G+D Computing

News.St7
Newsletter for Strand7 and Straus7 users
Strand7 is marketed as Straus7 in continental Europe

Issue 1, May 2003

Welcome to the first edition of our newsletter for 2003!

News.St7 aims to bring you the latest information about Strand7. It will highlight new Strand7 features,
provide modelling tips, which will be useful to both the new user and the experienced user alike,
summarise interesting work performed by Strand7 users, and in future editions will tell you more about
planned Strand7 developments and upcoming releases. News.St7 will also keep you up to date with
significant events on the Strand7 calendar, including exhibitions and courses.
We hope that you find this publication both interesting and useful. If you have any feedback or
suggestions regarding the content, or you would like to have your latest project included in the ‘Current
Projects’ section then please email News.St7@strand7.com. If you would like to receive your copy of
News.St7 directly by email, simply send a blank email to Newsletter-Subscribe@strand7.com.

Release 2.2 (both Sparse and the

Latest News
new Skyline) use a different

M arch 2003 saw the release of

Strand7 2.2, packed with new
features and options. One of two
major new features
is the Sparse solver.
The Sparse solver
100x speed
can solve very large
models in a much increases seen
shorter time than with the new
before, and models Sparse solver
which were too large
to be solved using
the Skyline solver, can now be
solved. The Sparse solver uses the
same basic procedure as the
Skyline solver for solving equations.
Therefore, it does not compromise
solution accuracy to achieve its
impressive speed increases. Faster
solutions are achieved due to a
much more efficient data storage
and manipulation strategy, based
on so-called “sparse matrix
methods”.
approach, based on the removal
from the global matrix, of
dependent degrees of freedom.
This leads to not only fewer
equations when solving
models containing links and
enforced displacements but
also removes the
deterioration of the condition
of the global stiffness matrix
that occurs when these
constraints are enforced with
Lagrange Multipliers.
Release 2.2 also introduces new Mirrored Angle Section
beam sections including trapezoid,
triangle and cruciform. This avoids
having to prepare a BXS for such In this issue…
sections.
Coupled with this is the new mirror • Sparse solver
option for beam sections, which
allows simple creation of composite • Strand7 API
beam sections, by mirroring a basic
shape about one or more mirror • Modelling bolted
planes. connections

Changes have also been introduced
to the constraint algorithms, which
are used for implementing enforced
nodal displacements and the Links.
Previously, Strand7 used the
Lagrange Multiplier approach for
such constraints. The solvers in
For those users who requested it we • Current projects
have now included Tonne Force
and kips in our force units and • Heat transfer benchmark
plates have the option to contour
Internal Angle Ratio to • Training calendar
accommodate mixed Tri and Quad
meshes. • Exhibitions calendar

www.strand7.com
G+D Computing

In the post-processing options, Did you know? Example API application

beam fibre stresses can be
enveloped and combination
envelope cases are now an option.
Also available using the peek tool is
reaction sum and node flux sum:
these reduce the need for hand
calculations and prove very useful
when calculating total reactions
(including moments) at structural
boundaries. It also greatly simplifies
global stability calculations.
The spectral response solver has
been reconfigured to allow for
multiple spectral solution cases,
with different spectral tables.
You can now also choose the
method of calculation for strain in
geometry nonlinear problems.
Options for nominal, engineering
or Green’s strain are provided.
This article presents just a sample of
what is on offer in Strand7 2.2;
more detailed information can be
found at www.strand7.com.
Strand7 API
Dynamic scaling
With results open press F4 to enter
the dynamic view environment.
You can dynamically
increase or decrease
the Displacement
Scale of deformed
models by clicking
Scale in the Dynamic
Rotation toolbar
(shown) and holding
the left mouse button
while moving the
mouse: up to increase
and down to decrease
the displacement
scale. When an
appropriate scaling
factor is obtained, click
OK. This will
automatically set the
displacement scale to
the desired value.
Shortcut
The Scale option can be quickly
accessed (while performing any of
the other dynamic functions - ie
Auto, rXY, rX, rY, rZ, Zoom and
Pan), simply by pressing and
holding down both the Ctrl key and
the left mouse button and following
the procedures explained above.
An example use of the Strand7 API
is a structural optimisation program.
Structural optimisation methods
that use FEA have seen increased
popularity in recent years due to
increases in computing power. An
example of such an optimisation
method is a Fully Stressed Design
(FSD) method. This method sizes
parts of a structure based on its
relative stress state. If part of a
structure is overstressed, it is added
proportionally more material to
compensate. Conversely, if part of
a structure is under stressed,
material is removed to increase the
structure’s efficiency.
In the case of a plate FE model, the
adding and removal of material can
be simulated by the varying of the
thickness of each element. This
implies that each ‘region’
considered for material
redistribution is represented by
each element. Hence, the plate
thickness of each element becomes
a design variable.
The ‘Plate Optimisation’ program,
which is distributed with the

R elease 2.2 introduces a new and Although these functions can be

very powerful Strand7 tool – the
Strand7 Application Programming
Interface (API). The Strand7 API
allows you to interact with Strand7
via your own computer programs.
This provides the capability to
completely bypass the Strand7
Interactive Environment or to
perform specialised functions that
are currently not available within
Strand7 itself.
In simple terms, the Strand7 API
allows you to incorporate access to
Strand7 within your own programs.
More specifically it can be used to:
1. Directly create new Strand7
model files (binary ST7 files)
from scratch.
2. Modify existing Strand7 model
files.
3. Execute the Strand7 solvers.
4. Extract and process the results
of a solution.
Strand7 API toolkit as example
code, is a simple program that uses
such a method.
performed by using the Strand7
GUI directly, and data can be The following figures demonstrate
transferred between Strand7 and how the optimisation program was
other applications via text-based used to optimise the design of a
import/export cantilever. The cantilever
functions, copy-paste, is modelled using plane
etc, such an approach stress elements and is
can require a API allows subject to an in-plane
significant amount of direct interface vertical tip load. Note that
additional data the cantilever is
processing or
between Strand7
‘discretised’ into many
repetitive manual and your elements – more than
procedures when the software would be necessary for
same operation is to appropriate stress recovery.
be performed many In the initial design, all
times. plates have the same
thickness.
The API approach provides a
totally transparent way of
interacting with Strand7 data so that
repetitive tasks can be easily
automated. Importantly, if the
format of the Strand7 data files
changes, such changes are fully
absorbed by the API so that no
changes to the external programs
are required to use a more up-to-
Initial Cantilever Design
date version of Strand7 or the API.

www.strand7.com
G+D Computing

Over the course of several
iterations, the design is modified by
the API program until the optimised
shape is obtained. This is shown in
the figure below.
Optimised Design
The design is visualised in Strand7 experienced, to form what almost
by contouring the plate elements appears as a truss structure.
according to their thickness (note
This example is based on a model
that elements with a very small
that contains over a 1000 elements
thickness are not drawn – this is a
and the solver was run over a
new feature in the contouring
dozen times to achieve the final
options of Strand7 2.2 – it is
‘shape’. This involved countless
possible to avoid drawing any
modifications to individual plate
element that falls outside specified
thicknesses. Hence, an unthinkably
limits).
slow manual procedure was
It can be seen that the design is automated through the use of the
similar to an I Beam, where a thin Strand7 API.
shear web is sandwiched between
Find out more about the Strand7
two thick sections that take the
API by visiting the Strand7
moment. These sections gradually
website.
increase in thickness towards the
root. Note that the web includes
lightening holes where little shear is

Current Projects

I n this section we present current projects where

Strand7 is being used to perform FE analyses. Projects
from both Strand7 users and our own Strand7 Consulting
will be presented.
If you would like to have your project included in the
next edition of News.St7, please email
News.St7@strand7.com

Use of Strand7 allowed the structure to be fully optimised

High speed Motor Yacht with regard to stresses and deflections whilst meeting all
Elan Design, based in Western Australia, is currently requirements of the governing rules.
using Strand7 to design a high-speed motor yacht which
must meet the requirements of Lloyd’s Register rules.
Measuring 240ft in length and rated at 31kts, it is to be
one of the largest and fastest of its kind.
The model was created in Pro-Engineer, exported as an
IGES file then imported, cleaned, automeshed and
analysed in Strand7.
The resulting plate/shell model contained 72 000
elements and 360 000 degrees of freedom.
The final solution was arrived at after over 70 complete
global solutions and more that 20 sub-models. During
these iterations, it was discovered that applying the
design rules straight out of the book would have
produced a structure which would most likely fail when
it first encountered a respectable storm. This pointedly
proved the value of the FEA.

www.strand7.com
G+D Computing
Nick Stark of Elan Design said:
“The whole process highlighted that while sound
engineering practice is vital, it can at best provide
guidelines by which to design. There is no substitute for
the detailed knowledge of a specific design provided by
FEA.”
If you would like more information then please contact
Elan Design: Tel +61 (0)8 9443 9170 Fax +61 (0)8 9443 9171
www.elandesign.com.au
Paper Lift Bucket Mechanism
WBM Gippsland has been using Strand7 to investigate
structural deformations that were occurring in their
clients’ paper roll lifting mechanism.
The bucket mechanism is used to lift 3 tonne rolls of
paper onto a conveyor system for further handling and
processing.
The client identified that permanent deformation of some
of the main structural members had occurred over
repeated operation of the lift. Some local pull-out of
holding down bolts had also occurred.
WBM Engineering was contracted to investigate the
potential cause of this deformation and devise remedial
changes.
www.strand7.com
The paper bucket is hinged about 2 support points and
lifted via 2 hydraulic rams.
WBM contracted the services of our own Strand7
Consulting team to create the initial model. This model
was supplied to WBM to be used as the basis of their on-
going investigation.
Strand7 was used to create a global model with sufficient
detail to investigate some of the local stresses close to the
bucket hinge and ram points.
The model was essentially divided into two zones. The
first zone was the area away from the zone of influence.
This incorporated the conveyor system behind the main
lift bucket. This was modelled using Strand7 beam
elements.
The second area, the zone of interest, was modelled in
detail with shell elements and the 3 tonne paper roll
represented as lumped mass.
A nonlinear static analysis was run. The load increments
were chosen such that the bucket was lifted through its
full operation cycle by applying appropriate strains to the
lifting rams.
The Strand7 model was used by WBM to identify the
most appropriate remedial measures to ensure ongoing
structural integrity.
Further details can be obtained from Bill Young at WBM
Pty Ltd, PO Box 888, Morwell, VIC 3840, Australia.
www.wbmconsulting.com
G+D Computing

Coal Reclaimer
Wave Engineering in Western Australia used Strand7 to
investigate the stress levels of a Coal Reclaimer.
The client had been experiencing abnormal operation of
their portal type reclaimer and as a part of their condition
monitoring program began an investigation into the
structural integrity of the machine. The investigation was
carried out in a stepped approach, namely:

1. Preliminary FEA model to generate 'first cut'

stress levels and identify high stress points for
location of strain gauges.
2. Strain gauge monitoring of the machine (at the
high stress points) during machine operation to
measure fluctuating stresses over time.
3. Refinement of the FEA model to calculate static If you would like more information on this project then
stress levels in the structure, which were contact Ryan Hanrahan at Wave Engineering by
combined with strain gauge results to determine telephone on +61 (0)8 9726 2422. Or alternatively you
total stresses in the structure over time. can email him at rhanrahan@waveeng.com.au.

