Zangeneh 2018 - Licentiate

Dynamic Soil-Structure
Interaction Analysis of
Railway Bridges
Numerical and Experimental Results
ABBAS ZANGENEH KAMALI
Licentiate Thesis, 2018

KTH Royal Institute of Technology
School of Architecture and the Built Environment
Department of Civil and Architectural Engineering
Division of Structural Engineering and Bridges
SE-100 44, Stockholm, Sweden
TRITA-ABE-DLT-182
ISBN 978-91-7729-713-0
© Abbas Zangeneh Kamali, 2018
Akademisk avhandling som med tillstånd av KTH i Stockholm framlägges till offentlig
granskning för avläggande av teknisk licentiatexamen fredagen den 13 april kl. 09:30 i
konferensrum M108, KTH, Brinellvägen 23, Stockholm. Avhandlingen försvaras på engelska.
In the memory of my father,
Bahram
Abstract
The work reported in this thesis presents a general overview of the dynamic
response of short-span railway bridges considering soil-structure interaction.
The study aims to identify the effect of the surrounding and underlying soil
on the global stiffness and damping of the structural system. This may lead
to better assumptions and more efficient numerical models for design.
A simple discrete model for calculating the dynamic characteristics of the
fundamental bending mode of single span beam bridges on viscoelastic
supports was proposed. This model was used to study the effect of the
dynamic stiffness of the foundation on the modal parameters (e.g. natural
frequency and damping ratio) of railway beam bridges. It was shown that
the variation in the underlying soil profiles leads to a different dynamic
response of the system. This effect depends on the ratio between the flexural
stiffness of the bridge and the dynamic stiffness of the foundation-soil
system but also on the ratio between the resonant frequency of the soil layer
and the fundamental frequency of the bridge.
The effect of the surrounding soil conditions on the vertical dynamic response
of portal frame bridges was also investigated both numerically and
experimentally. To this end, different numerical models (i.e. full FE models and
coupled FE-BE models) have been developed. Controlled vibration tests have
been performed on two full-scale portal frame bridges to determine the modal
properties of the bridge-soil system and calibrate the numerical models. Both
experimental and numerical results identified the substantial contribution of the
surrounding soil on the global damping of short-span portal frame bridges. A
simplified model for the surrounding soil was also proposed in order to define a
less complicated model appropriate for practical design purposes.
Keywords
Railway Bridge; Soil-Structure Interaction; Full FE modeling; Simplified modeling;

Coupled FE-BE modeling; Field measurements; Model updating
i
Sammanfattning
Arbetet i denna avhandling omfattar en översikt över den dynamiska

responsen av korta järnvägsbroar med beaktande av jord-bro interaktion.
Studien syftar till att identifiera jordens effekter på den globala styvheten
och dämpningen hos systemet. Från en praktisk synvinkel, kan detta
resultera i förbättrade antaganden och mer tillförlitliga och effektiva
numeriska modeller för praktiska ändamål.
För balkbroar på viskoelastiska stöd, en enkel diskret modell för beräkning
av den vertikala böjningsmodens dynamiska egenskaper togs fram.
Modellen användes för att studera effekten av fundamentets dynamiska
styvhet på den grundläggande egenfrekvensen och dämpningsförhållande
hos järnvägsbroar. Det visade sig att variationen i jordprofilerna kan leda till
väsentliga skillnader i systemets dynamiska respons. Detta beror på
förhållandet mellan brons böjstyvhet och fundamentets dynamiska styvhet
såväl som på förhållandet mellan jorden resonansfrekvens och brons
grundläggande frekvens.
Effekten av omgivande markförhållandena på den dynamiska responsen hos
plattrambroar studerades också, både numeriskt och experimentellt. För
detta ändamål har olika numeriska modeller (t.ex. FE modeller och
kopplade FE-BE modeller) utvecklats. Vidare har kontrollerade
vibrationsprovningar utförts på två fullskaliga plattramar för att bestämma
systemets modala egenskaper och kalibrera de numeriska modellerna. Både
experimentella och numeriska resultat identifierar den omgivande jordens
betydande bidrag på den globala dämpningen. I avhandlingen föreslogs
också en förenklad FE modell för den omgivande jorden för att uppnå en
mindre komplicerad modell lämplig för praktisk dimensionering.
Nyckelord
Järnvägsbro; Jord-bro interaktion; FE modellering; Förenklad modellering;

Kopplad FE-BE modellering; fältmätningar; Modell uppdatering
iii
Preface
The research work presented in this thesis has been carried out at the
Division of Structural Engineering and Bridges, Department of Civil
and Architectural Engineering at KTH Royal Institute of Technology.
The research has been financed by the Swedish Research Council
Formas. This support is gratefully acknowledged. The field
measurements presented in this thesis were partly financed by
Trafikverket and performed by Dr. Andreas Andersson, Mr. Stefan
Trillkott, Mr. Claes Kullberg and the author.
The research work has been conducted under the supervision of

Adjunct Prof. Costin Pacoste, to whom I am very grateful for his
precious guidance and encouragement in both academia and industry.
I would like to express my sincere appreciation to my co-supervisors,
Prof. Raid Karoumi and Prof. Jean-Marc Battini for giving me the
opportunity to work on this subject and for their constant support
during my studies. Special thanks to Docent John Leander for
reviewing the thesis and for his valuable comments.
I would like to thank ELU consulting company for giving me the

opportunity to continue my academic career as an industrial PhD
student. Much appreciation goes to my colleagues at KTH and ELU
for providing such a pleasant and stimulating work environment.
Last but not least, my deepest gratitude to my family, especially to

my parents for their supports and to my wife Roya for her precious
love, patience and understanding.
Stockholm, March 2018
Abbas Zangeneh Kamali
v
Publications
Appended papers:
I A. Zangeneh, C. Svedholm, A. Andersson, C. Pacoste, R.

Karoumi, Dynamic stiffness identification of portal frame bridge-
soil system using controlled dynamic testing, in: EURODYN
2017, Rome, September 10-13, Procedia Engineering 199 (2017)
1062-1067
II A. Zangeneh, C. Svedholm, A. Andersson, C. Pacoste, R.

Karoumi, Identification of Soil-Structure Interaction Effect in a
Portal Frame Railway Bridge through Full-Scale Dynamic
Testing, Published in Engineering Structures 159 (2018) 299-309
Both papers were planned, implemented and written by the first

author. The co-authors have participated in the planning of the work
and contributed to the papers with their comments and revisions. The
first author was also involved in performing field measurements.
Other Relevant Publications:
• C. Svedholm, A. Zangeneh, C. Pacoste, S. François, R. Karoumi,

Vibration of damped uniform beams with general end conditions
under moving loads, Engineering Structures 126 (2016) 40-52
• T. Arvidsson, A. Zangeneh, D. Cantero, A. Andersson, Influence

of Sleeper Passing Frequency on Short Span Bridges - Validation
against Measured Results, 1st Int. Conf. on Rail
Transportation, Chengdu, China, July 10-12, 2017
vii
List of abbreviations
SSI Soil-Structure Interaction
FEM Finite Element Method
BEM Boundary Element Method
2D Two Dimensional
3D Three Dimensional
HST High-Speed Train
PML Perfectly Matched Layer
DSM Direct Stiffness Method
SS Simply Supported
ES Elastically Supported
VS Viscoelastically Supported
FRF Frequency Response Function
CWT Continuous Wavelet Transform
ICE InterCity Express Train
HSLM High-Speed Load Model
SI System Identification
ix
Contents
Abstract i
Sammanfattning iii
Preface v
Publications vii
List of abbreviations ix
1. Introduction ................................................................................. 1
1.1. Background ................................................................................................. 1
1.2. Previous work ............................................................................................. 2
1.2.1. Simply supported beam bridges ......................................................................2
1.2.2. Portal frame bridges ........................................................................................3
1.3. Aims and scope .......................................................................................... 4
1.4. Research contribution ............................................................................... 5
1.5. Outline of the thesis ................................................................................... 6
2. Principles of dynamic soil-structure interaction ....................... 7

2.1. Wave propagation in 3D elastic media ..................................................... 7
2.1.1. Wave propagation in half-spaces ....................................................................9
2.1.2. Vibration attenuation .......................................................................................9
2.1.3. Presence of groundwater .............................................................................. 11
2.2. Dynamic soil properties ........................................................................... 12
2.2.1. Small strain shear modulus ........................................................................... 12
2.2.2. Small strain material damping ....................................................................... 13
2.3. General methods to analyze dynamic SSI ............................................. 13
2.3.1. Direct method using FEM .............................................................................. 13
xi
2.3.2. Substructure approach .................................................................................. 16
2.4. Concepts of period lengthening and foundation damping .................. 19
2.5. Impedance functions of shallow foundations ....................................... 19
3. Discrete model for beams on viscoelastic supports .............. 25

3.1. Proposed solution and verification ........................................................ 26
3.1.1. Elastically supported beams .......................................................................... 30
3.1.2. Verification..................................................................................................... 30
3.2. Parametric study ...................................................................................... 31
3.2.1. Definitions of the case studies ....................................................................... 31
3.2.2. Results of the parametric study ..................................................................... 33
3.3. Case studies: Train-induced vibration analysis .................................... 37
3.3.1. Effect of SSI on the train-induced resonant response ................................... 37
3.3.2. Equivalent frequency-independent spring/dashpots ...................................... 39
4. Dynamic SSI analysis of portal frame bridges ........................ 41

4.1. Coupled FE-BE model in the frequency domain ................................... 41
4.1.1. Verification..................................................................................................... 43
4.1.2. Case study bridge ......................................................................................... 44
4.2. Simplified modeling ................................................................................. 47
5. Experimental validation of numerical models ......................... 49

5.1. Description of the tested bridges ........................................................... 49
5.2. Experimental testing ................................................................................ 51
5.2.1. Forced vibration tests .................................................................................... 51
5.2.2. Monitored train passages .............................................................................. 54
5.3. FE model calibration ................................................................................ 55
5.3.1. Numerical models .......................................................................................... 55
5.3.2. System identification results .......................................................................... 56
5.3.3. Validation against monitored train passages ................................................. 58
6. Discussion and conclusion ...................................................... 59

6.1. Simple beam bridges ............................................................................... 59
6.2. Portal frame bridges ................................................................................. 60
xii
6.3. Experimental validation of numerical models ....................................... 61
6.4. Further research ....................................................................................... 61
6.4.1. Simple beam bridges ..................................................................................... 62
6.4.2. Portal frame bridges ...................................................................................... 62
References ..................................................................................... 65
Paper I ............................................................................................ 69
Paper II ........................................................................................... 77
xiii
1. Introduction
1.1. Background
The railway bridge superstructures traveled by high-speed trains may

experience extreme vibration levels due to the periodic nature of the
axle loads, particularly under resonance condition [1]. The resonance
occurs when the train load frequency coincides with a multiple of the
fundamental frequency of the bridge. The critical (resonant) speed of
a railway bridge traveled by a row of moving loads can thus be
estimated by Eq. 1.1 [2, 3].
f0d
ν cr = (1.1)
i
where f0 is the fundamental frequency of the bridge, d is the
characteristic axle distance and i is an integer multiple.
Resonance may result in ballast instability, track degradation,

structural failure or an increase in the maintenance costs of the
railway line [1]. Consequently, apart from the dynamic amplification
of deformations and stresses, the acceleration levels of the bridge
deck should also be maintained under certain threshold values to
guarantee traveling comfort and structural integrity. Typically, the
bridge deck acceleration level is the critical parameter in the case of
short-to-medium span bridges [4]. Therefore, the values of the
acceleration levels should be limited to 3.5 m/s2 and 5 m/s2 for
ballasted and ballastless tracks, respectively [5].
The frequency content and amplitude of the vibrations in railway

bridges is governed by the modal parameters (natural frequency and
damping ratio) of the structural system. Hence, to predict the resonant
response of railway bridges with reasonable accuracy, it is important
to properly estimate the modal parameters of the structural system,
2 | INTRODUCTION
considering the effect of the flexibility and dissipation capacity of the

supporting foundation-soil system.
Despite extensive research studies and guidelines concerning the

influence of Soil-Structure Interaction (SSI) on the seismic response
of bridges (see Sextos 2014 [6] for a summary), relative few studies
have been published about the effect of SSI on the train-induced
vibration of railway bridges. Due to the lack of detailed guidelines
and reliable simple models, the effect of SSI is usually neglected in
the dynamic analysis of railway bridges and classical boundary
conditions are assumed.
According to Eurocode 1991-2 [5], for the analysis of train-induced

vibrations of railway bridges the damping ratio of the bridge should be
at most 2.5 %. However, several numerical investigations [7-10] have
shown that the modal damping ratio of the fundamental vertical
bending mode of e.g. short-span railway bridges (L < 10 m) is higher
than the recommended values when the effect of SSI is considered.
This fact is primarily due to the considerable energy dissipation at the
boundaries. Particularly, in the case of portal frame bridges, the
dissipation capacity of the backfill soil could lead to an increase in the
global damping of the system and consequently reduce the amplitude
of the resonant response of the bridge [11, 12]. Therefore, disregarding
the influence of SSI in the dynamic analyses of short-span railway
bridges may result in conservative and unrealistic results especially in
the resonance regime.
1.2. Previous work
1.2.1. Simply supported beam bridges
Recent numerical studies on the SSI analysis of railway beam bridges

[7, 9, and 10] revealed that the resonance response can be
considerably influenced by the dynamic stiffness of the supporting
foundation-soil system. Romero et al. (2013) [7] have studied the
effect soil-structure interaction on the resonance response of a beam
bridge using a fully 3D multi-body boundary element-finite element
model (BEM-FEM) formulated in the time domain. The case study
bridge has a span length of 10 m and is simply supported at
INTRODUCTION | 3
embedded abutments in a homogenous half-space. The authors

concluded that both the period and damping ratio of the fundamental
bending mode of the system are higher when SSI is considered.
Doménech et al. (2015) [9] have analyzed the dynamic response of
simple beams traveled by moving loads considering soil-structure
interaction effects, using a 3D BEM-FEM coupled model. The case
study bridges are single-span beams with span length between 12.5
and 25 m simply supported at surface foundations on a homogenous
half-space medium. The analysis showed that considering soil effects
leads to a reduction of the bridge maximum free vibration amplitudes
and train-induced resonant response. This reduction is a consequence
of the inherent damping of the soil due to wave radiation. The effect
becomes higher, going from stiffer to softer soils. The study
concluded that the structures that are most affected by the soil
flexibility are those with highest natural frequency for all investigated
lengths. An analytical model for calculating modal properties of a
Euler-Bernoulli beam, supported by an elastic half–space, has been
derived by Svedholm et al. (2015) [10] and used to study the effects
of soil–structure interaction (SSI) for reinforced concrete bridges with
ballasted tracks. The numerical results presented in the study showed
that the expected foundation damping ratio for the 1st bending mode
of the system is between 0.1 % and 6 % for a long span bridge (L=60
m) supported by a dense soil and for a short span bridge (L=5 m)
supported by a loose soil, respectively. Based on the aforementioned
studies concerning single-span beam bridges simply supported at
shallow foundations on a homogenous half-space medium, the
following general conclusions can be drawn: (i) the flexibility of the
supports lowers the fundamental frequency, and consequently reduces
the resonance speed; (ii) the dissipation capacity of the foundation-
soil system (known as geometrical damping) increases the global
damping of the system and lowers the vibration levels at resonance;
and (iii) the effect of SSI increases as the ratio between the beam
flexural stiffness and the foundation-soil stiffness becomes higher.
1.2.2. Portal frame bridges
In the field of train-induced vibration analysis of railway bridges,

limited results have been published on the portal frame
bridges/underpasses considering the effect of soil-structure
4 | INTRODUCTION
interaction. Galvin et al. (2007) [13] developed a fully 3D coupled

BEM-FEM model in the time domain for the analysis of high-speed
train passage effects on local structures, taking dynamic SSI into
account. Using a rigorous numerical model, the vertical tractions on
the deck of a concrete underpass surrounded by embankment has
been studied. Ülker-Kaustell et al. 2010 [8] presented a simplified 2D
frame model for the dynamic SSI analysis of a portal frame railway
bridge, neglecting the effect of the backfill soil and modeling the
effect of the subsoil by using frequency-dependent impedance
functions. The authors concluded that the contribution of the soil-
bridge interaction to the modal damping ratios of the coupled system
is quite substantial, especially for soft soils. A simplified method for
dynamic SSI analysis of culverts traversed by high-speed trains has
been developed by Fraile et al. (2002) [11]. The method used in the
study was based on the modal superposition of the rigid body modes
associated to soil displacements and the flexible modes of the
structure supported by a Winkler type foundation. Vega et al. (2012)
[12] have studied the dynamic response of an underpass along a high-
speed railway line, performing on-site measurements and numerical
modeling. A full 3D FE model of the culvert and embankment has
been developed and calibrated based on the monitored train-induced
accelerations. A simplified 2D FE model has also been presented.
This model can be useful for practical design purposes, providing that
its limitations are well understood by the user (i.e. for cases where
higher plate modes are not important).
1.3. Aims and scope
The overall objective of the present research is to study the influence

of Soil-Structure Interaction (SSI) on the resonance response of
railway bridges, with focus on the modal properties of the coupled
bridge-soil system. The specific aims are as follows:
• Determine the factors that influence the modal properties of

the bridge-soil system.
• Develop both complex and simple numerical models for the

analysis and assessment of bridges on high-speed lines.
INTRODUCTION | 5
• Provide practical tools and guidance regarding the efficient

modeling of SSI for high-speed railway bridges.
• Implement structural identification techniques for inverse

estimation of the modal parameters of real bridge-soil systems
and experimental validation of the numerical models.
The following limitations apply:
• The dynamic Soil-Structure Interaction (SSI) is a broad

research field. In this thesis, only the effect of the flexibility
and damping of the foundation-soil system on the train-
induced dynamic response of railway bridges is studied.
• The present study is limited to linear time-invariant systems.

