanian standards (STAS) in use

STAS
Pos. Title
code
1 Materials 880-66
2 Cold-drawn steel wires for drag ropes 1298-80
3 Steel cables. Concept and classification. 1710-79
4 Steel cables. Simple structure cables. Shapes and dimensions. 1513-80
5 Steel cables. Technical requirements. 1352-78
Steel cables. Double cables. Combined structure. Shapes and
6 1689-80
dimensions
Steel cables. Combined cables. Double cables. Ordinary
7 1535-71
structure. Shapes and dimensions
Steel cables. Combined cables. Triple structure. Shapes and
8 2693-80
dimensions
Steel cables. Combined flexible double cables. Shapes and
9 1353-80
dimensions
10 Steel cables. Flat cables. Shapes and dimensions 1559-80
11 Alternate bending test for steel wires 1177-74
12 Testing of metals. Torsion test for wires 1750-80
13 Testing of metals. Coiling test for wires 6622-70
14 Testing of metals. Tensile test for wires 6951-76
15 Testing of metals. Tensile test for steel cables 2172-74
16 Combined double concentric cables 2590-80
17 Steel wire for prestressed concrete 6484-77
6. CONCLUSIONS
Bearing structures with cables are liable to generate unappointed phenomena in

comparison with conventional structures. Thus, beside the great displacements
originated from their custom designed structural shapes, displacements arise due to
the σ breaking / E ratio. While this ratio in case of conventional structures, considering
OL37 steel is 3700 / 2100000 = 1 / 568, in case of using cables becomes
15000 / 1650000 = 1 / 110, namely five times bigger.
Zinc coated cables are able to withstand to corrosion approx. 10 – 15 years. For
this reason, in case of highly corrosive environment should be remembered that
high quality ropes and cables made from synthetic materials with outstanding
endurance are available too.
The required anchoring of cables must be done by wrapping their end with cast
zinc inside threaded pipes instead of using clamps, because the cable may slip out
from the clamp at high dynamical stress.
Cables should be prestressed in a jointly manner with their anchors (thus the
anchoring is tested) before installation, with a force equal to approx. 1.10 times the
computed actual stress, in order to consume the significant remnant elongations
and to avoid subsequent loosening and relaxation.
References
1. Falk, S., Lehrbuch der Technischen Mechanik, Springer Verlag, Berlin-Heidelberg, 1967.
2. Leonhardt, F., Zellner, W., Vergleiche zwischen Hängebrücken und Schrägkabelbrücken für
Spannweiten über 600 m, International Association for Bridge and Structural Engineering, vol.
32, 1972.
3. ***, Bibliography and Data Cable-Stayed Bridges, Journal of the Structural Division, ST10,
1977.
4. Knut, G., Merkblatt-496, Ebene Seiltragwerke, Beratungstelle für Stahlverwendung, Düsseldorf,
1980.
5. Kopenetz, L. G., Ionescu, A., Lightweight Roof for Dwellings, International Journal for Housing
and its Application, vol. 9, No. 3, Miami, Florida, USA, 1985.
6. Haug, E., Engineering Contributions to the Design of Lightweight Structures via Numerical
Experiments, 2nd International Symposium Weitgespannte Flachentragwerke, University of
Stuttgart, 1979.
7. Jensen, J. J., Das Dynamische Verhalten Eines Vorgespannten Kabelnetz, University of
Trondheim, Norway, 1972.
8. Goschy, B., Dynamics of Cable-Stayed Pipe Bridges, Acier-Stahl-Steel, No. 6, 1961.
9. Szabó, J., Kollár, L., Függőtetők számítása, Műszaki Könyvkiadó, Budapest, 1974.
10. Otto, F., Das Hängende Dach, Bauwelt Verlag, Berlin, 1959.
11. Oden, J. T., Finite Elements of Nonlinear Continua, McGraw-Hill, 1972.
12. Krishna Prem, Cable-Suspended Roofs, McGraw-Hill, London & New York, 1979.
13. Fuller, B. W., Weathering of neoprene-coated Nylon fabric, E. I. Dupont de Nemours & Co.,
Report BL-327, 1956.
14. Otto, F., Trostel, R., Zugbeanspruchte Konstruktionen II, Ullstein Verlag, Frankfurt-Berlin, 1966.
15. O’Brien, T., General Solution of Suspended Cable Problems, Journal of the Structural Division,
ST1, Febr. 1967.
16. Sofronie, Ramiro, Vertical Deflection of Suspension Bridges, Revue Roumain Scientique
Technique-Mécanique Appliqué, Tome 24, No. 3, Bucureşti, 1979.
17. Mollmann, H., A Study in the Theory of Suspension Structures, Copenhagen, 1965.
18. ***, Recommendation for Guyed Masts, The Working Group on Guyed Masts of I.A.S.S., 1981.
19. Kopenetz, L. G., Cătărig, A., Cabluri structurale, Simpozionul Tehnologie şi Siguranţă, U.T.
Pres, Cluj-Napoca, 2004. (in Romanian)
20. Kopenetz, L. G., Cătărig, A., Teoria structurilor uşoare cu cabluri şi membrane, U.T. Pres, Cluj-
Napoca, 2006. (in Romanian)
21. Kopenetz, L. G., Cătărig, A., Probleme ale coroziunii cablurilor din oţel, Lucrările celei de a VII-
a Conferinţe de Construcţii Metalice, vol. 2, Timişoara, 1994. (in Romanian)
22. ***, GTU Documentation, USA (Nondestructive Testing Technical Diagnostics).

23. Knut, G., On the Fatigue Strength of Wires in Spiral Ropes, Journal of Energy Resources
Technology, vol. 107, 1985.
24. ***, Internal Report, Bethlehem Wire Rope, USA.
25. ***, Technical Report No. 107, Wire Rope Corporation, St.Joseph, MO, USA.
26. Bârsan, G. M., Kopenetz, L. G., Alexa, P., Theil, S., Gedanken zum Entwurf Leichter
Zusamengesetzter Tragwerke, Proceedings of the 4th Conference on Steel Structures, Timisoara,
Romania, 1985.
Computer aided building diary

Mária Kozlovská1 and Michal Danko2
1
Technical University of Kosice, Faculty of Civil Engineering, Slovakia
2
Inzinierske stavby, a.s. Kosice, Slovakia
Summary
One of the main and also generally obligatory tools for building site works
management is the site diary. The building code orders to building manager the
building diary leading. In the sense of the code, the building site diary is the
document, which is the part of documentation available on building site and in the
document are noticed all relevant events, which have happened on the site.
The article describes possibilities of better building diary leading, as the main
document for managing and monitoring of works on building site by computer
support. Also describes design, structure and architecture of electronic tool as well
as collaboration of database environments. The article presents real examples of
computer aided building site diary paper forms.
KEYWORDS: building, site diary, electronic tool
1. INTRODUCTION
The code also determines that the building diary has to be written from the first day
of preparation works till construction works finish and also who is competent to
write into the diary. According to law, by which are executed some dispositions of
the building code, into the building diary are written all important conditions
concerning a structure building up:
notes about divergences from project documentation verified by surveyors
office in building process or from conditions dedicated in building
permission or in other resolution or proceeding,
facts influencing building works and building process, especially time flow
of works, temperature in relation to building works, especially with wet
production process, weather,
date of building site visit, found facts and steps of the person competent to
do state building surveyor and the person doing state inspection,
records of project engineer and of partial projects designer or of

architectonic product author, geodesist and cartographer of the structure,
records of investor or his empowered person or the structure owner, if he is
not investor,
records of the person making building inspection or capable person, which
manages building process, building contractor,
data about elimination of building faults and back-logs according to
structure inspection decree conditions.
But the reality on many building site is such, that the records are not written daily,
often are made after longer time period, sometimes only before the building
disposition. Then they are only short information about weather, worker and
equipments amount, in better case are written performed building processes but
only in outline version.
2. THE BUILDING DIARY VALUE
The building site diary leading has its purpose, as well as irreplaceable value for all
building participants. That is why are in following text presented some principles
and good advices how write the diary in the manner, that the diary should be an
effective management tool. The building diary presents also the most effective
form of communication among building participants, which flows from law
practice principles, demanding reliable presentation of all facts connected with
building. The building diary in the sense of effective management tool must imply
complex data, which are necessary not only for real state monitoring, but also for
the building process management. The records written into building diary by
particular building participants are the basis for follow change management what is
inseparable part of every building activity. That is why the building site diary is the
main tool of operative works management on building site.
The areas, which management is for the building diary essential:
the building processes performance (data about workers, machines, products,
materials, practices, ….)
time schedule of works, building works quality
occupational health and safety
environment protection
conditions of building site (production, social, operative, climatic…)
financial controlling (performed works billing, internal accounting, …)
contract parameters accomplishment ( eventually change actions)
412 M. Kozlovská, M. Danko
In small or technically simple buildings can be the diary retrieved by simple record
about the structure, which investor writes and also person who special inspections
do, signs the record. In the record must be written also building works amount,
which perform on building site other corporate and natural persons besides the
investor.
RECOMMENDED STRUCTURE (MAP) OF THE SITE DIARY

INTRODUCTION PAPERS
à Contractor
Cover front à Diary number
page à Registration numbers of records sheets
à The building
à Main sheet – the building name, place and date of the
beginning, building number, contractor, project engineer,
building manager, technical supervyor of investor, investor,
subcontractors
à List of main evidences about building: contracts of
building participants, protocol of building site transmission,
additions and changes of contracts, …
General part à List of building documentation – evidence of
documentation available on building site
à Survey of performed exams and measurements – evidence
of protocols from exams and measurements of building
constructions, technological equipments, nets, …
à List of other documents originated in time of building
production – records from control negotiations, protocols,
…
DAILY RECORDS
à Record of contractor about works beginning
à Record of investor about empowered workers competent for
records into the diary writing (building supervisor) and for
the project engineer (author supervisor)
Site diary à Record of contractor about empowered workers to do notes
establishment into the diary
à Name of contractor employer responsible for OHS and FP on
building site
à Verification of investor about acceptance of the diary front
page
à Works beginning date

Beginning of
à Found facts, which influence negatively the working process
works
à Requirements on the faults elimination (limitation)
à Appeal to investor into preparation of materials, machines
Products,
and equipments, if they are not the part of contractor supplies
machines and
à Appeal to transport and supplying of supplies on building site
equipments
in limitation
taking
à Found inadequacies of investor supplies
à Skull session on building site including workers roster
Occupational à Record about control of workers certificates validity, special
health and works performance competences
safety and fire à Difference from safety arrangements in works race condition,
prevention with investor operations or with other contractors works
insuring à Records about work accidents
à Records about building site control performed by
occupational safety inspectorate
à Climatic conditions influencing works process and quality
à Type, amount and time schedule of performed works
Works on à Works assurance by workers, machines and products, ...
building site à Building site readiness for subcontractor taking up
performance à Building parts finishing
à Requirements on investor in connection with restrictions in
works performance
Works breaks à Writing of all facts, which menace or preclude works
and continuance
technological à Writing of all and also short-time break
breaks
à Date and time of interrupted works beginning and break
Continuation
reason record
of works
à Requirements on correction and revisions performance
Works à Appeal to control of quality and covered works integrity
covering à Record about control resuls
à Investor writes the works amount approved by him and by
Auxiliary
contractor
works
à Contractor specify amount and process of performed
specification
auxiliary works, capacity and terms point
à Record of all found divergences from executed and approved
Changes plans
specification à Record of performed divergences from previous plans,
including their approbation, documentation and evidence
Coordination à In field of transport - requirements on transport with object
of (sub) and term introducing
contractors on à In building works field – requirements on amount and terms,

building site verification about their performance
à Works performance requirements – building site equipments
demands
à Other works – lending of machines and equipments, method
of using of building site objects, building site patrol, …
à working out hours amount (works charger by hour tariff)
à undone machines-hours, time of machines and equipments
using
Facts
à increase of costs from reasons of impossibility of expected
necessary for
calculation incompliance because of changed organizational
billing
and technological conditions
à usage of energies (electrics, water, gas, …) taking method,
operation time, used equipments
Exams and à time, process and results of exams and measurements
measurements à investor statement into their result (in detail in control and
performance exam plan of building)
à raised damages on possession – reason writing (natural
Extraordinary forces, swag, …), damage amount, names and address of
events witnesses, performed arrangements
à illegitimate using of machines and equipments
à date of written appeal to transfer action beginning
à place, date and time of transfer action beginning
The building
à transfer action process
acceptance and
à place, date and time of transfer action finish (writing of
handover
protocol about transfer and acceptance between contractor
and investor, including list of faults and inadequacies
Competent à record of competent workers and institutes about their
institutions determinations
standpoints à competent institutes statements
à appeal on participation on equipments examination
Complex à exams process, including achieved results
examination à complex examination beginning and finish date
à list of complex examination participants
Nowadays the form of the building diary can be various. From standard “blind”
form, which can be bought in shops with paper forms (standardized notebooks) to
structured forms, which form and structure is selected by firm itself in dependence
on data management system. Basically the investor must accept the form of the
building diary.
In contract relations writing, concerning often used formulation of commercial

code “if contractor parties do not agree otherwise…” right and careful leading of
the building diary has got big importance. The records in building diary are often
only obligatory verification presenting real building process and accomplishment
of duties among building participants. Timely and good made record in the
building diary is also often an only entry basis for conventional sanctions, damages
reparation, changes or additions of contract applying.
3. E-BUILDING DIARY – APPLICATION ENVIRONMENT
The building diary can be considered as some collection of information and

records. In order to make work with these information and records easier and more
effective, it is necessary to use computers with suitable software. This software
must be dedicated specially for wok with information. Also it has to be able to
observe all functions and conditions for the building diary leading. A useful
alternative presents the product of company Microsoft – Access, which is part of
today often used office packs Microsoft Office. By this program is possible to
create database applications, which are the most suitable for work with
information. Particular database applications are made from database objects
namely from charts, into witch are inserted information, questions, paper forms,
combinations, data pages, macros and modules.
The main characteristic of the program Microsoft Office is simple and especially
wide possibility of work with information. Besides that it has got many other
advantages, which in the case of the site diary leading as fully functional program,
can do leading of the diary and work with it do more effective. At the present is
communication by net and internet connection a common stuff. Similar
possibilities provides also Access 2003, from which is possible to export data into
different external database systems. Also is possible to install all database software
Microsoft Access into computer net server or publish application data at the
internet. The advantage is also the form of outgoing records by program Word or
joining of tabs from program Excel. It is possible to join own databases of
companies with the site diary database, by what is user work easier. User will not
must write some data (about machines or employers), but only choose from
database.
Next advantage of Access is database application insurance. There are many
possibilities how to non-authorised entries. Database applications created in Access
2003 version is possible to sign by electronic signature. The user by his signature
confirms that information contained in this file is valid and from sign immediate
was not changed. If the file will not be changed, is possible to add signatures of
other users. Added digital signature informs that, signature author insure the
project safety. This possibility is in case of the building diary very important,
because records in the diary have to be signed.
4. E-BUILDING DIARY – APPLICATION OF STRUCTURE AND

VISUAL ASPECT
Application of building diary e-form structure comes from higher presented map
(tab. 1) of building diary. This application consists of 28 paper forms (fig. 1) which
are each other connected. They are divided into two groups, into forms for data
import into database and forms for browsing the data yet saved in database
The main structural element of electronic building diary is the forms. Each form
consists of three parts: head-line, body and heel. In head-line is presented the name
of database “Site diary” and “date” which is daily actualized. In place of heel of
forms generality is inserted command button “Back” which can switch over the
user into past form. Body of forms is different depending on contents and function.
The forms dedicated for data insert are similar, only fields describes are different.
In order to visual site of program will be agreeable for user, all forms are orange
with black texts. Pictures 2 – 11 present selected forms (screens).
DR of employers Insert employer into

Daily records list
DR of machines Insert machine into

list
Opening screens
DR of materials Insert material into
register
Daily records DR of exams and Insert exam or

measurements measuremnt into list
Daily register of Insert recesse into

recesses register
DR of accidents Insert accident into

list
DR of changes and Insert change or

divergences divergence into list
Next Insert
Entry form Main form Main sheet subcontractors subcontractors
Changes and
additions
List of main Insert buildings

buildings papers papers into list
List of building Insert building

documentation documentation into
Introductory lists
Survey of made
exams and
List of other Insert other

documents documents into list
Fig.1 Structure of building diary

Fig.2 Opening screen
Fig.2 Main formular

Fig.3 Basic information

Fig. 4 List of basic documents about building
Fig.5 Survey of made exam and measure

Fig. 6 List of workers
Fig.7 Formulary for workers inserts

Fig.8 List of machines
Fig.9 List of materials

Fig.10 List of accidents
Fig.11 Formulary for accident inserts

5. CONCLUSION
The high rate of construction, necessity of fast transport of information, including

their fidelity, create force on new tools development, which can effectively insure
activities connected with building realization. Development and design of
application supporting electronic leading of the site diary is a contribution,
supporting these necessities. Nowadays information and communication
technologies including personal computers are more often the natural “co-operator”
of building managers. Developed application presents the tool, which can
essentially do work of managers more effective.
The article is the part of the project KEGA 3/4004/06 MM program for building
educate support.
References
[1] Kozlovská, M., Hyben, I.: Works foreman – manager of the building process. 1. edition.
Bratislava: Eurostav, 2005.
[2] Building law No.76/1976 in later regulations version.
[3] Korínek, M.: Access 2000. České Budějovice : Kopp, 2000.
Determination of the heat flows, of the solar and luminous

characteristics of the glazing and of the devices for the solar
protection. Automaton calculus program “SOLAR”
Ioan Moga and Ligia Moga
Physics of Constructions, Faculty of Civil Constructions and Installations, University of Cluj-Napoca,
Cluj-Napoca, 400027,Romania
Summary
The automaton calculus program “SOLAR” has resulted from the activity for
establishing the annex A94 from “THE METHODOLOGY FOR CALCULUS THE
ENERGETICS PERFORMANCE FOR BUILDINGS - PART I”, for the applying of
the requirements of the law 372/2005.
The calculus program has at its basis the stipulations of the normative EN 13363-
2:2005.
With the help of the automaton calculus program, it can be determined for any type
of glazing with or without solar protection devices, the solar energy total
transmission factor, luminous transmission factor and the normalized thermal
emission flow.
KEYWORDS: heat engineering for constructions, solar protection devices, energy

savings, rehabilitation of buildings.
1. CALCULUS ALGORITHM
1.1. Calculus program description
The calculus program is realized using the programming language Delphi under
Windows system.
The calculus program was made out taking into consideration the stipulations from
the normative SR ENISO 10211/1-98 concerning the geometrical model, the layout
of the sectioning plans through the glazing system with the solar protection that
forms the calculus network. Regarding the convergence conditions of the obtained
solutions a more important requirement then the one in the normative SR ENISO
10211/1-98 was imposed. The requirement considers obtaining the convergence of
the numeric calculus results by assuring the energy balance condition, in every
426 I. Moga, L.Moga
node of the calculus network, according to relation (17), with a accuracy of

0,0001W.
1.2. Glazing system and the solar protection devices
The glazing unit and the solar protection devices are made from a sequence of
layers of material separated by spaces filled in with gas. The layers of the material
are considered homogeneous and with thermal conductivities, that are not varying
with temperature.
The heat and solar radiation flow are considered to transfer unidimensional. For the
ventilated spaces, the expressions for the two-dimensional convection are
converted into unidimensional formulas.
The layers for material and spaces are numbered with the subscript j from 1 to n,
space n representing the interior environment and space O the exterior
environment. The physical model does not limitate the number of layers.
The basis formulas indicated for the solar radiation and for the heat transfer express
the energy balance for each layer.
For resolving the non-linear dynamic system of equations obtained, it is
recommended the use of a iterative calculus process.
Figure 1 . System made of layers of material and spaces
Legend
Te Exterior air temperature 1 Exterior 8 Solar radiation
Tre Exterior radiating temperature 2 Layer 1 9 Direct solar and luminous
transmission factor
ve Exterior wind speed 3 Space 1

Ti Interior air temperature 4 Layer j 10 Direct solar and luminous
reflection factor
Tri Interior radiating temperature 5 Space j
6 Layer n 11 Radiative and convective
(direct and indirect) thermal transfer
7 Interior
NOTE: The interior and exterior environs are characterised by the interior air
temperature and the radiating temperature; the exterior environment is also
caracterised by the speed wind.
1.3. Solar and optical characteristics of the glazing
For each wave length λ and for each layer of the glazing j, next equilibrium
equations can be written (Figure 2) for the normalized spectral radiative flow Ij si
I’j:
I j (λ ) = τ j (λ ) ⋅ I j −i (λ ) + ρ ' j (λ ) ⋅ I ' j (λ )
I ' j (λ ) = ρ j (λ ) ⋅ I j −i (λ ) + τ ' j (λ ) ⋅ I ' j (λ )

(1)
where:
τ j(λ) spectral transmission factor of the face exterior oriented ;
τ’j(λ) spectral transmission factor of the face interior oriented ;
ρj(λ) spectral reflection factor of the face exterior oriented ;
ρ’j(λ) spectral reflection factor of the face interior oriented ;
Ij(λ) normalized spectral radiative flow interior oriented ;
I’j(λ) normalized spectral radiative flow exterior oriented ;
Figure 2. The scheme with the characteristic data of the layer j and related spectral flows
428 I. Moga, L.Moga
By writing the equation (1) for each layer j of the glazing, a system with n
equations will result, having as unknowns the values for Ij(λ) şi I’j(λ).
The limit conditions for resolving the system of equations will be:
I 0 (λ ) = 1 and I n (λ ) = 0 (2)
After resolving the system of equations and obtaining for each layer j the values for
the spectral radiative flow Ij(λ) şi I’j(λ), the next factor will be obtained:
spectral transmission factor : τ j (λ ) = I n (λ ) (3)
spectral reflection factor : ρ j (λ ) = I ' 0 (λ ) (4)
spectral absorption factor:

α j (λ ) = (1 − ρ j (λ ) − τ j (λ )) ⋅ I j −1 (λ ) + (1 − ρ 'j (λ ) − τ 'j (λ )) ⋅ I 'j (λ ) (5)
The direct solar transmission factor τe, the direct solar reflection factor ρe,and the
spectral absorption factor αej for each layer will be calculated taking into account
the spectral data, determined according to the methodology from SR EN 410:1998;
In the same way the direct luminous transmission factor τv and direct luminous
reflection factor ρv will be calculated.
1.4. Heat transfer
1.4.1 Thermal radiation
The heat flow through radiation depends upon the temperature of the system
double with other heat flows that appear in the system.
For the thermal radiation the scheme from figure 3 is used, where the next
characteristics are presented for each face of the layers that compose the system,
glazing or solar protection layer.
Tj absolute temperature ;
τth,j direct solar transmission factor ;
εj effective emissivity of the face exterior oriented ;
ε’j effective emissivity of the face interior oriented ;
qth,j radiative flow density towards interior ;
q’th,j radiative flow density towards exterior.
Figure 3. Characteristic data of the layer j and related radiative flow densities
The effective emissivity ε is derived from the normal emissivity εn, determined
with the help of a infrared spectophotometer, with the related correction from
the procedure described in SR EN 673 annex A.2.
For each layer j of the system of the glazing equations of radiative energy balance
will be written resulting a complex equations system, where the temperature are at
the forth value :
( )
qth , j = τ th , j ⋅ qth , j −1 + 1 − ε 'j − τ th , j ⋅ qth' , j −i + ε 'j ⋅ σ ⋅ T j4−1
qth' , j −1 = (1 − ε j − τ th , j ) ⋅ q th , j −1 + τ th , j ⋅ qth' , j + ε j ⋅ σ ⋅ T j4 (6)
Limit conditions are given by the interior and exterior radiative conditions, Tr,e şi
Tr,i :
qth , 0 = σ ⋅ Tre4 ; qth ,n = σ ⋅ Tri4 (7)
After resolving the system of equations and determining the temperatures Tj, the
next will be calculated :
the net radiative flow toward exterior
q e = qth' , 0 − qth , 0 (8)
the net radiative flow toward interior

qi = qth' ,n − q th ,n (9)
retained net heat, from thermal radiation, in layer j

( )
qth ,a , j = ε j ⋅ qth , j −1 − ε 'j ⋅ qth' , j + ε j + ε 'j ⋅ σ ⋅ T j4 (10)
430 I. Moga, L.Moga
1.4.2 Heat transfer through thermal conduction and thermoconvection in enclosed

spaces with glazing surfaces
Figure 4. Scheme presenting the characteristic data for an enclosed space and the heat flow
density through thermal conduction and thermoconvection
Legend
1 Layer j
2 Space of gas j
3 Layer j+ 1
(T j + T j +1 )
λj Thermal conductivity of gas in a space j at temperature Tm =
2
sj The thickness of the gas layer from the layer j
hg,j Thermal conductance of the gas in layer j
qc,j Heat flow density through thermal conduction and thermoconvection
starting from layer j to j + 1
Thermal conductivity of gas in a limited space j, at medium temperature
(T j + T j +1 )
Tm , j =
2 , enclosed between glazing surfaces (Figure 4), is given by the
relation:
Nu j ⋅ λ j
hg , j = (11)
sj
where:
λj thermal conductivity of the gas from the enclosed space j ;
sj the thickness of the gas layer ;
λ thermal conductivity of the gas at the temperature Tm;
Nu Nusselt dimensionless number, determined according to the stipulations

from the standard SR EN ISO 673:2000;
Limit conditions for exterior are:
for the air temperature : T0 = Te ;
for the thermal transfer through thermoconvection:
hg,0 = hc,e ; (12)
Limit conditions for interior are:
for the air temperature : Tn+1 = Ti ;
for the thermal transfer through thermoconvection:
hg,n = hc,i ; (13)
After resolving the system of equations and after determining the temperatures Tj,
in each node of the calculus network, the next will be calculated :
resulted net heat in layer j (through thermal conduction and thermoconvection), is
given by :
q c ,aj = hg , j −1 ⋅ (T j −1 − T j ) + hg , j ⋅ (T j +1 − T j ) (14)
heat flow density through thermoconvection toward exterior environs is given by :

q c ,e = q c ,a ,0 = hg ,0 ⋅ (T1 − Te ) (15)
heat flow density through thermoconvection toward interior environs is given by :

q c ,i = qc ,a ,n = hg ,n ⋅ (Ti − Tn ) (16)
A non-linear algebraic system of equations will result, after writting the equation
for the energy balance, in each node of the calculus network. Because of the
dynamic and non-linear nature of interaction between temperature and the
radiative and convective heat transfer, for resolving the resulted system of
equations, it is recommended the use of an iterative process. The dynamic nature is
assess a rewritting for the system of equations, because of the modification of
coefficients belonging to the system of equations, for each step of the iterative
calculus.
2. VALIDATION OF THE “SOLAR” AUTOMATON CALCULUS

PROGRAM
The validation of the automaton calculus program was made using the examples
from annex C of the norm EN 13363-2:2005.
432 I. Moga, L.Moga
Next entry data were used:

1. exterior blind: blind- 50 mm air space- 4 mm glazing- 13 mm air space- 4
mm glazing
2. interior blind: 4 mm glazing- 13 mm air space- 4 mm glazing- 50 mm air
space- blind.
3. inbuilt blind: 4 mm glazing - 13 mm air space- blind- 13 mm air space- 4 mm
glazing.
Propeties of the materials:
Glass:
• 4 mm clear normal glass;
• solar energy transmission factor τe= 0,82, reflection factor ρe= 0,07;
• thermal emissivity ε= 0,84.
Blind:
• solar energy transmission factor τe= 0,2, reflection factor ρe= 0,4;
• thermal emissivity ε= 0,9;
• without thermal radiation transmission.
The obtained results depend on several factors:
- number of decimals with which the thermotechnic parameters are taken
(2,3,4,...);
- number of decimals for working (algorithm calculus accuracy);
- number of calculus steps at which the thermotechnic parameters of air are
changing numarul;
- interpolation manner of those parameters, when this operations are made etc .
After performing the calculus in double precision, is inferred that the values
displayed in norm EN 13363-2:2005 in the annex C, are obtained as intermediate
calculus values before the energetic balancing of the glazing system layers with
solar protection.
After carrying out the calculus up to the energetic balance, until assurance with 3
decimals the energetic balance, the obtained results differ meaningful from the
displayed values in annex C of the norm EN 13363-2:2005. The obtained results
with the automaton calculus program “Solar” are the proper one by the
thermotechnic point of view, taking into consideration the energy balance in each
layer of the glazing assembly with solar protection.
The listing with the obtained results is given for the analyzed examples.
In the listing the results are obtained for: direct solar transmission factor τ,
radiation factor Gh, convection factor Gc, ventilation factor Gv, secondary thermal
transfer interior oriented Ghcv, total energy transmission factor Ghcv+T
On the right side of the tables in the listing is presented the energy balance for each
surface of the glazing assembly- solar protection. It can be observed that the results
displayed in tables in annex C of the norm EN 13363-2:2005, are the intermediate

calculus values before the energy balancing of the surfaces of the system.
3. CONCLUSIONS
The automaton calculus program “SOLAR” has a prompt practicability in the

designing phase and also in the phase for energetic rehabilitation for new and
existing buildings.
This program covers the required necessity of the experts in the field. Its
applicability is general and can be taken into consideration for any type of glazing,
filled with various gases, with the surfaces of the clear or treated glass, with or
without solar protection in three variants of emplacement: in exterior, in interior
and between leaves of glass.
References
1. SR EN 410:2003, Glass for constructions. Determination of the solar and luminous
characteristics of glazing.
2. SR EN 673:2000, Glass for constructions. Determination of the thermal transmittance U.
Calculus method.
3. SR EN 673:2000/A1:2002, Glass for constructions. Determination of the thermal transmittance
U. Calculus method.
4. SR EN 673:2000/A1:2002/A2:2004, Glass for constructions. Determination of the thermal
transmittance U. Calculus method.
5. SR EN ISO 10077-2:2004, Pthermal perfomance of windows, doors and shutters. Calculus for
the thermal transmittance- Second part: General method
6. SR EN ISO 10211-1:1998, Thermal bridges in constructonsi.Thermal flows and superficial
temperatures. First part: General calculus method.
7. SR EN ISO 10211-1:1998/AC :2003, Thermal bridges in constructonsi.Thermal flows and
superficial temperatures. First part: General calculus method.
8. SR EN 13363-2:2006, Solar protection devices applied for glazings. The calculus of the solar and
luminous transmission factor. Second part: The detailed calculus method.
434

435
2 I. Moga, L.Moga
436
434

Effect of moulding humidity on the properties of dry pressed

ceramic tiles
Radomir Sokolar
Department of Technology of Building Materials and Components, Faculty of Civil Engineering,
Brno University of Technology, City, Veveri 331/95, 602 00 Brno, Czech Republic
Summary
The aim of the work was to describe moulding humidity influence on the properties
of dry pressed ceramic tiles from fly ash – clay mixture (70 % fly ash and 30 %
stoneware clay). It was determined properties of green body (bulk density) and
fired body (water absorption, bending strength, bulk density and apparent porosity
according to EN ISO 10545 standards) by the mathematical functionalities.
KEYWORDS: Dry pressed ceramic tiles, pressing water content, bulk density,
water absorption, green ceramic body, fired ceramic body.
1. INTRODUCTION
The enormous amount of fly ashes generated during mineral coal burning is still far
from being used in its totality as a product or by-product, making technological
alternatives needed in order to reduce its possible environmental impact. The paper
present one of the possibilities how to use fly ashes – as a basic raw material for
the production of ceramic tiles. It is evident, that moulding moisture influenced
properties of dry pressed green or fired ceramic body very much. It was determined
optimal water content to get maximal compact green body and the best properties
of the firing fly ash - clay bodies according to EN ISO 10545 standards (e.g. water
absorption, bending strength) and properties of the fly ash body microstructure.
The properties of firing fly ash – clay body (water absorption, bending strength)
were compared in accordance with requirements of EN 14411 for dry pressed for
ceramic tiles (Table 1).
Table 1 - The choose properties of ceramic tiles - group B

Characteristics BIa BIb B IIa B IIb B III
Water absorption average ≤ 0,5 0,5 - 3 3-6 6 - 10 > 102)
[%] individually Max. 0,6 Max. 3,3 Max. 6,5 Max. 11 Min. 9
Bending strength average ≥ 35 ≥ 30 ≥ 22 ≥ 18
≥ 151)
[MPa] individually Min. 32 Min. 27 Min. 20 Min. 16
Breaking strength (t ≥ 7,5 mm) 1300 N 1100 N 1000 N 800 N 600 N
438 R. Sokolar
2. RAW MATERIAL PROPERTIES
The fly ash utilized in this work was originated in the burning process of mineral
brown coal in a pulverized coal-fired plant of Thermoelectrical Power Station,
which is located in Melnik (Czech Republic). This thermoelectrical process has
good efficiency in the combustion of the mineral coal and the residual carbon is
very low - 1,2 wt%. Stoneware clay (from region Postorna CZ) was used as a
binder for non plastic fly ash grains. The chemical composition of raw materials is
showed in Table 2.
Figure 1. Microstructure of used fly ash (2000x).
Table 2. Average chemical composition of used raw materials

[mass-%] SiO2 Al2O3 Fe2O3 TiO2 CaO MgO K2O Na2O S IL
Fly ash 55,9 29,3 4,7 1,7 2,2 1,4 1,6 0,1 0,1 1,2
Clay 62,2 18,8 4,2 1,5 1,4 1,4 2,3 9,3
Particle size distribution of fly ash was studied by using of the floating method
through the 0,063 mm sieve. The fly ash was grinded in the dry laboratory ball mill
with clay together to get minimal residue on the 0,063 mm sieve.
Table 3. Particle size distribution of fly ash

Original After Ball grinding with clay
Sieve residue 0,063 mm [%] 43,2 5,9
3. SPECIMEN SHAPING AND FIRING
The mixture of raw materials (70 % mass fly ash + 30 % mass clay) was milled in
dry ball mill together. Test specimens measuring 100 mm x 50 mm x 10 mm were
shaped in a laboratory press from granulate = moistened mixture with a different

water content were pressed through a 1 mm sieve. The pressing pressure
corresponds to used values for dry pressed ceramic tiles industrial production from
spray-dried granulates (40 MPa).
The firing of test specimens proceed in the industrial tunnel kiln – maximal firing
temperature 1020 °C with 5 hours soaking time in this temperature. After firing, It
was determined properties of firing body: E – vacuum water absorption, B – bulk
density (Bg – bulk density of green (not fired, only after drying) body, Bf – bulk
density of fried body), P – apparent porosity, T – apparent relative density
according to EN ISO 10545 – 3 and capillarity in dependence on moulding
humidity. Frost resistance of firing body was determined by using indirect method
- T – value (saturation value according to DIN 52253 – 3) describes the amount of
open pores capable of filing with water under atmospheric pressure, in relation to
the pores which fill with water at a vacuum of 30 mbar. Frost resisted bodies
embody low T – values (under 0,75).
5. RESULTS
According to figures 2, 3 and tables 4, 5 it is evident, that all determined properties

of green or fired bodies are depended on the pressing moisture of pressing
granulate – it is concerned linear dependence. With increasing of pressing moisture
the body is more compact and decreased the capillarity of fired fly ash – clay dry
pressed body.
19
Firing body water absorption [%]
18
y = -0,4522x + 23,791
2
R = 0,9978
17
16
15
14
10,0 11,0 12,0 13,0 14,0 15,0 16,0 17,0 18,0 19,0 20,0
Moisture of granulate [%]
Figure 2. Fired body water absorption in dependence on pressing moisture of granulate

440 R. Sokolar
1900
1800
Bulk density [kg.m ]
Firing body
-3
y = 13,713x + 1583,7
1700 2
R = 0,9977
1600
1500 Green body

y = 9,7219x + 1361
2
R = 0,9798
1400
10,0 11,0 12,0 13,0 14,0 15,0 16,0 17,0 18,0 19,0 20,0
Moisture of granulate [%]
Figure 3. Bulk density of green and fired body in dependence on pressing moisture of
granulate
Table 4. Properties of green and fired body

Moisture Bg Bf P E T T-value
[%] [kg.m-3] [kg.m-3] [%] [%] [kg.m-3] [-]
10,6 1458
11,4 1475 1741 32,4 18,6 2577 0,988
15,2 1505
15,6 1519 1798 30,0 16,7 2568 0,981
16,3 1523 1804 29,7 16,5 2569 0,976
19,1 1542
19,5 1548 1852 27,7 15 2563 0,963
Table 5. Capillarity of fired bodied in dependence on pressing moisture of pressing

granulate
Moisture Capillarity [mm] – after 5-10-20-30-40-50-60-70 min
[%] 5 10 20 30 40 50 60 70
11,4 47 64 76 85 all
15,6 40 54 65 74 83 95 all
16,3 36 50 60 69 77 83 all
19,5 34 45 54 62 69 75 86 all
6. CONCLUSION
Possibility of dry pressed ceramic tiles production explicitly on base of Czech fly
ashes is quite real. To get optimal properties of final fired body it must be
determined optimal water content of pressing granulate.
Acknowlegements
This Research project was financed with MSM 0021630511 „Progressive Building
Materials with Utilization of Secondary Raw Materials and their Impact on
Structures Durability“
References
1. MAAGE M. Frost resistance and pore size distribution of bricks 1-2, Ziegelindustrie
International, 1990, no. 9, p. 472-481, no. 10, p. 582-588.
2. BENTRUP H., FRANKE L.: Evaluation of the frost resistance of bricks in regard to long service
life, Ziegelindustrie International, 1993, no. 7-8, p. 483-492, no. 9, p. 528-536.
3. FRIESE P.: Predictions of the Frost Resistance of Bricks, Ziegelindustrie International, 1995, no.
12, p. 952-963.
4. SOKOLAR, R. Dry Pressed Ceramic Tiles on the Basis of Fly Ash. Interceram 2007. DVS-
Verlag GmbH, Vol. 56, no.1, p. 30-35, ISBN.
Application of damage model for analysis of masonry structures

Jerzy Szołomicki
Department of Civil Engineering, Wrocław Universityof Technology, Wrocław, 50-370, Poland
Summary
The predictive modelling of the masonry structures behaviour, in the non-linear
range represents a challenge due to their semi-discrete and composite nature. An
adequate computational model should include the fundamental mechanisms that
characterize the masonry behaviour at failure. This paper present the application
of a damage model, based on the finite elements method, to simulate the ultimate
response and the mode of failure of different masonry structures. To take into
account the mechanical properties degradation, a damage variable is introduced
in the constitutive law of the material.
KEYWORDS: Masonry structures, damage model, computational simulations
1. INTRODUCTION
The masonry structures are widely used in civil engineering. These structures are
characterised by a softening response. Indeed, the non-linear behaviour of masonry
is due to the damage and plastic micromechanical processes. From a microscopic
point of view the damage is linked to the growth and coalescence of micro-cracks,
leading to the formation of macro-cracks which can induce the collapse of the
structure. The plasticity is due to intergranular displacements and accounts for
inelastic deformations occurring during the loading process. Various nonlinear
models have been proposed in the literature to describe the softening response of
masonry structural elements. The available models adopted for structural
computations are mainly based on macro-mechanical approaches using damage
mechanics and plasticity theory.
As known, any macro-model of masonry structures always includes some
approximations, since the different failure mechanisms of masonry (cracking of the
mortar joints, sliding along bed and/or head mortar joints, cracking of the bricks
under direct tension, masonry crushing) are not exactly reproduced, but are
smeared out in the continuum.
This paper present the application of a damage model, based on the finite elements
method, to simulate the ultimate response and the mode of failure of different
masonry structures.
“Computational Civil Engineering 2007, International Symposium 443
2. THE CONCEPTION OF DAMAGE MODEL
The nonlinear behaviour of masonry can be modelled using concepts of damage

theory. In this case an adequate damage function is defined for taking into account
different response of masonry under tension and compression states. Cracking can,
therefore, be interpreted as a local damage effect, defined by the evolution of
known material parameters and by one or several functions which control the onset
and evolution of damage. The model takes into account all the important aspects
which should be considered in the nonlinear analysis of masonry structures such as
the effect of stiffness degradation due to mechanical effects and the problem of
objectivity of the results with respect to the finite element mesh.
A useful concept for understanding the effect of damage is that of effective stress.
The damaged σ d and effective undamaged σ stress tensors are correlated,
according to continuum damage mechanics, by the relation:
σ d = (1 - d) D ε = (1 - d)σ (1)
where d is a scalar value, ranging from 0 to 1 and representing the local damage
parameter, D is the elastic stiffness matrix and ε is the strain tensor.
The damage function g( τ, r ) defines the limit of the region of undamaged response
and is written at time t as:
(g(τ, r ))t = (τ )t − (r )t ≤ 0 (2)

where the undamaged complementary energy norm is defined as:
(τ )t = γ ( )
2Λ 0 (σ )
t
(3)
where Λ0 (σ ) is the elastic complementary energy.

For Simo’s damage model γ = 1 .
(r )t
in the damage function of Equation (2) is the current damage strength
measured with an energy norm and can be given as:
(r )t = max{(r )0 , (τ )t } (4)
where (r )0 denotes the initial damage threshold of the material.
The initial damage threshold (r )0 , can be considered to carry out a similar function
to the initial yield stress in an analysis involving an elasto-plastic material.
However, in a damage analysis, the value of the damage threshold influences the
degradation of the elastic modulus matrix. A value for (r )0 may be obtained from:
444 J. Szołomicki
σ dt
(r )0 = (5)
(E0 )1/2
where σ dt is the uniaxial tensile stress at which damage commences and E 0 is the
undamaged Young’s modulus. The damage criterion is enforced by computing the
elastic complementary energy function as damage progresses:
(
β σ T De σ )1/2
− (r ) ≤ 0 .
t
(6)
The damage flow rule defines the damage softening and is given by
( )
∂  G τ t , d 
t
d& = µ& t   (7)

∂τ
where µ& ≥ 0 is the damage consistency parameter and defines damage
loading/unloading conditions according to the Kuhn-Tucker relations
µ& ≥ 0, g (τ, r ) ≤ 0, µ& g(τ, r ) = 0 . (8)
In addition, to simplify the calculations in damage analysis, the damage multiplier
µ& is defined so that
µ& = r& . (9)
From the consistency of the damage condition in Equation (2) it is given that
τ& = r& = µ& . (10)
According to Equation (10), the definition (3) we have
γ 2 T −1
µ& = σ De σ& . (11)
τ
(∂G / ∂τ )t
defines the damage rate with respect to the undamaged elastic
complementary norm. If the damage potential function G is assumed to be
independent of d, substitution of Equation (10) into Equation (7) will lead to:
d=G (12)
with the undamaged condition being enforced so that
{G(r ) t
}
( r ) t = ( r )0 = 0 . (13)
Damage accumulation functions is given by:

( )
G (r ) = 1 −
t (r )0 (1 − A) − A exp[B((r )0 − (r )t )]. (14)
(r )t
For no damage, G(r)t = 0. The characteristic material parameters, A and B, would
generally be obtained from experimental data.
3. DAMAGE CRITERION
The damage criterion is defined as a function of the free energy Ψ0 of the

undamaged material, expressed in terms of undamaged principal stresses σ ip,0 :
( ) ( ) 2
∑ (σ )
3
p,0 K σ p,0 p,0
F=K σ 2ρ 0 Ψ0 − 1 = −1≤ 0 (15)
E0 i =1
where: ρ0 is the density in the material configuration.

The terms of the above equation have the following meaning:
( )
K σ p,0 =
r
+
1− r
, (16)
2ρ 0 Ψt0 L ( ) 2ρ 0 Ψc0( ) L
∑
i =1
σ ip,0
r= 3
, (17)
∑ i =1
σ ip,0
( ) =∑ ±σ
3
2ρ 0 Ψt,0c L
p,0
i εi , (18)
i =1
(Ψ0 )L = (Ψt0 )L + (Ψc0 )L . (19)
( )
In these equations Ψt,0c L
represent the part of the free energy developed when the
tension/compression limit is reached. Taking into account that the tension and
compression strengths are f t = (2ρ Ψ E )
0 t
0
0 L and f c = (2ρ Ψ E ) 0
0
c 0 L respectively,
and substituting the last definition in the Equation (16), the damage function can be
written as:
F = σ − fc ≤ 0 (20)
where
446 J. Szołomicki
∑ (σ )
3
σ = [1 + r (n - 1)] p,0 2
i (21)
i =1
f
with n = c . The advantage of the yield criterion written in Equation (21) is that
ft
any yield function F can be used always as long as it is homogeneous and of first
order in stresses (Mohr-Coulomb, Drucker-Prager).
4. GLOBAL DAMAGE IMPLEMENTATION
The idea for global damage indices definition stemmed from a macroscale analogy
with the microscale local damage index definition. Thus, the starting point for
deducing a global structural damage index is Equation (22), which defines local
damage as a relation between the actual free energy Ψ of the damaged material and
the elastic free energy Ψ0 of a fictitious undamaged material.
 1 T 0
Ψ (ε, d ) = (1 - d)Ψ0 (ε ) = (1 - d) ε σ  . (22)
 2ρ 0 
It seemed natural to reach this objective by integrating the pointwise Equation (22)
over a finite mass, as follows:
V
∫ ∫
Ψ = (1 - d)Ψ0 ⇒ Wp = ρ 0 ΨdV = (1 - d)ρ 0 Ψ0dV = (1 - D) Wp0
V
(23)
where D* is the global damage indices of the considered structural mass,

∫
Wp0 = ρ0 Ψ0dV is its fictitious ever-elastic potential energy due to the actual
V
strains and Wp is the actual potential energy. Solving Equation (23) for D, yields
the final expression:
*
Wp ∫ ρ Ψ dV − ∫ (1 - d )ρ Ψ dV ∫ dρ Ψ dV
0 0 0 0 0 0
D = 1- = V V
= V
. (24)
Wp0
∫
V
ρ 0 Ψ0 dV ∫
V
ρ 0 Ψ0 dV
In a finite element context, expression (24) takes the following operational form:
∑ a ∫ B σdV
e
T T
V (e)
D* = 1- (25)
∑ a ∫ B σ dV
e
T T 0
V (e)
where ∑e
denotes the sum over a number of finite elements, a is the mesh nodal
displacement vector, B is the strain displacement matrix, V (e) is the volume of

each finite element (e), σ is the actual stress vector and σ 0 is the stress vector
should the material preserve its original characteristics and undergo the actual
strains.
CONCLUSION
In this paper the author presented a damage model which can be applied
successfully to asses the structural conditions and estimate the safety level and
durability of historical masonry constructions under static and dynamic loading.
The global damage indices provides accurate quantitative data on the state of any
component subpart of a damaged structure and its importance to the overall
structural behaviour, being of invaluable help to the task of assessing the reliability,
safety and definition of adequate repair of retrofitting strategies.
References
1. Maier, G., Nappi, A. Papa, E. Damage models for masonry as a composite material: a numerical
and experimental analysis. Constitutive Laws for Engineering Materials, ASME Press, New
York, 1991.
2. Kachanov, L.M. Continuum model of medium with cracks. ASCE J. Engng Mech.,vol. 106:
p. 1039-1051, 1980.
3. Park, Y.J., Ang, A.S. Mechanistic seismic damage model for reinforced concrete. ASCE, J.
Struct. Engng, vol. 111: p. 722-739, 1985.
4. Pijaudier-Cabot, G., Bazant, Z.P. Non local damage theory. ASCE J. Engng Mech., vol. 113:
p. 1512-1533, 1987.
5. Gambarotta, L., Lagomarsino, S. Damage in brick masonry shear walls. Fracture and Damage
in Quasibrittle Structures: Experiments, Modelling and Computer Analysis, p. 463-472, 1994.
A New Form of the Active Moments Method

Constantin Ionescu
Civil Engineering, University, Iassy, Zip code: 700482, Romania
Summary
The paper desires to pay a tribute to Professors A. Şesan and N. Orlovschi, which
professed last century at the Faculty of Civil Engineering and Installations from
Iasi.
These professors imagined a calculation procedure for frames, named active
moment method, as a response to the displacement (deformation) method.
The structure calculation by displacement methods is conducted on a base system,
obtained by introducing fictive connections which deters the possible displacement
of the nodes – revolutions and translations.
The elements in question in the displacement method are the nodes real
displacements (written Zi). Under the exterior forces action and displacements, on
the elements in question direction, in the complementary connections appear
reactions. Total reactions from complementary connections must be equal to zero.
This way it can be obtained the condition equations.
Despite the active moments method, the basic system is similar to that in the
displacements method, but the element in question are “active moments” of nodes
Mi and displacement (kinematic chains), MA.
The active moment is defined as the moment which by its action on the base system,
on the node “i” or in the degree of freedom “a”, creates the real displacement of
the node “i”, and the real displacement of the nodes which are part of the
kinematic chain “a”.
The condition equations from the active moments method expresses the structure
static equilibrium and has two types: nodes equilibrium equations and kinematic
chains equilibrium equations. Must be mentioned the fact that none of the papers
written by the authors didn’t demonstrated the way in which the equilibrium
equations have been obtained.
In the present paper have been obtained the equilibrium equations by active
moments method by using the girder, node and nodes and beams chain equilibrium
conditions with the help of the methodology "Gh. Em. Filipescu". This way has
been obtained calculation relations for the extremity moments more general than
the ones used in the displacements method, relations (4.6.) and (4.19.).
Also, it has been distinguished an alternative of the active moments method, where
the elements in question are node active moments and active displacement forces.
KEYWORDS: structure calculus, condition equations, active moments, active

forces, displacements, rigidities.
1. INTRODUCTION
The engineering technique evolution, in different time periods, has been

determined by the necessity of refuge and attendance of the human activities which
imposed a diversified volume of constructions. Due to this fact, has increased the
number of problems connected to the search of different constructive forms which
could better undertake the actions with a minimal construction materials consume.
The resistance structures composition and calculation has been made, long ago,
based on the experience and only slightly on the theoretical elements.
Latter, based on the phenomena observation have been established constructive
rules and calculation models, repetitively verified on the constructions.
Generalizing the experience and development of the experimental researches have
been developed the theoretical calculation basis for the constructions.
The current stage of theory development and structure calculus is due to engineers
and science people who have discovered the solutions for the problems of the
resistance structures design and execution for different constructions by
experimental methods but also by theoretical ways.
Considering the investigations, have resulted general calculation methods, main
principles of construction organization and future horizons of the researches in the
area of one of the most interesting branches of the Construction Mechanics and
Construction Dynamics.
In the science people gallery, which have contributed to the development of
Construction Mechanics can be mentioned J.C.Maxwell, O. Mohr, K. Culman, L.
Cremona, pe S. Timoşenko, I.P. Prokofiev and others.
In Romania, great engineers and professors as Gh. Em. Filipescu, A. Beleş, M.
Hanganu, C. Avram, Alex. Gheorghiu, A. Şesan, M. Ifrim, N. Orlovschi and others
have obtained remarkable results in the area of Constructions Static and Dynamics.
2. STATEMENT OF EQUILIBRIUM CONDITION
In the paper “The frame calculation by circumvolution method” published by A.

Şesan and N. Popescu, in 1954, in the Journal of University “Al. I. Cuza” and
450 C. Ionescu
Polytechnic Institute in Iasi, are presented several definitions regarding the static
structures calculus undetermined by displacements method.
It is mentioned that in that moment were known two methods of expressing the
equilibrium equations. The first one had been published in Wien, in 1943, by R,
Guldan. The equilibrium condition was expressed by two types of equations:
• Equilibrium equations on node (ΣMi = 0):
diϕ i + ∑ K ikϕ i + K is ∆ is + K ij ∆ ij + Si = 0 (2.1)
k
• Equilibrium equations of floor (ΣTk=0) :

∑ Kϕ + ∑ K ϕ
u
s
a
j j + Du + Su = 0 (2.2.)
In these equilibrium equations have been made the notations:

φi, φk, φs – nodes unknown angular displacements;
∆k – unknown liniar displacements of the nodes;
Kik – rigidity at circumvolution of a bar extremity;
K s - rigiditatea la deplasare transversală a unei extremităţi de
bară;
Si, Du şi Su – terms resulted from the actions effect.
Starting from the same base system with stucked displacements, I.P.Procofiev has
published, in 1948, another form of static equilibrium equations:
Node equilibrium equations:
N G
∑r Z + ∑r Z
j =1
ij j
s=a
is s + Rip = 0 i = 1,2,...,N (2.3.)
Kinematic chain equilibrium equations:

N G
∑r Z + ∑r
i =1
si i
e= a
se Z e + Rsp = 0; s = a, b, . . ., G (2.4.)
Where: Z1, . . .,Zj are nodes angular unknown displacements ,

Zk, . . . .,Zn – nodes linear displacements.
Analyzing the two forms of the equilibrium equations statement, it can be observed
that the first way in which it is used the forces equilibrium (T = 0), expressed by
forces projections on perpendicular directions on beams, diminishes the application
area only to frames with floors parallel beams. The second way, presented by
Prokofiev, is based on effects overlapping (of the diagrams) and it is more general
than the first one, because it increases the methods area of application also to the
frames with inclined pillars.
It must be mentioned also the paper of P. Mazilu, in Bucharest, in 1946, entitled:
“The Frames Calculation – Displacements and Virtual Mechanical Work in Cross
Method”. The paper has been elaborated, as the author says, between 1942 –1943
and “introduces in Cross Method the principle of virtual displacements and virtual
mechanical work”.
In the same direction is written also the paper “Another Form of the Condition
Equations for the Deformation Method” with the authors: A. Şesan and N.
Popescu. The authors demonstrate that “the use of mechanical work for the
condition equations establishment – node and displacement – can be generalized
for any frames type”.
The cited article and the paper named “Variants and Simplification of the Moments
Distribution Method “, published by A. Sesan and N. Orlovschi, puts the basis of
active moments method by introducing the notions of node and chain active
moment.
In this paper, it is mentioned the use of “revolutions and displacements” as
unknown elements of the condition equations which is a “disadvantage”, because
it operates with “insignificant” values, “modified” by multipliers and because of
that the authors propose to express the equilibrium condition by elements in
question - “active moments”.
3. ABOUT THE “ACTIVE MOMENTS” METHOD
The “active moments” method was created in 1954 by professor A. Şesan and his
collaborators: N. Orlovschi and N. Popescu. This is an alternative of displacements
method which operates with unbalanced moments: “active moments“ of nodes, Mi
and displacement (kinematic chains), MA.
The active moment is defined as the moment which by activating, on the base
system, on “i” node or in the liberty degree “a”, determines the node “i” real
displacement, also the real displacement of the node which are part from the
kinematic chain “a”.
The condition equations in the active moments method express the structure of
static equilibrium and are of two types:
Nodes equilibrium equations:
N G
∑ M = 0;
i
− M i + ∑ µij M j + ∑ µia M a + mi = 0; (i = 1, 2,..., N ); (3.1.)
j =1 a
452 C. Ionescu
• Kinematic chains equilibrium equations:

N G
∑ M = 0;
a
− M a + ∑ µ ai M i + ∑ µ ab M b + ma = 0; (a = a,b,...,G); (3.2.)
i =1 b
Where have been done the notations:
µij = − d ji t ji ; µ ji = − dij tij ; µij ≠ µ ji ; (3.3.)
Kij β ij ,a
µia = ∑ ±ν 'ij ,a ; ν ij' ,a = ; (3.4.)
i ∑ (K
A
ij + K ji ) β ij ,a
1
µ ai =
βa
∑d a
ij (1 + tij ) β ij ,a ; µia ≠ µ ai ; (3.5.)
1
µ ab = −
βa
∑ (ν
a ,b
'
ij , a + ν 'ji ,b ) β ij ,a ; µ ab ≠ µba ; (3.6.)
1
mi = ∑ mij ; ma = (∑ (mij + mji )β i ,a + ∑ Pk δ k , a ); (3.7.)
i βa a k
For these coefficients calculation are used the rigidities for the rotation of the bars
extremities and for the transversal displacement of the bars extremities, Kij şi Kijψ,
the distribution and transmission: dij, and tij and two proportionality coefficients:
βij şi βA .
The proportionality coefficient of the bar rotation ij noted βij is calculated by

considering a bar rotation which is equal to the unit, by which can be determined
the other bars rotations.
If we consider that on a fictive bar acts an active moment, MA, than the
proportionality coefficient of this bar is βA. The calculus is done with the
relation:
∑ (Kψ + Kψ )β ij ji
2
ij , a
βa = a
(3.8.)
∑ (Kψ + Kψ )β
a
ij ji ij , a
4. DIRECT DEVICE FOR CONDITION EQUATION DEDUCTION

IN THE METHOD OF “ACTIVE MOMENTS”
The deformed position of a certain frame can be defined by the mean of nodes
displacements: rotations and translations.
The nodes angular displacements (zi, i = 1, 2, 3, ..., N) are independent variables,

while the linear displacements (zs, s = a, b, c, ..., G) are interdependent by the
mean of a number of parameters equal to the number of liberty degrees of the
kinematic system, obtained by introducing articulations in nodes and in bearing
constraint. The result is that the number of independent parameters, which
geometrically define the deformed position of a structure, equals the nodes number
to which is added the number of degrees of the frame elastic liberty. The nodes
translations are part of a kinematic chain which relies on a single parameter, zs.
For a bar ij, the rotation noted ψij is expressed considering the parameter zs with the
relation:
ψ ij , s = β ij , s zs (4.1.)
Where βij,s is the bar rotation ij when the parameter zs = 1.
4.1. The girder equilibrium conditions
We extract a bar ij from a base system, of a certain structure, corresponding to the

displacements method, activated by burdens and displacements (nodes rotation and
bar rotation), fig.4.1. The general expressions of the end moments can be expressed
as:
M ij = M ij ( zi ) + M ij ( z j ) + M ij (ψ ij ) + M ij ( p )
(4.2.)
M ji = M ji ( zi ) + M ji ( z jj ) + M ji (ψ ij ) + M ji ( p )
Or as notations in fig.4.1:
M ij = K ij zi + t ji K ji z j − K ijψψ ij − mij
(4.3.)
M ji = tij K ij zi + K ji z j − K ψjiψ ij + mji
In the case in which the bar ij, considered in the research, is component of different
chains and nodes corresponding to the liberty degrees a, b, c, ..., G, than the
expression of bar rotation can be written as a sum of rotations corresponding to
each liberty degree:
454 C. Ionescu
G
ψ ij = ψ ij ,a + ψ ij ,b + ψ ij ,c + ...... = ∑ψ ij ,s (4.4.)
s=a
or taking into consideration the relation (4.1):
G
ψ ij = ∑ β ij , s zs (4.5.)
s=a
zi zj
ψij zij
Mji(zi)=tijKijzi
zi
Mij(zi)=Kijzi
Mji(zj)=Kjizj
zj
Mij(zj)=tjiKjizj
Mij(ψij) = - Kijψ ψij
zij
Mji(ψij) = - Kjiψ ψij
Mij(P) = - mij Mji(P) = mji
Fig. 4.1 Situations of girder and node loads considering the statement of equilibrium
conditions – bar end moments ij for different load situations
The end moments, the relations (4.2.), as per relations (4.4.) and (4.5.) and
notations from fig.4.1, become:
G
M ij = K ij zi + t ji K ji z j − ∑ K ijψ β ij ,s zs − mij
s =a
G
(4.6.)
M ji = tij K ij zi + K ji z j − ∑ K ji β ij ,s zs + mji
ψ
s =a
4.2. Node equilibrium conditions
Considering a node i taken from a certain structure, fig, 4.2., in which compete
several bars, activated by a couple Mi. The node will rotate. The node rotation
measurement, zi, will equal each bar extremity rotation measure which competes in
the node. In each bar extremity the end moments, which appear are proportional to
the rotation measurement zi (Mij = Kijzi).
Mik(zi) = Kikzi
Mi
zi zi j
Mir(zi) = Kirzi
zi Mij(zi) = Kijzi Mji(zi) = tijKijzi
Mri(zi) = tirKirzi
r
Fig. 4.2 The end moments of the bars which compete in node i when the node is loaded
with a rotation of node, zi, or with an active moment of node, Mi
Expressing the node i equilibrium, by a moment’s equation:
Ni
∑ M i = 0;
i
⇒ M i = ∑ K ij zi
j =1
(4.7.)
Where:
456 C. Ionescu
1
zi = Ni
Mi (4.8.)
∑K
j =1
ij
Where - Ni represents the number of nodes connected by bars to node i.
4.3. Chain equilibrium condition
It is detached, from a certain frame, fig. 4.3.a., a chain of bars and nodes
corresponding to a liberty degree, depending on the parameter zs. We act on this
chain with an active moment, Ms, which determines the real displacement, zs, or,
the rotations, ψij,s, without the node to have the possibility to rotate.
We express the equilibrium of these bars and nodes chain by a moment’s equation:
Bs
∑M = 0
s
⇒ M s = ∑ ( K ijψ + K ψji )ψ ij , s
ij =1
(4.9.)
And considering the relation (4.1.) it is obtained:

Bs
M s = zs ∑ ( K ijψ + K ψji ) β ij , s (4.10.)
ij =1
Where:
1
zs = Bs
Ms (4.11.)
∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β ij , s
Where - Bs number of bars for the chain s.