Did you know?

Ctrl-Click the Entity Show/Hide
Toggle
Normally, by clicking the show/hide
toggles, the screen automatically
updates and redraws to effect the
changes. Often we need to turn on/off
more than one entity and for large
models this can take time when
redrawing after each selection.
Instead, hold the Ctrl key while
Stress results throughout the entire structure were
clicking the entity toggles to prevent
required so a full plate model was generated. The final Strand7 from redrawing after each
model consisted of 3,900 beam and 158,000 plate click. When finished select redraw to
elements. The model was required to take into account refresh the screen. (F3 or F5)
the additional structural stiffness of the scraper boom and
the effects of a rotational pivot on the non conveyor end
bogie. The boundary conditions were set up in order to
allow free movement of the bogies along the rails;
this in turn allowed calculation of the natural structural
deformation of the machine.

The final requirement was the reporting of total operating

stresses (a combination of calculated and measured
stresses) throughout the entire structure. Included in this
was a detailed error analysis around high stress areas to
fully justify the mesh quality and hence reported stresses.

The job highlighted a good case of using FEA and

experimental procedures together. It would have been
difficult to position strain gauges in meaningful locations
without performing preliminary FEA first.

www.strand7.com
G+D Computing

Modelling Tip

F requently we are asked ‘what is the most appropriate way to model a bolted connection?’. Unfortunately there isn’t a
straightforward answer to this question! There are several different modelling options which can be used but depend heavily
upon the level of accuracy and local detail required. These options will be presented in order of increasing complexity
including the pros and cons of each option.

1. Simple beam model. 2. Simple beam model with 3. Beam model with rigid 4. Beam model with bolt
contact elements between links at the ends and head modelled at the ends
the plates. contact elements between and contact elements
the plates. between the plates.

This modelling option is used if
only the global effects of the
bolt are of interest and the local
effects of the bolt on the plate
mesh are not. The beam is
modelled with the appropriate
bolt properties.
Pros
– Very simple to model;
– Good global
representation.
Cons
– Local Singularities;
– Plates may pass through
each other.
Although similar to Option 1,
this option would be more
realistic as the plates are
restricted from passing through
each other by using contact
elements.
Pros
– A bolt pretension can be
applied;
– More realistic than Option
1.
Cons
– Local Singularities;
– Requires nonlinear solver.
Use this option if you want to This is the most realistic model
distribute the bolt forces to the you can obtain short of
surrounding plates. Ensure that modelling a full brick model.
the hole in the plate mesh is
Pros
the same diameter as the bolt.
– Bolt head interaction with
To improve this model, you the plates can be seen;
could replace the rigid links on – No singularities.
the bolt ends with stiff beams.
Cons
Pros – Complex to model;
– Less Local Singularities; – Requires nonlinear solver.
– Bolt force is distributed to
surrounding plates.
Cons
– Assumes full bolt contact;
– Requires nonlinear solver.

As you can see there are considerable differences between the different modelling options. For models where you are only
interested in the global effects, Option 1 will be sufficient. For models where you are interested in knowing the interactions
of the bolt head with the plates then Option 4 would be adequate. As there is no clear-cut answer to which is the best
option, we encourage you to experiment and discover which option would best suit your model.
If your analysis requires even further detail then you will need to model your bolt with brick elements.

www.strand7.com
G+D Computing

Label Boundary Condition

Benchmark α T = 20oC, R = 0.2 m2K/W
β T = 15oC, R = 0.2 m2K/W

I n each newsletter we will introduce a new benchmark that users can add to
γ
δ
T = 0oC, R = 0.05 m2K/W
Q = 0 (Adiabatic)
Benchmark Table 1: Boundary Conditions
those already supplied with the Strand7 software in the form of the verification
tests. Target values:
In this edition we will look at a heat transfer problem. The source of this Temperatures at the points shown in
benchmark is EN ISO 10211-1:1995 Case 3. Benchmark Figure 1:
The heat transfer in a structure consisting of two walls meeting in a corner and a U = 12.9 oC X = 12.6 oC
single floor is analysed. The boundary conditions applied are labelled with Greek V = 11.3 oC Y = 11.1 oC
letters α, β, γ and δ (see Benchmark Figure 2 and the Benchmark Table 1). W = 16.4 oC Z = 15.3 oC
Results of interest are the temperatures at the six points labelled U to Z in Heat flow through surfaces shown in
Benchmark Figure 1 and the heat loss/gain through α, β and γ surfaces shown in Benchmark Figure 2:
Benchmark Figure 2. Material and geometry data are shown in Benchmark
Figure 2. α = -46.3 J/s
β = -14.0 J/s
γ = 60.3 J/s

The building corner was modelled

using linear and quadratic versions of
the TETRA, WEDGE and HEXA brick
elements. The number of elements
employed in the models that are
used for checking solution
convergence is shown in Benchmark
Table 2.

Element Type Model

TETRA 6576
WEDGE 2192
HEXA 1096
Benchmark Table 2: Number of elements used

The heat flow results were obtained

by summing the node fluxes on the
relevant surfaces. The results are
summarised in Benchmark Table 3.
Benchmark Figure 1: 3D Steady heat transfer problem

Benchmark Figure 3: Temperature contour

Exclusively to readers of News.St7:

you can download this benchmark
model from the News.St7 readers
page of our website:
Benchmark Figure 2: Horizontal and Vertical Sections showing geometry, material properties and
boundary conditions www.strand7.com/News.St7.html

www.strand7.com
G+D Computing

Strand7 Results summary: An important point about this benchmark is

that it demonstrates that Strand7 can be used
Element Temperature Results Heat Flow Results
to determine heat flows (gains/losses) through
(degree difference) (% difference)
structures. This can be very important for
U V W X Y Z α Β γ estimating energy requirements in buildings
TETRA4 13.0 11.5 16.5 12.5 11.0 15.3 -47.7 -14.3 62.1 and other structures or processes.
(0.1) (0.2) (0.1) (0.1) (0.1) (0.0) (3.0) (2.1) (3.0)
The results in Benchmark Table 3 (with
TETRA10 12.9 11.5 16.4 12.5 11.1 15.3 -46.3 -13.9 60.2
differences shown in brackets) indicate that
(0.0) (0.2) (0.0) (0.1) (0.0) (0.0) (0.0) (0.7) (0.2)
total heat flows can be modelled quite well
WEDGE6 12.8 11.7 16.5 12.3 11.1 15.3 -47.6 -14.3 61.9 even with low order elements. This is a little
(0.1) (0.4) (0.1) (0.3) (0.0) (0.0) (2.8) (2.1) (2.7) different to what one would expect when using
WEDGE15 12.9 11.5 16.4 12.5 11.1 15.2 -46.3 -13.9 60.2 low order elements for stress analysis.
(0.0) (0.2) (0.0) (0.1) (0.0) (0.1) (0.0) (0.7) (0.2)
Finally, actually calculating the heat flows α, β,
γ and δ is very simple in Release 2.2 when
HEXA8 13.0 11.6 16.6 12.5 11.3 15.3 -47.2 -14.2 61.3 using the Flux summation option of the Peek
(0.1) (0.3) (0.2) (0.1) (0.2) (0.0) (1.9) (1.4) (1.7) tool, provided that the option Calculate Node
HEXA20 12.9 11.4 16.4 12.5 11.2 15.3 -46.3 -13.9 60.2 Flux has been set at the commencement of the
(0.0) (0.1) (0.0) (0.1) (0.1) (0.0) (0.0) (0.7) (0.2) solution.
Benchmark Table 3: Summary of results obtained from the Strand7 models.

Thanks to Imagineering UK, the Strand7 agent in the UK, for sourcing this benchmark.
Imagineering UK can be contacted on: Website: www.iuk.co.uk Email: info@iuk.co.uk

Training

T his year we have already held courses covering ‘Introducing Training Calendar
Strand7’, ‘Nonlinear analysis’ and ‘Structural engineering’.
If you wish to attend one of the courses, then check the dates in the
calendar and make your booking before it’s too late. You can find
full details of the content of the courses, prices and booking form at
2003
www.strand7.com; then select ‘training’ from the contents menu.
7-10 July Introducing Strand7
Did you know?
11 July Heat transfer
Shear centres of beams
By default, Strand7 beam elements 14 July Automeshing
are positioned such that the
element’s nodes coincide with the
section’s centroid. In sections such 15-16 July Nonlinear analysis
as C-sections, where the shear
centre is not located at the centroid, 17-18 July Dynamic analysis
significant twisting displacements will
be generated due to lateral (shear)
loads. If your structure is 11-14 Nov Introducing Strand7
constructed in such a way that this
twist is restrained, due to the load 18-19 Nov Dynamic analysis
passing through the shear centre,
you can offset your section such that
the nodes are located at the shear 20-21 Nov Structural
centre of the section. You can do engineering
this automatically with the Auto
Assign Beam Offsets tool.

www.strand7.com
G+D Computing

Exhibitions

T his year we have so far planned to exhibit at the shows listed

below. If you get the chance then we would be delighted to meet
you at these exhibitions and discuss Strand7 with you. If not, we
hope to publish a few snapshots of each exhibition in the next issue
of News.St7.

May 29-31, 2003

2003 Structures Congress & Exposition
Booth #304
Seattle, USA
www.asce.org/conferences/structures2003

June 2-4 2003

Atlantic Design & Manufacturing Show
Booth #1256
New York
www.atldesignshow.com

Strand7 Head Office

G+D Computing Pty Ltd

Suite 1, Level 7
541 Kent Street
Sydney NSW 2000
AUSTRALIA
Website: www.strand7.com
Email: info@strand7.com

www.strand7.com
G+D Computing

News.St7
Newsletter for Strand7 and Straus7 users
Strand7 is marketed as Straus7 in continental Europe

Issue 2, May 2004

Strand7 Release 2.3 Launched!