Any type of non-linearity such as plastic behavior of the soil
material, cracking of the concrete or possible formation of
gaps at soil-structure interfaces is beyond the scope of this
research.
• In the numerical and analytical studies on beam bridges, only

single-span simply supported beam bridges are studied. The
analyses are limited to 2D models based on Euler–Bernoulli
beam theory.
• In the train-induced vibration analyses, the railway excitation

is simulated as a series of moving forces traveling at constant
speed, and thus the train-bridge interaction is neglected.
1.4. Research contribution
The theoretical, numerical and experimental investigations on the

dynamic behavior of railway bridges presented in this thesis have
resulted in the following research contributions:
• Closed-form expressions for evaluating natural frequency and

modal damping ratio for the fundamental bending mode of
single-span beams on viscoelastic supports.
6 | INTRODUCTION
• An analysis of the effect of different subsoil conditions on the

modal parameters and resonant response of beam bridges
through a comprehensive parametric study. One important
issue in this context is also to identify the situations for which
the effect of SSI is negligible.
• Recommendations on how to efficiently model the effect of

SSI on railway beam bridges.
• A proposal concerning a Winkler-based simplified model for

the analysis of the interaction between abutments and backfill
soil.
• Full-scale controlled dynamic tests on two short-span portal

frame bridges and experimental validation of the numerical
models for the prediction of the dynamic behavior of bridge-
soil systems.
1.5. Outline of the thesis
The reminder of this thesis is structured as follows. Chapter 2

explains the basic principles of dynamic Soil-Structure Interaction
(SSI) which are relevant to the work presented in subsequent
chapters. In Chapter 3, a closed-form solution for estimating the
fundamental modal parameters of beam bridges on viscoelastic
supports is presented, followed by a comprehensive parametric study
aimed at evaluating the effect of SSI on the modal properties of the
fundamental bending mode of beam bridges. Different modeling
alternatives for the dynamic SSI analysis of portal frame bridges are
discussed in Chapter 4 and a simplified procedure for modeling the
interaction between abutments and backfill soil is presented. Chapter
5 summarizes the experimental results of vibration testing on two
portal frame bridges and experimental validation of the numerical
models. Chapter 6 concludes the work in this thesis and presents
suggestions for future research work.
2. Principles of dynamic soil-structure interaction
The response of a structure to dynamic excitations is influenced by
the interaction between the structure, the foundation and the
surrounding soil medium. A dynamic Soil-Structure Interaction (SSI)
analysis assesses the combined response of the coupled structure-
foundation-soil system subjected to an incident wave field as well as
external dynamic loads. In this research, only the effect of flexibility
and damping of the soil-foundation system on the train-induced
resonance response of railway bridges is investigated. The effect of
incident wave field on the bridge due to rail transit at nearby
embankments or seismic ground motions is beyond the scope of this
report.
The chapter is organized as follows. First, a brief introduction to the

wave propagation in elastic medium is presented. Section 2.2 gives
some empirical formulas that can be implemented to estimate
dynamic soil properties at small strain levels. In section 2.3, some
general numerical approaches in dynamic SSI analyses are
introduced. Next, the concepts of foundation damping and period
lengthening are briefly discussed. In section 2.5, the impedance
functions of rigid surface foundations are presented, followed by
some practical recommendations.
2.1. Wave propagation in 3D elastic media
In contrast to seismic loading, railway-induced vibrations usually

result in low amplitude oscillations and consequently rather low strain
levels in the soil media. Therefore, the constitutive behavior of the
soil can be represented by a linear viscoelastic relation. In the absence
of body forces, the motion of an isotropic, homogeneous linearly
elastic medium is expressed by the following system of partial
differential equations (Navier's equation):
8 | PRINCIPLES OF DYNAMIC SSI
∂e ∂ 2u
(λ + G ) + G∇ 2ui =ρ 2i , i = x, y, z (2.1)
∂i ∂t
where e is the volumetric strain, ui is the displacement component

with respect to the i-th spatial coordinate, ρ is the density, G is the
shear modulus and λ is Lame’s first constant which can be expressed
as
2Gν
λ= (2.2)
(1 − 2ν )
In classical elastodynamics [14], the solution to Navier’s equation is

normally obtained by Helmholtz decomposition of the displacement
vector into dilatational and rotational parts. This results in a set of
uncoupled wave equations, defining the propagation of dilatational
(or longitudinal, P) waves and shear (or rotational, S) waves,
respectively.
1 ∂ 2e
∇ 2e = 2 2 (2.3)
Vp ∂t
∂ui ∂u j
2
1 ∂ wij
∇ 2 wij = , wij (
= − ) (2.4)
Vs 2 ∂t 2 ∂j ∂i
where wij is the rotation about the normal axis of the ij-plane and Vp
and Vs are the dilatational and shear wave velocity, respectively.
G λ + 2G 2 − 2ν
Vs = ,=Vp = Vs (2.5)
ρ ρ 1 − 2ν
The dilatational and shear waves, labeled as body waves, are the only
wave types that can propagate in a homogenous infinite medium. In
general, the shear waves are slower than the dilatational waves. Fig.
2.1a-b depicts the character of the motion associated with each of
these two waves. Table 2.1 presents the typical values of the shear
wave velocity and compression wave velocity for different materials.
PRINCIPLES OF DYNAMIC SSI | 9
Table 2.1. Indicative values of P- and S- wave velocity for different soils [15]
P-Wave velocity S-Wave velocity
Soil type
(m/s) (m/s)
Ice 3000-3500 1500-1600
Water 1480-1520 0
Granite 4500-5500 3000-3500
Sandstone, Shale 2300-3800 1200-1600
Fractured Rock 2000-2500 800-1400
Moraine 1400-2000 300-600
Saturated Sand and Gravel 1400-1800 100-300
Dry Sand and Gravel 500-800 150-350
Clay below GW level 1480-1520 40-100
Organic soils 1480-1520 30-50
2.1.1. Wave propagation in half-spaces
As a stress-free edge is defined in the elastic medium to model the

ground surface (half-space medium), an additional type of waves
denoted as surface waves is introduced. This type of waves is the
result of the interaction between inclined body waves and the free
surface of the half-space. The surface waves propagate along the free
surface with amplitudes that decline exponentially with depth. There
are several types of surface waves [16], with the Rayleigh waves
beeing the most important type. Fig. 2.1c shows the character of the
motion associated with Rayleigh waves. In the case of a
homogeneous perfectly elastic half-space, the phase velocity of
Rayleigh waves can be numerically approximated by Eq. 2.6 [14].
0.862 + 1.14ν
VR ≈ Vs (2.6)
1 +ν
2.1.2. Vibration attenuation
Due to material and geometrical damping, the intensity of vibrations

initiated at a source reduces as the waves propagate further away
from the source.
2.1.2.1. Geometrical damping
Geometrical (radiation) damping occurs because the vibration energy

diffuses over a larger area as the incident wave moves away from the
source and propagates to infinity. The solution for attenuation of

waves in a perfectly elastic half-space was presented by Lamb (1904)
[17]. According to Lamb’s solution, the body waves (compression
and shear waves) undergo substantial geometrical damping as they
propagate in every direction in the medium. However, since the
Rayleigh waves only radiate along the free surface, they do not
experience the same radiation damping [18]. To model numerically
the unboundedness of the medium, the solution to Navier’s equation
(Eq. 2.1) has to satisfy Sommerfeld’s radiation conditions [19] which
indicate that just an outward flow of energy is allowed at infinity.
Direction (a)
of particle
motion
Wavelength
(b)
Direction
of particle
motion
Direction
(c)
of particle
motion
Fig. 2.1. Particle motion variation for wave type: (a) compression wave; (b) shear
wave; (c) Rayleigh wave [20]
2.1.2.2. Material damping
Material damping represents the energy dissipation properties of the

soil material under cyclic stress. In the case of numerical modeling in
frequency domain, the hysteretic material damping in the soil can be
modeled by means of the correspondence principle of viscoelasticity
[21, 22], introducing complex Lamé coefficients:
G
= *
G (1 + 2iξs ) (2.7)
(λ + 2G )* =
(λ + 2G )(1 + 2iξ p ) (2.8)
where ξs and ξp are the frequency-independent hysteretic damping

ratios corresponding to the shear and compression waves,
respectively.
2.1.3. Presence of groundwater
The wave propagation in a saturated poroelastic medium can be

described by Biot’s poroelastic equations [23]. Based on Biot’s
theory, in the low frequency range a saturated poroelastic medium
acts as a frozen mixture in which the relative motion between the
solid skeleton and the pore fluid is negligible. Hence, the behavior of
a saturated poroelastic medium may be expressed as an equivalent dry
elastic medium by using the following equations [15];
G
Vs,eq = (2.9)
ρeq
(V Vs,eq ) − 2
2
ρ p,w
ρeq = ρ + (1 − ) ρ w , ν eq = (2.10)
ρg 2 (V Vs,eq ) − 2
2
p,w
where ρg =2650 kg/m3 and ρw =1000 kg/m3 are the mass density of
gravel grains and pore water, respectively and Vp,w=1500 m/s is the
compression wave velocity for water.
2.2. Dynamic soil properties
In the absence of dynamic laboratory and seismic in situ tests [15,

16], some empirical methods can be used to obtain a first estimation
of the soil properties. In the following, some simple empirical
expressions will be briefly mentioned. A summary of the different
empirical formulas and more detailed explanations can be found in
[15, 16 and 18].
2.2.1. Small strain shear modulus
The shear modulus is the most important parameter of the soil

medium that affects the wave propagations and subsequent vibrations.
The most accurate way to estimate the in-situ value of the shear
modulus is to measure the small-strain shear wave velocity, Vs. Then,
the maximum shear modulus can be calculated as shown in Eq. 2.11.
Gmax = ρVs2 (2.11)
Typical values for shear wave velocity for different materials are
presented in Table 2.1.
For granular soil, the following relation is often used to estimate the
small strain shear modulus Gmax in kPa [24, 25]:
Gmax = 1000 K σ m′ (2.12)
where σ´m [kPa] is the mean effective stress and K is an empirical

coefficient which depends on the soil type, void ratio and relative
density. The recommended values for K are about 7.5 for loose sand,
15 for very dense sand and 30 for gravel.
In Scandinavian clays, the small strain shear modulus depends mainly

on the undrained shear strength, cu [kPa], liquid limit, wL, and
plasticity index, Ip [25, 26]. The following empirical expressions can
be used to estimate the maximum shear modulus of normally
consolidated low-plastic clay (Eq. 2.13) and high- to medium-plastic
clay (Eq. 2.14).
504cu
Gmax = (2.13)
wL
G
=  208 + 250  c
max  I  u (2.14)
 p 
2.2.2. Small strain material damping
Even at very small strains, soil materials are not perfectly elastic and
reveal some energy dissipation capacity during cyclic loading. This is
referred to as soil material damping. The small strain material damping,
ξs, ranges between 2 and 6 percent for different types of soil [27, 28].
As the dissipation capacity of the soil medium through its material
damping may be considerably lower than radiation damping, the
accurate assessment of the former is typically irrelevant except for
some special cases (e.g. vertical oscillations of foundations on a
shallow soil stratum at low frequencies) [29].
2.3. General methods to analyze dynamic SSI
2.3.1. Direct method using FEM
The most straightforward method to model dynamic soil-structure

interaction is to include in the numerical model the entire structure-
soil system using the finite element (FE) method [30]. Using this
direct method, the dynamic analysis can be performed either in time
or frequency domain. In the time domain solution, material and
geometrical nonlinearities can also be included and are easy to
implement. As schematically shown in Fig. 2.2, the structural
elements including the foundations are modeled along with part of the
surrounding and underlying soil medium. The soil is usually modeled
as a continuum medium and transmitting boundaries are implemented
at the borders of the soil mesh to approximately model the
unboundedness of the soil medium. The dimensions of the finite
domain of the soil medium in combination with the implemented
transmitting boundaries should guarantee that the steady state
response of the system is not contaminated by wave reflection at the
boundaries.
FE model of
the structure
FE model of
the foundation FE model of
the soil
Transmitting
boundaries
Fig. 2.2. Schematic illustration of the direct method in the soil-structure analysis
2.3.1.1. Transmitting boundaries (absorbing boundaries)
Several possibilities exist to implement transmitting boundaries [31-

33]. In the following, two of the most popular transmitting
boundaries, available also in commercial finite element software, are
briefly discussed.
The simplest way to implement transmitting boundary is the viscous

boundary method [34]. In this method, at each node on the boundary,
two tangential and one normal dashpot are applied to absorb shear
and dilatational waves, respectively (see Fig. 2.3). The viscous
boundary perfectly absorbs waves at normal incidence while its
efficiency exponentially decreases as the wave’s impinging angle
deviates from the normal direction. This type of absorbing boundaries
is not applicable to static problems, as the dashpot has no static
stiffness.
A recently developed method which is very accurate is the Perfectly

Matched Layer (PML). In this method, an absorbing boundary layer
(layer of artificial absorbing material) is modeled at the edges of the
finite mesh. This boundary layer uses stretched internal spatial
coordinates to simulate the outward flow of energy to infinity. The
PML is relatively simple to implement in elastodynamic problems
and formulations exist in both the frequency and time domains [35-
38]. To the author’s knowledge, the frequency domain formulation of

PML is available in COMSOL Multiphysics [39] commercial
software.
Dt = rVs Atrib.
Dn = rVp Atrib.
Fig. 2.3. Schematic illustration of viscous boundary [40]
In the absence of proper transmitting boundaries, the scale of the

finite domain should satisfy the criterion in Eq. 2.15, to achieve a
steady-state response which is not disturbed by the wave reflection at
boundaries [41].
Vp
Lr ≥ (2.15)
f min
where Lr is the shortest path in the finite domain from the region of
interest to the boundaries and fmin is the lowest frequency of interest.
2.3.1.2. Finite element size
The maximum element size in the soil medium is controlled by the

minimum shear wavelength (λmin) and should be chosen in a way that
satisfies the following criterion [42].
1 1 Vs
le, max ≤  −  lmin , λmin = (2.16)
5 8 f max
where le,max is the maximum element size and fmax is the highest
frequency of interest. According to Lysmer (1978) [42], eight linear
elements (nine nodes) per wavelength will provide about 90%
accuracy on wave amplitudes in the highest frequency range of
interest.
2.3.2. Substructure approach
Due to high computational costs related to the FE modeling of the

unbounded soil medium, the direct solution is not very efficient and
alternative methods such as the substructure approach have been
developed. In the substructure approach, the structure-soil system is
decomposed into separate structure and soil subdomains and each
subdomain is solved by the best-suited computational technique. As
the response of the soil medium is generally frequency-dependent,
this approach is very efficient in the frequency domain, which in its
turn is limited to linear problems.
The general procedure of the substructure approach in the frequency

domain can be described as follows (see also Fig.2.4):
1- The stiffness and damping matrices of the ground at the

interface nodes for each frequency ( K gff and Cgff ) are evaluated
by Eq. 2.17, where Uf and Pf are the displacement and force
vectors at the interface, respectively (Fig. 2.4a).
(K g
ff (ω ) + iω Cffg (ω ) ) .U f (ω ) =
Pf (ω ) (2.17)
(b) P s , Us
Ps , U s
(c)
P f , Uf Ps , Us
= (a)
+ = Uf
Pf , Uf
Ug
Ug
Fig. 2.4. Schematic illustration of the substructure approach in the soil-structure

analysis
2- The equations of motion of the structure-interface system can

be written as shown in Eq. 2.18 (Fig. 2.4b).
 K ss K sf  M M sf 
 s 
− ω 2  ss
  K fs K ff   M fs M ff 
C Csf   U s (ω )  + Ps (ω ) 
+iω  ss  =  (2.18)
Cfs Csff   U f (ω )  −Pf (ω ) 
3- The stiffness and damping matrices of the ground are