In order to determine the reaction from the connection of liberty degree it turns to
the faulty in fig. 4.3.c. and the displaced in fig.4.3.d. The articulated scheme of the
nodes and bars chain loaded with the moments on the bars and the reaction Rs is in
equilibrium. The equilibrium is expressed by an equation of virtual mechanical
work by producing a kinematic displacement compatible with the connections. It
results:
Bs
Rs ∆ ar − ∑ ( K ijψ + K ψji )ψ ij , sψ ijar, s = 0 (4.12.)
ij
Or
Bs
Rs ∆ ar − ∑ ( K ijψ + K ψji ) β ij , s zs β ij , s ∆ ar = 0 (4.13.)
ij =1
Where
Bs
Rs = ∑ ( K ijψ + K ψji ) β ij2, s zs (4.14.)
ij =1
And
1
zs = Bs
Rs (4.15.)
∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β 2
ij , s
k
Ms Mkj
Mkr
a.
i
zs Mrk
Mjk
Mij(ψij,s)=Kijψ ψij,s r
j Mji(ψij,s)=Kjiψ ψij,s
zs
b.
ψik,s=βik,szs
ψkr,s=βkr,szs
ψij,s=βij,szs
Fig.4.3. Loading situations for a nodes and bars chain in order to express the equilibrium
condition: a. Bars and nodes chain loaded with the active moment of displacement, Ma;
b. the connection between the bar rotation, ψij,s and displacement zs;
458 C. Ionescu
k
Ra
c. Mkj
Mrk
j
Mjk
Mrk
ψ
Mij(ψij,s)=Kij βij,szs
r
i Mji(ψij,s)=Kjiψ βij,szs
∆ar
Ra
d.
Mij
Mjk Mrk
ψijar,s=βij,s∆ar
Mj
Fig .4.3 (continuance) Loading situations for a nodes and bars chain in order to express the
equilibrium condition: d. The articulated scheme for a bars chain having in the bars
extremities applied end moments, Mij and the reaction from the liberty degree hardcore; c.
Nodes and bars chain loaded with the real displacement zs, on the liberty degree direction
has been introduced the corresponding reaction Ra
Equalizing the relations (4.11.) and (4.15) it results:

Bs
Rs
∑ (Kψ + Kψ )β
ij =1
ij ji
2
ij , s
βs = = Bs
(4.16.)
Ms
∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β ij , s
Or
Bs
∑ (Kψ + Kψ )β
ij =1
ij ji
2
ij , s
Rs = β s M s = Bs
Ms (4.17.)
∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β ij , s
βs is, from the active moments method point of view, the rotation of a fictive bar on
which would act the active displacement moment.
Based on the relations, (4.8.), (4.11.) and (4.16.) can be written the general
expressions of bar moments loaded with burdens, with node active moments, Mi,
and active displacement moments, Ms, or with node active moments and active
displacement force, Rs:
K ij K ji G K ijψ β ij , s
M ij = Ni
M i + t ji Nj
Mj −∑ Bs
M s − mij (4.18.)
∑Kj =1
ij ∑K ji
s=a
∑ (K
ij =1
ψ
ij + K ji ) β ij , s
ψ
i =1
K ij K ji G K ψji β ij , s
M ji = tij Ni
Mi + Nj
−∑ Bs
M s − mji (4.19.)
∑K
j =1
ij ∑K ji
s =a
∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β ij , s
i =1
Also,
G
M ij = dij M i + t ji d ji M j − ∑ν ij ,s M s − mij
s =a
G
(4.20.)
M ji = tij dij M i + d ji M j − ∑ν ji ,s M s + mji
s =a
And
G
M ij = dij M i + t ji d ji M j − ∑ν ij' ,s Rs − mij
s =a
G
(4.21.)
M ji = tij dij M i + d ji M j − ∑ν Rs + mji '
ij , s
s =a
Where:
K ij K ijψ β ij , s K ijψ β ij , s
dij = Ni
; ν ij,s = Bs
; ν = '
ij Bs
; (4.22.)
∑K
j =1
ij ∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β ij , s ∑ (K
ij =1
ψ
ij
ψ
+ K ji ) β 2
ij , s
460 C. Ionescu
Observation: 1. the expressions 4.8 and 4.11 indicate the fact that the real
displacements of the nodes can be determined based on the node and displacement
active moments.
2. Analyzing the relations (4.3.), (4.6.) and (4.20.) we can say that the last
two relations are more general, because the end moments are applicable to any
straight bar from a static undetermined structure, with rigid nodes, loaded with
burdens, node rotations and parameters of liberty degrees, node active moments
and active displacement moments.
It is considered, a certain structure driven by a force system. The structure will
deform and in the bars extremities appear bending moments, determined by the
relations (4.5), (4.18) and (4.19). In order to determine the elements in question:
nodes different displacements (zi şi zs) or, node and displacement active moments
(Mi şi Ms) we will use the methodology of Gh. Em. Filipescu. In this method are
being used the labile base systems activated by the given forces and the end
moments (Mij, Mji). Are being used two types of equations:
Continuity equations. In the case studied, here, in the paper, these
equations reduced to the level of each bar, brought us to the
relations (4.5), (4.18) and (4.19);
Static equilibrium equations:
Ni
∑M
j =1
ij = 0; i = 1,2,3,.. ., N (4.23.)
LMV(s) = 0; s = a,b,c,...,G (4.24.)
Explaining the equations (4.23) and (4.24) by end moments, the relations (4.5),
(4.19), and (4.21), can be obtained these equation systems:
Ni Ni G Ni
∑Kij zi +∑tijKij z j −∑Kijψ βij,s zs −∑mij = 0

j=1 j=1 s=a j=1
(4.25.)
( )
Bs Bs G Bs
∑ Kij (1 + tij ) βij ,e zi + ∑ Kij (1 + tij ) βij ,e z j − ∑∑ Kijψ + K ψji βij ,s βij ,e zs +
ij =1 ij =1 s = a ij =1
Bs Np
ij
( )
+ ∑ − mij + mji β ij ,e + + ∑ Pk δ k ,e = 0; i = 1, 2,..., N :
k =1
i ≠ j; e = a, b,..., G;
(4.26.)
Or
rii zi + ∑ rij z jj + ∑ ris zs + Rip = 0; i ≠ j; i = 1,2,..., N; (4.27.)

j s
∑r z + ∑r z + ∑r
i
si i
j
ij j
s
se sz + Rsp = 0; i ≠ j; e = 1,2,...,G; (4.28.)
And
Ni Ni G Ni Ni
∑ dij M i + ∑ t ji d ji M j − ∑∑ν ij ,s M s − ∑ mij = 0;

j =1 j =1 s = a j =1 j =1
i ≠ j; i = 1,2,...,N;
(4.29.)
Bs Bs G Bs
∑ dij (1 + tij ) βij ,e M i + ∑ d ji (1 + tij ) βij ,e M j − ∑∑ (ν ij ,s + ν ji ,s ) βij ,s βij ,e M s +

ij =1 ij =1 s = a ij =1
Bs Np
ij
( )
+ ∑ − mij + mji β ij ,e + + ∑ Pk δ k ,e = 0; i = 1, 2,..., N ;
k =1
i ≠ j; e = a, b,..., G;
(4.30.)
Or
M i + ∑ µij z j + ∑ µis M s + mip = 0; (4.31.)
j s
∑µ
i
si M i + ∑ µ sj M j + ∑ µes M e + msp = 0;
j e
(4.32.)
Also
Ni Ni G Ni Ni
∑d M + ∑t
j =1
ij i
j =1
ji d ji M j − ∑∑ν Rs − ∑ mij = 0;
s = a j =1
'
ij , s
j =1
i ≠ j; i = 1,2,...,N;
(4.33.)
Bs Bs G Bs
∑ d (1 + t ) β
ij =1
ij ij ij , e M i + ∑ d ji (1 + tij ) β ij ,e M j − ∑∑ (ν 'ij ,s + ν ' ji ,s ) β ij ,s β ij ,e Rs +
ij =1 s = a ij =1
Bs Np
ij
( )
+ ∑ − mij + mji β ij ,e + ∑ Pk δ k ,e = 0; i = 1, 2,..., N ;
k =1
i ≠ j; e = a, b,..., G;
(4.34.)
Regarding the previous relations certain comments must be done, such as:
462 C. Ionescu
The relations (4.27) and (4.28) represent the condition equations

system for a structure static undetermined, in the displacement method.
The elements in question are the displacements. The free terms are
reactions, and the coefficients are unitary reactions and are determined
with the above relations:
Ni
rii = ∑ K ij ; rij =tij K ij ; (4.35.)
j =1
Ni Bs
ris = −∑ K ij β ij , s ;
ψ
rsi = ∑ K ij (1 + t ji ) β ij ,s ; (4.36.)
j =1 ij =1
Bs Ni
res = ∑ ( K ijψ + K ψji ) β ij , s β ij ,e ; R ip = −∑ mij ; (4.37.)
ij =1 k =1
Be Np
Rep = ∑ (−mij + mji ) β ij ,e + ∑ Pk δ k , p ; (4.38.)

ij =1 k =1
The relations (4.21) and (4.32) represent the condition equation system
for a structure static undetermined, in the method of active moments.
The elements in question are node and bars active moments. The free
terms and coefficients are determined with the relations:
Ni
µii = ∑ dij ; µij = t ji d ji ; i ≠ j; (4.39.)
j =1
Ni Bs
µis = −∑ν ij , s ; µsi = ∑ dij (1 + tij ) β ij ,s ; i≠ j (4.40.)
j =1 ij =1
Be Bs
µej = ∑ d ji (1 + t ji ) β ij ,e ; µes = −∑ (ν ij ,s + ν ji ,s ) β ij ,e ; (4.41.)
ij =1 ij =1
Ni Be Np
mip = −∑ mij ; mep = ∑ (− mij + mji ) β ij ,e + ∑ Pk δ k ,e (4.42.)

j =1 ij =1 k =1
The relations (4.33) and (4.34) represent the condition equations
system for a structure static undetermined, in an alternative of active
moments method. The elements in question are node active moments
and active displacement forces.
4.5. The indirect device for condition equation deduction from active
moment methods starting from the equation system of the displacement
method.
Corroborating the relations (4.8) with (4.35) and (4.11) with (4.16) and (4.37) for e
= s, can be obtained the expressions:
1
zi = Mi; (4.43.)
rii
βs
zs = M ; (4.44.)
rss s
Are introduced the relations (4.43) and (4.44) in the condition equation system of
the displacement method (2.3) and (2.4), results:
rii r r
M i + ∑ ij M j + ∑ is M s + Rip = 0; (4.45.)
rii j rjj s rss
rsi rsj rse

∑ r M +∑r
i
i
j
Mj +∑
s rss
β s M s + Rsp = 0; (4.46.)
ii jj
After dividing the equation (4.46.) by βs it is obtained:
∑µi
si M i + ∑ µ sj M j + ∑ µes M e + msp = 0;
j e
(4.47.)
M i + ∑ µij z j + ∑ µis M s + mip = 0; (4.48.)

j s
The equation system (4.45) and (4.46) is identical with (4.31) and (4.32), and the
system (4.47) and (4.48) is identical with the equation system formed from (2.3)
and (2.4). Consequently, between the coefficients and free terms of displacement
method and active moment method are the relations:
rii rij ris
µii = = 1; µij = ; µis = ; (4.49.)
rii rjj rss
rsi 1 rsj 1 rse

µ si = ; µsj = ; µse = ; (4.50.)
rii β s rjj β s rss
Rsp
mip = Rip ; msp = ; (4.51.)
βs
464 C. Ionescu
3. CONCLUSION
Have been obtained the equilibrium equations from active moment method by
using the equilibrium conditions for girder, node and nodes and bars chain by the
mean of "Gh. Em. Filipescu" methodology. The equation system obtained in
subchapter 4.2., and the relation (4.27.), has interchangeable lateral coefficients.
Has been identified a variant of active moments method, where the elements in
question are node active moments and displacement active forces.
Have been obtained calculus relations for end moments more general than the ones
used in displacement method, the relations (4.6) and (4.19).
References:
1. Orlovschi, N. I., Construction static, vol. II, Static undetermined structures, part 1,2,3, Rotaprint
I.P. “Gh. Asachi” Iaşi, 1975 (in Romanian).
2. Şesan, A., Orlovschi, N. I., Variants and simplifications of moments distribution method,
Construction theory and practice, Iaşi no. 1, 1954 (in Romanian).
3. Şesan, A., Orlovschi, N. I., A generalization of moments distribution methods. Studies and
researches, Academia, R. P. Romania, filial Iaşi, nr. 3-4, 1955 (in Romanian).
4. Şesan, A., Popescu, N., Another form of the condition equations of deformation method,
Construction Theory and Practice, Iaşi nr. 2, 1955 (in Romanian).
5. Amariei, C. I., Construction static – Static undetermined structures, vol. II, Rotaprint I.P. “Gh.
Asachi” Iaşi, 1981 (in Romanian).
Laboratory analysis in B.L.W.T. – priorities in insuring a safety

curtain walling design
Elena Carmen Teleman, Elena Axinte
Civil and Industrial Constructions Dept., Faculty of Civil Engrg., T.U. “Gh. Asachi” Iasi, Romania
Summary
In the last half of century, the mankind is confronted with a new, most serious
problem, determined by the global warming phenomenon and in particular, in the
last years, Romania has already been confronted with the severe consequences of
these climatic changes. A dramatic increase of the number of strong winds with
speeds over 100 km/h in temperate climatic zones had been the cause of disastrous
effects in terms of human lives and enormous material.
The design of very high buildings is a current modern concept but a greater
complexity is imposed even from the first stage because of the importance of a
realistic evaluation of the costs of investment. The balance between the investment
costs and the safety level imposed during the service life time or in fact the risk
involved by its exploitation is the designers most capital decision.
The latest solutions of steel glass façades are in fact glass panes of grate
dimensions sustained in various manners by steel lattice structures or steel rods.
Their important sensibility to every movement of the main structure, the influence
of temperature variations upon the fixing details and hence upon the glass itself
are almost exceeded by the effects of wind dynamic action directly supported by
this envelope.
The last decades brought many scientific studies on these intricate systems and
almost all of them are directed towards a more realistic evaluation of all the effects
of the wind pressure upon light curtain walling.
Along the conditions imposed in the design of these elements in order to cope with
wind pressure, the paper presents some of the results of several studies on models
in the wind boundary layer wind tunnel of the Laboratory of Aerodynamics from
the Faculty of Civil Engineering in Iasi. These results are directed towards a
thorough evaluation of extreme values of wind pressure coefficients and also focus
on the vibrations induced to the elements of the envelope by the dynamic effects of
wind action, a most important comfort and safety criteria.
KEYWORDS: steel-glass envelopes, maxima wind pressure coefficients, spectral

power, frequency of vibration, wind boundary layer tunnel
466 E.-C. Teleman, E. Axinte
1. INTRODUCTION
Neglecting due to lack of knowledge of the extension of wind effects upon the
antropic area had in the recent past year most disastrous consequences [1]. The last
decades of the XX–th century are marked by the development of the science of
building aerodynamics, a domain where the meteorology and the theory of
probabilities meet and mix together with the structural engineering concepts. The
design of a multi-storey structure is a process with a high degree of complexity and
the important effort involved in the project justifies the equilibrium between an
economic cost of execution and the risk assumed for the exploitation period of
time. After a rich experience has been achieved, it seems that the design of the
envelope of a multi-storey building affects in a highest degree the safety level for
the exploitation period of time. The light materials used for these envelopes, like
glass, thermo-insulating materials, steel or aluminium skeleton vary widely their
mechanical characteristics: dilatation coefficient, elastic modulus, strength and
strain but in the process of design of the perimetral cladding the common factor of
all these materials is that together they have to stand up against the effects of wind
dynamic action.
In these conditions the design of a multi storey structure to strong wind gusts and
their effects upon the envelope becomes an important preliminary stage because by
modelling the interaction between the wind and the future structure and
anticipating the structural response, benefic modifications of its aerodynamic shape
may improve the dynamic behaviour. Also, by studying the environmental
conditions imposed to the building important decisions may be taken in order to
avoid unpleasant effects of the urban implantations. Notorious laboratories for the
Aerodynamics of Structures from all over the world play an important role in this
preliminary process.
For at least 50 years the humanity is confronted with a new and most serious
problem: the dramatic climatic changes due to global warming, whose multiple and
severe consequences Romania has already experienced. A rather realistic
prognoses of wind motion at global scale along with the estimation of the local
effects inside the structure of the boundary layer in different environments became
priorities in insuring the security in the exploitation of the buildings because their
consequences at the economic scale are huge. The amplification of the number of
strong winds with speeds of more than 100 km/h in temperate zones and of the
number of hurricanes with wind speeds of more than 200 km/h in tropical and
subtropical zones increased dramatically the costs of reconstructions and repair in
the last decades and unfortunately they were the cause of important human
casualties, and sometimes disasters at human scale. In the same time, the
insurances costs increased in some countries, like Great Britain ten times in the last
50years [2].
Many of the environmental conditions imposed to a high rise building amplify the
effects of wind dynamic actions: increased speeds in shore areas, turbulences due
to perturbations of urban wind profile in the big cities area, vicinity with other
buildings of similar size etc., the implication being the distribution of the wind
pressure/suction on the surface of the building [3]. Due to an important scientific
knowledge accumulated in this respect, the architects and engineers are able to
master the shape and types of structures adopted for these high rise buildings. A
good example is the Milo Turning Tower, project of the Spanish architect Santiago
Calatrava and for the time being, the highest building in Europe (245 m), whose
torsioned shape reduces the vortices and diminishes the local pressures on the
surface of the envelope [4].
2. PARTICULARITIES IN THE DESIGN OF MODERN CURTAIN

WALLING
2.1. Types and design criteria
The requirements for transparent envelopes of the buildings are, no matter to which
codes for practice we are referring to, based on the following criteria [5], [6], [7]:
• dimensions and tolerances;
• strength and serviceability;
• fire protection;
• thermal and physical functions insured;
• other specific criteria imposed by serviceability functions.
As steel and glass are preponderant in the structure of transparent facades, a
fundamental need is to predict the glass behaviour in facades, in order to achieve
both performance and economy.
The conventional supporting structures of steel-glass facades are those in which the
glass pains carry their dead load and the afferent wind load depending on their
surface; the panes are supported by steel window frames which transfer the loads to
the main structure of the building. Normally, the glass frame producer designs the
panes including glass, sealing and fasteners and gives the necessary requirements
for the designer of the load bearing structure.
In the last decades, a very wide number of studies were directed towards the
behaviour of the glass panels under different exploitation conditions. Vertical
movements of the main structure of he building may be important but usually are
situated between +5.0 mm and -25 mm but the expansion/contraction of the glass
itself in vertical panels determine movements of 17.5 mm to -12 mm, important
tension stresses being accumulated at the corners of the panels or on the edges,
depending on the supporting system.
As the movements of the supporting system of the glass panels must follow the
movements of the main structure of the building, after a rich experience was
gathered in research studies, limitation of the maximum deformations of the glass
under wind action in the middle of the panels are imposed: δ1 is 1:175 from the
span; also, deformations of the supporting frame must be limited so δ2, δ3 are no
more than 1:175 of the span.
Also, preventing the lateral buckling phenomenon for the supporting elements
under local pressure/suction impose limits for their slenderness; it was assumed of
1:100 from the span for steel beams and 1:360 from the span in the case of lattice
girders [5].
Fig. 1. Deflections of the steel-glass panels: δ1- deflection in the middle of the panel; δ2, δ3
– deflections in the middle of the edges, on the supports [5]
The last decades imposed new solutions for transparent facades on the market,
being the result of applicability of new technologies combined with the demand of
higher performances. Hybrid –supported facades are greater surfaces of glazing
whose supporting structure is characterized by the replacement of either a
horizontal beam or a vertical column by a stainless steel cable or rod (fig.2.a., fig.
3) [6]. In the solution presented, the vertical cables take dead load of glass panes
and the wind load also. If greater surfaces (higher glass facades) are considered, a
horizontal truss (or more) is designed in order to take the wind action and to limit
the deflections of the glass consequently. A particular type represents the
suspended glass facades for which the cable supported mechanism prevents only
wind loads, the glass panes supporting their own weight (fig. 2.b).
The hybrid-supported structure concept is based on the fact that the cables and rods
remain in tension so one of the key structural design problems for these envelopes
is the thermal expansion, which might cause structural instability so normally, the
cables are pre-stressed. If the height becomes too big, the pre-stressing force into
the cables is not enough to overcome the effects of thermal expansion and for these
situations, springs are used for each vertical truss to pre-stress the cables and rods.
The design imposes for a non-linear analysis for obtaining the stresses,
deformations and the natural frequency response. The literature specifies that a
certain amount of trial and error is required to balance the rod and catenary system
with the floor spring (fig. 4.).
a b
Fig. 2.a, b. Steel-glass façades: a)-hybrid-supported glass façade; b)-cable supported
suspended glass façade [6]
For these types of envelopes, deformation becomes of capital importance and the
permissible deformation is established for the most frequent wind loads compatible
not only with the glass or cladding system involved but also with the joints and the
fastening systems. As the deflections of the cable truss are concentrated around the
holes drilled in the glass, uncontrolled values of efforts may increase the risk of
producing the fatigue phenomenon in the glass [6], [7].
Requirements for the design of the glass panels in limit states design impose the
validation of the relationships for maximum stresses and maximum deflection
defined with the following relationships [6], [7]:
2
a
σ max = k ⋅   ⋅ qd (1)
t
a4 qk
δ max = k ′ ⋅ ⋅ (2)
t3 E
where:
• a-length of shorter edge of the pane;
• t- thickness of the glass pane;
• qd – uniform pressure load on the panel in the ultimate state for design;
• k, k’- coefficient that takes into account the aspect ratio of the panel and also
the normalized value of the load necessary to reduce the calculations from
non-linear to linear case.;
• E-elastic modulus of the glass pane;
• qk- the value of wind load pressure in the limit state of serviceability.
2.2. Studies considering the security in exploitation of light envelopes

design
2.2.1. Accuracy in the determination of realistic values of wind pressure action
The cladding is generally designed to sustain the efforts due to wind dynamic
action. Several factors influence the design criteria in this respect [7]:
• wind attack angle;
• internal pressure inside the building;
• risks involved by the presence of accidental wide openings in the facade
which would lead to the alteration of internal pressure;
• amplification of wind suctions in the corners of buildings;
• effects of the alteration of the environmental conditions during the life time
upon the distribution of wind pressures on the surface of the building and also,
the effects of vicinity.
Imposed by the necessity of national standard regulations of design, execution and
exploitation of curtain walling systems in Romania and in order to insure the
security level imposed by the Romanian legislation a code for practice was
elaborated: NP 102-04, issue nr. 170/15.02.2005 [8], according to which,
whichever is the curtain walling system proposed by the designer, it must be tested
for the comfort criteria, particularly from the wind action point of view (chapter
5.2.5. Efforts upon curtain walls due to wind dynamic action). The tests will
consider for each system separately: local reference wind pressure according to the
height above the ground, gust factor considering the specific dynamic behaviour of
the structural main system and the peak factors applied to local pressure
coefficients, considering the dynamic characteristics of the facade panel.
Serious lack of data makes these requirements onerous. For example, if a safety
level is imposed, a reference wind pressure value superior to that corresponding to
a 10 years recurrence will be adopted. The recommendation is to extrapolate the
values of a Gumbel I or Frechet II distribution of extremes based on the

observation data from of I.N.M.H. Bucuresti, Filaret station. But since, the
distribution regards only one station it will not correspond with accuracy for all
over the surface of the Romanian territory and since the recommendation specify
that the designer should adopt any other type of data in order to insure the security
level, in fact it is non-operational for other zones than the one mentioned.
Fig. 3. Working system of a hybrid supported façade [5].
Fig. 4. Dissipative system for the wind: floor spring joint (Korean World Trade Center) [7].
Testing the curtain wall panels considering the cumulative effects of both dynamic
behaviour of the main structural system and the dynamic behaviour of the steel-
glass panel is difficult to achieve. The safety factor adopted in the limit state design
theory is a function of two other random variables: safety index, β and the
coefficient of variation of the wind action (speed/pressure), ν. Considering a log-
normal distribution, the safety factor is defined by [9]:
φ = e 0.75 βν (3)
It is important to observe that, if a normal distribution is considered, then the safety
index has usually values between 2.5 and 4.5; in this case the probability of failure
during the service life is placed between 1/100 and 1/106 .
The usual formulation of the wind pressure in the codes for design against wind
action all over the world is:
w = qref ⋅ ce ⋅ cg ⋅ c p ⋅ A (4)
where:
• qref- reference wind dynamic pressure;
• ce-exposing factor;
• cg-peak factor;
• cp-pressure coefficient;
• A-surface of the area exposed to wind pressure.
Bearing in mind that all the measured values are random and analyzed through the
statistics methods, a global coefficient of variation of wind action will be obtained
from the formulae:
ν = ν q2 + ν C2 + ν C2 + ν C2
ref e g p
(5)
In the process of design, the extreme values of wind pressure or suction themselves
influence the security level imposed. Modern techniques of measuring in boundary
layer wind tunnels have an increased degree of accuracy and the latest values of the
variation coefficients reported are considerably lower than the ones stipulated by
the various design codes, mostly due to the variation of the pressure coefficient cp.
As a general conclusion, the variation coefficient according to design code is 0.36
the partial safety factor being thus 2.25 in opposition with a variation factor of 0.18
obtained through wind tunnel measurements leading to a safety factor of 1.59 (β is
3.0 in both cases). An increased value of safety factor is based on local pressure
coefficients obtained from general tests which are not giving a certainty of the
pressure coefficients and offer only a rough and sometimes insecure evaluation of
the wind action on the surface of the building [9], [10].
As the instantaneous pressure coefficient cp is a process fully described through
statistic methods, the maxima/minima values of pressure are defined by the
algebraic sum between the mean pressure values and the product of the
amplification of the variation factor which is named in all the codes for structural
design to wind action as the peak factor, g or cg with the value of the standard
deviation (or the r.m.s.):
(
cˆ p ,k (t )(c p ,k (t )) − c p ,k (t )
g k (t ) = ± (6)
σ p ,k (t )
where:
• g k (t ) - local peak factor in the point k on the surface of the building at the
time t;
(
• cˆ p ,k (t )(c p , k (t )) - local extreme values of maxima/minima of instantaneous
pressure in the point k on the surface of the building at the time t;
• c p ,k (t ) - local mean value of the instantaneous positive (or negative) pressure
in the point k and at the time t;
• σ p , k (t ) - root mean square (or standard deviation).
Although the codes specify that a conservative value of the peak factor may be
taken as 3.5 to predict the extreme values of wind pressure from the mean values
obtained for a certain position on the surface of the building, numerous studies
developed showed that since in certain situations this peak factor may have values
three times bigger, it is not wise however, when it comes for the design of the wall
cladding of a high rise building to limit the evaluations of extremes only to the
value mentioned above.
2.2.2. Consideration of wind-induced vibrations in design of light envelopes

In the design of glass facades slenderness and long spans are often imposed for
both glass panes and frame structures and this may lead to dynamical problems
when a flexible structure with low natural frequency is exposed to gusty winds
[11]. Wind induced vibrations may damage the sealants, loosen fasteners and cause
inconvenience to occupants of the building. In ENV 1991-2-4, 1995 it is stated that
the fundamental natural frequency f0 should be always higher than 5 Hz, regardless
to the wind speed in order to avoid vibrations induced by dynamic wind loads [7].
Under the mentioned value, the dynamic behaviour of the structure must be
separately analyzed, though the methods are not specified. A F.E.M. modelling of
the glass pane and its frame provide an accurate analysis [6], [7], but in this case a
time-history description of the wind dynamic action is desired as an input.
2. STUDIES IN THE LABORATORY OF BUILDINGS

AERODYNAMICS CONCERNING CURTAIN WALLS DESIGN
During the last decade, the studies in the laboratory of Aerodynamics of the Civil
Engineering Faculty were directed towards developing the local instantaneous
pressure measurement techniques of sampling on rigid models of buildings placed
in urban environment and with different levels of height along with refining the
acquisition methods and processing the data obtained [13], [14], [15].
In parallel, a calibration of the results was developed, showing where the accuracy
of measuring must be increased. By elaborating a module of programs under
MATLAB the statistic data were analyzed, mainly for the coefficients of wind
pressure. Several programs were developed:
• TB1.M…TB3.M – program for the determination of the pressure/suction
coefficients-mean, peak and standard deviations, tracing the isopleths with
polynomial interpolation functions and graphics of variation of the
aerodynamic coefficients on the surface of the models;
• GPD.M- program for the determination of the peak factor cg of the local
instantaneous pressure (mean, r.m.s. and maxima values, histograms and
regression lines); also for the validation of the acquisition data (fig. 5.a, b);
• TREC.M – program of determination of the variation coefficient ν for
different averaging periods;
• TST_PSD.M – program for the determination of the spectral power and
spectral density in its normalized shape (fig. 6.a, b).
The last program enlighten the values of the frequencies that gather the majority of
the spectral densities, the frequencies around 5 Hz being very important for the
study as it was mentioned in the subchapter 2.2.1. This program was considered to
be the best way to identify the critical values of frequencies that might superpose
on the natural frequencies of the glass-steel panes.
27m_f...p09.....c_max....g_med = 2.543....s_g = 0.4494 27m_f...p12....g+ = 0.9127....g- = -3.739

15 3
10
a b
c_max-c_med(*)...c_min-c_med(o)
2.5
5
2
0
1.5 2 2.5 3 3.5 4
c_min.....g_med = -2.332....s_g = 0.3974 1.5
15
10 1
5
0.5
0.2 0.3 0.4 0.5 0.6 0.7 0.8
0 sigma
-3.5 -3 -2.5 -2 -1.5
a b
Fig. 5.a, b. Programs under MATLAB elaborated during the studies in the boundary layer
wind tunnel in the Civil Engineering Faculty in Iasi: a)-Program GPD.M for the
determination of peak factors; b)- Calibration of the peak factor values.
Fig. 6.a, b. Spectral densities normalized and traced in the program TST_PSD.M.: a)
pressure signal processed; b)- suction signal processed.
3. CONCLUSIONS
Modeling the turbulences localized on the surface of the buildings and responsible
of numerous accidents more or less severe was transferred into the numerical
simulation sphere. The analysis of the fluctuating pressures was put into evidence
by a spatial evaluation of the spectral density all over the surface of the structure,
allowing the elaboration of an accurate method of design to wind dynamic action at
least of the most common shapes of buildings [10] are presented in Eurocode 1 and
also in the Romanian code NP 082-04. Design Code. Basis of the structural design
and evaluation of the actions on structures. Wind action [16], [17].
Still, by refining the methods of evaluation of wind tunnel data of pressure on the
surface of the buildings many precious data for specific situations imposed by the
environmental conditions or architectural solutions are put into evidence.
The design of the transparent façades of glass and different solutions for the steel
framing call for simulations by F.E.M. methods of the system dynamic behaviour
using time history registrations of wind pressure. These simulations can be
validated afterwards by tests on larger scales (or natural scales) of specimens of
facade panels.
References
1. www.news softpedia.com, Sci/Tech News Staff - Uragane mai multe din cauza încălzirii
globale (in Romanian)
2. Brian Smith, Tom Wyatt, Structures, Dynamics and Wind a 10-th review, 8th Biennal Scruton
Lecture, Wind Engineering Society, 2003
3. Soligo, Michael, Vicepresident of Rowan Williams Davies & Irwin Inc.(RWDI) , Ontario,
Canada- High Rise Buildings in Hurricane Areas, Technotes, Issue No. 2b
4. http: //www.turningtorso.com
5. Mike Otlet, Head of Structural Engineers ATKINS-The role of a structural engineer in the
design of a façade, ATKINS, U.K. april 2006
6. Jusi Kallionieni-Joints and fastenings in steel glass facades, Helsinki University of Technology,
Depart. of Civil and Environmental Engineering, Laboratory of Steel Structures, Ph.D. Thesis,
1999, Helsinki, Fnland
7. Aki Vuolio- Structural Behaviour of Glass Structural in Facades, Helsinki University of
Technology, Laboratory for Steel Structures Publications, 2003, TKK TER 27, Helsinki, Finland
8. “Normativ pentru proiectarea si montajul peretilor cortina pentru satisfacerea cerintelor de
calitate prevazute de legea 10/1995”, NP 102-04, aprobat ord. nr. 170/2005
9. Peter A. Irwin- Developing Wind Engineering Techniques to Optimize Design and Reduce Risk,
Conference on Wind Engineering 2006 (WES 06), Glasgow, U.K., www.ukwes.bham.ac.uk
10. Scott Gamble P. Eng.- Wind Tunnel Testing – A Breeze Through, Structure Magazine, nov. 2003
11. Peter Irwin, president of Rowan Williams Davies & Irwin Inc.(RWDI) , Ontario, Canada-
Motion Criteria in High Rise Buildings, Technotes issue no. 2c
12. Scott L. Gamble, Roger J. Mittenburg, Michael D. Cicci, Marco Accardo (RWDI, Guelph,
Ontario, Canada)-Prediction of Local Exterior Wind Pressures from Wind Tunnel Studies Using
Time History Analysis Approach, Americas Conference on Wind Engineering, 2001.
13. Teleman, C.-Contributii privind alcatuirea si calculul cladirilor multietajate cu structura
metalica, Teza de doctorat, 2000, U.T. “Gh. Asachi” Iasi (in Romanian)
14. Teleman, E.-C., Axinte, E., Silion, R.- Trends of Actual Computer Assistance for Laboratory
Studies in Boundary Layer Wind Tunnel, “Computational Civil Engineering 2006”, International
Conference, Iasi, Romania 26 may
15. Carmen Teleman, Elena Axinte- Calibration Methods of Data Obtained in Boundary Layer Wind
Tunnel, Buletinul Institutului politehnic din Iasi, Tomul XLVIII (LII), fasc 1-2. Constructii.
Arhitectura, pp. 95-101
16. Cod de proiectare. Bazele proiectarii si actiuni asupra constructiilor. Actiunea vantului NP-082-04.
(in Romanian)
17. XP ENV 1991-2-4/2000- Eurocode 1: Bazele calculului si actiuni pe structuri – Partea 2-4:
Actiuni asupra structurilor- Actiunea vantului (P 06-102-4) (in Romanian)
FEM 3D analysis of RC frame foundations

of rotary cement kiln
Grzegorz Dmochowski, Przemysław Siwiec and Piotr Berkowski
Faculty of Civil Engineering, Wrocław University of Technology, Wrocław, PL50-370, Poland
Summary
Results of 3D finite element analysis of damaged foundations for cement tubular

oven are presented in a paper. The foundation consists of four RC frames modeled
for analysis using solid FE with 24 d.o.f. each. Serviceable and temperature loads,
as well as loads from pre-stressing are taking into account. The frames, primary
designed as 2D structures, are sensitive for out-off-plane and temperature loads.
The additional effects of stresses from these loads should be taking into
consideration during FEM analysis.
KEYWORDS: rotary cement kiln, RC foundation, spatial numerical model, FEM

analysis
1. DESCRIPTION OF THE FOUNDATION
The support structure of rotary kiln under consideration was built in the early
1950s. It consists of four reinforced concrete solid frames spaced at equal intervals.
The lowest frame is denoted by number 1 and the other by respectively 2, 3 and 4.
As part of rotary kiln retrofitting, frame 4 was to be torn down and an additional
frame was to be placed next to frame 1 whereby the latter practically would not
carry any loads. The other two frames: frame 2 and frame 3 were to continue to be
in service and be subjected to additional loads. Because the two frames were
considerably damaged it was necessary to determine their serviceability.
The foundation named frame 2 had been made as an RC frame embedded in the
foundation plate. A static scheme with one column in the form of a rocker with
both its ends articulated had been adopted. The other column had been rigidly fixed
in the plate. The frame’s spandrel beam had been thermally protected with an about
20 cm thick concrete overlay. In the course of its service considerable horizontal
displacements of the kiln’s foundation were observed. Therefore in the early 1980s
the structure was strengthened by tensioning. The rigid-rigid column, the frame’s
spandrel beam and the joint connecting the members were subjected to tensioning.
478 G. Dmochowski, P. Siwiec, P. Berkowski
The bowstrings were made from φ5 mm strings combined into L1+6+12+18 ropes,
with two ropes running on each side of the foundation and concreted.
The technical condition of frame 2 varied. The foundation plate was found to be in
good condition. The foundation had not undergone settlement and was undamaged.
The columns were in satisfactory condition. The concrete reinforcement cover was
found not to have loosened or cracked. The frame’s spandrel beam was in bad
condition. It was found to have sagged by 3-4 cm. The spandrel beam’s
undersurface was cracked and its reinforcement cover was loose over a
considerable area. When the loosened cover was removed, reinforcement in the
form of superficially corroded #35 square bars was revealed. Also the spandrel
beam’s entire top surface was found to be corroded.
The foundation named frame 3 had been built as a stiff frame, considerably long
(8.6 m) along the kiln’s axis, embedded in the RC plate. There were two platforms:
one for a motor and one for the kiln’s rotary bearing on the frame. The top surface
of the frame was thermally protected with a 20 cm thick concrete overlay. The
foundation plate was in good condition: no excessive settlements or damage were
found. The columns were in satisfactory condition. They were uncracked and their
reinforcement cover was found not to be loose. The frame’s spandrel beam was in
bad condition since it had been subjected not only to considerable service loads and
temperature impacts but also to the destructive action of grease escaping from the
motor driving the rotary kiln. Lubricant leaks were visible on the spandrel beam
plate’s undersurface.
2. FOUNDATION MATERIALS’ STRENGTH PROPERTIES AND

EXPLOITATION CONDITIONS
2.1. Concrete
On the basis of the fragmentary documentation which survived in the cement

plant’s archive it was found that the design strength of the foundation concrete had
been assumed to be Rw = 160 kG/cm2 which corresponds to current concrete grade
B12.5-B15. According to non-destructive test results, the concrete in the frames
could be classified as class B10 at the most. Therefore concrete parameters: Rbb =
4.8 MPa and Rb = 5.8 MPa were assumed for further calculations.
2.2. Reinforcing steel
In the surviving fragment of the design, smooth reinforcing steel φ20, φ30 and φ35
with allowable stress σd = 1200 kG/cm2, corresponding to a design strength Ra of
about 106 MPa, was specified. The lower steel strength had probably been adopted
in view of the considerable diameters of the structural reinforcement.
2.3. Foundation tensioning
The spandrel beam, the column and the joint of frame 2 had been tensioned using
double steel ropes, made from φ5 mm strings forming a steel rope of type
L1+6+12+18, running on both sides of the foundations. The nominal rope breaking
force of 1130 kN had been assumed. The ropes were tensioned with a force
amounting to 60% of the nominal force. Thus the tensioning force had been S =
2700 kN. From the way in which the ropes were secured it one could be concluded
that the spandrel beam had been tensioned above the neutral axis, as shown in
Figures 1 & 2.
2.4. Exploitation temperature
Considering that the concrete overlay protecting the foundation against temperature
effects was 20 cm thick the foundation had been assumed to be subjected at its
surface to a temperature field of +40°C, decreasing at every 10°C/m towards the
column’s base.
Directio ns of tensioning the structure
with circu mferential bowstrings
S S Prot ective concrete toppin g
20
S
S
165
Bowst rings a nchoring Art iculated joint

memb er
S
435
S
200
Dimen sio ns according

to existing des ign
170 4 40 120
Figure 1. Frame 2. Structural and tensioning scheme

Figure 2. Frame 2. Division into finite elements
3. ANLYTICAL MODELS OF FOUNDATIONS
Program MES DIANA was used for the numerical analysis [4]. Linear physical
and geometric relations were assumed for the material and the structure.
Isoparametric 3-D finite elements HX24L, described by eight nodes, were used to
build the models (Figure 3). In order to obtain more precise results a 3×3×3
Gaussian points integration scheme (triangles in Figure 3) was adopted.
In such finite elements stresses σyy and strains σzz are constant in direction x and
they can linearly change in directions y and z. Stresses σyy and σzz and strains σyy
and σzz respectively in directions x, y and z behave similarly. A uniformly
distributed load or a load linearly variable along each edge can be applied to each
wall of the finite element. The latter can also be loaded with temperature which
may be different in each of its nodes. The displacements and deformations of the
frames are given for the nodes while the stresses are given for the finite element’s
centre of gravity.
Figure 3. FEM element HX24L
The division of frame 2 model into finite elements is shown in Figure 2. 1022
nodes and 640 brick elements were used to create the model. The articulated joint
between the spandrel beam and the rocker column was modelled using short
articulated-articulated bars.
Protective con crete to pping
20
125 40
20
54 5
Dimensio ns acc ording

200
to existin g desig n
165 445 16 5
Figure 4. Frame 3. Structural scheme

Figure 5. Frame 3. Division into finite elements
The model of frame 3 was created using 1660 nodes and 1104 brick finite
elements.
The following load schemes were adopted for frame 2:
• the foundation dead load,
• the characteristic rotary kiln load,
• the design rotary kiln load,
• the frame tensioned with bowstrings,
• loading with a temperature field.
The following load schemes were adopted for frame 3:
• the foundation dead load,
• the characteristic rotary kiln load,
• the design rotary kiln load,
• loading with a temperature field.
The numerical 3-D model calculations were verified for a static flat bar scheme.
4. ANALYSIS OF RESULTS
The check of the deformations of frame 2 confirmed the latter’s susceptibility to

horizontal loads. The total horizontal displacements of the spandrel beam,
calculated in the elastic phase, exceeded 5 mm. Taking into account material
plasticization, deformations about 2.5 cm were yield, i.e. above the permissible
values.
The results of the numerical analysis, in the form of σxx and σyy graphs for frame 2
are shown in Figure 6 and for frame 3 in Figure 7. In both structures the stresses in
the elements lying on the spandrel beam’s edge and in its middle (along axis Y)
were analyzed. In addition, stresses in the region of the elevated spandrel beam
under the motor driving the kiln were taken into account in frame 3. The results are
presented for the design service loads, the tension of frame 2 and the temperature
increase load.
The total force tensioning the undersurface of the spandrel beam in frame 2 was
3326.4 kN. The load-bearing capacity of the reinforcement with a surface area as in
the design was to 2259 kN, which means that it was insufficient to carry the loads.
Similarly in frame 3, the total force tensioning the spandrel beam’s undersurface
was 3345 kN, which means that for the same assumed reinforcement it exceeded
the spandrel beam’s load-bearing capacity.
An analysis of stresses σxx in the spandrel beam of frame 2 shows only a slight
influence of the temperature load on the degree of strain of the structure. The
stresses due to an increase in temperature amount to merely 10% of the service
loads, which confirms that lowly over-rigid structures are resistant to this kind of
load.
1.6 h [m] Frame 2 1.6 h [m]

Load
temperature
1.2 tension 1.2
service load
0.8 Section 0.8

edge
middle
0.4 0.4
σxx [MPa] σyy[MPa]

0.0 0.0
-3 .0 -2.0 -1.0 0 .0 1.0 2.0 -0.4 -0 .2 0.0 0.2 0 .4 0.6
Figure 6. Frame 2. Stresses σxx and σyy
The effect of the stresses produced by tensioning was somewhat stronger. But due
to the point of application of the steel ropes in the spandrel beam the expected
reduction in tensile stress in the spandrel beam’s bottom zone was slight (below
10%) while the deflection of the spandrel beam increased. The increase in tensile
stress σyy in the middle of the spandrel beam thickness was significant: considering
that the design tensile strength is 580 kPa it may explain the appearance of cracks
propagating perpendicularly to the kiln’s axis. This means that the spatial aspect of
the behaviour of the spandrel beam should have been taken into account in the
design.
2.0 h [m] Frame 3 2.0 h [m]
Load
1.6 temperature 1.6
service load
1.2 1.2
Section
0.8 edge 0.8
middle
elevation
0.4 0.4
σxx [MPa] σyy [MPa]

0.0 0.0
-2 .0 -1.0 0.0 1 .0 2.0 -0.6 -0.4 -0.2 0.0 0.2 0.4
Figure 7. Frame 3. Stresses σxx and σyy
In frame 3 the temperature load impact was much greater. Stresses σxx produced by
temperature were nearly as high as the ones due to the design service load whereas
stresses σyy exceeded them significantly. Interesting was the appearance of tensile
stresses in the spandrel beam’s top zone thickened in order to mount the motor.
The stresses did not exceed the concrete’s tensile strength but as a result of the
long-lasting action of temperature they might cause cracking in and damage to the
spandrel beam’s top surface.
Figure 8. Deformation of frame 2

Similarly as in frame 2, the share of tensile stresses σyy (acting along the kiln’s
axis) produced by the service loads was found to be considerable. The stresses
were responsible for the appearance of cracks perpendicular to the rotary kiln’s
axis.
Figure 8 shows the deformations of frame 2 under the service load, the tension and
the temperature field. The deformations of frame 3 under the service load and the
temperature field are shown in Figure 9.
Figure 9. Deformation of frame 3
5. CONCLUSIONS
From the numerical analyses the following conclusions were drawn:

The load-bearing capacity of the spandrel beams in the investigated rotary kiln
support frames was insufficient to carry the design loads.
The originally adopted static scheme in the form of a frame with one rocker
column was susceptible to horizontal displacements which might increase if
disturbances in the operation of the kiln occur or if the kiln components would be
nonaxially assembled.
The cracks in the spandrel beam of frame 3, running perpendicularly to the kiln’s
axis, were the result of the spandrel beam’s spatial work neglected in the original
design.
When designing rotary kiln support frames it should be taken into account the
increase in stress, due the thermal field, particularly when the static scheme should
be changed for one with two rigid columns.
The 3-D frame model calculations have confirmed the fact that tensile stresses in
the column and in the spandrel beam were concentrated closer to their axes and not
at their outer surfaces. Therefore additional reinforcement should be provided to
carry these stresses.
References
1. Lipiński, J., Foundations for machines, Arkady, Warsaw, 1985, (in Polish).
2. Polish Standard PN-84/B-03264 – Concrete, reinforced-concrete and prestressed structures.
Structural analysis and design, (in Polish).
3. Kobiak, J., Stachurski, W., RC structures, Vol. 1, Arkady, Warsaw, 1984, (in Polish).
4. DIANA, Finite element analysis, Element library, User’s manual – release 6.1, TNO Building and
Construction Research.
5. Comparison of design loads, POLTEGOR-Projekt Ltd., (in Polish).
6. Design of rotary kiln foundations (in Polish), B.B.P. No. 4 in Gliwice, the Opole Branch, 1950,
Extracts, (in Polish).
7. Dmochowski, G., Siwiec, P., Spatial analysis RC frame foundations of rotary cement plant kiln,
Current scientific and research problems in building, Ed. Warmińsko-Mazurski University, 2000,
(in Polish).
Flood routing using ‘Femme’

L. De Doncker1, P. Troch1, R. Verhoeven1,
K. Buis2, P. Meire2
1
Hydraulics Laboratory, Ghent University, Ghent, 9000, Belgium
2
Ecosystem Management Research Group, Universityof Antwerp, Wilrijk, 2610, Belgium
Summary
The environment ’Femme’ (’a flexible environment for mathematically modelling

the environment’) is used to model ecological processes as the transport of
nutrients and polluents. The program is user-friendly and based on a Fortran
modular open source code and uses a lot of integrated integration tools. For the
study of the interaction of ecological processes and flow in the river, a realistic
modelling of the surface water flow is necessary. Here, the implementation of a one
dimensional hydrodynamic model for surface water flow in ‘Femme’ is reported.
Because of the lack of hydrodynamic information in the ecological model, the
kinematic and parabolic equations are implemented first. They describe accurate
the time shift of the wave. The simplification to the parabolic and the kinematic
equations allow a faster and easier solution. The parabolic model is known as the
convection-diffusion equation and describes the translation and attenuation of the
wave in open channels and is valid for short stretches with mild slopes. The
kinematic model doesn’t take into account the flattening of the wave and
corresponds less with the realistic deformation of waves.
First, the context of the research is presented in the introduction. Further, the
importance of hydraulic modelling is illustrated. The study area and the modeling
environment are further developed. As a conclusion, some calculation results are
enclosed.
KEYWORDS: ecosystem modelling, environmental engineering, flood routing,

vegetated rivers.
1. INTRODUCTION
This paper is part of the multidisciplinary research project ’A fundamental study on

exchange processes in river ecosystems’ (University of Antwerp, Vrije Universiteit
Brussel, Ghent University, 2004 - 2007). The overall research aim is to study the
diverse physical and biological processes in margins and inundation areas of water
courses and how their interactions determine the exchange of water, dissolved
488 L. De Doncker, P. Troch, R. Verhoeven
compounds and particulate matter. This paper will give an introduction on

hydraulic modelling. The first action is exploring the possibilities of surface water
modelling, to add in a later phase the vegetational processes. Simplified modelling
equations are used.
2. FLOOD ROUTING
2.1 Importance of Hydraulic Modelling
River hydraulics is characterised by changing discharges and water levels due to

rain fall. Studies about this topic have although to take into account this non-
permanent character of the flow. Time shift and flattening of the peak of the wave
can be remarked by studying waves at two different places in rivers as can be seen
on Figure 1. The hydrograph of the wave is shown in section I as well as in the
more downstream section II.
Figure 1: Hydrograph in 2 sections of the river
Description of this phenomenon, i.e. non-permanent flow of surface water is done

by the Saint-Venant equations which include the continuity equation (1) and the
momentum equation (2):
∂Q ∂h
+B =0 (1)
∂x ∂t
∂Q ∂  Q 2 
(
+   = gA S 0 − S f − ∂h
∂t ∂x  A  ∂x
) (2)
with Q = discharge [m³/s], B = section width at water surface [m], h = water depth
[m], g = gravity [m/s²], A = wetted cross section area [m²], S0 = tan α = bottom
slope [m/m], Sf = energy gradient needed to overcome frictional resistance of
channel bed and banks in steady flow = friction slope [m/m].
2.2 Study Area: the river Aa
Focus of the study is the downstream part of the river Aa (Fig. 2), this is the stretch
between weir 3 and weir 4, a distance of 1.4 km, near the village of Poederlee. In
this area the interaction between groundwater, surface water and vegetation will be
studied. Regular measurements of discharge and water level allow to gather data to
calibrate the model.
The catchment basin of the river Aa is situated in the region of Antwerp and is
hydrographical part of the Nete basin. More than 40 % of the water in the Nete
basin is going to the river Aa, which is although an important river. The river Aa
flows into in the Kleine Nete near the city of Grobbendonk. The origin of the river
Aa is found near the communities of Merksplas and Turnhout and is streaming
through Turnhout, Gierle, Gielen, Poederlee and Vorselaar. The river Aa has a total
length of 36.8 km and the drainage area is about 23,700 ha.
Figure 2: The river Aa

2.3 Simplified Saint-Venant Equations
The Saint-Venant equations can be simplified in some typical cases which can be
easily solved. For rivers with a small slope, the acceleration terms are negligible
and the equation is known as a parabolic model, which is the convention-diffusion
equation. The model describes the attenuation of the wave due to storage and
friction.
The maximum discharge will decrease and the minimum discharge will increase in
the downstream direction, due to the influence of the diffusion parameter D. So, the
wave will become flatter and wider, because the total amount of water has to be the
same.
When also the distance between two sections is short, the peak flattening is
negligible and the most simplified model is the kinematic model which includes
only the friction and the translation of the wave. Peak flattening is not
incorporated.
Equations (3) and (4) present respectively the parabolic and kinematic wave model,
with D = diffusion parameter, responsible for peak flattening and C = convection
parameter, responsible for the translation of the wave. D and C are variables and
depend of the values of the discharge and the water level. Therefore, the solution
can only be obtained by numerical models.
∂ 2Q ∂Q ∂Q
D =C + (3)
∂x 2
∂x ∂t
∂Q ∂Q
C + =0 (4)
∂x ∂t
 ∂S f 
 
C=−
1  ∂h  = − 1 ( ∂Q ) (5)
B  ∂S f  B ∂h
 
 ∂Q 
 
 
1 1 
D= (6)
B  ∂S f 
 
 ∂Q 
The values of D and C depend of the discharge and the water height and can be
found in a numerical way. Values mentioned in (Verhoeven, 2006) are 104 to 106
m/s for D and C = 1.5 to 1.8V (with V the average velocity). Calculations in this
study resulted in other values as D = 250 m/s and C = 0.29 m/s, which is the
translation velocity of the wave and can also be seen on Figure 3. Best values are
obtained by fitting the calculation result on measured values or on more detailed
simulation results as obtained with the Saint-Venant equations.
In a first phase, the cross sections are simplified to rectangular or trapezoidal
sections. Comparison with the effective geometry of the river Aa has shown that
the differences are only small (De Doncker et al., 2005).
2.4 The Hydraulic Model ‘Femme’
’Femme’ or ’a flexible environment for mathematically modelling the

environment’ is developed by NIOO (Netherlands Institute of Ecology) (Soetaert et
al. (2004)). ‘Femme’ is a modelling environment for the development and
application of ecological time dependent processes by use of numerical integration
in the time of differential equations. The program is written in Fortran.
‘Femme’ consists of a wide range of numerical calculations and model
manipulations (as integration functions, forcing functions, linking to observed data,
calibration possibilities, etc.). These technical possibilities allow the user to focus
on the scientific part of the model and detailed research of the model without the
confrontation with real program linked problems.
‘Femme’ is focused on ecosystem modelling, is open source (no black box) and
exists of a modular hierarchical structure (implementation of different models next
to each other). What was missing up till now was the implementation of a
hydrodynamic surface water model to couple ecology and surface water in each
timestep.
For the study of the interaction of ecological processes and flow in the river, a
realistic modelling of the surface water flow is necessary. Here, the implementation
of a one dimensional hydrodynamic model for surface water flow in ‘Femme’ is
reported.
2.5 Implementation of the Flow Equations in ‘Femme’
The hydrodynamic model for the transmission of the wave in the river has to be
implemented in ‘Femme’. Differential equations were converted into algebraic
equations using the finite difference scheme. Forward discretisation of the
equations in time and central discretisation in space is carried out.
Implementation of the hydrodynamic model in ‘Femme’ can be done by using the
characteristics of this environment. Discretisation in space has to be implemented
in the code, the integration routine for timestepping is incorporated in the code.
Further, ‘Femme’ will be used for the integrated hydrodynamic – ecological

modeling. Therefore, it is not advisable to model complex ecological processes
with simplified conceptual hydrodynamic models. In this multidisciplinary
approach, different research areas have to be integrated to study properly the
interaction between surface water and vegetation.
2.6 Calculation Results

2.6.1 Kinematic versus parabolic wave model
The kinematic and the parabolic model are built into the ‘Femme’ environment. As
a first verification of the model, a stretch of a river is modelled. The wave,
measured upstream, has a triangular hydrograph Q(t) as indicated in gray dashed
line at Figure 3 and Figure 4.
Figure 3: Use of the kinematic wave model: results at different distances
The kinematic as well as the parabolic model calculates the transmission of the
wave. The resulting hydrograph at 20.4 m, 388 m and 796 m from the upstream
boundary is indicated at Figure 3 and Figure 4. The total length of the channel is
1000 m. The channel is rectangular, has a bottom width of 10 m and a water height
of 1 m. The Manning coefficient is 0.05 m-1/3s.
In Figure 3, the use of a kinematic model is illustrated. The time shift indicates that
the wave moves from upstream to downstream with a constant celerity, without
peak flattening. The small decrease of the peak discharge is due to the numerical
(non-physical) dispersion as a consequence of the discretisation. Small wiggles at
the base flow are due to the triangular hydrograph, which is not natural and
contains discontinuities.
Figure 4: Use of the parabolic wave model: results at different distances
In numerical modeling, a good choice of timestep ∆t and cell width ∆x is

necessary. These values have to be sufficient small to not to miss any effects (e.g.
peak discharges) and to be sufficient large to minimize the calculation time. The
Courant-Friedrichs-Lewy condition (CFL, Cunge et al.) determines a relation
between the time step and the grid size separation to solve the partial differential
equations in a convergent way.
For an explicit scheme, it means that the solution will be numerical stable. So, for
using the explicit solution scheme of the kinematic and parabolic model, attention
have to be paid at this criteria.
This condition can be avoid by using an implicit scheme for solving the Saint-
Venant equations (as the Preissmann scheme). Time steps can be taken larger,
which is useful for long simulation periods, and the solution will be stable.
The parabolic model, used for the results of Figure 4 contains an extra diffusion
term. Next to the time shift, also the peak flattening in the downstream direction
can be seen.
2.6.2. Variation of the Manning coefficient
Further, some calculations about the influence of the roughness coefficient n are
carried out. Three different values of this friction factor are used: n = 0.05 m-1/3s, n
= 0.2 m-1/3s and n = 0.5 m-1/3s. These are values derived from the research carried
out in the river Aa. Winter values were low, due to the low amount of vegetation in
the river. In spring and summer, biomass starts to increase and results into higher
Manning coefficients.
Figure 5: Use of the kinematic wave model: result at the same distance (x = 388 m) for
different values of the Manning coefficient n
Figure 6: Use of the parabolic wave model: result at the same distance (x = 388 m) for
Differences between winter and summer up to a factor 10 are not rare. As can be
seen, the presence of more macrophytes cause obstruction in the channel. For a
certain amount of water, it is more difficult for the water to pass and so the
discharge will be lower, just as the peak value of the discharge. More vegetation in
the river results in a higher friction coefficient of Manning.
2.6.3 Use of the Saint-Venant equations
After having modeled the simplified kinematic and parabolic equations, the
proliferation of the triangular hydrograph is also modeled using the Saint-Venant
equations. These results can be seen on Figure 7 and Figure 8. Figure 7 shows the
results at different sections when the Manning coefficient is set constant (n = 0.05
m-1/3s).
Figure 7: Use of the Saint-Venant equations: result at different distances
At Figure 8, the influence of this friction coefficient in a certain section (x = 299

m) can be seen. The total stretch is 10,000 m and upstream and downstream
boundary conditions have to be added, which is different compared to the
simplified equations.
The parabolic and kinematic model use the upstream boundary condition and start
to calculate one cell after the other. In contrary, the Saint-Venant equations use
both boundaries and need well balanced initial conditions. Calculation of the
surface water level for permanent steady state flow over the total length of the
channel can be a good start and can serve initial conditions. Further more,
boundary conditions are important because both of them have influences in the
different cells.
Figure 8: Use of the Saint-Venant equations: result at the same distance (x = 299 m) for
Concerning the morphology of the channel, the slope is important for the water
levels in the channel. This bottom slope has an influence on the friction slope Sf
(water surface slope). A constant slope of 0 can eliminate that influence.
3. CONCLUSIONS
‘Femme’ has been used as environment for the description and propagation of
waves in rivers. As a study area, the river Aa was selected and its characteristics
were implemented in ‘Femme’.
By this, three kind of models to simulate the propagation of waves in channels are
considered. As can be seen, the one model is more accurate as the other, but
accurateness includes also a higher degree of complexity. So depending of the
situation and the kind of results that have to be obtained, one model can be chosen.
While studying a variety of ecosystem cases, different wave models (based on the
kinematic, parabolic and Saint-venant equations) or even a constant discharge can
be used in the hydraulic calculation.
Acknowledgements
This research is funded by the FWO (Fund for Scientific Research) - Flanders
(G.0306.04). It is part of the multidisciplinary research project ’A fundamental
study on exchange processes in river ecosystems’ (University of Antwerp, Vrije
Universiteit Brussel, Ghent University, 2004 - 2007). The overall objective is to

study the physical and biological exchange processes in margins and inundation
areas of water courses and how their interactions determine the exchange of water,
dissolved compounds and particulate matter.
Thanks to Mr. Martin Van Daele and Mr. Stefaan Bliki for their assistance with the
discharge measurements.
This research is funded by the FWO (Fund for Scientific Research) - Flanders
(G.0306.04).
References
1. Abbott, M. B. et al., Unsteady Flow in Open Channels, Vol. 1, Ed. by Mahmood, K. and
Yevjevich, V., 1975.
2. Buis, K., Anibas, C., Banasiak, R., De Doncker, L., Desmet, N., Gerard, M., Van Belleghem, S.,
Batelaan, O., Troch, P., Verhoeven, R. and Meire, P., A multidisciplinary study on exchange
processes in river ecosystems, W3M, Wetlands: Monitoring, Modelling, Management, 22 – 25
September, Wierzba, Poland, 2005.
3. Cunge, J., Holly, F. and Verwey, A., Practical aspects of computational river hydraulics, Pitman
Advanced Publishing Program, London, 1980.
4. Chow, V.T., Open Channel Hydraulics, McGraw-Hill, New York 1959.
5. Chow, V.T., Maidment, D. R., Mays, L.W., Applied Hydrology, McGraw-Hill, New York 1988.
6. De Doncker, L., Troch, P., Verhoeven, R., Influence of aquatic weed growth on the flow
resistance of the river Aa, 6th FirW PhD Symposium, 30th November 2005, Ghent University,
Belgium.
7. Soetaert, K., de Clippele, V., Herman, P, ‘Femme’, a flexible environment for numerically
modeling the environment, Manual, NIOO-CEME, Yerseke, The Netherlands.
8. Verhoeven, R., Water beheer en waterbeheersing, course for civil engineers, part C: ‘Afvoer en
berging van water’.
Simulation of cooling tower collapse on the basis of non-linear

concrete model and FEM analysis
Przemysław Siwiec, Sebastian Toś
Department of Civil Engineering, Wrocław University of Technology, Wrocław, Poland
Summary
In this paper authors attempt to take advantage of a non-linear model of material

to create a mathematical model of the cooling tower, and further, taking advantage
of the non-linear FEM analysis, they try to recover numerically the course of the
collapse. Both for concrete, as well as for steel, an elasto-plastic model of material
characterized by specific plasticity criteria has been assumed.
KEYWORDS: cooling tower, non-linear analysis, elasto-plastic material model,