This month sees the launch of a new major release of 2.3 New Features Summary
Strand7.
New IGES Import Options
This represents yet another leap forward in functionality, New 2-Point Link Element
with new features in virtually all areas of the software. Load and Freedom Case Enhancements
While many of the new features have been developed in Inertia Relief
response to user requests, there are also a number of novel Quasi-static Seismic Analysis
additions such as the string group. New Element Attributes
In this newsletter we highlight two of the new features. For New Model Construction Tools
full details please take a look at the “What’s New” article Create Composite Material Libraries
which you can find on our web site: Identify Link Free Ends
www.strand7.com/whatsnew23.htm Load Patch Element
String Group
We value your feedback and have set up a dedicated Convert Time Domain Data to Spectral Curves
service to handle your comments. We would be very New Harmonic Solver Options
pleased to hear from you with any comments on this Powerful Enveloping of Results
release and also features that you would like to see FFT Analysis for Dynamic Problems
developed in the future. Please call us or send your 100 New API Functions
feedback to: news.st7@strand7.com Functionality and Help Improvements

Editorial Theory Manual In this issue…

W elcome to the first edition of our newsletter for 2004. T he Strand7 Theoretical • Release 2.3 highlights 2

We would like to thank all of our readers for the positive Manual is almost complete • Exhibition USA report 4
and useful feedback that we received after the publication and will be available in the
of our last issue. very near future. • Recommended hardware 5

We are delighted to announce Strand7 Release 2.3 and The manual will provide • Current projects 7
know that you will benefit from the new features. As users with a thorough
• Modelling tip 11
always, enormous effort has gone into the development, foundation in all of the
testing and documentation, all of which you will find theory utilised by Strand7. • Benchmark 12
meets the usual high standards that you have come to
Please keep an eye on our • Training calendar
expect from Strand7. 14
web site for a release
In this issue of News.St7 we have continued with the announcement. • Exhibitions calendar 14
theme of issue 1, by providing information and articles
that hopefully you will find interesting and useful. This
issue sees the continuation of the benchmarks and
modelling tips and at your request includes even more
‘did you know’ items!
If you have any feedback or suggestions regarding the content of News.St7, or you would like to have your latest project included in the “Current Projects”
section then please email: news.st7@strand7.com. If you would like to receive your copy of News.St7 directly by email, simply send a blank email to:
newsletter-subscribe@strand7.com

www.strand7.com
G+D Computing

Release 2.3 Highlights The conversion of plate load to beam load depends on
the load patch attribute assigned to the load patch plate.

T here are far too many new features and functions in

There are six different load patch attributes (shown below)
and these control both the proportion of load applied to
Strand7 Release 2.3 to adequately describe them all in each edge and the distribution.
this issue of our newsletter, so instead we focus on just
two of them. Detailed descriptions of the others can be
found in the latest Strand7 Online Help.

The Load Patch

The load patch is a plate element that is used for
transferring area loads and mass (such as pressure and
non-structural mass) onto a framework of beam elements.
In its simplest usage, a pressure load applied to a load
patch is transferred to four beam elements connected to
the four edges of the plate.
For example, Type 1 distributes the patch load to the
To use a load patch, you set the property type for the
beams on all four edges of a quad element (or three edges
plate to Load Patch and then assign one of the load patch
of a tri element) according to the tributary area associated
attributes to the plate. You can see that the property
with each edge. In the case of a rectangular load patch,
dialog below has none of the usual property data (such as
two opposite edges will be assigned a triangular beam
modulus).
load and the other two edges will be assigned a
trapezoidal load. The figure below illustrates this.

An important feature of the Strand7 load patch is that if

the beams along its edge are subdivided, a single patch is
still sufficient to correctly distribute the load to the beams.
In the figure below, the long beams have been subdivided
into 8 elements and the short beams into 5.

The load patch has no stiffness and no mass (although

mass attributes can be assigned to it), so its inclusion in a
model does not affect the stiffness or mass of the structure
you are analysing. This implies that without the
application of plate attributes, a load patch element is
basically redundant.

The way in which the patch loads are transmitted
differentiates the load patch from a normal plate/shell
element. The plate/shell element converts its loads to a
consistent set of forces and moments that are applied only
to its nodes. The load patch however, converts its load
(and mass) attributes to distributed loads (and mass)
applied directly to the beam elements (not the patch
nodes) connected along the edge of the patch. In other
words, you will not obtain the load patch result by using
a plate/shell element with a zero modulus to transfer
pressure loads to beams.
Conversely, the load patch may rest on part of a beam
and the appropriate load is then applied to that part of the
beam covered by the load patch.
Many of the attributes that are supported by plate/shell
elements can be assigned to load patch elements for
conversion to equivalent beam attributes (including
pressure, shear and non-structural mass). This conversion
is performed automatically (and transparently) at the
commencement of the solve. There is also a tool to
convert the patch attributes to actual beam attributes if
you prefer to see the beam loads before solving – in this
case you could delete the patch plates after conversion.
2

www.strand7.com
G+D Computing
Finally, to facilitate the application of patch loads to
complex frameworks, especially 3D frameworks, the
Create Load Patches tool can be used. This tool searches
for convex connections of (at least 3) beam elements in a
framework and automatically creates one or more patch
plates, depending on the geometry.
The String Group
This is a very exciting addition to the capabilities of
Strand7, which allows easy modelling of a string or cable
type of element that is fed through one or more pulleys
within a structure. Consider the following very simple
example:
Point A is fixed and at points B and C there are frictionless
pulleys through which a cable (string) passes. To model
this situation in Strand7, we connect together three truss
elements to represent the segments of the string and then
group these elements together by assigning a String Group
attribute to them. This information tells the Strand7 solver
to treat the three truss elements as a continuous string that
can slide freely over the pulleys. The pulleys in this
simple case are just fixed node restraints.
The following figure shows the solved Strand7 model. A
single point mass acted upon by a 1g acceleration
represents the weight. The blue lines represent the string
(truss elements) and the red lines are the axial force
diagram.
As the figure illustrates, the axial force in the three
elements is the same. Had this been run without the
string group attribute, then we would have found zero
force in the two elements on the left hand side.
There are many other more important applications for the
string element, including the following:
External post-tensioning of a concrete bridge beam
Post-tensioning is a popular technique used in the
construction industry to improve the performance of a
www.strand7.com
structure and increase its load carrying capacity.
Consider the figure below, showing the simple
arrangement of a concrete beam post-tensioned by a
single ungrouted tendon.
A Strand7 model of this beam and tendon is now very
easy to construct because no special procedures are
needed to model the freely sliding tendon. Simply define
the tendon by a series of truss elements (four were used
for this model) and group these all together with a string
group attribute. The figure below shows an exaggerated
display of a Strand7 brick and tendon (string group)
model, subject to the tendon tension only. The tension in
the tendon is constant.
Erection of a deployable structure
The figure below shows three steps during the erection
procedure of a frame structure. In this example, a tendon
(shown in red) is fixed to one end of the structure and
passes though a series of frictionless pulleys through the
base of the structure. The tendon is tensioned from the
free end and this allows the structure to take up its desired
shape.
In this model, the tendon is modelled with eight separate
truss elements, all grouped together under the same string
group. The tension is applied via a prescribed
displacement pulling the free end of the tendon
downwards.
In this first release of the string group, the pulley through
which the string passes is assumed to be frictionless. A
future release of Strand7 will allow pulleys with friction to
be modelled
3
G+D Computing

Exhibitions - Report In conjunction with our recent shows in the USA and
China, we are pleased to announce the appointment of
two new Strand7 distributors:
I n the past few months we have exhibited Strand7 at Beaufort Analysis Inc has been appointed as a Strand7
several conferences in the USA, China and India. distributor in the USA. They can be contacted at:
We are pleased to report that Strand7 was very well Beaufort Analysis, Inc.
received all round. We have had strong interest from 111 Safrit Drive
engineering companies operating in all of the mainstream Beaufort, NC 28516, USA
engineering disciplines and also in several of the more Tel: +1 252 422 0104
specialised fields. Email: info@beaufort-analysis.com
Website: www.beaufort-analysis.com
Strand7’s popularity stems, in part, from the fact that
whilst it is described as “general purpose” FEA software, it Sizhuanda Co. Ltd. has been appointed as a Strand7
contains advanced features that are usually only available distributor in China. They can be contacted at:
in premium priced FEA packages.
Sizhuanda Co. Ltd.
115 Fucheng Road,
Haidian District,
Beijing 100036, China
Tel: +86 10 8811 2908
Fax: +86 10 8811 0350
Email: info@strand7cn.com
Website: www.strand7cn.com

Find out more about future exhibitions in our section on

the back page. Our exhibition schedule is also regularly
updated on the Strand7 website: www.strand7.com

NASCC 2004 Steel Conference, Long Beach, USA.

Many of the interested parties were very pleasantly

Chinese and Japanese Interfaces
surprised to discover the depth of analysis that Strand7
can perform. As always, the Windows-native and
intuitive feel of Strand7, combined with functionality and
W ith the ever increasing popularity of Strand7
incredible value for money were big winners! worldwide, we are pleased to announce that the Strand7
interface is now available in Chinese and Japanese.

Chinese Interface

Pacific Design & Manufacturing Show, Anaheim, USA.

In the USA, China and India, Strand7 has been purchased

by clients working in areas of marine, structural
composites, structural and automotive engineering, and
by a number of universities and polytechnics. The list is
ever growing.
Our move into these new markets is coupled with a
significant increase in resources available to our research
and development and technical support teams. Therefore
Japanese Interface
we have many new exciting features planned for future
Strand7 releases.
4

www.strand7.com
G+D Computing

The RAM
Recommended PC Hardware

T he objective of this article is to identify those

components of a computer system that have the greatest
effect on the performance of Strand7. This article is The Random Access Memory (RAM) is where most of the
aimed at the Strand7 user who is on a budget and wants CPU temporary data is stored. Data which fills up the
to spend money on the components that count. RAM will spill onto the computer’s hard drive for
temporary storage. Access to data stored in RAM is
Our support team frequently gets enquiries regarding the extremely fast compared to access to data stored on the
best computer specification to purchase. There is no hard drive. If possible we want to minimise spillage onto
doubt that with an unlimited the hard drive and exclusively use the RAM
spending budget, you will get as temporary storage. This simply means
the best computer with the Did you know? that the more RAM you have, the better.
best performance. However
with the rapid rate of
Transparent User Stress Limits At the start of a solver run, Strand7 reports
technology progression, what If you’re interested in the sections of your the optimum amount of RAM needed for
is the best now will most
model that exceed the design limit, you can the model. If you have significantly more
certainly be superseded
contour your results such that anything than the reported optimum amount then
outside the defined limits is transparent. you can be sure that your model
within months if not weeks. This makes it easier to check complex calculations are not being slowed down by
With this in mind, it is usually internal components without having to access to data on the hard drive. The
manually hide outer elements.
not cost effective to pay for reason you may need significantly more is
the fastest and most powerful In Results/Settings under the Style tab, clear that Windows itself utilises a portion of
the < Min and > Max options.
computer system on the RAM for the operating system and
market as they attract a higher Now, under the Limits tab, set the user background tasks.
defined limits to a minimum value of say the
premium. Consider buying
the next best. They will often
design limit and the maximum to the A point to note is that Windows can only
maximum calculated. utilise a maximum of 4GB of RAM.
be substantially cheaper yet
suffer only slightly in the
Only those elements with stress over the However a single process, such as Strand7,
allowable limit will be contoured with the rest is only allocated a maximum of 2GB RAM,
performance stakes. of the model being transparent.
no matter how much RAM is on the system.
The two components that Installing more than 2GB is still an
affect the performance of advantage however, as this provides
Strand7 the most are the CPU Windows with memory which can be used
and the RAM. for other functions such as disc caching,
Other factors contributing to multitasking etc.
the overall performance of Windows also provides its own virtual
Strand7 are also discussed. memory management via a so-called swap
file so that if insufficient RAM is available,
the hard disc is used instead. The system
The CPU setup tools in Windows offer the option to
manage the swap file automatically or to fix
its size manually. In most cases, Strand7
(and other applications) will run more
efficiently if the swap file is set to a fixed
size. Fixing the swap file size to twice the
amount of RAM on your system is usually a good setup.
In terms of Strand7 performance, the most significant
attribute of the CPU is its clock speed.
The Video Card
The two dominant CPU manufacturers are Intel and
AMD. At the time of writing the Intel Pentium4 2.8GHz
HT and the AMD XP2800+ appear to represent the best
value for money.
Strand7 includes code specifically optimised for each of
these CPUs so you can rest assured that no matter which
CPU you choose, Strand7 will be performing to the best
of the CPU’s ability. In Strand7, go to File/Preferences Installing a good quality video card will noticeably
and click on the Solver tab to see how Strand7 has decrease the time it takes to refresh a model on screen in
identified your CPU. the Strand7 pre and post-processing environments.
5