introduced in the structure-interface equations of motion by
inserting Eq. 2.17 into Eq. 2.18 (see also Fig. 2.4c). This
results in the coupled equations of motion as follows:
 K ss K sf  2 
M ss M sf 
 − ω
  K fs K sff + K ffg (ω )  M
 fs M ff 
C Csf   U s (ω )  Ps (ω ) 
+iω  ss  =  (2.19)
Cfs Csff + Cffg (ω )   U f (ω )   0 
2.3.2.1. Rigorous modeling: Coupled FE-BE methods
An efficient and accurate way to calculate the dynamic stiffness

matrices of the soil medium is to use the Boundary Element Method
(BEM), provided that the fundamental solutions (Green’s functions)
at the interface nodes are available. In the BE method, the boundary
of a domain is discretized with a proper number of boundary elements
to numerically solve a boundary integral equation [43]. The Green’s
functions represent the response of a finite or infinite medium to a
unit harmonic load. The Direct Stiffness Method (DSM) [44, 45] is
one of the methods which can be used to calculate Green’s functions
for a horizontally layered isotropic elastic medium. Once the Green’s
functions at the interface nodes are evaluated, the BE method can be
implemented to calculate the interaction force-displacement
relationship of the interface nodes (Eq. 2.17). However, this approach
is limited to linear analysis and difficult to use for practical
engineering purposes since the BEM and DSM are not available in
commercial software.
2.3.2.2. Approximate modeling: impedance functions/distributed spring-

dashpot systems
In this approximate approach, the dynamic interaction between the

structure and the subgrade soil system is modeled using impedance
functions at the bottom interface, assuming rigid behavior of the
foundations (see Fig .2.5). The basics of impedance functions will be
discussed in section 2.5.
In the case of embedded non-rigid structural elements such as bridge

abutments when the flexibility of the structural element at interface is
important, the stiffness and damping characteristics of the soil-
structure interaction can be approximately modeled using distributed
springs and dashpots (see Fig. 2.5). Several approximate methods
have been proposed to model the seismic response of vertical wall-
soil systems, assuming elastic behavior of the soil medium and a
perfect contact at the wall-soil interface. Scott (1973) [46], proposed a
simple model representing the interaction between a rigid cantilever
wall and the soil by a set of massless, linear horizontal springs.
Scott’s method was modified by adding a series of linearly distributed
dashpots to consider the radiation damping capacity of the medium
[47]. Veletsos and Younan (1995, 1997, and 2000) [48-50] developed
more complicated models to consider the frequency dependency of
the stratum and flexibility of the wall.
k n , dn
k t , dt
Z=K+i.wC Z=K+i.wC
Fig. 2.5. Schematic illustration of the simplified FE model
2.4. Concepts of period lengthening and foundation

damping
In the field of seismic soil-structure interaction, the concepts of period

lengthening and foundation damping are frequently used to describe
the effect of the SSI on the modal properties of the structures. These
concepts originate from the flexibility and dissipation capacity of the
foundation-soil system. The flexibility of the soil-foundation system
may considerably contribute to the global stiffness of the coupled
system and increase the natural period of the structure. Moreover, the
deformations of the foundation-soil system enable energy dissipation
through hysteretic soil damping and radiation damping, which may
lead to an increase in the overall damping of the system. Several
approximate closed-form expressions for evaluating these two
quantities for building structures have been presented by a number of
researchers in the field of seismic SSI [51, 52].
However, in the field of railway bridge dynamics the aforementioned

concepts are somewhat new, and it is necessary to study them in more
detail. The period lengthening (frequency shortening) of the
fundamental vertical bending mode of a railway bridge due to SSI,
may result in a reduction of the critical speed at resonance. On the
other hand, neglecting the effect of foundation damping in the train-
induced vibration analyses of certain types of railway bridges could
lead to conservative and unrealistic results in the resonance regime. In
chapter 3, closed-form expressions aimed at evaluating the period
lengthening and foundation damping of the vertical bending mode for
railway beam bridges are proposed.
2.5. Impedance functions of shallow foundations
The foundations of railway bridges are generally designed as shallow

foundations whenever the quality of the subsoil is good. Assuming a
rigid massless surface foundation, the dynamic stiffness and damping
properties of the foundation-soil system (decoupled from the
structure) is characterized by frequency-dependent impedance
functions. The complex-valued impedance function for each degree
of freedom can be represented by Eq. 2.20. A comprehensive
presentation of the theoretical background and derivation of the

impedance functions can be found in [29].
(ω ) K (ω ) + iωC (ω )
Z= (2.20)
The real part of the impedance, K is denoted as the dynamic stiffness.

It describes the stiffness and inertia properties of the soil. The
imaginary part is the product of the frequency ω and the dashpot
coefficient C. The dashpot coefficient describes the dissipation
capacity of the system due to material and radiation damping (Eq.
2.21).
C ≈ Cradiation + Chysteretic (2.21)
2K
Chysteretic = ( )ξs (2.22)
ω
where ξs is the damping ratio of the soil material. The expression for
Chysteretic in Eq. 2.22 is an approximation based on the correspondence
principle of viscoelasticity [21].
Numerous impedance functions for different foundation

configurations (surface, embedded), and subsoils condition (uniform
or layered half-space, stratum over rock) have been published by
pioneers in the field of soil dynamics [29, 53-55]. In this report, the
impedance functions for rigid surface foundations are calculated
using the Direct Stiffness Method [44, 45] and the Boundary Element
method [43] which are implemented in an elastodynamic toolbox [56]
for MATLAB.
In 3D, the full impedance matrix is represented by a 6×6×ω matrix

where the off-diagonal terms correspond to the coupling between
different degrees of freedom. In the following subsections, only the
vertical component of the impedances of surface foundations is
discussed.
2.5.1.1. Surface foundations on homogeneous half-space
The normalized dynamic stiffness and radiation damping coefficients

of a rigid rectangular surface foundation with L/B=2 resting on a
homogenous half-space are shown in Fig. 2.6.
1 1.5
(b)
Cv,rad / Cv,0
0
Kv / Kv,0
(a)
1
-1 Dry ν =0.3
Saturated νeq , ρeq
-2 0.5
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
ωB/Vs ωB/Vs
Fig. 2.6. Normalized vertical dynamic stiffness and radiation damping coefficients
for rectangular surface foundations with L/B=2, resting on a homogenous half-space
The dynamic stiffness in Fig. 2.6a is normalized to the static

component, Kv,0. The elastic solution for the static stiffness of the
system can be defined as shown in Eq. 2.23 [29].
ρVs 2 Af B 3
K v,0 (halfspace)
= (0.73 + 1.54( ) 4 ) (2.23)
2 B(1 −ν ) L
where B, L and Af are the semi-width, semi-length and the area of the
foundation, respectively.
The radiation damping coefficient in Fig. 2.6b is normalized to the

perfect viscous P-dashpot coefficient, Cv,0, analog to infinite 1D
compression wave propagation.
Cv,0 = ρVLa Af (2.24)
In Eq. 2.24 VLa is the Lysmer’s analog wave velocity [57] which is
defined based on Lysmer's frequency-independent analog for a
circular rigid foundation vibrating vertically on the surface of a half-
space [58]:
3.4Vs
VLa = (2.25)
π (1 −ν )
Dry soil
As can be seen from Fig. 2.6a, in the case of a shallow foundation

with practical dimensions (L/B=2-3) the variation of the normalized
dynamic stiffness and radiation damping coefficients in the
interesting range of frequencies is moderate. Hence the vertical
impedance of a shallow foundation resting on a dry homogenous half-
space can be considered as frequency independent.
Saturated soil
In the presence of ground water, the dynamic stiffness shows a

downward parabolic tendency, as shown in Fig. 2.6a. The reason for
this tendency is the trapped mass, ∆M, of the nearly incompressible
soil beneath the foundation, which oscillates as a rigid body in phase
with the foundation [59]. In the presence of this mass, the real part of
the impedance function in Eq. 2.20 is replaced by Eq. 2.26. For
Poisson’s ratio larger than 0.3, the trapped mass for vertical motion
can be approximated by Eq. 2.27 [59].
) K (ω ) − ω 2 ∆M
Re( Z= (2.26)
Af
∆M ≈ 2.4π (ν − 0.3) ρ ( )1.5 (2.27)
π
The variation of the normalized radiation damping coefficient is not
affected by the presence of ground water, as shown in Fig. 2.6b.
However, due to higher values for Lysmer’s analog wave velocity as
well as equivalent soil density for the saturated soils, the magnitude
of the radiation damping coefficient, Cv,radiation is slightly higher than
in the case of dry soils.
2.5.1.2. Surface foundations on homogeneous stratum over bedrock
In practice, natural soil deposits are normally resting on bedrock or

very stiff layers at a shallow depth, instead of extending to infinity as
for the half-space medium [29]. The vicinity of such a stiff layer to
the oscillating surface alters the static and dynamic stiffness as well
as the dissipation capacity of the soil-foundation system. Fig. 2.7
shows the dynamic stiffness and radiation damping coefficients of a

rectangular surface foundation with L/B=2 resting on a stratum of
depth H, for various H/B ratios. The bedrock is considered as a fully
rigid surface.
1 2
Cv,rad / Cv,0
0.5
Kv / Kv,0
1
0 H/B=16
H/B=8
(a) (b)
-0.5 H/B=4 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
f / fc f / fc
Fig. 2.7. Normalized vertical dynamic stiffness and radiation damping coefficients
for rectangular surface foundations with L/B=2, resting on homogenous stratum
over bedrock
In this case, the static stiffness of the system (based on an elastic

solution) is increased as the depth of the stratum H decreases [29].
 rVs 2 Af B 3  B H 
K v,0 (stratum)=  (0.73 + 1.54( ) 4 )  1 + ( 2.28)
 2 B(1 −ν ) L   (0.5 + B L) 
The dynamic stiffness exhibits a strong frequency-dependency and

drops sharply at frequencies near the natural frequency of the deposit
(see Fig. 2.7a). The first resonant frequency of a homogenous stratum
for compressing modes (vertical, rocking) can be estimated by Eq.
2.29. At operating frequencies below the first resonant frequency of
the stratum, radiation damping is negligible since the downward
propagation of body waves is hindered by the bedrock and
consequently the foundation damping originates only from the
hysteretic damping of the soil medium. However, as operating
frequencies become higher than the resonant frequency of the
stratum, the radiation damping capacity of the stratum is restored and
gradually approaches the “half-space” values, as shown in Fig 2.7b.
VLa
fc = (2.29)
4H
3. Discrete model for beams on viscoelastic
supports
The dynamic response of short-span railway beam bridges at
resonance is governed by their fundamental vertical bending mode
[60-62]. To study the influence of soil-structure interaction on the
modal properties of railway beam bridges in a systematic manner,
derivation of closed-form expressions to calculate the natural
frequency and modal damping ratio of the bridge-foundation-soil
system is helpful. In this chapter, a discrete 2 degrees of freedom
model is proposed, and analytical closed-form expressions are
derived to estimate the natural frequency and modal damping ratio of
single-span beam bridges with viscoelastic supports. These
expressions provide a tool that can be used for a better understanding
of the phenomena, for determining when the effect of SSI is
important and for parametric studies of the effect of various
foundation-soil types on the dynamic behavior of the bridge.
The chapter is organized as follows. In the first section (Section 3.1),

closed-form analytical expressions for the evaluation of the
fundamental natural frequency and the corresponding damping ratio
of a simple beam on viscoelastic supports are derived using a
simplified discrete system. Second (Section 3.2), the variation of the
fundamental natural frequency and modal damping ratio of bridge-
foundation-soil systems are analyzed through a comprehensive
parametric study. Different bridge lengths and soil properties were
chosen considering practical and realistic case studies. This gives
valuable information on how SSI can affect the dynamic response of
railway beam bridges and when the effect of SSI is negligible. In the
last section (Section 3.3), the train-induced resonant response of a
bridge case study is analyzed. The variation of the response
amplitudes at resonance with different foundation-soil is studied and
its relationship with the evolution in the modal parameters of the
26 | DISCRETE MODEL FOR BEAMS ON VISCOELASTIC SUPPORTS
system is discussed. With a view to practical design applications,

some simple modeling techniques and recommendations for the
dynamic SSI analyses of railway beam bridges at resonance are
proposed in the last part of the chapter.
3.1. Proposed solution and verification
The studied continuous system (single–span 2D Euler–Bernoulli

beam) is approximated by an equivalent 2 degrees of freedom system,
as shown in Fig. 3.1. The fundamental vertical bending mode shape
of the system is approximated by Eq. 3.1 which is based on the
superposition of the flexural sine mode of a simply supported beam
and the rigid body mode shape of the beam supported by dynamic
impedances. Note that in [62] the same assumption has been adopted
to estimate the natural frequency of elastically supported beams.
 π x  u 
, t ) Φ.d ≅ 1 sin( )   f 
u ( x= (3.1)
 L  ub 
The Lagrange’s equation for the system can be written as:
d  ∂T  ∂U ∂D
  + + 0,
= j = f, b (3.2)
dt  ∂u j  ∂u j ∂u j
kv cv EI, m, L cv kv
ub
EI, m, L
uf uf
kv cv m, L cv kv
Fig. 3.1. Model of free vibration mode shape as a superposition of simply

supported (SS) and rigid beam on viscoelastic supports
DISCRETE MODEL FOR BEAMS ON VISCOELASTIC SUPPORTS | 27
The kinetic and potential energy of the system are evaluated using Eq.
3.3 and 3.4, respectively. The dissipation energy corresponding to the
material and geometrical damping of the foundation-soil system is
given by Eq. 3.5:
2
1 L  du  1  4u u u 2 
T
= m ∫   d= x mL uf 2 + f b + b  (3.3)
2 0  dt  2  π 2 
2
L d u  1  π 4 EI 2 
2
1 1
U= EI ∫  2  dx + (2k v )uf =  3 ub + 2k v uf 2 
2
(3.4)
2 0
 dx  2 2  2L 
1
D = (2cv )uf 2 (3.5)
2
It should be noted that in the derivation of these equations, the

structural damping of the beam is neglected and only the foundation
damping is considered. Applying Eq. 3.2, term by term for both
degrees of freedom results in the following equation of motion for the
free response of the 2 DoF system at its first bending mode:
 + Kd + Cd =
Md 0 (3.6)
where
 2k v 0 
 uf   1 2 π  2cv 0
d =   , M = mL   , K= 
π 4 EI  , C =  
ub  2 π 1 2   0
 0 0
 2 L3 
The modal properties associated with Eq. 3.6 are determined by

assuming a solution of the form uj(t)=ujeλt where λ is the complex
eigenvalue. This results in the complex eigenvalue problem in Eq.
3.7.
(K + λC + λ 2 M )d = 0 → det(K + λC + λ 2 M ) = 0 (3.7)
Inserting the obtained mass, stiffness and damping matrices into Eq.
3.7, gives the quartic polynomial in Eq. 3.8 as the complex
eigenvalue equation.
(π 2 − 8) 2 4
2
γ λ + βγλ 3 + γ (1 + D)λ 2 + 2β Dλ + 2 D =
0 (3.8)
2π
mL π 4 EI c
where γ = , D= 3 , β = v
kv 2 L kv kv
Providing that the studied system is underdamped, which is mostly

the case, the studied eigenvalue problem has always a pair of complex
conjugate roots of the form [63]:
λ =± −ξnω n ± iω n 1 − ξn 2

λr iλi = (3.9)
By calculating the real and imaginary parts of the complex roots for
Eq. 3.8 (λr and λi) using general solution for quartic polynomials, the
closed-form expressions for the undamped natural frequency and
modal damping ratio of the fundamental bending mode of the system
is obtained.
ω n
= λr 2 + λi 2 (3.10)
ξ n = − λr ω n (3.11)
According to the general solution of the quartic polynomial [64], the

four roots of Eq. 3.8 can be expressed in general format as follows:
π2 β 1
λ1,2 =− −S ± −4 S 2 − 2 p + q S (3.12)
2(π − 8)γ
2
2
π2 β 1
λ3,4 =− +S± −4 S 2 − 2 p − q S (3.13)
2(π − 8)γ
2
2
Where:
2
2π 2 (1 + D)  π2 β 
=p − 1.5  2 
(π 2 − 8)γ  (π − 8)γ 
3
π6  β  2π 4 β (1 + D) 4π 2 β D
q =2   − +
(π − 8)3  γ  (π 2 − 8) 2 γ2 (π 2 − 8) γ 2
12(π 2 − 8) 2
D 0 = ( D + 1) γ − 6 β γ D +
2 2 2
γ D
π2
  432   72(π 2 − 8) 3
D= 2( D + 1) γ
3 3
+ β γ
2 2
D  36 +  36 − D − γ D(1 + D)
π 2  
1
  π2

( 
)
2
3 0.25 ∆ + ∆ 2 − 4∆ 3 + ∆0 
1 2 2π  2 1 1 0
S= − p+  
3(π − 8)γ 2 
( )
2
2 3 2
3 0.5 ∆ + ∆ − 4∆ 3 
 1 1 0 
 