FEM analysis
1. INTRODUCTION
The main purpose of design process is to bestow an object a form, shape and
guarantee proper technical parameters in order to fulfill its functions and safety of
service. As soon as the erecting process begins the object is under influence of
destructive factors and interactions tending to ruin it. The purpose of the structure
designer is to predict these factors and to counteract properly. Nevertheless crashes
do occur. Therefore the new assignment appears: to determine causes and failure
sequences. Being in the know with the failure course we are able to determine the
safety conditions of a similar already existing structure and to undertake proper
improving actions.
In this paper the authors have tried to create a proper numerical model of the
reinforced concrete (RC) cooling tower and to investigate the failure course of RC
cooling tower in 1987.
2. THE FAILURE SEQUENCE
The failure of over 100 m high RC cooling tower was one of the most spectacular
crashes in Poland. On 7th February 1987 at 7 p.m. the RC cooling tower suddenly
collapsed. At the level of 38th concreting cycle almost the whole perimeter of its
shell was cut. Only from the north-east side a deep interstice appeared (Fig. 1). At
the moment of the failure the RC cooling tower was under regular service. Till that
day the object had been exploited uniterruptedly for 23 years and it was considered
to be in the best technical state of all six identical RC cooling towers in the battery.
The most characteristic in this failure was the lack of clear cause. On that day it
was light windy: 3–5 m/s, so the wind factor could not be the main reason. The
unusual influences such as sound waves caused by supersonic airplanes or
underground earthquakes were excluded, too. The RC cooling tower collapsed as if
under its own weight.
According to the initial investigations [8] — carried out from February to April
1987 — it has been established that at the most of the fracture the concrete was
crumbled, cracked and not condensed. At the whole of the fracture perimeter
reinforcement bars without any traces of rapture were found. Some of them were
strongly crumpled, some cambered downwards only. A few of them were
cambered upwards, too.
Fig. 1 RC Cooling tower after the collapse
The examination of the concrete and reinforcement strength of the RC cooling

tower confirmed their satisfactory quality. The strength of most of concrete
samples reached over 20 MPa.
The initial estimation of collapse causes has revealed concrete corrosion of the
operating contact zone leading to loss of the shell stability [8]. However this
phenomenon seamed to be too simple and incomplete especially in consideration of
the fact of strong winds occurring only a few days before the failure.
A following hypothesis has been proposed: the proximate cause of failure was the
improperly performed operating contact zone which in result of variable winds and
500 P. Siwiec, S. Toś
corrosion factors led to loss of the whole shell stability. In order to prove this
hypothesis the nonlinear FEM analysis on the numerical model of RC cooling
tower has been performed, with use of the DIANA system.
3. DESCRIPTION OF THE NUMERICAL MODEL
3.1 Geometry
In order to properly analyze the proposed thesis the numerical model of 100 m high
RC cooling tower has been created.
Eight-node shell elements of CQ40 [13] type have been used for creating the
numerical model of the bicurvature shell structure. These elements allow to
introduce the variable shell thickness in every node so the numerical model has no
rapid stiffness changes.
elementy
łącznikowe
Fig. 2 Numerical model Fig. 3 Elements of upper shelf
The structure has been divided into 17 rings with 12 final elements in every ring
(for the half of the structure) — see Fig. 2 [3].
The stiffening ring has been created with the three-node rod elements of CL18B
type (Beurnoulli) with 6 degrees of freedom in every node. Their cross-section has
been eccentricitaly founded (shifted in relation to) on the net of nodes in order to
model the stiffening shelf properly. Due to possibility of unrestricted cross-section
shape modeling, the actual trapezoid-shaped shelf has been created — see Fig. 3.
The load carrying poles have been modeled with the rectilinear elements of L13BE
type (circle-shaped in cross-section) and they have been connected in a rigid way
with the shell elements and with the fundaments.
The DIANA system allows to model the reinforcement as the settled inserts, bound
with the mother element and make its stiffness increase. The reinforcement
elements have no degrees of freedom of their own so their deformation is
determined by the deformation of the mother element. It means that there is no slip
phenomenon between the concrete and the reinforcement. Generally it is possible
to insert the reinforcement in the mother final element in unrestricted way —
regardless of its geometry.
Two kinds of the reinforcement have been used in the numerical model of the RC
cooling tower: the linear inserts (rods) settled in rod elements, and the net inserts
settled in shell elements. Tab. 1 shows the area of used inserts [8].
All final elements used in the numerical model have been numerical integrable and
for each of the groups the modified integral-points-set has been used:
• Poles elements had eight Gauss’s integral points at their cross-section surface,
evenly distant from the rod center. Three integral zones with 8 points in each
one at the rod length have been used (according to the Gauss’s rule).
• Upper shelf elements had nine Gauss’s integral points at their cross-section
surface (3×3) and three integral points groups distributed at the rod length
(according to the Gauss’s rule).
Tab. 1. The reinforcement area

Element Area
Load carrying poles 0.34%
Upper shelf 0.28%
Outer horizontal reinforcement 0.25%
Inner horizontal reinforcement 0.25%
Outer vertical reinforcement 0.20%
Inner vertical reinforcement 0.20%
• The reinforcement settled in elements had the integral schemes analogical

with their mother elements: three Gauss’s points at the length for rods and
3×3 Gauss’s integral points for reinforcement nets of shell elements.
In order to investigate the simulation of the failure course, the FEM numerical
model has been created in which the connecting elements have been used in order
to model the linear damage zone properly [13].
The characteristic feature of these elements is the possibility of physical
disconnection into two parts after exceeding the tensile strength. The connecting
elements of the N6IF type have defined stiffness against compression/tension and
shearing.
Tab. 2. The material parameters

Parameter
Young’s modulus 26⋅106 kPa
Poisson ratio 0.1667
Density 23 kN/m3
Compression strength 11 500 kPa
sin φ 0.5
sin Ψ 0.5
Tensile strength 900 kPa
Damage energy Gf 500 kNm/m2
Weakenig model Hordijk
Tangential stress ratio D 0.01
In order to create the numerical model properly in accordance with the actual
structure, the connecting elements in damage zone (3 to 5 elements long) had the
stiffness of the weakened concrete (Hordijk’s model - see Tab. 2) and their
shearing stiffness has been equal with the reinforcement bar shearing stiffness. The
stiffness of the rest of the connecting elements has been defined as equal with the
stiffness of the shell elements.
3.2. The material
All final elements had nonlinear characteristics [2] and their parameters have been
based partly on published investigations results of the analyzed object and partly on
the design project documentation and expert opinions on the cooling tower [7, 8].
There has been taken into consideration the material ductility described by the
Drucker-Prager’s constitutive model in the compressed zone for the concrete
elements.
For the tensile zone the smeared cracks model has been used [14, 15]. The
winding effect of the load carrying RC poles has been modeled by increasing the
strength parameters of concrete. On the basis of the in situ investigations the
concrete strength of fc=20 MPa has been assumed whereas for the RC poles it has
been increased to fc=25 MPa [8]. The Tab. 2 shows the nonlinear concrete
parameters.
Tab. 3 Values of the C parameter
d max [mm] α
8 4
16 6
32 10
The most important parameter describing the material in the cracked zone is the
damage energy Gf. Its value has been based on the recommendations of the CEB-
FIB Model Code (1990), which allowed to estimate the damage energy on the basis
of the strength investigations according to the relationship:
Gf = αf ck0,7 (1)
where fck is the compression strength, depends on the maximum aggregate size
(Tab. 3).
Fig. 4 shows the concrete performance for the biaxial state of stress. Fig. 5 shows
the H-M-H criterion for steel.
σ
[MPa]
σ1/σc
250
1
σ2/σc
190
5.e-6 ε [m]
Fig. 4 Biaxial state of stress for concrete Fig. 5 The steel stiffening model
4. THE LOADS
In the calculation carried out only the deadweight and the wind have been
considered [1]. They were the most important factors and most significant for the
structure effort. Other loads such as: temperature, hydraulic pressure, unequal
insolation effects, unequal foundations subsiding and other exceptional loads have
been neglected. As far as the general question is considered taking into account all
of these loads will allow to determine the global state of the structure effort.
During the numerical analysis process the structure has been loaded in a few
phases:
• First the deadweight was loaded (incrementally) till the nonlinear load
coefficient Lf reached the value equal to 1.0.
• Then the wind load was added (incrementally) till the nonlinear load
coefficient Lf reached its maximum value.
5. THE TECHNIQUE OF THE NONLINEAR PROBLEM

TRACING
In order to solve the nonlinear problem the standard Newton-Raphson’s algorithm

and the energetic norm of the results convergence control have been used. In
nonlinear process the arc-length technique based on tracing the selected node
displacement has been used in order to control the load growth. At each stage of
the load increment the node displacement (degree of freedom) towards the selected
direction has been analyzed. Every time the most deformated node - in the previous
stage of the analysis - has been taken into account. Calculations have been stopped
as soon as the lack of convergence has occurred.
construction without damages

nonlinear load coefficient Lf
node 20481
node displacement [m]
Fig. 7 The deformation development in

Fig. 6 The structure answer curves RC cooling tower under the wind load
The structure performance under the increasing load has been investigated with use
of the structure response curve, which described the chosen node displacement as a
function of the load increase (defined as the nonlinear load coefficient Lf ).
6. THE NUMERICAL SIMULATION OF THE COLLAPSE

SEQUENCE
According to the proposed hypothesis it has been assumed that the primary damage
area of the operating contact zone reached up to 21 m (three elements of the half-
ring). In order to fulfill the symmetry conditions the interstice has been located at
the windward side. This is a particular case yet it is sufficient for the essence of the
analyzed matter. The numerical analysis indicates the possibility of the primary
shell damage development caused by the wind loads without collapse of the whole
structure.
Assuming the variable wind directions, it is probable that the progressive interstice
growth up to the whole perimeter would occur. At the next calculation stage the
interstice has been assumed as long enough and the cooling tower has been
deadweight loaded only. At further stages the size of the totally damaged material
in the interstice has been varied. The run of the structure response curve is shown
in the Fig. 8 and the deformation development - in the Fig. 9.
Fig. 8. The structure response curves due Fig. 9. The deformation development in
to development of initial crack. RC cooling tower under the deadweight
The deformation character confirms the conclusion of locally violent variable
displacement character. Displacement occurs within the area of one ring above and
one ring below the interstice. What is worth noticing is the deformation form of the
upper part of the shell.
At the area close above the interstice the shell goes outwards from its perimeter
while the rest of the shell goes towards the inside. According to the above scheme
only a small part of the shell should collapse outside the cooling tower and the
whole rest of it should collapse towards the inside.
7. THE FAILURE SEQUENCE DETAILS
Phase 1 — improper composition of the concrete mixture, its impurities or not

sufficiently careful assembly phase caused the material imperfection which
increased during the service of the cooling tower [5, 6]. The negative coincidence
was localization of this imperfection in the zone of the highest structure effort —
i.e. at of the tower height [3]. At this moment it is impossible to exclude the
geometrical imperfection in the analyzed zone, too.
Phase 2 —increasing material degradation process caused by accelerated corrosion

of the cracked concrete brings about the interstice or the weakened material zone.
Then the variable winds and frost penetration cause enlarging of this zone. It is
possible that the weakened zone still holds its geometrical shape in spite of total
loss of its strength parameters due to internal cracking. This would explain the fact
that the Electric Power Station employers had not noticed the shell defect.
Undoubtedly the interstice initiation took place at the operating contact zone.
Phase 3 — when the strong winds cause the increasing of the interstice up to its
critical length, further interstice length growth begins which results from the
deadweight of the cooling tower only. At first this process was not violent as it was
restrained by the reinforcement and could last even for a few days until the moment
of a violent collapse. If only the stronger puff of wind occurred during this time,
the collapse would undoubtedly have taken place sooner.
The confirmation of the presented above hypothesis of the failure sequence is the
occurrence of the most of the concrete debris at the eastern side of the cooling
tower. According to the presented model the failure process started exactly above
the area where the concrete debris was found outside the cooling tower.
8. CONCLUSION
In this paper the numerical model of the RC cooling tower has been created with
consideration of material and geometrical nonlinearity. This model has been
investigated on the basis of 100 m high RC cooling towers commonly built in
Poland. The presented numerical model enables simulation of the process
occurring during regular RC cooling tower exploitation. The numerical model
analysis allows the precise determination of the causes and the failure course. Due
to possibility of failure results verification — especially the concrete debris
localization — the presented numerical model has been proved reliable.
References
1. 1. Siwiec P., Szechinski M.: Zelbetowa chlodnia kominowa obciazona wiatrem — Inzynieria i
Budownictwo, nr 3, 1997.
2. 2. Siwiec P., Strouven P., Moczko A.: Study of steel bar corrosion by experimental and
analytical simulation — DIANA Computational Mechanics ’94, ed. Kusters M. A., Hendriks
A. N., Kluwer Academic Publishers, Dordrecht — 1994.
3. 3. Siwiec P.: Nieliniowa analiza statyczno-wytrzymalosciowa uszkodzonej, zelbetowej chlodni
kominowej — doctor’s thesis, The Institute of the Building Engineering, Wroclaw 2001.
4. 4. Zabiello Z., Hawre L.: Problemy oceny stanu technicznego chlodni kominowych w kontekscie
awarii chlodni nr 6 — The Science-technical Conference: The durability of the RC cooling
towers, Czocha — 1989.
5. 5. Persona M.: Z badan doswiadczalnych chlodni kominowych Elektrowni Turow — The
Science-technical Conference: The durability of the RC cooling towers, Czocha — 1989.
6. 6. Persona M., Wybrane zagadnienia z badan zelbetowych chlodni kominowych — Chlodnie

Kominowe. Efektywne systemy ochrony i napraw, Sympozjum, Jelenia Gora 1991.
7. 7. Persona M., Szechinski M.: Remark on failure of a reinforced concrete cooling tower —
IABSE Colloquium, Stuttgart 1991.
8. 8. Persona M., Stus R., Szechinski M.: Okreslenie przyczyn katastrofy budowlanej chlodni
kominowej nr 6 oraz ustalenie stanu technicznego czesci niezawalonej pod katem mozliwości
wykorzystania dla ukladu chlodzacego Elektrowni Turow — Reports of the Institute of the
Building Engineering, Wroclaw University of Technology, SPR no. /87.
9. 9. Persona M. i zespol: Wyniki badań i ocena stanu technicznego konstrukcji plaszcza chlodni
kominowej nr 6 w elektrowni Turow — Reports of the Institute of the Building Engineering
Wroclaw University of Technology, SPR nr 43/89.
10. 10. Owczarzy J.: Katastrofy zelbetowych chlodni kominowych — The Science-technical
Conference: The durability of the RC cooling towers, Czocha — 1989.
11. 11. Niku S. M.: Finite Element Analysis of Hyperbolic Cooling Towers — Lecture Notes in
Engineering, ed. Brebbia C. A., Orszag S. A., Springer-Verlag 1986.
12. 12. Golczyk M.: Analiza techniczna przyczyn i skutkow uszkodzen i zniszczen hiperboloidalnych
powlok chlodni kominowych — The Science-technical Conference: The durability of the RC
cooling towers, Czocha — 1989.
13. 13. DIANA Finite Element Analysis, User’s Manual, Release 6.2 — TNO Building and
Construction Researche, Delft 1996.
14. 14. Blaauwendraad J.: Numerical Methods in Structural Mechanics — Delft University of
Techniology — 1993.
15. 15. De Borst R.: Computational Methods in Non-linear Solid Mechanics — Delft University of
Technology, series B20, Delft 1993.
Unsteady axial Poiseuille flow of a Bingham fluid in an annulus

Irene Daprà, Giambattista Scarpi
DISTART, University of Bologna, 40136, Italy
Summary
Several fluids which are of interest in civil engineering, as bentonite,

montmorillonite, slurries, behave as Bingham fluids, i.e. fluids which present a
yield stress. This paper investigates numerically the start-up, the cessation and
some pulsating flows of a Bingham plastic between two coaxial cylinders. The
constitutive law presents a discontinuity for zero shear rate which introduces
severe difficulties in solving any problem of unsteady motion, both analytically and
numerically. A suitable way to avoid the obstacle is to regularize the constitutive
equation using a smooth function to approximate the Bingham law. The adopted
function depends usually on a parameter: as it tends to infinity the model tends (in
distribution theory sense) to the true Bingham law. Several models have been used
in literature to regularize the rheological behaviour of a viscoplastic fluid: in this
paper a new constitutive equation is proposed, based on the error function (erf).
The velocity fields have been calculated using an implicit finite difference method.
The numerical results confirm that the time required to reach the final steady state
is infinite for start-up, whereas it is finite for stopping. In the pulsating flow the
mean velocity of the plug in a period and the mean rate of flow increase owing to
the non linearity of the material.
KEYWORDS: Bingham fluid, unsteady flow, Poiseuille flow, flow in annulus.
1. INTRODUCTION
Bingham fluids such as bentonite and montmorillonite are used to solve many
engineering problems for example in drilling, in pile works and in rock
consolidation. Bentonite is the base material in drilling: it cools the drill and
reduces wear, consolidates the excavation, goes back up in the annulus between
wall and stick bringing to the surface the debris.
Bingham fluids require the application of a greater shear stress than a critical value,
the yield stress, to begin to move. After yielding, the rheological behaviour is that
of a Newtonian fluid, while in the core region the material moves as a solid body.
Many unsteady flows of Newtonian fluids admit analytical solution in closed form:
the same problems for a Bingham fluid are obviously more difficult to solve
because of the non linearity of the constitutive law.
The discontinuity in the constitutive equation at zero shear rate introduces severe
difficulties for analytical and even for numerical solutions. To overcome that,
Glowinski [1] suggested regularising the constitutive law, substituting the
discontinuous function with a continuous one, which, when a given parameter
tends to an assigned limit, usually zero or infinity, tends (at least in the sense of
distributions theory) to the true Bingham law. Several proposals for regularisation
have been made in literature; the most used is that proposed by Papanastasiou [2].
An interesting analysis of advantages and disadvantages of continuous models with
exhaustive literature is given by Frigaard and Nouar [3].
Characteristic of a Bingham fluid is the fact that its stopping time is finite, whereas
Newtonian fluids and generally any fluid without yield stress need an infinite time
to stop. Glowinski [4] and Huilgol et al. [5] give a theoretical upper bound for the
stopping time for pipe and plane flow. Chatzimina et al. [6] study numerically the
cessation of the Couette and Poiseuille flow of a Bingham fluid using the
Papanastasiou model comparing the numerical results with the theoretical
predictions. Many authors have written in the past about unsteady flow of Bingham
fluids; among these we recall for example Duggin [7] which gives a numerical
solution for the start-up flow in a circular pipe; Papanastasiou and Boudouvis [8]
solve numerically flow problems in generalised conduits. Hammad [9] studies the
pulsatile flow for different waveforms of the pressure gradient. Daprà and Scarpi
give an analytical solution to the start-up problem in a circular pipe.
This paper investigates numerically the start-up, the cessation and some pulsating
flows of a Bingham plastic between two coaxial cylinders. A new model based on
the error function is proposed to regularize the constitutive equation.
2. CONSTITUTIVE EQUATIONS
Being v , T and γ& the velocity vector, the stress tensor and the rate of strain
tensor, the constitutive equation of a Bingham fluid is
 γ& = 0 τ ≤ τ0

  τ0 & (1)
T =  γ& + µ  γ τ ≥ τ 0
  
where τ 0 is the yield stress and µ is the viscosity of the fluid. To avoid the
singularity in the constitutive equation, it is possible to substitute the rheological
law (1) with a continuous function. In 1987 Papanastasiou suggested the equation
510 I. Daprà, G. Scarpi
 τ 0 1 − exp ( − mγ& ) 
T=   + µ  γ& (2)
 γ& 
 
The authors propose the following relation which is continuous and infinitely
derivable everywhere
τ 
( )
T =  0 erf kγ& + µ  γ& (3)
γ& 
where, as usual,
x
2
erf ( x ) = ∫e
−t2
dt (4)
π 0
is the error function.
3. AXIAL FLOW IN A CONCENTRIC ANNULUS

3.1. Steady flow
The geometry is shown in Fig. 1. The flow is one-dimensional and the momentum
equation is:
∂p 1 ∂ ( rτ ) ∂u
− + =ρ (5)
∂x r ∂r ∂t
where x is the direction of the motion, r the radial coordinate, p the pressure, τ
the shear stress, u the velocity, ρ the fluid density and t the time. Eq. (1)
becomes
 ∂u
 =0 τ ≤ τ0
∂r
 (6)
τ = τ 0 sgn  ∂u  + µ ∂u τ ≥ τ 0
  ∂r  ∂r
where
x u
0
Ri R1 R2 Re r
0
Fig. 1. Flow geometry
 ∂u
−1 <0
 ∂u   ∂r
sgn   =  (7)
 ∂r   1 ∂u
≥0
 ∂r
∂p
Putting − = P , the steady solution for a Bingham fluid is then
∂x
 1  Pr 2 
u = u1 =  − + a ln r − τ 0 r + b1  Ri ≤ r ≤ R1
 µ 4 

 1  Pr 2

u = u2 =  − + a ln r + τ 0 r + b2  R2 ≤ r ≤ Re (8)
 µ 4 

u = u ( R ) = u ( R ) R1 ≤ r ≤ R2
 1 1 2 2

R1 ≤ r ≤ R2 being the unyielded region, where the material behaves as a solid. The
equation
a ln
(
Re −τ 0 + τ 02 + 2aP ) + τ  R + R − τ 02 + 2aP  P
 − (R2
)
− Ri2 = 0 (9)
R (τ )
0 e i e
+ τ + 2aP
2  P  4
i 0 0  
allows to evaluate a and

PRi2
b1 = − a ln Ri + τ 0 Ri (10)
4
PRi2
b2 = − a ln Ri − τ 0 Ri (11)
4
−τ 0 + τ 02 + 2aP
R1 = (12)
P
τ 0 + τ 02 + 2aP
R2 = (13)
P
the amplitude of the plug is then
2τ 0
R2 − R1 = (14)
P
The rate of flow Q , which allows to calculate the mean velocity V , is
R1 Re
∫
Q = 2π ru1dr + π ( R22 − R12 ) u + ∫ 2π ru dr
1 2 (15)
Ri R2
π ( R22 − R12 ) R2 − R1
Using the hydraulic radius R = = as reference length, eq. (8)
2π ( R1 + R2 ) 2
and (15) give
R2 P   2 2 η1ηe ηe4 η22ηe2 η12ηe2 

V=  c  η η ln − + + η η
2 2
− +
µ (ηe2 − ηi2 )   η2ηi 2
2 e i e
2 2 
ηe4 1 2ηi4 1 1
( η2 − η12 ) − η22 (ηe2 − ηi2 ) + (η2 − η1 ) ⋅
2
+ − + (16)
8 8 8 4 2
η η η
3 3 3
2  
⋅  1 − e − i − η1η22 − η23 + η2ηiηe + η22ηe  
 3 3 3 3  
a
where ηe = Re / R , ηi = Ri / R , η1 = R1 / R , η2 = R2 / R and c =
Re2 P
3.2. Unsteady flow
Introducing the following dimensionless quantities: ξ = x / R , η = r / R , v = u / V ,

θ = τ R / µV , θ 0 = τ 0 R / µV = Bn , P = PR 2 / µV , K = kV / R , T = t µ / ρ R 2 , eq. (5)
becomes:
1 ∂ (ηθ ) ∂v
P+ = (17)
η ∂η ∂T
The constitutive equation
 ∂v  ∂v
θ = Bn erf  K + (18)
 ∂η  ∂η
is plotted in Fig. 2.
1.2
0.8
0.4
0
θ
-0.4
-0.8
-1.2
-0.004 -0.002 0 0.002 0.004
∂ v/ ∂η
Fig. 2. Regularized constitutive equation
Substituting (1) in (17)

 2 KBn   ∂v  2   ∂ 2 v 1  ∂v  ∂v   ∂v
P+ exp  −  K   + 1 2 +  + Bn erf  K  = (19)
 π   ∂η    ∂η η  ∂η  ∂η   ∂T
  
For a coaxial pipe of dimensionless internal radius ηi and external radius ηe , the
boundary conditions are:
v (ηi , T ) = v (ηe , T ) = 0 ∀T (20)
Moreover
∂v ∂v
= =0 v (η1 ) = v (η2 ) (21)
∂η η1 ∂η η2
3.2.1. Start-up flow
A suddenly positive pressure gradient P = P0 is applied to the fluid at rest. P0

gives rise to a unit non-dimensional mean velocity in steady state and must be
greater than 2Bn in order for the fluid to move. The initial condition for eq. (19) is
v (η , T ) = 0 ηi ≤ η ≤ η e T < 0 (22)
3.2.2 Cessation of flow
The constant pressure gradient P0 applied to the fluid in steady flow is suddenly
cancelled.
Conditions at T = 0 are
 P0η 2
v = v1 = − + A ln η R − η Bn + B1 ηi ≤ η ≤ η1
 4
 P0η 2
 v = v2 = − + A ln η R + η Bn + B2 η2 ≤ η ≤ ηe (23)
 4

v = v1 (η1 ) = v2 (η2 ) η1 ≤ η ≤ η2

a b b
where, referring to eqs. (9)-(11), A = , B1 = 1 and B2 = 2
µV µV µV
3.2.3 Pulsatile flows
Two different pressure wave are imposed to a fluid at rest

a) rectangular wave
∞
P = P0 ∑  H (T − n (T1 + T2 ) ) − H (T − n (T1 + T2 ) − T1 )  (24)
n =0
where H (T ) is the unit step function; in a period T0 = T1 + T2 the pressure

gradient P is equal to P0 for the first time interval T1 and P = 0 for the
second one T2 .
b) sinusoidal wave superimposed to a constant pressure gradient P0 :
P = P0 (1 + ε sin 2π f T ) (25)
where f is the frequency.

4. NUMERICAL RESULTS
The numerical calculations have been carried out for coaxial pipes with ηi = 8 and
ηe = 10 : the annulus width is thus 20% of the radius of the external pipe; the
dependence of the unyielded core extension η2 − η1 on the Bingham number in
steady state is plotted in Fig. 3.
2
1.8
1.6
1.4
1.2
η 2-η 1
1
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50
Bn
Fig. 3. Solid plug amplitude versus Bingham number
Eq. (19) with boundary conditions (20), (21) and the appropriate initial conditions
is solved using an implicit finite difference method. The numerical procedure has
been first validated for a Newtonian fluid, for which the analytical results are well
known.
Numerical experiments indicated that the maximum value for parameter K in
η −η
order to obtain valid results can be estimated as K = 5000 2 1 . If K exceeds
Bn
this value, the numerical procedure can exhibit instabilities. An uniform spatial
grid of 1000 points has been adopted, using a time step equal to 4 ⋅ 10−6 . Smaller
time steps give the same results, whereas greater ones can generate instabilities.
The difference between the steady-state velocity profile obtained with the
numerical solution and the theoretical one for a true Bingham fluid is less than 1%.
a) Start-up
Fig. 4 illustrates the numerical steady velocity profile for Bingham fluids, with
Bn = 1, 2, 3,10 and for a Newtonian one, which can be considered as a Bingham
fluid with Bn = 0 . Fig. 5 shows the development of the plug velocity versus time
for the same values of Bn .
1.6
1.4
1.2
0.8
v
0.6 Bn=10
Bn=3
0.4 Bn=2
Bn=1
0.2 Bn=0
0
8 8.5 9 9.5 10
η
Fig. 4. Steady velocity profiles
1.6
1.4
1.2
1.0
0.8
v
Bn=10
0.6
Bn=3
Bn=2
0.4
Bn=1
Bn=0
0.2
0.0
0 1 2 3 4
T
Fig. 5. Start-up: solid plug velocity versus time
b) Cessation of flow
Unlike Newtonian fluids and generally any fluid without yield stress which need an
infinite stopping time, Bingham fluids require a finite time Ts to reduce to zero the
velocity. Glowinski [5] [6] and Huilgol et al. [7] gave a theoretical upper bound for
the stopping time for Poiseuille and plane flow. Fig 6 shows the rate of flow as
function of time at different Bingham numbers. As expected, the rate of flow
decreases the more rapidly the greater Bn is. Fig 7 shows the time needed to
reduce the rate of flow to 10−3 and 10−5 respectively, as function of Bn : the
difference between the two obtained values of time becomes appreciable only if
Bn > 10 . The time required to reduce the rate of flow to 10−5 can be considered as
the stopping time.
120
100
80
Bn=1
Bn=2
60
Q
Bn=3
Bn=10
40
20
0
0 0.1 0.2 0.3 0.4 0.5
T
Fig. 6. Cessation of flow: discharge as a function of time
10
Q=0.001
1 Q=0.00001
Ts
0.1
0.01
0.1 1 10 100
Bn
Fig. 7. Cessation of flow: stopping time Ts versus Bingham number
c) Pulsatile flows
For a rectangular wave, Fig 8 illustrates the core velocity versus time for some
values of Bn ; referring to eq. (24) T1 = 2 and T2 = T1 /10 . For Bn = 10 the fluid
completely stops for a time length about equal to T2 / 2 .
For a sinusoidal wave like eq. (25), putting Bn = 5 , Figs 9-10 show the time
development of plug velocity for some values of ε and frequency f = 0.05 and
f = 0.5 respectively. For f = 0.05 and ε = 0.5 the movement stops for a time
length of about 15% of the period. Calling vs , vmax , vmin the steady, the maximum
and the minimum plug velocity respectively, vmax − vs is always greater than
vs − vmin , for all values of ε . The quantity δ = ( vmax − vs ) − ( vs − vmin )  / vs can be
considered as an index of the asymmetry of the velocity with respect to the steady
value. Fig. 11 shows δ versus frequency at different ε . Being always δ > 0 it is
easy to deduce that the mean rate of flow in a period increases owing to the non
linearity of the constitutive equation.
1.6
1.2
0.8
v
Bn=0
0.4 Bn=1
Bn=3
Bn=10
0
0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6 4 4.4
T
Fig. 8. Rectangular wave: plug velocity versus time
3.5
ε=0
3 ε=0.1
ε=0.2
2.5 ε=0.3
ε=0.4
2 ε=0.5
v
1.5
0.5
0
0 10 20 30 40 50
T
Fig. 9. Sinusoidal wave: plug velocity versus time; Bn = 5, f = 0.05

3
ε=0
ε=0.1
2.5
ε=0.2
ε=0.3
2 ε=0.4
ε=0.5
1.5
v
0.5
0
0 1 2 3 4 5
T
Fig. 10. Sinusoidal wave: plug velocity versus time; Bn = 5, f = 0.5
80
70
ε=0.1
60 ε=0.2
ε=0.3
50 ε=0.4
ε=0.5
δ (%)
40
30
20
10
0
0.0 0.5 1.0 1.5 2.0
f
Fig. 11. Sinusoidal wave: δ versus frequency
5. CONCLUSIONS
Unsteady motion of a Bingham fluid in coaxial pipes has been examined

numerically. The momentum equation has been written using a regularized
constitutive equation proposed by the authors, based on the error function. The
solution has been carried out via an implicit finite difference method; numerical
experiments indicated that the maximum value for the parameter of the rheological
η −η
model in order to obtain valid results is K = 5000 2 1 . If K exceeds this value,
Bn
the numerical procedure can exhibit instabilities. An uniform spatial grid of 1000
points has been adopted, using a time step equal to 4 ⋅ 10−6 . Four different motions
have been considered: start-up, cessation and two pulsatile flows. The analysis of
start-up flow has been useful to validate the numerical procedure comparing the
obtained asymptotic solution to the well known analytical one. In the cessation the
rate of flow decreases the more rapidly the greater Bn is: the stopping time which
is infinite for Newtonian fluids becomes finite for Bingham fluids and decreases
for increasing Bn . In pulsatile flow, applying a rectangular pressure gradient the
flow can completely stop for some time if the Bingham number is sufficiently
great. The superposition of a sinusoidal wave to the steady pressure gradient gives
rise to an increment of the mean discharge in a period, owing to the non linearity of
the constitutive equation.
References
1. Glowinski, R., Numerical methods for nonlinear variational problems, Springer-Verlag, New
York, 1984.
2. Papanastasiou, T.C., Flows of materials with yield, J. Rheol. 31 (5), 385-404, 1987.
3. Frigaard, I.A., Nouar, C., On the usage of viscosity regularisation methods for visco-plastic fluid
flow computation, J. Non-Newtonian Fluid Mech. 127, 1-26, 2005.
4. Glowinski, R., Sur l’écoulement d’un fluide de Bingham dans une conduite cylindrique, J. de
Méc. 13, 601-621, 1974.
5. Huilgol, R.R., Mena, B., Piau, J.M., Finite stopping time problems and rheometry of Bingham
fluids, J. Non-Newtonian Fluid Mech. 102, 97-107, 2002.
6. Chatzimina, M., Georgiou, G.C., Argyropaidas, I., Mitsoulis, E., Huilgol, R.R., Cessation of
Couette and Poiseuille flows of a Bingham plastic and finite stopping times, J. Non-Newtonian
Fluid Mech. 129, 117-127, 2005.
7. Duggins, R. K., The commencement of flow of a Bingham plastic fluid, Chem. Eng. Sci., 27,
1991-1996, 1972.
8. Papanastasiou, T.C., Boudouvis A. G.,Flows of viscoplastic materials: models and computation,
Comp. and Struct., 64, (1-4), 677-694, 1977.
9. Hammad, K. J., Unsteady pipe flows of a viscoplastic non-Newtonian fluid: effects of pressure
gradient waveform, Proc. ASME FED, 247,197-204, 1998.
10. Daprà, I., Scarpi, G, Start-up flow of a Bingham fluid in a pipe, Meccanica, 40, 49-63, 2005.
Formation of lime efflorescence on concrete and concrete

products
Tomáš Fojtík1, Jiří Brožovský2
1,2
Faculty of Civil Engineering, Brno University of Technology, Brno, 602 00, Czech Republic
Summary
Many specialists from all over the world occupied themselves with the problem of
efflorescence formation on concrete and on concrete products. This problem is
nearly as old as the concrete itself. The fact that to this very day, the problems of
efflorescence formation course and of formed efflorescence elimination is not yet
clearly defined up to now, shows how extensive and difficult is this task. This paper
concerns the problems of efflorescence formation on concrete and concrete
products depending on some conditions to which the concrete is exposed.
KEY WORDS: concrete, concrete product, lime efflorescence, primary, secondary
1. INTRODUCTION
Is the efflorescence on the surface of concrete the demonstration of its degradation?
This form of degradation is in contrast with the sulphate corrosion or with other
forms of corrosion not dangerous in light of structural properties and the
efflorescence has not major effect on the durability of concrete structures.
Nevertheless the efflorescence which occurs on the surface of concrete products
mostly in the form of white “spots” or layers is an esthetic defect which can
significantly influence the quality of concrete soffit products.
The efflorescence consists from inorganic salts or hydroxides which are leached
out from concrete (from the cement stone or from the aggregate). The hydrous
solution of these salts is transported by the system of capillary tubes in concrete. It
evaporates on the surface of the concrete and reacts with atmospheric CO2 or SO2
(produced in a high degree by the industrial production). The efflorescence is
formed largely in the form of salts (chiefly as sodium, potassium and mainly as
calcium sulphates or carbonates). The main component is as a rule CaCO3.
The occurrence of efflorescence is typical on the surface of concrete alternately
soaked and dried or exposed to the action of pressure water. Less frequent factors
of efflorescence formation on concrete surface are:
- The proportion of individual concrete components and their total concentration
in the bearing hydrous solution,
522 T. Fojtík, J. Brožovský
- The relative air humidity in the surrounding atmosphere (φ [%]). It depends on

the solubility degree of individual components and on RH whether the given
salts crystallize in the pores or on the surface of the material. The
crystallization in pores is one of the building materials significant destruction
factors [1, 2, 3 ].
2. LIME EFFLORESCENCE
It is formed by calcium carbonate (CaCO3) which is formed by the action of
atmospheric carbon dioxide (CO2) in humid medium on calcium hydroxide
(Ca[OH]2) liberated during maturing (hydration) of concrete [3].
2.1. Primary lime efflorescence

The primary lime efflorescence is formed only for a limited period from the
concrete or concrete products production by the reaction of calcium hydroxide (this
is the component of both, the hardened concrete and the hardened lime mortar)
with the atmospheric CO2 under moisture, in the surface layers of concrete
products.
2.2. Secondary lime efflorescence

The secondary lime efflorescence is formed during the whole service life of
concrete by reaction of the calcium hydroxide leached out on the surface of
products with the atmospheric carbon dioxide under presence of moisture.
By the effect of atmospheric CO2 the in water insoluble calcium carbonate is
transformed after a longer period in the in water soluble calcium hydrogen
carbonate Ca[HCO3]2.
The secondary efflorescence does not in contrast to the primary efflorescence form
only lighter spots on the surface of concrete but it constitutes on the surface
crystalline substance, which dusts off as removable dust.
Consequently, after this transformation of the calcium carbonate the lime
efflorescence disappears. The time of calcium carbonate conversion into calcium
hydrogen carbonate is approximately 1,5 till 3 years (in dependence on internal and
outer conditions, first of all on climatic effects) [2, 3].
3. MECHANISMS INFLUENCING FORMATION OF LIME

EFFLUESCENCE
The frequency of lime efflorescence formation on concrete depends on many
variables. These are first of all the water content (w and φ) and the content of the
atmospheric CO2 [3].
It is necessary to mention that no less than from 75 % the formation of all
efflorescence on concrete is caused by the “lack of technological discipline” in the
production and the remaining 25 % is the result of the concrete components
chemism.
3.1. The main conditions supporting the formation of lime efflorescence on

concrete:
- The high w/c ratio,
- Temperature,
- Porosity,
- Improper admixtures,
- Mixing and compacting conditions,
- Additives,
- Water tightness,
- Treatment conditions,
- Capillarity,
- Content of free lime,
- Proportion of cement and aggregate in concrete,
- CO2 content in the atmosphere,
- Time of products storing,
- Type of construction,
- The storing of products [1,2,3].
4. MEASURES ELIMINATING THE FORMATION OF LIME

EFFLORESCENCE IN THE PRODUCTION
4.1. Measures and methods limiting the efflorescence formation used in

production
• Elimination or limiting of causes mentioned in chapter 3.1,
• Utilization of additives which contribute to following possible effects:
- Interruption of capillary tubes and on this way the undesirable mass
transfer,
- Increase of water tightness, waterproofing of capillary pores,

- Chemical fixing of lime to insoluble product [3].
4.2. Means and methods removing the already formed efflorescence

As a rule the lime efflorescence can be removed sufficiently mechanically, for
instance by sweeping it up. Also special means can be used. The concentrated
preparation is diluted with water and after the concrete surface was saturated with
water and the superfluous water was removed, the diluted preparation is applied on
the surface of the product by means of a brush or sponge. After the efflorescence is
dissolved the surface is washed down with water.
If it is not possible to remove the lime efflorescence by one of readily available
products, it is possible to apply a special procedure. We can use for instance the
3% solution of hydrochloric acid, phosphoric acid or the solution of some organic
acid (for instance acetic acid or oxalic acid etc.) It is necessary to take into account
that in the case of lime efflorescence dissolution with acids the shade of concrete
products colour and the structure of the surface can change, because the acids
attack and disintegrate together with the efflorescence also the cement stone [1, 3].
5. EXPERIMENTAL MEASUREMENTS
5.1. Division of experimental works
Considering the fact, that the formation of efflorescence is influenced by quite a
number of factors, the most important effects were selected for testing, namely the
general effect of the surroundings and of the concrete composition.
5.1.1. Determination of efflorescence formation conditions

The aim was the laboratory testing of concrete samples of concrete with exactly
defined variant composition, exposed to different types of surroundings.
5.1.2. Determination of concrete composition on the formation of the efflorescence

Test pieces with exactly defined proportion of concrete components were prepared
in this phase. The test pieces were tested consequently for the dependence of lime
efflorescence formation on the type of used concrete, on the method of concrete
compacting during the production of these test pieces, on the proportion of cement
and aggregate and on the value of w/c ratio.
5.1.3. Determination of efflorescence formation conditions in dependence on the

composition of concrete
The aim was to make test pieces with the constant w/c ratio and with variable
proportion between cement and aggregate and to expose it to different types of
surroundings.
5.2. Methods of testing
5.2.1. Determination of conditions for efflorescence formation

The made concrete prisms were stored in following surroundings:
- Water bath (t = 20 ± 2°C),
- Laboratory surroundings (t = 20 ± 2°C, RH = 40 – 60 %),
- Air tight chamber with 98 % of CO2 concentration,
- Hot air drier (t = 50°C),
The influence of these surroundings was monitored on the formation of
efflorescence. The extent of concrete surface coverage with efflorescence was
determined and evaluated by optical microscopy and its mineralogical composition
was detected by means of X-ray diffraction analysis.
5.2.2. Determination of the concrete composition influence on the formation of

efflorescence
The made test pieces were exposed to surroundings in which the most
efflorescence was formed during the preceding phase and the effect of the concrete
composition was described on the efflorescence formation at varying:
- Kind of cement (locality),
- Way of concrete compacting,
- Proportion of cement and aggregate,
- Water-cement ratio.

concrete composition
The prepared concrete prisms with the cement and aggregate proportion 1:5 or 1:6
were placed into following surroundings:
- Climatic chamber – for the period of 24 hours, temperature 28°C and RH 95
%,
- Normal atmospheric surroundings (outdoor) – for the period of 24 hours,
- Drier – for the period of 6 hours, temperature 60°C.
The effect of climatic conditions on the formation of efflorescence was
subsequently monitored on these test pieces and the results were compared with
results obtained in the phase I (article 5.2.1.)
5.3. Results of individual phases of work
5.3.1. Determination of efflorescence formation conditions

All efflorescence which arises on the test pieces during this work is so-called
primary efflorescence. This is formed only during the limited period from the
concrete or concrete products production by the reaction of calcium hydroxide with
atmospheric CO2 in humidity on surface layers of concrete products.
The secondary efflorescence is formed by the effect of atmospheric CO2 and it is
transferred after longer period from the in water insoluble CaCO3 into in water
soluble calcium hydrogen carbonate Ca [HCO3]2. This process lasts longer time.
The test pieces made from the cement CEM I 42,5 R from the cement plant Mokrá,
and from the aggregate Bratčice with constant proportion of cement and aggregate
1:3 and with the water-cement ratio w = 0.5 with vibration and by manual
stamping. The test pieces were exposed for 28 days to individual surroundings and
we can state that:
- The formation of efflorescence does not take place in surroundings with
fluctuating pressure and relative air humidity with the temperature 50°C and
higher,
- In laboratory surroundings at t = 20 ± 2°C, constant atmospheric pressure and
in relative air humidity varying from 40 to 60% the efflorescence is
admittedly formed but not so quickly as in the water medium with the
temperature t = 20 ± 2°C. On the other hand the efflorescence which was
already on the surface of the test pieces before they were placed into this
surroundings increased moderately their volume,
- In the chamber with constant atmospheric pressure and the temperature
t=20±2°C with the content of 98% of CO2 is the efflorescence formed only in
small quantity, but when we have put into the chamber a sample which
already contained on the surface efflorescence the efflorescence quantity
increased (volumetrically, no the surface area),
- Unambiguously the most favorable conditions for the formation of
efflorescence is the aqueous medium with the temperature of t = 20 ± 2°C. In
this medium even after 14 days of storing always new efflorescence was
formed while in other media the growth of it was stagnating,
- The most important factor influencing the formation of efflorescence is
besides the temperature of surroundings the value of the relative humidity
(RH) and the rate of air circulation,
- Quick overlapping of the surface with efflorescence takes place after placing
the prisms in the chamber with elevated concentration of CO2 but in a definite
moment the formation of new efflorescence stops. The elevated CO2
concentration in the atmosphere results in the closing of pores on the surface
by the formed efflorescence and in the stagnation of its further formation,
- We observed the rapid increase of efflorescence after placing the concrete into
water bath at the temperature t = 20 ± 2°C for the period of at least 28 days,
- When we have taken out the samples from the water storing after 28 days,
have removed the efflorescence and placed the samples once more into this
surroundings, the growth of efflorescence did not start again.
5.3.2. Determination of the concrete composition influence on the formation of

efflorescence
- We can state that the intensity of efflorescence formation is not direct
proportional to the increasing value of water ratio, i.e. the value of water ratio
doesn´t affect the size of efflorescence formation,
- We can further state that with the increasing proportion of cement and
aggregate the intensity of efflorescence growth increases,
- The type of cement has an effect on the formation of efflorescence. The
intensity of concrete surface coating with efflorescence was in our case with
both types of cements nearly identical but in some cases we observed still a
higher intensity of efflorescence formation in the case of concrete made from
cement CEM I 42.5 R from the cement plant Mokrá. The contributing case
can be the higher content of CaO in the cement CEM I 42.5 R Mokrá in
comparison with the cement CEM I 42.5 R from the cement plant Hranice,
- The way of compacting in the concrete manufacture can have a significant
influence on the formation of efflorescence. The efflorescence during water
storage was formed practically on all samples, naturally the concrete
compacted by vibro-pressing was much more resistant against the formation
of efflorescence in comparison with concrete made only by ramming.

concrete composition
The test was carried out in laboratory conditions to precise the hypothesis
concerning the efflorescence formation in hot air driers which is connected with the
velocity of surface desiccation and with the water vapour diffusion. Two sets of
prisms with different ratio of cement and aggregate were placed into the hot air
drier, in the damping box and in the laboratory surroundings.
- In the phase I more efflorescence was formed on the concrete surface in the
case of the cement : aggregate proportion 1,5 contrary to the proportion 1:6,
- The main effort was of course to describe the processes in the hot air drier
during the period of 24 hours since the placing of samples. The rapid
desiccation of concrete during the first 8 hours caused intensive formation of
efflorescence. Of course the surface of concrete is after 24 hours practically
without efflorescence,
- The obtained efflorescence was submitted to X-ray diffraction analysis, which
confirmed the similarity of efflorescence concerning the mineralogical
composition. In cements of CEM I 42.5 R type from both the cement plant
Mokrá and the cement plant Hranice the prevailing mineral component in
efflorecence is calcite.
6. CONCLUSION
We can state following the theoretical analyses and practical realized tests that the
less favourable atmospheric conditions under which the formation of efflorescence
takes place in a high degree are the temperatures 0 – 5°C in the morning hours of
autumn months. The period of fresh concrete products storing has no effect on the
quantity of efflorescence, it is rather the measure against their formation which can
be easily realized in every manufacturing plant or during the production.
The first condition for the minimization of the lime efflorescence formation is the
dosing of optimal batch water quantity for perfect hydration and for achieving of
concrete high compactness with small rate of absorption. This can be achieved by
the addition of adequate additives and admixtures.
It is further necessary to prevent during the hardening of concrete both its
desiccation and the water condensation on its surface owing to the cold
surrounding atmosphere. This can be achieved by the use of so-called maturing
chambers during the hardening of concrete products.
The type of the used cement with different chemical and mineralogical
compositions can affect the formation of possible lime efflorescence.
The tests which we have performed verified the influence of some variables on the
possibility of lime efflorescence formation on concrete. It is necessary to remember
that the efflorescence formation is influenced by a number of further factors which
could not be included into this work.
The further development in this area should be directed towards the deeper
understanding of the efflorescence formation reasons under the effect of further
outer factors. This research should help to prevent or at least to minimize the
formation of efflorescence on concrete and in this way to increase the esthetic
value of concrete.
Acknowledgements
The work was supported by the MSM 0021630511 plan: Progressive Building
Materials with Utilization of Secondary Raw Materials and their Impact on
References
1. Matoušek, M., Drochytka, R.: Atmospheric Corrosion of Concretes, IKAS Praha, 1998, ISBN
80-902558-0-9.(in Czech)
2. Pytlík, P., Concrete Technology, Brno: Naše vojsko, 1994. ISBN 80-206-0434-0. (in Czech)
3. Kresse, P., Efflorescence-Mechanism of occurrence and possibilities of prevention, Research
Centre of the company BAYER AG Krefeld-Uerdingen, published in the Journal: Betonwerk-
Fertigteil-Technik (BFT).
The use of computer means

in representation of the ruled surfaces
Monica Gheorghiu and Mirela Chelcea
Descriptive Geometry, Engineering Graphics and Graphics Computer Programs Department,
Technical University of Civil Engineering, Bucharest, Romania
Summary
Teaching the Descriptive Geometry by the means of the computer is a recent

attempt to make the subjects more understandable and to facilitate the approach of
the students to some difficult subjects. This work makes a connection between a 3D
representation of some examples of this chapter and there orthogonal projections,
the final target of the Descriptive Geometry using the possibility of animation
provided by an adequate "software", for example the program "AutoCAD 3D".
KEYWORDS: ruled surfaces, Descriptive Geometry, orthogonal projections,

paraboloid, cylindroid, helicoid, hyperboloid.
1. INTRODUCTION
Teaching Descriptive Geometry is, without any doubt, unless a first for the system
of university Rumanian education, but a recent attempt to align itself with the last
tendencies of education. A particular effort of those who teach is presumed in order
to provide such a course of Descriptive Geometry, regarding the adaptation of the
means of presenting the matters using an adequate "software", for example the
program "AutoCAD 3D".This is why this work is a first step in attempting to build
a system of teaching and of evaluation based on the computer.
The subject of this work belongs to the applicable part of this matter and, although
it attracts the interest of the student, it is considered as having a high level of
difficulty. The presentation in 3D of the ruled surfaces, axonometric, apply the
intuition of onlooker and this is why provides a complete sight of the subject
however complicate would be.
The successive transformation of that, in images 2D, meaning in orthogonal
projections, can easily be done by using the Autocad 3D or other programs which
makes possible to use animation 3D. This manner of presenting simultaneously the
two methods to visualize can provide a certain easiness to discern the main
orthogonal projections in a unit space and can lead to obtain the capacity to
represent in orthogonal projection on any plane of the objects however complicated

they are.
The authors also consider that the 3D representation based on the computer of the
various types of ruled surfaces, as well as the connection between these views of
the examples suggested to be solved and their orthogonal projections could make
this subject not only understandable but also more attractive. The elements, which
compose certain surfaces, are proposed to be identified in the assembly and to be
represented in the usual system of projections.
The following stages were established in the presentation of the subject, in order to
fulfil the requests of the problem: the view of the final representation of the defined
surface, the 3D representation of it, the turn around of the 3D view, the orthogonal
projections due to this proceeding, the 3D and 2D view of the constituent elements
and finally, the way to form the surface.
Figure 1. Paraboloid – axonometric representation and projection on the plane [W]
2. EXAMPLES
2.1. Paraboloid
2.1.1. Paraboloid. First example
The figured paraboloid could have the following parameters: first directix – line
AB, second directrix – line CD, with A, B and D in [H] and C in [V]. AB is
parallel to Oy, BD parallel to Ox (see Figure 1).
532 M. Gheorghiu, M. Chelcea
2.1.2. Paraboloid. Second example

The application could also require the representation of the symmetrical surface in
relation with base plane (see Figure 2).
Figure 2. Paraboloid – axonometric representation and projection on the plane [W]
2.2. Conoid
2.2.1. Conoid. First example
This complex problem demands to draw the ruled surface with directrix I a
semicircle with the centre the point M and radius 30mm, which belongs to
the plane that makes a 600 angle to plane [H], and directrix II the segment
ST, 20 mm long, on a line perpendicular to [W].
Figure 3. Conoid – axonometric representation

Figure 4. Conoid – projection on the plane [W]
The other request is to represent also the symmetrical surface to the plane
that includes points s and T and is parallel to plane [V] and finally to solve
the problem of intersection between the rolled surface and plane [H], the
requirement could modify as follows: (see Figures 3 and 4).
Figure 5. Cylindroid – axonometric representation
Figure 6. Cylindroid –projection on the plane [V], [W] and [H]
2.2.2. Conoid. Second example

The surface which is suggested to represent is formed by two symmetrical conoids.
The first one has the first directrix in a plane inclined at an angle of de 600 to the
horizontal, composed of 2 circle arcs with a 65 mm radius, passing through points
S and T in [H], positioned between [H] and their point of intersection, above [H]
with S and T while directrix II segment MN, which belongs to a line which is
perpendicular to [V].
The second conoid is the symmetrical of the first relative to [W] passing through M
and N (see Figures 5 and 6).
Figure 8. Cylindroid – projection on the plane [V], [W] and [H]
2.2.3. Conoid. Third example

The surface of this exemple is made by three units, from among the first conoid has
the first directrix in a plane oriented at an angle of de 600 to the horizontal,
composed of 2 circle arcs with a 65 mm radius, passing through points S and T in
[H], while directrix II segment S1T1 from [H], (see Figure 7).
The second conoid has directrix I in a plane oriented at an angle of de 600 to the
S1 and T1, positioned between [H] and their point of intersection, above [H] while
directrix II segment S2T2 in [H].
The third conoid has directrix I in a plane oriented at an angle of de 600 to the
S2 and T2, positioned between [H] and their point of intersection, above [H] while
directrix II segment S3T3 in [H] (see Figure 8).
Figure 9. Proposed application. Cylindroid – axonometric representation
2.3 Cylindroid
2.3.1. Cylindroid. First example
The request of this problemis to make an axonometric representation of the
cylindroid with directrix I – two semicircles with centers B and D respectively, in a
plane parallel to [H], directrix II – two semicircles with centers b and d
respectively, in plane [H].
Figure 10. Cylindroid – the rotation of the surface until we see the projection to plane [W]
The semicircles with centers b and D are between plane [V] and a plane parallel to
[V] determined by points A, E and a. The semicircles with centers B and d are on
the other side of the plane determined by A, E and a, mentioned above (see Figures
9, 10, 11 and 12).
Figure 11. Cylindroid – the rotation of the surface until we see the projection to plane [V]
Figure 12. Cylindroid – the rotation of the surface until we see the projection to plane [H]
The problem may be extended by adding the request to repeat the obtained
cylindroid surface several times, or by rotating the surface and so changing the data
(see Figures 13 and 14).
2.3.2. Cylindroid. Third example

It is required to draw the ruled surface that has both curves circle arcs with a radius
of 30 and 50mm respectively, in a plane that makes a 600 angle to plane [H]. Point
M is the middle of the segment representing the cord sub-stretched by both
directrixes. It is also required to represent the symmetrical surface to this one

relative to the plane that includes point N and is parallel to plane [V] (see Figure
15).
Figure 15. Cylindroid – axonometric representation and projection on the plane [W]
2.3.3. Cylindroid. Fourth example

The problem requires the representation of a ruled surface that has the following
parameters: first directrix the segment ST, second directrix a circle arc with a
radius of 225mm and the corresponding cord SR. It is also required to represent the
surface that is symmetrical to it relative to the parallel plane [V] that includes
points S and T (see Figure 16).
Figure 16. Cylindroid – axonometric representation and projection on the plane [W]
2.3.4. Cylindroid. Fifth example

The surface to be represent is a cylindroid that has directrix I is a circle arc on the
semicircle with the radius 45 and the center S, in a plane parallel to [V], above [H],
inscribed between its point of intersection with [H] and point R on the semicircle,
directrix II is also a circle arc, in a plane perpendicular to [H], that makes with the
plane of directrix I an angle of 150, derived from a semicircle with a radius of 70,
above [H], that intersects directrix I in point R. Also draw this cylindroid’s
symmetrical relative to the plane of the first directrix (see Figures 17, 18, 19 and
20).
Figure 19. Cylindroid – the rotation of the surface until we see the projection to plane [H]
Figure 20. Cylindroid – the rotation of the surface until we see the projection to plane [W]
2.3.5. Cylindroid. Sixth example

The problem demands to represent a ruled surface with directrix I in a plane
parallel to [W] determined by points S and T, made up of circle arc derived from a
semicircle with a 20mm radius, centre S, with all points more distanced from [H]
than from S and the tangents to this circle, symmetrical to the sS segment, starting
from point T and T1 respectively, T’s symmetrical relative to s. Directrix II has the
same shape as directrix I, is parallel to it at a distance of 20 mm, in a plane parallel
to [W], at a distance of 90 mm from it. Figure also the ruled surface symmetrical to
this one relative to the plane of the second directrix. Figure also the ruled surface
which is symmetrical to the one obtained up to now, relative to the plane of the last
figured directrix (see Figure 21 and 22).
2.4 Helicoidal Surfaces

2.4.1. Helicoidal Surfaces. First example
The problem target is the axonometric representation of the group formed by 4
identical adjoining helicoidal surfaces (see Figures 23 and 24), with directrixes
belonging to a cylindrical surface generated thus:
▪ surface I, directrix I starting from point A and directrix II from B,
▪ surface II, directrix I starting from point B and directrix II fromC,
▪ surface III, directrix I starting from point C and directrix II from D,
▪ surface IV, directrix I starting from point D and directrix II from A.
Figure 23. Helicoidal surfaces – axonometric representation and projection on plane [W]
Figure 24. Helicoidal surfaces – axonometric representation and projection on plane [H]
The first rulings of each surface belong to plane [H] and make up a square ABCD
with a side of 30 mm. The plane containing the second rulings for each surface is at
a distance of 30 mm from [H], while these rulings together make up a square,
rotated to the previous one in [H] by 300. Each of the next ruling planes is at a
distance of 30 mm from the previous one and the square made up by these rulings
is rotated by 300 to the previous plane.
The last directrixes are in a plane situated at a distance of 120 mm from [H].
Figure 25. Helicoidal surfaces – axonometric representation and projection on plane [W]
Fig. 26. Helicoidal Surface – rotation of the surface until the projection to plane [H] is seen
2.4.2. Helicoidal Surfaces. Second example

It is demanded to represent the helicoidal surface with a directrix plane in plane
[H], which has directrix I a line perpendicular to [H] through the point H(60,60,0)
and directrix II a curve belonging to a conoidal surface, with the basis a circle with
center H and radius 48 mm. Each ruling is to be in a plane situated at a distance of
5 mm to the previous ruling’s plane, rotated by 450, and with a length 2mm less
than it (see Figures 25 and 26).
Figure 27. Hyperboloid – axonometric representation
Figure 28. Helicoidal Surface – the rotation of the surface until we see the projection to
plane [H]
2.5. Hyperboloid
2.5.1. Hyperboloid. First example
It is required to trace the ruled surface with both basic curves circles, directrix I is
the circle with centre S and radius 50 mm in [H], directrix II a circle with radius 30
and centre T in a plane parallel to [H] at a distance of 90 mm from it. The
divergence between the horizontal projections of the ends of the segment
representing a directrix, is 1200 for each directrix (see Figure 27 and 28).
2.5.2. Hyperboloid
If the ruled surface described by the above data, but with a divergence of 1500 is
represent (see Figure 29 and 30) some differences will be noticed, the main one
would be that the interior diameter is smaller in the second case than in the first.
Figure 30. Helicoidal Surface – the rotation of the surface until we see the projection to
plane [H].
3. CONCLUSIONS
The new manner of teaching provides not only a maximum efficiency of the study,
but, essential more, it radically changes the approach of the students to the
geometrical representation. It also makes them involved and responsible in the
relationship with their own training to fulfil the university conditions and
stimulates simultaneously their creativity.
References
1. Gheorghiu, M., Chelcea, M., Descriptive Geometry, Problems, MATRIX ROM Publisher, ISBN
973-755-015-3, Bucharest, 2006. (in Romanian).
2. Gheorghiu, M., Chelcea, M., Famous buildings as models for ruled surfaces, The 2nd
International Conference on engineering graphics and design – ICEGD - Galaţi, Romania – 2007.
Numerical studies on the seismic performance of three structural

systems
Horaţiu Mociran, Eugen Panţel
Faculty of Civ. Eng., Technical University, Cluj-Napoca, 400027, Romania
Summary
For many years, the basic intent of the building code seismic provisions has been
to provide buildings with an ability to withstand intense ground shaking without
collapse, but potentially with some significant structural damage.
The damage to buildings, transportation structures and lifelines wrought by the
1989 Loma Prieta earthquake, the 1994 Northridge earthquake, and the 1999
Hanshin earthquake near Kobe, has forced structural engineers, disaster response
agencies and building officials to carefully consider seismic response of the
building environment in terms of performance rather than life-safety.
Civil engineering structures normally rely on their ability to dissipate energy to
resist dynamic forces such as strong earthquakes. In recent years, to keep the
vibration of these structures within the functional and serviceability limits and to
reduce structural and architectural damage caused by extreme loads, different
passive protective systems have been proposed. Addition of energy dissipation
devices (EDDs) is considered one of the viable strategies for enhancing the seismic
performance of building structures. For many building structures, EDDs may
provide considerable performance improvement or cost saving.
This paper provides a comparison of the performance indices of a three story
building with three different structural systems – moment frame, viscously damped
frame and a base isolated frame. Fluid viscous damping and base isolation have
the same objective of significantly decreasing the response of a structure to
earthquake excitation. With both fluid viscous damping and base isolation it is
possible to have a structure remain within the elastic region, so there is no
permanent deformation from a seismic event. Each of the models were analyzed as
linear structures and subjected to time histories for 3 different earthquakes of
Vrancea type. The non-dimensional performance indices considered for the models
are: Peak Drift ratio, Peak Base Shear and Peak Level Acceleration.
In summary, the viscously damped frame has the best overall relative performance
of the three framing schemes. The base isolated frame is better than moment frame.
KEYWORDS: steel frames, passive control, base isolation, viscous damper, time
history analysis.
1. INTRODUCTION
Earthquakes represent one of the most damaging disasters. Repairing an area that
has experienced an earthquake is an expensive and difficult process. Engineers
have introduced different techniques to prevent a large amount of the damage
typically caused by earthquakes. The use of passive energy dissipation devices
represents a feasible alternative to improve seismic behavior of structures by
reducing the structural damage resulting from environmental disturbances [1].
Damping is one of many different methods that have been proposed for allowing a
structure to achieve optimal performance when is subjected to an earthquake. The
level of damping in a conventional elastic structure is very low, and hence the
amount of energy dissipated during transient disturbances is also very low. During
strong motions, conventional structures usually deform beyond their elastic limits,
and eventually fail or collapse.
The concept of supplemental dampers added to a structure assumes that much of
the energy input to the structure will be absorbed by supplemental devices [2]. An
ideal damper will be able to reduce both stress and deflection in the structure. Fluid
viscous dampers operate on the principle of fluid flow through orifices. A stainless
steel piston travels through chambers that are filled with silicone oil. The pressure
difference between the two chambers cause silicon oil to flow through an orifice in
the piston head and seismic energy is transformed into heat.
Seismic isolation is another alternative for protecting structures from earthquakes.
The philosophy behind this method is based on minimizing the earthquake induced
forces transferred to the superstructure. During the past 20 years, seismic isolation
has emerged as one of the most promising retrofitting strategies for improving the
seismic performance of existing buildings. It is also an attractive approach for new
construction when conventional design is not suitable or economical [3], [4].
Japan, New Zealand, and a number of European countries pioneered the use of
seismic isolation in civil engineering structures. More recently, the United States
has begun to implement isolation technology in bridges. In the seismic isolation
approach, the superstructure mass is uncoupled from seismic ground motions. This
is also referred to as "superstructure" isolation. It uses special types of bearings
called "seismic isolation bearings," which are placed below the superstructure and
on top of the substructure (foundation). In the event of a strong earthquake, they
add flexibility to the structure, by elongating its period and dissipating energy. It
could also give rise to large relative displacements across the isolator interface; this
can be controlled by incorporating damping elements in the bearing or by adding
supplemental dampers.
546 H. Mociran, E. Panţel
2. THE STRUCTURE, DAMPERS, BASE ISOLATORS, LOADS