www.strand7.com
G+D Computing

Dynamic rotations will also be much smoother.
Speed enhancement can be achieved by reducing your
screen colour depth to 16 bit from 24 or 32 bit. There is
very little noticeable difference to the naked eye and the
processing overheads are dramatically reduced.
The Motherboard
terminate and an error is reported.
If the files being written are large then the disk speed will
noticeably affect the solution time. The disk speed also
has an effect when there is insufficient RAM available and
the hard drive is used as “virtual memory”.
With the price of hard drives falling, it is advisable that a
minimum size of 80 GB is selected with disk speeds
≥7200 rpm.
To get the best performance out of your hard disc, we
recommend that you regularly perform a disc
defragmentation. You should also be aware that if you
write to result files across a network then the write time
can significantly increase.

The motherboard is responsible for the communications
between the various components of your computer
system.
The Front Side Bus (FSB) speed is the most significant
attribute of the motherboard specification.
In terms of Strand7 performance, the faster the better.
Purchasing
Although building your own computer system can give
cost savings, if you are not confident, it is advisable that
you let an expert computer technician do it for you. If
you purchase a complete system, you usually obtain more
comprehensive warranties and benefit from more
comprehensive testing of component compatibility.
It is wise to purchase from companies that will provide
good after sales support. These companies should have
well supported software/drivers and are more likely to
release service fixes if problems are found.

The Hard Drive After reading this article you should know how a specific
component of a computer system affects the performance
of Strand7. If you need any further information then
please contact our support team

API - Users Application Catalogue

The hard drive is used to store files. Strand7 needs to
store the model and results files and also the files that are
created during the analysis. Most of the analysis files are
T he Strand7 API was released mid 2003 and is
temporary and hence their creation and deletion is already being used by many users.
“invisible” to the user. We have received several requests to set up a data
Did you know? exchange catalogue containing examples of user
Your hard drive needs to be
large enough to store these applications. We are now in the process of
Space - Shortcut to toggle
files. selection arrow compiling this catalogue which will soon feature on
the Strand7 website.
In large models, the solver Have you ever tried to select
and results files can be quite an element only to discover We will provide summary details of the API
large – especially in transient
that the entity selection arrow application developed and details of the developer
is not active? who can be contacted for further information.
dynamic analyses.
The largest temporary analysis Unfortunately, at this stage, we are not in a position
file is usually the matrix file. Well, there is no need to use to verify the accuracy of each user’s application.
At the beginning of the the mouse pointer to toggle
If you would like to include your application in this
the arrow. Simply hit the
solution the solver reports the catalogue then please send us your contact details
space bar.
matrix size and also the “free and we will forward you a short Application
scratch space”. If the free With Release 2.3 single key
shortcuts are also available to Description Form to fill out.
scratch space is not large toggle the entity selection
enough then the solution will Please send the information to: api@strand7.com
buttons: node, beam, plate,
brick, link, vertex, geometry.
6

www.strand7.com
G+D Computing

The tank and support structure was modelled as a

Current Projects
complete unit which allowed for accurate footing loads to
be extracted for all load cases.
I n this section we present current projects where Strand7
is being used to perform FE analyses. Projects from both
Strand7 users and our own Strand7 Consulting are
presented.
Our thanks go to Outokumpu Pty Ltd, Crib Point
Engineering and Enmin Pty Ltd for their contributions to
this issue.
If you would like to have your project included in the
next edition of News.St7, please email:
news.st7@strand7.com

Goro Nickel Project

Outokumpu is the largest provider of mineral processing
technology in Australia. In 2002 they were awarded a
contract for 9 paste thickeners for the Goro Nickel Project
in New Caledonia. These thickeners are 20m in diameter
with a 10m sidewall above a conical floor section which
makes the tanks almost 20m tall.
All structural aspects of the thickener were modelled
using Strand7 including the high torque rake drive unit,
the drive shaft, raking mechanism, the bridge structure,
tank and support structure.
FEA seismic reaction (inertia force) contour plot
This was particularly useful when the client requested the
calculation of the individual footing loads resulting from
the application of the predicted settlement of the floating
foundation slab. G+D Computing recommended using
prescribed displacements, as specified by the client, to
tackle this challenge. This method then allowed
Outokumpu to quickly check for any stress concentrations
and also create a datasheet showing individual footing
loads for a series of slab settlement cases. At the
completion of the project, the client expressed their
appreciation in writing for the timely manner in which
Outokumpu was able to answer the numerous requests
for detailed footing information. This would not have
been possible had the structure not been modelled using
FEA.

Once fabrication of the equipment

Cutaway general arrangement of thickener
The drive shaft and raking mechanism
model was created in AutoCAD Inventor as
a solid model and imported into Strand7
where it was automeshed with plate
elements. This method was perfectly
suited to the geometry as it had numerous
pipe to pipe intersections at different
angles which would have been particularly
difficult to model manually. Making
changes to pipe sizes and orientations was
extremely easy and a great deal of time
was saved using the automeshing tools.
Did you know?
Applying Attributes
When applying attributes to an
element, for example Beam Rotational
End Releases, selecting the beam
before opening the attribute dialog box
will apply the release to both ends.
But, what if you only want the release
at one end?
There is a simple solution. Before
selecting the beam, bring up the
required dialog box. Now select the
end of the beam you want to apply the
release to and you will find that only
half the beam is selected. Only this
end will be assigned the attribute.
Strand7 uses this approach for the
other attributes, such as brick face
and plate edge attributes.
had started in the Philippines, the
client requested that Outokumpu
assess the possibility of transporting
the tank and mechanism as a
complete unit to site in New
Caledonia. As the component
models were already constructed,
they were joined together and this
allowed Outokumpu to inform the
client of the exact centre of gravity of
the structure and also design
transport beams that would allow
two large crawler trucks to transport
them over the final journey by road
to the processing plant.

www.strand7.com
G+D Computing

significantly to the governing loads, acceptance criteria,

engine mass and panhard rod configuration.
The FE model consisted predominantly of Quad4 shell
elements. Rigid links and lumped masses were used to
model the engine, transmission and drive shafts.
Spring/damper elements were used to model the isolation
bushes used at six connection points within the drive
system. Contact elements were introduced to accurately
model contact faces and restrict excessive movement in
certain locations under the ultimate loads.

At present, the fabrication portion of the contract is

complete and is awaiting client approval for final
assembly and shipping to site.

You can obtain more information about this project

from Hamish Mackie, Outokumpu Pty Ltd.
Tel: +61 2 9984 2500 Fax: +61 2 9984 2501
www.outokumpu.com.au
An initial FE assessment was performed. 26 load
combinations were considered for the proof loads and a
further 26 for the ultimate loads.
VLocity Traction Engine Beams These loads considered all of the possible combinations
of positive and negative accelerations in the longitudinal,
Crib Point Engineering were contracted by Cummins to
transverse and vertical directions.
design and manufacture the main traction engine support
beams for the new VLocity trains which will be operated Following the initial assessment, areas were identified
throughout Victoria, Australia. that required modification and optimisation to meet the
design requirements.
Crib Point Engineering employed the specialist services of
Strand7 Consulting to perform a Finite Element Analysis Under close cooperation between G+D, Crib Point
to investigate the required proof and ultimate loads and Engineering and Cummins, various remedial options
also to perform a fatigue evaluation. were investigated.
All of these options considered constructability, code
compliance, potential
clash with ancillary Did you know?
components and
significantly the maximum CTRL-Click Edit Nodes
permissible weight for the
Moving nodes can be
entire system. done with 2 simple clicks
This iterative process was in Strand7 just like drag
and drop.
first completed for the
proof loads and Bring up the Edit/Nodes
dialog box.
subsequently for the
ultimate loads. Holding down the CTRL
key, click the node you
Proof load acceptance was wish to move with the left
governed by maximum mouse button. Now
stress criteria whereas the release the CTRL key
and click the node you
ultimate analysis wish to move it to.
considered both material and
All Done! But don’t forget
The preliminary design configuration was taken from geometric nonlinearities to to clean the mesh after
previously designed traction engine beams. For this ensure that gross failure and you’re finished.
project, changes were made to several areas, but most progressive collapse of the
8

www.strand7.com
G+D Computing
structure did not occur.
Finally, a fatigue assessment was made in accordance
with BS7608. Principal stress ranges were determined for
the applied fatigue loads. Each structural detail was
classified in accordance with the code and a fatigue life
calculated based on the summations of the fatigue
damage caused at each location under each load case.
The final configuration of support beams has resulted in a
design that has been optimised for fatigue life and
structural performance whilst minimising construction
costs, manufacturing time and overall weight of the
system.
You can obtain more information about this project from
Sean Turnbull at Crib Point Engineering.
Tel: +61 3 5979 1703 Fax: +61 3 5979 3323
www.cpe.net.au
Did you know?
Right-Click Attributes
Once you have applied an attribute to a particular element
you can easily apply a different type of attribute by right
clicking over the existing attribute dialog box.
For example, you have just assigned restraints to a node
and now you want to assign a force.
In the Node Restraints box, right-click to bring up a list of
attributes to choose from. Select Force and you will find
that the Nodal Force Attribute is now the active dialog box.
It’s that easy.
This works for all attribute types in Strand7.
www.strand7.com
Vibratory Feeder
Strand7 Consulting was employed by Enmin Pty Ltd to
investigate the structural integrity and estimate the design
life of a new support frame design for one of their
vibratory feeders.
Enmin provide materials handling solutions via a range of
vibrators and flow aid devices which are designed and
manufactured in-house.
The vibratory feeder under investigation is driven by a
pair of eccentric mass vibrators. These are operated in-
phase but in opposite directions. This results in the
oscillating vibratory action occurring predominantly in
one plane.
Enmin supplied Strand7 Consulting with CAD files which
were converted into mid plane models and subsequently
imported into Strand7 where the geometry was cleaned
and automeshed. A predominantly Quad4 plate mesh
was created. Isolation bushes were modelled with
spring/damper elements.
The new design required that there be clear space under
the front of the feeder. The design should also allow the
feeder to be easily cleaned, be visually appealing and
provide minimal areas for spilt material to collect.
The general arrangement of the feeder model is shown
below:
A nonlinear transient dynamic analysis was performed.
This allowed an investigation of the full start-up and
operational cycle.
A factor versus frequency table was created to replicate
one start-up cycle followed by constant speed operation.
The graph over the page shows the changing amplitude
(force) and frequency (motor speed) with time over the
start-up period.
G+D Computing

Stress ranges were calculated for the start-up and The final optimised design is now in successful
continuous operation speeds. These values were production.
assessed in accordance with the British Standard for
fatigue assessment. Further details can be obtained from Glenn McTaggart at
Enmin Pty Ltd.
The fatigue life was estimated based on the cumulative Tel: +61 3 9753 3633 Fax: +61 3 9753 3796
effects of start-up, shutdown and continuous operation. www.enmin.com.au

Some areas were identified that required modification to

achieve the desired design life. It was also found that the
structure had a significant mode of vibration with a
frequency close to the operational speed of the feeder.
In conjunction with Enmin, the feeder design was
optimised to reduce stress ranges and shift the vibrational Strand7 Consulting provides expert FEA services for
modes away from those encountered at the normal engineers. For more information on the services provided
operational speeds. This was achieved by redistribution please contact:
of mass using local plate thickness changes and also Steve Raynes at Strand7 Consulting
changes to the stiffness of the main joint connections. Tel: +61 2 9264 2977 Fax: +61 2 9264 2066
www.strand7consulting.com
An example of the plate principal 11 stress plot is shown:

Did you know?