Based on the sign of the discriminant (4∆03−∆12), the equation has one
or two pairs of complex conjugate roots [65, 66]. As it will be proved
in the verification studies in section 3.1.2, the real and imaginary
parts of the lowest eigenvalue for the fundamental bending mode of
the system can be obtained by Eq. 3.14 and 3.15, considering only the
real part of the term S.
π2 β
λr =
− + sgn S (3.14)
2(π 2 − 8)g
1 2 q
λi
= 4 S + 2 p + sgn (3.15)
2 S
where
+1 if (−4 S 2 − 2 p − q S ) < 0,

sgn =  2
−1 if (−4 S − 2 p − q S ) > 0,
Once the real and imaginary parts of the complex eigenvalue are
evaluated, closed-form expressions for the natural frequency and
modal damping ratio of viscoelastically supported (VS) beams can be
obtained from Eq. 3.16 and Eq. 3.17, respectively:
2
 π2 β  2 p q
ω0,VS =  − + sgn S  + S + + sgn (3.16)
 2(π − 8)g
2
 2 4S
1  π2 β 
ξ 0,VS =
−  − + sgn S  (3.17)
ωn ,VS  2(π − 8)g
2

Noteworthy to mention, the proposed solution is not valid when both

terms of the impedance function (stiffness and damping)
simultaneously approach zero, in which case the system may
experience undamped rigid body vertical modes. Theoretically, the
damping coefficient of the impedance function is always higher than
a certain limit. Particularly, as the dynamic stiffness of a foundation-
soil system becomes zero or even negative at resonance, the damping
simultaneously approaches its maximum value, see section 2.5.
3.1.1. Elastically supported beams
The closed-form expression in Eq. 3.16 is not valid for the case of
beams on elastic supports (ES beams). Neglecting the damping terms
in Eq. 3.8, results in a real-valued eigenvalue problem as shown in
Eq. 3.18. The fundamental natural frequency of the ES beam is
always the lowest root of the quartic equation in Eq. 3.19.
(π 2 − 8) 2 4
2
γ ω − γ (D+ 1)ω 2 + 2 D =
0 (3.18)
2π
D + 1 − ( D + 1) 2 − 4 D(π 2 − 8) π 2
ω0,ES = (3.19)
γ (π 2 − 8) π 2
3.1.2. Verification
A benchmark example which was previously solved by Svedholm et

al. [67] is considered to verify the accuracy of the proposed closed-
form expressions. Referring to Fig. 3.1, a beam with a total length
L=42 m, mass per unit length m=18400 kg/m and flexural rigidity
EI=122 GNm2 is considered. The complex modal properties of this
system are calculated for two different cases; a free-free beam (kv≈0)
and an elastically supported beam (kv=550 MN/m); while the dashpot

coefficient cv is varied from 1 kNs/m to 110 MNs/m.
The comparison between the calculated complex modal properties of

the 1st bending mode of the free-free beam (kv≈0) and the elastically
supported beam (kv=550 MN/m) is presented in Fig. 3.2 and Fig. 3.3,
respectively. As can be seen, the calculated undamped natural
frequency and modal damping ratio using the previously introduced
closed-form expressions are in perfect agreement with the results of
the analytical solution.
6 100
(a) (b) Closed-from
ξ0 (%) Analytical [67]
[63]
ffn0 (Hz)
4 50
2 0
0 20 40 60 80 100 0 20 40 60 80 100
cv .L mEI (-) cv .L mEI (-)
st
Fig. 3.2. Modal properties of the 1 bending mode for a free-free beam (kv≈0) for
increasing cv, (a) undamped natural frequency, (b) modal damping ratio
2.3 3
(a) (b)
2.25 2
ffn0 (Hz)
ξ0 (%)
Closed-from
2.2 1
Analytical [67]
[63]
2.15 0
0 20 40 60 80 100 0 20 40 60 80 100
cv .L mEI (-) cv .L mEI (-)
st
Fig. 3.3. Modal properties of the 1 bending mode for an elastically supported beam
(kv=550 MN/m) for increasing cv, (a) undamped natural frequency, (b) modal
damping ratio
3.2. Parametric study
3.2.1. Definitions of the case studies
The parametric study focuses on short to medium size beam bridges,

considering the usual span lengths prone to experience high vertical
accelerations of the deck at resonance, see Table 3.1. The mass per
unit length m and the flexural rigidity EI were assumed to vary with
span length L and estimated based on collected data from existing

single-track ballasted concrete bridges [68]. No structural damping is
assigned to the beam itself. Eliminating the internal damping of the
beam permits a better understanding of the effect of SSI on the global
damping of the coupled system.
Table 3.1. Cross-sectional data for studied single-span beam bridges

L (m) 10 20 40
m (ton/m) 11.2 16.95 28.6
EI (GNm2) 8.74 36 165
*
f0,SS (Hz) 13.8 5.7 2.2
* The natural frequency of the simply-supported (SS) beam
The selected beams are assumed to rest on shallow foundations. The

evolution of the modal properties of the beams is studied for two
different types of subsoil conditions: (i) homogenous half-space
medium; (ii) homogenous shallow stratum over bedrock. In the case
of the half-space medium, the shear wave velocity of the soil is varied
between 150 to 600 m/s in increments of 25 m/s, to cover the lower
and the upper limit of indicated values for granular soils (see Table
2.1). For the case studies supported by shallow stratum over bedrock,
the shear wave velocity of the soil medium is assumed to be constant
at 300 m/s, while the depth of stratum is varied between 1 and 120 m
in small enough increments.
Identical 3.5×7 m2 massless surface foundation slabs are considered

for all cases. Poisson’s ratio, mass density and hysteretic damping of
the soil medium for all cases are assumed constant at νs=0.3, ρs=1800
kg/m3 and ξs=2.5%, respectively. There are some limitations in the
current parametric study: (i) the size of the foundation slab is
considered constant; (ii) the geometry is limited to surface rectangular
foundations while in reality the shallow foundations are practically
embedded; (iii) in the case of very soft deposits, the shallow
foundations are replaced by piled foundations which are not studied
in the current analyses; (iv) the effect of the foundation mass is
neglected.
As the response of the bridge-soil system is governed by the

viscoelastic properties of the beam and supporting subsoil, the
following dimensionless parameters are defined to study the evolution

of the modal properties of the coupled systems:
(i) The relative stiffness parameter, k, defined as the ratio of

the flexural rigidity of the beam to the static stiffness of the
elastic support.
k = π 4 EI (2 L3kv,0 ) (3.20)
(ii) The wave parameter, 1/σ which represents the beam-to-soil

dynamic stiffness ratio.
1 s = L. f 0,SS Vs (3.21)
(iii) The relative frequency parameter, f, defined as the ratio of

the fixed-based natural frequency of the beam to the
resonant frequency of the supporting stratum for
compressing modes. This parameter is used to study the
sensitivity of the modal properties of the system to the
depth of the stratum.
=f f=
0,SS f c 4 Hf 0,SS VLa (3.22)
It should be noted that, since the dynamic stiffness of the foundation

kv (ω) and corresponding damping coefficient cv(ω) are frequency
dependent variables, an iterative procedure should be used for
calculating the modal properties of the system using the closed-form
expressions.
3.2.2. Results of the parametric study
3.2.2.1. Beams supported by surface foundations on half-space medium
In what follows, the evolution of the modal properties of beams

supported by surface foundation rested on a homogenous half-space
medium is studied using the proposed closed-form expressions. In the
analysis, the shear wave velocity of the soil medium is varied
between 150 to 600 m/s in increments of 25 m/s. The variation of the
frequency of the coupled systems as a function of the relative

stiffness parameter k is presented in Fig. 3.4a.
0.95
(a)
f0 / f0,SS
0.9
L=10 m
0.85 L=20 m
0.8 L=40 m
ES beam
0.75
15
10
ξξ0 (%)
(%)
(b)
5
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
k (-)
Fig 3.4. Variation of the (a) fundamental frequency and (b) foundation damping
ratio vs. relative stiffness parameter (k) for beams supported by half-space medium
The dashed black line represents the natural frequency reduction of

the corresponding ES beams (beams on only elastic springs). Fig 3.4b
and 3.5 shows the variation of the foundation damping as a function
of the relative stiffness parameter k and the wave parameter (1/σ),
respectively.
15
L=10 m
L=20 m
10 L=40 m
(%)
ξξ0(%)
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1/σ (-)
Fig 3.5. Foundation damping ratio variation as a function of wave parameter (1/σ)
for beams supported by half-space medium
The following conclusions can be drawn from the analysis of the

results presented in Fig. 3.4 and 3.5:
(i) It is well known that the flexibility of the supports lowers

the fundamental natural frequency of the ES beams, as
shown in Fig. 3.4a. While this is also the case for the studied
viscoelastically supported (VS) beams when SSI is included,
the deviation between the natural frequency shortening of
the VS beam with respect to the ES beam becomes higher
for shorter beams and also softer soil conditions. The
presence of the foundation damping results in higher natural
frequencies compared to the case of ES beams in which the
effect of foundation damping is neglected.
(ii) The dissipation capacity of the foundation-soil system is

negligible for a flexible beam on very stiff soil. An increase
in the additional damping of the system due to SSI becomes
significant as the ratio between the flexural stiffness of the
beam and the foundation-soil stiffness becomes higher.
Finally, considering for the supporting soil a mean shear wave

velocity Vs of 300 m/s the calculated foundation damping and
natural frequency shortening of the system for the three cases are
presented in Table. 3.2.
Table 3.2. Comparison between frequency shortening and foundation

damping of the studied beams on half-space medium with Vs=300 m/s
L (m) ξ0 (%) f0,VS/f0,SS f0,ES/f0,SS
10 3.1 0.95 0.93
20 0.9 0.97 0.96
40 0.3 0.98 0.98
3.2.2.2. Beams supported by surface foundations on stratum over bedrock
In this section, the effect of SSI on the dynamic response of beams

supported by shallow foundations rested on homogenous stratum is
studied. The shear wave velocity of the soil medium is assumed to be
constant as 300 m/s, while the depth of stratum is varied between 1 to
120 m using small enough increments.
The variation of the frequency shortening and foundation damping

ratio as a function of the relative frequency parameter f (ratio
between natural frequency of the beam and the stratum) is presented

in Fig. 3.6a and 3.6b, respectively. As can be seen, for cases where
the resonance frequency of the beam (f0) is lower than the resonance
frequency of the stratum (fc), the foundation damping term decreases
and can be considered as zero if the f is small enough (f0 < 0.5fc). As
discussed in section 2.3, this is due to the fact that at operating
frequencies lower than the stratum’s resonant frequency, the radiation
damping is zero and the foundation damping only originates from the
hysteretic damping of the soil medium.
0.95
f0 / f0,SS
0.9
(a)
0.85
15 L=10 m
L=20 m
(b)
10 L=40 m Half-space
(%)
ξξ0 (%)
values
5 Negligible
0
0 0.5 1 1.5 2 2.5
f (-)
Fig 3.6. Variation of (a) fundamental frequency and (b) foundation damping ratio as
a function of the relative frequency parameter (f) for beams supported by stratum
over bedrock
In the range of operating frequencies near the resonant frequency of

the deposit (1< f <1.5), the estimated foundation damping of the beam
on a stratum is higher than for the case of a beam on a half-space
medium, as shown in Fig. 3.6b. This is because the dynamic stiffness
of the foundation-soil system significantly decreases near its
resonance frequency while at the same time the radiation capacity of
the deposit is resumed. Consequently, the combination of low
dynamic stiffness and high damping coefficient at supports results in
lower natural frequency and higher modal damping ratio for the
fundamental bending mode of the system.
For beams with fundamental natural frequencies beyond 1.5fc, both

the frequency shortening and foundation damping of the system are
very close to the “half-space” values. In other word, the depth of the
stratum becomes large enough to be approximately considered as a
half-space medium.
3.3. Case studies: Train-induced vibration analysis
3.3.1. Effect of SSI on the train-induced resonant response
In order to examine some of the main findings outlined during the

parametric analysis, one of the case study bridges (L=10 m;
f0,SS=13.9 Hz) is analyzed under the passage of German ICE3 train.
The structural damping of the beam which is neglected in the
parametric studies in section 3.2, is incorporated in the train-induced
vibration analysis and assumed to be 1 % for all cases which is the
recommended lower bound for reinforced and prestressed concrete
bridges. It should be noted that, in the train-induced vibration
analyses of this chapter, the longitudinal distribution of the axle loads
by the rails has been considered by spreading each axle load over
three rail supports [5].
First, the effect of the shear wave velocity of the supporting half-
space medium on the train-induced resonant response of the beam is
studied. Four different shear wave velocities for the half-space
medium are considered, including an infinitely rigid soil which
represents fixed supports. The results are presented in Fig. 3.7. As can
be seen, the predicted maximum acceleration for the case with
infinitely rigid soil exceeds the serviceability limit state for the
vertical acceleration [5]. In the presence of SSI, the resonant speeds
are slightly decreased due to the frequency shortening of the system.
Furthermore, the resonant amplitudes gradually decrease as the soil
becomes softer, due to an increase in the dissipation capacity of the
foundation-soil system.
10
C
Vss=inf
= inf
8
C
Vss=600
= 600m/s
m/s
amax (m/s2)
6 C
Vss=300
= 300m/s
m/s
4 C
Vss=150
= 150m/s
m/s
2
0
150 200 250 300 350 400 450
v (km/h)
Fig. 3.7. Effect of the shear wave velocity of the half-space medium on the
maximum acceleration at mid-span for the case study bridge (L=10 m), under
passing of the ICE3 train
Further, the effect of the depth of the supporting stratum on the train-
induced resonance response is also investigated. In this analysis, the
shear wave velocity of the soil is assumed to be constant as Vs=300
m/s. Four different depths of the stratum are considered, including
zero depth which represents fixed supports. As shown in Fig. 3.8, the
behavior of the system is consistent with the findings in section
3.2.2.2.
10
m (f =0)
H=0(fs/fc=0)
H=0
8
m (f =0.4)
H=4(fs/fc=0.4)
H=4
amax (m/s2)
6 H=9 m (f =1.0)
H=9(fs/fc=1.0)
H=16
H=16(fs/fc=1.8)
m (f =1.8)
4
0
150 200 250 300 350 400 450
v (km/h)
Fig. 3.8. Effect of the stratum depth (VS=300 m/s) on the maximum acceleration at
mid-span of the bridge (L=10m), under passing of the ICE3 train
For H=4 m, since the frequency of the bridge is far lower than the
resonance frequency of the stratum, f0,SS/fc < 0.5, the amplitude of the
peak response was only affected by the hysteretic damping of the soil
while the peak was shifted to lower speeds due to frequency
shortening. In the case of H=16 m, when the f0,SS/fc >1.5, similar
results are obtained as for the beam supported by half-space (see Fig.
3.7). Furthermore as discussed in section 3.2.2.2, if the natural
frequency of the beam was close to the resonance frequency of the

stratum, f0,SS/fc≈1 (H=9 m), the frequency shortening and damping
ratio of the systems were at their peak values (see also Fig. 3.6).
3.3.2. Equivalent frequency-independent spring/dashpots
As it was mentioned in the section 2.4, in the case of a surface

foundation on dry half-space medium with usual dimensions (i.e.
L/B=2-3), the variation of the impedance functions’ coefficients in the
range of interesting frequencies is negligible (see also Fig. 2.6).
Therefore, equivalent constant values can be used for a surface
foundation on half-space as follows.
ξs kv,eq
kv,eq = kv,0(halfspace) , cv,eq
= cv,0 + (3.23)
πf 0,SS
In the case of shallow foundation on a stratum with depth H, the

following simplified approach can also be implemented:
a) If f0,SS>1.5VLa/4H, then the case may be simplified by using

the equivalent constant values of the half-space case (Eq.
3.23).
b) If f0,SS<0.5VLa/4H, then the case may be simplified by using

the following equivalent constant values for a surface
foundation on stratum:
ξs kv,eq
kv,eq = kv,0 (stratum) , cv,eq = (3.24)
πf 0,SS
Comparisons between train-induced vibrations computed based on