2.1. The structure
Because the majority of buildings in Romania are less than three stories in height, a
three story steel frame was selected as a reference frame. The structure was
designed as a conventional SMRF (Figure 1) to provide a benchmark for seismic
performance comparison with passive control systems. The following design
parameters according to P100-1/2006 were used to design SMRF [5]:
- importance class II γ=1,2
- design ground acceleration ag=0,24cm/sec2
- TC=1,6 sec
- behavior factor q=6
All connections between beams and columns are assumed rigid connections.
Figure 1. 3 Storey Moment Frame
The structure was then modified by the addition of fluid viscous dampers (Figure
2) or base isolators (Figure 3) to improve the seismic performance, with no attempt
made to redesign the main frame elements.
Figure 2. 3 Storey Frame with Linear Viscous Dampers
Figure 3. 3 Storey Frame with Base Isolators
2.2. The dampers
The linear viscous dampers (Figure 4) will be installed with a diagonal brace
configuration. The inherent damping ratio of the structure is assumed to be 5%, and
the total effective damping ratio of the whole system is designated at 20% of
critical.
The damping cofficient used for all dampers is c=245kN-sec/m. In this example,
the linear effective stiffness is set to zero so that pure damping behavior is
achieved.
Figure 4. TAYLOR DEVICES Viscous Damper
2.3. The base isolators
The isolators are of elastomeric type (Figure 5) and have been chosen from FIP
INDUSTRIALE catalog, according to the vertical reactions requirements [6].
Figure 5. FIP INDUSTRIALE Base Isolator
The selected isolators SI-S 400/78, have the following specifications:

Mechanical properties: Vertical load V=570kN; Horizontal stiffness=0.64kN/mm.
Geometrical characteristics: Diameter D= 400mm; Total height Z=450mm.
2.4. The loads
The seismic performance of these structural systems was studied using linear
response-history analysis. Three scaled artificial earthquakes of Vrancea type, that
matched on average a P100-1/2006 Provisions response spectrum were used for
analysis.
3. NUMERICAL RESULTS
The results of the time history analyses are presented and discussed. Comparisons
are made of estimated base shear, interstory drift and floor acceleration.
Figure 6. Base Shear for Vrancea type earthquake
Moment
Frame
1.0
Base
Isolated
0.9
0.8
Base Shear/ Moment Frame Base Shear
0.7
0.6
0.5
Viscous
Damped
0.4
0.3
0.2
0.1
0.0
Structure Type
Figure 7. Normalized Maximum Value of Base Shear

3.1. Base shear
Figures 6 and 7 show a comparisons of base shears for the three structural schemes
that were studied. Viscous dampers are extremely effective in reducing the base
shear of the frame, whereas the addition of base isolators results in a modest
reduction in base shear.
3.2. Interstory drift
This is a code design parameter and is something most engineers focus upon during
the design process. From a damageability perspective, it is a measure that impacts
damage to the framing system, building façade and windows, piping, ductwork and
partitions.
Figure 8 Top floor displacements
11.5 Roof
Moment Frame
Viscous Damped
Base Isolated
nd
8.0 2 Floor
Story Height (m)
st
4.5 1 Floor
0.0 Ground
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Interstory Drift %
Figure 9 Interstory Drift

Results proved that control methods were successful in reducing the floor
displacements of the structure (Figure 8). The viscously damped frame reduced the
maximum drift by 63.30%, whereas the base isolated frame reduced the maximum
drift by 23.90%. Figure 10 shows a comparison of maximum values of interstory
drift for all framing schemes, normalized to the moment frame.
Moment
Frame
1.0
0.9
Base
Isolated
0.8
0.7
Peak Drift/ Moment Frame Drift
0.6
0.5
Viscous
Damped
0.4
0.3
0.2
0.1
0.0
Structure Type
Figure 10 Normalized Maximum Value of Interstory Drifts
3.3. Floor accelerations
This parameter is almost never taking into account in the design process, because it
requires a time history analyses to obtain it. From the damageability perspective it
is the measure that impacts damage to the ceiling and lights, electrical and
mechanical equipment, elevators and the building contents.
Figure 11 Top floor accelerations

In assessing the acceleration performance, the viscously damped frame has the best
performance (Figure 11). The viscously damped frame reduced the maximum floor
acceleration by 66.90%, whereas the base isolated frame reduced the maximum
floor acceleration by 9.20% (Figure 12).
Moment
Frame
1.0
Base
Isolated
0.9
Peak Floor Accn./ Moment Frame Accn.
0.8
0.7
0.6
0.5
Viscous
0.4
Damped
0.3
0.2
0.1
0.0
Structure Type
Figure 11 Normalized Peak Floor Accelerations
3.4. Performance indexes
For each control approach considered (i.e. viscous damper and base isolation), the
uncontrolled response of the structure was compared with the controlled response
[7].
The non-dimensional performance indexes considered in this study are Peak Base
Shear (PI1), Peak Drift ratio (PI2), Peak Level Acceleration (PI3).
Fbmax δ max &x&max
They are defined as follows: PI1 = ,c
; PI 2 = cmax ; PI 3 = cmax ; where:
max
Fb ,u δu &x&u
Fbmax max
,u ; Fb ,c - the maximum base shear of the uncontrolled/controlled frame;
δ umax ; δ cmax - the maximum inter storey drift ratio of the uncontrolled/controlled
frame;
&x&umax ; &x&cmax - the maximum absolute floor acceleration of the uncontrolled/controlled
frame.
The results presented in Table 1 show that the viscouly damped frame was the most
effective in reducing the building’s response.
Table 1. Performance index for peak base shear, peak drift and peak acceleration
Cases Performance Index % Reduction
PI1 PI2 PI3 Base Shear Story Drift Acc
Viscous Damped 0.367 0.367 0.331 63.30 63.30 66.90
Base Isolated 0.869 0.761 0.908 13.10 23.90 9.20
4. CONCLUDING REMARKS
The objective of the paper is to evaluate the efficiency of two passive control
systems. This paper has presented a comparative study of seismic performance
(base shear, interstory drift and floor acceleration) of a three story steel building.
The three story moment resisting frame was designed according to P100-1/2006
Provisions and than retrofitted with viscous dampers and base isolators. The
structure was subjected to linear time history analyses for different earthquakes of
Vrancea type.
The above numerical results presented for the three story one bay frame can not be
generalized. The analyses show that the seismic performance of the viscously
damped frame significantly exceeds that of the other framing schemes. The three
story base isolated frame had better performance than the three moment frame.
These parameters indicate that structural and non-structural damages are
significantly reduced in the case of viscously damped frame.
The results emphasis that seismic performances of the building depend upon the
structural framing scheme chosen for that building.
References
1. Miyamoto, H.K., Singh, J.P., Performance of Structures with Passive Energy Dissipators,
Earthquake Spectra, 2002.
2. Haskell, G., Lee, D., Fluid Viscous Damping as an Alternative to Base Isolation, Technical
papers from Taylor Devices, 2004.
3. Alhan, C., Gavin, H., Parametric Analysis of Passive Damping in Base Isolation, 16th ASCE
Engineering Mechanics Conference, Seattle, 2003.
4. Alexa, P., Mociran, H., Mathe, A., Passive control of semirigid steel structures, Computational
Civil Engineering, Iasi, 2005.
5. P100-1/2006, Code for seismic design of residential buildings, agrozootehnichal and industrial
structures. (in Romanian).
6. www.fip-group.it
7. Rama Raju, K., et al., Optimum Distribution of Viscous Fluid Dampers in Multi-Storied
Buildings, International Conference on Structural Stability and Dynamics, Florida, 2005.
Identification Methods of the Nonlinear Systems

Subjected to Seismic Actions
Doina Stefan, Violeta-Elena Chitan
1
Structural Mechanics, Technical University “Gh. Asachi” Iasi, 700050, România
Summary
The need to identify the nonlinear systems subjected to stochastic actions leaded to
the development of the classical and/or deterministic methods toward some
methods capable to take into account the improved models for structure and
excitation.
The identification of the nonlinear systems can be carried out using specific
approaches that are based on the reparability hypotheses or on the change of the
nonlinear system with an equivalent linear system that has the features closely
related to the initial one.
Current nonlinearities that occur in the structural response during a strong
earthquake are geometrical and/or physical (nonlinearity of the stiffness and the
damping). In order to underline the nonlinearity of the stiffness, in this paper we
take advantage of the modeling method for the structural response using a non-
stationary linear model and the parameter identification of the equivalent
stationary linear model, both for the SDOF and the NDOF systems.
Explaining the nonlinearity of the stiffness can be done in two specific ways: the
structure response modeling through the nonstationary linear model and the
parameter identification of the equivalent stationary linear model; the real stiffness
valuation of the structure during each vibration cycle using the internal force
calculus of the structure.
KEYWORDS: nonlinear system, stationary linear model, nonlinearity of the

stiffness, structural response.
1. INTRODUCTION
In the context of continuous efforts to improve the knowledge in the field of

dynamic-seismic behavior of a building, the direction is toward the study of
structural behavior starting from the tests in situ.
This kind of approach gives the means for: structural quality and security level
checking in the conditions of real exploitation; validation of the existent analytical
models of behavior, also for the a-priori hypotheses of the structural response
prediction; knowledge accumulation, necessary to improve the dimensioning
methods and the well argued working out of the new behavior models.
Knowing precisely or often good enough the features of a system implies the
discerning use of various theoretical and experimental approaches. The system
identification doesn’t mean anything else but the convenient processing of some
information achieved from the tests, in order to obtain a physical or mathematical
model, as sample and easy-to-use as possible.
An identification process consists of: the choice of the model structure; the choice
of the criterion to compare the model with the real structure; estimation of the
model parameters.
The need to identify the nonlinear systems subjected to stochastic actions leaded to
the development of the classical and/or deterministic methods toward some
methods capable to take into account the improved models for structure and
excitation. For this reason, the non-linear system identification can be carried out
using specific approaches that are based on the reparability hypotheses or on the
change of the nonlinear system with an equivalent linear system that has the
features closely related to the initial one.
earthquake are physical and/or geometrical. In the case of physical non-linearity
the equations of motion settled in the hypothesis of visco-elstic linear behavior do
not remain valid already, and generally, the forces that occur are expressed using
non-linear functions, with respect to displacement and velocity, this leading to the
non-linear stiffness and non-linear damping.
To underline the non-linear stiffness, two methods are usually carried out: the
structural response modeling using a non-stationary linear model and the parameter
identification of the equivalent stationary linear model; the estimation of the
effective structural stiffness during each cycle of vibration using the computation
of internal structural forces.
2. THE EQUIVALENT LINEARIZATION METHOD
Using the fact that linear systems are solved easily, the study of the nonlinear
systems can be performed with sufficient accuracy by changing the nonlinear
system with an equivalent linear system having the features closely related to the
initial one. This technique of linearization for the systems subjected to stochastic
actions was proposed for the first time by Booton in 1953.
556 D. Stefan, V.E. Chitan
2.1. SDOF Systems

The simplest mechanical system is the SDOF oscillator, which has the equation of
motion as follows:
my&&(t ) + h( y , y& ) = f (t ) (1)
where:
y are denoting the displacement, velocity and acceleration;
y , y& , &&
denotes a function that describes the nonlinearity of the oscillator;
h ( y , y& )
f(t) denotes the stochastic action.
The linearization method consists of the change of nonlinear oscillator with an
equivalent linear one. Equation (6.6) becomes:
my~
&& (t ) + c ~ & ~
eq y ( t ) + k eq y ( t ) = f ( t ) (2)
where:
y are denoting the displacement, velocity and acceleration of the
y , y& , &&
equivalent linear system;
ceq denotes the viscous damping of the equivalent system;
keq denotes the stiffness of the equivalent linear system;
(ceq and keq must be found out).
It is normaly to say that the nonlinear system and the equivalent linear system do
not have identically motions, so the solutions y , y& , &&y do not verify exactly the
equation (2), this becoming:
my~
&& (t ) + h( ~
y, ~
y& ) = f (t ) + ε (t ) (3)
where: ε(t) is the error that allows the measurement of the difference between the
equations (1) and (2).
ε ( t ) = h( ~y , ~y& ) − c eq ~y& (t ) − k eq ~
y (t ) (4)
In order to find out ceq and keq, the error ε(t) must have the minimum value. This
method suggests a minimization process for ε(t), starting from the identity:
∂E[ ε (t )2 ] ∂E[ ε (t )2 ] (5)
= = 0 (E[ε (t )2 ] − math. operator)
∂ C eq ∂ K eq
Developing the equation (5) we obtain the system:

E [h ⋅ ~
y ]E[ ~
y& 2 ] − E [h ⋅ ~y& ][ ~y⋅~ y& ]
K eq =
E[ ~y 2 ]E[ ~y& 2 ] − E [ ~
y⋅~ y& ] 2
(6)
E [h ⋅ ~
y& ]E [ ~
y 2 ] − E [h ⋅ ~y ][ ~y⋅~ y& ]
Ceq =
E[ ~ y 2 ]E[ ~y& 2 ] − E [ ~
y⋅~ y& ] 2
where:
h is as a matter of fact h(y,y) & ;
~ ~ ~
& ~
& ~ ~
&
E[ y , h( y , y )] and E[ y , h( y , y )] can be expressed as functions of y and
y& and ceq and keq are functions depending on the terms of the covariance matrix,
denoted cov[Y] and having the following format:
 E[ ~
y, ~y& ] E[ ~
y, ~
y& ]
cov[Y ] =  ~ ~ (7)
& ~& ~ 
 E[ y , y ] E[ y , y ]
2.2. NDOF Systems

The nonlinear NDOF system has the following equation of motion:
My&&(t ) + h( y , y&) = f (t ) (8)
where:
f(t) is a representative gaussian process of the excitation;
W(t) is a stationary gaussian process;
g(t) is a vector with terms as functions that are measurable in time;
y is the displacement vector.
f ( t )= W( t ) ⋅ g( t ) (9)
The equivalent linear system has the form:
My~
&& + C ~ & ~
eq y + Keq y = f ( t )
(10)
where:
Ceq is the equivalent damping matrix that depends on the time (because g(t) is
not constant);
Keq is the equivalent stiffness matrix.
The vector of errors is defined by the relation:
ε ( t ) = h( ~y , ~y& ) − Ceq (t ) ~
y& − Keq (t ) ~
y (11)
This vector is computed as a function of variables of the linear system and it is

minimized using the least square criterion. The Cij(t) and Kij(t) coefficients from
the Ceq(t) and Keq(t) matrices are computed from:
∂E [ εT ( t )ε( t ) ]
=0 i = 1, N ; j ≤ i
∂ Cij ( t ) (12)
∂E [ εT ( t )ε( t ) ]
=0
∂ K ij ( t )
Using equation (11) we can write:
E[h( ~
y, ~
y& ) ~
y T ] − Ceq (t ) E[ ~
y& , ~
y& T ] − Keq (t ) E[ ~
y, ~
y& T ] = 0
(13)
E[h( ~
y, ~
y& ) ~
y& T ] − C (t ) E[ ~
eq y, ~ y T ] − K (t ) E[ ~
eq y, ~
yT ] = 0
From above, the Cij(t) and Kij(t) terms can be obtained.

The excitation is a gaussian one, this leading to a gaussian response and allowing
[ ] [ ]
to write the E h( ~y , ~y& ), ~y T and E h( ~y , ~y& ), ~y& T terms as functions of the Y(2n x 2n)
[ ]
covariance matrix terms. The computation of the E h( ~y , ~y& ), ~y T and E h( ~y , ~y& ), ~y& T [ ]
terms is easily carried out if we assume the fact that the displacements and
velocities are two by two independent.
 ∂h (~
y, ~
y& )   ∂h (~
y, ~
y& ) 
Cij (t) = E i ~  Kij (t) = E i ~&  (14)
 ∂y   ∂y 
 E [ ~~
yy T ] E [ ~~
yy& T ]
Y =  ~& ~ T & ~ T ]
(15)
 E [ yy ] E [ ~ yy
The linear system has n(n+1)/2 unknowns, which are the Cij(t) and Kij(t) i=1,n
terms and they are computed with respect to the terms of the Y(t) matrix.
3. NONLINEAR SYSTEM IDENTIFICATION. THE MODEL OF

THE NONLINEAR OSCILLATOR
It is considered the fact that the measured response at the roof level can be modeled
using the nonlinear oscillator, which has the equation of equilibrium as follows:
y + f ( y , y& ) = − py&&ter
&& (16)
Assuming for the mass and the damping to remain constant, and for the stiffness to
be variable, the force f has the expression:
f = c0 y& + k 0 (t ) y (17)
where:
co is the damping constant over the unit of mass;
ko(t) is the structural stiffness over the unit of mass.
The equation (16) can also be written:
&y& + c0 y& + k 0 (t ) y = − p&y&ter (18)
In order to compute the response of the assumed model, the non-linear differential
equation (17) must be solved for each iteration step of the numerical algorithm for
minimization.
The equation of equilibrium in incremental variables is:
∆&&
y + c 0 ∆y& + k 0 (t ) ∆y = − p∆&&
y ter (19)
where:
∆&&
y = &&(
y t + ∆t ) − &&(
y t)
∆y& = y&(t + ∆t ) − y&(t )
(20)
∆y = y (t + ∆t ) − y (t )
∆&&
y ter = &&
y ter (t + ∆t ) − &&
y ter (t )
There are many ways to solve the equation (19). The best one is the step-by-
step approach. The dynamic equilibrium is stable at the beginning and at the
end of each time step ∆t, inside in which it is assumed a linear variation of
acceleration and the structural features remaining constant during all the
range (fig.1).
Figure 1. The dynamic equilibrium
The incremental values for velocity and displacement are:

∆t
∆y&(t ) = &&(
y t ) ∆t + ∆&&(
y t) (21)
2
∆t 2 ∆t
∆y (t ) = y&( t ) ∆t + &&(
y t) + ∆&&(
y t) (22)
2 2
Using the incremental displacement as a basic variable, one can obtain the
incremental acceleration from equation (22), which is after substituted in equation
(21), this leading to:
6 6
∆&&(
y t) = 2 ∆y (t ) − ∆y&(t ) − 3&&(
y t) (23)
∆t ∆t 2
3 ∆t
∆y& ( t ) = ∆y( t ) − 3y& ( t ) − &&(
y t) (24)
∆t 2
2
Substituting equations (23) and (24) in equation (19) we obtain:
 6 6  3 ∆t  (25)
 ∆t 2 ∆y (t ) − ∆t 2 ∆y& (t ) − 3 &y&(t ) + K0  ∆t ∆y (t ) − 3 y& (t ) − 2 &y&(t ) + K 0 (t ) ⋅ ∆y (t ) = − p&y&ter
   
or:
~ ~
K (t ) ⋅ ∆y (t ) = ∆&&
yter (26)
where:
~ 6 3
K (t ) = 2 + C0 + K0 (27)
∆t ∆t
~ 6  ∆t  (28)
∆&y&ter = − p∆&y&ter + y& (t ) + 3 &y&(t ) + C0 3 y& (t ) + &y&(t )
∆t  2 
Therefore, the incremental displacement is computed starting from equation (26)
and dividing the effective incremental load by the effective stiffness.
4. CONCLUSIONS
A nonlinear system can be defined as a system that doesn’t obey to the

superposition principles. This “non-property” occurs to the extremely various
systems for which it doesn’t exist general solving methods, but only special
methods, adapted to each class of problems.
The identification of the nonlinear systems can be carried out using specific
approaches that are based on the reparability hypotheses or on the change of the
nonlinear system with an equivalent linear system that has the features closely
related to the initial one.
earthquake are geometrical and/or physical (non-linearity of the stiffness and
damping). In this paper we take advantage of the modeling method for the
structural response using a non-stationary linear model and the parameter
identification of the equivalent stationary linear model, in order to underline the
nonlinearity of the stiffness.
References
1. Afra H., P.Argoul, Identification du comportement sismique des batiments et comparaison avec
les donnees reglementaires, 1990, Annales des Ponts et Chausees, pag.50-65. (in French)
2. Eykhoff P., Identificarea sistemelor, 1977, Ed. Tehnica, Bucuresti. (in Romanian, translation
from English)
3. Guihot Pascal, Analyse de la reponse de structures non lineaires sollicitees par des sources
d’excitation aleatoire. These de doctorat de l’Universite de Paris VI, 1999. (in French)
4. Stefan Doina, Identificarea sistemelor mecanice cu aplica]ii la calificarea seismia a unor cladiri
si echipamente din componenta Centralelor Nuclearo-Electrice, Teza de doctorat, Univ. Tehnica
“Gh. Asachi”, ian., 1996, Iasi. (in Romanian)
Iaşi, Romania, May 25, 2007
GRAITEC – CAD and Analysis Software

for the Construction Industry
GRAITEC ADVANCE software
GRAITEC Roumanie, Bucharest, 051702, Romania
Summary
GRAITEC, a major worldwide software developer for CAD and structural design
solutions, with over 20 years of experience, offers integrated CAD / CAE solutions
for the construction engineering.
With the Graitec Advance software packages, the customers benefit of the best
integrated solutions for obtaining higher quality projects in a shorter time:
Advance Design – finite elements analysis software for reinforced concrete and
steel structures dimensioning, Advance Concrete - professional reinforced
concrete design software and Advance Steel - professional steel design software.
KEYWORDS: GRAITEC, Advance, Concrete, Steel, Design, CAD, CAE,

Software
1. ADVANCE DESIGN: PROFESSIONAL FINITE ELEMENT

ANALYSIS SOFTWARE FOR REINFORCED CONCRETE /
STEEL STRUCTURES DIMENSIONING
After several years of market research, the specialists concluded that 25% of
structural engineer project time consists in modeling, while the rest of 75% is spent
for results postprocessing, taking into account that the projects complexity
increases and the deadlines are shorter day by day.
GRAITEC proposes a superior solution to these needs: a new-generation software
developed with the most recent CAD / CAE technologies. Advance Design offers a
top solution for the structural design and dimensioning: easy and intuitive
modeling; powerful design engine; advanced functionalities for reinforced concrete
and steel structures dimensioning, verification and optimization; automated post-
processing of results and automated reports… all within a unique environment!
Fast modeling: advanced CAD tools

• Automatic tools for the input of structural
elements (portal frames, vaults)
• Materials and sections libraries
• Various CAD features ands functionalities
• Multiple viewports allowing to work
simultaneously with several view points and
display modes
• Advanced display options (filtered elements
display, display by colors, realistic rendering,
transparency...)
• Commands for automated creation of loads and
loads combinations according to standard
Advance Design – screen capture regulations
• An advanced reports generator
Meshing and finite elements analysis

• Advanced meshing engine: " grid" or
Delaunay mesh algorithm, finite
elements with 3, 4, 6 or 9 nodes,
progressive meshing, refinement tools
• Static, dynamic, non-linear analyses,
large displacements, generalized
buckling, etc.
• Extended solver capabilities due to
enhanced memory management
Modeling - meshing - postprocessing
• Various post-processing methods:
graphical results, result curves, stresses
diagrams, calculation reports
Reinforced concrete and steel design
Reinforced Concrete Design module:

• Calculation of theoretical reinforcement sections
• Determination of 3D interaction curves for all types
of sections
• Obtainment of the theoretical reinforcement ratios
Steel Design module:

• Verification of buckling, lateral buckling and
deflections
• Shapes optimization after specified parameters
Interaction curve for a column • Generate shape sheets for each element
564 GRAITEC ADVANCE CAD/CAE SOFTWARE
Thanks to the new GTC exchange format, Advance

Design interacts with Advance Concrete and
Advance Steel to insure the coherence of your
projects. When a section in Advance Design is
modified, the Advance Concrete or Advance Steel
user receives a notification and therefore will be able
to automatically update all the drawings!
2. ADVANCE CONCRETE: AUTOCAD BASED REINFORCED

CONCRETE DESIGN SOFTWARE
Advance Concrete was specifically developed for structural engineers and detailers
seeking a simple to use and completely integrated package in AutoCAD® to create
all their construction drawings. Perfectly integrated into AutoCAD®, providing a
vast range of structural elements tools and modeling functionalities, Advance
Concrete significantly accelerates the initial design phase of buildings and
generates all the necessary construction drawings.
Reinforced concrete structures modeling

• Easy 3D and 2D modeling
• Complete library of structural elements (slabs,
beams, columns, walls, foundations)
• Piles libraries, providing parameters according to
rules
• Complete library of openings (doors, windows,
niches), hole shapes (rectangular, circular,
polygonal), fabrics
• Reinforcement libraries (straight bars, stirrups,
rectangular and circular frames)
• Parametric reinforcement of common elements
(beams, columns, openings...)
• Reinforcing styles manager, that allows the
automatic reinforcement of several structural
elements
• 3D reinforcement viewer
• A large selection of automatic dimensioning and
Advance Concrete reinforcing styles locating tools
Creation of construction drawings

• Automated creation of formwork and
reinforcement drawings: elevations, perspectives,
sections, foundations drawings.
• Fast creation of bars and fabrics which are
automatically distributed on elements
• Automated creation of material lists and steel
ratios.
• Improved drawing quality and faster finishing with
the "dynamic block" concept Advance Concrete: detail and section
• Drag and drop of drawings on the paper plans
Intuitive and easy to use interface

• Isometric plan or perspective views
• Dynamic scales and rotation for views
• All AutoCAD functionalities for viewing and
modeling
• Toggle metric / imperial systems anytime
during the design process
• Real time preview of model modifications
• "Model Watch": a new technology that
controls the 3D model consistency in real
time, and warns immediately when errors
occur
Advance Concrete – screen capture
2. ADVANCE METAL: AUTOCAD BASED STEEL DESIGN

SOFTWARE
Developed for steel construction companies and design offices, Advance Steel is a
software for steel structural modeling (profiles and connections) that automatically
creates general arrangement drawings, fabrication drawings, bill of materials and
NC files. Advance is integrated in the AutoCAD® platform.
566 GRAITEC ADVANCE CAD/CAE SOFTWARE
Steel structures modeling

• Complete library of steel structural elements (straight
stairs, spiral stairs, ladders, railings...)
• Customizable automatic joints with the possibility to
create specific joints
• A vast range of materials, sections, bolts libraries
• Polygonal cuts on plate and beam (cope, along
centerline, polygonal, on UCS, with an angle)
• Placement of bolts, holes and shear studs
• Welded connections
Stairs example
Drawings creation
• Automated functions to create all the fabrication and general arrangement drawings starting from
the model
• Additional dimension lines (e.g. Arc, Angle….) can be created and modified
• The drawing layout may be customized with specific requirements using the Drawing Style
Manager.
• Automated creation of an isometric
view from the project, with many
different label possibilities.
• Function for automated view creation
on the drawing for special situations
• The details are automatically labeled
and dimensioned.
• Ability to change scale, clipping and
presentation of the details within the
drawing
• Creation of bills of material (BOM), Joints and section details
lists and NC files
Various Advance Metal features and facilities

• Automatic numbering of single parts and assemblies
• Checking functionalities: element collisions, tool access and
consistency of element marks (fabrication, delivery and
assembly).
• Centralized documents
• CIS/2 compliancy
Advance Steel – screen capture
• Advance Concrete and Advance

Steel can be used simultaneously
during the same AutoCAD® work
session: a unique environment for
general drawings creation!
3. GRAITEC ADVANTAGES
• Graitec Advance – design and detailing, completely integrated
solution for reinforced concrete and steel structures
• High quality projects finished in lesser time.
• Effective project assistance service offered by our specialists in
structural modeling, CAD and Design
• All the required resources of software research and development for
all construction engineering branches
Graitec advantages and • Periodical version updates of acquired software
solutions • Assistance for an optimal use of Graitec software considering the
particular needs of each customer
• The safety of your investment
Contact
For more information, please visit www.graitec.com or contact the Graitec Roumanie Sales
Department, Str. Samuil Vulcan No. 10, Sector 5, Bucharest, Tel./Fax: 021.410.01.19 /
021.410.01.24.
The use of the bispectrum for analysis dynamic parameters

of rail fastening
1
2
Summary
“Higher Order Statistics” (HOS) is extension of second-order characteristics such

as the auto-correlation function and power spectrum. HOS analysis is emerging as
a new powerful technique in signal analysis, offering insight into non-linear
coupling between frequencies and potential applications in many areas where
traditional linear analysis provides insufficient information. This contribution
describes the HOS theory and possibilities application to experimental data
acquisition from measurements of rail fastening parameters.
KEYWORDS: Higher Order Statistics, Cumulants, Bispectrum, Railway

Superstructure
1. INTRODUCTION
The basic claim put in particular components of the railway track is their functional
reliability and the interrelated minimum costs of their manufacturing, repair and
maintenance. Each part of the rail structure is exposed to an extensive static and
dynamic stresses. This fact is of great importance especially in the design and
construction of high-speed railway lines.
Different methods and various criteria have been applied to test the railway
superstructure, and different mathematical methods are used to evaluate the signals
measured. The method exciting the structure by a mechanical impact is often used
in dynamic tests. The excitation by impact is advantageous for the determination of
proper frequencies of a given system. Another possibility is the loading of the
structures specimens by a continuous excitation by means of vibrators. This makes
it possible to focus attention on certain frequencies or to the interval of the
frequencies. For the evaluation, frequency spectra transfer functions, frequency
response functions, etc. have been composed. From the viewpoint of the railway
operation security, it is expedient to focus our attention also on new methods that
will offer the information about the quality of particular components (or of the
whole structure) both from the viewpoint of different failures and from the
assembly viewpoint.
It should be said that until the present time practical applications of the evaluation
of the results of measuring by “elementary” methods such as typical statistic and
frequency analyses have prevailed. At the present time, there exist a number of
mathematical methods through which discrete data may be processed and which
are especially suitable for processing the signal measured. However, in the
technological practice these methods appear only rarely although these may offer
more information about the signals measured and also about the structure tested.
These methods could not be used because of low efficiency of the hardware and the
technology of measuring. At the present time, these methods are not fully utilised
owing to the insufficient awareness of technologists.
2. THEORETICAL BACKGROUND
“Higher Order Statistics” (HOS) is extension [1, 2] of second-order characteristics

such as the auto-correlation function and power spectrum to higher orders. HOS
analysis is emerging as a new powerful technique in signal analysis, offering
insight into non-linear coupling between frequencies and potential applications in
many areas where traditional linear analysis provides insufficient information.
The second-order analyses work fine if the signal has a Gaussian (Normal)
probability density function, but many real-life signals are non-Gaussian. The
easiest way to introduce the HOS measures is just to show some definitions so that
the reader can see how they are related to the familiar second-order measures. In
the text to follow, are definitions for the time-domain and frequency-domain third-
order HOS measures, moments, cumulants, high-order spectra or polyspectra, etc.
[3, 4].
2.1 Moments
The time process x(n) can be characterized in many ways, for example by its
amplitude, its energy or its waveform. The probability density function (pdf) of the
process provides a detailed information about the distribution of the amplitudes of
the process which can be used to characterize the process. A set of quantities which
describes the shape of this “pdf” are the moments. The first-order moment m1 of
process x(n) is just its mean, and it provides a measure of location of the „pdf”. The
second-order moment is the variance, a measure of the spread of the “pdf”. Higher-
order moments exist too, such as the skewness and kurtosis. Now, moments are
statistical measures which characterize signal properties. The first four moments of
the process are defined by the following equations:
m1 = E{x n } (1)
m 2 (τ ) = E{x(n ) ⋅ x(n + τ )} (2)
m3 (τ 1 ,τ 2 ) = E{x(n ) ⋅ x(n + τ 1 ) ⋅ x(n + τ 2 )} . (3)
m4 (τ 1 ,τ 2 ,τ 3 ) = E{x(n ) ⋅ x(n + τ 1 ) ⋅ x(n + τ 2 ) ⋅ x(n + τ 3 )} (4)
2.2 Cumulants
Cumulants are specific non-linear combinations of these moments. The first-order

cumulants of process is the mean. The second, third and fourth-order cumulants of
process are defined by the following equations [5]:
c 2 = m2 − m12 (5)
c3 = m3 − 3 ⋅ m2 ⋅ m1 + 2 ⋅ m13 . (6)
c 4 = m4 − 4 ⋅ m3 ⋅ m1 + 3 ⋅ m22 + 12 ⋅ m2 ⋅ m12 − 6 ⋅ m14 (7)
2.3 Polyspectra
This term is used to describe the family of all frequency-domain spectra, including
the those of the 2nd order. Most HOS work based on polyspectra and focus their
attention on the bispectrum (third-order polyspectrum) and the trispectrum (fourth-
order polyspectrum). Polyspectra consist of higher order moment spectra and
cumulant spectra and can be defined for both the deterministic signal and random
processes. Moment spectra can be very useful in the analysis of deterministic
signals (transient and periodic), whereas cumulant spectra can play a very
important role in the analysis of stochastic signals. The kth-order polyspectrum
[3, 5] is defined as the Fourier transform of the corresponding cumulant sequence:
∞
C2 ( f ) = ∑ c (k ) ⋅e
2
−i 2πfk
(8)
k = −∞
∞ ∞
C3 ( f1, f 2 ) = ∑ ∑ c (k , l ) ⋅e3
− i 2πf 1 k
⋅ e − i 2πf 2 l . (9)
k = −∞ l = −∞
∞ ∞ ∞
C 4 ( f1 , f 2 , f 3 ) = ∑ ∑ ∑ c (k , l , m ) ⋅ e
4
− i 2 π ( f 1k + f 2 l + f 3 m )
(10)
k = −∞ l = −∞ m = −∞
which are the power spectrum, respectively the bispectrum and trispectrum. These
can be estimated in a way similar to the power spectrum, but more data is usually
needed to get reliable estimates. Note that the bispectrum is a function of two
frequencies, whereas the trispectrum is a function of tree frequencies. In contrast to
the power spectrum which is real-valued and non-negative, bispectrum and
trispectrum are complex valued and contain a phase information.
Bispectrum a technique that is often used, can be estimated in a way similar to
Welch periodogram method for the power spectrum estimation but the lengths of
data required to obtain consistent estimates are longer than those required for the
power spectrum estimation.
The variance of the estimate can still be reduced by averaging over multiple
segments of data. To extract an information from the non-normalized bispectrum is
often difficult because the variance of the bispectrum estimate at a particular
frequency pair (k,l) depends not only on the data length but also on the power of
the signal at frequencies k, l and k+l. The bicoherence is a useful normalized
bispectrum. The bicoherence at any frequency pair k, l can be interpreted as a
fraction of power at frequency k+l which is a phase coupled to the component at
k+l. Bicoherence is estimated as [4]
C3 ( f1 , f 2 )
2
b( f1 , f 2 ) == (11)
P( f1 ) ⋅ P( f 2 ) ⋅ P( f1 + f 2 )
where C3(f1, f2) is the estimate of the bispectrum, and P(f) is the estimate of the
power spectrum.
3. EXPERIMENTAL RESULTS
The text to follow presents an example of the laboratory measurements and the
analyses of the dynamic parameters of the rail fastening specimen. The
measurements were carried out mainly with respect to the transfer properties of this
fastening of rails under different assembly and operational states (the effects of
tightening and releasing).
The test specimen was assembled of a part of the concrete sleeper B 91, on which a
rail of structural shape UIC 60 using the elastic fastening of VOSSLOH SKL14
type was fastened. To test the specimen of the rail grid, the method of measuring
the response to the random excitation was used. The excitation was implemented in
the horizontal direction towards the rail head by means of an efficient vibration
exciter B&K 4818. The frequency range of the observed area was determined
ranging from 5 Hz to 1500 Hz.
The measuring system PULSE 3560C by the firm Brüel& Kjaer was used for the
experiment control. The acceleration responses were measured by an
accelerometric sensing unit fixed to the sleeper 10 cm from the fastening. Let us
remark that the acceleration sensing unit was fixed to the structure measured by
means of the bee-wax.
The measurement resulted in the scanned and digitally recorded electrical signal
proportional to the instantaneous value of the acceleration in place of the sensing
unit fixation. Frequency spectra and bicoherence were developed from the time
responses.
The graphs in Figures 1 and 3 represent a dynamic response of the rail fastening
with the standard tightened bolts. Figure 1 shows a graph of the amplitude
spectrum calculated by means of a direct application of Fourier´s transformation to
the signal measured. Significant components on frequencies 15 Hz, 20 Hz, 75 Hz,
95 Hz, 150 Hz, 220 Hz and 550 Hz may be read from this graph. It is evident from
this graph that particular frequency peaks are relatively uniformly distributed in the
frequency region. Figure 3 shows the graph of the bicoherence. There are no
distinctive maximum values noticeable here.
The graphs in Figures 2 and 4 represent dynamic responses of the rail fastening
with loosened bolts. Figure 2 represents a frequency spectrum, and Figure 4 the
bicoherence graph. From graph on Figure 2, significant components in frequencies
15 Hz, 75 Hz, 100 Hz, 160 Hz, 210 Hz and 550 Hz may be read. So, if the graphs
of frequency spectra (Figures 1 and 2) are compared, it may be stated that the
spectra are almost identical and the differences in significant components are not
excessively important.
On the contrary, great differences are evident if the graphs of bicoherence in
Figures 3 and 4 are compared. In the graph in Figure 4, significant components in
frequencies and frequency pairs of 15 Hz, 75 Hz, 95 Hz,
120 Hz, 150 Hz, and 190 Hz are evident. Compared with the graph in Figure 2,
there are significant maximum values in the graph in Figure 4.
These results from the fact that the nonlinearity of the transmission of vibration
waves from the rail to the sleeper increases in the specimen with the rail
improperly fixed. Thus, this phenomenon is very well detected by the calculated
bicoherence.
Figure 1 Frequency response function, rail fastening with the standard tightened bolts
Figure 2 Frequency response function, rail fastening with loosened bolts

Figure 3 Bicoherence function, rail fastening with the standard tightened bolts
Figure 4 Bicoherence function, rail fastening with loosened bolts

4. CONCLUSION
In conclusion, it may be stated, based upon the measurements and analyses carried
out that the experiment carried out successfully verified the possibilities of utilizing
the given method for assessing the quality of particular components of the rail
structure. It follows from the text mentioned above that the method of higher order
spectra may play an important role in future when analyzing railway structures.
The results obtained show that this method proves to be prospective for the
analysis of the reaction of the rail fastening under different technological
(constructional) state to the dynamic load, then in detecting different types of
defects in fastening rails, sleepers and also the defects of rails proper.
Another important fact is that this method can be used to detect structural defects
and to assess the quality of the superstructure, which is not essentially made
possible when both the calculation models and the typical methods used until now
are applied. Laboratory tests should also be verified by measurements in the field.
Based upon the results obtained, it may be assumed that the method presented is
feasible in the region of the railway traffic. For example this method may be used
in continuous diagnostics of various components of the rail structure or in the
creation of a diagnostic system directly applied to commercial railway carriages
etc.
Acknowledgements

("The investigation of dynamic effects due to the rail transport by the method of
quadratic time and frequency invariant transformations ") and by research project
MSM 0021630519 (“Progressive reliable and durable load-bearing structural
constructions”)
References
1. G. Bessios, C. L. Nikias, FFT-Based Bispectrum Computation on Polar Rasters, IEEE
Transactions on Signal Processing, 39 (11), pp. 2535-2539, 11/1991
2. M. L. Williams: The use of the bispectrum and other order statistics in the analysis of one
dimensional signals, PhD Thesis, Imperial College of Science, Technology and Medicine,
University of London, 1992
3. J. W. A. Fackrell, S. McLaughlin: Quadratic phase coupling detection using Higher Order
Statistics, IEE Colloquium on Higher Order Statistics, Savoy Place London, 1995
4. Wang W. Y., Harrap M. J.: “Condition monitoring of ball bearings using envelope autocorrelation
technique,” Machine Vibration, Vol. 5, pp. 34-44, 1996
5. Rivola A., White, P. R.: “Bispectral Analysis of the Bilinear Oscillator with Application to the
Detection of Fatigue Cracks,” Journal of Sound and Vibration, Vol. 216(5), pp. 889-910, 1998
Subpixel Measurement in Close-Range Photogrammetry

Petr Kalvoda1, Jiri Vondrak1, Miroslava Sucha1, Radovan Machotka1
1
Department of Geodesy, University of Technology in Brno, Faculty of Civil Egineering, Brno
Veveří 95, 60200, Czech Republic
Summary
Digital image processing provides option to enhance the image, to automate

measurements of the points and locate targets with subpixel precision. The main
idea has been to create application to determine the center of the target. The whole
processing has been run by professional university mathematical pack MATLAB.
Well-known algorithms for subpixel location were applied and consequently, the
results were compared.
KEYWORDS: subpixel measurement, close-range photogrammetry.
1. INTRODUCTION
In photogrammetry, we often solve problem, how to reach precision, which would

be comparable with geodetics methods. Resultant precision is affected by image
scale, parameters of camera, configuration of images and accuracy of
determination control points and of course accuracy of determination of image
coordinates also [3].
2. DESCRIPTION OF PROJECT
This project is processed in frame of doctoral thesis called: Subpixel measurement

of images data. The goal is to develop application, which allows us to carry out
measurement of images coordinates with subpixels precision [3]. University
program package MATLAB [6] with usage of toolbox Image Processing appears
optimal. Matlab stores images as two-dimensional arrays (i.e.matrices). Each
element of the matrix corresponds to a single pixel in the display image.
GOAL OF THE PROJECT
APPLICATION DEVELOPMENT FOR

SUBPIXEL DETERMINATION OF THE
TESTING ON IDEAL EXAMPLE
TESTING WITH USAGE CALIBRATION

POINT FIELD
COMPARISON WITH RESULTS FROM

PHOTOMODELER AND ORPHEUS
3. DEVELOPMENT OF THE APPLICATION
At first was created graphical user interface (GUI), which allows carrying out self
measurement in MATLAB. There were necessary to solve these basic targets:
- to open and display image
- to indicate area, which is determined for searching ⇒ creation submatrix
of the original image. Corners coordinates of the indicated area are displayed in left
column see fig.1.
- to program single algorithms to locate center and use it on single
submatrix.
- output of coordinates to a text-file, which will be further possible to
process.
Figure1. Illustration of created GUI

578 P. Kalvoda, J. Vondrak, M. Sucha, R. Machotka
3. SUBPIXEL LOCATION OF TARGETS
There are known few methods to locate object center on the image [3]. Among the
most famous ones belongs certainly method of gravity center and method of
gradient. We can achieve precision up to 1/30 of pixel by using these methods,
minimal size of the target is 5 pixels [3]. Disadvantage of the method of the gravity
center is necessity of manual setting up threshold [3].
3.1. Method of gravity center

n
∑ (r T g ) i i i
rM = i =1
n
∑ (T g ) i i
i =1 (1)
n
∑ (c T g ) i i i
cM = i =1
n
∑ (T g ) i i
i =1 (2)
→ 0 pro gi ≤ t
Ti =
→ 1 pro gi ≥ t
t is selected threshold value, n is number of picture elements, gi is grey level and ri
and ci are image coordinates.
3.2. Method of gradient
This method includes into computation only pixels, in which a change of gray
levels happen.
n
∑r ∇ i
2
( gi )r
rM = i =1
n
(3)
∑∇
i =1
2
( gi )r
∑c ∇ i
2
( gi )c
cM = i =1
n
(4)
∑∇
i =1
2
( g i )c
∆g r ,i ∆g c ,i
∇ ( gi )r = ∇ ( gi )c =
∆ri ∆ci
In the first stage of processing was created M-file for computation center of the
measurement mark (using the gradient). It was decided that method of gradient will
be combined with edge detectors [8]:
- Roberts
- Sobel
- Canny
- Prewitt
- Zerocross
4. TESTING ON TARGET WITH DIAMETER 75 PIXELS
Nine circular targets, with diameter 75 pixels, were created in software

Microstation. Coordinates of the center are known exactly for the targets. Thanks
to the exact known coordinates, it was possible to determine mean errors of the
measurements. Center of these marks were gradually computed for individual gray
levels-g: 0, 100, 200, 250 in new application of MATLAB. For computation were
used gradient by itself and gradient combinated with individual detectors. We can
see dependence of mean errors of single determination on gray levels of the marks.
The goal of this experiment is to find impact of gray levels of the mark on
individual edge detectors.
The file *.dgn was converted to file *.tif, deformation of individual marks
happened, so the marks aren’t identical due to conversion. Mean errors 0.017 pixel
of the gradient by itself is caused by the deformation. Differences of the others
mean errors from this value are caused by individual edge detectors. It was decided
to carry out testing on the ideal example. Targets with diameter 5 pixels (minimal
size for the method) and with guaranteed symmetry, which have gradually different
gray levels, were created.
Table 1.
m0 [px]
g 0 100 200 250
gradient 0.017 0.017 0.017 0.017
grad + Roberts 0.017 0.017 0.017 0.017
grad + Sobel 0.146 0.167 0.111 0.123
grad + Canny 0.026 0.026 0.026 0.026
grad + Prewitt 0.067 0.067 0.142 0.163
grad + Zerocross 0.018 0.018 2.119 5.385
5. TEST FOR IDEAL EXAMPLE
5.1. Mark with diameter 5 pixels
In practice, used targets will not have the same color on a whole surface. Because
of irregular illumination affect, part of the mark on the image can have different
grey levels. It can really influence the results, so the “substitute „of the circular
mark with diameter 5 pixels with different gray levels were created. These marks
were used for test.fig.2.
1.type 2.type 3.type 4.type
Figure 2. Illustration of used marks
5.2. Computation of marks centers and their comparison
In application in MATLAB centers of the marks were determined gradually only

with use gradient by itself and than in combination with single edge detectors.
Computation of center of mark-1st type with use of gradient by itself is possible to
consider errorless. With these centre coordinates are compared all others.
There was evident systematic shift for all combinations, in all marks 1-9 in
coordinates between double determinations of the mark center except to Zerocross
detector. Thereby, this edge detector was foreclosed from further computation.
Determination of the center by Canny detector is same as the gradient alone.
Table 2.
1.type of mark-g:0
grad.-Roberts grad.-Sobel grad.-Canny grad.-Prewitt grad.-Zerocross
marks
1-9 δy (px)δx (px) δy (px)δx (px)δy (px) δx (px)δy (px)δx (px) δy (px) δx (px)
avarage 0,500 0,500 -0,100 -0,100 0,000 0,000 -0,100 -0,100 -0,180 0,253
Table 3.
2.type of mark-g:120-90-60-30-0
grad.1-grad.2 grad.1-Roberts grad.1-Sobel grad.1-Canny grad.1-Prewitt
marks δy δx δy δx δy δx δy δx δy δx
1-9 (px) (px) (px) (px) (px) (px) (px) (px) (px) (px)
avarage 0.000 -1.003 0.500 0.500 0.000 -0.125 0.000 -0.125 0.000 -0.125
From table 3 is evident that change of gray levels has effect not only on
computation by gradient itself but also on computation by gradient in combination
with some edge detectors. Similar to 1st type mark, Roberts edge detector shows
systematic shift 0.5 pixel on all marks 1-9. It is caused by mask size 2x2 [8].
Table 4.
3.type of mark-g:120-0
avarage 0.000 -0.933 0.500 0.643 0.000 -0.333 0.000 -0.375 0.000 -0.333
In 3rd type of mark, there is still systematic shift in computation by Roberts edge
detector on all marks 1-9. But in this case the shifts are not exactly 0.5 pixels.
Table 5.
4.type of mark-g:120-0
avarage 0.000 0.056 0.500 0.500 0.050 -0.125 0.000 0.000 0.000 -0.077
From these results we can see, that the most stable is Roberts and Canny edge
detector. Question is, if it applies also for arbitrary gray levels. Thereby, graphical
test was done for individual edge detectors.
5.3. Graphical test for edge detectors
The goal for detected edges is to display the reality as true as possible and also to
make the picture symmetric. Gray levels were changed gradually from 30 to 120 on
marks mentioned above.
5.3.1. 2nd type of mark: interface g: 120-90-60-30-0
Roberts detector: Sobel detector:
Canny detector: Prewitt detector:

In this case the best is Roberts detector. Experimentally was found, that for this
operator is necessary to have searching area 9 pixels from mark edges.
5.3.2. 3rd type of mark
Grey levels of mark were gradually changed. Changes of edge determinations by

means of individual edge detectors were studied. Figures are same always for
extend of gray levels, which are listed below.
Roberts detector
Interface g: 0/0 – 60/0 70/0 – 100/0 110/0 – 120/0

Sobel detector
Interface g: 1/0 – 60/0 90/0 120/0

Canny detector
Interface g: 0/0 – 110/0 120/0

Prewitt detector
Interface g: 1/0 – 60/0 90/0 120/0

Canny detector appears to be the most stable. Mark image is deformed only in
g=120. For this detector is small searching area sufficient, it means 2 pixels from
mark edges. Roberts detector deforms image from g=70 and also it is necessary to
indicate area 7 pixels from mark edges.
5.3.3. 4th type of mark
In this case the intensity of pixels was again changed gradually to 30, 60, 90, 100,
110 and 120. The best is Canny detector. It detects edges correctly up to g=130
with searching area 2 pixels. Roberts detector detects edges up to g=120 but with
searching area 11 pixels.
6. CONCLUSION
Image coordinates were measured by created applicatio. The goal of experiment

was to determine and test the most suitable method for computation of marks
centers for its further use in practice.
The most stable results gave Canny and Roberts detector in all tests. Effect of
Canny detector is given by three parameters: width of Gaussian mask and upper
and lower thresholds used by tracker [8]. Increasing width of Gaussian mask
reduces the sensitivity of detector to noise [9]. The disadvantage of Canny detector
is that there can be a problem with Y-junctions. The advantage of Roberts detector
is fast computation.
Real images will be use for testing. It is possible, that the other detector will be
more suitable and will provide more stable results.
References
1. Dušek, F., Matlab a Simulink, Univerzita Pardubice 2000, ISBN 80-7194-273-1
2. Sojka, E., Digitální zpracování a analýza obrazů, Vysoká škola báňská - TU Ostrava, 2000, ISBN
80-7078-746-5, S.133 (in Czech)
3. Luhmann, T.: Nahbereichs – photogrammetrie, 2000, ISB 3-87907-321-X (in Germany)
4. Kager, H. - Rottensteiner, F. - Kerschner, M. - Stadler, P., Orpheus 3.2.1,-User manual. Vienna
university of technology, 2002
5. Eos systems INC., PhotoModeler Pro – User Manual. Canada, Vencouver 2000,
http://www.photomodeler.com/
6. The Mathworks, Getting started with Matlab,

http://www.mathworks.com/access/helpdesk/help/pdf_doc/matlab/getstart.pdf
7. Kropatsch, W. - Bischof, H., Digital image analysis, Vienna, 2001, ISBN 0-387-95066- 4
8. The Mathworks, Image processing toolbox user’s guide, 1999
9. Bartoněk D., Vybrané kapitoly z počítačové grafiky, VUT v Brně FAST 2002, ISBN 80-214-
2083-9 (in Czech)
10. Hanzl V. - Sukup K., Fotogrammetrie I, VUT v Brně FAST 2002, ISBN 80-214-2049-9 (in
Czech)
The Generalised Mohr-Coulomb (GMC) Yield Criterion and

some implications on characterisation of pavement materials
Mircea Balcu1, Ştefan Marian Lazăr2
1
Strength of Materials Department, Technical University of Civil Engineering Bucharest, RO-020396,
Romania
2
Department of Roads and Railways, Technical University of Civil Engineering Bucharest, RO-
020396, Romania
Summary
In this study a generalization of Mohr-Coulomb Yield Criterion has been

developed, implemented, and tested in practice. The practical usefulness of the
proposed model is demonstrated with a case study
In the Generalised Mohr-Coulomb (GMC) Yield Criterion one introduces in the
Plasticity Mohr-Coulomb Theory the complete tridimensional stress state
( σ I , σ II , σ III ), thus generalising the well-known Mohr-Coulomb Theory
( σ I , σ III ). Indirectly, GMC also takes into account the influence of spherical
tensor σ~" , therefore σ~ = σ~ '+σ~" ( σ~ ' = the deviatoric tensor).
According to Soil Plasticity Theory, and the constitutive characterisation, the
material model can be described by the cohesion c and internal friction angle Φ ,
or, alternatively, by the uniaxial strengths Rc and Rt . In the GMC model, the
material is described by the generalised parameters c * and Φ * .
KEYWORDS: yield criterion; plasticity; cohesion and internal friction angle;

pavement materials
1. INTRODUCTION
The object of the mathematical theory of plasticity is to provide a theoretical

description of the relationship between stress and strain for a material which
exhibits an elasto-plastic response. In essence, plastic behaviour is characterised by
an irreversible straining which is not time dependent and which can only be
sustained once a certain level of stress has been reached.
In order to formulate a theory which models elasto-plastic material deformation
three requirements have to be meet:
586 M. Balcu, Ş.M. Lazăr
• An explicit relationship between stress and strain must be formulated to

describe material behaviour under elastic conditions, i.e. before the onset of
plastic deformation;
• A yield criterion indicating the stress level at which plastic flow commences
must be formulated;
• An incremental relationship between stress and strain must be developed for
post-yield behaviour, i.e. when the deformation is madeup of both elastic and
plastic components.
The Mohr-Coulomb yield criterion is a generalisation of the Coulomb (1773)
friction failure low defined by:
τ = c − σ n tgΦ (1)
where τ is the magnitude of the shearing stress, σ n is the normal stress (tensile
stress is positive), c is the cohesion and Φ is the angle of internal friction.
Graphically Eq. (1) represents a straight line tangent to the largest principal stress
circle as shown in Fig. 1 and was first demonstrated by Mohr (1882).
Figure 1. Mohr criterion
From Fig. 2, and for σ 1 ≥ σ 2 ≥ σ 3 Eq. (1) can be written as
(σ 1 − σ 3 ) + (σ 1 + σ 3 ) sin Φ = 2 ⋅ c ⋅ cos Φ (2)

Figure 2. Mohr-Coulomb Yield Criterion
A complete derivation of the Mohr-Coulomb (MC) yield criterion is given in an

forcoming paper [2].
2. NOVOZHILOV’S APPROACH OF MC YIELD CRITERION
The equivalent value of MC – yield criterion will now be expressed as follows:

σ ech
MC
= c ⋅ cos Φ (3)
or, in a Plasticity Theory format,

f (σ 1 , σ 2 , σ 3 , c, Φ ) = σ ech
MC
− c ⋅ cos Φ = 0 (4)
The use of spheric/deviatoric decomposition of the stress tensor, σ~ = σ~ '+σ~" ,

yields the following expressions for principal normal stresses used in Eq. (2) (see
[2] for details):
1
(
σ 1 = σ "+ σ − sin Θ + 3 cos Θ
3
)
1
(
σ 3 = σ "+ σ − sin Θ − 3 cos Θ
3
)
Therefore,
2 3
σ1 −σ 3 = σ cos Θ
3
2
σ 1 + σ 3 = 2σ "− σ sin Θ
3
and σ ech
MC
writes as:
 1 1 
σ ech
MC
= σ " sin Φ + σ  cos Θ − sin Θ sin Φ  = c ⋅ cos Φ (5)
 3 3 
Comment: The MC - yield criterion is not a pure “shear” theory, because it
contains both spheric part σ " , and the “deviatoric” part, σ .
3. THE GENERALISED MOHR-COULOMB (GMC) YIELD

CRITERION
In Fig. 3 the effective stress σ is given an interesting geometric interpretation, due

to V.M. Rosenberg [3].
∧
Figure 3. Rosenberg’s method: σ = VB ; Θ = MVB ; VM = σ cos Θ
A generalised Mohr-Coulomb yield criterion follows the line of classical MC, with
the vortex point V following the straight line equation:
X Y
+ =1 (6)
d * c*
where (see Fig. 4): c * and d * denote the generalised cohesion and generalised
abscissae, respectively.
Figure 4. The Generalised Mohr-Coulomb (GMC) Yield Criterion
The Eq. (6) writes, succesively, as follows:

σ1 + σ 3
2 σ cos Θ
+ =1
c * ctgΦ * c*
σ1 + σ 3
tgΦ * +σ cos Θ = c *
2
σ1 + σ 3
sin Φ * +σ cos Θ cos Φ* = c * cos Φ * (7)
2
Now, following the trigonometric procedure of Novozhilov, one writes

succesively:
2
σ 1 + σ 3 = 2σ "− σ sin Θ
3
and, therefore,
σ1 + σ 3  1 
sin Φ* =  σ "− σ sin Θ  sin Φ * (8)
2  3 
Eq. (7) finally writes as follows:
 1 
σ ech
GMC
= σ " sin Φ * +σ  cos Φ * cos Θ − sin Θ sin Φ * = c * ⋅ cos Φ * (9)
 3 
To compare, one observes that σ ech
GMC
(Eq. (9)) is a generalisation of σ ech
MC
(Eq.
(5)).
4. APPLICATION – CASE STUDY: ASPHALT MIXTURES
Laboratory triaxial tests enable one to find out the intrinsic material characteristics
of an asphalt mixture MASF 16 in the ambient temperature T = 23 ºC, lateral
pressure σ 3 = 2 to 4 daN/cm2, and a vertical loading in a regim of v = 0,46
mm/min.
The following MC – parameters were found:
Φ = 36,90 º
c = 2,02 daN/cm2
Following the Generalised Mohr-Coulomb theory (GMC) the corresponding values
are found as follows:
( ) (
Φ* = arctg 3 sin Φ = arctg 3 sin 36,90 = 46,12 º )
c* = 3 ⋅ c ⋅ cos Φ = 3 ⋅ 2,02 ⋅ cos 36,90 = 2,80 daN/cm2
5. CONCLUSIONS
For yield criteria with non-smoothly intersecting multiple yield surfaces e.g. Mohr-
Coulomb (MC) and Generalised Mohr-Coulomb (GMC), a good return scheme has
to be supplemented by proper care to account for the non-regular regions in the
yield surface. The problem of determining if multiple yield surfaces are active, has
recently received some attention. If these predictor-corrector algorithms has to
incorporate the Koiter’s generalisation [4], some singularities should be taken into
account [2].
References
1. Owen, D.R.J., Hinton, E., Finite elements in Plasticity: Theory and Practice, Pineridge Press.
Limited, Swansea, U.K., 1980.
2. Balcu, M., Lazăr, Ş.M., A Generalised Mohr-Coulomb (GMC) Plasticity Theory for solving non-
linear analysis of pavement materials and structures, Buletinul Ştiinţific al UTCB, nr. 4, 2007. (in
English and Romanian)
3. Smirnov-Aliaev, G.A., Rozenberg, V.M., Theory of plastic deformations in metals, Lenizdat,
1951. (in Russian)
4. Koiter, W.T., Stress-strain relations, uniqueness and variational theorems for elastic-plastic
materials with a singular yield surface, Q. Appl. Math., vol. 11, 350-354, 1953.
Quality costs in Bridge Engineering

Alina Nicuţă, Constantin Ionescu
Faculty of Civil Engineering and Installations, University Gh. Asachi, Iasi, 700050, Romania
Summary
The current paper desires to create an introduction towards the research in quality
costs for bridges.
The superior quality of bridges is a matter of great importance. The quality
administration is determined by different instruments which are components of ISO
9000 standards. As for Romania the quality in the area of engineering is guided by
the Law 10 of quality. The concept of quality costs has the main objective the
identification, evaluation and comparison of costs. These costs administration must
integrate all the phases of creation and exploitation of a bridge construction.
KEYWORDS: quality costs, bridges, Law 10 – quality law, design, execution and
exploitation.
1. INTRODUCTION
The bridges network is ageing and so the agencies take more and more into
consideration the maintenance and rehabilitation of the existent infrastructure.
The quality of bridge structures is influenced by different types of factors such as
technical, economical, social and natural. Most of the times, their influence isn’t an
isolated phenomena but a simultaneous one. They interfere on the quality
characteristics, creating the rise and fall of the quality general level.
Quality costs are the costs associated with preventing, finding, and correcting
defective work. These costs may be huge. Many of these costs can be significantly
reduced or completely avoided. One of the key functions of a Quality Engineer is
the reduction of the quality total cost associated with a construction work.
2. GENERAL APPROACH OF TRANSPORTATION QUALITY

COSTS
Considering the transport capacity already valuated, the agencies must focus more
than before on the work zones effects on the users.
The costs in a different approach can be connected to the transport infrastructure

elements quality.
The most common is the traditional approach of the costs regarding the quality. It
allows the determination of the source where they appear, the cause and quality
costs measurement.
A certain time period the quality improvement and the quality costs reduction were
considered differently. This problem was settled in 1979, in the paper „Quality is
Free” by Ph. Crosby, where the quality costs are analyzed in detail. In its book,
Crosby demonstrated the fact that the lack of quality determines complementary
expenses. Even if in practice is difficult to identify and divide all the quality costs,
an attempt in this direction brings into evidence looses due to the reparations,
replacements, stagnation etc.
The determination of quality costs takes to profit increase by the identification of
the most proper solutions of costs reduction.
Quality costs are grouped this way:
1. prevention
2. inspection
3. internal costs of unconformity
4. external costs of unconformity
Fig. 1: Reduction of quality costs
Source: Marieta Olaru, “Costurile referitoare la calitate”, in “Quality Management” Second

Edition
Quality costs have different classifications as seen in Figure 1.