Verification Manual
Confused about how to set up your nonlinear
model?
Example models can be found in the Strand7
Verification Manual.
You can find the “Verification Manual” folder
within the main Strand7 program folder on
your computer.
Here you will find around 350 Strand7 models
set up and ready to run.
The Verification Manual PDF contains the
documentation that describes each model and
compares Strand7 results with target values.
There are examples for all of the solver
options available in Strand7, divided into
separate sections for easy identification. It’s a
great first reference for how to set up your
model.

10

www.strand7.com
G+D Computing

Modelling Tip

I n this issue we will briefly discuss the modelling techniques that can be employed to model a fillet weld connecting two
perpendicular plates. There are generally several approaches to model any connection. The connection configuration and
level of accuracy required will dictate the appropriate modelling technique. The methods outlined below are the most
commonly used to model a fillet weld and a brief appraisal of each method is given.
*Please note that the mesh densities shown are for illustration purposes only.

1. Two intersecting plates 2. Increase thickness of 3. Model the fillet weld 4. Use a full brick model
plates in the weld zone material with plates

This method of modelling a
welded joint is by far the
simplest and most efficient. If
you had a large model with
many welded connections then
this would be the most practical
method.
Note that the stresses should be
considered at a specified
distance from the weld/plate
intersection.
Pros
• Very simple to model;
• Good global
representation.
This method is still relatively
simple, yet takes accuracy levels
a step further.
The additional plate thickness
more accurately models the weld
stiffness and the effective plate
spans.
Pros
• Simple to model;
• Good global
representation;
• Additional stiffness in
weld zone is accounted
for.
This method is used when you This method is used when
want to distribute the load to the considerable detail about the
actual area of the weld and connection is required.
consider the additional stiffness
For clarity, the sketch shows a
of the weld material.
crude mesh. Mesh refinement
It is also possible to model the would be required for model
free edge between the red/blue accuracy.
plates where the weld is not a
full penetration type.
Pros Pros
• Simple to model;
• Can obtain highly
• Allows stresses to be
accurate results;
extracted for weld
• Removes the singularity
material itself;
at the plate intersection;
• Weld configuration can
• The ability to extract
be more accurately
results at the weld toe
modelled.
and in the weld material.

Cons Cons Cons Cons

• Singularities at plate • Singularities at plate • Singularities at plate • Complex procedure to
intersection; intersections. intersection (although model lots of welds;
• Extra stiffness of weld stress estimation in the • Solve times increase
material ignored. critical zones is often dramatically.
simpler than with
methods 1 and 2).
Different design codes specify different distances from the weld toe to extract the calculated stresses. Please refer to your
specific design code for more information.
If you require detailed results around the weld zone, then it is advisable that you use brick elements to model the welded
connection as shown by method 4. Simple comparative models can be run to estimate percentage difference between each of
the modelling options. This can be used to justify results calculated via the simpler methods that can be used in larger global
models

11

www.strand7.com
G+D Computing

Benchmark
The displacement and stress at
various locations are given as a
I n this edition we will look at a benchmark problem that formed part of the function of applied load. In each
case, the x-axis values are non-
submission to the Hong Kong Buildings Department. This gained Strand7 the dimensionalised by plotting Pa4/Eh4,
technical approval required to support structural design submissions to the Hong instead of P, where:
Kong Building Authority. Subsequently, Strand7 has been used extensively for the
design and analysis of various structures throughout Hong Kong. P = Applied pressure
This particular benchmark problem makes use of the Strand7 geometric nonlinear a = Width of plate (2m)
solver. The solver is capable of analysing structures that undergo large E = Material modulus (2.0x1011)
displacements. The problem is defined by a square plate that is clamped at all
four edges and is loaded by a uniform normal pressure. The nonlinear h = Plate thickness (5mm)
displacement and stress response are compared with a published analytical This means that the x-axis values of
solution. pressure are multiplied by 0.128 for
The reference solution is given in National Advisory Committee for Aeronautics graphing purposes
(NACA) Technical Note No. 847, “Square Plate with Clamped Edges Under (24/(2.0x1011*0.0054) = 0.128).
Normal Pressure Producing Large Deflections”, by Samuel Levy, circa 1940 [1]. In the graph showing displacement,
The plate is a square plate of edge length 2m and thickness 5mm. The modulus is the y-axis is non-dimensionalised by
2x1011 Pa and the Poisson’s ratio is 0.316. It is clamped at all edges. A normal dividing the displacement by the
pressure of 3000Pa is applied incrementally in 16 equal steps. plate thickness, h = 0.005.
In the graph showing stress, the y-
axis is non-dimensionalised by
plotting σa2/Eh2 instead of σ. This
means that each stress value is
multiplied by 8.0x10-7 for graphing
purposes
(22/(2.0x1011*0.0052) = 8.0x10-7).
Using the customised graphing
options within Strand7, it is quite
straightforward to normalise the
results obtained in the manner
described above.
As shown in the figures on the
following page, an excellent
agreement is found between
Reference 1 and the Strand7 results.
In each newsletter we will introduce
a new benchmark that users can add
to those already supplied with
Strand7.
Benchmark Figure 1 – Symmetric quarter Strand7 model consisting of 4x4 mesh of Quad4
elements
Exclusively to readers of News.St7:
The Strand7 model consists of a 4x4 mesh of Quad4 elements modelling one You can download this benchmark
symmetric quarter. The Quad4 element in Strand7 is a thin plate/shell element model from the News.St7 readers’
based on Kirchhoff theory. It is suitable for the analysis of linear and nonlinear page of our website:
plate/shell problems and has the ability to accurately represent thin shell www.strand7.com/news.st7.htm
behaviour, even for elements with a significant amount of warping [2]. This
makes it particularly useful in geometrically nonlinear problems because in
general, as the elements deflect, the amount of element warping increases.

12

www.strand7.com
G+D Computing

Benchmark Figure 2 – Displacement vs Load from Reference 1. Benchmark Figure 4 – Stress vs Load from Reference 1.

Benchmark Figure 3 – Displacement vs Load from the Strand7 Result. Benchmark Figure 5 – Stress vs Load from the Strand7 Result.

References Stress Graph Legend

A: Extreme fibre stress ratio at the centre of the plate.
1. “Square Plate with Clamped Edges Under Normal C: Extreme fibre stress ratio at the mid-point of an edge of the
Pressure Producing Large Deflections”, National plate.
E: Membrane stress ratio at the centre of the plate.
Advisory Committee for Aeronautics (NACA) F: Membrane stress ratio at the mid-point of an edge of the
Technical Note No. 847, by Samuel Levy, circa 1940. plate.
Note that B and D are not shown on the Strand7 stress graph
because they are not automatically calculated.
2. “Strand7 Verification Manual”, Edition 1, G+D
Computing, 2000.

13

www.strand7.com
G+D Computing

Training Training Calendar

T his year we have already held the “Introducing Strand7” course. 2004
We have also conducted a number of onsite courses, tailored
1-4 June Introducing Strand7
specifically to the clients’ requirements. Over recent years the
8-9 June Nonlinear Analysis
demand for these tailored courses has increased, as it allows the 10-11 June Dynamic Analysis
course to be focused on the application of Strand7 to the clients’ 15-16 June Structural Engineering
specific requirements. In addition, there are cost savings for an 17 June Automeshing with Strand7
organisation that needs to train several engineers. 18 June Heat Transfer Analysis
12-15 October Introducing Strand7
If you are interested in tailored onsite training courses then please 19-20 October Nonlinear Analysis
contact us for more information. 21-22 October Dynamic Analysis
In June and October we will be running our full series of courses 26-27 October Structural Engineering
28 October Automeshing with Strand7
ranging from Introduction through to Structural Analysis.
29 October Heat Transfer Analysis
If you wish to attend one of the courses then check the dates in the
calendar and make your booking before it’s too late. You can find
full details of the content of the courses, prices and booking form at:
www.strand7.com/training.htm

Exhibitions

W e plan to exhibit at the following conference:

17 – 20 China International Steel Beijing Exhibition

August 2004 Construction Expo/Conference Centre, China

If you get the chance then we would be delighted to meet you and
discuss Strand7 with you. You will have the opportunity to meet
representatives from Strand7 and from our distributor in Beijing.
We will be keeping you updated on the latest developments in
future editions of News.St7

Strand7 Head Office

G+D Computing Pty Ltd

Suite 1, Level 7
541 Kent Street
Sydney NSW 2000
AUSTRALIA
Tel +61 2 9264 2977
Fax +61 2 9264 2066
Website: www.strand7.com
Email: info@strand7.com

14

www.strand7.com
G+D Computing

News.St7
Newsletter for Strand7 and Straus7 users
Strand7 is marketed as Straus7 in continental Europe

Issue 3, October 2004

Strand7 Solutions in Athens 2004!

Last month saw the XXVIII Olympic
Games held in the birthplace of the
original Games, Greece. The
preparation leading up to the Games
was highly scrutinised by the media,
with the involvement of various
Australian companies forming the basis
of some of their stories. This newsletter
reports on the work undertaken by one
of the engineering consultancies
involved in the design and analysis of
the centrepiece of the Games, the main
stadium, together with the vital role that
Strand7 played in the analysis.
Also in this issue we focus on some
details of Strand7 including Links, Points
and Lines, a new section on the Strand7
API and how a thermal stress analysis
can be completed efficiently.
As always, the newsletter contains a new benchmark and is packed with helpful Did You Know? items.