the simplified constant values and reference model are presented in
Fig. 3.9 and 3.10. As can be seen, the proposed approximation of the
frequency dependency of the impedances is in good agreement with
the results of the reference model.
3
kv(ω), cv(ω)
Closed-from
2
kv,eq, cv,eq[63]
Analytical
(m/s2)
max
1
a
0
150 200 250 300 350 400 450
v (km/h)
Fig. 3.9. Maximum acceleration at mid-span of the beam (L=10m) under
passing of ICE3 train, based on equivalent constant impedance values for a
supporting half-space with VS=300 m/s
6 3
Equi. static ksv, csv (a) kv,eqstatic
Equi. , cv,eqksv, csv (b)
k(f), c(f) kv(ω
k(f), ), cv(ω)
c(f)
4 2
amax (m/s2)
amax (m/s2)
2 1
0 0
150 200 250 300 350 400 450 150 200 250 300 350 400 450
v (km/h) v (km/h)
Fig. 3.10. Maximum acceleration at mid-span of the beam (L=10m) under
passing of ICE3 train, based on equivalent constant impedance values for a
supporting stratum (a) f 0,SS < 1.5VLa 4 H , (b) f 0,SS > 1.5VLa 4 H
4. Dynamic SSI analysis of portal frame bridges
Portal frame (integral) bridges are the most common type of
underpasses along modern railway lines. This type of bridge is
designed as a reinforced concrete rigid frame with integral wing walls
and is surrounded by an embankment. Theoretically, the dynamic
response of these partially-buried stiff structures is expected to be
governed by the surrounding soil. However, the effect of the soil is
usually disregarded in the train-induced vibration analysis due to the
high computational cost and lack of efficient simple models. The
presence of backfill soil may provide an additional stiffness to the
system which leads to an increase in the fundamental natural
frequency and thus shifts the critical resonant speed to a higher value.
The dissipation capacity of the backfill soil could also lead to an
increase in the global damping of the system and consequently reduce
the amplitude of the resonant response of the bridge.
In paper II, the results of the SSI analysis of a portal frame bridge
using a direct approach (full FE modeling) have been presented in
detail. In this chapter, the effect of the surrounding soil, especially
backfill soil, on the dynamic response of portal frame bridges is
investigated using two different substructure approaches. First, a 3D
coupled Finite Element-Boundary Element (FE-BE) model is
developed (section 4.1). Using this model, the effect of the
surrounding soil conditions on the modal properties and resonant
response of a case study bridge is investigated. In section 4.2, the
reliability of a proposed simplified model aimed at representing the
interaction between embankment and abutments is examined.
4.1. Coupled FE-BE model in the frequency domain
To study the effect of the surrounding soil on the dynamic response of

frame bridges as well as their modal properties, a 3D coupled FE-BE
model was used. The bridge was modeled with finite elements, while
42 | DYNAMIC SSI ANALYSIS OF PORTAL FRAME BRIDGES
a boundary element model simulated the infinite extent of the backfill

and underlying soil. The general outline of the coupled system, 3D
FE model of the frame bridge and boundary element discretization of
the bridge-soil interfaces, is shown in Fig. 4.1. The FE model of the
bridge was defined using a FE toolbox for Matlab [69]. The
discretization used 4-node rectangular thin shell (Kirchhoff) elements.
A BE mesh was also defined. The BE mesh matched the finite
element mesh on the surface of the abutment walls as well as the
bottom plates (see Fig. 4.1). 4-node boundary elements were used
with 2 × 2 Gaussian integration points per boundary element. Matlab
toolbox BEMFUN [70] was used to compute the BE system matrices.
Green’s functions for a horizontally layered soil medium are
integrated in the BE formulation [45, 56], avoiding the need for
discretizing the free surfaces. The coupling procedure in [70, 71] (see
also sections 2.3.2 and 2.3.2.1) was implemented to couple the two
numerical models. The BE system matrices relate the displacement
and traction vectors at the BE nodes, using Eq. 4.1 to calculate the
dynamic stiffness matrix of the soil.
Tq U -1 (ω ) ( I +T(ω ) )
K gff (ω ) + i ωCgff (ω )  = (4.1)
In Eq. 4.1, U and T are the displacement and traction boundary

element system matrices, respectively. Tq and I are the stress transfer
and identity matrices, respectively [70, 71].
z
y
(a)
x
(b)
= +
(c)
Fig 4.1. (a) Configuration of the coupled FE-BE models, (b) FE mesh, (c) BE
mesh
DYNAMIC SSI ANALYSIS OF PORTAL FRAME BRIDGES | 43
It should be noted that the geometry of the bottom slabs was not
properly modeled in the current coupled model. This is due to the
coupling issues related to the embedded parts of the foundation slabs.
In addition, the effect of the integrated wing-walls was also
disregarded. In future studies, the coupled model should be developed
to properly model these components.
4.1.1. Verification
In this section, a benchmark example previously solved by Coulier et

al. [72] is considered to verify the numerical implementation of the
coupled FE–BE model. The vertical receptance transfer function at
the center of a flexible square surface foundation resting on a
horizontally layered half-space was calculated within a frequency
range of 0-50 Hz. The dimensions of the concrete foundation were
5 m x 5 m with a thickness of 0.25 m. Young’s modulus, Poisson’s
ratio and the density of the concrete were assumed to be 33 GPa, 0.2
and 2500 kg/m3, respectively. Rayleigh damping coefficients
corresponding to a modal damping ratio equal to 3% were used. The
configuration and properties of the soil layers are presented in Fig.
4.2a. The density, Poisson’s ratio and material damping ratio for all
soil layers were assumed to be 1800 kg/m3, 1/3 and 2.5%,
respectively. The foundation slab was discretized using 30 by 30 thin
shell (Kirchhoff) elements, coupled to the corresponding BE mesh on
the surface of the soil medium. The vertical displacement at the center
of the slab is shown in Fig. 4.2b.
Psin( ωt) 4
FE model
d (mm/MN/Hz)
BE mesh Real
2
2m Vs = 150 m/s 0
--------------------------
2m Vs = 250 m/s
------------------------- Imag.
Half -2
Vs = 350 m/s
-space 0 10 20 30 40 50
(a) (b) f (Hz)
Fig 4.2. (a) Configuration of the flexible surface foundation on a layered half-
space, (b) calculated real and imaginary part of the vertical receptance at the
center of the flexible foundation using the implemented FE-BE model (black solid
line) and the solution by Coulier et al. [72] (red circles)
As can be seen, a perfect match is obtained between the results from

the implemented coupled model and the benchmark solution. The
peak at around 19 Hz corresponds clearly to the resonance of the
foundation on the layered half-space [72].
4.1.2. Case study bridge
The case study is a short-span portal frame bridge, which is partially

embedded in a half-space medium. The geometry of the bridge is
presented in Fig. 4.3a. The maximum element size was set to 0.5 m.
Young’s modulus, Poisson’s ratio and the density of the concrete
were assumed to be 30 GPa, 0.2 and 2500 kg/m3, respectively.
Rayleigh damping coefficients with a modal damping ratio equal to
2 % for a frequency range of 20-70 Hz were used. The mass of the
ballast layer was added to the bridge deck. The configuration and
properties of the surrounding soil layers are presented in Fig. 4.3b.
Hysteretic damping ratio in the soil layers was assumed to be constant
at 2.5 %.
In order to identify the influence of the surrounding soil on the

dynamic behavior of the portal frame bridge, a comparison was
performed between calculated modal properties and train-induced
resonant response of the selected bridge for the following cases: (1)
frame bridge, neglecting the effect of surrounding soil; (2) frame
bridge, considering the effect of backfill soil only (fixed bottom
slabs); (3) frame bridge, considering the effect of backfill soil and
subsoil.
t=0.5 6
VS=250 m/s
t=0.5 z
y
ν=0.2
x
ρ = 1800 kg/m3
5
Rigid VS=400 m/s, ν=0.48
ρ = 2000 kg/m3
6
2 (a) (b)
Fig 4.3. (a) Geometry of the case study bridge, (b) configuration and properties
of the surrounding soil
4.1.2.1. Eigenvalue analysis
The presence of soil-structure interaction in cases 2 and 3 leads to a

non-proportionally damped system. The quadratic eigenvalue
problem for a non-proportionally damped system is given by Eq. 3.7.
Due to the frequency-dependence of the dynamic stiffness matrix of
the soil medium, an iterative approximate method is implemented to
estimate the modal properties of the studied non-proportionally
damped system. The general steps to estimate the modal parameters
for each mode of vibration can be summarized as follows:
1- Initial estimation of the nth natural frequency, ωn, based on the

corresponding standard eigenvalue analysis of the model
without SSI (case 1).
2- Calculate frequency-dependent stiffness and damping matrices

of the surrounding soil using Eq. 4.1.
3- Assemble the global stiffness and damping matrices
K(ω ) = K ss +K ffg (ω )
 (4.2)
C(ω=
) Css + Cffg (ω )
4- Numerically solve the quadratic eigenvalue problem using

polynomial eigenvalue solver in Matlab [73, 74]
[ϕ , l ] =polyeig ( K(ω ), C(ω ), Mss ) (4.3)
where ϕ and λ are the complex-valued eigenvectors and eigenvalues

of the system, respectively.
Having solved Eq. 4.3, the natural frequency and damping ratio of the
nth mode are evaluated using Eq. 3.10 and 3.11, respectively. The
procedure repeats steps 2 to 4 until the difference between initial and
estimated natural frequencies becomes negligible.
Fig. 4.4 illustrates some of the calculated mode shapes of the studied
bridge for case 1. The estimated modal parameters of the system, for
the three different cases, are presented in Table 4.1.
(a) 1st bending (b) 1st torsional (c) 2nd bending (d) 1st plate mode
Fig 4.4 Calculated mode shapes for case 1 (No SSI)
Table 4.1. Calculated natural frequencies and damping ratios for the
studied bridge
Case 1 Case 2 Case 3
Mode (No SSI) (No subsoil) (Full model)
f(Hz) ξ(%) f(Hz) ξ(%) f(Hz) ξ(%)
1st bending 24.5 26.6 6.6 25.4 13.1
1st torsional 31.9 34.4 6.1 36.2 8.1
2.0
2nd bending 58.4 -* > 30 -* > 40
1st plate mode 63.6 64.1 2.4 64.1 2.4
* was not converged
As can be seen, the stiffness and damping of the surrounding soil lead
to considerable changes in the modal properties of the fundamental
(vertical) bending mode of the case study bridge. The calculated
modal damping ratios for the vertical bending and the torsional modes
were considerably higher than for the plate mode. This is most likely
due to the stronger effect of soil-structure interaction for the bending
and torsional modes. Comparison between calculated modal
properties of the plate mode for different cases showed that this mode
was not considerably affected by SSI. This is due to the small relative
motion between deck and abutments. On the other hand, the
configuration of the vertical bending modes of the deck, by structural
compatibility, is associated to bending of the abutment walls. Hence,
the presence of the backfill soil will significantly affect the modal
properties of these modes. This effect is obviously depending on the
relative stiffness between deck and walls. A detailed discussion
concerning the effect of the SSI on the modal parameters of different
modes in portal frame bridges was presented in the appended papers.
4.1.2.2. Train-induced resonance response
Fig. 4.5a presents the calculated accelerance transfer function at the

mid-point of the deck for different cases, whereas the results of the
train-induced vibration analyses are also presented in Fig. 4.5b. As
shown in Fig. 4.5a-b, the effects of the surrounding soil, especially
the radiation damping, substantially reduce the vibration amplitudes
of the bridge.
Case 1 Case 2 Case 3
0.6 10
(a) Case 1 (b)
a (m/s /kN)
8
0.4 Case 2 amax (m/s )
2
6
2
Case 3
4
0.2
2
0 0
10 15 20 25 30 35 40 150 200 250 300 350 400 450
f (Hz) v (m/s)
Fig. 4.5. Influence of surrounding soil conditions on the dynamic response of the
studied bridge, (a) FRFs at mid-span, (b) maximum train-induced acceleration at
mid-span due to the HSLM-A2 train
4.2. Simplified modeling
The objective of this section is to propose a simplified modeling

alternative for the surrounding soil which can predict the dynamic
response of portal frame bridges with reasonable accuracy. Such
simplified models are useful for practical design purposes as well as
parametric studies. As discussed in section 2.3.2.2 and chapter 3,
assuming a rigid behavior of foundation the dynamic interaction
between the bottom plates and the subsoil can be modeled using
impedance functions. However, modeling the interaction between
abutment and backfill soil is not as straightforward as the foundation-
subsoil interaction.
As it was briefly discussed in section 2.3.2.2, for the case of linear

elastic analyses, several approximate methods can be used to model
the dynamic interaction between abutment walls and backfill soil [46-
50]. In this research, the abutment–backfill interaction in the normal
and tangential directions was modeled using linear frequency-

independent spring/dashpots distributed over the whole area of the
abutment (see Fig. 2.5). The properties of the spring/dashpot system
in the normal direction were defined based on the methods proposed
in [47, 48]. The properties of the spring/dashpots in the two tangential
directions were roughly estimated to be a fraction of their values in
the normal direction. A detailed explanation concerning the
configuration and properties of the implemented spring/dashpot
system can be found in Paper II.
Fig. 4.6 shows the comparison between the results evaluated using the
simplified model on one hand and the coupled FE-BE model on the
other hand. A reasonably good agreement between the results
obtained using the two models can be noted in as far as the overall
shape and vibration amplitudes are concerned. However, a more
detailed comparison between the predicted vibration amplitudes (see
Fig. 4.6a-b), indicates that the simplified model slightly overestimates
the modal damping ratios. The reason for this discrepancy lies in the
approximate nature of the simplified model which should be
investigated more in future studies. Note also that, the comparison
was restricted to case 2 due to the improper geometry of the
foundation slabs in the current coupled FE-BE model.
0.8
a (m/s /kN)
Case 2 (FE-BE model) 4

0.6
amax (m/s )
Case 2 (Simplified model)

2
2
0.4
(a) 2 (b)
0.2
0 0
10 20 30 40 50 60 70 150 200 250 300 350 400 450
f (Hz) v (m/s)
Fig. 4.6. Comparison between predicted results by the coupled FE-BE model and
simplified model, (a) vertical accelerance at mid-point, (b) maximum train-induced
acceleration at mid-span due to the HSLM-A2 train
5. Experimental validation of numerical models
Despite recent development in advanced mathematical models aimed
at evaluating the dynamic stiffness of the structure-soil systems, the
lack of reliable validation of numerical models against in-situ
measurements is still noticeable. To overcome this shortcoming,
system identification (SI) techniques are helpful. SI is a powerful tool
for inverse estimation of the dynamic properties (i.e. frequency
response functions and, modal properties) of a real bridge-soil system
[75]. The principal aim of SI is to obtain information from an
experiment and use it to validate a numerical model by adjusting its
parameters and minimizing the deviation between simulated and
measured results [76].
In this chapter, a summary of the experimental results of the full-scale

testing on two portal frame bridges and experimental validation of the
numerical models is presented. The chapter is organized as follows.
In section 5.1, a short description of the tested bridges is presented.
Section 5.2 briefly describes the experimental testing procedure. The
results of the post-processing of the experimental data are also
concisely presented. Finally, a summary of the experimental
validation of the numerical models for the case study bridges is
presented in section 5.3.
5.1. Description of the tested bridges
The Faresmyran and Degersmyran bridges are two portal frame

railway underpasses with a single span of 8 m and 7 m, respectively.
Both bridges are reinforced concrete rigid frames with integral wing
walls and were designed for both freight trains (33 ton/axle) and high-
speed trains with the maximum speed of 250 km/h. They are located
on a single track ballasted railway line in the North of Sweden.
Photos of the Faresmyran and Degersmyran bridges are shown in Fig.
5.1 and Fig. 5.2, respectively. The general arrangement and
50 | EXPERIMENTAL VALIDATION OF NUMERICAL MODELS
significant features of the structure and surrounding soil for both case
studies are summarized in Fig. 5.3. The basic information about the
geotechnical properties of the geo-materials at the site, such as type of
backfill material, elevation of bedrock and ground water level, were
found in the geotechnical reports and as-built drawings of the project
(see Fig. 5.3).
(a) (b)
Fig. 5.1. (a) photo of the Faresmyran bridge, (b) view of the hydraulic exciter
(a) (b)
Fig. 5.2. (a) photo of the Degersmyran bridge, (b) view of the hydraulic exciter
According to the geotechnical data, the foundation slabs of the

Faremyran bridge were placed on a 1.2 m thick layer of compacted
crushed rock. The geotechnical survey indicated that the subsoil is an
approximately 5 m dense moraine stratum over bedrock. In the case
of the Degersmyran bridge, the foundation slabs were placed on a
layer of compacted crushed rock, approximately 0.5-1 m thick and
laid directly on the bedrock. In both cases, the abutments have been
surrounded by compacted layers of crushed rock and the ground
water level has always been kept lower than the bottom of the
EXPERIMENTAL VALIDATION OF NUMERICAL MODELS | 51
foundation slabs. As the geometry of both bridges as well as the

properties of the backfill soil are very similar, the main difference
between the two cases is the depth of the subsoil stratum underneath
the foundation slabs (see Fig. 5.3).
5.2. Experimental testing
The experimental investigation aimed at validating the accuracy of

the numerical models used for evaluating the dynamic stiffness and
modal properties of the bridge-soil system. For this purpose, apart
from monitoring the dynamic response from passing trains, a set of
controlled vibration tests has been performed.
5.2.1. Forced vibration tests
The controlled vibration tests were performed on the bridge deck by

using a hydraulic bridge exciter [77] in the vertical direction. Fig. 5.3
shows the locations of the exciter and the arrangement of the
accelerometers. Based on the results of pre-test FE analyses, it was
decided to locate the exciter at two different points (P1 & P2), as
shown in Fig 5.3, in order to excite several modes of vibration.
Sixteen uniaxial accelerometers (SiFlex-SF1500S) denoted as a1 to
a16 were installed on the bridge in such a manner that the bending,
torsional and plate modes of the deck could be captured with
reasonable accuracy.
In each dynamic test, the force amplitude was constant, and a linear
frequency sweep was carried out at an adequately small rate to
achieve a steady-state response. A HBM MGCPlus data acquisition
system was used for collecting data with a sampling frequency of
1200 Hz. A low-pass filter at 80 Hz was used to remove the high
frequency contents from the measured signals. To obtain high-
precision measurement with good coherence, most of the tests were
repeated at least two times. The dynamic tests were performed in the
time intervals between regular traffic in service. Thus, the
experimental results were not affected by train passages. In this
report, a summary of the experimental results and identified modal
parameters based on the forced vibration tests is presented. More
detailed information about the post-processing of the experimental
data, precision of the measured FRFs and amplitude dependency of

the results can be found in papers I and II.
(a)
a5 a6 a7 a8
3.65 (3.55)
P2
a9 a10 a11
3.65 (3.55)
P1
a1 a2 a3 a4
)
.2
(6
L/5 L/5 L/5 L/5 L/5