594 A. Nicuţă, C. Ionescu
The prevention costs in the area of bridges represent the investments done before
the execution works or maintenance in order to prevent the defection in the next
steps. Here are being included also the quality costs determined by the
infrastructure impact on the environment. The lately transportation fast
development had a great impact on the environment so that the transportation has
become the main source of aggression on the environment and human health. The
environment pollution due to a bridge structure takes place in two steps: the
execution and exploitation.
Evaluation costs are inspection costs (testing, inspection, examinations) and tests
which have been considered if the quality requests have been fulfilled. In the
bridge area the evaluation costs represent the value of the effort done in order to
determine the conformity degree of the construction works with the specified
quality.
Internal unconformity costs in bridges execution and maintenance represents the
costs of unconformity adjustments discovered after the works creation but before
the element exploitation. These are expenditures caused by the fact that certain
works quality level doesn’t fulfill the specified requests.
External unconformity costs in the bridge area are the expenses determined by
the unconformities correction discovered after the execution and maintenance
works and after bringing into service the elements. These are costs caused by the
fact that the quality level for certain constructions doesn’t fulfill the specified
requests.
Fig 2 The connection between the quality and quality costs

Source: Marieta Olaru, “Costurile referitoare la calitate”, in “Quality Management” Second
Edition
The connections between the quality costs categories is suggested in fig 2 by the
graphical representation of the quality “optimal level“. It can be seen that in the
situation of a relatively small increase of the investments for the prevention and
evaluation measures will result a sensitive reduction of the unconformity costs, so
that, totally, the quality costs will decrease. This fact is also presented in the
previous figure but with a different structure. On the other hand, in the case of
neglecting the preventive quality costs component, can result important increases
of the other expenses, considering the internal and external unconformity
multiplication.
The decrease of any of the three expenses categories will lead towards the decrease
of quality total cost, but the minimal level is in the equilibrium point between these
three costs categories. As seen in figure, there is a point on the total expenses curve
where these are minimal and this is a result of the combination between the
unconformity costs and those of prevention and evaluation.
Being outside the optimal level means obtaining a low quality level, when the
quality loses are greater than the investments, or obtaining a high level of quality,
when unconformity costs reduction is done with great expenses for the prevention
and evaluation and also by more important investments.
The quality costs are interconnected:
QUALITY COSTS
VERIFICATION COSTS
Unconformity Unconformity
prevention costs Verification costs
Special verification Internal External
Deflection costs
Accordance costs (Unconformities)
(Conformity)
Normal expenses for Resource waste
performance creation
Fig 3 Quality costs

Identification of the quality costs is necessary for operative and administrative

procedures supervision. They are used as a strategic instrument to create projects
and for measures argumentation.
Quality costs gathering and calculation is included into a quality
management system. It is recommended a preventive treatment of the quality
costs.
The basic point in quality costs treatment is represented by the unconformity
prevention and costs, in the first steps of the construction and not the
unconformities improvement.
A different vision on the quality costs starts from the Law 10 regarding the Quality
in Constructions. For a bridge construction work are being taken into consideration
as research points not only the quality regarding the bridge construction but also
complementary elements such as:
design
execution
exploitation
with the proper expenses for the quality.
If previously in general view, the accent is on the quality costs in the context of
construction work exploitation, in the current view are being considered also the
quality costs in the phase of design and execution.
3. QUALITY SYSTEM IN BRIDGE ENGINEERING
In accordance with Law no. 10 regarding the quality in construction, the quality
system is composed from:
a. Technical regulation;
b. Quality of the products used for the construction creation;
c. Technical agreements for new products and technologies;
d. Projects verification, works execution and projects and constructions
expertise;
e. Assurance of quality in construction area;
f. Authorization and accreditation of analysis laboratories and testing in
construction activity;
g. Metrology activity in constructions;
h. Construction reception;
i. Behavior in exploitation and time intervention;
j. Construction post utilization;
k. State quality control in construction.
In the area of Bridge Engineering, quality assurance, control and maintenance for a
bridge has some particularities, specially, considering the exploitation period, the
fact that this is designed, executed and exploited by different companies, by the
long time contact with the environment and by its social importance.
This way, for every company must by developed and implemented a quality system
with the purpose of realizing the quality policy objectives. In order to achieve the
quality objectives are realized the mobilization of all the technical, administrative
and human factors which can influence a construction quality. The quality system
must be guided towards the reduction, elimination and most important prevention
of qualitative deficiencies.
In order to implement the quality policies and deficiencies, in a company must be
created proceedings regarding the personnel training, design activities
coordination, bridge creation and maintenance.
Also, it is necessary to select all the activities which lead directly towards the
assurance, administration and maintenance of the quality, reported to the design,
creation and maintenance activities. Based on the specified data there must be
determined as right as possible a bridge quality costs on the existence period.
4. AGENTS WHICH DETERMINE THE QUALITY IN BRIDGE

ENGINEERING
Considering a critical analysis of the quality system components and the existence
steps for a bridge such as: design, execution and exploitation can be determined the
factors by which it can be assured the quality.
4.1. In the design phase, the factors which assure the bridge quality are:
a. The professional level of each member of the design collective
(including the collective operational discipline);
b. The experience in the area of the design collective (the general
experience and the specific experience for the design theme);
c. The informational system complexity (standardization,
instructions, normative, treaty, similar projects etc.);
d. The used computer field system performance (data base,
INTERNET, INTRANET, HARD, Soft);
e. The durability of the adopted constructive system (bridge on
frames, girder bridge, etc.);
f. The site identification (information, technical and
environmental knowledge, etc.);
g. The capacity to satisfy the basic requests: resistance, stability,
safety in exploitation, environment protection;
h. Optimization of the adopted constructive solution;

i. Project inspection, project expertise in the case of frictions
between designer and verifier, approval of the internal
technical department of the beneficiary.
4.2. Factors specific to the bridge construction phase:

a. Level of professional background of each and every member
of the engineers collective (including the collective operational
discipline);
b. Experience in the area of bridge execution and experience
specific to the designed bridge type;
c. The complexity of used information system (standardization,
instructions, technological methodologies etc.);
d. Performances of the analysis and testing laboratories;
e. Information related to the execution project acquired by the
executor (relation between designer – executor);
f. Quality of the products used for the construction;
g. Bridge testing (if mentioned in the project);
h. Bridge reception.
4.3. In the exploitation phase the bridge quality maintenance factors are:
a. The level of professional background of each member of the
experts collective in bridge maintenance, at beneficiary
(including the collective operational discipline);
b. Experience in bridge exploitation area and experience specific
to the received bridge type;
c. The bridge exploitation in accordance with the project
specifications (weight, load class etc.);
d. Pursuit in exploitation of the received bridge behavior;
e. Processing and stocking of data and information from the
determination in time of the bridge behavior;
f. Intervention in due time on the bridge: maintenance, repairing,
consolidation etc.;
g. Construction expertise.
5. QUALITY COSTS IN BRIDGE ENGINEERING
Considering the factors specific for the creation of the bridge quality, the quality
cost can be detailed in three main costs groups:
a. The cost for quality assurance and damages prevention

(constructive and technological, personnel training
improvement, costs of the used equipments etc.);
b. Quality cost control (evaluation, remuneration, endowment
etc.);
c. Costs of losses due to the non quality (remunerations, replaced
elements, costs of intermissions in assembly delivering, looses
in exploitation phase etc.).
Total cost of a bridge quality, C TC is:
C TC = C PA + C C + C I + C P
where:
C PA - Prevention cost of the damages and assurance of the quality level;
C C - Total quality control cost;
C I - Maintenance costs at the beneficiary;
C P - Looses due to the lack of quality.
6. CONCLUSIONS
In order to determine the quality costs it is necessary to study all the factors which
influence the quality in a bridge life time, to detail these factors and to distinguish
the quality costs considering the other activities costs which fulfill for the bridge
edification.
At the bottom of quality costs evaluation we must consider the quality system
correlated with the systems engineering concept by which the total creation of a
bridge is determined by a system consisting from the design subsystem, creation
(construction), exploitation and evaluation of the bridge.
References
1. Kelada, J., La gestion intégrale de la qualité. Pour une qualité totale. Edition Quafec, Québec,
1990
2. "Quality Costs" in Juran, J.M. & Gryna, F. M. (1988, 4th Ed.), Juran’s Quality Control Handbook,
McGraw-Hill
3. Marieta Olaru, Costurile referitoare la calitate, Managementul calitatii, editia a II-a
4. Voicu M., Irina Severin, Iniţiere în ingineria calităţii, Editura Bren, Bucureşti 2000
5. Principles of Quality Costs, ASQC Quality Press, Appendix B, "Detailed Description of Quality
Cost Elements."
Investigation of mechanical properties of castable with respect to

their application as ceramic anchor elements
Lenka Nevrivova
University of technology, Faculty of Civil Engineering,
Institute of Technology of Building Materials and Components, Brno, 602 00, Czech Republic
Summary
Burnt refractory building materials are commonly used as ceramic anchor

elements, particularly ordinary and high-aluminous fire clay. One of the most
important material properties affecting their reliability are sufficient bending
strength and shear strength both in normal and increased temperatures.
The study compares properties of high quality compact fire clay with fire clay
castable of ULCC, LCC, MCC, and RCC type. Influence of burning temperature
has been watched for high-temperature concrete on resulting physical mechanical
properties of the material. The use of computer technology is important
particularly for reviewing of measured data and optimization of material mixture.
KEYWORDS: castable, ceramic anchor elements, medium cement castable, low

cement castable
1. INTRODUCTION
Putty, mortar, daub, castable and also castable shaped bricks concreted and pre-
burnt in production plants belong among non-shaped refractories.
Anchor elements anchoring the walling made of refractory monolith together with
curtain wall of a kiln belong among castable elements used during construction of
kiln structures, as well. Anchor material is being selected particularly according to
its usability temperature. It is possible to use anchors made of high-temperature
steel, high-quality fire clay and/or castable anchor elements. Anchoring system and
achor material have major influence on service life of kiln walling.
The submitted paper deals with material suitable for production of castable anchor
elements being used in practice more and more often. This is why their utility
properties are subjected to interest not only of producers and consumers of such
materials. The paper deals with material utility properties of castable anchors under
both ordinary and increased temperatures and its goal is examination of
applicability of various kinds of castables for these special purposes.
It is necessary to be stressed that the mentioned work would be very problematic

without use of computational technology.
2. CASTABLE ANCHOR ELEMENTS
Castable monolithic walling is anchored to a curtain wall of a kiln by a system of

anchors which enables exchanging individual segments without affecting kiln
stability. There exist a whole range of anchor elements serving for this reason such
as anchor fire clay stones, steel anchors, prefabricated cones and anchor blocks.
Anchor system is decisive for service life and efficiency of monolithic walling.
Steel anchor could be used up to 1.200 °C, ceramic or castable anchors for
temperatures over 1.200 °C. Distances of anchor elements are specified.
Figure 1. Steel anchor elements
3. PREPARATION OF TESTING SAMPLES
The submitted work observes various kinds of castables varying mainly by amount
of alumina cement and plasticizer. Alumina cement is being gradually substituted
by appropriate finely grounded component in form of micro-filler.
The following four types of castables have been selected for our study as well as
high-quality compact fire clay – traditional anchor material – for comparison. The
examined castables:
• regular cement castable (RCC), testing bodies marked R
• medium cement castable (MCC), testing bodies marked M
• low cement castable (LCC), testing bodies marked L
• ultra low cement castable (ULCC), testing bodies marked U
602 L. Nevrivova
Figure 2. Ceramic anchor elements
The preparation of testing bodies has been carried out in a traditional way. First,
individual components of castable mixtures were weighted accurately, each
mixture was poured into a mechanical mixer and after partial dry mixing, batch
water was added. After mixing, the mixture was poured into forms and vibrated.
After 24 hours, the bodies were formed and marked with high temperature colour
for further identification.
Figure 3. Vibrating into forms and set of testing bodies
Before burning process started, the testing bodies were measured and weighted in a
dry state. Further, they were heated in a chamber kiln under temperatures 300, 800,
1.300, and 1.450 °C in form of cubes with 70 mm edges. Always three pieces of
each mixture under each temperature.
The cones were burnt under temperature 600 °C as well as the plate for cutting the
samples for strength in bending test under increased temperatures and cubes with
100 mm edge for testing of carrying capacity in heat were burnt under temperature
1.450 °C. Temperature raise in a kiln was linear 5.5 °C per minute, holding time
under maximum temperature 5 hours.
Cooling was autogenous. After being burnt, the testing bodies were measured again
for assessment of permanent length changes in heat.
4. PROCEEDED LABORATORY TESTS
The range of physical-mechanical tests and examinations were carried out in order
to enable assessing convenience of various types of castables as materials for
ceramic anchor elements. Permanent length change of testing bodies burnt under
temperatures 300, 800, 1.300, and 1.450 °C was observed. Volume mass was
determined and apparent porosity by means of hydrostatic weighting method on
fragments of the testing bodies, too. Compressive strength in cool state was
determined on the testing bodies dried under 110 °C and burnt under 300, 800,
1.300, and 1.450 °C. Tensile strength in bending in cool state and under increased
temperature was observed for the testing bodies pre-burnt under 600 °C.
Temperature of fracture was 20, 300, 800, and 1.300 °C. The testing bodies pre-
burnt under 600 °C were determined on bearing capacity in heat and resistance
against sudden temperature changes (Fig. 4).
Figure 4. Determination of resistance against sudden temperature changes
Figure 5. ULCC material texture after its Figure 6. ULCC material texture after
drying (enlarged 40x) burning under 1.450°C (enlarged 40x)
Microscopic observation of individual types of castables was also made.

Computational technology enabling image transfer from microscope to PC monitor
was used for observation of microstructure of the testing bodies in variable
enlargements which enabled further image processing. Using the software LUCIA,
size and occurrence of individual pores were measured, pores with minimum and
604 L. Nevrivova
maximum radius etc. Also influence of burning process on microstructure of the

material was observed.
ULCC material microstructure (Fig. 5) is relatively softened after drying, aggregate
grains are bound by a hydraulic bond. After castable burning under temperature
1.450 °C (Fig. 6), it is possible to observe significant material sintering and glass
phase content. Aggregate grains are bound by a ceramic bond.
5. RESULT DISCUSSION
The testing bodies from four castable mixtures varying in cement amount were
examined. Mixture R belongs to a category of common castables contained 20 %
of alumina cement. Mixture M belongs to a category of castables with medium
amount of cement, it contained 12 % of alumina cement. Mixture L belongs to a
category of castables with low cement amount, it contained 5 % of alumina
cement. Mixture U belongs to a category castables with ultra-low cement contents;
it contained 2 % of alumina cement.
The testing bodies produced from these castables were examined and compared
together with burnt refractory material, high-quality, compact fire clay of which
most of anchor elements are ordinarily made.
2,60
2,50
Volume mass [g.cm -3]
2,40
2,30
2,20
2,10
2,00 RC
MCC
1,90 LCC
ULCC
1,80
0 200 400 600 800 1000 1200 1400 1600
Burning temperature [°C]
Figure 7. Effect of burning temperature on material compactness
The Fig.7 shows that raising cement content in the mixure decreases volume mass
and compactness of castable. Drop in volume mass was observed particularly for
samples pre-burnt under temperature 800 °C which corresponds to temperature of

hydraulic bond transformation to a ceramic bond. The result of sintering process
over temperature 800 °C is raising volume mass with burning temperature (Fig. 7).
Difference between the mixture with cement content 5 % and 2 % (LCC and
ULCC) is minimum.
0,00
Permanent length changes [%]
-0,20
-0,40
-0,60
-0,80
RC
MCC
-1,00 LCC
ULCC
-1,20
0 200 400 600 800 1000 1200 1400 1600
Figure 8. Effect of burning temperature on permanent length changes
Material compactness corresponds with its permanent length changes apparently

from the following graph (Fig. 8). Increase of length changes related again with
initiation of material sintering is apparent under temperatures higher than 800 °C.
Influence of cement content on permanent length change is not significant enough
and all values range in tolerance ± 0,15 % up to temperature 1.000 °C, after that,
cement content shows more significantly.
Cement content has, however, significant effect on material strength in
compression (Fig. 9). Compression strength in cool state rises with burning
temperature and under temperatures over 1.000 °C is the strength higher for the
mixtures with ultra low and low cement content than for the rest two mixtures.
Strength of ULCC mixture burnt under 1.450 °C temperature is more than three
times higher than for ordinary castable.
Compression strength in bending under increased temperature could be evaluated
similarly as compression strength in cool state; again, the highest strengths were
reached for LCC and ULCC mixtures. These mixtures with lowered cement
content proved good both for fracture temperatures over 1.000 °C and lower
temperatures. Ordinary castable and high-quality fire clay which has also been
subjected to this test, reached strengths ten to fifteen MPa lower. Under fracture
606 L. Nevrivova
temperature 1.300 °C, the differences between materials are not so significant and
it can be assumed that they will become even less significant with rising
temperature.
RC
200
Compression strength [MPa]
MCC
LCC
ULCC
150
100
50
0
0 200 400 600 800 1000 1200 1400 1600
Figure 9. Effect of burning temperature and cement on compression strength
All castable mixtures stand the testing of resistance against sudden changes of
temperatures (Fig. 4) and their failure was minimum, nearly insignificant in
comparison with fire clay material; it manifested only with tiny cracks after 30th
testing cycle. Fire clay material fell into two parts already after 20th testing cycle.
The best evaluated with respect to this test was material ULCC.
5. CONCLUSION
After finishing all tests and after comparing all measured data by means of
computational technology, material with ultra-low cement content (2 %) ULCC
was recommended for production of ceramic anchor elements. This material stand
good in all proceeded tests and therefore it is possible to be recommended as very
convenient material for production of ceramic anchor elements.
This type of refractory shows low permanent length changes and reaches high
volume masses due to its cement amount. Due to excellent sintering of aggregates
and micro-filler with matrix during burning process, hard ceramic bond originates
which provides high both compression strengths in cool state and tensile strengths
in bending also under increased temperatures. Castable of a type ULCC is also
possible to be used in applications over 1.500 °C.
Acknowledgements
This outcome has been achieved with the financial support of the research plan
MSM 0021630511.
References
1. Kriegesmann, J. Technical ceramic production materials. Köln: Deutscher Wirtschaftsdienst,
Ellerau 2005, ISBN 3-938595-00-0
2. Wachtman J. B. Mechanical properties of ceramics. Wiley, N.Y.USA 1996, ISBN-13: 978-
0471735816
Sensitivity study of a model for the stability analysis of

continuous welded rail
Valentin-Vasile Ungureanu, Marius Comanici
Department of CivilEengineering, “TRANSILVANIA” UNIVERSITY, Braşov, Romania
Summary
In 1992÷1999 period the International Union of Railways (UIC) commissioned a

research program from European Rail Research Institute (ERRI) about improving
the knowledge of continuous welded rail (CWR) track, including switches [2]. This
research was necessary for revision and update of Leaflet UIC 720 which regulate
the problems concerning the laying and maintenance of CWR track, which was
from January 1986 [3]. In the new Leaflet UIC 720 [4], wich was from March
2005, was introduced concepts and criteria for the CWR buckling safety
assessment and it were shown cases studies which appeal to the two analysis of
CWR stability software, one developed at TU Delft (Holland) for ERRI – software
called initialy CWERRI, and nowdays LONGSTAB – and the other developed at
Foster&Miller company for Federal Rail Administration of United States of
America (FRA) – sftware called CWR-BUCKLE [1, 2]. In this context, at Civil
Engineering Faculty from Brasov was developed a software for simulation of lost
of track stability using a non-linear discret model for CWR buckling analysis, in
presence of thermal and vehicle loads, model called SCFJ [5, 8]. A presentation of
SCFJ model can be found in [5].
In this paper is presented a comparison of numerical experiment results which
were achieved with SCFJ and CWERRI softwares.
KEYWORDS: Continuous welded rail, Non-linear stability analysis, Temperature

loading, Sensitivity analysis.
1. INPUTS
For this analysis which used the same inputs like in comparative studies of
CWERRI and CWR-BUCKLE softwares presented in [6, 7, 9].
Hereby, it was considered a sector of L=47,5 m in the central zone of a CWR track
located in a curve with R=400 m, which has in the middle a misalignment
characterized by a half sine wave with a length of λ = 9,144 m and an amplitude of
δ = 0,0381 m (fig. 1). These values are characteristic for USA railway [6, 7, 9]. The
track is composed from AREA 136 type rails and reinforced concrete sleepers
laying at 0,61 m spacing. For the fastenings it was considered a linear-elastic
behaviour with a torsional resistance of the fastenings Rt = 11250 N/rad per meter
track. The vertical behaviour of railway track is supposed to be linear-elastic with a
stiffnes of Rz = 68900 kN/m per meter track. The longitudinal resistance is
supposed to be linear-elastic. The lateral resistance is tri-linear with o pick value Fp
wich coresponding at Wp displacement and a limit value Fl for Wl displacement. It
was supposed that the values of lateral resistance are in function of the vertical
force between the sleeper and ballast (fig. 2) generated by the vertical loads on
train axle. The model is vertically loaded by a hooper wagon with two bogies,
represented by four vertical axle forces of Fz= 293 kN each (fig. 3). The centre
spacing between bogies is 12,85 m. The space between the axle in a bogie is 1,78
m. The vehicle is placed on track in such a way that the middle of center spacing
between bogies is coicident with the centre of the misalignment. Ori: The vehicle is
centered on the misalighment. The value of the friction coefficient between
sleepers and ballast is tan Ф = 0,86 – this being an average value for reinforced
concrete sleepers.
Fig. 1. Plan view of CWR track model
Fig. 2. The lateral behaviour of ballast incuding the corrected value due to vertical loads
610 V.V. Ungureanu, M. Comanici
Fig. 3. Vertical axle loads on track
2. THE COMPARISON OF THE RESULTS
In the sensitivity study the cantities of interests are the increases of the superior
critical temperature Tmax, respectively inferior Tmin, which result from the lateral
displacement curve for the middle of the CWR model in function of the increase of
the temperature (fig. 4). These were obtained through variation of each parameter
in an interval, while the other parameters are remaining constants, as shown in
Table 1. It was considered that in the initial position the track is strain free.
Fig. 4. Lateral displacement at the middle of CWR model versus temperature increasing
Table 1. The parameters of the model for the sensitivity analysis

Parameter Reference Range
value
Radius (R) [m] 400 100 ÷ ∞
Lateral peak resistance (Fp) [N/m track] 17508 8754 ÷ 26262
Lateral limit resistance (Fl) [N/m track] 9630 4815 ÷ 14445
Longitudinal resistance (Rx) [N/m/m track] 1,378·106 1,0·105 ÷ 1,0·107
Torsional resistance (Rt) [Nm/rad/m track] 1,1125·105 0,0 ÷ 3,0·106

Misalignment amplitude (δ) [m] 0,0381 0,008 ÷ 0,05
Wave length of misalignment (λ) [m] 9,144 1,2 ÷ 9,6
It is observed that for the critical temperature increases in function of radius

(fig. 5), the difference between SCFJ and CWERRI results is maximum 7%, and
the biggest differences are for the smallest radius values.
Fig. 5. The critical temperature increasing versus radius
The differences between SCFJ and CWERRI for the critical temperature increasing
in function of the lateral peak resistance (fig. 6) are maximum 2,6% when the track
is loaded only with thermal loads, therefore in absence of vehicle vertical loads.
Fig. 6. The critical temperature increasing versus lateral peak resistance (without vertical
loads)
Fig. 7. The critical temperature increasing versus longitudinal resistance

Fig. 8. The critical temperature increasing versus torsional resistance of the fastenings
Fig. 9. The critical temperature increasing versus the amplitude of the misalignment
For critical temperature increasing in function of lateral resistance (fig. 7) resulted

3% maximum differences between SCFJ and CWERRI results.
There were analysed the differences between SCFJ and CWERRI results for
critical temperature increasing in function of torsional resistance of the fastenings
(fig. 8). It results a maximum difference of 1,7%.
Fig. 10. The critical temperature increasing versus the amplitude of the misalignment
With regards to the amplitude of the misalignment, the differences between SCFJ
and CWERRI results for critical temperature increasing (fig. 9), it results a
maximum difference of 2,7%.
Finally, it was observed that for the critical temperature increasing in function of
the length of initial misalignment (fig. 10), the differences between SCFJ and
CWERRI results were maximum 1,9%, if the amplitude is 1/240 of the length of
initial misalignment.
In conclusion, it was observed a good correspondence between the results of the
two models, therefore the SCFJ model can be used for implementation of Leaflet
UIC 720R regulation in Romania or in stability of CWR track analysis with a
precision of the results closed to those developed by others.
References
1. *** - ERRI D202-RP3 Improved knowledge of forces in CWR track (including switches) -
Theory of CWR track stability, European Rail Research Institute, Utrecht, February 1995.
2. *** - ERRI D202-RP12 Improved knowledge of forces in CWR track (including switches) –
Final report, European Rail Research Institute, Utrecht, February 1999.
3. *** - UIC 720 R - Laying and maintenance of CWR Track, 1st edition, January 1986.
4. *** - UIC 720 R - Laying and maintenance of CWR Track, 2nd edition, March 2005.
5. Dósa, A., Ungureanu, V.V. - SCFJ – model discret de pierdere a stabilităţii căii fără joante,
Infrastructuri eficiente pentru transporturi terestre – Zilele Academice Timişene, 24-25 mai 2007,
Timişoara.
6. Esveld, C. - A better understanding of continuous welded rail track, Rail Engineering
International, No. 4, 1996.
7. Esveld, C. – Improved knowledge of CWR track, Delft University Press, 1997.
8. Ungureanu, V.V. - Modele de pierdere a stabilităţii căii fără joante, Referatul nr.3 din cadrul
stagilului de pregătire pentru doctorat, Universitatea “TRANSILVANIA” din Braşov, Facultatea
de Construcţii, Braşov, 2006.
9. Van, M.A.- Stability of Continuous Welded Rail Track, PhD Thesis, Delft University Press, 1997,
ISBN:90-407-1485-1.
Parametrical study of the effect of the torsional resistance of the

fastenings on the stability of continuous welded rail
Valentin-Vasile Ungureanu, Adam Dósa
Department of civil engineering, “TRANSILVANIA” UNIVERSITY, Braşov, Romania
Summary
This work contains a parametrical study on the stability of continuous welded rail
(CWR), taking in account the torsional resistance effect which appears between
rail and sleepers through fastenings system. The effect of the torsional resistance
was analyzed in the presence and in the absence of vehicle loading. It was
considered both linear and nonlinear behaviour of the fastenings. The analysis was
done using a simplified bi-dimesional model of CWR, with beam elements and
nonlinear lateral, longitudinal and torsional spring elements.
It was shown the influence of the different parameters which governs the behaviour
of the fastenings on the stability of CWR.
KEYWORDS: Continuous welded rail, Track stability, Fastening system.
1. GENERALITIES FOR THE TORSIONAL RESISTANCE OF

FASTENINGS
The continuous welded rail track is loaded by vehicles and weather factors,
especially by temperature variations. These loads generate vertical, transversal and
longitudinal displacements and deformations of railway track panels. Each of these
displacements and deformations are restricted by one or more resistances. All of
these resistances are of friction forces kind. The rotation of rail in the fastenings is
restricted, mainly by the torsional stiffness of fastenings. In the literature the
resistent moment of fastenings is marked by Mo or Mr.
This moment is experimentally determined in laboratory by graduate loading of a
rail clamped in fastening with a force F placed at a a distance from the fastening
axis, wich generate a θ rotation of the rail [1, 2, 8, 10, 11]. The resistant moment
and the rotation are determined by following relations:
Mo= F·a (1)
θ = ∆/b (2),
where b is the distance between fastening axis and the gauge which measure ∆
displacement of the rail (fig. 1).
Fig. 1. Experimental determination of the torsional resistance of the fastenings Mo [1, 2, 8].
The torsional resistance (Mo) – rotation (θ) diagram, called characteristic curve of
fastener, can have different shapes (fig. 2), depending on the relative position of the
rail in fastening.
So, if the rail is oblique positioned (1st case by fig.2), between rail and the lateral
shoulder of fastening do not exist clearances, the rail is tangent to lateraly shoulder
of fastening – for example like a K type indirect fastening – the fastening will be
restrict the rotation until the fastening elements will be destroyed, and the
characteristic curve will be linear. When the rail is in the normal position in
fastening (2nd case by fig.2), between rail and the lateral shoulder of fastening exist
equal clearances to both sides of rail, for small rotations will be mobilized the
rotational friction resistances between rail foot and seating plate, as well as
between rail foot and K clamping claw fastening, and the concordant branch will
be approximate linear. If all of these resistances will be integral mobilised it will be
remarked increases of rail rotation for small increases of rotational moment, until
the clearances between rail and lateraly shoulder of fastening will be passed off,
and the characteristic curve will be approximate constant on this branch. After the
rail will be tangent to the lateral shoulder of the fastening, the fastening will be
restrict the rotation until the fastening elements will be destroyed, and the
characteristic curve will be linear, again. The size of the branch whichever the
characteristic curve is constant is a function of the size of clearance between the
rail and the lateral shoulder of fastening. Hereby, if the rotation of rail in fastening
will be in that direction which lead the rail from oblique to normal position, it will
618 V.V. Ungureanu, A. Dósa
obtain the maximum size of the level of the characteristic curve , for the situation
in which the rail is tangent to the lateral shoulder of fastening and it exist
maximum values of clearances between the rail foot and the lateraly shoulder of
fastening (3rd case by fig. 2).
Fig. 2. The torsional resistance (Mo) – rotation (θ) diagram for diverse relative positions of
the rail in the fastening [4, 8, 10].
The value of resistent moment Mo, for which the characteristic curve of fastening is
constant depends on the vertical force with wich the fastening push on the rail foot
by fastening elements (for example, by the clamping claw, in case of K type
fastening), it being directly proportional with the size of this vertical clamping
force. In the fig. 3 is shown one example for this situation [8, 11].
The regulations and some authors consider a linear characteristic curve of fastening
which result by linearisation of the first branch of characteristic curve [1, 2, 5, 6, 7,
8, 11]. The explanation for this supposition is that, for the allowed values of
misalignment, the rotation of rail in the fastening does not exceed the characteristic
values of the first branch of the experimental characteristic curve.
However, for the common fastenings used in Romania, a tri-linear characteristic

curve is most proper for the real behaviour of the fastening.
Fig. 3. Characteristic curves of K 49 type fastening [10, 11].
Fig. 4. Characteristic curve of the fastening in function of vertical load due to vehicle [12].
2. THE MODELING OF TORSIONAL RESISTANCE OF

FASTENINGS
The majority of models consider a linar variation of characteristic resistent moment

– rotational angle of rail in the fastening (fig. 5 a), but exist models in which it is
considered a bi-linear, tri-linear (fig. 5 b) or multi-linear characteristic curve [2].
All linear [2, 6, 7, 13], bi-linear [3] or tri-linear [4, 8, 9, 10, 11] common models
are not taking in account the influence of vehicle vertical load on the rotation of the
rail in the fastenings, but in the model given in [5, 12] it is included. The behaviour
of fastening to rail rotation can be appropriately corrected according to this model
[5, 12], taking account of the pushing force on sleeper, as it is shown in fig. 4. Such
as approach is more close to the real behaviour of the fastening and it allows the
estimation of the influence of other parameters of fastening, which are not
considered in the other models.
Fig. 5. Torsional resistance of fastenings [5].
Hereby, the resistance to rotation of rail in fastening will be Mro for unloaded
railway track into the tri-linear model, and for the loaded railway track it will be
with values between Mrmax, concordantly with the most vertical loaded sleeper,
and Mrmin, concordantly with the less vertical loaded sleeper or unloaded sleeper –
which is on the sector of railway track between the most distant axle where coming
on the upplift of track.
The SCFJ model was presented in [5] and it was used in numerical experiments
depicted in the following. It allowes the use of a linear or tri-linear characteristic
curves (fig. 5).
3. THE COMPARISON OF THE RESULTS WITH THOSE

PROVIDED BY THE OTHER SIMILAR MODELS
The values of critical temperature are very important in the analysis of the stability
of continuous welded rail (CWR), i.e. the maximum increase of temperature Tmax -
over which the buckling of CWR will be surely coming on - and the minimum
increase of temperature Tmin – under which the buckling of CWR will surely not be
coming on [2, 3, 5, 6, 7, 12, 13].
The numerical experiment shown in [6, 7, 13] for testing of the SCFJ model – in
wich the torsional resistance of fastenings is linear and it is independent of the
vertical vehicle load – was repeated. It resulted maximum 1,5 % difference in
regard to the other models for the torsional resistance of fastening like in [7].
Fig. 6. The increases of critical temperature versus the torsional resistance of the fastenings
[6, 7, 13].
A graphical representation of the results is shown in fig. 6, where the maximum

and minimum critical temperatures by [7] were marked by Tmin A and Tmax A, and the
maximum and minimum critical temperatures by SCFJ were marked by Tmin B and
Tmax B, respectively. It results a very good correspondence between both results.
Torsional resistance of fastenings increase, minimum critical temperature Tmin
increase more than maximum critical temperature T max.
The explanation of this behaviour is that the rotations and transversal dispacements
of rails are smaller for minimum critical temperature T min, while maximum critical
temperature Tmax have more important values, therefore the mobilization of
torsional resistance of fastenings is bigger for minimum critical temperature Tmin.
It is imporatant to point out that the common values of torsional resistance
of fastenings are in the range of 0 ÷ 500 kNm/rad/m track.
4. THE INFLUENCE OF THE TORSIONAL RESISTANCE OF

FASTENINGS ON THE STABILITY OF CWR TRACK
For former example the vehicle was eliminated to distinguish the influence of
torsional resistance of fastenings when the track is loaded only with temperature
variations. The analises of results relieve, in absence of vehicle load, a growth of
Tmin with about 2%, and a growth of Tmax with about 3,5%, comparatively with the
results obtained in presence of vehicle loads.
Fig. 7. The variation of critical temperature versus the value of rotation angle of rail in
fastenings θ2 for a constant value θ1 = 3·10-3 rad.
It was analised the influence of level θ1 ÷ θ2 size, for that the value of torsional
resistance of fastenings is constant, on the value of the critical temperature is
increasing. It was considered a reference value of resistant moment Mo = 333.75
Nm/rad/m track, adjusted in terms of vertical load from vehicle. It was considered
that the value of fastening stiffness for the rotations of rail in fastening which
exceed θ2 is equal with the stiffness for the first branch – i.e. it is that which
characterise the rotations between 0 and θ1. It was obtained the results shown in
fig. 7, for a constant value θ1 = 3·10-3 rad – which correspond at a K type fastening
– and for values of θ2 in range 3·10-3÷ 30·10-3 rad. The values of θ2 were considered
for a step by 1·10-3 rad. It was observed that Tmax stays constant, the corresponding
rotation being less than θ1, and Tmin decreases with about 4% for θ2 = 28·10-3 rad,
thereupon the value of Tmin being constant because the appropiate rotation is
smaller than θ2, i.e. it can say that the behaviour of model is bi-linear, not tri-linear.
Fig. 8. The variation of critical temperature versus the decrease of lateral resistance.
In order to mark out the effect of torsional resistance of fastenings on stability of

CWR track it was also analised the situation with reduced lateral resistance on the
central zone of the misalignment. Hereby, it was investigated the situation in which
the lateral resistance is reduced up 100%, 75%, 50% and 25% at 4 sleepers placed
on central zone of the misalignment, as for the situation in which the torsional
resistance is zero, as for the situation in which that has the same values like in the
previous numerical experiment. It was considered θ1 = 3·10-3 rad and θ2 = 12·10-3
rad. The results of this numerical experiment are shown in fig. 8. Though the
influence of torsional resistance is small, if the results are compared it can be
observed that Tmin is more influenced by the torsional resistance of the fastenings
than Tmax.
Almost all developed models neglect the influence of torsional resistance because
they are taking in account the small influence this parameter.
However, it was observed in numerical and laboratory experiments that in some
cases the torsional resistance of the fastenings can have a significant influence on
the track stability. Therefore a more trusty analysis of the track stability must take
in account the effect of this factor, also.
References
1. *** - ERRI D170-RP5 Track component dimensions. Standardisation of characteristics and
acceptance tests. Study of fastening system characteristics and test methods, European Rail
Research Institute, Utrecht, September 1994.
2. *** - ERRI D202-RP3 Improved knowledge of forces in CWR track (including switches) -
Theory of CWR track stability, European Rail Research Institute, Utrecht, February 1995.
3. Bao, Y. L. - Three-dimensional stability/lateral shift analysis of continuous welded rail (CWR)
track and innovative methods to enhance CWR track performance, PhD Thesis, University of
Illinois at Urbana-Champaign, Urbana, Illinois, 1998.
4. Chatkeo, Y. – Die Stabilität des Eisenbahngleises im Bogen mit engen Halbmessern bei hohen
Axialdruckkräften, Mitteilungen des Prüfamtes für Bau von Landverkehrswegen der TU
München, Heft 46,1985.
5. Dósa, A., Ungureanu, V.V. - SCFJ – model discret de pierdere a stabilităţii căii fără joante,
Infrastructuri eficiente pentru transporturi terestre – Zilele Academice Timişene, 24-25 mai 2007,
Timişoara.
6. Esveld, C. - A better understanding of continuous welded rail track, Rail Engineering
International, No. 4, 1996.
7. Esveld, C. – Improved knowledge of CWR track, Delft University Press, 1997.
8. Herman, A. – Contribuţii la calculul căii fără joante, Teză de doctorat, Universitatea “Politehnica”
Timişoara, Facultatea de Construcţii, Timişoara, 1998.
9. Radu, C., Poştoacă, S., Sianu, C. - Unele aspecte privind stabilitatea căii fără joante,
Consfătuirea pe ţară a lucrătorilor de drumuri, poduri şi căi ferate - vol. II , Tuşnad, 1982.
10. Rangosch, R. S. - Lagestabilität lückenloser Meterspurgleise in kleinen Bogenradien, Dissertation
ETH Nr. 11322, Schriftenreihe des IVT Nr. 108 Zürich, Oktober 1995.
11. Rădulescu M. - Calea fără joante – Stabilitatea şi calculul eforturilor, Editura Transporturilor şi
Telecomunicaţiilor, Bucureşti, 1963.
12. Ungureanu, V.V. - Modele de pierdere a stabilităţii căii fără joante, Referatul nr.3 din cadrul
stagilului de pregătire pentru doctorat, Universitatea “TRANSILVANIA” din Braşov, Facultatea
de Construcţii, Braşov, 2006.
13. Van, M.A.- Stability of Continuous Welded Rail Track, PhD Thesis, Delft University Press, 1997,
ISBN:90-407-1485-1.
Concrete-filled steel tubular columns.

Shear connection design
Călin Moga, Ştefan I. Guţiu and Gabriel Urian
Faculty of Civil Engineering, Technical University, Cluj-Napoca, 400027, Romania
Summary
This paper presents an analytical evaluation method for the longitudinal shear
stresses over the interface between the structural steel and the concrete filled steel
tubular section.
A working example is also given in this paper, which can be useful in the design
activity of such type of structures.
KEYWORDS: composite columns, concrete filled tube, shear connection.
1. INTRODUCTION
Where composite columns of the type concrete filled rectangular hollow sections
are subjected to significant transverse shear, provision shall be made for the
transfer of the corresponding longitudinal shear stress at the interface between steel
and concrete.
If the natural bond (adhesion, interface interlocking and friction) is not enough to
achieve the required shear resistance, the mechanical connectors will be used.
Shear connectors should be provided, based on the distribution of the design value
of longitudinal shear, where this exceeds the design shear strength τ Rd , given in
Table 1.
Table 1
CROSS-SECTION TYPE τ Rd (N/mm2)
Concrete filled circular hollow section 0.55
Concrete filled rectangular hollow section 0.40
626 C. Moga, Şt. I. Guţiu, G. Urian
2. LONGITUDINAL SHEAR STRESSES
To determine the longitudinal shear at the interface between concrete and steel, the
filled concrete, Figure 1, is transformed into an equivalent steel section, Figure 2.a,
using the modular ratio n.
The equivalent in steel width of the concrete, Figure 2.a, will be::
bc
te = (1)
n
where the modular ratio can be taken as:
Ea
n ≈ 2 ⋅ ni = 2 (2)
E cm
Ea - modulus of elasticity of structural steel;
Ecm - secant modulus of elasticity of concrete.
Figure 1 Figure 2
In the case of the columns of type concrete filled rectangular hollow steel section,
in the absence of connection at the interface between steel and concrete, the shear
stresses in accordance with Juravsky formula, Figure 2.b, will be:
VSd Sa
τ1 = (3.a)
b⋅I
VSd Sa
τ2 = (3.b)
2t ⋅ I
VSd Sa
τ3 = (3.c)
(2t + t e ) ⋅ I
1
VSd (Sa + Sc )
τ max = n (3.d)
(2t + t e ) ⋅ I
where:
Sa – section modulus of the steel
component referred to neutral axis y-y:
 h + tw  h2
• Sa = bt f  w  + 2t w (4)
 2  8
Sc - section modulus of the concrete
referred to neutral axis y-y:
h 2w
• Sc = b c (5)
8
Figure 3
I - moment of inertia of the equivalent in steel of the entire cross-section:

1
I = Ia + Ic (6)
n
The longitudinal shear force between steel and concrete will be:
L = (τ 2 − τ3 ) t e = ∆τ ⋅ t e (7)
By replacing the shear stresses values τ 2 ,τ3 into eq. (7) it results:
VSd Sa 1 1  V S t e2
L=  −  t e = Sd a [F/L] (8)
I  2t 2t + t e  I 2t (2t + t e )
The average shear stress at the interface between steel and concrete is:
L
τF = [F/L2] (9)
∑ bi
hw
where: ∑b i = bc + 2
2
= b c + h w is the half perimeter of the contact surface
between steel and concrete.
If: τf > τ Rd , the natural bond is not enough to achieve the shear resistance and will
be necessary to use mechanical shear connectors.
The number of the mechanical shear connectors for the entire length l of the
column results from the condition:
l V S t e2 l
N≥L = Sd a (10)
PRd I 2 t (2 t + t e ) PRd
3. WORKING EXAMPLE
The connection for a composite column of type concrete filled rectangular hollow
steel section is designed.
Design data (Figure 4)
Materials and characteristics:
Box section - Steel: S 355
- fy=355 N/mm2
- Ea= 210 000 N/mm2
25 ⋅ 40 3 22 ⋅ 37 3
Ia = − = 40 470 cm 4
12 12
Concrete - Class: C 25/30:

fck = 25 N/mm2
Ecm = 30 500 N/mm2
Figure 4 22 ⋅ 37 3
Ic = = 92 864 cm 4
12
Connectors: headed welded studs:
d = 19 mm; h= 100 mm
fu = 450 N/mm2
γ v = 1.25
Connection design
The mechanical design model is presented in Figure 5.
It results:
Sa =1235 cm2
1 1
I = Ia + I c = 40 470 + 92 864 =
n 13.8
= 47 200 cm 4
where:
n = 2·ni =13.8
E 210 000
ni = a = = 6.88
Figure 5 E cm 30 500
The longitudinal shear force will be:

VSd Sa t e2 80 000 ⋅ 1235 1.6 2
L= = = 388 daN / cm
I 2t (2t + t e ) 47 200 2 ⋅ 1.5(2 ⋅ 1.5 + 22 / 13.8)
where: te = 22/13.8 = 1.6 cm
The adhesion-friction stress over the interface between concrete and steel is:
L 388 −1
τf = = 10 = 0.66 N / mm 2 > τ Rd = 0.40 N / mm 2 ,
∑ bi 59
so it results that are necessary mechanical connectors,

where: ∑b i = 2 ⋅ 18.5 + 22 = 59 cm .
The design shear resistance of a headed stud automatically welded, in accordance

with EC 4, is:
 0.8f u π d 2 / 4 0.8 ⋅ 450 π ⋅ 19 2 / 4 −3
 = 10 = 81.65 kN
 γv 1.25
PRd = min  2
 0.29 α d f ck ⋅ E cm 0.29 α19 25 ⋅ 30 500
2
 = = 73.13 kN
 γv 1.25
The number of connectors will be:

388 ⋅ 1000
N≥ = 53 .
7 313
For two rows of connectors the maximum distance between these will result:
10 000
e= = 385 mm .
1+ N/2
The proposed arrangement of the headed stud connectors on the composite cross-
section is presented in figure 6.
Figure 6
4. CONCLUSIONS
This paper presents an analytical method for the evaluation of the shear stresses
over the interface between concrete and steel component for the composite
columns of type concrete filled rectangular hollow steel section.
If the natural bond is not enough to achieve the required shear resistance, it is
possible to use mechanical shear connectors, the widely used types being headed
studs.
The numerical example which is also given in the paper can be useful for the
design of composite steel-concrete columns.
References
1. Bia,C., Ille,V.,Soare,M.V.:Rezistenţa materialelor şi teoria elasticitţii.E.D.P.Bucureşti.1983
2. Moga,C.: Stâlpi cu secţiune compusă oţel-beton. Referat de doctorat. U.T.Cluj-Napoca.2006
3. *** EUROCODE 2. Design of concrete structures. EN 1992.
4. *** EUROCODE 3. Design of steel structures. EN 1993.
5. *** EUROCODE 4. Design of composite concrete and steel constructions. EN 1994.
Composite columns with circular section.

Shear transfer mechanism
Petru Moga, Ştefan I. Guţiu and Cătălin Moga
Faculty of Civil Engineering, Technical University, Cluj-Napoca, 400027, Romania
Summary
The mechanisms of the shear transfer over the interface between the circular steel
tube and the concrete core as well as the design of the shear connection are
presented in this paper.
A numerical example for the evaluation the longitudinal shear stresses over the
interface between structural steel and concrete is also presented here.
KEYWORDS: composite columns, shear stresses, shear transfer, connection.
1. INTRODUCTION
A composite column may either be a concrete partially or completely encased

section or a concrete- filled section, Figure 1.
Figure 1
632 P. Moga, Şt. I. Guţiu, C. Moga
In case of the concrete-filled hollow steel sections, there are three mechanisms
which are often referred to as the natural bond, by which shear stresses can be
transferred over the interface between the steel tube and the concrete core, these
are: adhesion, interface interlocking and friction, Figure 2.
Figure 2
If the natural bond is not enough to achieve the required shear resistance there is
the possibility of using mechanical shear connectors, the most widely used types
being headed studs and the shot-fired nails.
The shear stresses, which take place on the interface steel-concrete, are shown in
Figure 3.
Figure 3
Taking into account that the structural

steel and the concrete have different
mechanical characteristics, the concrete core is
transformed into an equivalent steel section
using the modular ratio n.
The mechanical model is presented in
Figure 4.
The equivalent in the steel area of the
concrete core will be:
Figure 4
Ac
A c.echiv = (1)
n
where the modular ratio can be taken as:
Ea
n ≈ 2 ⋅ ni = 2 (2)
E cm
2. TANGENTIAL STRESSES
The state of tangential stresses caused by the shear force T=VSd.c is presented in
Figure 5, where τ xz is the shear stress given by the Juravski formula:
TS y
τ xz = τ zx = (3)
bI y
where:
- Sy – section modulus of the slipping portion referred to neutral axis:
Sy =
2
3
(R 2
− z2 )3
(4)
- b – width of the section at the distance z from the neutral axis:
b = 2 R 2 − z2 (5)
- Iy – moment of inertia of the cross-section:
πR 4 πD 4
Iy = = (6)
4 64
By replacing the above parameters into relation (3) it results:
T(R 2 − z 2 )
τ xz = (7.a)
3I y
4T z2 
τ xz = 1 − 2  (7.b)
3 A  R 
Figure 5
3. LONGITUDINAL SHEAR BETWEEN

STEEL TUBE AND CONCRETE
The longitudinal shear stress τl , equal with the normal component τ n , of the shear
stress τ xz , will be:
4T z2  z 4T 1
τ n = τ l = τ xz sin α 0 = 1 − 2  ⋅ , τl = z(R 2 − z 2 ) (8)
3 A  R  R 3 A R3
The maximum value of τl and τ n is obtained from the condition:

R
f ' (z) = 0 ⇒ z = .
3
It results:
8 T T
τ n . max = = 0.51
9 3A A
The longitudinal and the normal shear stresses variation are presented in Fig 6.
Figure 6
The sum of τ n divided by a quarter of the interior surface of the steel tube will give
the value of the longitudinal tangential stress on the interface between steel pipe
and concrete core:
2 
4 R ⋅ τ n . max 
τf =
4 ∑ τn
= 
3  = 689 ⋅ 10 −4 T (9)
2πR c 2πR c R cR
where:
Rc - the interior radius of the steel tube;
R - the radius of the concrete core taking into account the modulus ratio and
results from:
πR c2 Rc
A c.echiv = = πR 2 ⇒R= (10)
n n
By replacing the value R in the relation (9) it is obtained:

n n
τ f = 689 ⋅ 10 −4 2
T = 0.275 2 T (11)
Rc dc
The longitudinal shear force for the entire length of the composite column and
respectively for a half of cross-section, will be given by:
1 n
L l = τf πd c ⋅ l = 0.43 T⋅l (12)
2 dc
If the natural bond is not enough to achieve the shear resistance, it will be
necessary to use mechanical shear connectors and their number will result from the
condition:
2 ⋅ Ll
N≥ (13)
PRd
In accordance with Eurocode 4, the design value of the longitudinal shear strength
τ Rd in case of a concrete filled circular hollow sections is τ Rd =0.55 N/mm2.
From the relation (12) the maximum value of the shear force when it is not
necessary to use mechanical shear connectors will result:
d c2
Tmax = 2 (results T in [N] for dc in [mm] ) (14)
n
It is necessary to underline that the shear force T represents the part of the shear
force, taken over by the concrete component of the composite column, which can
be evaluated with the relation:
A cv
VSc = VSd = VSd − VSa (15)
A av + A cv
2 A
where: A av = A a and A cv = A c.echiv = c
π n
4. NUMERICAL EXAMPLE
For a composite steel-concrete column the longitudinal shear stress over the
interface between the circular steel tube and the concrete core is evaluated.
4.1. Design data
Composite column cross-section and loading (Figure 7)
Cross-section characteristics:
Circular pipe: 325 × 10 mm

Steel: S 355:
- fy=355 N/mm2
- Ea= 210 000 N/mm2
- Aa=99 cm2
Concrete: C 25/30:
- fck = 25 N/mm2
Figure 7 - Ecm = 30 500 N/mm2
- A c = 731 cm 2
4.2. Longitudinal shear
Modulus ratio n:
Ea
n = 2 ⋅ ni = 2 = 13.8
E cm
Equivalent in steel area of the concrete core:
731
A c.echiv = = 53 cm 2
13.8
Shear force of the concrete component:
2 2
A av = A a = 99 = 63 cm 2
π π
53
VSc = 80 = 36.55 kN
63 + 53
The shear force, which can be achieved by the natural bond:
d c2 305 2 −3
Tmax = 2 =2 10 = 50 kN > VSc = 36.55 kN ,
n 13.8
so it results that are not necessary mechanic shear connectors.
The longitudinal shear stress:
n 13.8
τ f = 0.275 2
T = 0.275 2
36 550 = 0.40 N / mm 2 < τ Rd = 0.55 N / mm 2
d c 305
5. CONCLUSIONS
The method to design the longitudinal shear connection between the circular steel
tube and the concrete core, which is presented in this paper, is very simple and easy
for applying in the design activity.
The presented method is one originally method, based on the elasticity theory, so
experimental tests are useful.
References
1. Bia,C., Ille,V.,Soare,M.V.:Rezistenţa materialelor şi teoria elasticitţii.E.D.P.Bucureşti.1983
2. Moga,C.: Stâlpi cu secţiune compusă oţel-beton. Referat de doctorat. U.T.Cluj-Napoca.2006
3. *** EUROCODE 2. Design of concrete structures. EN 1992.
4. *** EUROCODE 3. Design of steel structures. EN 1993.
5. *** EUROCODE 4. Design of composite concrete and steel constructions. EN 1994.
The meshing influence upon the design model of reinforced

concrete chimneys structures
Ioana Entuc1, Mihai Budescu2, Ioan Ciongradi2 and Gabriel Oprisan1
1
Civil Engineering Department, “Gh. Asachi” Technical University, Iasi, 700050, Romania
2
Structural Mechanics Department , “Gh. Asachi” Technical University, Iasi, 700050, Romania
Summary
Designing of the chimney structures assumes the establishing of a mathematical

model made up of a certain number of elements named finite or discrete elements
interconnected by a finite number of links. A mathematical model supposes a fine
meshing of the structure which allows a structural analysis using algebraically
equations systems.
The authors have analyzed the meshing influence of a chimney structure in
accordance with different type of finite elements.
KEYWORDS: finite element; calculus model, industrial chimney, fundamental

modes, natural frequency.
1. INTRODUCTION
Modeling procedures represent an important issue in structural design when taking

into account the particularities of structural system.
A dynamical analysis of a RC industrial chimney from Portugal (fig.1) using two
different types of models as compared with experimental measurements has shown
important differences between the results [3].
The first design model consists in a vertical cantilevered beam divided into 25
linear element bars (fig. 2).
The second model consists in spatial, solid, brick type finite elements, the structure
being divided on vertical direction in 25 rings, each ring being divided into 48
elements in the horizontal plane (fig. 3).
An experimental procedure has been realized by measuring on two perpendicular
directions x and y of acceleration at three different levels on the height of the
chimney, at every 20 …30 meters placed platforms. The structural vibrations were
caused by the wind action.
640 I. Entuc, M. Budescu, I. Ciogradi, G. Oprisan
The differences between the fundamental frequency values for the analyzed
chimney are presented in table 1.
a b
Figure 1. Industrial reinforced concrete chimney

a. Industrial Chimney, Pego 216 m height; b. Exterior platform
Figure 2. Fundamental modes of vibration – Model with beam finite elements

Figure 3. Fundamental modes of vibration – Model with solid finite elements
Tabel 1 Fundamental frequency for the industrial reinforced concrete chimney

Modes Natural frequency, (Hz)
of Model with beam type Model with solid type Experimental
vibration finite elements finite elements results
1 0.233207 0.247273 0.253
2 0.233207 0.247273 0.247
3 1.1915 1.187399 1.265
4 1.1915 1.187399 1.282
5 3.027815 2.862939 2.859
6 3.027815 3.063163 3.100
7 3.542195 3.063163 3.194
8 5.48583 3.709982 4.350
9 5.48583 5.564148 4.400
10 8.357798 5.564148 4.590
As a result of the case study presented before, the investigations have been
continued with the influence of the model meshing for the dynamic analysis of the
reinforced concrete industrial chimneys.
Choosing the adequate finite elements represents an important step in the
structure’s discretization. Therefore, the type of the selected finite element depends
on the following parameters: number of dimensions (1D, 2D, 3D); number of links
and the associated functions approximations. A proper choice of the finite element
type is important both for the quality of results and the necessary calculus effort.
The model represents a chimney which has 100 meters height and 10 meters the
exterior diameter at the base. In the analysis of the seismic response there have
been utilized different types of finite elements such as: linear finite elements –
beam, bar; plane finite elements – plate, shell; spatial finite elements – solid, brick)
In figure 4 – 13 they are presented the proper modes of vibration obtained for each
model after post processing.
Figure 4. Calculus model of chimney using 5 beam finite elements
Figure 6. Calculus model of chimney using 25 beam finite

50
49
48
...
...
3
2
1
100
99
98
...
...
...
3
2
1
Figure 9. Calculus models of chimney using plate elements (50 elements in vertical
direction and 12 elements in horizontal direction)
Figure 10. Calculus models of chimney using plate elements (50 elements in vertical
direction and 24 elements in horizontal direction)
Figure 11. Calculus models of cone shape chimney using plate elements (50 elements in
vertical direction and 24 elements in horizontal direction)
Figure 12. Calculus models of cylinder shape chimney using plate elements (50 elements in
Figure 13. Calculus models of cone shape chimney using brick elements (50 elements in
The variation of the values in case of the fundamental periods for the calculus
models analyzed using FEM is shown in table 2.
Tabel 2 Fundamental periods (in seconds) - The comparative study in case of using of
different type of finite elements applied on RC industrial chimney
beam beam beam beam beam plate plate plate brick brick
5 10 25 50 100 50 x 50 x 50 x 50 x 50 x
Fundamental modes
eleme eleme eleme eleme eleme 12 24 24 24 24

nts nts nts nts nts eleme eleme eleme eleme eleme
nts nts nts nts nts
Nodes Nodes Nodes Nodes Nodes Nodes Nodes Nodes Nodes Nodes
: 12 : 17 : 32 : 57 : 107 : 612 : 1224 : 1224 : 2448 : 2448
Fig. 1 Fig. 2 Fig. 3 Fig. 4 Fig. 5 Fig. 6 Fig. 7 Fig. 8 Fig. 9 Fig 10
1 1.5306 1.5099 1.5041 1.5033 1.5031 2.4866 2.4432 2.2791 1.5321 1.4317
2 1.5306 1.5099 2.4044 2.3998 2.3987 2.4866 2.4432 2.2791 1.5321 1.4317
3 0.25488 0.24364 0.11660 0.11659 0.11658 0.43173 0.42583 0.47474 0.26645 0.29785
4 0.25488 0.24364 0.086011 0.085742 0.085675 0.43173 0.42583 0.47474 0.26645 0.29785
5 0.11706 0.11670 0.043966 0.043773 0.043725 0.41143 0.42520 0.28517 0.26022 0.17538
6 0.094446 0.087894 0.038919 0.038875 0.038865 0.41143 0.42520 0.28517 0.26022 0.17538
7 0.094446 0.087894 0.026642 0.026491 0.026454 0.38691 0.39893 0.22603 0.24347 0.14153
8 0.050601 0.045325 0.023413 0.023341 0.023323 0.38691 0.39893 0.22451 0.24347 0.13837
9 0.050601 0.045325 0.017866 0.017741 0.017710 0.30999 0.30549 0.22451 0.18907 0.13837
10 0.040337 0.039223 0.016790 0.016688 0.016663 0.30023 0.30549 0.19539 0.18784 0.12252
11 0.034535 0.027727 0.013128 0.012997 0.012964 0.30023 0.30202 0.19539 0.18784 0.12252
12 0.034535 0.027727 0.010812 0.010651 0.010612 0.20501 0.20659 0.18405 0.12808 0.11367
3. CONCLUSIONS
The realized study is based on the fundamental modes of vibration of a reinforced

concrete industrial chimney. For each case, the analyzed parameters were the type
of the finite elements and the geometry shape of the structure as well as the
increase in the number of finite elements. The variation of the finite elements
amount has been realized by meshing of the structure both on vertical (5, 10, 25, 50
elements) and horizontal (12 and 24 elements) directions.
Behind the comparative study it has been observed that the discretization of the
flexible structures like industrial chimneys show good results when using the 1D
finite elements (beam or bar) and 3D finite elements (solid or brick).
Therefore, the fundamental period range between 1.53006 – 1.5031s for the
chimneys’ model divided in 5, 10, 25, 50 and 100 bar elements; for the chimney
divided with plate elements in 50 and 100 of rings the fundamental period
decreased between 2.4866s to 2.2791s; in case of structure model with brick
elements divided in 50 and 100 of rings the numerical results have been increased
from 1.4317s up to 1.5321s.
It has been observed from the numerical analysis that the using of plate type finite
element lead to around 60% longer periods. Using of bar type elements having the
length H/10 … H/50 is proper in designing for seismic and wind actions. Brick
finite element is proper to establish the state of stress and strain for all type of
actions (radial displacement - chimney lifting, temperature differences).
References
1. ***ALGOR 2000 (1998). “Concrete Design Manual,” Chapter III “Design for ACI 318-95,”
Chapter VI “Design for BS 8110-85 R1989,” Chapter VII “Design for Eurocode 2,” and Chapter
VIII “Design for NZS 3101-95,” Ed. Computers and Structures, Inc., Berkeley, California, USA.
2. Diaconu D., Hobjilă V., Ciongradi I. (2000). “Principii şi aplicaţii în calculul automat al
structurilor termodeformabile” Ed. Gh. Asachi Iasi,
3. Entuc, I.- (2004). “Contributii la calculul si reabilitarea cosurilor de fum industriale” Teza de
Doctorat, Universitatea Tehnica “Gh. Asachi” Iasi, Facultatea de Constructii şi Instalatii.
Temperature Distribution Computing System for 2D domains