100m. Each covering is

2004 Athens Olympics
supported by a curved arch In this issue…
A highlight of the recent Olympic Games held in Greece, structure comprising an
was the Olympic Athletic Centre of Athens Stadium, upper arch tube and a ● Feature Article 1
named the OAKA Stadium. This stadium was used for both lower arch torsion tube.
● Strand7 Element Use 3
the opening and closing ceremonies of the Games, as well The upper arch is
as for numerous high-profile events such as athletics. constructed from a circular ● Strand7 Functions 4
hollow section 3m in
As part of the redevelopment of the Olympic site for the diameter with a variable ● Modelling Tip 6
Games, a new roof structure was built over the main wall thickness of up to
stadium. This roof was designed by the renowned Spanish ● API in the Field 11
100mm. The lower arch is
architect and engineer Santiago Calatrava. Sinclair Knight similar to the upper arch ● Benchmark 13
Merz (SKM) were Engineering Consultants on the project. but is 3.6m in diameter.
SKM’s brief was to engineer the entire OAKA roof structure. ● Announcements 15
Working out of their London office, a team of SKM The two arches on each
engineers from the UK, Australia and New Zealand used covering are connected by ● Training Calendar 16
Strand7 to undertake this project. a row of steel cables
arranged in a vertical plane. ● Exhibitions calendar 16
Details of the Roof Structure Steel diaphragm plates are ● User Profile 16
The roof structure consists of two coverings, each spanning installed at every cable
over 300m in length and varying in width from 60m to connection point. The
If you have any feedback or suggestions regarding the content of News.St7, or you would like to have your latest project included in the “Current Projects”
or User Profile section then please email: news.st7@strand7.com. If you would like to receive your copy of News.St7 directly by email, simply send a
blank email to: newsletter-subscribe@strand7.com
1

www.strand7.com
G+D Computing
main roof rafters are built into the lower arch and cantilever
on both sides of the arch. An additional row of cables
connected to the top arch supports the external rafters
along the perimeter of the stadium, while another row of
cables supports the internal rafters at approximately mid-
span. Transparent cladding (polycarbonate roof tiles) are
installed over the rows of purlins, perpendicular to the
rafters.
Figure 1.1 - In-progress construction of arch.
Erection
The entire roof structure is pin-supported at only four
points. It was initially erected in two symmetrical halves at
temporary locations at each side of the stadium; this
allowed construction on the roof to proceed without
hindering works on the main stadium.
Figure 1.2 - Aerial view showing the east side arch in position and
the west side arch ready to be moved into position.
Each half was then moved into its final position by sliding it
along temporary supports where they were connected
together at only two points. Finally the bases were fixed to
permanent supports.
An impressive animation of this procedure can be found at
www.athens2004.com/en/Videos/multigallery.
Strand7 Modelling
SKM used Strand7 extensively for the modelling and
analysis of the OAKA Stadium roof structure. Because of
Strand7’s comprehensive suite of finite elements and
sophisticated user interface, Strand7 was used not only for
modelling the complex stadium roof structure but also for
simulating the entire erection process.
For modelling the erection process, one of the most useful
features of Strand7 was its ability to start a new analysis of
www.strand7.com
the model using the results of a previous run as the initial
conditions. “Using this and other features of Strand7, we
were able to simulate the stages of construction by applying
new conditions and loads to a model that was already
solved in a previous run,” said Ljubisa Petrovic of SKM.
Another use of this feature was to establish the pre-tensions
in the cables required to achieve a desired final roof shape
and final levels of tension in the cables. This procedure
can be easily automated using the Strand7 API.
Figure 1.3 - Strand7 model showing connection of roof girders to
lower torsion tube.
Figure 1.4 - Strand7 model showing the arrangement of cables
between the upper and lower arches.
Another extremely useful feature for SKM was Strand7’s
ability to save the deflected geometry of the model and
reuse it as a new and unloaded model for a subsequent run.
This approach was used for calculating the preset roof
geometry and to determine the initial shape that would
produce the required final shape under the applied loading.
As the two sections of the roof structure were to be
assembled using a temporary support structure, these tools
enabled SKM to establish the initial geometry such that after
the complex erection procedure was completed, the entire
roof deflected into the final shape as designed by the
architect.
SKM engineers involved in the Strand7 simulation included
Nick Ling (project manager) together with Ljubisa Petrovic
(senior structural engineer, structural analyst) and Louise
Harrod (structural engineer), the engineers responsible for
the development of the Strand7 models and their
execution. SKM may be contacted through their web
address www.skmconsulting.com
2
G+D Computing

Master-Slave vs. Rigid Links

When these two models are solved using the linear static

S trand7 supports seven different link types to allow the

solver, the different behaviours of the two links can be
clearly observed.
relationship between the displacements/rotations of two
nodes to be precisely specified. Among the more
commonly used link types in Strand7 are the Master-Slave
links and the Rigid links.
A common misconception concerning these two link types
is that a Master-Slave link with all degrees of freedom
linked is equivalent to a Rigid link set in Global XYZ. This
article discusses the differences between these two links.
Master-Slave link Definition
This link defines a master-slave relationship between two
nodes. One node follows the selected displacement
components of the other node (and vice versa) in the global
Cartesian system or in a User Coordinate System. Any Figure 2.2 – Difference between Master-Slave link and Rigid link.
combination of DX, DY, DZ, RX, RY and RZ may be linked. The Master-Slave link ensures that the global (or UCS)
displacements and rotations of the linked nodes are the
Rigid link Definition
same. The Rigid link ensures that the relative distance and
This link provides an infinitely stiff connection between two rotation between the linked nodes does not change.
nodes.
Example 2
Comparison
Two cantilevered beam models are shown in Figure 2.3.
The difference between the two links is that a Master-Slave The model on the left has a Master-Slave link with all
(set in all d.o.f) links all freedoms (both translation and degrees of freedom set. The one on the right has a Rigid
rotation) and therefore cannot rotate, but only translate as a link set in Global XYZ. Both models are fully restrained at
rigid body. The displacements and rotations are identical the base and loaded with an identical vertical force.
and so the two nodes will translate in space such that the
orientation of the displaced link remains parallel to the
original orientation. A Rigid link can be thought of as an
infinitely stiff beam and can rotate as a rigid body.
Displacements and rotations are coupled together by
considering the rigid body rotation of the link and the fact
that in the direction along the axis there is no relative
displacement, (i.e. the link cannot extend or shrink). As the Figure 2.3 – Cantilever Beam Model.
length of the link approaches zero, the behaviour of the
Left: Master-Slave link; Right: Rigid link.
Master-Slave and Rigid link becomes the same.
These two models are solved using the linear static solver
The following two examples illustrate the differences
and their response is shown in Figure 2.4.
between the two link types.
Example 1
Two simple beam models are shown in Figure 2.1. The
model on the left has a Master-Slave link with all degrees of
freedom set. The one on the right has a Rigid link set in
Global XYZ. Both models are fully restrained at the base
and loaded with an identical lateral force. Figure 2.4 – Response of beam under vertical loading.
From Figure 2.4, it seems that there is no difference
between the two link types. From the dimensions given in
Figure 2.3, we expect the moment reaction about the fixed
base to be:
M Base = F × d
M Base = 100 × 625
M Base = 62500 Nmm
However using the Strand7 Peek tool, a moment reaction of
50000Nmm is obtained for the Master-Slave link model.
This is shown in Figure 2.5.

Figure 2.1 – Laterally loaded beam model.

Left: Master-Slave link; Right: Rigid link.

www.strand7.com
G+D Computing
Figure 2.5 – Moment reaction for Master-Slave link model.
Strand7 does not consider the length of the Master-Slave
link in its calculation. For this model, the link is simply
used to transfer the node restraints at the base node to the
other linked node and hence the beam is considered
cantilevered about its left end with a moment arm of
500mm instead of 625mm.
Using the Strand7 Peek tool again, a moment reaction of
62500Nmm is obtained for the Rigid link model as shown
in Figure 2.6.
Figure 2.6 – Moment reaction for the Rigid link model.
Strand7 does consider the length of the Rigid link in its
calculation and therefore the expected moment reaction is
obtained.
Conclusion
Hopefully the two examples shown have clearly illustrated
that a Master-Slave link with all degrees of freedom set is
generally not equivalent to a Rigid link set in Global XYZ.
A Master-Slave link with all degrees of freedom set is a
mathematical tool and you should exercise caution when
using this arrangement. If used, the length of the element
should be minimised to reduce any moment-loss effects.
Master-Slave link summary:
● Transfers displacements/rotations between
linked nodes.
● The length of the link is not considered in any of
the calculations.
● Particularly useful when overlaid on point contacts
to ensure that they remain perpendicular to the
contact surfaces (linking only the degrees of
freedom required to keep the contact
perpendicular).
● Also very useful in construction sequence as the
state of the Master-Slave can be changed in a
nonlinear run between load increments, when
used in conjunction with the nonlinear restart
procedure.
Rigid link summary:
● Provides an infinitely stiff connection between the
linked nodes.
● The length of the link is considered in calculations.
● Particularly useful for connecting an offset lumped
mass to some part of a structure.
www.strand7.com
Details of this and other theoretical issues can be found in
our new Theoretical Manual.
Happy modelling!
Points and Lines
T he Points and Lines function provides an invaluable
modelling tool for mesh generation and modification.
Located within Tools, Points and Lines allows the user to
generate and modify a mesh by creating or moving nodes
and beam elements according to the chosen geometric
option.
Straight and Parabolic Lines
Straight or parabolic lines consisting of nodes or a series of
beams (depending on whether the Create Beams check
box is set) may be generated using the Straight Line and
Parabola options. These options are particularly useful for
defining beams prior to extrusion or for defining nodes for
the manual creation of plates and bricks.
Straight Line Parabola
Figure 3.1 displays the Parabola option. A quadratic curve
is created as a series of beams between three points to
define the geometry.
P3
P2
P1
the
Figure 3.1 – Points and Lines/Parabola.
Fillets
Fillets of a specified radius consisting of nodes or beam
elements can be easily generated using the Three or Four
Point Fillet options. Uses of the fillet options include
modelling the details of welds in plate sections prior to
extrusion or generating user-defined beam sections.
Three Point Fillet Four Point Fillet
Figure 3.2 shows the use of the Three Point Fillet option. A
fillet is created, followed by editing the beam elements
using Edit-Element to create an intersection between the
fillet and straight beams.
G+D Computing
Figure 3.2 – Points and Lines/Three Point Fillet.
Line Extensions, Normals and Intersections
Lines may be extended or trimmed or beams created
tangentially to a line direction using the Line Extend option.
An intersection between three or four nodes may be found
using the Line Intersection tool. The Line Normal tool
allows nodes or beams to be defined normal to a line in a
specified plane. These tools are particularly useful for
defining beams normal to a surface that will later be
defined as point contact elements.
Line Extend Line Normal Line Intersection
The use of the Line Normal tool is shown in Figure 3.3. A
beam normal to the blue plates is created. Using
Edit/Element, the nodes of the blue plate elements are
aligned with the beam, followed by smoothing of the plates
using Tools/Smooth Plates.
Figure 3.3 – Points and Lines/Line Normal.
Node Averaging
The Node Average tool allows nodes to be created or
relocated to a midpoint between two nodes in terms of the
specified coordinate system. The Cartesian Node Average
tool produces a node at an averaged position according to a
www.strand7.com
Cartesian coordinate system. The UCS Node Average tool
produces an averaged node position according to the
selected UCS.
Cartesian Node Average UCS Node Average
Figure 3.4 illustrates the use of the UCS Node Average
option. A node is aligned such that the coordinates are an
average of the selected corner nodes according to a
cylindrical UCS. The positioning of the node allows a
Quad8 element to be correctly defined using
Create/Element, with the repositioned node defined as a
mid-side node of the element, still lying on the correct
radius.
Figure 3.4 – Points and Lines/UCS Node Average.
Circles and Ellipses
Circles and ellipses consisting of nodes or a series of beams
may be defined using the Two Point Circle, Three Point
Circle, Ellipse and Variable Radius Curve. The Two Point
Circle and Three Point Circle tools allow the definition of a
partial or full circle. The centre of a circle may be located
using the Find Circle Centre tool. The Ellipse and Variable
Radius Curve options allow a partial or full ellipse to be
defined.
Two Point Three Point Find Circle Ellipse Variable
Circle Circle Centre Radius Curve
Figure 3.5 uses the Find Circle Centre tool with the Create
Beams option set. By selecting three points on the circle, a
node defining the circle centre and three beam elements
are created. Using the Entity Inspector (Shift), the circle
radius may be determined from the beam length. This is
particularly useful for QA purposes.
5
G+D Computing

Figure 3.6 – Points and Lines/Circle Line Fillet.