1
6.
0.5 a12
L=8 (7) (b)
~ 0.6, Ballast
~ 0.6, Sub-Ballast 0.62 (0.55)
0.5
a13 a15
1.7
a14 a16
Exciter
Backfill material
5.5 (5.9)
1:
1.
5
Filling
Material 2.8 (3.4) 0.7
G.W.L
Compacted fill ~ 1.2 (0.75)
~ 5 (0) Moraine
Fig. 5.3. Geometry, surrounding soil properties and instrumentation, (a) Plan, (b)
Elevation (units are in m). The values in parentheses belong to Degersmyran bridge.
5.2.1.1. Summary of the identified modal parameters
A simple peak-picking technique [78] was used to estimate the modal

parameters of the system. The Half-Power Bandwidth method was
used to estimate the modal damping ratio. The identified mode shapes
were estimated based on cubic interpolation between measured
accelerations. Fig. 5.4 illustrates the experimentally identified mode

shapes of the system. The mean values of the estimated modal
parameters, frequencies f and modal damping ratios ξ, for the
identified vibration modes are presented in Table 5.1.
(a) (b) (c) (d)
Fig. 5.4. Experimental mode shapes (a) 1st bending mode, (b) 1st torsional mode, (c)
1st plate mode, (d) 2nd plate mode
Table 5.1. Estimated modal parameters for case study bridges

f1 ξ1 f2 ξ2 f3 ξ3 f4 ξ4
Bridge
(Hz) (%) (Hz) (%) (Hz) (%) (Hz) (%)
Faresmyran 29.8 16.4 44.9 6.5 -- -- 67.3 3.5
Degersmyran 31.6 9.1 47.7 6.9 67.7 2.6 71.5 2.5
By comparing the identified modal parameters for the two bridges,

the following can be concluded:
1- As it was mentioned in section 5.1, the geometry and

surrounding soil condition of the tested bridges were relatively
similar, except the depth of subsoil stratum. Consequently, the
shape of the measured FRFs and identified modes for the two
bridges were rather close. However, the Degersmyran bridge
evidently showed stiffer behavior as its span length is 1 m
shorter.
2- A large difference in the modal damping ratios for the 1st

bending mode of the bridges has been detected. This is
probably due to the effect of different subsoil conditions. Thus,
for the Degersmyran bridge with shallow subsoil condition the
damping ratio for the 1st bending mode was about 9%. For the
Faresmyran bridge, placed on a deeper subsoil layer, the
damping ratio for the 1st bending mode was about 16 %.
3- No considerable difference between the modal damping ratios

of the torsional modes has been detected. Probably, this mode
is mainly affected by the backfill soil and subsoil conditions do
not play an important role for this mode of vibration.
4- For both bridges, the estimated modal damping ratios for the
bending and torsional modes were considerably higher than for
the plate modes. This is most likely due to soil-structure
interaction effects which are more pronounced for the bending
and torsional modes. As shown in Fig. 5.4 and 4.5, the
configuration of the plate mode of the deck was rather
independent of the movement of the abutments.
5.2.2. Monitored train passages
In addition to the controlled vibration tests, the dynamic response of

the bridge from passing trains was also monitored. The train-induced
vibrations were measured under the passage of X62 commuter train.
The configuration of the train can be found in paper II. The speed of
the passing trains at the Faresmyran and Degersmyran bridge sites
was estimated to be about 160 km/h and 180 km/h, respectively.
The monitored time responses of the vertical deck acceleration at

sensor location 2 are shown in Fig. 5.5a-b. The signals were filtered
at 200 Hz. Fig. 5.5 also depicts the time-frequency scale-space of the
signal based on a Continuous Wavelet Transform (CWT) [79, 80].
This gives a contour of the energy density of the monitored signal in
the time-frequency domain. As can be seen, the monitored signals for
both bridges contain a high energy concentration in the range of 60-
90 Hz, corresponding to high acceleration levels particularly during
the passage of the first and last bogies.
As discussed in Paper II, the phenomenon is related to the close

match between the sleeper-passing frequency, the wheel–track
resonance frequency and natural frequency of the plate mode of the
bridge. For a sleeper distance of 0.6 m and the monitored train speeds
previously mentioned, the sleeper passing frequencies are about 73
Hz and 83 Hz for Faresmyran and Degersmyran bridges, respectively.
These frequencies are relatively close to the identified frequency of
the plate modes of the deck for two bridges. In both cases, the
amplification is especially noticeable for the first and last bogies
which have a wheelbase (2.4 m) that is an exact multiple of the
sleeper distance (0.6 m).
5 5
a2 (m/s )
a2 (m/s )
2
2
0 0
-5 -5
160 160
140 140
Pseudo - f (Hz)
120
Pseudo - f (Hz)
120
100 100
80 80
60 60
40 40
20 20
0 0.5 1 1.5 2 0 0.5 1 1.5 2
t (s) t (s)
(a) (b)
Fig. 5.5. Time history of the vertical acceleration during train passage at sensor
location 2 and corresponding time-frequency scale-space of the signal by CWT,
(a) Faresmyran bridge, (b) Degersmyran bridge
5.3. FE model calibration
In this section, the summary of the identification results and

experimental validation process of the FE models are presented. The
detailed results of the system identification are presented in Paper I
and II.
5.3.1. Numerical models
In the case of the Faresmyran bridge, a 3D solid model for the bridge
and part of the track over the bridge deck was used (see Fig. 5.6). The
model was built in Abaqus/Standard (V6.14) [81]. The simplified
substructure method, using impedances and Winkler-based
spring/dashpots (as discussed in sections 2.3.2.2 and 4.2) was used to
model the SSI.
For the Degersmyran bridge, a full 3D solid model involving the

entire bridge-track-soil system (direct approach) was implemented.
(see Fig. 5.7).
Fig. 5.6. Configuration of the FE model & boundary conditions, Faresmyran bridge
Fig. 5.7. General configuration of the full FE model of the bridge-soil system,
Degersmyran bridge
5.3.2. System identification results
Based on the preliminary sensitivity analyses, the following

parameters of the model were selected as identification parameters:
• Reinforced concrete – Young’s modulus, mass density and

hysteretic damping ratio
• Ballast – mass density and shear modulus
• Soil layers – shear modulus
The model updating was performed as a nonlinear optimization

problem to minimize the error between the experimental and
calculated FRFs using the Genetic Algorithm. Detailed information
about the model updating procedure and sensitivity analyses is
presented in Papers I and II. Table 5.2 presents the updated values of
the identification parameters for both bridge-soil systems. Note that
the updated values of the identification parameters are similar for the
two bridges. The comparison between the measured FRFs and
predicted ones based on the calibrated FE models is shown in Fig.
5.8a for the Faresmyarn bridge and 5.8b for the Degersmyran bridge.
Table 5.2. Updated identification variables for two calibrated bridge-soil models
Subsoil
Concrete Ballast Backfill
(Saturated)
Bridge
Ec ρc ξc Ebl ρbl VS,bk* VS,s
3 3
(GPa) (t/m ) (%) (GPa) (t/m ) (m/s) (m/s)
Faresmyran 39 2.3 2.5 0.11 1.4 318 385
Degersmyran 37.5 2.3 1.7 0.12 1.5 311 445
* Average value
0.8
Measurement (Test P1)
(a) Updated FE model
|a| (m/s2/kN)
0.6
0.4 (b)
0.2
0
20 30 40 50 60 70
f (Hz)
Fig. 5.8. Comparison between experimental and numerical FRFs at sample sensor
location 2, test P1, for (a) Faresmyran bridge (b) Degersmyran bridge
Table 5.3 also presents the comparison between experimentally and

numerically identified modal parameters for the two bridges. As
intended, the calibrated FE models can successfully replicate the
experimental results.
Table 5.3. Experimentally and numerically identified modal parameters

Faresmyran bridge Degersmyran bridge
Mode Test Calib. model Test Calib. model
f ξ f ξ f ξ f ξ
1st bending 29.8 16.4 29.2 13.9 31.6 9.1 32.1 6.1
1st torsional 44.9 6.5 43.9 7.1 47.7 6.9 48.5 6.9
1st plate mode - - - - 67.7 2.6 - -
2nd plate mode 67.3 3.5 68.5 3.6 71.5 2.5 72.3 2.9
5.3.3. Validation against monitored train passages
The reliability of the calibrated FE models was also validated by

correlating the monitored acceleration signals due to passing trains
and those simulated by the FE models. Fig. 5.9 compares the
monitored and predicted results for the Degersmyran bridge. As can
be seen, the calibrated FE model can also simulate the train-induced
vibration close to its measured counterpart if the vibration of the
wheel mass traveling on the discretely supported rail is included in
the model. The information about the modeling details of the train
and track is presented in paper II.
5
(a)
a2 (m/s )
2
-5
0.15 0.5 1 1.5 1.85
t (s)
0.6
Measurement
(b) Simulation, wheel model
0.4 Simulation, moving force
a2 (m/s /Hz)
2
0.2
0
0 50 100 150 200
f (Hz)
Fig. 5.9. Comparison between measured and simulated train-induced accelerations
at sensor location 2 for Degersmyran bridge (a) time signal (b) frequency content
6. Discussion and conclusion
In this chapter, the results of the work on dynamic soil-structure
interaction analysis of railway bridges presented in this thesis will be
discussed and some general conclusions will be presented.
6.1. Simple beam bridges
In this research, a novel analytical method for calculating the

dynamic characteristics of the fundamental bending mode of beam
bridges on viscoelastic supports was proposed. Using this method, the
effect of the dynamic stiffness of the foundation on the fundamental
modal parameters (e.g. natural frequency and damping ratio) of
railway beam bridges was investigated and discussed. The proposed
closed-form expressions, in combination with the impedance
functions of different foundation-soil systems, can clarify the main
features of dynamic SSI analysis of these structures and lead to
review the recommended modal damping ratios in the code
provisions and design manuals. The main conclusions of the study
are:
• In the case of bridges rested on half-space medium, the additional

damping of the system due to SSI increases as the ratio between
the flexural stiffness of the beam and the foundation-soil stiffness
becomes higher. The presence of the foundation damping in VS
beams results in higher natural frequencies in comparison to the
case of ES beams. As a general conclusion, the effect of SSI in
this case can be considered as beneficial.
• In the case of bridges rested on stratum over bedrock, the dynamic

response of the structural system is essentially governed by the
ratio between the resonant frequencies of the beam and the
subgrade deposit. If the fundamental frequency of the beam is
sufficiently larger than the resonant frequency of the deposit,
60 | DISCUSSION AND CONCLUSION
both frequency shortening, and foundation damping of the system

approach the “half-space” values. On the contrary, if the
fundamental frequency of the beam is lower than the resonance
frequency of the deposit, the radiation damping capacity of the
soil deposit is almost zero. Consequently, the resonance may
occur at lower speeds in comparison to the case of a SS beam,
which makes the effect of SSI detrimental. In the intermediate
region, the estimated frequency shortening and foundation
damping can be higher than in the case of a half-space medium
due to the fact that the dynamic stiffness of the foundation-soil
system significantly decreases in the vicinity of its resonance
frequency while at the same time the radiation capacity of the
deposit is resumed.
6.2. Portal frame bridges
The dynamic soil-structure interaction of portal frame bridges was

investigated by using coupled FE-BE methods as well as full FE
models. The results of the numerical analyses on short-span portal
frame bridges revealed that SSI has a significant influence on the
dynamic response of these structures. It was found that the
surrounding soil, especially backfill soil, leads to an increase in the
natural frequency and even more important a significant increase in
the damping ratio of the vertical bending modes which reduces the
response at resonance. Hence, neglecting the effect of SSI in the
train-induced vibration analysis of this type of bridges might be
considerably conservative.
A simplified model for considering the effect of the surrounding soil

was proposed which may be useful for practical design purposes and
parametric studies. The advantage of the proposed simplified method,
compared to simplified 2D models, is that it can be used in the train-
induced vibration analysis of a wider class of 3D FE models where
plate modes are also important. While the agreement between
predicted responses by the simplified and full FE model or coupled
FE-BE models were promising, the simplified model slightly
overestimated the dissipation capacity of the backfill soil, in
comparison to the results of the reference models.
DISCUSSION AND CONCLUSION | 61
6.3. Experimental validation of numerical models
This work also includes the in-situ dynamic testing on two portal
frame railway bridges and the experimental validation of their
numerical models. A series of controlled vibration tests were
performed to identify the modal parameters of the tested bridge-soil
systems. In both case studies, the measured modal damping ratios of
the bending and torsional modes were considerably higher than the
plate modes, due to the stronger effect of soil-structure interaction.
While the measured modal damping ratios for torsional and plate
modes of the two investigated bridges were rather close, the
difference between the modal damping ratios for the 1st bending
modes was substantial. The reason for this is probably related to the
depth of the subsoil stratum which was the main dissimilarity
between the two bridges.
An automated FE model updating procedure was successfully

implemented to calibrate non-proportional damped FE models. The
calibrated FE models produced frequency response functions which
were in good agreement with their measured counterparts, despite
some discrepancies at high frequencies (i.e. above 60 Hz). Note that,
the reliability of the calibrated FE models was also verified against
monitored response from passing trains.
Using the calibrated FE models, it was shown that neglecting the

dynamic stiffness of the surrounding soil resulted in substantial
deviation from experimental results. The identified modal damping
ratio for the fundamental bending mode of the tested bridges was
about 5-6 times higher than the recommended design values [5]. The
experimentally identified modal damping ratios of the tested bridges
have brought attention to the fact that the recommended damping
ratios in EN1991-2 [5] are noticeably conservative for short-span
portal frame bridges and need to be revised.
6.4. Further research
Based on the results in this report, the following research work is

suggested for the future:
62 | DISCUSSION AND CONCLUSION
6.4.1. Simple beam bridges
• The proposed closed-form expressions for estimating the

fundamental modal properties of the beam bridges should be
improved by considering some other important parameters
such as foundations’ mass and structural damping of the beam.
The reliability of this analytical solution should be verified
based on in-situ measurements.
• The effect of piled foundations on the dynamic properties of

railway bridges should be studied more in detail. The
uncertainties in the geotechnical properties of the underlying
soil, e.g. heterogeneity and level of ground water, should also
be studied.
6.4.2. Portal frame bridges
• The reliability and accuracy of the proposed simplified model

for abutment-backfill soil interaction needs to be verified and
improved based on more numerical analyses as well as in-situ
measurements.
• In this research, the interaction between backfill soil and

abutments was assumed as a fully tied interaction. This
simplification may overestimate the dissipation capacity of the
backfill soil for vibration modes which are dominantly
governed by the motion of the abutment. The influence of the
interaction between abutments and backfill soil, modeled as
nonlinear compression-only behavior, must be investigated
further.
• The implemented coupled FE-BE model should be enhanced

for proper modeling of the portal frame bridges considering a
detailed model of the structure/soil/track.
• The efficiency of the enhanced transmitting boundaries such as