Dan Diaconu-Şotropa1, Ioan Păvăloi2
1
Faculty of Civil Engineering, “Gh. Asachi” Technical University, Iasi, 700050, Ramania
2
Institute for Computer Science, Romanian Academy, Iasi branch, Iasi, 700506,Ramania
Summary
Many problems involve the numerical solution of a set of equations in a

complicated shape domain. The solution of such problems requires the domain to
be discretized to produce a set of points on which the numerical algorithm can be
based. For some problems, the generation of a suitable grid/mesh can be as
demanding as the effort required to perform the computations for which the grid
was intended.
This paper show the way of technical accomplishment of objective meant to
analyze stationary heat transfer at discretizeable two-sized structures with
quadrilateral type finite elements.
KEYWORDS: finite elements, mesh generation, quadrilateral mesh, quadtree, heat

transfer, thermic deformation, elastic deformation
1. INTRODUCTION
The finite element method (FEM) is a fundamental modeling technique in

widespread use in the engineering analysis community.
Significant progress over the last two decades in the area of numerical methods and
software technologies has enabled to solve still more and more demanding and
complex problems.
Researches were developed in order to study structures’ behavior at actions
resulted from environment’s temperature variations (this is the situation of climatic
areas with very high or very low temperatures) or from technological reasons (at
equipments from a nuclear plant, for example), and also from fires.
So the research focused to study the domain of structures deformation capability,
conductive heat transfer and thermic deformation capacity (respectively, elastic
deformation of structures created by dynamic variation of temperature) using finite
elements method (MEF)[1].
648 D. Diaconu-Şotropa, I. Păvăloi
As experimental researches needed to determine the behavior of structures are

quite expensive, especially those who use fire, researches were focused to develop
numerical experiment with applications in extreme temperatures area.
The presented work is motivated by current trends in present legislative
environment regarding fire security of buildings exposed to large variations of
temperature, especially to fire hazard [2].
2. GENERAL PRESENTATION OF WINFIDIAS SYSTEM
The WinFidias system is a set of modules that provides:

- 2D and 3D polygons input; The input is obtained through two modules, that
are parts of the WinFidias system:
- PolyRead – gets the polygons description through text files (.txt),
binary files (WinFidias internal files with .iid extension) or .dxf files.
- PolyEdit – gives the possibility to edit 2D and 3D polygons that will
be subsequently processed;
- 2D and 3D polygons processing. This can be done through QuadGen module.
QuadGen, gives the possibility to get a division made of quadrilaterals and is
based on Quadtree Method.
- polygons editing and polygons’ internal elements (points, segments and holes)
editing.
- analysis of the mesh result for one or more polygons. The result of this
analysis will be saved as mesh reports.
- Mesh results visualization. The results of the mesh process can be viewed and
the unexpected errors reported by the analysis module can be fixed. The mesh
results can also be printed.
- the calculculus of temperature distribution in one-dimensional, two-
dimensional and three dimensional structures, not only in stationary but also
in non-stationary conditions;
- visualization of data processing results and reports issuing.
2.1 PolyRead Module
This module gets the 2D and 3D polygons description from:

- .txt files (text files);
- .iid files (bynary files; are WinFidias’ own files);
- .dxf files.
Text files are very easy to manage and contain the elements description using
absolute 3D coordinates. These files can be manually edited, using any text editor,
like Notepad. Elements can be points, segments, holes and polygons. Moreover,
these text files can contain a mesh division already done and this make them very
useful to import in WinFidias the mesh results obtained with other mesh programs
and packages.
Binary files used (files with .iid extension) are WinFidias system’s specific files.
Using .dxf files as input gives the possibility to use programs such as AutoCAD to
define the input for mesh processing.
When the polygons description is not available in the input files (polygons as
distinct entity) being given only the description of all polygons edges (as
segments), this module is able to process these segments in order to obtain
polygons description that will be subsequently used in mesh processing.
2.2 PolyEdit module
Using this module can be edited polygons and all polygons’ internal elements [3].
The editing of 2D polygons will be made in the plane that contains the polygon.
There are several methods to define the polygon’s plane [4]:
- selecting three (non-collinear) points that belongs to the plane;
- selecting a point and the plane’s normal;
- selecting a point and a parallel plane;
- selecting a point and a perpendicular plane;
- selecting a polygon;
- selecting some non-collinear elements that belong to that plane.
Polygons editing will be made in the selected plane using a set of graphic tools able
to design points, segments, holes and polygons. User has the possibility to select
2D coordinates (local coordinates), 3D coordinates (global coordinates) or relative
coordinates. A remark should be made here: will be visible all elements of
polygons already edited in other planes, that intersect or are included in editing
plane. Moreover, a set of tools is available to scale and view the image from
another angle by zooming and rotating the edited image.
When new elements are edited and at exit from editing module the program
analyses the editing results in order to report any unusual situations, such as:
- the presence of points or segments inside holes;
- the presence of points on polygon vertices;
- two holes with a common area;
- two polygons with a common area.
Note: Except polygons, all other edited elements must be inside a polygon or on a
polygon edge.
User is informed at the appearance of above-mentioned situations and such a

situation is distinctly marked on display so the user can easily identify and correct
it.
The default view mode is 3D but the system has a set of tools to pan, zoom and
rotate the image, to view only some selected elements, to project selected elements
on a defined plane, etc., that make the system more easy to use. During editing in
2D and viewing in 3D user can select only a set of polygons to be visible. This
option is useful, for example, in complicated building designing, as it gives the
possibility to process only one level of such a building.
2.3 QuadGen Module
Finite element analysis for pre-stressed concrete structures is an important part of

the whole design process [5, 6, 7]. Advantages of quadrilateral isoparametric
elements versus triangles are numerous and well-known in this domain. However,
in civil and building engineering applications, a commercial-quality mesh
generator must have additional capabilities, such as:
- ability to process domains with numerous & arbitrarily placed internal points,
segments and holes ("point supports", "line supports" and “hole supports”);
- ability to generate highly regular meshes, in order to minimize the number of
transitional (distorted) quadrilaterals, whose shapes are significantly different
from a rectangle.
The method used inside WinFidias system to get a quadrilateral mesh division is
Quadtree Method. This method is implemented inside QuadGen module. The
implementation satisfies the two above-mentioned requirements.
The implemented algorithm has four steps:
- initialization of the quadtree;
- quadrant and boundary smoothing;
- finite element mesh topology from the modified quadtree;
- mesh finalization.
2.4 Processing
The WinFidias system divides the polygons in quadrilaterals [8]. The user can
select also a so called level for division, a constant value that represents a value
considered at edge division. Depending on this value the system will generate a
certain number of elements (the user cannot choose the number of final elements
obtained as a result of division process) [9, 10].
The user can process all polygons or only a subset of them. The processing order
can be:
- set by the user, for all polygons or for a subset of them;

- set by the system (depending on some criteria, the optimal mesh order will be
established), for all polygons or for a subset of them.
Note: The processing order could seem to be insignificant, but is under study now
to see how the mesh order affects the final mesh quality.
Two or more polygons can have a common edge. This is a common situation met
in stress analysis in building field. So vertices of polygon’s elements placed on an
edge must be the same with elements’ vertices of another polygon that contains the
same edge (the edge is a common edge).
2.6 Mesh Analyzing
There are many criteria to evaluate the mesh quality in 2D/3D. Some of them are:
1. The variation in the area of the elements should not be too large.
2. The aspect ratio of triangular elements should be as close to 1 as possible.
The aspect ratio of a triangular element is defined as the ratio of the
circumradius of the triangle to twice its inradius. Hence the aspect ratio of an
equilateral triangle is exactly 1.
3. The ratio of the largest to the smallest edge of the element should be close to
1.
4. The ratio of the largest to the smallest angle of the element should be close to
1.
5. The ratio of the area of the largest/smallest element to all the immediate
neighbors should not be drastically low/high.
At every moment the user can decide to analyze the mesh result of the processing
of one or more polygons. Using this module he can appreciate the mesh quality and
if the quality is enough good for his purposes. The quality criteria were presented
above.
Using one of the above-mentioned criteria, the user can have a look of the elements
that don’t respect a criterion. There are several pre-defined thresholds but the user
can define his own thresholds for every criteria. Using the result of this module, the
user can decide to reprocess one or more polygons or only a part of them.
2.7 Correction module
This module offers the possibility to examine the result of the analysis module for
one or more polygons division [11, 12]. The results of the analysis could be
examined in sequential manner and in every moment some of the following actions
could be done:
- reprocessing of one or more polygons;
- editing of new elements inside polygons. If a new element will be inserted

inside an already processed polygon, the mesh result will be deleted and the
polygon has to be reprocessed again.
- the mesh result can be manually edited;
- a part of an already processed polygon can be reprocessed.
2.8. Module for Temperature Distribution Calculus
The calculus of temperatures distribution in bars structures, both in 2D and in 3D

structures, in both in stationary and non-stationary state is based on routines
package Fidias, developed in Iasi, at Faculty of Civil Engineering.
Structures could be homogeneous or non-homogeneous ones. Generally, a set of
parameters may vary from one element to another, but are constant inside it. This is
the case of inner heat sources flow, thermic conductivity coefficients, unit thermic
flow, environmental temperature.
Fidias package accepts for conductivity coefficient a variation law dependent on
temperature : λ = λ(T).
The boundary conditions are of the following types:
- Dirichlet, when the temperatures at some points located on surface border are
given;
- Neumann, when the heat flow on a limited area is given;
- Cauchy, when the heat transfer by convection, on a limited part of outline, is
considered (the environmental temperature and the convection coefficient are
known).
These conditions are considered to be constant during the surface processing.
According to the circumstances, both XYZ general system of axes and the local
system of axes will be considered.
The characteristic element matrices are based on the equilibrium equations
expressed in a matrix shape (1, 2):
[ K e ]{Te } = {Qe }−stationary phenomena (1)
[Ce ]{Te } + [ K e ]{Te } = {Qe }− non −stationary phenomena (2)
often based on a variational formulation or a residual formulation (Galerkin

variant). By expanding and assembling the element equations, the matrix equation
for the entire structure is obtained.
The calculation modules that solve the same type of heat transfer problem are
basicly identical, the difference consisting in the way of determining the element
matrix equations of thermal equilibrium.
- calculus of temperatures distribution in stationary-liniary heat transfer;

- calculus of temperatures distribution in stationary-non-lineary heat transfer;
- calculus of temperatures distribution in non-stationary-non-lineary heat
transfer.
WinFidias has modules for the analysis of structures modelled by using one-
dimensional and two-dimensional elements.
2.8.1. Module for analysis of structures modelled by using one-dimensional

elements
The accepted approximation function (shape function) used to describe

temperatures distribution, in 3D system of axes, is (3):
T( x , y, z, t ) = N1 ( x , y, z )T1 ( t ) + N 2 ( x , y, z )T2 ( t ) (3)
The problem of function continuity will be studied only at the nodes, as the
connections between the bars of the structure are made only at these points.
With an adequate coordinates transformation, the function representation in local
system of axes is (4):
T(ξ, t ) = N1 (ξ)T1 ( t ) + N 2 (ξ)T2 ( t ) (4)
where N1 and N2 are defined (5, 6) as:
N1 (ξ) = 0.5(1 − ξ) (5)
N 2 (ξ) = 0.5(1 − ξ) (6)
The elements assessment in the matrix equilibrium equation is done by obtaining
an exact solution of the involved integrals.
2.8.2. Module for the analysis of structures modelled by using with two-
dimensional elements
The finite element used in such an analysis is the plane quadrilateral, with only one
degree of freedom at each node.
The accepted approximation function (shape function), which describes the
temperatures distribution, in the general system of axes, is (7):
T( x, y, z, t ) = N1 ( x, y, z)T1 ( t ) + N2 ( x, y, z)T2 ( t ) + N3 ( x , y, z )T3 ( t ) + N4 ( x , y, z )T4 ( t ) (7)
With an adequate coordinates transformation, the function representation in the
local system of axes is (8):
T1 (ξ, η, ξ) = N1 (ξ, η)T1 ( t ) + N 2 (ξ, η)T2 ( t ) + N 3 (ξ, η)T3 ( t ) + N 4 (ξ, η)T4 ( t ) (8)
Read/edit model
description
analyze complete model

description and fill
data structures
Yes
Correct errors Erros ?
No
choose sub-domains
and mesh them
analyze sub-domains results
No Satifies
Reprocess conditions
Yes
Read/edit temperature
model description
Temperature distribution process
write output report

and data
stop
Figure1. Processing scheme

where N1, N2, N3, N4 are defined as (9, 10, 11, 12):
N1 (ξ) = 0.25(1 − ξ)(1 − η) (9)
N 2 (ξ) = 0.25(1 − ξ)(1 − η) (10)
N 3 (ξ) = 0.25(1 − ξ)(1 + η) (11)
N 4 (ξ) = 0.25(1 − ξ)(1 + η) (12)
References
1. Diaconu, D., Raspunsul sistemelor structurale la variatii dinamice din temperatura (teza de
doctorat), Universitatea Tehnica “Gh. Asachi”, Iasi, 1997.
2. Diaconu, D., Hobjila, V.,Ciongradi, I., Principii si aplicatii in calculul automat al structurilor
termodeformabile, Ed. “Gh. Asachi”, Iasi, 2000.
3. Farin, G., Curves and Surfaces for Computer Aided Geometric Design - A Practical Guide, 2nd
Edition, Academic Press, Boston, New York, 1990.
4. Gross, M., Visual Computing, Springer Publishing Company, Berlin, 1994.
5. Jones, M.T., Plassmann, P.E., Adaptive refinement of unstructured finite element meshes. Finite
Elements in Analysis and Design, vol. 25 (1-2), pp. 41-60, 1997.
6. Kreiner, R., Kroplin, B., Unstructured quadrilateral mesh generation on surfaces from cad, Proc.
of 4th Inter. Conf. on Numer. Grid Generation in Comp. Fluid Dyn. and Related Fields, Cromwell
Press, pp. 211-221, 1994.
7. Pavaloi, I., Bolea, C., Automatic Two-Dimensional Mesh Generation, ECIT 2002.Conference,
Iasi, Romania, 2002.
8. Pavaloi, I., Unstructured Mesh Generation on Planes based on a quadtree, Romanian Academy
Days, Oct. 2002.
9. Rivara, M.C., New mathematical tools and techniques for the refinement and/or improvement of
unstructured mesh, Proceedings 5th International Meshing Roundtable, Pittsburg, pp 77-86, 1996.
10. Samset, H., The quadtree and related hierarchical data structures, Computer Surveys, vol. 16,
no.2, pp.187-260, 1984.
11. Shepard, M.S., Georges, M.K., Reliability of automatic mesh generation. Comp. Meth. Appl.
Mech. Engng, pp. 443-462, 1992
12. George, P.L., Automatic Mesh Generation : Applications to Finite Element Method. John Wiley
& Sons, 1991.
Possibility increase cooling tower service life in terms of long-

term monitoring
Jiří Bydžovský1, Amos Dofka2 and Zdeněk Šnirch3
1
Institute of Building Materials and Elements, University of Technology, Faculty of Civil
Engineering, Brno, Czech Republic
2
3
Summary
In terms of long-term monitoring of records from long-term cooling towers

construction monitoring tower we have to be able to unite the records and use them
for new progressive complex technology, susceptible to define time of service life
and dependability of cooling tower as reinforced construction. Paper more
precisely deal with influence of repair action on service life and degradation
processes at construction. Monitored dependencies are then reflect on degradation
curves each for a singles construction characteristics (compression strength of
concrete, degradation of surface, volume of carbonatation etc.).
KEYWORDS: degradation, cooling tower, service life, monitoring.
1. INTRODUCTION
Cooling towers with natural gravity flow are designed in principle, in all countries,
as reinforced concrete structures acting in a flexible state.
The durability of cooling tower casings is influenced not only by operational
loading, but also by manufacturing defects. The most important defects are an
insufficient concrete layer covering the steel reinforcement, poor workmanship or
insufficiently treated construction joints and low quality concrete [2].
Cooling towers are further repeatedly loaded by wind and temperature.
Temperature effects on concrete structures and their components become evident
by changes in volume. As the results of these effects, stress gradients are formed
and lead to damage of the structure. This stress can cause crack formation,
separation of the surface layer, or excessive deflection.
The temperature cycle causes continuous damage to cooling towers. The most
serious damage occurs during interrupted operation of towers in the winter season,
when the water cyclically freezes in the shell and the result is the destruction of the
concrete. The consequence of repeated freezing and thawing is the disintegration or
crumbling of the concrete. The freezing water in capillaries of the structure
increases in volume during transformation into ice. The dilatation causes local
stress, which tears the surrounding concrete.
The next significant influence, which decreases the service life of cooling tower
casings, is degradation caused by aggressive chemicals. Some chemical substances
in solution can attack individual components of concrete. Aggressive chemicals
can be divided into: inorganic acids, organic acids, alkaline solutions, solutions of
salts and different mixtures in solution. Concrete Damage caused by acids is the
result of the reaction between acid and alkaline calcium hydroxide of hydrated
Portland cement. Soluble calcium compounds are formed in water during the
reaction and they are leached out. However, acids can dissolve even limestone and
dolomite aggregates.
Among the important aggressive factors belong aggressive gases such as CO2 and
SO2, which cause carbonation and sulphatation of concrete. Concrete carbonation
is the reaction between carbon dioxide (CO2) in the atmosphere, which has acidic
characteristics, and the products of cement hydration. Normal air contains carbon
dioxide in relatively low concentrations (0,03 %). The level of CO2 in industrial
atmospheres is generally higher.
Carbon dioxide penetrates concrete pores by diffusion and reacts primarily with
calcium and the products of cement hydration.
Carbon dioxide penetrates concrete pores by diffusion and reacts primarily with
calcium hydroxide, which is dissolved in water. The result of this reaction is the
decrease of concrete alkalinity, in the value of about pH 9.6, which causes the loss
of reinforcement steel protection which was initially secured by concrete. The layer
of passive reinforcement protection is disturbed and the corrosion of reinforcement
takes place [3].
It is known that one of the main consequences of degradation is the decrease of
concrete pH value, which leads to corrosion of reinforcement. The reinforcement
corrosion is mostly an electrochemical process which needs anode, cathode and
electrolyte. Just moisture in concrete mass of the cooling tower is the appropriate
electrolyte and the steel reinforcement forms the anode and the cathode. The
electric current flows between cathode and anode and the reaction results in
increase of metal volume because (Fe) iron oxidizes in Fe(OH)3 and Fe(OH).2 and
it is present as rust FeO(OH). For the progress of this reaction the presence of
water and oxygen are necessary. In concrete of good quality the corrosion process
is slow. Accelerated corrosion takes place when the ph value (alkalinity) of the
concrete decreases for, instance by carbonation, sulphation etc. or when aggressive
chemicals are added into cement and/or when the reinforcement is made from
different metals. Further reasons can be vagabond currents and local concentration
cells given by irregular chemical medium.
Without executing rehabilitation measures is it impossible to maintain the
reliability of a degraded structure. Every delay of rehabilitation lowers its
effectiveness and increases the costs for the realization of rehabilitation. In the case
that the rehabilitation measure is not realized the ageing of the concrete will
accelerate and the result, owing to concrete degradation, will be the end of building
service life.
2. SERVICE LIFE OF THE TOWER
The service life of an object is the ability to fulfill the functions demanded to the
limit stated under prescribed system of maintenance and repairs. The useful limit of
the structure is the state in which the further utilization must be interrupted.
For service life calculation of the structure, more methods can be applied. One of
these methods is the cubic method [4] which has the following calculation formula:
 S3  Q
Z = S +  ZZ + -S  × (1)
 2 × ZZ 2
 100
where ZZ is basic service life of the building; Z is total service life; S is age of the
building; Q is state of the object (for building without faults is Q = 100 %).
A great advantage of this method for the calculation of building service life is its
simplicity under sufficient correlation with the actual state of the building. At the
same time the method enables consideration of the immediate state of the building
and of its maintenance.
3. METHODS FOR DETERMINING THE STATE OF

CONSTRUCTION
For successful and permanent rehabilitation of cooling towers casings, it is

necessary to execute a detailed technical diagnosis of the state of concrete and
reinforcement. On the one hand, from the physio-mechanical point of view,
(chiefly compression strength, tensile strength of surface layers etc.) and on the
other hand from the point of view of physio-chemical parameters (mainly
carbonation degree, contamination grade etc.). By performing diagnostic tests, it is
possible to determine the actual state of the structure, to record the total extent and
intensity of failures and defects, and define the main reasons for its failure.
The normal method used for diagnosing these problems is the detailed building-
technical survey. By building-technical survey we mean activities, which are able,
by means of different field and laboratory methods and by expending adequate
costs, obtain information concerning the selected structure.
The extent of the building-technical survey is given primarily by the aim of the
survey, by the state of the examined building and by other conditions. The extent of
survey should be defined by the designer of reconstruction or rehabilitation who is
the most competent to evaluate the extent of information required [1].
Pursuant to long-term monitoring of building technical inquiry records we are able
to compile degradation curves. The cooling tower condition calculated on principle
of degradation curves ( Value of Q paragraph no.2) provide a very exactly
prediction of cooling tower service life.
4. METHODS OF PROCESSING DEGRADATION CURVES FOR

ITTERSON COOLING TOVER
Curves will be create for this parameters:

• Concrete surface failures
• Carbonatation depth
• Concrete compression strength
• Armature corrosion
• Modulus of elasticity
Surface of the shell
Curves will be created separately for outer and inner surface of the shell for
processing curves will be the shell divide in the following way (see next figure).
As a basis of processing degradation curves were used records from building
technical inquiryes from several periods. Degradation curves will be process apart
for every level of tower shell. From building technical inquiryes results in
each year will be created for every level a set containing valuation from each
parameter. The curves will be created from average value from several periods.
I
II
III
IV
V - Diagonal columns
Conditions of degradation curves for each parameter separately:
• Concrete surface failures

Degradation curves will be process for each caregory of failure depth defined
below:
Caregory of failure depth for Shell and diagonal columns
. from 0 to 5 mm
. from 6 to 25 mm
. from 26 to 40 mm
. from 41 to 60 mm
. other categories by 20mm levels, depends on
degradation level
• Carbonatation depth
Degradation curves which represent concrete carbonatation depths will be
processed from average value for each tower levels and for whole construction
(shell, diagonal columns). From each period will be processed one curve.
• Concrete compression strength and Modulus of elasticity

Degradation curves which represent compression strength will be processed from
average value for each tower levels and for whole construction (shell, diagonal
columns). From each period will be processed one curve.
Modulus of elasticity values will be calculated from compression strength on the

ČSN 73 0038 basis.
• Armature corrosion
Caregory ofcorrosion for Shell and diagonal columns
. 0 - 5 % profile decrease
. 31 – 50 % profile decrease
. more than 50 % profile decrease.
5. DEGRADATION CURVES
With regard to plenitude of degradation curves there are listed only chosen curves
for example. Figures number 1 and 2 containe degradation curves of concrete
failures of the tower shell include all categoryes of failures. These curves include
line for each category and line for general degradation which is sum of all
cateroryes. Important information for corect interpretation of our curves is fact of
complete repair of the monitred cooling tower in the year 1994. The monnitored
cooling tower was built in the year 1971.
Inside shell - whole tover

60,00
50,00 0-5 mm
6-25 mm
40,00 26-40 mm
40-60 mm
% failures
30,00 general degradation
20,00
10,00
0,00
1992 1994 1996 1998 2000 2002 2004 2006
year
Fig. 1. Degradation curves for Inside shell – whole tover( levels I-IV)
Level V. - Diagonal columns
20
18 0-5 mm
16 26-40 mm
14 26-40 mm
% of failures
12
40-60 mm
10
general degradation
8
6
4
2
0
1992 1994 1996 1998 2000 2002 2004 2006 2008
year
Fig. 2. Degradation curves for Diagonal columns - levelV)
Outside shell - whole tover
24
22
20
18
16
Depth (mm)
14
12
10
8
6
Carbonatation depth Concrete layer above armature [mm]
4
2
0
1992 1994 1996 1998 2000 2002 2004 2006 2008
year
Fig. 3. Degradation curves comparing carbonatation depth with concrete layer above
armature avarage values from whole tower shell- Outside side.
6. CONCLUSIONS
The rehabilitation of cooling towers is technically and economically a very difficult

task. Therefore it is necessary to consider carefully towers conditions. Long-therm
monitoring of cooling towers condition is definitely the right way of a very exactly
service life prediction. By optimization of results from these curves select we
should be able to propose an appropriate rehabilitation method from both the
technological point of view and from the point of view of rehabilitation materials
Values of already completed degradation curves will be used as basis to next
research.
Aknowledgements
The work was solved with the support of FF-TA/078 „Technology of

Determination of Life Cycle Residue of Cooling Towers and Smoke-stack with the
Use of Probability Analysis“ and with the support of MSM 0021630511
„Progressive Building Materials with Utilization of Secondary Raw Materials and
their Impact on Structures Durability“.
References
1. Drochytka, R., Dohnálek, J., Bydžovský, J., Pumpr, V. Technické podmínky pro sanace
betonových konstrukcí TP SSBK II. 1. vydání. (Technical conditions for rehabilitation of concrete
structures) Sdružení pro sanace betonových konstrukcí, Brno 2003, 210 stran, ISBN 80-239-
0516-3
2. Emmons, P., H., Drochytka, R., Jeřábek, Z. Sanace a údržba betonu v ilustracích. (Rehabilitation
and maintenance of concrete in pictures) 1.vydání. Brno: Akademické nak-ladatelství CERM,
1999, 334 stran, ISBN 0-87629-286-4.
3. Matoušek, M., Drochytka, R. Atmosférická koroze betonů. (Atmospheric corrosion of concrete)
1. vydání. Praha: IKAS, 1998, 171 stran, IBSN 80-902558-0-9.
4. BRADÁČ, A.: Teorie oceňování nemovitostí. 2. vydání.Akademické nakladatelství CERM, Brno,
1994, 482 s.ISBN 80-85867-36-2.
Some general considerations about the behaviour and retrofitting

solutions for the existing buildings with masonry structures
P.Tragakis 3, D.Stoica 1, S. Majewski2 and M. Voiculescu1
1
Civil Engineering Department, Technical University of Civil Engineering, Bucharest, Romania
2
Structural Engineering, Silesian University of Technology, Gliwice, Poland
3
Summary
Being in countries with high seismic risks and vulnerability or mining subsidence
we have the legacy of an existing buildings stock (with masonry and gravitational
frame structures) which must became safety from all the view points.
Depending of each historical period of erection, buildings and structures are
tributary to the knowledge level of that time including small or big mistakes, or
spectacular technical solutions.
Passing from country to country, especially in Europe, you may observe a lot of
similarities between architectural époques or structural styles, especially because
the gravitational rules were imposed, using French, Italian or German Design
Codes, more or less with the same content.
This may be a worse point but also an interesting idea, because being anywhere in
Europe (Germany, France, Italy, Romania, Greece or Poland) for the same
historical periods you may find the same architectural style of buildings and the
same design for their structures. More or less it mean that will be a good reason to
extend some of the information obtain for a country to the others, taking into
consideration the each item.
In the followings, only the buildings with masonry structures were considered in
the idea to optimize the possible retrofitting solutions.
Incremental-iterative procedures were used for approximate assessment of the
response of unreinforced masonry buildings to seismic loads. Working within the
framework of the shear-type structural behaviour, we propose a constitutive model
for the resisting elements able to account for both the variation in lateral stiffness
induced by cracking, and their sort of failure (i.e. ductile or brittle). Application to
a practical case demonstrates how, when modelling walls with generic windowing,
the manner in which they are subdivided into smaller resisting elements strongly
influences the predicted structural response and the manner of failure of each wall.
KEYWORDS: masonry structures, structures, earthquakes, ductility, stiffness,

strength, deflections, drifts
1. INTRODUCTION
Almost in the entire Europe the existing buildings stock may be divide in several
typological groups as followings:
• Very old buildings (included in the national heritages) with timber and/or
masonry (stone or solid bricks) structures made without any rules excepting the
traditions transmitted as an experience from masters to masters;
• Buildings erected until the very beginning of the 20th century, without specific
design codes, with timber and/or masonry structures which sometimes may
include also structural elements from simple or reinforced concrete;
• Buildings erected during the world wars, designed with very poor codes (only
gravitational) with masonry and gravitational frame structures;
• Buildings erected immediately after the second world war in concordance with
some seismic pre-codes with masonry and frame structures;
• Buildings erected during the 70’s period in concordance with some former
seismic or mining subsidence design codes(actually passed as knowledge –
technically and scientifically) with masonry and frames structures (cast in
place or pre-cast technology);
Each typological group consist in a large number of different possibilities and

masonry structures group is not different from this point of view.
Is very difficult for a small team to carry out in a short time all these possible cases
and in the followings just a part of the entire entity is presented.
In these papers our attention was focused only to the masonry structures, included
in the 2nd typological group of buildings, previously presented.
Depending of the computational tools of each period, after about 20 years of work
and about several hundred case studies of different building types (from different
typological groups) the authors attend to find some general considerations about
the behaviour of the buildings with masonry structures and their optimal possible
retrofitting solution. In this trend several idealised study cases were used.
2. STATE OF THE ART AND OF THE PRACTICE FOR THE DESIGN

PERIOD OF THESE TYPES OF BUILDINGS
For the preparedness of the idealised study cases, the state of the art and of the
practice tributary to that period were considered:
• These buildings present a lot of irregularities including architectural and
structural ones, in the plan or in the elevations;
• The masonry structures, in parallel with the gravitational frame structures
are probably the most irregular cases.
666 P.Tragakis, D.Stoica, S. Majewski and M. Voiculescu
• From the general considerations the masonry structures may be:

o Unitary – meaning practically one building, made from the
beginning and kept in this shape;
o Attached – meaning that in time the owners (different or not)
attached in plane (other bodies of the building) or on the vertical
(other levels);
• From the structural point of view the masonry structures may be:
o Optimum developed in plane and space – with similar stiffness
characteristics on each principal direction and on the vertical;
o „in lane or wagon type” buildings – where usually there is a huge
difference between the stiffness characteristics on principal
directions.
• Without any specific design tools, masonry structures were more or less
the architects appendage. The masonry walls thickness were generally in
accordance with the solid brick dimensions (7x14x28 cm) made from a
good class of bricks (about C50-C100) and a worse class of mortar
(between M4-M10);
• There was not any computations for these kind of structures, excepting the
centric compression stresses;
• Usually there are not any small columns or ties in these masonry structures
meaning that we may consider a URM solution;
• The coupling beams (lintels) are from masonry and timber or steel at the
bottom;
• The considered slabs thickness was about l/40 (where l=minimum
dimension of the slab, from both directions);
• Always the foundation soil was considered with a conventional pressure
between 200-300 KPa.
3. STUDY CASES FOR THE BEHAVIOUR OF THE EXISTING

GRAVITATIONAL FRAME STRUCTURES AND THEIR
POSSIBLE RETROFITTING SOLUTIONS
Taking into consideration all this “rules” (well known) 48 idealised study cases
were carried out to obtain some general considerations about the masonry
structures. The idealised cases consist into one exterior masonry resistance line
(one way frame) with 2 equal spans of 500 cm) with the tributary areas for
masonry piers and coupling beams and also for the level weights.
The used parameters were the levels number and the horizontal acceleration
(PGA). Three cases of PGA were considered (0.16g, 0.20g and 0.25g) and three
types of building behaviour (masonry piers with virgin coupling beams, masonry
piers with double hinged coupling beams and masonry piers with middle span
hinged coupling beams). For the levels number 6 variants were considered,
between 1-6 levels, with a level height of 300 cm).
All the beams were considered with 28x140 cm transverse section and for the piers
were considered two different sections: a „L” section for the both marginal piers
and a „T” section for the median one, with 28 cm thickness.
The chosen seismic excitation was INCERC NS 1977 scaled for different types of
PGA (as mentioned before).
In order to assess the seismic response of simple masonry buildings, one of the
most widely used structural models is the shear-type, in which floors behave as
rigid diaphragms elastically connected by interposed walls. If the vertical
displacements of the structure induced by horizontal actions are ignored, the
model's degrees of freedom are extremely reduced. In such a case, the quality of
the information obtained by applying the model depends upon the accuracy of the
assumptions regarding the behaviour of the single walls. In this regard, satisfactory
descriptions can be provided of the walls' behaviour during the brief initial loading
stage (phase-I), where each results uncracked and behaves elastically (Figure 1,
line O-A), as well as when they reach the corresponding limit condition (line A-B).
On the other hand, very little is known about the subsequent stage (phase II).
If the resisting element (either a wall or a masonry element of walls with openings)
is restrained well enough to its bases and is not excessively thin, the cracking
which forms where tensile stress is at a maximum ceases almost immediately, and
a new elastic phase may precede ultimate failure (line B-C). Under such conditions,
the state of stress begins to grow again and reach values well above those of the
preceding stage. Finally, collapse of the element may come about through brittle
failure (line C-D), if, that is, the compression exceeds the material's strength, or,
more frequently, through 'slow frictionful slipping of one part over another (line C-
E), in which case, the failure turns out to be ductile.
In the transition from phase I to faze II, the element's stiffness undergoes an abrupt
decline; thus, the stress distribution among the elements may vary considerably,
with consequent worsening of the state of stress in the as yet uncracked elements.
Such a situation favours a more uniform stress distribution throughout the entire
structure while this is still behaving elastically and wall cracking is not evident.
Greater uniformity of the stress distribution may delay failure of the elements. In
any case, their manner of ultimate failure, either ductile or brittle, depends on the
material's strength parameters and on the panel-aspect ratio.
Figure 2 – A masonry panel under

Figure 1 – Two-phase behaviour
compression and shear
With the aim of accounting for the essential features of both phenomena described
in the foregoing, in the following we propose a simple mechanical model able to
reproduce the phase-II elastic equilibrium state attained in rectangular masonry
elements. This permits the overall seismic response of a building to be deduced via
standard incremental procedures. In doing so, each of the walls with generic
windowing is treated as an assemblage of elements, some of which may be
cracked. In conclusion, the main characteristics of the resulting algorithm are
demonstrated in an example application.
A rectangular panel of unreinforced masonry is fixed at its bases, the other sides
being load-free (Figure 2). Let w be the width, h the height and t the thickness,
assumed constant and small with respect to w and h. The panel is subjected to a
relative displacement δ between its bases and parallel to them.
Faze I -In the absence of perceptible cracking, the stress distribution in the panel
can be determined, within a technically acceptable margin of error, through linear
elastic analysis by disregarding the effects of the material's dishomogeneities and
anisotropies. In particular, by extending the beam model to the panel [1] and
denoting by E and ν the material's Young's modulus and Poisson's ratio, the lateral
stiffness results to be
(1)
Where χ is the shear factor (6/5 for a rectangular-shaped cross-section) and η=

h/w. In order to account to some extent for the actual degree of constraint, a
reduced value of KI is usually assumed for masonry elements located near the top
of the building.
Phase II - When cracking becomes evident, the simplest mechanical model for
assessing the stress state in the panel is represented by the graph in figure 3 below.
Figure 3 – The load-bearing zone
In particular, it is assumed that the reactive elastic zone is made up of the

parallelogram E-B-F-D, whose dimensions for given values of w and h are
determined solely by angle α. Within the zone, the state of stress is constituted by a
uniform compression σ, with resultant C (α) =σ t d (α) =σ t (w cosα - h sinα). The
triangular regions A-E-D and B-C-F are instead held to be stress-free. The lower
the material's tensile strength, the closer such a model will conform to reality. In
order to determine α, we assume a value such that the lateral stiffness of the
reactive zone is at a maximum:
(2)
Thus, angle α coincides with the maximum point of the function

, with o≤α≤ά, where =arctan(w/h).
Figure 4 shows how α(t) is monotone decreasing between values 45° and 0°.
Figure 5 instead plots the stiffness of the wall elements in phases I and II, from
which it can be seen how the loss of stiffness is always significant, as long as η<2.
In deducing the interaction curves for the panel, we consider only cases in which
failure occurs either by crushing of the weaker masonry components, or by
slipping. Denoting by N=Ccosα and T=Csinα, the vertical and horizontal
components of C, respectively, and by ν and u, the corresponding displacements

(Figure 2), the interaction curves traced in the N-T plane are deduced by recourse
to the static theorem of limit analysis. In doing so, we first deal with elements for
which N can be considered a constant, and then an element for which v is zero and
N is unknown.
Figure 4 – Angle α versus ratio h/w Figure 5 – Panel stiffness
Figure 6 – Masonry crushing interaction curve for fm=50 tf/m2
Figure 8 – The limit strength Figure 9 – The limit strength

Figure 7 – Interaction curves
domain for generic masonry domain for maximum-stiffness
for slipping
panels panels
Figure 10 – Interaction curve N-M for fm=1 tf/m2 Figure 11– Interaction curve N-M for fm=10 tf/m2
Figure 12– Interaction curve N-M for fm=20 tf/m2 Figure 13– Interaction curve N-M for fm=30 tf/m2
Figure 20– Interaction curve N-M for fm=100 tf/m2
One type of retrofitting solution using RC jackets on both sides of the piers and
coupling beams was accepted.
Of course there are also other theoretical retrofitting solutions (which will be study
in the following stage) like: steel systems, base seismic isolators (almost all the
time is not enough to use these kinds of dampers without to execute some
retrofitting operation also at the superstructure), FRP systems and RC frame system
implants.
For the existing structure, in each type of case the gravitational and seismic safety
were verified:
• Gravitational safety
The cross sections for the masonry piers were determined in concordance with the
state of the art and state of the practice for that period, like a “correct” designer. In
this situation, for all the study cases the masonry piers offer gravitational safety and
from this point of view there is not necessary to take supplementary measures.
• Brittle failure mechanisms: t≥0.5 with t=τo/Rt and τo =Q/Ashear
For all the study cases we looked always both at the strength capacity and the
stiffness of the building structures.
From the stiffness point of view starting with the proper period of vibration (figure
21) and ending with the drifts (figure 22 a, b and c) we observe that the lateral
stiffness of the existing structures is very rich. As a consequence the drifts are
normal or in the normal limits.
The masonry structures, because of the structural walls offer always a medium to
large lateral stiffness for the horizontal actions, but not also enough strength. So for
these type of structures usually all the problems are coming from the strength
capacities.
The solution of RC jacketing of all the masonry structural elements, without to take
any other measures seems to be adequate or reasonable for low and medium rise
buildings (1-6 levels) only, because as we know the masonry structures are
conceptually stiff structures.
In figure 21 are presented the fundamental period of vibrations for the existing
structures and for the retrofitting case studies.
Figure 21 - Fundamental period of vibrations
The ductility demands highlight that for PGA less than 0.25g there will be not any
plastic hinges in the masonry piers. Increasing the number of the levels, plastic
hinges will occur in the masonry coupling beams.
This was exactly the reason to choose 3 series of models. Depending of the ratio
between the height and the length of these masonry coupling beams, the occurrence
of plastic hinges (which mean practically severe”X” cracks) may be at both ends or
in the middle spans of these coupling beams. In this situation the immediate
behaviour of the initial uncraked structure will be like the considered models with
plastic hinges at the edges or in the middle span.
For the retrofitted model we considered a RC jacketing at one or both faces of the
masonry piers and coupling beams and to rehabilitate the initial model with
functional coupling between the masonry piers.
In the figures 27-28-29 (a, b and c) there are presented base seismic shear
coefficient and both CBNL/CBL and Safety indexes (NL=non linear, L=linear and
B=base) for all the study cases, for all three seismic intensity values.
The CBNL/CBL increase with the seismic intensity increasing and always the lowest
value is for the middle span plastic hinges models in time when the higher values
are touch by the double hinged coupling beams models.
Similar commentaries concerning the safety index which have the values less than
0.500 always for the existing situations.
And generally speaking the best behaviour is of the RC jacketing structural walls
retrofitting solution.
Drifts (‰) Deflections (cm)
Figure 22a – Drifts and deflections for PGA=0.16g

Figure 22b – Drifts and deflections for PGA=0.20g
By applying the shear-type model, determination of the structural response of a

building to seismic loads can be performed via sub-structuring techniques, by
which the main structure is comprised of floors and the walls are secondary. In
such a framework, in order to account for possible openings in the walls, each wall
can be thought of as an assemblage of smaller rectangular elements, each of which
satisfies the constitutive equation illustrated in Figure 1 and the limit strength
conditions deduced in the foregoing. Finally, with regard to the transition between
the two phases, although the cessation of uncracked stage should more properly be
evaluated in the context of fracture mechanics theory (in order to better account for
failure of the brick-mortar interface), here for the sake of simplicity, considering
the extremely low levels of stress, it was assumed that cracking starts when the
maximum tensile stress component reaches the material's flexural strength.
Under the hypothesis that the loads increase proportionally to a single multiplier λ,
the structural response can be obtained by proceeding incrementally. In each step,
an iterative procedure must be used in order to assure that the stiffness considered
for each resisting element corresponds to its actual state. The resulting algorithm
can be easily implemented.
Figure 22b – Drifts and deflections for PGA=0.20g
Figure 23a - Ductility demands for masonry piers Figure 23b - Ductility demands for masonry
– PGA=0.16g coupling beams – PGA=0.16g
Figure 26 – Velocities (m/s)

Figure 27a– CB for PGA=0.16g Figure 27b– CB for PGA=0.20g Figure 27c– CB for PGA=0.25g
Figure 28a – CBNL/ CNL index Figure 28b – CBNL/ CNL index Figure 28c – CBNL/ CNL index
for PGA=0.16g for PGA=0.20g for PGA=0.25g
Figure 29a – Safety index for Figure 29b – Safety index for Figure 29c – Safety index for
PGA=0.16g PGA=0.20g PGA=0.25g
4. CONCLUSIONS AND RECOMMENDATIONS

This ample study is just at the very beginning, but even after the things presented
above, we shall say that the buildings with masonry structures are stiff structures.
If normally in a simplified manner we appreciate the fundamental period of
vibration for a frame structure at T= (0.03-0.1) n, for these existing structures we
found an average period of T=0.081n+0.069 (where n is the storeys number). For
the retrofitted masonry structures buildings the fundamental period of vibrations is
about T=0.033n+0.028 (sec.).
Usually the worse responses aspects became because of the mortar quality and not
from the solid bricks.
Several retrofitting solutions are viable to increase the strength capacities (because
usually the stiffness is more or less enough) but in these studies just one (extended
to two) was carried out.
When a designer must expertise and retrofit this type of structures, from the
beginning he will know the structural responses and expectations of them in
comparison with all the demands.
Having these ideas will be more comfortable and easy to have discussions with the
architects involved in the same project and of course with the owner of the
building, in the idea to choose the better way to put in safe the existing building
structure. Sometime other problems may occur such are the soil characteristics, the
neighbourhood buildings.
Acknowledgements
We dedicate this article to the memory of the wonderful professor which
Dan Ghiocel was.
Its content is a partial result of the second stage of MTKD-CT-2004-509775 Marie
Curie grant from the REPROCITY program in the Silesian University of
Technology – Gliwice - Poland
References
1. Th. Pauley, H. Bachman, K. Moser – The Aseismic Design for R/C – Technical Editor,
Bucharest, 1997;
2. A. Pretorian – Civil Engineering Aseismic Design –Matrix Rom, Editor, Bucharest, 2000;
3. P100-92 – Romanian Aseismic Design Code
4. P85-96 – Structural R/C wall Design Code
5. NP007-97 – Structural R/C Fames Design Code
6. P13 – 63 s and P13 - 70 – Romanian Aseismic Design Codes
7. ETABS2000 – User Guide
8. ANELISE – User Guide
9. Daniel Stoica – Contributions To The Improvement of Constructive Solutions For Tall Buildings
In Seismic Zones – PhD thesis;
10. Dabija, Fl. – Structuri de rezistenta la cladiri civile – Institutul de Constructii Bucuresti – 1980
11. Pretorian, A., Titaru, E., Stoica, D. – Expertize tehnice conform normativ P100-92 pentru
imobile din Bucuresti
12. Pretorian, A., Titaru, E., Stoica, D. – Aspecte tehnice si economice cu privire la punerea in
siguranta a cladirilor civile existente – Conferinta Nationala de Inginerie Seismica – septembrie
1997 - Bucuresti
13. S. Majewski – Elasto – Plastic model for concrete – Archives of Civil Engineering, L, 1, 2004 –
Gliwice-Poland
14. Plumier,A., C. Doneux, V. Caporaletti, F. Ferrario, D. Stoica – Guide Technique Parasismique
Belge Pour Maisons Individuelles – L’Universite de Liege – Belgium
15. D. Stoica, A. Plumier – Seismic behaviour of non-engineered masonry structures / First part –
Synthesis on numerical modelling of masonry walls behaviour / First draft – March 2003 –
University of Liege – Belgium
16. Plumier A., A., V. Warnote, D. Stoica – A simple method in order to take into consideration the
additional eccentricity for general torsion – University of Liege Review – June 2003Plumier, D.
Stoica – About connection forces between slabs and structural walls during the horizontal and
vertical seismic action - University of Liege Review – October 2003
17. HENDRY, A. W.: "Reinforced & Prestressed Masonry", Longman, Harlow, 1991.
18. PAULAY, T. M. and PRIESTLEY, J. N.: "Seismic Design of Reinforced Concrete and Masonry
Buildings", John Wiley & Sons, New York, 1992.
19. FAVDLLI, A., LIGARO', S.:Un modello semplificato per Panalisi lineare di edifici in muratura
soggetti al sisma, Atti XII Congr. AIMETA '95, Napoli, 1995, pp. 223-228
20. MENGI, A. et al.: "A Model for Nonlinear Earthquake Analysis of Unreinforced Brick Masonry
Buildings", Computers & Structures, 41, 1991, pp. 801-812.
21. MODENA, C. et al: "Esame comparative della normativa europea e della normative italiana
sulle strutture in muratura portante", ANDIL, Ed. Lambda, Padova, 1992.
22. R. Barsotti, S. Ligaro – Non-linear analysis of URM buildings under horizontal loads -
State of the art in the pounding mitigation techniques

V.Warnotte 1, D.Stoica 2, S. Majewski3 and M. Voiculescu2
1
M&S Department, University of Liege, Liege, Belgium
2
3
Structural Engineering, Silesian University of Technology, Gliwice, Poland
Summary
Pounding introduces impact loads that have to be superimposed on those caused
by the ground acceleration itself. When these impact loads from pounding are too
high, the structural system has to be modified to reduce the response.
Several methods have been proposed to avoid pounding induced collapse of
buildings. The methods may be classified according to their approach to the
problem of pounding: methods to avoid pounding, methods to strengthen structures
to withstand pounding, and techniques to reduce pounding effects in the structures.
KEYWORDS: pounding, seismic gap, damping, energy dissipation, impact
1. POUNDING MITIGATION TECHNIQUES

Pounding introduces impact loads that have to be superimposed on those caused by
the ground acceleration itself. When these impact loads from pounding are too
high, the structural system has to be modified to reduce the response.
Several methods have been proposed to avoid pounding induced collapse of
buildings. The methods may be classified according to their approach to the
problem of pounding: methods to avoid pounding, methods to strengthen structures
to withstand pounding, and techniques to reduce pounding effects in the structures.
1.1. Seismic gap

The first work for pounding prevention was to establish a good and reliable
estimate of the minimum gap required for the design earthquake so that pounding
between the structures will not occur.
Providing a sufficient gap has been the commonly accepted strategy adopted by
building codes throughout the world. The value of the separation distance between
two structures which is sufficiently large to prevent pounding is known as the
seismic gap.
682 V.Warnotte, D.Stoica, S. Majewski, M. Voiculescu
1.1.1 Building Code Requirements

Building codes in zones of active seismicity around the world have recognized the
destructive effects that pounding may induce in constructions. The approach
commonly adopted in building codes has been to avoid contact interactions
between the structures by providing sufficient separation between them.
The criterion has been defined by Valles (1997) using four different expressions:
gap ≥ factor (sum of individual displacements D1 and D2 of buildings ) (1)
gap ≥ coefficient (height) (2)
gap ≥ fixed distance (3)
gap ≥ SRSS (of the displacements D1 and D2 ) (4)
where D1 and D2 are the maximum displacement of the individual buildings
(Figure 1).
Figure 1 – Definition of the notations
The first criterion may be considered as equivalent to the absolute sum of

maximum, multiplied by an amplification factor. The amplification factor in most
cases comes from the increase in displacements due to the inelastic response of the
structures. This criterion does not take into account that the maximum
displacements in the structures, in general, will not occur at the same time.
The second condition may be easily justified, since; in general, building codes
specify a maximum inelastic drift related to structures (story height). Using this
approach, the dynamic characteristics of the structures are not relevant to the gap
computation, since the lateral deformations are always checked. This approach is
the easiest. Nevertheless, this form of specifying the gap, by not considering the
dynamic properties of the adjacent structures, may be overly conservative for
buildings that tend to respond in phase; for example, buildings with high
percentages of critical damping, or inelastic buildings for some ratios of
frequencies to the characteristic earthquake frequencies.
The third one is specified for construction considerations, that is, to allow for
adequate space to place the formwork for beams and columns, to build masonry
walls, or place elements of the facade.
The last one takes into account the fact that the maximum displacements in the
structures will not occur at the same times. It uses the SRSS (Square Root of the
Sum of the Squares) modal combination rule that is assuming that the input motion
is stationary, and the response of each structure is uncorrelated with the others. It
refers to the theory of probability which indicates that the SRSS combination
provides the most probable value of the maximum of [D1 (t) + D2 (t)].
Therefore, it yields conservative results when the response of the structures is
somewhat too perfectly correlated.
First example of code requirements. In the Eurocode 8, the specification against
pounding phenomenon is:
Building pounding is a phenomenon that occurs when adjacent structures are
separated at distances less than the differential lateral displacements that occurs in
each structure as a result of their earthquake response. Buildings shall be protected
from earthquake-induced pounding from adjacent structures or between structurally
independent units of the same building. This principle is considered to be satisfied
if:
1. For buildings, or structurally independent units, that do not belong to the same
property, if the distance from the property line to the potential of impact is not less
than the maximum horizontal displacement Ds of the building at the corresponding
level (Recall: displacements Ds in Eurocode 8 are q times the elastic displacements
De computer under the design earthquake, the latter being on earthquake "reduced"
by the behaviour factor q: Ds=q.De);
2. For buildings, or structurally independent units, belonging to the same property,
if the distance between them is not less than the square root of the sum of the
squares (SRSS) of the two maximum horizontal displacement of the building or
units at the corresponding level.
A specification is however made: if the floor elevations of the building or
independent unit under design are the same as those of the adjacent building or
unit, the above referred minimum distance may be reduced by a factor of 0,7.
Second example of code requirements. The Uniform Building Code (UBC)
specifies that the separation distance between two buildings shall be at least
(0.375xRw) times the displacement due to seismic forces, where Rw is the US
equivalent to the behaviour factor q. Pantelides and Ma (1998) have found that the
UBC code values for the seismic separation distance to avoid pounding seems to be
conservative.
Anagnostopoulos and Spiliopoulos in 1992 obtained that the Eurocode requires

more conservative gap sizes than the Uniform Building Code (UBC) one.
Only one building code was found that allowed some level of pounding if the
effects do not jeopardize the integrity of either construction (Valles and Reinhorn,
1997). However, no procedure was outlined to indicate how the collision effects
are to be calculated.
1.1.2 Double Difference Combination rule

Series of dynamic time history analyses can estimate the seismic gap by obtaining
the maximum relative distance of the adjacent structures. However, an alternative,
called the "spectral difference (SPD) method," based on a spectrum approach has
been proposed by various authors. The method accounts for phasing associated
with vibration of adjacent structures, defined as "vibration phase." They also
propose simplified rules to predict the inelastic vibration phase, and demonstrate
the accuracy of the SPD method for a variety of adjacent building pairs and
earthquakes. Unlike the time history analysis method, the SPD method clarifies the
effects of various parameters on the relative displacement through a closed-form
solution. The SPD method is also useful for various relative displacement problems
between adjacent structures, such as bridges. It is more accurate than the current
code calculation method that ignores the vibration phase.
An estimate of the required separation, based on random vibration considerations,
has been proposed by Jeng, Kasai and Maison (1996). They proposed a spectral
difference method to calculate the minimum gap to avoid pounding for linear
structures. The method was named Double Difference Combination rule (DDC):
(5)
Where uA, uB = design peak displacements of buildings A and B.