A comprehensive description of each geometric option is
available within the Strand7 Online Help
Figure 3.5 – Points and Lines/Find Circle Centre.
Circle Intersections and Tangents
The intersections and tangents of circles may be determined Modelling Tip – Thermal Stress Analysis
using the Two Circles Intersect, Circle Line Intersect, Circle
Line Fillet, Two Circle Fillet, Two Circle Tangents and One
Circle Tangent tools. The Two Circles Intersect and Circle
I t is well known that a variety of thermal factors can
greatly affect the structural behaviour of a body or
Line Intersect tools allow the intersections between circles component. With service temperatures having an influence
and lines to be determined in the form of nodes or beams. on material properties as well as leading to load
Fillets may be created between circles and lines using the considerations in the form of thermal expansion or
Circle Line Fillet and Two Circle Fillet tools. The Two shrinkage, it is important to take these thermal effects into
Circle Tangents and One Circle Tangent tools may be used account when performing a finite element stress analysis.
to create nodes or beams tangent to the circles.
In this issue, we will look at a number of features available
in Strand7 that allow users to progress from a thermal
analysis to stress analysis in a few simple steps. Also
discussed is the manner in which the Strand7 structural
Two Circles Circle Line Intersect Circle Line Fillet
Intersect
solvers use applied temperature load case data and Strand7
heat results.
1. Using Heat Results in the Strand7 Structural Solvers
Strand7 is equipped with solvers that are able to analyse a
Two Circle Fillet Two Circle One Circle Tangent variety of heat transfer problems including steady-state heat
Tangents problems - where the equilibrium temperature distribution
Figure 3.6 shows the Circle Line Fillet tool. A fillet is at infinite time is calculated, and transient heat problems -
created by selecting the circle centre, two points lying on where the temperature distribution is calculated as a
the straight line and defining the circle and fillet radii. The function of time.
circle and straight line beams were joined to the fillets Both the steady-state and transient temperatures calculated
using the Edit-Element option. from the heat solvers can be used as the input temperatures
in a thermal stress
analysis.
Did you know?
For example consider
the model shown in Subdivide only Normal Beams
Figure 4.1. Here, a When a model contains both
composite material is normal beams and other line
embedded in a high- elements, e.g. trusses and
springs, you may not want to
thermal conductivity subdivide the truss/spring
material maintained at elements when the beams are
400 oC. The upper subdivided as it could lead to
surface is exposed to a singularities.
convection Preferences within Strand7 can
environment at 30 oC be changed so that when all
with convection beams in a model are selected,
coefficient h = 25 only those that are normal beams
W/m2 oC. are subdivided.
Choose Tools/Options. Under
Subdivide set Subdivide only
Normal Beams.

www.strand7.com
G+D Computing

Tamb = 30 oC h = 25 W/m2 oC

400oC 400oC

Outer Inner Composite Figure 4.4 – Node temperature attribute dialog box.
Composite 400oC
material:
material: If the results from a steady-state heat solution are required,
make sure the *.sha file is selected for importing. The
k = 2.0 W/m oC
k = 0.3 W/m oC nodal temperature results from the steady-state heat
ρ = 2800 kg/m3 solution will then be applied to all the nodes in the model
ρ = 2000 kg/m3 C = 900 J/kg oC
C = 800 J/kg oC as a temperature attribute for the currently active load case.
If, instead, the results from the transient heat solve are
Figure 4.1 – Thermal stress analysis example. required for importing, ensure that the .tha file type is
After specifying the model’s boundary conditions and selected in the Import files of type drop down list.
material properties, either of the heat solvers can be used to Following this, you may choose to either import the
determine the temperature distribution through the temperatures from a single time step or from all the saved
structure: time steps from the transient heat analysis.
Point A Choosing the temperature distribution from one specific
time step, as shown in Figure 4.5, will import nodal
temperatures, for that time step, into the current active load
case.

Figure 4.2 – Steady state heat solution for the equilibrium

temperature field.
The contour plot in Figure 4.2 shows that the temperature
at Point A is approximately 242oC. The transient heat
solution reinforces these results with the evolution of the
temperature at Point A to its steady state value after
100 minutes of 242 oC shown in Figure 4.3.
Figure 4.5 - Importing nodal temperatures from a specific step of
the transient heat results file.
Alternatively, selecting to import all the saved steps in the
transient heat solution results in the nodal temperatures
from each of the time steps being allocated to different load
cases. The dialog is shown in Figure 4.6.

Figure 4.6 - Importing nodal temperatures from all steps of the

Figure 4.3 – Transient heat solution for the variation in
temperature with time.
To use the nodal temperatures calculated by the heat
solvers in a structural analysis, with the model open choose
Attribute/Node and select Temperature. Then in the
dialog box, shown in Figure 4.4, click Import:
transient heat results file.
Note: If the model does not contain enough load cases to
assign all of the steps, new load cases are automatically
created during the import.

www.strand7.com
G+D Computing
Once the temperature results have been imported back into If it is necessary to determine the development of the
the Strand7 model, as load case dependent nodal thermal stresses with time, the transient heat results
temperature attributes, the finite element model needs to be obtained earlier can be used in conjunction with the
set up for a structural analysis. This includes all the usual nonlinear transient dynamic solver. Choose
considerations of defining the correct structural material Solver/Nonlinear Transient Dynamic to bring up the
properties (i.e. Young’s solver dialog box. Then click Temperature. This will
Did you know? modulus E, Poisson’s bring up the dialog box shown in Figure 4.8, where the
ratio v etc.), applying user has the option to input nodal temperatures via load
Brick Cutting Plane Animations case data or by choosing the results of a transient heat
restraints and
Viewing internal stresses in a brick specifying the thermal solution.
can often be difficult and so the expansion coefficient
cutting plane option is used to In the current case study the nodal temperature input is
for each property set
view the stresses through a slice. from the Transient Heat File. The time steps for the
You can then create an animation (note that these are
transient dynamic analysis can be the same or different to
in the usual manner, aspects of the structural
those of the transient heat analysis. If they are different,
(Results/Create Animation), analysis and are not
then Strand7 will interpolate the transient heat steps to
which will create an animation of necessary for the initial
determine the values to use in the stress analysis.
results for this slice. steady-state or transient
But what about if you want to see heat analysis).
this slice progress through the
brick structure for a single load The thermal strains εo
step? Choose to create an resulting from the
animation and set Move Cutting applied node
Plane. When animated the slice temperature attributes
will progress through the structure are calculated from the
for that load step, between the following equation:
extremities of the model.
ε o = α ( T − Tref )
where α is the
Figure 4.8 – Temperature input for the nonlinear transient
coefficient of thermal
dynamic solver via transient heat file.
expansion, T is the
temperature of the Once the reference temperature has been specified, the
body and TR is the variation of the thermal stress can be calculated by the
reference (or strain- Strand7 nonlinear transient dynamic solver. The results
free) temperature. obtained for point B are shown in Figure 4.9:

It is therefore also
important to properly
specify the correct
reference temperatures
for the model. This is
found under Global/
Load and Freedom Cases and is load case dependent. The
manner in which the different Strand7 solvers determine
and use this reference temperature is discussed in more
detail later.
For the embedded composite material model, shown in
Figure 4.1, the thermal stress results obtained from the Figure 4.9 – Nonlinear transient solution for variation of stress
with time.
Strand7 linear static solver based on the steady-state
temperature field are shown in Figure 4.7. It can be seen that the thermal stress at point B relaxes with
time to its steady-state value of 62 MPa.
Note: With this method the Reference Temperature
specified in each load case is ignored. Instead it is
enforced in the dialog shown in Figure 4.8 since there
could be multiple load cases with different values of
Reference Temperature.
2. Modelling material properties that change with
Point B temperature.
Material properties such as Young’s modulus and the
Figure 4.7 – Thermal stress field based on steady state
thermal expansion coefficient can be a function of material
temperatures.
temperature. In Strand7, such dependency is described
It can be seen that the highest value of approximately through the use of a Factor vs. Temperature table (choose
62MPa for the maximum principal stress is predicted at Tables/ Factor vs. Temperature to define these).
point B.
8

www.strand7.com
G+D Computing
As an example, if the stiffness (Young’s Modulus) of a
material decreases as the temperature increases, the table EO
shown in Figure 4.10 could be used.
A nominal value of Young’s Modulus is defined via the
Structural tab in the Element Property dialog box and the
table of Factor vs. Temperature type is assigned to the
property via the Tables tab (see Figure 4.11 and Figure
4.12).
At a given temperature the Young’s Modulus is calculated
by multiplying the nominal value by the table factor. In
other words, the table defines the ratio between the
nominal value of modulus and the value of a specific
temperature. It is not a table of Modulus as a function of
temperature.

Figure 4.11 – Specifying nominal value EO via structural tab.

Figure 4.10 – Factor vs. Temperature Table for Young’s Modulus.

Did you know?