PMLs in the dynamic SSI analyses using a direct approach
should be examined.
DISCUSSION AND CONCLUSION | 63
• The effect of the span length and abutment height (relative

deck/abutment stiffness) on the dynamic abutment-backfill
interaction needs to be investigated numerically and
experimentally. Theoretically, the influence of SSI should be
proportional to the relative deck/abutments stiffness.
References
[1] ERRI D-214, 1999b. Rail bridges for speeds over 200Km/h, Final Report. European Rail Research
Institute (ERRI).
[2] L. Fryba, 1999. Vibration of Solids and Structures under Moving Loads, London, Thomas Telford
[3] H. Xia, N. Zhang, W. Guo, 2006, Analysis of resonance mechanism and conditions of train–bridge
system. Journal of Sound and Vibration 297 (3), 810–822.
[4] M. Zacher, M. Baeßler, 2009, Dynamic behaviour of ballast on railway bridges, In: Dynamics of
High-Speed Railway Bridge, Taylor & Francis Group, London, pp. 99-112.
[5] EN1991-2, Eurocode 1, Actions on Structures. Part2: Traffic loads on Bridges, 2003.
[6] A. Sextos, ICT applications for new generation seismic design, construction and assessment of
bridges, Struct. Eng. Int., 24(2), 173–183 (2014).
[7] A. Romero, M. Solís, J. Domínguez, P. Galvín, Soil-structure interaction in resonant railway bridges,
Soil Dynamics and Earthquake Engineering 47 (2013) 108-116.
[8] M. Ülker-Kaustell, R. Karoumi, C. Pacoste, Simplifed analysis of the dynamic soil-structure
interaction of a portal frame railway bridge, Engineering structures 32 (11) (2010) 3692-3698.
[9] A. Doménech, M. D. Martínez-Rodrigo, A. Romero, P. Galvín, On the basic phenomena of soil-
structure interaction on the free vibration response of beams : Application to railway bridges,
Engineering structures 125 (2016) 254-265.
[10] C. Svedholm, C. Pacoste-Calmanovici, R. Karoumi, Modal properties of simply supported railway
bridges due to soil-structure interaction, in: Conference of 5th ECCOMAS Thematic Conference on
Computational Methods in Structural Dynamics and Earthquake Engineering, COMPDYN 2015, pp.
1709-1719.
[11] A.Fraile, L.Hermanns, F.Chirino, E.Alarcon, Dynamic effects of high speed trains on underpasses,
Proceedings ICCST’02 Sixth Conference on Computational Structures Technology, Civil-Com Press,
Edinburgh,UK, 2002
[12] J. Vega, A. Fraile, E. Alarcon, L. Hermanns, Dynamic response of underpasses for high-speed train
lines, Journal of sound and vibration 331 (2012) 5125-5140.
[13] P. Galvin, J. Dominguez, High-speed train-induced ground motion and interaction with structures,
Journal of Sound and Vibration 307 (2007) 755–777.
[14] J.D. Achenbach. Wave propagation in elastic solids, volume 16 of North-Holland Series in Applied
Mathematics and Mechanics. North-Holland, Amsterdam, The Netherlands, 1973
[15] J. Houbrechts, M. Schevenels, G. Lombaert, G. Degrande, W. Rucker, V. Cuellar, A. Smekal, 2011,
Test procedures for the determination of the dynamic soil characteristics, RIVAS (Railway Induced
Vibration Abatement Solutions Collaborative) project.
[16] S. L. Kramer, 1996, Geotechnical Earthquake Engineering, Prentice-Hall
[17] H. Lamb, 1904, On the propagation of tremors over the surface of an elastic solid, Philosophical
Transaction of the Royal Society, London, Serie A, Vol. 203, pp. 1-42.
[18] L. Hall, Simulation and analyses of train-induced ground vibration: a comparative study of two and
three dimensional calculations with actual measurement, PhD dissertation, Stockholm (Sweden):
KTH Royal Institute of Technology, 2000.
[19] A.C. Eringen and E.S. Suhubi. Elastodynamics, Volume 2, Linear theory. Academic Press, New
York, USA, 1975.
[20] B. A. Bolt, 1976, Nuclear Explosions and Earthquakes: The Parted Veil, W. H. Freeman and Co.
66 | REFERENCES
[21] DR. Bland,1960, The theory of linear viscoelasticity, New York, NY: Pergamon Press
[22] F.J. Rizzo and D.J. Shippy. An application of the correspondence principle of linear viscoelasticity
theory. SIAM Journal on Applied Mathematics, 21(2):321–330, 1971.
[23] M.A. Biot. Theory of propagation of elastic waves in a fluid-saturated porous solid. I. low-
frequency range. Journal of the Acoustical Society of America, 28(2):168–178, 1956.
[24] H.B. Seed, I.M. Idriss, 1970, Soil moduli and damping factors for dynamic response analyses,
Report EERC 70-10, Earthquake Engineering Research Center, University of California, Berkley,
CA, US.
[25] Swedish Transport Administraion, Trafikverkets tekniska råd för geokonstruktioner-TR Geo 13 (in
Swedish), 2016.
[26] R. Larsson and M. Mulabdic. Shear moduli in Scandinavian clays; measurement of initial shear
modulus with seismic cones. Technical Report 40, Swedish Geotechnical Institute, 1991.
[27] R.V. Whitman, F.E. Richart, 1967, Design procedures for dynamically loaded foundations, Journal
of Soil Mechanical and Foundations Division, ASCE, Vol. 93, No. SM6, pp. 169-192
[28] W. P. Stewart, R. G. Campanella, 1991, In situ measurements of damping in soils, Proceedings, 2nd
International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil
Dynamics, St. Louis, MO, pp. 83-92
[29] G. Gazetas. 1998, Foundation vibrations, in: H. Y. Fang (Eds.), Foundation Engineering Handbook,
Springer Science+Business, Newyork 1998, pp. 553-593
[30] J. P., Wolf, 1989, Soil-structure interaction analysis in time domain, Nucl. Eng. Des., 111(3), 381–
393.
[31] J. P., Wolf, C. Song, 1996, Finite Element Modelling of Unbounded Media, Wiley
[32] E. Kausel, 1988, Local transmitting boundaries, Journal of Engineering Mechanics , ASCE, 1011-
1027
[33] D. Givoli, 1992, Numerical Methods for Problems in Infinite Domains, Elsevier, Amesterdam.
[34] J. Lysmer, R.L. Kuhlemeyer, Finite dynamic model for infinite media, Journal of Engineering
Mechanic Division, ASCE 95 (1969) 859–877
[35] U. Basu, A.K. Chopra. Perfectly matched layers for time-harmonic elastodynamics of unbounded
domains: theory and finite-element implementation. Comput. Methods Appl. Mech. Engrg. 192
(2003) 1337-1375.
[36] S. Kucukcoban, L.F. Kallivokas. Mixed perfectly-matched-layers for direct transient analysis in 2D
elastic heterogeneous media. Comput. Methods Appl. Mech. Engrg., 200 (2011) 57-76.
[37] E. Kausel, J.M. Barbosa. PMLs: A direct approach. Int J Numer Meth Engng 90 (2012) 343-352.
[38] J. Josifovski. Analysis of wave propagation and soil-structure interaction using a perfectly matched
layer model. Soil Dynamics and Earthquake Engineering 81 (2016) 1-13.
[39] COMSOL Multiphysics Reference Manual, version 5.3, COMSOL, Inc, www.comsol.com
[40] E. Kausel, 2015, Short course on dynamic soil-structure interaction, UME graduate school, Pavia,
Italy
[41] J. Zhang, Y. Tang, 2006, Evaluating radiation damping of shallow foundations on nonlinear soil
medium for soil-structure interaction analysis of bridges, US-Japan Bridge Engineering Workshop,
Seattle, WA.
[42] J. Lysmer, 1978, Analytical procedures in soil dynamics, Proceedings, ASCE Geotechnical
Engineering Division Specialty Conference, Pasadena, CA, Vol. 3, pp. 1267-1317
[43] M. Bonnet, 1995, Boundary Integral Equation Methods for Solids and Fluids. John Wiley and Sons:
Chichester, United Kingdom.
[44] E. Kausel. Fundamental solutions in elastodynamics: a compendium. Cambridge Univer-sity Press,
New York, 2006.
[45] E. Kausel and J.M. Roësset. Stiffness matrices for layered soils. Bulletin of the Seismo-logical
Society of America, 71(6):1743–1761, 1981.
[46] R.F. Scott, 1973, Earthquake-induced pressures on retaining walls, Proc. 5th World Conf on
Earthquake Eng., Rome, Italy, II, 1611-1620.
[47] S. K. Jain, R. F. Scott, Seismic analysis of cantilever retaining walls, in: Proc. of Structural
Mechanics in Reactor Technology, Anaheim, Calif., (1989) pp. 241-246
REFERENCES | 67
[48] A.S. Veletsos, A.H. Younan, 1995, Dynamic soil pressures on vertical walls, in: Proceedings: 3rd
Int. Conf. on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, (1995)
Volume Ill, St. Louis, Missouri
[49] A.S. Veletsos, A.H. Younan, 1997, Dynamic response of cantilever retaining walls, J.Geotech
Engrg., ASCE, 123 (2) 1090-0241
[50] A.H. Younan, A.S. Veletsos, 2000, Dynamic response of flexible retaining walls, J.Earthquake
engineering and structural dynamics vol.29, 1815-1844
[51] J. P. Stewart et al., 2013, Soil-structure interaction for building structures, report NISTGCR-12-917-
21, ATC-83, Rep. Prepared for the Applied Technology Council, Redwood City, CA.
[52] Givens, M.J., 2013, Dynamic Soil-Structure Interaction of Instrumented Buildings and Test
Structures, Ph.D. Dissertation, Department of Civil and Environmental Engineering, University of
California, Los Angeles, California.
[53] J. G. Sieffert, F. Cevaer, 1992, Handbook of Impedance Functions, Surface Foundations, s.l.: Ouest
Éditions Presses Académiques.
[54] A. Pais, E. Kausel, 1988, Approximate formulas for dynamic stiffnesses of rigid foundations, Soil
Dynamics and Earthquake Engineering, Vol. 7, No. 4, pp. 213-227.
[55] G. Mylonakis, S. Nikolaou, G. Gazetas, 2006, Footings under seismic loading: Analysis and design
issues with emphasis on bridge foundations, Soil Dynamics and Earthquake Engineering, Vol. 26, pp.
824-853.
[56] M. Schevenels, S. François, G. Degrande, EDT: An ElastoDynamics Toolbox for MATLAB.
Computers & Geosciences 2009; 35(8):1752–1754. Available from:
http://dx.doi.org/10.1016/j.cageo.2008.10.012.
[57] R. Dobry, G. Gazetas,1986, Dynamic response of arbitrarily shaped foundations, J. Geotech.
Engrg., ASCE, 112(2), 109-135.
[58] J. Lysmer, F. E. Jr. Richart, 1966, Dynamic Response of Footing to Vertical Loading, Journal of the
Soil Mechanics and Foundations Division, ASCE, Vol. 92, No. SMI, pp. 65-91.
[59] J. P., Wolf, 1994, Foundation Vibration Analysis Using Simple Physical Models, Prentice Hall
[60] C. Svedholm, 2017, Efficient Modelling Techniques for Vibration Analyses of Railway Bridges, In
TRITA-BKN. Doctoral thesis 145, Stockholm (Sweden): KTH Royal Institute of Technology.
Available online: http://www.diva-portal.org/.
[61] P. Museros, E. Moliner, M. D. Martínez-Rodrigo, 2013. Free vibrations of simply-supported beam
bridges under moving loads: Maximum resonance, cancellation and resonant vertical acceleration.
Journal of Sound and Vibration 332 (2), 326–345.
[62] J.D. Yau, Y.S. Wu, Y.B. Yang, 2001, Impact response of bridges with elastic bearings to moving
loads, Journal of Sound and Vibration 248 (1) (2001) 9–30.
[63] G. Lallement, D.J. Inman, A tutorial on complex eigenvalues, Proceedings of 13th IMAC,
Nashville, TN (1995), pp. 490-495.
[64] G. Cardano, 1545, Ars magna The Rules of Algebra, Dover, 1993, ISBN 0-486-67811-3
[65] E. L. Rees, 1922, Graphical Discussion of the Roots of a Quartic Equation, The American
Mathematical Monthly. 29 (2): 51–55. doi:10.2307/2972804. JSTOR 2972804.
[66] D. Lazard, 1988, Quantifier elimination: Optimal solution for two classical examples, Journal of
Symbolic Computation. 5: 261–266. doi:10.1016/S0747-7171(88)80015-4.
[67] C. Svedholm, A. Zangeneh, C. Pacoste, S. François, R. Karoumi, Vibration of damped uniform
beams with general end conditions under moving loads, Engineering Structures 126 (2016) 40-52
[68] C. Johansson, A. Andersson, J. Wiberg, M. Ülker-Kaustell, C. Pacoste, R. Karoumi, 2010,
Höghastighetsprojekt-bro: Delrapport i: Befintliga krav och erfarenheter samt parameterstudier
avseende dimensionering av järnvägsbroar för farter över 200 km/h. Tech. Rep. 139, KTH
Brobyggnad (in Swedish)
[69] D. Dooms, M. Jansen, G. De Roeck, G. Degrande, G. Lombaert, M. Schevenels, S. François, A
finite element toolbox for Matlab (StaBIL). Technical report BWM-2010–**. Department of Civil
Engineering, KU Leuven; June 2010
[70] S. François, M. Schevenels, G. Degrande, Matlab toolbox for boundary elements in elastodynamics
(BEMFUN), report BWM-2009-**, Department of Civil Engineering, KU Leuven; May 2009
68 | REFERENCES
[71] S. François, 2008, Nonlinear modelling of the response of structures due to ground vibrations, PhD
dissertation, Department of Civil Engineering, KU Leuven
[72] P. Coulier, S. François, G. Lombaert, G. Degrande, 2014, Coupled finite element–hierarchical
boundary element methods for dynamic soil–structure interaction in the frequency domain, Int. J.
Numer. Meth. Engng 2014; 97:505–530
[73] F. Tisseur, K. Meerbergen, 2001, The quadratic eigenvalue problem, SIAM Rev. Vol. 43, Number 2,
2001, pp. 235–286.
[74] MATLAB 2013a, Natick, Massachusetts, The MathWorks Inc., 2013
[75] A. Sextos P. Faranonis, V. Zabel, F. Wuttke, T. Arndt, P. Panetos, Soil–Bridge System Stiffness
Identification through Field and Laboratory Measurements, J. Bridge Eng., 2016, 21 (10)
[76] A. Morrasi, S. Tonon, Dynamic testing for Structural Identification of a Bridge, J. Bridge Eng.,
2008, 13 (6), p. 573-585
[77] A. Andersson, M. Ülker-Kaustell, R. Borg, O. Dymén, A. Carolin, R. Karoumi, Pilot testing of a
hydraulic bridge exciter, in: MATEC Web of Conferences, Vol. 24, EDP Sciences, 2015, p. 02001.
[78] D. J. Ewins, 2000, Modal testing: Theory, practice and application, 2nd Ed., Research Studies Press
Ltd., Baldock, Hertfordshire, U.K.
[79] Mallat S, Wavelet A. Tour of signal processing - the sparse way. 3rd ed. Academic Press, 2008.
[80] MATLAB Wavelet Toolbox, The MathWorks, 2010.
[81] Dassault Systmes, ABAQUS/Standard User’s Manual, 6.14 ed., Providence, RI, USA, 2014.
Paper I
Dynamic stiffness identification of

portal frame bridge-soil system using
controlled dynamic testing
Procedia Engineering 199 (2017) 1062-1067, EURODYN 2017,
Rome, September 10-13,
Paper II
Identification of Soil-Structure
Interaction Effect in a Portal Frame
Railway Bridge through Full-Scale
Dynamic Testing
Engineering Structures, volume 159, pages 299-309, 2018
List of Bulletins from the Department of Structural
Engineering, The Royal Institute of Technology, Stockholm
TRITA-BKN. Bulletin.
Pacoste, C., On the Application of Catastrophe Theory to Stability Analyses of

Elastic Structures. Doctoral Thesis, 1993. Bulletin 1.
Stenmark, A-K., Dämpning av 13 m lång stålbalk − "Ullevibalken". Utprovning av

dämpmassor och fastsättning av motbalk samt experimentell bestämning av
modformer och förlustfaktorer. Vibration tests of full-scale steel girder to determine
optimum passive control.
Licentiatavhandling, 1993. Bulletin 2.
Silfwerbrand, J., Renovering av asfaltgolv med cementbundna plastmodifierade

avjämningsmassor.
1993. Bulletin 3.
Norlin, B., Two-Layered Composite Beams with Nonlinear Connectors and

Geometry − Tests and Theory.
Doctoral Thesis, 1993. Bulletin 4.
Habtezion, T., On the Behaviour of Equilibrium Near Critical States.

Licentiate Thesis, 1993. Bulletin 5.
Krus, J., Hållfasthet hos frostnedbruten betong.

Wiberg, U., Material Characterization and Defect Detection by Quantitative

Ultrasonics.
Lidström, T., Finite Element Modelling Supported by Object Oriented Methods.

Hallgren, M., Flexural and Shear Capacity of Reinforced High Strength Concrete
Beams without Stirrups.
Krus, J., Betongbalkars lastkapacitet efter miljöbelastning.
1994. Bulletin 10.
Sandahl, P., Analysis Sensitivity for Wind-related Fatigue in Lattice Structures.

Sanne, L., Information Transfer Analysis and Modelling of the Structural Steel
Construction Process.
Zhitao, H., Influence of Web Buckling on Fatigue Life of Thin-Walled Columns.

Kjörling, M., Dynamic response of railway track components. Measurements

during train passage and dynamic laboratory loading.
Yang, L., On Analysis Methods for Reinforced Concrete Structures.

Petersson, Ö., Svensk metod för dimensionering av betongvägar.

Lidström, T., Computational Methods for Finite Element Instability Analyses.

Krus, J., Environment- and Function-induced Degradation of Concrete Structures.

Editor, Silfwerbrand, J., Structural Loadings in the 21st Century.