The correlation coefficient (ρAB) is calculated according to the simplified formulas
for white noise input:
(6)
ξ = damping ratio and TB/TA = ratio of the fundamental periods of the two
buildings.
Good results are obtained, at least for elastic systems.
Garcia (2004) evaluates four different methods to calculate critical separation

distances. All methods use Double Difference combination rule, but follow
different approaches to calculate the correlation coefficient ρ. Then he proposes a
new one which consists of using values of parameter ρ derived from empirical
estimates obtained through numerical simulations.
1.2 Increasing the stiffness of one or both buildings

Since the gap between two buildings usually cannot be increased, increasing the
stiffness of one or both buildings may reduce the seismic deformations to the point
where impact is precluded with the existing gap. Increasing the stiffness of the
building, reduce the period of this one and lead to a decrease of its displacement.
1.3 Supplemental energy dissipation

Another method to avoid pounding is the use of supplemental energy dissipation
devices in the buildings.
Using supplemental energy dissipation devices reduces the maximum lateral
deflections of the building. Even if the reduction in the maximum energy levels
provided may not be sufficient to avoid pounding, the amplification effects of
impacts in the structures will be smaller.
1.4 Alternative load paths

If increasing the stiffness is not feasible (ex. stiff shear wall buildings) FEMA-172
proposes to provide alternative load paths for the vertical load-resisting members
that may be damaged or destroyed by the impact. These alternative load paths
would include supplementary columns or vertical shoring to support the floor or
roof systems. These supplementary supports would be installed at sufficient
distance from the vulnerable exterior walls or columns to be protected when the
existing elements are damaged.
1.5 Impact absorbing materials

A measure for reducing the effects of pounding, while maintaining small separation
distances, would be to fill the gap with a special, shock absorbing material (bumper
dampers). Bumper damper elements must be dissipative to reduce pounding
accelerations and forces during contact. Under bumper damper elements, all energy
dissipation devices available that can be placed between the structures, but
connected only to one of them, are considered. Bumper dampers are therefore
energy dissipation links that are activated when the gap is closed. The presence of
the bumper damper element will reduce the impulsive forces transmitted from one
structure to the other.
If the element provides only stiffness the impulse loads will still be reduced since
the impacting bodies will encounter a spring element reducing the kinetic energies
of the structure before the stroke of the element is reached, at which point the full
pounding of the masses will take place, but the impacting velocities will be
smaller. Although the velocities at the onset of pounding are smaller, the linear
spring will increase the velocities after pounding, but the high frequency
accelerations observed will be reduced.
Anagnostopoulos (1988) showed that a soft viscoelastic material filling gap
between two adjacent structures can reduce the effects of pounding significantly.
This solution, however, did not reduce the building response below the values
reached without pounding. Another inconvenient of this solution is that it does not
solve the problem for the last building in a row. This building will behave like a
pendulum.
Wolf and Skrikerud (1980) also studied a tuning device which completely occupies
the existing gap size. It is assumed that this tuning device consists of a visco-elastic
material which is modelled as an elastic spring with a stiffness coefficient kt and a
viscous damper (working in parallel) with a coefficient ct. They found that
introducing a tuning device between the two involved structures reduces the
maximum impact force and thus local response somewhat.
1.6 Connecting buildings together

As an alternative to seismic separation, permanent connections of the adjacent
buildings have been investigated. A permanent linkage would provide a continuous
force to the structures which is proportional to the stiffnesses and thus more in-line
with the dynamic behaviour of the unlinked frames.
Linking adjacent buildings has a number of disadvantages, including possible high
forces in the link, the fact that the dynamic characteristics and the design failure
mechanisms are changed, and the uncertainties inherent when the two structures of
different characteristics must become one.
Nevertheless if those problems are solved, linking two structures will reduce the
possibility or the influence of pounding interactions (Plumier et al. (2005)). If an
energy dissipation device is used between the two structures, pounding will occur
if the stroke of the element is not sufficient.
2. CLASSIFICATION OF DEVICES USED TO CONTROL THE

SEISMIC MOVES OF STRUCTURES
2.1. Introduction
A great number of protective systems for structures built on seismic areas have
been developed since the years 1970's. These developments have been made to
mitigate action effect in individual structure, without special consideration on
pounding problem.
After a brief summary of these energy dissipaters, the next chapter will be on
studies on pounding mitigation using these devices.
Modern structural protective systems can be divided into three groups as shown in
Table 1.
Table 1 - Modern Structural Protective Systems

Semi-active and Active
Seismic Isolation Passive Energy Dissipation
Control
Elastomeric Metallic Dampers Active Bracing Systems
Friction Dampers Active Mass Dampers
Variable Stiffness or
Lead Rubber Bearings Viscoelastic Dampers
Damping Systems
Viscous Fluid Dampers Smart Materials
Sliding Friction Pendulum Tuned Mass Dampers
Tuned Liquid Dampers
A seismic isolation system is typically placed at the foundation of a structure. By
means of its flexibility and energy absorption capability, the isolation system
partially reflects and partially absorbs some of the earthquake input energy before
this energy can be transmitted to the structure. As it cannot be used either to reduce
the individual displacement of one structure or to connect two neighbouring
structures, the seismic isolation system will not be presented in this section.
Passive energy dissipation devices for structural applications are similar to
seismic isolation technology. Their basic function is to absorb or consume a
portion of the input energy, thereby reducing energy dissipation demand on
primary structural members and minimizing possible structural damage. Contrary
to semi-active or active systems, there is no need for an external supply of power.
Semi-active and active structural control is an area of structural protection in
which the motion of a structure is controlled or modified by means of the action of
a control system through some external energy supply. However, semi-active
systems require only nominal amounts of energy to adjust their mechanical
properties, and unlike fully active systems, they cannot add energy to the structure.
2.2. Passive control devices

Passive control techniques are based on the artificial increase of the dissipation
capacity, obtained by means of the insertion, in proper positions, of special devices
of the which both the stiffness and strength have to be defined in order to achieve:
- a limitation of the relative move of buildings one toward the other;
- energy dissipation.
Recentering of the system after the earthquake is another identified goal, for
practical use of the structure after the earthquake.
Devices with force-displacement response characteristics that are primarily a
function of the displacement amplitude are classified as rate (in fact strain rate)
independent devices. The behaviour of these devices is generally independent of
the relative velocity or the frequency of motion. They include friction and metallic
devices.
2.2.1 Metallic dampers

In recent years, a variety of mechanical devices that incorporate the yielding
deformation of mild steel to provide supplemental damping have been
implemented in earthquake-resistant designs of buildings and other structures.
Some particularly desirable features of these devices are their stable hysteretic
behaviour, low-cycle fatigue property, long term reliability, and relative
insensitivity to environmental temperature.
They typically exhibit hysteretic force-displacement behaviour, which can be
approximated as bilinear or trilinear. These devices tend to be inexpensive to
produce and their properties will remain stable over the long lives of buildings.
Unfortunately, they often have a limited number of working cycles, which may
require them to be replaced after large seismic events.
The yield strength of the connector is difficult to decide because if the yield
strength is too high, the connector may not function properly but if the yield
strength is too low, the energy absorbing capacity may be too small during a strong
earthquake. However, some possibility to tune the device exist by the fact that
stiffness, which is related to section, and yield strength, which is mater of section
and yield stress, can be decided independently, to some extend.
Several of the devices considered included torsional beam, flexural beam, and U-
strip dampers are shown schematically in Figure 2.
Many new designs have been proposed, including the X-shaped and triangular
plate dampers displayed in Figure 3. With this shape yielding is spread almost
uniformly throughout the material.
Figure 2 - a) Torsional Beam, b) Flexural Beam, c) U-strip (Soong and Dargush 1997)
Figure 3 - X-plate Metallic damper (Soong and Dargush 1997)

An inconvenient is that they are designed to deform so much when building
vibrates during an earthquake that they cannot return to their original shape and
have permanent deformation.
2.2.2 Friction dampers

Friction dampers utilize the mechanism of solid friction to provide the desired
energy dissipation. The friction develops between two solid bodies sliding relative
to one another. When the parts slide over each other, they create friction, which
uses some of the energy from the earthquake that goes into the building. Friction
between dry surfaces produces a constant force independent of velocity, always
opposed to the direction of motion that is proportional to the contact forces
between the sliding surfaces and the coefficient of friction of the materials. The
behaviour of the devices are nearly unaffected by amplitude, frequency,
temperature, or the number of applied loading cycles.
Friction devices generally exhibit rigid-plastic behaviour so the force-displacement
curves of the devices are rectangular loops as shown in Figure 4. These devices can
be characterized by their displacement amplitude and slip-load. The friction force
in the damper can be adjusted through appropriate torque of the bolts that control
the pressure on the friction surfaces.
Figure 4 Figure 5 Figure 6

Force displacement relation Force-displacement for a Force-displacement for a
for friction device viscoelastic damper fluid viscous damper
Practical engineer must pay attention that frictional devices must be loaded beyond
the slip threshold, and thus are likely to be ineffective under small to moderate
shaking.
Several types of friction dampers have been developed for the purpose of
improving seismic response of structures. They are illustrated in Figure 7.
Figure 7 - a) Limited Slip Bolt Joint, b) X-braced Friction Damper c) Sumitomo Friction
Damper d) Energy Dissipation Restraint e) Slotted Bolted Connection (Soong and Dargush
1997)
In recent years, there have been a number of structural applications of friction
dampers aimed at providing enhanced seismic protection of new and retrofitted
structures. For example, the applications of friction dampers to the McConnel
Library of the Concordia is shown in Figure 8.
Figure 8 - Pall Friction Damper installed in the Webster Library of Concordia University in
Montreal, Canada.
One important thing is to ensure that the contact forces between the sliding
surfaces and the coefficient of friction do not change with the long-term periods.
Another inconvenient is that friction dampers restrict a structure from restoring
itself to its original position after seismic events. They need to be supplemented by
a restoring force mechanism.
2.2.3 Viscoelastic dampers

Viscoelastic materials used in structural application are typically copolymers or
glassy substances which dissipate energy when subjected to shear deformation. A
typical viscoelastic damper is shown in Figure 9 which consists of viscoelastic
layers bonded with steel plates. When mounted in a structure, shear deformation
and hence energy dissipation take place when the structural vibration induces
relative motion between the outer steel flanges and the centre plate.
They provide extra damping, alleviate mechanical vibration, and thus improve
overall system dynamic responses.
Figure 9 - Viscoelastic Damper

Viscoelastic devices have an output that is somewhere between that of a damper
and a spring. Under high-level seismic inputs, the spring response dominates,
producing a response that increases stresses of adjacent columns at any given
deflection. Viscoelastic dampers are quite linear in their response and are able to
dissipate energy under low levels of shaking.
The designer has to take into account the effect of ambient temperature and
excitation frequency for an effective design of viscoelastic dampers in structural
applications. These devices can be modelled with Maxwell model, which consists
of a spring and dashpot in series. The force-displacement curves of the devices are
shown in Figure 5.
2.2.4 Fluid viscous dampers

Viscous damping involves taking advantage of the high flow resistance of viscous
fluids. The forces developed in a viscous damper are proportional to the velocity of
its deformation. Fluid viscous dampers put out virtually zero force at the low
velocities associated with thermal motion. The force-displacement curves of the
devices are shown in Figure 6.
Viscous fluid dampers are similar to shock absorbers in a car. They consist of a
closed cylinder containing a viscous fluid like oil. A piston rod is connected to a
piston head with small holes in it. First, the piston moves in the cylinder, then the
oil is forced to flow through holes in the piston head, causing friction. When the
damper is installed in a building, the friction converts some of the earthquake
energy going into the moving building into heat energy. The force depends on the
size and shape of the orifices and the viscosity of oil.
Figure 10 - Example of fluid viscous damper
The restoring force in an elastic structure damped by viscous dampers is

proportional to the velocity at which the material is strained, so that energy is
dissipated via viscous damping.
(7)
where C = damping constant, V = velocity and α = velocity exponent
with (0,3 ≤ α ≤ 1,0).
When α is equal to unity, the damper is a linear viscous damper. The fluid damper
commercially available consists of a cylinder and a stainless-steel piston with a
bronze orifice head and accumulator. The orifice utilizes a series of specially
shaped passages to alter flow characteristics with fluid speed.
Fluid viscous damping reduces stress and deflection because the force from the
dampers is completely out of phase with stresses due to flexing of the columns.
This is only true with fluid viscous damping, where damping force varies with
stroking velocities. Consider a building shaking laterally back and forth during a
seismic event. Column stress is at a maximum when the building has flexed a
maximum amount from its normal position. This is also the point at which the
flexed columns reverse direction to move back in the opposite direction. If we add
a Fluid Viscous Damper to the building, damping force will drop to zero at this
point of maximum deflection. This is because the damper stroking velocity goes to
zero as the columns reverse direction. As the building flexes back in the opposite
direction, maximum damper force occurs at maximum velocity, which occurs when
the column flexes through its normal, upright position. This is also the point where
column stresses are at a minimum. It is this out of phase response that is the most
desirable design aspect of fluid viscous damping.
Fluid inertial dampers have several inherent and significant advantages: linear
viscous behaviour, insensitivity to stroke and output force; easy installation; almost
free maintenance; reliability and longevity. Fluid viscous dampers allow the
structure to re-center itself perfectly at all times.
2.2.5 Tuned mass dampers

The objectives of incorporating a tuned mass damper into a structure is basically
the same as those associated with metallic dampers and other energy dissipation
devices, namely to reduce energy dissipation demand on the primary structural
members under the action of external forces.
This reduction, in this case, is accomplished by transferring some of the structural
vibrational energy to the tuned mass damper which, in its simplest form, consists of
an auxiliary mass-springdashpot system anchored or attached to the main structure.
In tuned mass dampers, typically a solid concrete or metal block acts as the
secondary mass. Additional spring and dampers are used to attach this secondary
mass to the primary structure, and to provide the restoring and dissipate
mechanisms needed to tune the system for near-optimal response under various
types of dynamic excitations.
It is noted that a passive Tune Mass Dampers (TMD) can only be tuned to a single
structural frequency. While the first-mode response of a MDOF structure with
TMD can be substantially reduced, the higher mode response may in fact increase
as the number of stories increases. For earthquake-type excitations, the response
reduction is large for resonant ground motions and diminishes as the dominant
frequency of the ground motions gets further away from the structure's natural
frequency to which the TMD is tuned.
2.2.6 Tuned liquid dampers

Tuned Liquid Dampers work according to the same principle as Tuned Mass
Dampers; a tank filled with water replaces the mass and the sloshing of the liquid
creates the mass-springdashpot system, due to grids with holes that slow down the
liquid moves.
2.3 Active control devices

The term "active" is used to indicate that the operation of these systems requires a
significant amount of external power. Active control devices utilize the feed back
from sensors measuring the response of a structure to control the behaviour of
structural elements through mechanical actuators. Records from the sensors are
then fed into a controller (computer) that activates devices for modifying the
structure's response continuously during its excitation. There are several different
types of active control devices presently in use, some of which are as follows:
active mass damper, active base isolator and active tendons.
Active tendons are devices wherein tension in the prestressed tendons is varied
during the earthquake excitation in a way to reduce the structure's response.
The active mass damper is a combination of a passive Tuned Mass Damper and an
active control actuator. The ability of this device relies mainly on the natural
motion of the TMD. The forces from the control actuator are employed to increase
the efficiency of the device.
One of the problem with this devices is that since active control systems depend on
power supply, it has to be ensured that this supply will not be interrupted during a
strong earthquake, otherwise the whole system will remain idle exactly at the time
that it will required to function.
2.4 Semi-active connectors

Nestled between passive and active structural control technologies is an emerging
area of research addressing the possible use of innovative or "smart" materials for
sensing and control purposes. This class of smart materials can be incorporated into
structural elements, capable of modifying structural behaviour in response to
external stimuli. Semi-active control systems have been developed to take
advantage of the best features of both passive and active control systems.
The term "semi-active" is used to indicate that the operation of these systems
requires a very small amount of external power. As in an active control system, the
mechanical properties are typically adjusted based on feedback from the structural
system to which they are attached. As in a passive control system, semi-active
control systems utilize the motion of the structure to develop control forces. The
control forces are developed through appropriate adjustment of damping or
stiffness characteristics of the semi-active control system. Semi-active are fail-safe

because they can be designed to exhibit either prescribed damping or prescribed
stiffness characteristics in the event of a complete loss of power.
The most predominant materials that have been examined as actuation devices in
recent years are shape memory alloys, piezoelectic elements, electrorheological
fluids and, more recently, magnetoreheological fluids.
One means of achieving a semi-active damping device is to use a controllable,
electromechanical, variable-orifice valve to alter the resistance to flow of a
conventional hydraulic fluid damper.
Another class of semi-active devices uses controllable fluids. For example, the
magnetoreheological fluids are based on special fluids which are able to change
their viscosity from liquid to semi-solid state within milliseconds, under the action
of an applied magnetic field, so to produce, according to a preset control strategy
the desired control forces.
Acknowledgements
The content of this article is a partial result of the first stage of MTKD-CT-2004-
509775 Marie Curie grant from the REPROCITY program in the Silesian
University of Technology – Gliwice – Poland.
References
1. Anagnostopoulos, S.A. (1988), "Pounding of buildings in series during earthquakes" ASCE
Journal of Structural Engineering and Structural Dynamics, Vol. 16, pp. 443-456.
2. Anagnostopoulos, S.A., Spiliopoulos, K.V. (1992), "An investigation of earthquake induced
pounding between adjacent buildings", Earthquake Engineering and Structural dynamics, Vol.
21, pp. 289-302.
3. Beskos, D. E., Anagnostopoulos, S. A. (1997), "Computer Analysis and Design of Earthquake
Resistant Structures, a Handbook", Computational Mechanics Publications, Southampton.
4. Chau, K.T., Wei, X.X. (2001), "Pounding of structures modelled as non-linear impacts of two
oscillators" Earthquake Engineering and Structural dynamics, Vol. 30, pp. 633-651.
5. Eurocode 8 (2003), "Design of structures for earthquake resistance", prEN 1998-1.
6. FEMA-172 (1992), "Techniques for the Seismic Rehabilitation of Existing Buildings", Building
seismic safety council, Washington D.C.
7. GOSCA (2001), Denoël, V., "Calcul sismique des ouvrages d'art" Travail de fin d'études de
l'Université de Liège.
8. Jankowski ,R. (2004), "Non-linear viscoelastic model of structural pounding" 13th World
conference on Earthquake Engineering, Paper No. 3082.
9. Maison, B.F., Kasai, K. (1992), "Dynamics of pounding when two buildings collide" Earthquake
Engineering and Structural dynamics, Vol. 21, pp. 771-786.
10. NEHRP, October 1997 "Guidelines for the seismic rehabilitation of buildings —Provisions:
FEMA 273 — Commentary FEMA 274", Federal Emergency Management Agency,
Washington, DC.
11. Pantelides, C.P., Ma, X. (1998), "Linear and Non-linear pounding of structural systems"
Computers and Structures Vol. 66 (1) pp. 79-92.
12. Papadrakakis, M., Mouzakis, H., Plevris, N., Bitzarakis, S. (1991), "A Lagrange multiplier
solution method for pounding of buildings during earthquakes" Earthquake Engineering and
Structural dynamics, Vol. 20, pp. 981-998.
13. Plumier, A., Warnotte, V., Van Alboom, W (2005), "A case study of flexible reconnection over
expansion joints in a buildings" 9th World Seminar on Seismic Isolation, Energy Dissipation and
Active Vibration Control of Structures.
14. Rosenblueth, M., Meli, R. (1986), "The 1985 earthquakes: causes and effects in Mexico City",
Concrete International 8, American concrete institute, pp. 23-24.
15. SAP2000 (2004), "Three dimensional static and dynamic finite element analysis and design of
structures, Version 9.0" Computers & Structures, inc, Structural and Earthquake Engineering
Software, Berkeley, California, USA.
16. Soong, T. T., and Dargush, G. F. (1997), "Passive Energy Dissipation Systems in Structural
Engineering", Wiley, Chichester.
17. UBC (1992) "Uniform building codes" International conference of building officials Whittier
CA.
18. Valles, R.E., and Reinhorn, A.M. (1997), "Evaluation, Prevention And Mitigation Of Pounding
Effects In Buildings Structures", National Center of Earthquake Engineering Research Technical
Report.
19. Wolf, J. P., Skrikerud, P. E. (1980), "Mutual pounding of adjacent structures during earthquakes"
Nuclear Engineering and Design, Vol. 57, pp. 253-275.
20. Warnotte V. Structural problems due to earthquake action at joints between building structures -
D.E.A. Graduation Work
21. Plumier,A., C. Doneux, V. Caporaletti, F. Ferrario, D. Stoica – Guide Tehnique Parasismique
Belge Pour Maisons Individuelles – L’Universite de Liege – Belgium
22. D. Stoica, A. Plumier – Seismic behavior of non-engineered masonry structures / First part –
Synthesis on numerical modelling of masonry walls behaviour / First draft – March 2003 –
University of Liege – Belgium
23. Plumier A., A., V. Warnote, D. Stoica – A simple method in order to take into consideration the
additional eccentricity for general torsion – University of Liege Review – June 2003Plumier, D.
Stoica – About connection forces between slabs and structural walls during the horizontal and
vertical seismic action - University of Liege Review – October 2003
The calculation of a cable network using the tangent stiffness

matrix
Dragos Stan Caciula and Nicolae Ungureanu
1
Structural Mechanics Department, Gh.Asachy University, Iassy, 700050, Romania
Summary
In our days, in many study fields the finite element method is a very useful tool in
the analyze of the structures. In this analyze all the structure’s elements such as
plates, beams and cables can be represented thorough the finite element method.
This article follows the methods and the equations for the finite nonlinear elements
of a cable network. Also will be calculated and analyzed the finite displacements,
non-linear equations, the virtual mechanical work for the total finite displacements
for the non-tensioned calculation state.
KEYWORDS: stiffens matrix, iterative methods, incremental method, Newton

Raphson method.
1. INTRODUCTION
There are many numerical methods to solve the nonlinear static problems. Thus we
can specify o series of methods such as: incremental method which uses the step by
step linearization. It is assumed that the fixed solution is close enough by exactly
solution of the nonlinear equation as long as the step is small enough. Newton
Raphson method allows determination thorough approximations of a convergent
solution. The stiffness matrix must be generated at every step. Modified Newton
Raphson method keeps the stiffness matrix in the domain of certain iteration steps.
2. SOLUTION FOR THE NONLINEAR ALGEBRAIC

EQUATIONS
We assume a set of N nonlinear equations with the number of variable N, yi, i = 1,

N:
gi(yj)=o , i, j=1, N (1)
698 D. Căciulă, N. Ungureanu
There are two approximately classes of solution for the nonlinear equations:
iterative and incremental. These are consisted by the application of an incremental
operator relation with an initial state assumes known. The load is applied with
successive step. The solutions obtained are stable as long as the matrix remains
positive defined but as a drawback is the fact that there are no errors control
methods.
Exact solution class has his advantages and his drawback one from each others.
The iterative process has an initial drawback and is: the initial solution 0yi is used
step by step for the obtaining the right solution yi.
2.1. Iterative methods
There are a couple of iterative methods for solving the equation (1). The easiest
method is the Gauss Seidel methods of the repeated substitutions. After all the
steps are followed, the solution (1) can be rewritten thus:
y j = g 'j ( y j ) (2)
and these successive approximation are based on:

p +1
yi = gi' ( y) p
i
(3)
Convergence is presumed when a suitable norm of the unknowns is within a
prescribed tolerance, the Euclidean norm is often used:
( p +1
yi − p yi )( p +1
)
yi − p yi ≤ ε (4)
The “total step” calculation phase involves the use of the values only on the
previously obtained values pyi for p+1yi values calculation. A scheme much more
efficient is the one step Gauss - Seidel methods use in which the most recently
value of the yi is used for a better solving of the p+1yi step:
p +1
yi = gi' ( p +1
y j , j < i; p y j , j ≥ i ) (5)
The work method’s simplicity is the major advantage of the Gauss Seidel method.
A much more stable method and much more used is the Newton Raphson method.
This method allows the determination, through iterations, of convergent solution.
The nonlinear algebraic equations are replaced with first order Taylor series and
iterations are made until is accomplished the convergence criteria. Stiffness matrix
must be generated at every iteration step and each load step. Between the real
solution yi and the most recently approximation pyi the error defining is:
∆ p yi = yi − p yi (6)
Regarding the Taylor series for equations we may write:

 ∂g 
0 = gi ( yi ) = gi ( p yi ) +  i  (∆ p yi ) (7)
 ∂yi 
 yi = p yi 
Also we can calculate ∆ p yi :
 ∂g 
 i  ∆ p y j = − gi ( p y j ) (8)
 ∂y j 
 yi = p yj 
which represent a classic set of linear equations which gives as a more exactly
estimation for the value of pyi. The classic method for linear algebraic solution can
be used for the calculation of the ∆pyi value. We may write the next definition:
p +1
yi = p yi + ∆ p yi (9)
Afterwards the approximation can be calculated from equations (6) and (7).
Convergence is defined by either a chosen norm of gi ( p +1
)
y ≤ ε or by chosen
norm ∆ p +1
y ≤ ε . An analogical dimension of this method is represented in figure
1.
g(0 y)
g(1 y)
dg
dy 0 y
dg
g(2y) dy 1 y
Solutia 3 2 1
exacta
0
y y y y y
0 1 2
y y y
Figure 1. Newton Raphson method
We can define the tangent stiffness matrix:

 ∂g 
p
J ij =  i  (10)
 ∂y j 
 yj = yj 
p
p
Calculation of matrix J ij values are necessarily to be done in each iteration and
for all the process needed for calculation is a long term process. In the case in
which the nonlinear calculation doesn’t need a high precision the stiffness matrix
can be let in his initial state. This method is called the modified Newton Raphson
method. And it is represented in figure 2.
g(0 y)
dg
0 dy y =0 y
g
1
g
Exact
0 1 2
y y y
Figure 2. Modified Newton Raphson methods
Rate convergence is just linear but is much more efficient from the point of view of
calculation because Jij doesn’t need to be recalculated. A non - regularity in the
case of modified Newton Raphson method is the change of tangent stiffness matrix
only in those stages of iteration process figure 3.
2.2. Incremental methods
Incremental schemes are extension of linear analysis. In a finite element model the
loads are made in small incremental steps and the structure is assumed to have a
linear response at every incremental load.
Rewriting equations (1) is obtained:
fi ( yi ) = Pi ( yi ) (11)
Where:
fi(yi) are interval forces;
Pi(yi) are external forces.

The fi and Pi are both yi nonlinear functions.
g g
0 y y =0 y y y =3 y
g
Exact
2 2
0
y
1
y y 3
y y
Figure 3. Modified Newton-Raphson methods, periodic updates of Jij
Let’s assume that we obtain the solution 0yi after q increments and we are looking
for a p∆y increment after a load increment q∆Pi:
q +1
yi = q yi = q ∆yi (12)
q +1
q
∆Pi = Pi ( q +1 y j ) − q Pi ( q y j ) (13)
Equation (11) has to be satisfied for every q+1yi. That’s why we can write:
fi ( q +1 y j ) = fi ( q y j + q ∆y j ) = Pi ( q +1 y j )
 ∂f  (14)
.............. = fi ( y j ) +  iq  q ∆yi = Pi ( q +1 y j )
 ∂y j 
 yj = q yj 
Regarding equation (11) and (14) we have:
 ∂f 
 i  q ∆y j = q ∆Pi (15)
 ∂y j 
 yj = q yj 
Equations (15) represent a linear equation set for q∆yi and again are obtained the
tangent stiffness matrix of fi gradient which was evaluated at the end of prior
increment qyi. If Pi(yj) is a non-preserving forces set it must be used the
approximation for q∆Pi use equation (13) is a q∆yi function. If q∆Pi is low enough,
equation (15) gives a good approximation for ∆yi.
Incremental method’s drawback is the errors tend to be cumulative, figure 4. An
improvement of this incremental method is so called auto - correction incremental
scheme. This method instead of using the increment for the q∆Pi (figure 4) as it is
defined by equation (13) it is replaced by qfi(qyj) from equation (11) and we obtain:
 ∂f 
 i  q ∆y j = q +1
Pi ( q +1 y j ) − q fi ( q y j ) (16)
 ∂y j 
 yj = q yj 
E ro a re
p
4
p
dy
dp 3p
3
p
dy
dp 2p
dy
dp 1 p
dy
dp 0p
y
0 1 2 3 4
y y y y y
Figure 4. Incremental scheme
These schemes allow a more precisely control an error but instead calculation
method is a method which needs a longer time for calculation.
2.2. Iterative incremental combined method
For obtaining results as correct as possible it is recommended the combined

iterative-incremental method’s use. Incremental method supplies correct initial
solutions which are used by iterative solution. Assume that for p used as an
exponent in equation is iteration and q used as index it is in fact increment.
Equation (1) can be written at the end of q + 1 increment thus:
q +1
gi ( q +1 y *j ) = 0 (17)
q +1 *
Where y is the real solution at the end of q + 1 increment.
j
Assuming that:
q +1 *
j y = q y j + 0 ∆y (18)
in which qyi is calculated solution after q increment and 0∆y is the error after p
iterations. Developing equation (1) with Taylor series it is obtained:
 ∂g 
q +1
gi ( q +1 y *j ) = q +1
gi ( q y j ) +  i  0 ∆y j = 0 (19)
 ∂y j 
 yj = q yj 
3
p
2
p
1
p
3
E xact y
y
0 1 3
y y y
3
E xact y
y
3 0
y
Figure 5. Combined incremental-iterative scheme
q +1
Thus the first iteration for y jp Newton Raphson scheme is:
q +1
yi0 = q yi + 0 ∆yi (20)
p + 1 iteration can be written:

p +1
q +1
yip +1 = q yi + ∑ p ∆yi (21)
p =0
In which p∆y is obtained by solving:
 ∂g 
 i  p ∆y j = − q +1 gi ( q +1 y Pj ) (22)
 ∂y j 
 yj = q+1 yj 
In figure 5 we may see the way the Newton Raphson works after three increments.
3. CABLE NETWORK ANALYSES THROUGH FINITE

ELEMENT METHOD
Finite element method is very easy to use for solving a special cable network. We
assume as finite element a cable hanged between two fixed points. Stiffness matrix
and load vector will be formed for every cable network’s element. Global stiffness
matrix of Kg structure will be created from stiffness matrix of every element and
the global load vector Fg is consisted by load vectors of each network’s element.
Forta
Kt 1 F2 F1
Supraincarcare
T1
T2
O u2 Deplasare
u1
Figure 6. Diagram of Newton Raphson’s method for solving nonlinear equations. 0 ,1, 2,
index of iteration. Kto, Kt1 – tangent stiffness matrices. T1, T2 – tensile forces in cable. F1, F2
– nonequilibrium nodal forces. ∆u1, ∆u2 – increments of displacement
Appling Lagrange variation principle, an algebraic equation system will be used

which use the shape conditions. Regarding the geometric nonlinearity effect, this
equation system it is in fact a nonlinear equations system. Using tangent stiffness
matrix and Newton Raphson method (figure 6), nonlinear equation system solution
is transformed into algebraic linear equation repeated solution.
Solving the equation system we find the total displacement ∆g vector from the
network nodes.
3.1. Calculation hypothesis
It is assumed:
a. Finite elements (cable segments) of the net are direct between the nodes, having
constant diameter and zero bending stiffness;
b. For each element, there are know initial values of axial forces T and the lengths
of cable element when the axial forces are being exerted.
3.2. Finite element cable stiffness matrix
Under the hypothesis assumed earlier, finite element is considered to be a bar,

figure 7.
T T x ui i j x
j uj
vi wi
vj wj
Y Z Y Z
a) b)
Figure 7. Element of a cable net in the local coordinate system. a) Tensile forces T of the
element with length l. b) Positive components of nodes displacement.
Figure 7 shows the element in the local Cartesian system. Initial axial force T of
element is shown in figure 7.a), positives components of displacement in i and j
nodes are shown more detailed in figure 7.b. These six components of
displacement represent the vector of nodal displacement of the element ∆ which
has the following form:
T
∆ = ui , vi , wi , u j , v j , w j  (23)
Where the superscript T denotes vector transposition.

Elastic stiffness matrix kE of the element, having been included in the local
coordinate system, characterizes state of the element under the axial force T = 0,
having the following form:
1 0 0 −1 0 0
0 0 0 0 0 0
EA 0 0 0 0 0 0
kE = (24)
l0 −1 0 0 1 0 0
0 0 0 0 0 0
0 0 0 0 0 0
where E is modulus of elasticity of the cable;
A is cross-sectional area;
l0 is the length of the element at T = 0.
The elastic matrix kE is a part of the tangent stiffness matrix kt of the finite cable
net element:
kt = kE + kG (25)
where kG is the geometric stiffness matrix of the element..
The tangent matrix kt can be derived by means of imparting unit displacements to
the nodes of a bar with the length l, whereas the bar element is tensioned due to
axial force T. The element before and after imparting displacements to the nodes is
shown in figure 8.
l
X T i j T
O
ui vi uj
vj
T wi
i wj
Y l
Z
j
T
Figure 8. Cable net elements before and after displacing the nodes
The vector of nodal forces F of the element in the following form corresponds to
the vector of nodal components of the element displacements ∆:
T
F =  Fix , Fiy , Fiz , F jx , F jy , Fjz  (26)
O X
Y
Z
l
i
Fix l
a)
j
l
i
l
Vj=1
b)
l
Fiy T
j
l
i c)
l
wj=1
l
T
Fiz
Figure 9. Unit components of the displacement vector in direction X (a), Y (b), Z (c) and
corresponding nodal forces
The values of nodal forces initiated by individual unit displacements in direction X,

figure 9.a):
EA EA
ui = 1 , Fix = , Fjx = −
l l
EA EA
u j = 1 , Fix = − , F jx = (27)
l l
in Y direction, figure 9.b):
T T
vi = 1 , Fiy = , F jy = −
l l
T T
v j = 1 , Fiy = − , Fjy = (28)
l l
And in direction Z , figure 9.c):
T T
wi = 1 , Fiz = , F jz = −
l l
T T
w j = 1 , Fiz = − , F jz = (29)
l l
If we write the relationship between the vector of nodal forces F and displacement
vector ∆ we can write:
 EA EA 
 0 0 − 0 0 
 l l 
 T T 
0 0 0 − 0 
 Fix   l l  ui 
F    v 
 iy   0 T T 
−   i
 Fiz   0
l
0 0
l   wi 
 = X   (30)
 F jx   EA EA  u j 
 F jy   − l 0 0
l
0 0   vj 
     
 
 F jz   0 −
T
0 0
T
0   w j 
 l l 
 T T 
 0 0 − 0 0 − 
 l l 
And in a shorter way can write:
F = kt ∆ (31)
Also we can make the following approximation:
l l, (32)
T ≈T (33)
Then the tangent stiffness matrix kt of the element can be rewritten according to
(25):
1 0 0 −1 0 0 0 0 0 0 0 0
0 0 0 0 0 0  0 1 0 0 −1 0 
 
EA  0 0 0 0 0 0 T 0 0 1 0 0 −1
kt =  +   (32)
l  −1 0 0 1 0 0 l 0 0 0 0 0 0
0 0 0 0 0 0 0 −1 0 0 1 0
   
0 0 0 0 0 0 0 0 −1 0 0 1
References
1. Kadlčák J., The Calculation of Suspended Cable Roofs, Proc.1st Conf.on Mechanics, Praha, vol.3,
1987.
2. Kopentetz L., Cătărig A., Teoria Structurilor Uşoare cu Cabluri şi Membrane, U.T.Press, Cluj
Napoca, 2006.
3. Leonard J.W., Tension Structures, Behavior and Analysis, McGraw Hill Book Company, New
York, 1997.
4. Stefănescu B., Soluţii Eficiente de Alcătuire a Structurilor Metalice de mare Deschidere, Teză de
doctorat, Bucureşti, 2000.
5. Tibert G., Numerical Analyses of Cable Roof Structures, Royal Institute of Technology,
Stockholm, Licentiate Thesis.
City of Iassy’s monumental constructions

Constantin Ionescu
Civil Engineering, University, Iassy, Zip code: 700482, Romania
1. TOWN OF IAŞI (IASSY)
Iassy is the administrative center and the greatest town of the County of Iassy. It is
the second municipality of the country after Bucharest and it was the ex-capital of
Moldavia and Romania.
The town has been mentioned for the first time in a commercial benefit issued in
1408 by the Ruler of Moldavia, Alexandru the Good. Still, considering the fact that
exist buildings older than this date (for example the Armenian Church built in
1395), it is believed that the town is older than that.
Armenian Church
The town had the initial centre on the first terrace of Bahlui, close to the place
where the river receives the waters of Nicolina creek. In time, it has enlarged its
area on Bahlui Banks and on the descents of seven hills (Copou, Şorogari,
Miroslava-Galata, Cetăţuia etc.).
Many archeological discoveries have proved that these places have been lived
since Paleolithic. There were also identified many settings from VII th – XIV th
century, which proves that the territory of the future town has been strongly
habituated.
Considering the last discoveries, the birth process of the medieval town had two
phases: the first one, before establishing here the residence of the Moldavia rulers,
when Iassy became the economical centre of the region; the second, after the
construction of Royal Court, at the beginning of XV th century, when took place
the acceleration of urbanization process. The official documents mentioned the
town as fair in the second half of the XIV th century, as custom point in 1408 and
as town in 1415 and 1434, when existed here a royal court.
In 1564, the ruler Alexandru Lăpuşneanu has moved here the Capital of Moldavia
from Suceava. In 1640, Vasile Lupu has set up here the first school in Romanian
language and put the basis of a printing - house in the church Trei Ierarhi. In 1643,
the first book published in Moldavia appeared at Iassy.
Between 1565 and 1859, the town was the Capital of Moldavia, than between 1859
and 1862, Iassy and Bucharest were the capitals de facto of the United Principality
of Moldavia and Valahia.
Capital of Moldavia from XV th century till 1859, Iassy knew a more important
development from the second half of the XVI th century, when the reign has
abandoned the ruling system of the state from periodical displacements at different
courts in the country, establishing permanently in this town.
The Vasile Lupu ruling (1634-1653) influenced the town appearance, which was
enriched with new edifices. The setting development has lasted also in the XVIII th
century, in 1776 were discovered the existence of 30 crafts of artificers and
chapmen’s from Iassy. From the XVII th century, Iassy was an important cultural
centre: here has functioned the first printing house from Moldavia and the first
superior school (Vasilian Academy); here has lived and written Grigore Ureche,
Miron Costin, Dimitrie Cantemir, Ion Neculce. In the second half of the XIX th
century took birth the literary - cultural grouping Junimea, which has reunited
important names such as Titu Maiorescu, Mihai Eminescu or Ion Creangă.
In the first decades of the XIX th century have been taken the first measures of
setting modernization, which had the aspect of a big village, with large green
spaces. In 1806 appeared the first lanterns for the city lighting; in 1833 has started
the streets pavement; in the same time period have been taken measures for Bahlui
course regularization and was created the security for fire protection.
712 C. Ionescu
2. MONUMENTAL CONSTRUCTIONS IN TOWN IASSY
“Saint Nicolae Domnesc“ Church
“Saint Nicolae Domnesc“ Church was built between 1491-1492 by Ştefan the
Great in the most pure and beautiful Moldavian style. Situated close to the old
Royal Court and because of the fact that here have been made rulers almost all of
the Moldavia Rulers, since Despot till Alexandru Ioan Cuza, the church was told
Saint Nicolae – Domnesc. Till the end of XVII th century, the Saint Nicolae-
Domnesc Church has served as Metropolitan Cathedral.
By the end of XVI th century it was the most beautiful from the Iassy’s Churches.
In 1675-1676 the ruler Antonie Russet starts the restoration works, continued in
1680 by the ruler Gheorghe Duca.
As a consequence of the earthquakes and burnings which took place, the church
was demolished in 1888.
It has been rebuilt by the end of XIX th century (between 1890 - 1904) by the
French architect Andre Lecompte de Nouy, and currently is the oldest religious
edifice in Iassy which lasted till today.
Galata Monastery
It was build between 1579 and 1584 by the ruler Petre Schiopul, after the victory
from Bogdaneşti. The monastery has a strong wall made by stone which together
with the bell tower, by its strategic position, has insured an asylum place for the
ruler family.
It is also known that in its first reign, 1574-1579, Petru Schiopul has built another
church at the Galata hill laps. That church has felt down during an earth flow. It
remained from the old church the bell with the script 7087 (1579), month of March
25 and which is being exposed in the monastery.
The present painting was restored in 1811 by Vasile Dubrovski, because the
original one has been destroyed in 1762 by fire.
714 C. Ionescu
Golia Monastery
The monastery was built between 1650 and 1660 by the ruler Vasile Lupu. The
frontage is in Late Renaissance Style, which gives it a certain note in the
Moldavian architecture of the XVII th century. The Golia Tower (30 m height)
gives a large panorama over the town.
On a grave block, dated 1515, is introduced an information by which a certain
Teodora is counted by the founders of a church which existed in that place. After
the renovation, Ana, the widow of Ioan Golia, gave in 1606, the church to the
Vatoped Monastery.
In 1650 Vasile Lupu decides the total transformation of the ruined place. The
rebuilt took place after the architectural creation of the old church and was finished
by his son, Stefaniţă Vodă, in 1660.
Cetăţuia Monastery
It was founded by Duca Vodă. The monastery construction has started in 1668 and
was finished in 1672
It is one of the most beautiful art creations from Moldavia. Here has functioned the
famous Greek printing – press which has printed books from all the Greek space.
Among other works here appeared in XVIII th century the full work of Saint
Simion of Thessalonica.
The walls of the enclosure, tall of almost 7 m have protection walls and defense
guns. The monastery contains an art collection and is visited by many churchly and
laic delegations. In the church there are the graves of ruler Gh. Duca and his
daughter Maria.
716 C. Ionescu
Dosoftei House
The building where the Metropolitan Dosoftei has adjusted in 1679 the second
printing – press from Moldavia, is a construction from the second half of the XVII
th century.
The house has three frontages, curved and rectangular windows, an opened
threshold, with tall and slight columns and rural turrets. It has been reconstructed
between 1966 and 1969. In august 1970 it was opened here the section of old
literature of the Romanian Literature Museum.
The museum has different important pieces: a manuscript Slavonic from the XIV th
century (approximately 1350-1380); the Apostle, the oldest dated Romanian
manuscript (1559 –1560); the writing of Metropolitan Varlaam, the first printing in
Romanian language from Moldavia (1643); a copy realized by Al. Beldiman of
“Moldavia Chronicle” by Ion Neculce. Also the pattern of printing - press of
Dosoftei, religious paintings collections from XVI th and XVIII th centuries.
Lipovean Church
The first lipovens who came to Moldavia by the middle XVIII th century, hunted
by the Russians was due to the fact that they opposed to the clerical reform
initiated by tsar Alexei Mihailovici Romanov.
The community of Russians from Iassy has built in 1780 their first church on the
left side of Bahlui by wood, right on the current place. Damaged due to the time
and weather, the church was rebuilt by bricks in 1830, but the tower and the bell
tower are still in wood. In 1866, takes place a fire which has ruined the church
entirely.
The restoration and development of Russians cult bezel took place between 1872
and 1882, after the plans of Romanian architect Ştefan Emilian and by public
subscription.
Currently this church in Iassy has several valuable religious paintings created in the
traditional style of religious Russian art.
718 C. Ionescu
Union Museum in Iassy
The building where exists the Union Museum was built at the beginning of XIX th
century (1806), in Moldavian neoclassical style.
The archeological researches made by the Institute of Archeology in Iassy have
confirmed the information received from documentary sources as for the Museum
building it’s no older than the beginning of XIX th century, but it was build on an
area lived long before this date.
In the summer of 1944 the construction had to suffer due to the fact that the north -
east wing was bombed, the current Union Museum, remained only half of the
initial built area.
After the Union proclamation in 5 / 24 January 1859, the construction was hired by
the state to serve as residence of ruler Alexandru Ioan Cuza as long as the capital of
Romanian Principality stayed at Iassy (from 1859 till 1862).
In 1956 the construction became the Union Museum. Here can be found objects
belonging to the ruler, exhibits connected to his activity regarding the Union,
documents, and interior decorations.
Cathedral of Moldavia and Bucovina
The idea of building a greatest church, monumental one, at Iassy, belongs to

Archbishop Veniamin Costachi. Lordly Charter from 8 august 1826, regarding the
design and construction works of the new church, is the act of birth of the
Cathedral. It was built on the place where first were two smaller churches: the
White Church – XV th century and Stratenia Church – XVII th century. First the
work has been done between 1833 and 1839, after the plans of architects Johan
Freywald and Bucher, but due to the fall of the central arch, the church has
remained in ruin till 1880. The Archbishop Iosif Naniescu with the help of the state
authorities will finish the construction works in 1887. The architect Alexandru
Orăscu, rector of the Bucharest University, will restore the project of the great
church, by abandoning the huge central dome and the painting will be realized by
Gheorghe Tatarescu. The Cathedral Blessing took place on 23 of April 1887.
720 C. Ionescu
Roznoveanu Palace – Iassy Municipality City Hall
Roznoveanu Palace is an architecture monument, built in 7 - 10 decade of the

XVIII- th century, in Iassy.
The building accommodates the current headquarter of the Iassy city hall.
Impressing by richness and the interiors pomp, it was built between the decades 7 -
10 of the XVIII th century and renovated between 1830 and 1833 by the well
known architect Johan Freywald, the one that projected the Cathedral in Iassy. As
architecture but also by the interior mural and the statues that decorated in 1830 the
exterior, the palace belongs to the neoclassic style, with baroque ornamental
elements. Till 1891 when it was sold to the state, the building was the domicile of
the Rosetti – Roznovanu family, the strongest and most influential family in the
Moldavian political area.
Frumoasa Monastery
The monastery is placed on the place of an older church, created in the XVI th
century. Situated on a terrace with a height of 8-10 meters compared with Bahlui
flat, the Church Frumoasa dominates the view by its white and slim silhouette and
by its four domes, monitored by an imposing tour with a stone wall destroyed by
weather.
Ioasaf Voinescu, rebuilt the church in 1836.
The entrance in the church is preceded by a beautiful portico Doric, composed by
four tall columns, the ones in the middle being connected between them by an
arcade in arch bond, which supports a triangular fronton.
722 C. Ionescu
The Iassy Train Station
The plans of the main building, with the palace design, had a length of 133,8 m and
with 13 rooms, which were inspired by the architecture of Dogi Palace on Canal
Grande in Venice.
The Iassy station was placed on the swampy flat of Bahlui. During the time the
construction has suffered important alignments; due to the creation of foundations
have been done forcemeats of almost two meters. This way due to important
cracks, the south wing was separated from the main body.
Today we can consider that the birth of CFR Building Station for the town Iassy to
be signed at 21 September 1868, when was also emitted the “act for the rail road
construction and exploitation from Suceava to Roman with ramification at Iassy
and Botoşani”. In that moment the works have been taken by the Company of
Victor von Ofenheim.
Old Custom House
The Old Custom House construction initiated its history as a building in 1873,
when the king Carol I, found out about the visit in Iassy, and started the
construction of "Custom House" in Iassy Train Station. The engineer for this
building was Grigore Eliad. Together with the first stone for the building
construction were also placed a few coins from that time, golden carols and silver
ducats, in order to confess to the next generations about the dead past.
The building, rebuilt a few years ago has even now the same purpose as in the past.
It is a store – warehouse, place were, as mayor Vasile Pogor said, “the dealers from
the villages will come to buy merchandise”.
724 C. Ionescu
National High school
The National High School building is built between 1890 and 1894, after the plans
of architect N. Gabrielescu.
The final aspect of the building is a total of three bodies with detailed destinations,
distributed considering a predetermined zonal functioning, typical for the area form
and dimensions.
Centre of Gravity is represented by the study rooms, under the form of wings
around an internal court, after the model of academic architecture practiced in the
Central and West Europe but received by the majority representative edifice of the
époque.
From the architectural point of view, the building is observed reported to other
époque study buildings, by the originality of the exterior but also by the human
aspect in interior, by the way in which was projected the interior space for the
students, the main beneficiaries of the school.
“Alexandru Ioan Cuza” University
The current building of the University has been built between 1893 and 1897 after
the designs of architect Louis Blanc and inaugurated in the presence of King Carol
I and Queen Elisabeta. The building has a combination of classic and baroque
styles, its monumental entrance leading in the famous “Room of Lost Steps”,
which was decorated with paintings of Sabin Bălaşa.
726 C. Ionescu
“Vasile Alecsandri” National Theatre
The building plans belong to the famous Viennese architects Fellner şi Helmer,
which also designed similar constructions in Wien, Praga, Odessa, Zürich.
The National theatre building is an important architectonic jewelry which holds
real art monuments: the Curtain painted in 1896 by the Viennese master M. Lenz
and finished by one of his disciples, presents in the centre the mythology of life,
with the three ages and on the right side, the Romanian Principalities Union
(Moldova, Transilvania and Ţara Românească); the Iron Curtain, depicted by Al.
Goltz, with ornamental motives symmetrically displaced, which detaches the scene
from the rest of the room; the roof depicted by Al. Goltz, in pastel colours,
represents allegory paradisiacal, being illustrated with nymphs and angels and lined
in rococo style; the crystal chandelier from Venice with 109 lights.
Râpa Galbenă
The Elisabeta Esplanade was the name known even today as the “Yellow Coulee”.
The space was created by the time when the mayor of Iassy was Alexandru
Badarau.
The “Yellow Coulee” construction is impressive even today after a century by its
monumentality. It was realized between 1898 and 1900.
The place were it was settled the art work called the “Yellow Coulee”, previously
existed a place were the engineers in the area and the potters used to take yellow
clay. This explains the name of the place, transmitted from generation in generation
till today.
728 C. Ionescu
Culture Palace
Culture Palace from Iassy was inaugurated in 1926, by King Ferdinand of the
Romania. Realized upon the plans of the architect I.D. Berindei, the construction
lasted two decades. A first monument has been raised on the ruins of the old royal
court (1434), which has been rebuilt in neoclassic style by Price Alexandru Moruzi
(1806-1812). A fire at the end of 19th century has destroyed this palace and so at
the beginning of the next century has been taken the decision of building the
present construction.
Central Academic Library in Iassy – The Royal Cultural Foundation
"Mihai Eminescu" Central Academic Library

As a culture element, the Central Academic Library assures the documentary base
necessary to the study process.
Today it hosts the branch office of three institutions: British Council, German
Library Iassy, and Centre of Educational Advisory of the Fulbright Commission.
The library has been built between 1930 and 1932 by the engineer Emil Pragner
after the designs of architect Constantin Jotzu and it is one of the four central
libraries in Romania.
The interior of this architectonic monument is built in Carrara marble and Venetian
mosaic, and the exterior is decorated with columns in ionic style, neoteric pillars,
small triangular frontons and medallions with important personalities of our
national culture.
730 C. Ionescu
National Bank Palace in Iassy
The construction has been designed in the spring of 1946, by the architect Radu
Dudescu together with his colleagues from the Architecture Service, Nicolae
Creţoiu and Spiridon Spirescu, a new place for the Iassy residence. The project has
been approved by the Administration Board of BNR and in the autumn of the same
year has started the construction works.
The building architecture, designed in neoclassical style, was created in a manner
in order to be introduced in the proximity of the historical monuments of Iassy. The
construction of BNR Iassy was made by the firm of Eng. A. Şesan, professor at the
Politechnical School here. For the construction had worked the students: Matei
Botez, Virgil Focşa, Tudor Silion, Nicolae Orlovschi, future professors of the
Construction Faculty in Iassy.
Hotel Europa
732 C. Ionescu
“Petre Andrei” University Building

“Traian” Hotel in Union Square, built by Eiffel Company
References
1. Cihodaru, C. ş.a., Documenta Romania Historica A Moldova, vol. I (1384-4886), Ed. Acad. RSR,
Bucureşti, 1975.
2. Simanschi, Leon ş.a., Documenta Romania Historica A Moldova, vol. II (1449-1486, Ed. Acad.
RSR, Bucureşti, 1976.
3. Cihodariu, C. ş.a., Istoria oraşului Iaşi, Editura Junimea, Iaşi, 1980.
4. Mic dicţionar enciclopedic, Editura Enciclopedică Română, Bucureşti, 1972.
5. Mihai Sorin Rãdulescu, Arhitectul Radu Dudescu – un episod semnifictiv din arhitectura
românească interbelică, Rev. Banca Naţională a României, ISSN 1582-7550.
6. Sabina Mariþiu, Arhitectul Radu Dudescu şi Banca Naţională a României, Rev. Banca Naţională
a României, ISSN 1582-7550.
The design of large sized floors

Ludovic Kopenetz, Alexandru Cătărig, Titu Hodişan
Department of Civil Engineering, Technical University of Cluj-Napoca, 400020, Romania
Summary
The large sized floors can frequently be found in multistory car parks structures,
bank buildings and industrial buildings. These floors, whose plan dimensions are
between 30x30 square meters and 50x50 square meters (or larger) , need special
attention considering the deformation of the structural concrete and the function of
the building. The paper presents some aspects of the structural analysis that the
structural designers have to consider.
KEYWORDS: floor, slab, structure system, drying shrinkage, concrete,

reinforcement, tensile unit strain, salt water, oil, chlorides.
1. INTRODUCTION
The growth rate of our country’s economy has been, for the last years, above 8%.
In the field of civil construction the rise is over the mean value of the national
economy. This has led to the appearance of large dimension structures.
Unfortunately a lot of these structures have shown delicate structural problems, as
for example: cracks, flaws, in many cases, in the horizontal load resisting system
(floors) (Fig.1,2,3).
Overground and underground parking structures belong to this category, [1], [2].
There are a lo of large span parking structures in combination with other facilities
like: flats, offices, markets, placed on the upper levels.
The main structural material for these structures is the reinforced concrete and such
will be the case in the following years due to the use of high resistance concrete,
[3].
Structural system selection should satisfy strength, flexibility, durability, easy
maintenance to vertical and horizontal loads. If the vertical load structure selection
is not difficult because of the multiple structural alternatives (frames, frames with
structural walls, tube systems), the horizontal load structure selection requires
special analysis because of the large spans.
Figure 1
736 L. Kopenetz, A. Cătărig, T. Hodişan
2. STRUCTURAL ANALYSIS
The selection of the structural system is influenced by the basic function of the
building. As compared with the constructions of the 80’s and the 90’s today large
span construction (2500-3000sqm) prevail. An important problem appears in the
case of high values of the axial forces engendered by the vertical loads (even in
case of live load of 200daN/sqm).
The selection of the structural system is also influenced by horizontal

displacements which must fulfil the condition
∆r
≤ 0,007 (1)
He
where:
- ∆ r is the relative level displacement excluding nonstructural walls,
- H e is the height of the level.
Figure 2
Figure 3
Note: This condition for displacements is applicable in the case of buildings with
deformable wall panels.
The horizontal load resisting system consists of ceilings with bays, slabs on beams
and girders and slabs on columns with prestressed or nonpresstresed reinforcement.
In the case of nonpresstresed reinforcement in areas with big tensile stress, cracks
appear.
Considering the great width of the slab due to the large spans (15 – 25m),
the value of drying shrinkage is ε c = 0.6 ‰.
This value is much bigger than the value given by STAS 10107/0-9, [5],
( ε c = 0,2 − 0,3 ‰) 400% bigger than the maximum concrete tensile strain
( ε t = 0,15 ‰).
In structural analysis,the use of the equation of drying shrinkage with a variation

of a temperature ∆t = −10 0 is a fundamental error because none of the effects can
be considered similar even formally, [4]. Thus, considering the variation of
temperature, both concrete and reinforcement shrink. In the case of drying
shrinkage, only concrete does.
Note: Existing software programs are not able to observe these aspects.
Figure 4 - Stresses in a slab of large dimensions. Concrete of different classes and

similar reinforcement
In the case of drying shrinkage there appears a change of the induced stress, even
to exterior statically determinate structures. Considering the relaxation of concrete
these phenomena are balanced in time, but for the green concrete they are
extremely dangerous. For the calculation of the induced stress of drying shrinkage,
several methods can be applied such as Trost, Dischinger- Fritz, or the practical
method.
If considering a reinforced concrete linear element:
ε = ε c − ε bt = ε a (2)
where:
- ε is the unit strain of the linear element,
- ε c is the initial drying shrinkage unit strain,

- ε bt is the concrete tensile unit strain,
- ε a is the reinforcement unit strain.
As concerns the application of the relation (2), in figures 4,5,6 are shown the
values of drying shrinkage stresses, considering the value of drying shrinkage unit
strain ε c = 0,6 ‰ and ε c = 0,3 ‰ for different qualities of concrete and
reinforcement percentage.
Figure 5 - Compression stresses in a RIC beam with Symmetrical reinforcement

versus concrete class, contraction coefficient and reinforcement percentage.
3. THE IMPACT OF SALT WATER AND OF OIL IMBUING

ON THE HORIZONTAL RESISTING SYSTEM
In the case of multilevel parkings the influence of salt water and oil on the
horizontal resisting system is a serious problem.
In winter cars leave snow on the floors and sometimes oil drips.The cracking state
ot the superior part of the floor in column zones above beams and girders will
allow the migration of the salt water and oil in the structure.
The effect of salt water is the corrosion of the reinforcement (rust foil) and the
presence of chlorides above normal values (300-500mg/kg), in the concrete.
Figure 6 - Tension stresses in the reinforcement of symetrically reinforced beam

versus concrete class, contraction coeficient and reinforcement percentage.
The presence of chlorides speeds the corrosion of the reinforcement, leading to its
disappearance( it is a self propagating process in progress as long as there steel
exists).
In combination with the calcium hydroxide [Ca (OH)2] in the concrete, oil
produces salts (non resistant) and if this is on a long term the concrete resistance
decreases.
If these zones coincide with bending pressure zones in beams and slabs, there is
considerable shrinkage because the resistance of concrete is diminished and cracks
open in the streched zones.
4. CONCLUSIONS
* Initial tensile stresses of concrete and pressure stresses of reinforcement

are proportional with the concrete quality, in the case of a strong reinforcement
(Fig.4).
* Initial tensile stresses of concrete are proportional with the reinforcement

percentage, with concrete quality and with the drying shrinkage unit strain (Fig.6).
* Initial pressure stresses of reinforcement are in inverse ratio to the

increase of the reinforcement percentage and proportional with the quality of the
concrete.
* It is necessary to take special measures to diminish the effects of the

shrinkage of the concrete.
References
1. CEB-FIP (1990), Model code for concrete structures (MC-90), CEB Thomas Telford, London,
1993.
2. ACI – COMMITTEE 109 (1992), Prediction of creep, shrinkage and temperature effects in
concrete structures – 209R-92,ACI, Detroit, Michingan.
3. Paulay, T., Priestley, M.I.N., Proiectarea structurilor din beton armat la actiuni seismice. Editura
Tehnica, Bucureşti, 1997.
4. Eurocode 2 (1991), Design of concrete structures, European Prestandard, ENN 1992-1, 1991E,
Brussels, Belgium.
5. STAS 10107 / 0 -90, Calculul si alcatuirea elementelor structurale din beton armat si beton
precomprimat.
6. Herczeg, K., Gepkocsi parkolok, muszaki, konyvkiado, Budapest, 1978.
7. Dabija, F., Vierescu, R., Proiectarea structurala a cladirilor pentru parcaje, Buletin AICPS, nr.4,
1999.
Structural analysis of sandwich type composites
Alexandru Cătărig, Ludovic Kopenetz, Titu Hodişan

Department of Civil Engineering, Technical University of Cluj-Napoc , 400020, Romania
Summary
The light structures with sandwich elements have a large use in the civil and
industrial buildings. There are a lot of materials that are used to make the
sandwich elements, and thus results a large variety of elements. Considering the
composition, the structural analysis must have in view the general and local
phenomena occurring both in the elements and at the level of the whole , with a
view to obtain the standard structural security.
KEYWORDS: sandwich composites, layer, bending stiffness, torsion stiffness,

polymers
1. INTRODUCTION
The sandwich-type composites are part of the light but very efficient structures.
The specifically qualities of the sandwich-type composites are:
• The accessibility for design of a large variety of complicated
shapes.
• Durability.
• A very good dynamical response (in case of earthquake for
example).
• A high capacity of thermo-insulation.
• A reduced cost production for a large variety of in-situ production.
• An independence of environmental condition, [1],[2],[3].
The sandwich-type composites were used for the first time in Aeronautical
structure (The Mosquito – 1940) and for the production of some light wood
building constructions (in Finland).
From 1960 the sandwich-type composites, started to be used as layered elements,

(bonded to each other) having the properties as a combination of all components,
[4],[5],[6],[7].
The ASTM C274-53 (US STANDARD) considers the sandwich-type composites

as a laminar material including some elementary components bonded to each other
in order to accord superior properties to the final item (Fig.1,2,3).
Figure 1
Figure 2
Figure 3
The very first idea of creating sandwich-type composites belongs to Leonardo da

Vinci, and the name “sandwich” was used for the first time by Fairbairn in 1849.
Choosing the proper materials, especially the facing layers (with decorative, but
particularly high durability properties) involves many pre-conditions as:
• Physical properties – density, color, thermal conductibility,
reflection, environmental durability (excessive heat/cold),
chemical stability, corrosion, etc.
744 A. Cătărig, L. Kopenetz, T. Hodişan
• Mechanical properties – stress resistance (tension, shear, fatigue)

resistance to distortions, impact, etc.
2. SANDWICH-TYPE COMPOSITES COMPONENTS
Due to diversity of materials used in the sandwich-type composites there are many
available products. The most widespread sandwich-type composites consist in two
external metallic or concrete (sometimes pre-compressed) layers and, an
intermediate layer (usually with a lower elasticity modulus) who connects the two
external layers.
The external layers, usually metallic or structurally reinforced concrete may,

include local decorative strips. The reinforcement of the lateral layers may vary
according to the local requirements such as.
If the external layer is made of steel, the usual thicknesses are 1.5mm for a 0.5mm
span. The bonding (medial) layer is usually made of Polymers with reduced density
( ρ = 10 − 50daN / m 3 ) such as PVC, polyurethanes, polystyrenes or other light
materials as porous rubber, wood, fibers, stabilized pulps, etc.
The thickness of the interior layer may vary between 40 − 150mm according to the
used material.
The external layers are meant to resist to the bending moments (compression
/tension) and the inner connecting layer must be capable to resist to the inter-
laminar, and torsion forces.
The sandwich-type composites are generally delivered as panels having the

dimensions (50 − 150cm) ∗ (300 − 1000cm) .
The sandwich-type panels are a bit more expensive than the classical
building materials (200–300 €/m2 or 30–45 € /m3).
Considering the thermal insulations it must be mentioned that a sandwich-type
panel of 80 − 100mm is equivalent with a brick wall of 37.5cm, for a mass to
surface ratio of 13/800 for external layers of steel, and 8/800 for external layers of
aluminum.
Apart from steel or concrete, for the external layers may be used wood panels,
reinforced gypsum, or other fiber reinforced plastics materials.
The connection among the layers is performed by an adhesive material.

If, from the initial stages of production the three layers were designed to match
smoothly each other, this procedure will allow a perfect continuity of the whole
structure (Fig.4,5,6).
Figure 4 Figure 5
Figure 6
3. EXPERIMENTAL STUDY FOR SANDWICH-TYPE

COMPOSITES
The tests were performed both for the individual layers and for the whole structure
(sandwich-type composites).
For each layer the tests are the usual ones, such as:
4
• Tension test for determining the limit of connection loads between the
external and internal layers. This test is performed perpendicular on the
item's surface.
• The bending test is performed at 4 points and afterwards at 3 points in
order to get the direct bending stiffness ratio along with the tangent
stiffness ratio. It is very important to consider the sample's dimensions in

order to get the precise data about the transversal curvatures.
• The shear test will be performed on the sandwich product in order to get
the shear resistance of the core layer.
• The torsion test is performed on the sandwich product for determining the
torsion stiffness ratio.
4. STRUCTURAL ANALYSIS
Considering the basically structure of the sandwich-type composites, the analyses

must include the general comportment of the whole structure along with the local
influences of each layer including their interaction. Only on this general base, the
analysis will be performed according to required standards.
For the smooth external layers the local bending stiffness may be neglected and
such, the moment of inertia for the whole section will be:
2
 t
I sandwich = 2 Bt  H − 
 2 
where:
- B is the panel's width,
- t is the thickness of the external layer,
- H is the high of the whole panel.
For the deformation calculation, the main task will be performed by the
core's shear resistance. The displacements that occur due to the composite's
bending or due to shear have similar values. On this ground, the analysis
must be performed very attentively considering in each case, the main
impact from all load cases, in order to eliminate all the false presumptions.
5. EXAMPLE FOR SANDWICH-TYPE COMPOSITES

CALCULATION
How big it will be the displacement in the point 1 for a beam made of a sandwich
panel with the following characteristics: E steel = 2.1 ⋅ 10 6 daN / cm 2 ,
G polyurethan foam = 50daN / cm 2 .
Solution
f = f M + fT
2
 0.06  4
I t = 2 ⋅ 60 ⋅ 0.06 ⋅  5 −  = 177.85cm
 2 
AFOAM = 60 ⋅ (10 − 0,06 ) = 596.4cm 2
100 ⋅ 400 3
fM = = 0.36cm
48 ⋅ 2.1 ⋅ 10 6 ⋅ 177.85
100 ⋅ 400
fT ≅ = 0.34cm ≅ f M
4 ⋅ 50 ⋅ 596.4
l
f = 0.36 + 0.34 = 0.7cm < = 1.33cm
300
6. CONCLUSIONS
* The light sandwich-type composite structures confer the possibilities of

an unlimited variety of constructions unattainable by any other building materials.
With the nowadays ecological conditions, since the beginning of the new Millennia
all the classical structures must be up-dated, and consequently the composite
materials started to be competitive.
* They may be used for a large variety of constructions: in the earthquake

areas, in harsh environmental condition on no soil pre-preparations, everywhere.
On those premises it might be predicted that the use of composed structure will
increase exponentially (Fig.7,8,9,10).
Figure 7 Figure 8
Figure.9 Figure.10
* The light sandwich-type composite may be used also for the

consolidations of already existing buildings without demolitions simply by
replacing the old items with new lighter but stronger non-structural items.
References
1. Kunst – Ghermanescu, M., Contribuţii la calculul construcţiilor foarte uşoare alcătuite din arce
tubulare pneumatice, Teză de doctorat, ICB, 1983.
2. Hagiescu, S. , Crăilă , E. , Dragomir , E. , Construcţii uşoare aeroelastice pentru acoperiri
provizorii, INCERC, 1977, 1979.
3. Sofronie, R., Aeroelastic stability of wide – span surface structures, Romanian Journal of
Technical Sciences, Applied Mechanics nr.6, Tome 22, 1977.
4. Kopenetz L. G., Ionescu A., Lightweight roof for dwellings, International Journal for Housing and
Its Application, vol. 9, Nr.3, Miami, Florida, USA.
5. Frei, O., Zugbeanspruchte konstruktionen, Band I und III, Ullstein, Fachverlag, 1962.
6. Davies, J.M., Cassette Wall Construction, Proceedings of Steel Constructions Conference,
Timişoara, Romania, 2000.
7. Kopenetz, L.G., Catarig, A., Teoria structurilor uşoare cu cabluri şi membrane, UT Press, 2006,
Cluj-Napoca, Romania.
8. Blumberg, H., “Anforderungen an Technische Chemiefasern in textilen Bauen “ , Chemiefasern /
Textilindustrie, April, 1979.
9. Ivan, M., Mih T., Fem modeling of centered compressed sandwich panels, Proceedings of Steel
Constructions Conference, Timişoara, Romania, 2000.
10. Davies, J.M., Steel framed house construction, Structural Engineer, 78, nr.6, 2000.
11. Herzog, T., “ Pneumatic Structures – A Handbook for the Architect and Engineer “, Crosby
Lockwood Stamples, London, 1977.
12. Williams, C.J.K., “Form Finding and Cutting Patterns for Air Supported Structures, The State of
the Art, The Institute of Structural Engineers, London, June, 1980.
13. Newmark, N.M., A Method of Computation for Structural Dynamics, Journal of Engineering
Mechanic Division, ASCE, vol 85, nr. EM3, 1959.
14. Hinton, E., Razzaque, A., Zienkiewicz, O.C., Davies, J.D., A simple finite element solution for
paltes of homogeneous, sandwich and cellular construction, Procceedings, Institute Civil Engineers,
Part 2, 1975, pag. 43-65.
15. Ivan, M., Cozma, D., The effect of the sandwich panels Diaphragm on the analysis, Proceedings
of Steel Constructions Conference, Timişoara, Romania, 2000.
Resistance of Concrete with Fine-Ground Granulated Slag

Admixture to Sulphates and Acidic Aggressive Media
Jiří Brožovský1, Tomáš Fojtík2 and Amos Dufka3
1,2,3
Institute of Building Materials and Components, Faculty of Civil Engineering, Brno University of
Technology, 602 00 Brno, Czech Republic
Summary
Blast-furnace slag granules may be used not only in mixed Portland cement
making but it is also suitable as a partial substitution of Portland cement in
concrete. Substitution of 20 – 50 % Portland cement by fine-ground blast-furnace
slag granules increases concrete resistance to aggressive media in comparison
with concrete made of mere Portland cement (CEM I). In the given case, concrete
resistance to sulphates (concentration SO42- - 10 000 mg/l) and to acidic
aggressive media (pH = 3) was monitored.
KEYWORDS: Concrete resistance, sulphates, acid(ic) aggressive medium, blast-

furnace slag.
1. INTRODUCTION
Utilization of process waste in making of building materials is a permanent trend.