Sven Sahlin Workshop, June 1996. Proceedings. Bulletin 19.
Ansell, A., Frequency Dependent Matrices for Dynamic Analysis of Frame Type
Structures.
Troive, S., Optimering av åtgärder för ökad livslängd hos infrastruktur-

konstruktioner.
Karoumi, R., Dynamic Response of Cable-Stayed Bridges Subjected to Moving

Vehicles.
Hallgren, M., Punching Shear Capacity of Reinforced High Strength Concrete
Slabs.
Hellgren, M., Strength of Bolt-Channel and Screw-Groove Joints in Aluminium

Extrusions.
Yagi, T., Wind-induced Instabilities of Structures.

Eriksson, A., and Sandberg, G., (editors), Engineering Structures and Extreme
Events − proceedings from a symposium, May 1997. Bulletin 26.
Paulsson, J., Effects of Repairs on the Remaining Life of Concrete Bridge Decks.
Olsson, A., Object-oriented finite element algorithms.

Yunhua, L., On Shear Locking in Finite Elements.

Ekman, M., Sprickor i betongkonstruktioner och dess inverkan på beständigheten.

Karawajczyk, E., Finite Element Approach to the Mechanics of Track-Deck

Systems.
Fransson, H., Rotation Capacity of Reinforced High Strength Concrete Beams.

Edlund, S., Arbitrary Thin-Walled Cross Sections. Theory and Computer

Implementation.
Forsell, K., Dynamic analyses of static instability phenomena.

Ikäheimonen, J., Construction Loads on Shores and Stability of Horizontal

Formworks.
Racutanu, G., Konstbyggnaders reella livslängd.

Appelqvist, I., Sammanbyggnad. Datastrukturer och utveckling av ett IT-stöd för
byggprocessen.
Alavizadeh-Farhang, A., Plain and Steel Fibre Reinforced Concrete Beams

Subjected to Combined Mechanical and Thermal Loading.
Eriksson, A. and Pacoste, C., (editors), Proceedings of the NSCM-11: Nordic

Seminar on Computational Mechanics, October 1998. Bulletin 39.
Luo, Y., On some Finite Element Formulations in Structural Mechanics.

Troive, S., Structural LCC Design of Concrete Bridges.

Tärno, I., Effects of Contour Ellipticity upon Structural Behaviour of Hyparform

Suspended Roofs.
Hassanzadeh, G., Betongplattor på pelare. Förstärkningsmetoder och

dimensioneringsmetoder för plattor med icke vidhäftande spännarmering.
Karoumi, R., Response of Cable-Stayed and Suspension Bridges to Moving

Vehicles. Analysis methods and practical modeling techniques.
Johnson, R., Progression of the Dynamic Properties of Large Suspension Bridges

during Construction
− A Case Study of the Höga Kusten Bridge.
Tibert, G., Numerical Analyses of Cable Roof Structures.

Ahlenius, E., Explosionslaster och infrastrukturkonstruktioner - Risker, värderingar

och kostnader.
Battini, J-M., Plastic instability of plane frames using a co-rotational approach.

Ay, L., Using Steel Fiber Reinforced High Performance Concrete in the
Industrialization of Bridge Structures.
Paulsson-Tralla, J., Service Life of Repaired Concrete Bridge Decks.

Billberg, P., Some rheology aspects on fine mortar part of concrete.

Ansell, A., Dynamically Loaded Rock Reinforcement.

Forsell, K., Instability analyses of structures under dynamic loads.
Edlund, S., Buckling of T-Section Beam-Columns in Aluminium with or without

Transverse Welds.
Löfsjögård, M., Functional Properties of Concrete Roads − General

Interrelationships and Studies on Pavement Brightness and Sawcutting Times for
Joints.
Nilsson, U., Load bearing capacity of steel fibree reinforced shotcrete linings.
Silfwerbrand, J. and Hassanzadeh, G., (editors), International Workshop on

Punching Shear Capacity of RC Slabs − Proceedings. Dedicated to Professor Sven
Kinnunen. Stockholm June 7-9, 2000. Bulletin 57.
Wiberg, A., Strengthening and repair of structural concrete with advanced,

cementitious composites.
Racutanu, G., The Real Service Life of Swedish Road Bridges - A case study.
Alavizadeh-Farhang, A., Concrete Structures Subjected to Combined Mechanical

and Thermal Loading.
Wäppling, M., Behaviour of Concrete Block Pavements - Field Tests and Surveys.
Getachew, A., Trafiklaster på broar. Analys av insamlade och Monte Carlo
genererade fordonsdata.
James, G., Raising Allowable Axle Loads on Railway Bridges using Simulation and
Field Data.
Karawajczyk, E., Finite Elements Simulations of Integral Bridge Behaviour.

Thöyrä, T., Strength of Slotted Steel Studs.

Tranvik, P., Dynamic Behaviour under Wind Loading of a 90 m Steel Chimney.

Ullman, R., Buckling of Aluminium Girders with Corrugated Webs.

Getachew, A., Traffic Load Effects on Bridges. Statistical Analysis of Collected

and Monte Carlo Simulated Vehicle Data.
Quilligan, M., Bridge Weigh-in-Motion. Development of a 2-D Multi-Vehicle

Algorithm.
James, G., Analysis of Traffic Load Effects on Railway Bridges.

Doctoral Thesis 2003. Bulletin 70.
Nilsson, U., Structural behaviour of fibre reinforced sprayed concrete anchored in

rock.
Wiberg, A., Strengthening of Concrete Beams Using Cementitious Carbon Fibre

Composites.
Löfsjögård, M., Functional Properties of Concrete Roads – Development of an

Optimisation Model and Studies on Road Lighting Design and Joint Performance.
Bayoglu-Flener, E., Soil-Structure Interaction for Integral Bridges and Culverts.

Licentiate Thesis 2004. Bulletin 74.
Lutfi, A., Steel Fibrous Cement Based Composites. Part one: Material and
mechanical properties. Part two: Behaviour in the anchorage zones of prestressed
bridges.
Johansson, U., Fatigue Tests and Analysis of Reinforced Concrete Bridge Deck
Models.
Roth, T., Langzeitverhalten von Spannstählen in Betonkonstruktionen.

Licentitate Thesis 2004. Bulletin 77.
Hedebratt, J., Integrerad projektering och produktion av industrigolv – Metoder för

att förbättra kvaliteten.
Österberg, E., Revealing of age-related deterioration of prestressed reinforced

concrete containments in nuclear power plants – Requirements and NDT methods.
Broms, C.E., Concrete flat slabs and footings New design method for punching and
detailing for ductility.
Wiberg, J., Bridge Monitoring to Allow for Reliable Dynamic FE Modelling - A

Case Study of the New Årsta Railway Bridge.
Mattsson, H-Å., Funktionsentreprenad Brounderhåll – En pilotstudie i Uppsala län.

Masanja, D. P, Foam concrete as a structural material.

Johansson, A., Impregnation of Concrete Structures – Transportation and Fixation

of Moisture in Water Repellent Treated Concrete.
Billberg, P., Form Pressure Generated by Self-Compacting Concrete – Influence of

Thixotropy and Structural Behaviour at Rest.
Enckell, M., Structural Health Monitoring using Modern Sensor Technology –

Long-term Monitoring of the New Årsta Railway Bridge.
Söderqvist, J., Design of Concrete Pavements – Design Criteria for Plain and Lean
Concrete.
Malm, R., Shear cracks in concrete structures subjected to in-plane stresses.

Skoglund, P., Chloride Transport and Reinforcement Corrosion in the Vicinity of

the Transition Zone between Substrate and Repair Concrete.
Liljencrantz, A., Monitoring railway traffic loads using Bridge Weight-in-Motion.

Stenbeck, T., Promoting Innovation in Transportation Infrastructure Maintenance –

Incentives, Contracting and Performance-Based Specifications.
Magnusson, J., Structural Concrete Elements Subjected to Air Blast Loading.

Pettersson, L., G., Full Scale Tests and Structural Evaluation of Soil Steel Flexible
Culverts with low Height of Cover.
Doctoral Thesis 2007. Bulletin 93
Westerberg, B., Time-dependent effects in the analysis and design of slender

concrete compression members.
Mattsson, H-Å, Integrated Bridge Maintenance. Evaluation of a pilot project and

future perspectives.
Andersson, A., Utmattningsanalys av järnvägsbroar. En fallstudie av stålbroarna

mellan Stockholm Central och Söder Mälarstrand, baserat på teoretiska analyser
och töjningsmätningar.
Licentiate Thesis 2009. Bulletin 96
Malm, R., Predicting shear type crack initiation and growth in concrete with non-
linear finite element method.
Bayoglu Flener, E., Static and dynamic behavior of soil-steel composite bridges
obtained by field testing.
Gram, A., Numerical Modelling of Self-Compacting Concrete Flow
Wiberg, J., Railway bridge response to passing trains. Measurements and FE model
updating.
Athuman M.K. Ngoma, Characterisation and Consolidation of Historical Lime

Mortars in Cultural Heritage Buildings and Associated Structures in East Africa.
Ülker-Kaustell, M., Some aspects of the dynamic soil-structure interaction of a

portal frame railway bridge.
Vogt, C., Ultrafine particles in concrete

Selander, A., Hydrophobic Impregnation of Concrete Structures

Ilina, E., Understanding the application of knowledge management to safety critical

facilities
Leander, J., Improving a bridge fatigue life prediction by monitoring

Andersson, A., Capacity assessment of arch bridges with backfill – Case of the old
Årsta railway bridge
Enckell, M., Lessons Learned in Structural Health Monitoring of Bridges Using

Advanced Sensor Technology
Hansson, H., Warhead penetration in concrete protective structures

Gonzalez Silva, I., Study and Application of Modern Bridge Monitoring Technique
Safi, M., LCC Applications for Bridges and Integration with BMS
Guangli, D., Towards sustainable construction, life cycle assessment of railway
bridges
Hedebratt, J., Industrial Fibre Concrete Slabs – Experiences and Tests on Pile-
Supported Slab
Ahmed, L., Models for analysis of shotcrete on rock exposed to blasting

Le, T., N., Corotational formulation for nonlinear dynamic analysis of flexible
beam structures
Johansson, C., Simplified dynamic analysis of railway bridges under high-speed

trains
Jansson, R., Fire Spalling of Concrete – Theoretical and Experimental Studies

Leander, J., Refining the fatigue assessment procedure of existing steel bridges
Le, T., N., Nonlinear dynamics of flexible structures using corotational beam
elements
Ülker-Kaustell, M., Essential modelling details in dynamic FE-analyses of railway

bridges
Safi, M., Life-Cycle Costing. Applications and Implementations in Bridge

Investment and Management
Arvidsson, T., Train-Bridge Interaction: Literature Review and Parameter

Screening
Rydell, C., Seismic high-frequency content loads on structures and components

within nuclear facilities
Bryne, L. E., Time dependent material properties of shotcrete for hard rock
tunnelling
Wennström, J., Life Cycle Costing in Road Planning and Management: A Case
Study on Collisionfree Roads
Gonzalez Silva, I., Application of monitoring to dynamic characterization and

damage detection in bridges
Sangiorgio, F., Safety Format for Non-linear Analysis of RC Structures Subjected

to Multiple Failure Modes
Gram, A., Modelling of Cementitious Suspensial Flow – Influence of Viscosity and

Aggregate Properties
Du, G., Life Cycle Assessment of bridges, model development and Case Studies
Zhu, J., Towards a Viscoelastic Model for Phase Separation in Polymer Modified
Bitumen
Chen, F., The Future of Smart Road Infrastructure. A Case Study for the eRoad.
Ahmed, L., Models for analysis of young cast and sprayed concrete subjected to
impact-type loads
Albrektsson, J., Durability of Fire Exposed Concrete – Experimental Studies

Focusing on Stiffness & Transport Properties
Wallin, J., Systematic planning and execution of finite element model updating
Wadi, A., Flexible Culverts in Sloping Terrain

McCarthy, R., Self-Compacting Concrete for Improved Construction Technology

Veganzones Munõs,J.J., Bridge Edge Beams: LCCA and Structural Analysis for the
Evaluation of New Concepts
Abbasiverki, R., Analysis of underground concrete pipelines subjected to seismic

high-frequency loads
Gasch, T., Concrete as a multi-physical material with applications to hydro power

facilities
Ali, M., M., Use of Macro Basalt Fibre Concrete for Marine Applications
Döse, M., Ionizing Radiation in Concrete and Concrete Buildings – Empirical

Assessment
Khan, A., Fundamental investigation to improve the quality of cold mix asphalt
Zhu, J., Storage Stability and Phase Separation Behaviour of Polymer-Modified

Bitumen
Chen, F., Sustainable Implementation of Electrified Roads

Svedholm, C., Efficient modelling techniques for vibration analyses of railway

bridges
Solat Yavari, M., Slab Frame Bridges, Structural Optimization Considering

Investment Cost and Environmental Impacts
Tell, S., Vibration mitigation of high-speed railway bridges, Application of fluid

viscous dampers
Elgazzar, H., End-Shield Bridges for Highspeed Railway – Full scale dynamic
testing and numerical simulations
Sousa Neves, A. C., Structural Health Monitoring of Bridges: Model-free damage
detection method using Machine Learning
Sjölander, A., Analyses of shotcrete stress states due to varying lining thickness and
irregular rock surfaces
Heng, P., Simplified Mechanical Models for the Nonlinear Dynamic Analysis of
Elasto-plastic Steel Structures Impacted by a Rigid Body
Mohammadi Mohaghegh, A., Structural Properties of High-Performance Macro

Basalt Fibre Concrete;
Flexure, Shear, Punching Shear and Fire Spalling
The department also publishes other series. For full information see our homepage
http://www.byv.kth.se
ISBN
TRIT LT
www.kth .se

Zangeneh 2018 - Licentiate

Uploaded by

Document Information

Original Title

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

Zangeneh 2018 - Licentiate

Uploaded by

Copyright:

Available Formats

Dynamic Soil-Structure

ABBAS ZANGENEH KAMALI

Licentiate Thesis, 2018

© Abbas Zangeneh Kamali, 2018

Railway Bridge; Soil-Structure Interaction; Full FE modeling; Simplified modeling;

Arbetet i denna avhandling omfattar en översikt över den dynamiska

Järnvägsbro; Jord-bro interaktion; FE modellering; Förenklad modellering;

The research work has been conducted under the supervision of

I would like to thank ELU consulting company for giving me the

Last but not least, my deepest gratitude to my family, especially to

Stockholm, March 2018

Abbas Zangeneh Kamali

I A. Zangeneh, C. Svedholm, A. Andersson, C. Pacoste, R.

II A. Zangeneh, C. Svedholm, A. Andersson, C. Pacoste, R.

Both papers were planned, implemented and written by the first

Other Relevant Publications:

• C. Svedholm, A. Zangeneh, C. Pacoste, S. François, R. Karoumi,

• T. Arvidsson, A. Zangeneh, D. Cantero, A. Andersson, Influence

2. Principles of dynamic soil-structure interaction ....................... 7

3. Discrete model for beams on viscoelastic supports .............. 25

4. Dynamic SSI analysis of portal frame bridges ........................ 41

5. Experimental validation of numerical models ......................... 49

6. Discussion and conclusion ...................................................... 59

The railway bridge superstructures traveled by high-speed trains may

Resonance may result in ballast instability, track degradation,

The frequency content and amplitude of the vibrations in railway

considering the effect of the flexibility and dissipation capacity of the

Despite extensive research studies and guidelines concerning the

According to Eurocode 1991-2 [5], for the analysis of train-induced

1.2. Previous work

1.2.1. Simply supported beam bridges

Recent numerical studies on the SSI analysis of railway beam bridges

embedded abutments in a homogenous half-space. The authors

1.2.2. Portal frame bridges

In the field of train-induced vibration analysis of railway bridges,

interaction. Galvin et al. (2007) [13] developed a fully 3D coupled

1.3. Aims and scope

The overall objective of the present research is to study the influence

• Determine the factors that influence the modal properties of

• Develop both complex and simple numerical models for the

• Provide practical tools and guidance regarding the efficient

• Implement structural identification techniques for inverse

The following limitations apply:

• The dynamic Soil-Structure Interaction (SSI) is a broad

• The present study is limited to linear time-invariant systems.

• In the numerical and analytical studies on beam bridges, only

• In the train-induced vibration analyses, the railway excitation

1.4. Research contribution

The theoretical, numerical and experimental investigations on the

• Closed-form expressions for evaluating natural frequency and

• An analysis of the effect of different subsoil conditions on the

• Recommendations on how to efficiently model the effect of

• A proposal concerning a Winkler-based simplified model for

• Full-scale controlled dynamic tests on two short-span portal

1.5. Outline of the thesis

The reminder of this thesis is structured as follows. Chapter 2

The chapter is organized as follows. First, a brief introduction to the

2.1. Wave propagation in 3D elastic media

In contrast to seismic loading, railway-induced vibrations usually

where e is the volumetric strain, ui is the displacement component

In classical elastodynamics [14], the solution to Navier’s equation is

2.1.1. Wave propagation in half-spaces

As a stress-free edge is defined in the elastic medium to model the