Among others, blast-furnace slag granules have been often used in this process.
Most frequently, blast-furnace slag has been used in making of mixed cements
based on Portland clinker. Lately however, blast-furnace slag also has been utilized
as concrete admixture. The paper deals with results of monitored resistance related
to concretes, in which Portland cement (CEM I) was partially substituted by fine-
ground blast-furnace slag granules. 10 –50 % of cement were substituted by blast-
furnace slag. Resistance of those concretes to sulphates (concentration SO42- - 10
000 mg/l) and aggressive acidic media (pH = 3) was monitored, whereas the
concretes have been exposed to aggressive medium for 12 months. Concrete
containing 100 % of cement based on Portland clinker served as comparative
specimen. The resistance was assessed in conformity with standard demands of
ČSN 73 1340.
2. DEGRADATION MECHANISM OF AGGRESSIVE MEDIA
2.1. Degradation of Cement Composite by Sulphates
Sulphate corrosion of concrete caused by weather effects or by industrial waste

water with higher ionic SO42- concentration has been studied by many a prime
experts. Statements based on such research say that, dependent on preconditions,
degradation produced by sulphates may have a form of gypsum corrosion.
Preconditions for sulphicoaluminate corrosion are created with concentrations
containing more than 250 mg.l-1 sulphates in 1 litre of dilution. Ca(OH)2 reacting
with SO42- produces gypsum. This process may be expressed by equation (1):
Ca( OH )2 + Na2 SO4 + 2 H 2O ⇒ CaSO4 .2 H 2O + 2 NaOH (1)
The produced gypsum further reacts with high alkali calcium hydro aluminates,
producing ettringite. This process may be expressed by equation (2):
3CaO.Al2O3.6H2O + 3( CaSO4 2H2O ) +19H2O ⇒ 3CaOAl2O3.3CaSO4 .31H2O (2)
The emerging, insufficiently dissoluble ettringite, substantially swells and pressing
against porous walls and capillaries it causes disruption and formation of cracks. In
viewpoint of a possibility to eliminate negative affects of sulphates on concrete,
resp. on polymer cement mortar, there is a very significant fact that formation of
ettringite conditioned by a Ca(OH)2 concentration value of ca 0.46 mg.l-1
(conversion to CaO) as a minimum, i.e. ettringite only arises, when cement stone
contains aluminates with high percent of CaO. In case of poor concentration, no
reaction between SO42- and Ca(OH)2 occurs or it is very subtle and so constitution
of expansive pressures is limited. It univocally follows that there is a possibility to
improve concrete resistance to sulphates, when choosing a suitable type of active
hydraulic admixture diminishing the quantity of Ca(OH)2.
Another aspect, very important in the viewpoint of cement matrix stability in
sulphate medium, is chemical and mineralogical composition of cement. Higher
content of C2S in cement decreases concentration of forming Ca(OH)2 and so the
resistance to sulphates increases; and vice versa, higher content of C3S initiates
higher content of arising Ca(OH)2, whereas concrete resistance using such cement
decreases.
Resistance of cement matrix to sulphates can be also substantially influenced by
presence of tricalciumaluminate C3A in cement. According to Solacolm´s research,
cement resistance to impact by sulphate dilutions can be derived from calculation
of its aluminate modulus Mhl. The aluminate modulus of cement is given by
relation (3):
752 J. Brožovský, T. Fojtík and A. Dufka
Al2O3
M hl = (3)
Fe2O3
On the basis of aluminate modulus values, cements are ranked into five classes,
whereas very resistant cements with Mhl = 0.0 – 0.7 belong to the first class and
cements no resistant to sulphates, where Mhl = 2.4, fall into the utmost class five. It
univocally follows that, should pozzuolana admixture be able to eliminate
degradation of concrete resp. mortar exposed to sulphates impact, the content of
active Al2O3 may not be too high.
2.2. Degradation of Cement Composite by Acids
Concrete based on Portland cement is a strong basic substance. When exposed to

impact of acids (eventually to certain salt solutions thereof), the material reactions
are principally neutralizing. A consequence of such reactions is degradation of
cement matrix, attended by descent of concrete strength characteristic. The effect
of single acid types is determined not only by acid strength (resp. its dissociative
constant), but also by the present cation type etc. Principally, the impact of single
acid types may be described as follows:
Acids causing decomposition of cement matrix are those, destructing the cement
matrix step by step (so called concrete „acid extraction“). To exemplify, we can
mention the affect of mild carbonic acid, de facto present in „famished“ waters.
Initially, such waters quickly leach Ca(OH)2 present in concrete, later it comes to
slower decomposition of hydrated silicates and aluminates. This fact leads to
decreasing of pH value, reflecting in negative impact on corrosion of reinforcement
as well as on stability of some hydrated clinker minerals.
Acids causing cement matrix destruction – this group particularly encompasses
acids, creating voluminous crystalline new formations in concrete microstructure,
when reacting with cement mastic. Initially, the impact of such acids may reflect in
soft decrease of compactness and impermeability or, even in strength increasing.
Continued swelling of crystalline new formations creating in cavities and pores,
however, leads to crakes and, in extreme cases, to total destruction of concrete.
3. WORK METHODOLOGY
To investigate concrete resistance to corrosion, there were prepared test specimens

(blocks 40 × 40 × 160 mm each) conformable to the ČSN 73 1340.
After manufacturing, the test specimens were placed into moist environment for 24
hours, followed – after form removal and before exposure to corrosive medium –
by placing into water bath at 19 to 21°C. After 28 hours of hardening, the

specimens were exposed to corrosive medium action. Selection of parameters
under evaluation is based on the ČSN 73 1340.
Tested specimens were monitored in the light of parameters as follows:
− appearance (visually)
− density D (EN 12390-7)
− ultrasonic pulses velocity V (EN 12504-4)
− dynamic modulus of elasticity Ebu (ČSN 731371)
− compression strength fcc (EN 1015–11)
− tensile strength under flexure fcf (EN 1015–11)
Specimen strength was destructively tested before exposure to corrosive medium
and after 2, 4, 6, 9 and 12 months of exposure.
− Corrosive environment effect during other time periods was evaluated,
based on both, ultrasonic pulse speed variation and dynamic modules of
elasticity.
Corrosive media:
− sulphates - sodium sulphate solution (10,000 mg of SO42- in 1 litre)
− acidic environment – solution with pH 3±0.2 value.
4. CONCRETE COMPONENTS AND COMPOSITION

− Binder: Portland cement CEM I 42.5R
− Aggregates: fraction 0/4 mm hoisted from Bratčice locality, crashed;
fraction 4/8 mm from Olbramovice locality (granodiorit) – basic parameters,
see Table 1.
− Admixtures: fine-ground blast-furnace slag granules (Nová Huť, Ostrava) –
parameters, see Table 2; chemical composition, see Table 3; result of X-Ray
analyse, see Figure 1. Slag contains a low portion of crystalline phases)
− Plasticizer: Stachement 2000
− Mixing water: fulfils demands of EN 1008
−
Table 1. Basic Parameters of Aggregates

Bratčice Olbramovice
Parameter Unit
0/4 mm 4/8 mm
Volume weight [kg/m3] 2530 2610
Bulk weight, jolted [kg/m3] 1650 1625
Porosity [%] 34.8 37.7
Table 2. Basic Parameters of Blast-Furnace Slag

Parameter Unit Slag
Specific surface S [m2/kg] 380
Specific weight ρ [kg/m3] 2830
Basicity modulus Mz [-] 1.07
140
120
100
80
I [-]
60
40
20
0
0 10 20 30 40 50 60 70
úhel [°]
Figure 1. Result of X-Ray analyse blast-furnace slag
Table 3. Chemical Composition of Slag

Component SiO2 Al2O3 Fe2O3 MnO CaO
Content [%] 32.7 6.59 0.19 0.5 41.4
Annealing Non-
Component MgO S2- Loss Decomposable
Portion
Content [%] 12.2 0.55 1.08 6.99
− Concrete composition and basic physical and mechanical characteristics are

shown in Tables 4 and 5 below.
Table 4. Composition of 1 m3 Concrete

Quantity (BS0-BS50
Component
concrete)
CEM I 42.5 R Hranice Cement 416 kg
Fine-Ground Slag 0-50% of cement weight
Sand 0 to 4 mm from the Bratčice
725 kg
Gravel Pit
1087 kg
Olbramovice Gravel Pit
Plasticizer Stachement 2000 4.16 kg
Mixing Water 159 kg
Cement/Water Ratio 0.382
Table 5. Basic Parameters of Concretes

Parameter Unit BS0 BS10 BS20 BS30 BS40 BS50
fcc [MPa] 55.8 53.3 49.7 46.4 43.2 41.1
fcf 7 days [MPa] 7.7 8.5 8.3 8.9 7.4 8.2
D [kg/m3] 2364 2325 2335 2346 2305 2315
fcc [MPa] 63.1 66.8 62.5 65.9 55.3 59.4
28
fcf [MPa] 8.9 8.7 9.0 9.5 9.6 9.4
days
D [kg/m3] 2366 2309 2303 2313 2303 2323
5. FINDINGS OF CONCRETE RESISTANCE MONITORING
Concretes have been exposed to aggressive environment impact for 12 months.

Variation results of parameters monitored in the course of time are graphically
demonstrated in:
− Figure 2: progress of tensile strength under flexure during the exposure
period – concretes stored in aggressive acid environs
− Figure 3: progress of compression strength during the exposure period –
concretes stored in aggressive acid environs
− Figure 4: modulus of elasticity progress during the exposure period –
concretes stored in aggressive acid environs
− Figure 5: progress of tensile strength under flexure during the exposure
period – concretes stored in sulphates
− Figure 6: progress of compression strength during the exposure period –
concretes stored in sulphates
− Figure 7: modulus of elasticity progress during the exposure period –
concretes stored in sulphates.
12
11,5
11 0
10,5 10%
fcf [MPa]
20%
10
30%
9,5 40%
9 50%
8,5
8
0 2 4 6 9 12
Exposure Time [months]
Figure 2. Tensile Strength under Flexure of Concretes Stored in Aggressive Acid Environs
85
80
0
75 10%
fcc[MPa]
20%
70
30%
65 40%
50%
60
55
0 2 4 6 9 12
Figure 3. Compression Strength of Concretes Stored in Aggressive Acid Environs

56000
55000
Ebu [MPa]
54000
53000
52000
51000
0 1 2 3 4 5 6 9 11 12
0 10% Struska 30% 40% 50%
Figure 4. Dynamic Modulus of Elasticity Related to Concretes Stored in Aggressive Acid

Environs
12,5
12
11,5 0
11 10%
fcf [MPa]
10,5 20%
10 30%
9,5 40%
9 50%
8,5
8
0 2 4 6 9 12
Figure 5. Tensile Strength under Flexure of Concretes Stored in Sulphates

85
80
0
75 10%
fcc [MPa]
20%
70
30%
65 40%
50%
60
55
0 2 4 6 9 12
Figure 6. Compression Strength of Concretes Stored in Sulphates
58500
57000
55500
Ebu [MPa]
54000
52500
51000
0 1 2 3 4 5 6 9 11 12
0 10% 20% 30% 40% 50%
Figure 7. Dynamic Modulus of Elasticity Related to Concretes Stored in Sulphates

6. CONCLUSION
Partial substitution of Portland cement by blast-furnace slag granules had a positive

influence on resistance of concretes exposed to aggressive acid and sulphate
environment for the period of 12 months.
Specifically, concretes using slag amounting to 40 or 50% of cement weight show
increased resistance to aggressive acid and sulphate environment, whereas slag
content amounting to 20 – 50% of cement weight increases concrete resistance to
sulphates.
Contributions occurs in the environmental viewpoint (liquidation of wastes,
diminished power consumption for manufacturing of cement, heavy on energy) as
well as in relation to economy (lower consumption of expensive cement).
Acknowledgements
The work was supported by the MSM 0021630511 Plan: Progressive Building
Materials with Utilization of Secondary Raw Materials and Impact their on
References
1. Drochytka, R.et al. Progressive Building Materials with Utilization of Secondary Raw
Materials and their Impact on Structures Durability. Brno University of Technology, Final
report of the project VVZ CEZ MSM: 0021630511, Brno 2006. Brožovský, J.: Subtask 3 (in
Czech).
2. Brožovský, J. and Martinec, P. Durabbility of Concrete with Fly Ash. In Proceedings of the
2d International Conference on Concrete and Reinforced Concrete: Concrete and
Reinforced Concrete – Development Trends. NIIZHB, Moscow, Russia, Vol. 4, 2005. (in
Russian).
3. Brožovský, J., Zach, J., Brožovský, J., Jr.: Durability Of Concrete Made From Recycled
Aggregates. In Proceedings Of The International Rilem Jci Seminar Concrete Durability
And Service Life Planning: Curing, Crack Control, Performance In Harsh Environments,
Ein-Bokek, Dead Sea, Israel, 2006
4. EN 1015-11 Methods of Testing Cement - Part 1: Determination of Strength Methods of
test for mortar for masonry - Part 11: Determination of flexural and compressive strentgth
of hardened mortar
5. EN 12390-7 Testing Hardened Concrete – Part 7: Density of Hardened Concrete
6. EN 12504-4 Testing concrete – Part 4: Determination of Ultrasonic Pulse Velocity
7. CSN 73 1371 Method of Ultrasonic Pulse Testing of Concrete
8. CSN 73 1340 Concrete Constructions. Tests of Corrosion Resistance of Concrete. General
Requirements
9. EN 1008 Mixing Water for Concrete. Specification for Sampling, Testing and Assessing the
Suitability of Water, Including Water Recovered from Processes in the Concrete Industry,
as Mixing Water for Concrete
Comparative Analysis of Some Methods for Dynamic Behavior

and the Structural Design of Multistory Structures
Mihai Coveianu1, Olimpia Marchelov2 and Daniel Bîtcă3
1
Civil Steel Structures Department, Technical University of Civil Engineering, Bucharest, Romania
2
3
Summary
This paper compared different calculation methods for the linear dynamic analysis
of high multistory steel structures.
In order to exemplify the calculations, we choose a 26 levels structure which is up
to be build now in Victoria Place, Bucharest.
Many tests were accomplished, on different ways, such as: modal analysis using
vibrating eigenmodes; spectral acceleration using fundamental vibrating mode,
two vibrating eigenmodes,…,12 vibrating eigenmodes – in order to relieve the
contribution of eigenmodes; for the spectral analysis using 12 modes another set of
tests was made for modal combination (SRSS, CQC, ABS, GMC, 10Pct, Dbl Sum);
structural analysis using different combinations of the ground acceleration for
Vrancea North-South and Vrancea East-West, the computations being realized in
direct method and also using the vibrating eigenmodes.
The results of all the tests were registered in order to dignify the similitude of the
results using more accessible methods with a more laborious calculation method.
KEYWORDS: eigenmodes, structural design, spectral analysis, dynamic analysis,
multistory building, seismic.
1. MODAL ANALYSIS
Figure 1. Model Discretisation

After the modal analysis the following values were obtained:
Table 1. Periods
Mode Period
Mode 1 3.033
Mode 2 2.473
Mode 3 2.055
Mode 4 1.487
Mode 5 1.306
Mode 6 1.126
Mode 7 1.044
Mode 8 0.969
Mode 9 0.969
Mode 10 0.879
Mode 11 0.866
Mode 12 0.815
2
Period
0
1 2 3 4 5 6 7 8 9 10 11 12
Mode
Figure 2. First 12 periods of vibration
The main vibrating eigenmodes are:

• mode 1: translation
• mode 2: torsion
• mode 3: torsion + translation
After the modal analysis, the deformed shape corresponding to the vibrating mode
1, mode 2, ….. mode 12 was obtained and we could educe the following: mode 1 is
concordant with the translational effect on the oy direction, the mode 2 is
concordant with the torsion effect, mode 3-correlated torsion with translational
effect on the x direction and the rest of the modes are concordant with the
individual vibrating modes.
762 M. Coveianu, O. Marchelov, D. Bîtcă
2. SPECTRAL ANALYSIS
For this analysis a conventional statical load was used. It was considered a place in
Bucharest with: Ks=0.2, Tc = 1.5sec., Y=0.25, a = 1.
2.1 Eigenmodes Contribution
That design was made in many cases: with first eigenmode, with two eigenmodes,
three eigenmodes…..12 eigenmodes, in order to relieve the eigenmodes
contribution.
One could notice:
Table 2. Base Reactions

Mode Global Fx [tf] Global Fy [tf]
Mode 1 1747 3077
Mode 1 – Mode 2 2225 3116
Mode 1 – Mode 3 2998 3470
Mode 1 – Mode 4 2999 3471
Mode 1 – Mode 5 3000 3519
Mode 1 – Mode 6 3000 3519
Mode 1 – Mode 7 3009 3540
Mode 1 – Mode 8 3076 3540
Mode 1 – Mode 9 3186 3548
Mode 1 – Mode 10 3192 3556
Mode 1 – Mode 11 3194 3556
Mode 1 – Mode 12 3194 3629
The superposition of vibrating modes was made according to SRSS method, using
the next formula:
2
SRSS = Σyi
Although the main three parameters were followed (Shear Base Force, internal
forces on elements, displacements) only the Shear Base Force was graphically
represented:
Figure 3. Base Shear Force
One can notice that the contribution of the first vibrating eigenmode was
approximately 82% and for the first three modes the contribution was
approximately 95%.
Although in many of the study cases it is marked the fact that on the flexible
structures the first eigenmode is reduced, according to the previous analysis it was
observed that of first eigenmode is important (80-82%).
3.2 The Different Methods for Superposition of Eigenmodes
The results of the spectral analysis (Base Shear Force) for the model with 12
eigenmodes and the main methods of superposition are presented in the next table:
Table 3. Base Reactions using different methods

Mode Global Fx [tf] Global Fy [tf]
CQC 3615 3770
ABS 6884 6977
GMC 3624 3779
10Pct 3416 3678
DblSum 3750 3831
SRSS 3194 3629
In the engineering literature the notions concerning the superposition of

eigenmodes are different. For that reason the SAP engineers (from CSI Berkeley)
gave the possibility to use many superposition methods.
In that paper all that possibilities were studied. Although the fundamental point of
view of the methods are disparate in a practical point of view the results show close
values between these methods (with an exception for the ABS method which gives
different results but it is stated that this method is much too safe).
ABS = y1 + y2 + yi
12000
10000
8000
(T)
6000
4000
2000
0
1 2 3 4 5 6
CQC ABS GMC 10 PCT Dbl Sum SRSS
Figure 4. Base Reactions using different methods
For the 12 eigenmodes model we made the analysis with another 5 superposition
methods: CQC, ABS, GMC, 10Pct, DblSum. The shear force values according to
methods mentioned above is similar; the value is approximately 3800tf (for CQC
method- 3770tf, for GMC method – 3779tf, 10 Pct methods– 3678tf, Dbl Sum
Method – 3831tf, SRSS method – 3629tf) and approximately 6980tf for ABS
method. The value of the member forces keeps the same similitude, approximately
30tf (for CQC method -26.7tf, for GMC method – 26.7tf, 10 Pct method– 30.7tf,
Dbl Sum Method – 28.1tf and for ABS method – 67tf) and maximum displacement
at the top of the structure for the ABS method is 375mm.
Table 4. Displacements using different methods

Displacement ox Displacement oy Displacement xy
Mode
[mm] [mm] [mm]
CQC 161.45 251.50 295.68
ABS 255.07 375.00 438.60
GMC 162.28 251.20 295.33
10Pct 161.05 255.53 302.47
DblSum 160.94 247.11 285.03
SRSS 160.23 256.00 303.58
4. TIME HISTORY ANALISYS
4.1 Direct Integration Time History Analysis
For this analysis the Vrancea North-South Time History was considered and also
Vrancea East-west Time History, each of them on ox direction and oy direction,
and after that a combination was made for the both direction (North-South Time
History on ox direction combined with oy direction, North-South Time History on
oy direction combined with East-West Time History on ox direction).
Table 5. Base Reactions using different Time-History Accelerations

Case Step type Global Fx [tf] Global Fy [tf]
vnsx max 10691 5906
vnsx min -13136 -4088
vnsxewy max 10836 5398

vnsxewy min -13040 -4489
vnsy max 5906 10444

vnsy min -4088 -8230
vnsyewx max 5340 10471

vnsyewx min -4430 -8280
• Time-History Acceleration “Vrancea North-South” on ox direction

• Time-History Acceleration “Vrancea North-South” on oy direction
• Time-History Acceleration “Vrancea North-South” on ox direction + Time-
History Acceleration “Vrancea North-South” on oy direction
• Time-History Acceleration “Vrancea North-South” on oy direction + Time-
History Acceleration “Vrancea North-South” on ox direction
One could notice that the shear force values concerning Vrancea North-South Time
History on ox direction is approximately the same with the shear force value of the
Vrancea North-South Time History on ox direction correlated with the Vrancea
East-West Time History on oy direction.
From the picture above it can be noticed that the shear force value of the Vrancea
North-South Time History on the oy direction is approximately 50% lower than the
shear force value of the Vrancea North-South Time History on the ox direction. In
the same time very close results appeared when only the Vrancea North-South
Time History was used on the oy direction compared to the model where Vrancea
North-South Time History was used on the oy direction and the Vrancea East-West
Time History was used on the ox direction.
Feigenmodes SRSS /y = 4834/0.25 = 19336 T (comparison with 13757 T the value for
the base shear force in case of design with Time History Acceleration).
As at conclusion, the calculation based by eigenmodes (Spectral Analysis) and Ψ=1
seems to be the worse case scenario in comparison with calculation based by Time
History Acceleration (38%).
So a simplified design with elastic response of the structure using a value of Ψ= 0.7
is enough for the design.
The maximum results are observed when Vrancea North-South Acceleration is
applied for x direction (13136T).
Also it can be said that applying in the same time Vrancea East-West Acceleration
and Vrancea North-South Acceleration brings insignificant increase of the results.
This fact is known in the case of Vrancea Acceleration that using simultaneous
North-South and East-West doesn't bring important increase.
A particularity for this structure is that seismic action is more important in case of
considering Vrancea North-South accelerogram for x direction instead of using
Vrancea North-South accelerogram for y direction (15%).
This is due to the fact that the structure is more rigid on ox direction (the mode 1 is
on oy direction). The period in mode 1 is 5.44sec (on oy direction) and 3.71sec (on
ox direction). That means that for the typical Vrancea earthquake the seismic
response is smaller if the structure is more flexible.
4.2 Eigenmodes Analysis
By using the Time History Acceleration VNSXEWY the value for Shear Base
Force results 8839 tf.
Although in theory it is said that for linear calculations the results must correspond
between Time History Analysis (laborious calculation) and Modal Analysis
Method (much easier calculation) for the studied structure the differences are major
(8839tf comparative with 13700tf).
5. CONCLUSIONS
The structure that we analysed has a self vibration period of 3.03 sec corresponding
to a translation on y direction. The disposition of structural elements could not be
done symmetrically in order to have identical vibration periods on x and y
directions. Because of that more detailed dynamic analysis is needed. For that
reason in this paper we used more calculation methods: the spectrum analysis
method and the calculation with accelerograms.
Concerning the spectrum analysis method two conclusions can be mentioned:
Even if in the specific literature it is suggested that for flexible structures the
influence of the 1st vibration mode is not the most important, from what we
analysed the 1st vibration mode proved to be the most significant (as we mentioned
80-82%).
No matter what discrepancies can be in the academic environment concerning the
different correct methods of composing for the vibration modes we did not find
significant differences. For an exact calculation the use of the accelerograms
method is recommended.
Concerning the analysis using accelerograms and the procedure of direct
integration the conclusions are:
The calculation version using the vibration modes (spectrum analysis) and Ψ= 1 is
more on the safe side than the accelerograms calculation (38%).
It is noticed that the maximal results appear when the accelerogram s North-South
Vrancea is used on x direction (13136 T).
We can also say that the simultaneous application of the accelerograms Vrancea
Est-Vest and Vrancea Nord-Sud does not bring significant increase on the results
(1%).
This is known that concerning the Vrancea accelerogram using N-S and E-W does
not bring significant increase of the results.
The seismic action is more important when we introduce the North-South Vrancea
accelerogram directly on x direction, instead of y direction. The difference is
around (15%). This is due to the fact that the structure is more rigid on the x
direction (1st vibration mode on y direction) . The 1st vibration mode has a period
of 4.55 sec. on y and 3.71 sec. on x direction. The result is that for the seismic
action type Vrancea the seismic action is smaller if the structure is more flexible.
Even if in theory it is said that in a linear calculation there should be no differences
between the accelerogram calculation method using direct integration (elaborated
calculation) and the vibration modes calculation (a more simple calculation) for
the structure that we analysed the differences are significant 8839 instead of 13700
Tf.
Using the analysis described above it resulted that the consideration for the design
of the wind forces is not necessary because the resultant of the wind forces is
smaller than the seismic action (5400 instead of 13700 Tf).
Using dynamic analysis of the structure in this paper it can be recommended that
the analysis of the structure to be made in the elastic domain using the
accelerogram method by direct integration without the danger of
overdimensionning (the base shear force using accelerograms is estimated at 15.7%
of the weight of the entire structure).
References
1. Coveianu, M. Contribuţii la calculul geometric neliniar al structurilor metalice alcătuite din
bar(Teza de Doctorat), Universitatea Tehnica de Construcţii, Bucureşti, 1998.
2. Bănuţ, V. Calculul neliniar al structurilor, Editura Tehnică, Bucureşti, 1981.
3. Mazilu, P. Statica construcţiilor vol.1, E.S.A.C., Bucureşti, 1955.
4. Mazilu, P. Statica construcţiilor vol.2, Editura Tehnică, Bucureşti, 1959.
5. Pătrâniche, N., Siminea, P., Chesaru, E. Construcţii metalice, Editura Didactică şi Pedagogică,
Bucureşti, 1982.
Wind Dynamic Response of a Suspension Pedestrian Bridge

Oana Mihaela Stan
Department of Mechanics, Statics and Dynamics of Structures, Technical University of Civil
Engineering, Bucharest, Zip code 72302, Romania
Summary
The paper presents wind dynamic response of a suspension pedestrian bridge.

Dynamic response of the structure was obtained by performing two analyses:
modal superposition (linear analysis) and direct integration of equation of motion
(geometric nonlinear analysis). The aerodynamic instability of the suspension
structure was studied too. All the analysis was carried out by LUSAS (FEM)
program.
KEYWORDS: Geometric nonlinear analysis, Dynamic response
1. INTRODUCTION
After the disaster of Tacoma Narrows Bridge in 1940 attention focused on the
dynamic problems of suspension bridges. The great span length of suspension
bridges makes their static and dynamic behavior under the action of lateral forces
an important problem. The most significant forces are due to wind and to the
earthquakes.
In order to obtain a correct response of the structure subjected to dynamic loads it
is absolute necessary to carry out nonlinear dynamic analysis.
Time history response of the suspension bridge will be obtained in two load cases:
mean velocity wind action and for gust of wind.
2. PEDESTRIAN SUSPENSION BRIDGE
Pedestrian Suspension Bridge was designed as a link between Burjuc and Tisa,
over Mures River.
The superstructure consist in a steel continuous lattice girder over three spans
(40,00 m + 120,00 m + 40,00m) (Figure 1.).
770 O.M. Stan
The deck is a triangular system with 1,55 m high and 2,20 m width and stiffening
frame at superior and inferior sole.
The suspension system is formed by the main parabolic suspension cables and
vertical hangers. Each main suspension cable is it made from three steel carrying
cables with 60 mm in diameter. The vertical hangers are made from 6 round steel
bars with 21 mm in diameter, placed at 5,00 m one to another along the central
span.
The steel towers has A shape in longitudinal and transversal plane. The high of the
towers is 25,40 m. The towers are supported on two caisson indirect column
foundations.
Y
Z
X
Figure 1. Finite element model
3. FINITE ELEMENT MODEL
The structure has a pronounced spatially behavior. This is the reason that lead to a
discrete spatially model. The finite elements (3D beam) were chosen to allow
running nonlinear geometric analysis. Discrete model of the structure has 3088
nodal points and 1330 beam finite elements, [6].
4. STRUCTURE ANALYSIS
4.1. Static analysis
The geometric nonlinearity of the suspension system of the structure require static
nonlinear analysis (modified Newton – Raphson), [3].
The wind pressure acting in global direction Z is 2 kN/m2 according to the

standards.
A comparative study between linear and nonlinear geometric static analysis is
illustrated in Figure 2. Lateral flexibility of the suspension system is revealed by
the evaluation of the displacements of the main suspension cables in both linear
and geometric nonlinear static analysis.
Maximum displacement for the middle of central span

0,248 m – linear static analysis
0,246 m – geometric nonlinear static analysis
Maximum displacement for the cables

5,300 m – linear static analysis
1,206 m – geometric nonlinear static analysis
Figure 2. Maximum displacement for wind pressure
4.2. Modal analysis
The necessity of running a modal analysis result from the prescription of different
European codes for vibration response of the suspension structure, [1], [4].
The 3D discrete model was subjected to modal analysis is order to obtain the first
15 eigenvalues and mode shapes.
In Figure 3. are presented the most significant mode shape of the suspension
bridge: the first symmetric mode of lateral vibration of the deck (Mode No. 1), the
first anti-symmetric mode of vertical vibration (Mode No. 2), the first symmetric
mode of lateral vibrations of the towers (Mode no. 9) and the first mode of
torsional vibrations (Mode no. 11).
4.3. Wind effect according to different European prescription
Once eigenvalues and mode shape obtained from modal analysis, it can be use to
evaluate wind velocity corresponding to specific dynamic effects.
Not all the specific effects were reflected in European prescription. This is the
reason that wind critical values are evaluated only according to some prescription.
772 O.M. Stan
Evaluation of wind response involve studies about limited amplitude oscillations,

divergent amplitude oscillations and non-oscillatory divergent movements, [1], [4],
[8], [9].
Modal shape 1 T = 1,190 s
Figure 3. The most significant modal shapes

Conclusion: Pedestrian suspension bridge present aerodynamic stability for wind

turbulence, for von Karman gust and flutter effect (the most dangerous dynamic
effect of the wind acting on suspension bridges), [6].
4.4. Resonance amplitude
The first 15 natural frequencies of the structure were in the range of 0,00 – 6,00
Hz. Wind response will be studied in the same range of frequency.
The structure response is presented for two nodal points which offer the most
significant information.
For the tower - the maximum lateral displacement of a nodal point lead to the top
of the tower.
For the deck - the maximum lateral displacement of a nodal point corresponding to
the middle of the central span.
The structure response in frequency domain due to wind excitation is presented in
Figure 4. and Figure 5.
The resonance amplitude of 4,39 Hz at the top of the tower is related to the first
natural frequency of symmetric lateral vibration mode (Mode no. 9). In the same
way, the resonance amplitude of 0,84 Hz at the middle of the central span is related
to the first natural frequency of symmetric lateral vibration mode (Mode no. 1).
The amplitude in X and Y global directions can be neglected instead of Z global
direction.
Figure 4. Resonance amplitude for the top of the tower due to wind excitation
774 O.M. Stan
Figure 5. Resonance amplitude for the middle of the central span due to wind excitation
4.5. DYNAMIC WIND ANALYSIS
4.5.1. Wind mean pressure
The real wind forces were considered acting in one step. Time history response of
the structure is presented in Figure 6 and Figure 7 for the same nodal point (top of
the tower and middle of the central span). Lateral displacements of nodal points
oscillating around the same values obtained in static analysis for wind pressure.
Figure 6. Time history response in displacements for the top of the tower.
Figure 7. Time history response in displacements for the middle of the central span.
4.5.2. Gusty wind
The real wind velocity is permanently changing in time and space. A result of
different study and measurements on wind velocity is illustrated in Figure 8.
Figure 8. Wind velocity space and time distribution
Dynamic response on wind excitation can be obtained by experimental or by

analytical approach.
776 O.M. Stan
Mathematical method can be a stochastic approach in frequency domain –

unacceptable for suspension bridges (cannot evidence the flutter effect) or a
deterministic approach in time domain – which require time history of air flow in
time and space for direct integration of equation of motion at different moment of
time.
Suspension bridge was subjected on a deterministic approach. Air flow was
modeled in time steps of 0,05 s for 12,0 s range (Figure 9). The 12,0 s range is long
enough to cover the dynamic response of the structure which has first natural
period 1,19 s.
Figure 9. Wind velocity time distribution
Space air flow was neglected – considering constant pressure of the wind along the
bridge.
Figure 10. Time history in acceleration of the top of the tower in global direction Z shows a
maximum acceleration of 0,12g.
Dynamic analysis include 12,0 s of wind pressure and 3,0 s without any excitation.
For gusty wind it was defined a specific load function which multiplied static wind
load for all the members of the structure (Figure 9.).
Time history response of the structure is illustrated in Figures 10 - 13 for

different quantities.
Figure 11. Time history in acceleration of the middle of the central span in global direction
Z shows a maximum acceleration of 0,15g.
Figure 12. Time history displacement shows an oscillation around displacement value
evaluated in static analysis for top of the tower.
Figure 13. Time history displacement shows an oscillation around displacement value
evaluated in static analysis for middle of the central span.
778 O.M. Stan
5. CONCLUSIONS
The geometric nonlinearity of the suspension system requires static nonlinear

analysis.
Dynamic response of suspension bridges is highly influenced by the damping ratio.
Choosing a wrong value for damping ratio can corrupt the response.
Geometric nonlinearity has to be taken into account for dynamic analysis too.
Direct integration of equation of motion is the solution for flexible structure.
The pedestrian suspension bridge is well conformed from static and dynamic point
of view.
References
1. W. C. ARROL - Implications of the Rules on Bridge Design - Past and Future, Bridge
Aerodynamics, London, March, 1981.
2. J. M. W. BROWNJOHN - Estimation of Damping in Suspension Bridges - Proc. Instn. Civ. Engrs
Structs & Bldgs, Vol. 104, Nov., 1994.
3. M. A. CRISFIELD - Nonlinear Finite Element Analysis of Solids and Structures, Vol. 1 -
Esentials; John Wiley & Sons, 1997.
4. EUROCODE 1: Basis of Design and Actions on Structures, 1994.
5. D. GHIOCEL, D. Lungu - Acţiunea vântului, zăpezii şi a variaţiilor de temperatură în construcţii,
Ed. Tehnică Bucureşti, 1972. (in Romanian)
6. O. M. STAN - Efecte dinamice specifice la poduri de mare deschidere – Teză de doctorat,
U.T.C.B., 1998. (in Romanian)
7. Shigeki UNJOH, Yukio ADACHI - Damping Characteristics of Long - Span Bridges, IABSE
SYMPOSIUM, KOBE 1998 (Long - Span and High Rise Structures), p.335-340.
8. T. A. WYAT - The Effect of Wind on Slender Long - Span - Bridges, Symposium on Wind
Effects on Buildings and Structures, Paper 36, London, 1984.
9. T. A. WYATT, C. SCRUTON - A Brief Survey of the Aerodynamic Stability Problems of
Bridges, Bridge Aerodynamics, London, March, 1981.
The modeling, simulation and numerical analysis of a fire in a

construction with light structure
Ion Anghel
"Alexandru Ioan Cuza" Police Academy – Fire Officers Faculty
Summary
This paper presents a numerical simulation of a light construction structure by

using of the high level dedicated software Fire Dynamics Simulator (FDS). The
important temperatures, Heat Release Rate, Total Heat Flux were computed along
with the numerical fire visualization.
Keywords: Fire Model, Fire simulation, Numerical simulation, Heat Release Rate,
Fire visualization.
1. INTRODUCTION
The present study refers to the modeling, simulation and numerical analysis of a
fire developed in a construction with a light structure. The research had as starting
point a practical requirement the realization of a prototype with a light structure
resistant to fire. In this idea tests were made on the structural model chosen, on a
numerical way with the help of the advanced platform “FIRE DYNAMICS
SIMULATOR” (FDS) and also experimental (by ICECON S.A.) by the effective
building of a scale model and the preparing of it for the fire test.
The experimental results obtained were concretized by temperature measurements
in divers points on the burning structure. The measurements were executed at
ICECON S.A., in the context of a mutual research contract with a part of the staff
from The Firefighters Faculty.
2. THE MODELING, SIMULATION AND THE NUMERICAL

ANALYSIS OF THE FIRE
The numerical analysis was made on a geometry with a light structure identical
with the one obtained experimentally, observing the way of fire behavior of it and
the underlining of 20 numerical monitoring points of the temperature on a
780 I. Anghel
determined time interval. The structure and the calculus network utilized for the
discretization of differential equations of the fire is shown in figure 1.
For the numerical modeling of the fire the “field zoning” type was adopted or more
precise the tridimensional discretization of the transitory characteristically
equations. The 3D modeling of the fire can be shown in principal by the input
values and the output values:
Input values:
¾ The detailed form of the structure, the construction of the
chambers, including walls, the ceiling and the floor, the number of
air vents and their sizes (if necessary, etc;
¾ The characteristics of the furniture (if necessary), the
characteristics of the fuel, the parameters of buoyancy and
radiation.
Output values:
¾ The thermal flux, the movement and speed of smoke;
¾ Temperatures measured in divers points of the structure;
¾ Previsions on the time until sprinkler and heat detectors activation
(if necessary), the time to flashover;
Figure 1. The geometry of the light structure and the calculus network
A FDS simulation determines the thermal behavior of the whole structure by

calculating the thermodynamically quantities in each cell of the discretization. The
number of quantities previsioned is greater then the number of quantities that are
normally measured in validation experiments.
The following is a list of thermodynamically values that will be obtained for every
cell, list that can be compared with the experimental measurements. Their order
underlines their importance.
¾ The temperature of the gas;

¾ The speed of the gas;
¾ The concentration of major gases ;
¾ The concentration of smoke/ obscurization;
¾ The pressure.
For solid surfaces, the FDS results are usually:
¾ The surface temperature;
¾ The total thermal flux (by conduction, convection and
radiation).
These global quantities are used to evaluate the total energy budget, the
confirmation of the conservation of the total energy in the model, and the
evaluation of the incertitudiny of the values obtained experimentally.
FDS resolves the mass conservation equations, of the momentum and energy for a
multi-component, thermal expandable mixture of ideal gases, like:
Mass conservation
∂ρ (1)
+ ∇ ⋅ ρu = 0
∂t
Component conservation (species)

∂
(ρYl ) + ∇ ⋅ ρYl u = ∇ ⋅ ρDl ∇Yl + m& l′′′ (2)
∂t
Momentum conservation
 ∂u  (3)
ρ + (u ⋅ ∇ )u  + ∇p = ρg + f + ∇ ⋅ τ
 ∂t 
Energy conservation
∂
(ρh ) + ∇ ⋅ ρhu = Dp − ∇ ⋅ q r + ∇ ⋅ k∇T + ∑ ∇hl ρDl ∇Yl (4)
∂t Dt l
To note that the extern force of the fluid, represented by the term f in the above
equations, is the friction exercised by the water drops from the sprinklers and by
other external forces. The term Dp = ∂ρ + u ⋅ ∇p represents a material derivate. The
Dt ∂t
conservation equations are accompanied by a state equation that links the
thermodynamically quantities. It is used an approximation of the characteristically
equation of the ideal gas by decomposing the pressure in a “background”
component, a hydrostatic component and a perturbation induced by the flow:
p = p0 − ρ∞ gz + ~
p (5)
782 I. Anghel
For application p 0 is a constant and the other two components are relatively small.
This assumption can be modified if the pressure rises due to the fire in a small an
isolated compartment, or when the height of the domain is very high and p 0 can no
longer be considered a constant and must by considered a function of altitude.
The scope of decomposing the pressure is that for the flows with a small Mach
number the temperature and density can be approximated with the state equation:
p0 = ρTR ∑ (Yi / M i ) = ρTR / M (6)
These equations (1-6) are simultaneous resolved in time (transient regime) for all
the calculus cells assimilated to the light structure.
The geometry of the structure together with the location of the 20 point where the
temperature is measured is shown in figure 2:
a) b)
Figure 2. a)-b) The geometry and the positioning of the points of numerical evaluations of
the temperature
The location of the 20 calculus points of the temperature can be easier represented
by groups of 5 points, in Figure 3:
a) T1 ÷ T5 Thermocouples interior ceiling:

784 I. Anghel
b) T6 ÷ T10 Thermocouple ceiling:

c) T11 ÷ T15 Thermocouples walls:

786 I. Anghel
d) T16 ÷ T20 Thermocouples door:
Figure 3. a)-d) The variation of temperature in the 20 considered points

It is observed that for all the thermocouples the value of the temperatures at that
location was calculated, in the interval 0...250 [s] and the graphics of temperature
variations were created. The obtained values indicate a very high level of the
temperature on in inside ceiling and that it drops in order: wall, door, and ceiling.
The thermal flux by conduction, convection and radiation, respective the total
thermal flux is given in figure 4 a)-d):
a)
b)
c)
d)
Figura 4. a)-d) The thermal flux (conduction, convection, radiation, total)
788 I. Anghel
The rate of burning is shown in figure 5.
Figure 5. The burning rate
In figure 6 some instantaneous numerical visualizations of the evolution of the fire

in the light structure are shown:
a) b)
c) d)
Figure 6. a)-d) The numerical visualization of the fire at diverse moments in time
3. CONCLUSIONS
In this paper the modeling, simulation and the numerical analysis of the fire
behavior of the small scale model of a construction with a light structure. The
values of the temperatures obtained, the thermal flux, the burning rate and the
numerical visualization of the fire were later compared with the experimental ones,
obtaining very good accordance.
The modeling, simulation and the numerical analysis of fires can be utilized with
success in the following sub- domains of fire security engineering:
- Regenerates of some fires based on the fire’s print, the cause of the fire and
other data (the characteristics of the materials, the way the firefighters
approached the localization and extinguishing of the fire);
- The electronic resolve of the tactical situations;
- The building of fire security scenarios for buildings;
- Providing the necessary parameters for the fire intervention through numerical
modeling.
References
1. Anghel I., Darie E. şi Flucuş I. - „Evaluarea securităţii la incendiu a utilizatorilor unei clădiri
rezidenţiale – partea a II-a ” Revista Electricianul nr. 5/2006, Bucureşti, pag. 42-47, ISSN 1223-
7426, Revista Instalatorul nr. 7-8/2006, Bucureşti, pag. 10-15, ISSN 1223-7418. (in Romanian)
2. Anghel I., Darie E. - „Modelarea transferului termic în condiţii de incendiu într-o încăpere
închisă” a X-a sesiune de comunicări ştiinţifice cu participare internaţională „Leadership şi
management la orizonturile secolului al XXI-lea” a Academiei Forţelor Terestre „Nicolae
Bălcescu”, Sibiu 24-26 noiembrie 2005, ISBN 973-7809-29-7. (in Romanian)
3. Darie E., Anghel I. şi Popescu G. - „Evaluarea fenomenelor cuplate de transfer de căldură ” A
IX-a Sesiune de Comunicări Ştiinţifice cu Participare Internaţională a Academiei de Poliţie
“Alexandru Ioan Cuza” – Facultatea de Pompieri „SIGPROT 2006”, Bucureşti, 26 Mai 2006, pag.
133-138, ISBN 973-718-513-7. (in Romanian)
4. McGrattan K. and Forney G. - "Fire Dynamics Simulator (Version 4) User's Guide", NIST
Special Publication 1019, Fire Research Division, Building and Fire Research Laboratory in
cooperation with VTT Building and Transport, Finland , NIST, 2006.
5. McGrattan K. B. - "Fire Dynamics Simulator (Version 4) Technical ReferenceGuide", NIST
Special Publication 1018, Fire Research Division, Building and Fire Research Laboratory in
cooperation with VTT Building and Transport, Finland , NIST, 2006.
6. Forney G. P. and McGrattan K. B., -"User's Guide for Smokeview Version 4 – A Tool for
Visualizing Fire Dynamics Simulation Data", NIST Special Pub. 1017, NIST, 2004.
7. Contract AMTRANS Nr. 7B07/2004, „Sisteme eficiente de realizare a unor structuri uşoare
rezistente la foc pentru construcţii civile”. (in Romanian)
Capacity based earthquake design of reinforced concrete shear

wall according to EC8
Laura Fraifeld1, Ramona Agachi1, Horst Werkle2, Franz Zahn2
and Gabriela Atanasiu1
1
Faculty of Civil Engineering, Technical University «Gh. Asachi» Iasi, Roumania
2
Faculty of Civil Engineering, University of Applied Science, Konstanz, Germany
Summary
Any structural design can be considered satisfactory if it meets the three principal
goals of structural performance: safety, economy and simplicity of construction.
The first two require a sufficient margin of safety against collapse. Also, the
construction should be serviceable under working load conditions through its life
time.
The topic of this paper is the analysis of the seismic response of reinforced
concrete building using displacement based design model. For this purpose a
typical Romanian residential building has been considered as case study. First, the
equivalent static loads acting on the shear walls was determined by using simple
hand calculations taking into account linear elastic behavior. An analysis of the
most critical shear wall for the determined lateral forces was made by using two
different nonlinear analysis programs for comparison, namely the program ADINA
[9] and the program ATENA 2D [10]. Seismic demands estimated in this structure
using an equivalent single degree of freedom system derived from the first
fundamental mode of vibration of the building.
KEYWORDS: safety against collapse, serviceability, seismic response, linear

elastic behavior, nonlinear analysis, seismic demands, single degree of freedom
system
1. INTRODUCTION
The design of structures for earthquake resistance in engineering practice is

generally based on a response spectrum analysis. The ductility of the structure is
considered by a behavior factor depending on the construction type of the building.
The present simplified approach neglects all the redistribution of the sectional
forces due to yielding of structural elements and other nonlinear effects occurring
during an earthquake. A rigorous nonlinear computation, how-ever, requires a time
history analysis of the building including the effect of material non-linearity. This
type of analysis is time consuming and the interpretation of the results is not
always obvious.
Capacity design represents a method developed for practical earthquake design,
including the overall nonlinear behavior of the structure. It implies some major
simplifications, mainly modeling the structure as a single degree of freedom
system.
The analysis of a reinforced concrete building according to the capacity design
consists in a conventional design by the response spectrum method. The nonlinear
behavior of reinforced concrete walls in pushover analysis has been modeled by
plane stress elements in the finite element program ADINA [9]. The following
steps have to be accomplished:
- -Design of the reinforcement by a RESPONSE SPECTRA ANALYSIS
(RSA)
- -Determination of the safety factor according to the capacity design for the
reinforcement for response spectra analysis by using ADINA[9] and
MathCAD[11].
2. ANALYSIS OF A REINFORCED CONCRETE WALL USING A

SHELL MODEL
2.1. Numerical analysis of reinforced concrete wall. The design of the

structural walls
In this work a structural reinforced concrete wall has been analyzed by
means of more methods and then the results are compared. The computation
of the element took into account a structural wall from the first storey of a
building P + 4, denoted Structural Wall TW1.
In the process of design using the multi-storied one-bay frame mode, the horizontal
afferent loads to the structural wall had the value q1, y = 3228.08 daN/m.
2.2. Numerical analysis of reinforced concrete walls using ADINA Software
The numerical simulation has been performed using a 3D element acted on top by a
force equal to 400000N, at level 3: 0.75 x 400000N, at level 2: 0,5 x 400000N and
at level 1 : 0,25 x 400000N. The reinforcement consists of 18ø16 bars disposed on
both lateral sides of the wall (9 bars on a part and 9 on the other part). The element
collapses at a load of 328200 N. The Program ADINA [9] determines the
792 L. Fraifeld, R. Agachi, H. Werkle, F. Zahn, G. Atanasiu
maximum capacity of the element by several iterations based procedure. The

maximum displacement obtained is 0.004675 m.
Fig.1. Structural Wall TW1 Topology
2.3. Computation of the Structural Wall TW1 using ATENA 2D Software
For the numerical experiments of the wall we use as materials the concrete class
C25/30 and reinforcement steel S550. The element was introduced as a
bidimensional element having the geometrical configuration presented in Figure 2.
The horizontal and vertical loads are linearly applied in order to avoid the
singularity of stresses and displacements.
As horizontal load a load of value of 2500 KN has been applied in the following
manner: 10% of the load for the first floor, 15% for the second one, 20% for the
third, 25% for the fourth and 30% for the fifth floor in the direction of the line that
represents the axis of the slab at the level.
The obtained values are then divided by the length of the wall, respectively of
12.3m, being expressed in [MN/m]. The vertical load is also considered, being
linearly applied, thus the values obtained are divided by the length of the wall.
Fig.2. Geometrical Model of the Element
Fig.3. Applying the load on the member

There have been introduced 30 steps with the step multiplier equal to 0.1. In
accordance with the Output Data given by ATENA Software, the load capacities of
the member with their corresponding displacements have been deducted and are
given in Table 1.
Table 1: Load Capacities of the Structural Wall TW1 and the corresponding displacements
Forces Displacement Forces Displacement
KN m KN m
250 1,27E-04 4000 3,30E-03
500 2,55E-04 4250 3,62E-03
750 3,82E-04 4500 3,85E-03
1000 5,10E-04 4750 4,60E-03
1250 6,38E-04 5000 4,90E-03
1500 7,66E-04 5250 6,02E-03
1750 8,94E-04 5500 7,31E-03
2000 1,02E-03 5750 9,37E-03
2250 1,39E-03 6000 1,24E-02
2500 1,56E-03 6250 1,73E-02
2750 1,88E-03 6500 2,97E-02
3000 2,07E-03 6750 6,85E-02
3250 2,26E-03 7000 1,13E-01
3500 2,67E-03 7250 1,97E-01
3750 3,08E-03 7500 3,26E-01
3. ANALYSIS OF A REINFORCED CONCRETE WALL USING A

BEAM MODEL
3.1 The moment-curvature analysis
The plastic moment capacity of all ductile concrete members is computed using the
moment curvature (M- Φ) analysis based on expected material properties shown in
Figure 4.
Moment-curvature analysis derives the curvatures associated with a range of
moments for a cross section based on the principles of strain compatibility and
equilibrium of forces.
The (M- Φ) curve can be idealized with an elastic perfectly plastic response to
estimate the plastic moment capacity of a member’s cross section. The elastic
portion of the idealized curve should pass through the point marking the first
reinforcing bar yield.
The idealized plastic moment capacity is obtained by balancing the areas between
the actual and the idealized (M- Φ curves beyond the first reinforcing bar yield
point.
Fig.4. Moment-Curvature Representation
3.2 Determination of plastic hinge
The plastic hinge length is required to convert the moment-curvature relationship

into a moment-plastic rotation relationship for the nonlinear pushover analysis.
The plastic rotation, θP, can be determined from the following Equation 1, by using
moment-curvature analysis and applicable strain limitations. The plastic rotation is:
θ p = L pφ p = L p (φ m − φ y )
(1)
Where: Lp is the plastic hinge length, Φ p, the plastic curvature, Φm, the maximum
curvature and Φ y the yielding curvature.
3.3 Beam bending
The wall TW1 has been loaded with a force N equal to 2023.5 KN. In the Figure 5
the cross-section of the shear wall TW1 is shown, as an extract from FAGUS-4
Software.
Fig.5. Cross-section of the shear wall TW1
The relationship (M- Φ) has been obtained by using the software FAGUS-4 [12]
shown in Fig. 6.
Mz [kNm]
• -2023,5
∆
80000
64000
∇
48000
32000
16000
0,0002 0,0006 0,0010 0,0014 0,0018

χz [m -1]
Fig.6. Moment-Curvature Diagram

From the moment-curvature diagram, trying to equal the area cut off and added, the
following results for the ultimate moment, yielding moment and yielding curvature
are obtained: Φu = 0.0020 m-1, Φy= 0.00048 m-1 and My = 84000 KN*m.
3.4 Displacement computational aspects
The base shear force of the wall resulted is: Vb = 8000 KN. The actual moment
distribution will be somewhere between parabolic and linear, and a better
approximation would be to use the average of 1/4 (for parabolic) and 1/3 (for linear
curvature distribution). The length of the plastic hinge is: LP=1.52 m.
The displacement corresponding to the yielding moment is determined using
Equation 2:
h
⌠
∆ y :=  φ y ( x) ⋅ M ( x) dx (2)
⌡
0
After yielding the wall will be very heavily cracked and therefore there will also be
a plastic (additional) shear displacement.
The plastic curvature is computed with the following equation:
p= Φu- Φy (3)
The plastic rotation of the plastic hinge is:
θp=Lp* Φp (4)
The plastic displacement is given by the following relationship:

θ∆p=θp*(h-Lp/2) (5)
Table 2: Displacement values of the shear wall TW1

No. Type of displacement Displacement value (m)
1 Yield displacement(∆y) 0.02744
2 Shear displacement (∆shear) 0.0044
3. Total yield displacement (∆yieldtot) (1+2) 0.03184
4 Additional shear displacement (∆asdy) 0.035
5 Plastic displacement (∆p) 0.03059
6. Ultimate displacement (∆u) (3+4+5) 0.09743
4. RESULTS INTERPRETATION
For the capacity curve of the wall, a bilinear approximation has been used
considering a linear elastic part up to the point where the shear capacity of the wall
Vb. This point was reached and a perfectly plastic part with zero stiffness. The
bilinear capacity curve of a wall has been defined by three parameters, the shear
capacity of the wall Vb, the nominal yield displacement ∆y and the nominal
ultimate displacement at the top of the wall ∆u. The comparison made with the
results given by ATENA Software is graphically represented in Fig.7.
Load-Displacement Diagram
9000
8000
7000
Force [KN]
6000
Load Displacement obtained by using
5000 Atena Softw are
4000 Load Displacement obtained by using
Fagus Softw are and hand computations
3000
2000
1000
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Dislacements [m]
Fig.7. Numerical interpretation of displacement – shear force relationship
5. FINAL REMARKS
Being a benchmarking methodology, the paper presented the comparison between

two methods of analysis of the shear wall nonlinear behavior.
Using ADINA [9] the maximum displacement obtained for the base shear force
equal to 328200 N is 0.004675 m. By using ATENA [10], for a base shear force of
7500 KN it results a displacement of 0.0326 m.
Considering the 2nd method based on FAGUS[12] software and hand
computations, the ultimate displacement obtained for a shear force equal to 8000
KN is equal to 0.09743 m.
The difference of numerical results between the displacements obtained in the
analyses made with ATENA [10] and the displacements by means of the second
method consists in the methodology accuracy. In the first case, the base shear force
is of a smaller value than the value obtained using the second method.
ADINA [9], a very complex program, was used for analysis of structures using
many fields of study, while ATENA [10] is specialized in analysis of construction
works. By comparing the numerical results obtained, one considers that ADINA
[9] is a program with insufficiencies regarding the civil engineering field. ATENA
[10] is therefore easier to approach and for purpose of obtaining numerical results
following a simple analysis of structures, as the case of the analyzed concrete wall.
ADINA [9] is more suitable for very complicated structures, as in the case of the
San Francisco Oakland Bay Bridge or the performed crush analysis test of
automobiles for Ford Windstar. Using ATENA [10] software one can analyse only
bi-dimensional elements, while with ADINA [9] is suitable to perform 3D analyses
of various structural elements. The software gives appropriate results, illustrated by
the moment – curvature analysis obtained in the final step of computation.
As a final conclusion, the results of the numerical experiments done with ADINA
[9], FAGUS [12] and MATHCAD [11] software shows give the corresponding
results of load – displacement curve for the analyzed structural reinforced concrete
wall. The main difference consists in the way the software is suitable for numerical
simulations in the construction field analyses.
References
1. Werkle, H., Finite Elemente in der Baustatik, Statik und Dynamik der Stab- und
Flächentragwerke, Vieweg Verlag 2001
2. Werkle, H, Avak, R., Mathcad in der Tragwerksplanung , Vieweg Verlag 2002
3. Atanasiu, G., Structural Dynamics, 2nd version, Vasile Goldis University Press, Arad ,2005
4. Atanasiu, G., Vlad, I. Bratianu, C., Finite Elements in Structural Analysis, Cermi, Iasi, 2005
5. Bindseil, P., Massivbau Bemessung im Stahlbetonbau, Vieweg Verlag 2005
6. Schneider Bautabellen, Bautabellen für Ingenieure, Auflage 2004
7. Wommelsdorf, O., Stahlbetonbau Teil 1: Grundlagen – Biegebeanspruchte Bauteile, Werner
Verlag, 2004
8. Wommelsdorf, O., Stahlbetonbau Teil 2: Stützen und Sondergebiete des Stahlbetonbaus
Bemessung und Konstruktion, Werner Verlag, 2004
9. ADINA, Finite Element Software, Registered Trademark of K.J. Bathe/ADINA R & D Inc.
USA, version 2002
10. ATENA 2D, Finite Element Software, version SBETA, 2001
11. Mathcad, Mathematical Computations Software, version 2001 Professional
12. FAGUS Software, Statical Computations Software, version 4, 2004
Appendix
• Finite Element Analysis – a computer simulation technique used in engineering

analysis, which uses a numerical technique called the Finite Element Method
(FEM).
• Finite Element Method (FEM) – a structural mechanics method in which the
object is represented by a geometrically similar model consisting of multiple,
linked, simplified representations of discrete regions – i.e. finite elements.
• Response Spectrum – a plot of the peak or steady state response (displacement,
velocity or acceleration) of a series of oscillators of varying natural frequencies,
that are forced into motion by the same base vibration.
• Modal Analysis – represents the study of the dynamic properties of structures
under vibration excitation.
• Pushover analysis - is a non-linear static analysis carried out under conditions of
constant gravity loads and monotonically increasing horizontal loads.
• Linear Dynamic Analysis – is the analysis in which the response of the structure
to the ground motion is computed in the time domain, and all the phase information
is therefore maintained.
• Moment curvature analysis - derives the curvatures associated with a range of
moments for a cross section based on the principles of strain compatibility and
equilibrium of forces.
• The length of plastic hinge - required to convert the moment-curvature
relationship into a moment-plastic rotation relationship for the nonlinear pushover
analysis.
ISBN 978-973-8955-14-1
M
S
MANIFESTARI STIINTIFICE

CCE2007 Proceedings

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CCE2007 Proceedings

Uploaded by

Copyright:

Available Formats

CIVIL ENGINEERING

C. Ionescu, F. Paulet-Crainiceanu, R. Scinteie, A. Nicuta

EDITURA SOCIETATII ACADEMICE "MATEI - TEIU BOTEZ"

C. Ionescu, F. Păuleţ-Crăiniceanu, R. Scînteie, A. Nicuţă

EDITURA SOCIETĂŢII ACADEMICE "MATEI-TEIU BOTEZ"

Descrierea CIP a Bibliotecii Naţionale a României

Civil engineering / ed.: C. Ionescu, F. Păuleţ-Crăiniceanu,

I. Ionescu, Constantin (ed.)

Computerized processing: F. Păuleţ-Crăiniceanu, R. Scînteie

1. Păuleţ-Crăiniceanu, F., Ionescu, C., Scinteie, R., Nicuta, A.M.

Civil Engineering, a domain based on computation

In Civil Engineering, the computation is at the base of almost all activity. An

KEYWORDS: Civil Engineering, computers, research, education, computation in

In Civil Engineering, as in other engineering fields, information processing is vital.

2. COMPUTER-BASED CIVIL ENGINEERING

considered: the underground structure in rock mass with one-sided dip of

graphical representations and introduce animations, and also to solve practical

Ciobanu, Silisteanu and Radinschi [31] are presenting a project focussed on

Non-linear effects in traffic simulations

Growing requirements on transport of passengers and goods have a principle

KEYWORDS: non-linear dynamics; traffic flow; congestions; traffic modeling;

1. TRAFFIC CHAOS AS A TECHNICAL TERM

Nonlinear dynamics as a tool for understanding chaos emergency already gets

2. REAL TRAFFIC DYNAMICS OBSERVING

2.1. Measuring of the traffic variables

2.2. Measurement with induction loops

Figure 2. Location of induction loops on the part of A5-Nord highway [5].

2.3. Free flow and congestions

2.3. Fundamental diagrams

3. TRAFFIC FLOW SIMULATIONS

It is possible to see dynamic properties of traffic also through the numerical

3.1. Car following models

Vehicle-following models, exactly Car Following Models (CFM) are investigated

3.2. Model simulations

For the simulation itself we have created a program environment in Java

3.2.1. System visualization

Direct system animation, when a movement of each vehicle is animated by an

Another type of visualization of system development could be the local density

3.2.2. Obtained results

The main question could be the shape of fundamental diagram. In simulations we

Figure 9. Dependency of simulated traffic intensity on average traffic density.

3.2.3. Non-linear phenomena

The basic simulation experiment was performed on a system with a maximum

Figure 11. Comparison of dependency of velocities congestion on average density at

– 53 vehicles/km at least two stable situations exist side by side – homogeneous

Influence of deformation at a heritage building support on

Masonry arch stability criteria definition has been developed by numerical

Figure 1. North front of Riga Cathedral

developed to define safety exploitation of deformed structure. Proposed

2. MASONRY COMPUTING EVOLUTION

2.1. Geometrical approach

2.2. Elastic analysis

2.3. Kinematic approach and arch collapse mechanism

Hinge forming by loading, thermal variations or geometric changes is most of arch

3. MASONRY MATERIAL APPROACH

In present research are proposed two step computational approaches. A complete

3.1. Macro modeling approach

Homogenizations of material included in most of building codes give simplify

Compression elasticity modulus E, [MPa] for solid clay brick

909 909 909

Figure 3. Three point brick bending test results

3.2. Micro modeling approach

Micro modeling of masonry material is precise stone material interaction

4. MONITORING DATA COLLECTION