Auto-Assign restraints
When creating a rigid link array there is
no need to go through the process of
creating rigid links one at a time.
Instead you can make use of the auto-
assign tool in Strand7. Figure 4.12 – Assigning Factor vs. Temperature Tables to
properties.
Select all the nodes you wish to be
connected via rigid links then choose
Tools/Auto Assign/Restraints. 3. How each Strand7 solver uses the applied load case
Select Rigid Connections as the temperature data.
Type and click Apply. This will
connect all the nodes via rigid links to Due to the fundamental differences in their underlying
an automatically generated master theories, the manner in which load case temperature data is
node, which will lie at the average of used varies quite markedly between each of the Strand7
the selected nodes. Alternatively you solvers. Table 4.1 details how the static and transient
can select your own master node for dynamic solvers make use of load case temperature data.
the rigid links to radiate from.
Further to the details given in Table 4.1, it is important to
be aware of the treatment of applied temperature load case
data by the other Strand7 solvers (not listed in Table 4.1).
For example, while the linear buckling solver does not
support temperature dependence options directly, any
temperature dependence used in the initial conditions file
(either from a linear or nonlinear static analysis) will be
effectively used in the buckling analysis.
This tool is particularly useful for
models that have bolted connections or
areas where a point load is to be
applied over an area.

www.strand7.com
G+D Computing
Solver Temp. Dependent Material Properties
Linear Static In the solver dialog box, the user can
specify any one load case which
contains nodal temperatures (Tmat) to be
used to calculate temperature
dependent material parameters (eg. E0
and α0). This means that all load cases
use the same value of E.
Nonlinear At each load step a temperature
Static distribution is calculated using the load
factors λ l and temperature distribution
Tl of included load cases according to:
1. A combined temperature change is
derived from the linear combination of
temperature change in all of the
included thermal load cases:
∑ λ (T )
NLC
∆T = l l − Tl ref
l =1
2. The combined temperature value is
defined as the combined temperature
change plus the reference temperature:
Tcomb = ∆T + T ref
Linear In the solver dialog box, the user can
Transient specify a load case that contains nodal
Dynamic
temperatures (Tmat) to be used to
calculate temperature dependent
material parameters. All material
models are assumed to be linear and
material parameter values will remain
unchanged during the solution. The
temperature distribution at t = 0 will be
used to determine temperature
dependent material parameters.
Nonlinear 1 - Transient Heat File: material
Transient parameters are determined based on the
Dynamic temperatures at any time t from the
transient heat results.
Note the two
2 - Load Case Data: The material
options for
defining the temperature distribution at a time
material instance t is calculated by first
temperatures: determining the combined temperature
change ∆T from all included load cases l
1 - Transient using equation [5]. The current
Heat File combined material temperatures Tcomb
are then calculated from equation [3]
or and used to calculate temperature
dependent material parameters.
2 - Load Case
Data
Table 4.1 – Details for the use of applied load case temperature data by the static and transient dynamic solvers.
1 – If the thermal strain ε calculation includes the effects of temperature dependence of the thermal expansion coefficient, the following integral is used
where α0 is the nominal expansion coefficient and fα represents the Factor vs. Temperature table
2 – When the reference temperatures of different load cases are not the same, a warning message is written to the log file.
www.strand7.com
Thermal Loads
For each load case (l), the thermal
loading or strain is determined
from:
(ε 0 )l (
= α 0 f α (Tmat ) Tl − Tl ref [1] )
Each load case uses its own
temperature distribution to
determine the temperature loading
Tl.
The temperature loading1 is
determined from the temperature
change calculated in equation [2]
and based on the material
parameters evaluated at the actual
combined temperatures given by
equation [3].
[2]
[3]
The combined temperature change
at a time instance t, is calculated as:
∑ f (t ) (T (t ) − T )
NLC
∆T (t ) = l l
l =1
where fl represents the Factor vs.
Time table assigned to load case l
evaluated at time t.
1 - Transient Heat File: the
temperatures used to determine the
thermal loads1 at any time t are
obtained directly from the transient
heat results.
2 - Load Case Data: The
temperature loading1 is determined
from the combined temperature
change calculated in equation [5]
and based on the material
parameters evaluated at the
combined temperature values given
by equation [3].
Reference Temperature
Each load case (l) uses its own
reference temperature (Tl ref )
specified in Global/Load and Freedom
cases data.
Each load case (l) uses its own
reference temperature (Tl ref )
specified in Global/Load and Freedom
cases data to determine the combined
temperature change using equation
[2].
However, the reference temperature
used in equation [3] to determine the
combined temperature value is the
smallest reference temperature of the
included thermal load cases2:
T ref = min
l =1, 2,K NLC
(T )
l
ref
[4]
where NLC is the number of included
ref
thermal load cases, and Tl is the
reference temperature for load case l.
Each load case (l) uses its own
reference temperature (Tl ref )
ref specified in Global/Load and Freedom
l [5]
cases data to determine the combined
temperature change using equation
[5].
1 - Transient Heat File: the reference
temperature is specified directly by
the user.
2 - Load Case Data: Each load case (l)
uses its own reference temperature
(Tl ref ) specified in Global/Load and
Freedom cases data to determine the
combined temperature change using
equation [5]. However, the reference
temperature used in equation [3] to
determine the combined temperature
value is the smallest reference
temperature of included thermal load
cases2 (i.e. determined using equation
[4] ).
ε = ∫ α 0 f α (T ) dT ,
10
G+D Computing
Similarly, if an initial condition is used for a natural
frequency analysis, the temperature dependence in the
initial file (either from a linear or nonlinear static analysis) is
effectively used. Otherwise (i.e. if an initial condition is not
used), the temperature distribution in any load case may be
selected to determine temperature dependent material
parameters in the same way as the linear static solver.
Did you know?
Ctrl-Click Retrieve
Have you ever found yourself wanting to assign a
restraint to a node similar to an adjacent node, or
perhaps assign a common value of normal pressure to a
brick face?
Ctrl-Click
Within Strand7 is a simple procedure that can retrieve
this information, ready for it to be assigned elsewhere.
Open the required attribute dialog and then ctrl-click the
node or element. The dialog will be automatically filled
out with all the attribute values of the clicked entity, which
can then be applied to other elements.
For example, open the node restraint dialog and ctrl-click
a node with translational restraints. The node restraint
dialog box will change to set the translational X, Y and Z
boxes. This procedure can be used for all element
attributes.
While the harmonic response and spectral response solvers
do not directly consider material temperature dependence
these effects can be included via the results of a previous
natural frequency analysis
Various material parameters required for the heat solvers
(such as conductivity, convection coefficients and radiation
coefficients) can also be temperature dependent - their
treatment follows the description given in Section 2. In
nonlinear steady heat analysis, the solver iterates on
temperature, continually updating material parameters
based on the temperature at the current iteration until
convergence is achieved. For transient heat analysis, the
material parameters are updated, according to the
temperature distribution at the commencement of the time
step based on the matrix update option selected in the
solver dialog.
4. Discussion
When planning a quasi-coupled thermal stress analysis in
Strand7, it is important to remember that the mesh
requirements for obtaining accurate stress results might
www.strand7.com
differ to those required for calculating the temperature
distribution. In the heat solvers, temperature is the
fundamental quantity that is calculated at the nodes.
Therefore, depending on the nature of the problem, a
relatively coarse mesh might suffice for determining an
accurate temperature distribution. However, when
undertaking the stress analysis, a much more refined mesh
may be needed to properly determine the resulting thermal
strain and stress fields.
Also, when performing a thermal stress analysis in Strand7 it
is very important to be aware of how Strand7 uses different
applied load case temperature data (see Section 2), setting
the appropriate solver dialog box accordingly. As in any
analysis, check the log file for any warnings before post-
processing
User Defined Contour Files
T he Strand7 API was released mid 2003 and received a
major update earlier this year with Strand7 Release 2.3
incorporating over 100 new API functions.
A relatively unknown feature of Strand7 is the ability to
contour a model according to a user defined text file. The
text file has to conform to a specific format that is described
in detail in the Strand7 Online Help (for nodal contours the
text file consists of simply a title followed by a list of node
number – contour value pairs). Creating these text files for
big models can be a repetitive and tedious task, a task
perfectly suited for the Strand7 API.
With the power of the Strand7 API, user defined contour
files can be created with ease. This article shows you how
to produce your own user defined contour file.
Introduction
A simple fatigue analysis will be carried out on a structural
component undergoing cyclic loading. The material S-N
curve is shown in Figure 5.1.
Figure 5.1 – Typical S-N Curve.
Two contour files will be produced – endurance stress and
fatigue life – and contoured within the Strand7 graphical
environment.
The Model
The component is fixed at the base as shown in Figure 5.2.
Two load cases, unequal in magnitude, representing the
11
G+D Computing
limits of the operating cycle are specified – a compressive - Find the averaged alternating and mean stresses at
load case and a tensile load case. each node.
- Use the ultimate material strength and the
Goodman equation:

S alternating
S endurance =
 S 
1 − mean 
 S 
 ultimate 

to evaluate the endurance stress at each node.

- Use the endurance stress and find the
corresponding fatigue life at each node from the
S-N curve.
- Export two Strand7 Node point contour text files:
- endurance stress.txt
- fatigue life.txt
Contour with User Defined Text Files
The two exported text files can be used to contour the
model within Strand7.
- Choose View/Entity Display
Figure 5.2 – The model with the tensile load case active.
- Under Contour Type select User Defined (file) as
Initial Processing shown in Figure 5.4.
A linear load case combination, to derive the principal stress
range, is specified in Strand7 as shown in Figure 5.3.

Figure 5.3 – Linear Load Case Combinations dialog.

The two load cases are combined according to their
fundamental stress components – that is, XX, YY, ZZ, XY,
YZ and ZX – to create the combination load case. For this Figure 5.4 – Entity Display.
analysis, the principal stresses from the combination load
- Select the endurance stress.txt file.
case are used to predict fatigue life.
Ensure that for Files of type, Plate Node Contour
The Coding File (*.txt) is selected as shown in Figure 5.5.
Click Open.
The code used to calculate the endurance stress and fatigue
life has been written with Microsoft Excel VBA XP. This
VBA code is available for News.St7 readers to download
and run on their own computer from our website
www.strand7.com/News.st7.htm.
The pseudo code is:
- Initialise the Strand7 API.
- Open the Strand7 model.
- Run the Linear Static solver.
- Open the Linear Static results and generate the
linear load case combination.
- Extract the principal stresses and evaluate the
alternating stress at each node for each plate.
Where Salternating = (σ11 - σ22) / 2 Figure 5.5 – Selecting endurance stress.txt file.
- Extract the principal stresses and evaluate the mean The model will now be contoured according to endurance
stress at each node for each plate. stress as shown in Figure 5.6.
Where Smean = (σ11 + σ22) / 2
12

www.strand7.com
G+D Computing

Benchmark

I n this edition we will look at a benchmark problem

dealing with seismic analysis.
There are three common methods accepted by design codes
around the world for performing seismic analysis of
structural systems. These are:
● Equivalent Static Analysis
● Spectral Analysis
● Transient Dynamic Analysis
The equivalent static analysis is a straightforward method to
evaluate horizontal force distributions in simple structures.
This method has been
common practice in the Did you know?
past and remains a good,
quick tool for sizing Run Time Improvements
members in a conceptual There are a few tips that you
design. Spectral and should keep in mind when
transient dynamic analyses trying to improve the run time
Figure 5.6 – Endurance Stress contour. of your Strand7 solve.
are more accurate methods
Similarly, the fatigue life contour can be displayed as shown of analysis and, these days, Firstly store your temporary
in Figure 5.7. files locally, not across a
much more straightforward
network. Choose
with the tools available in File/Preferences, to change
Strand7. Throughout this the location of your temp
discussion we implicitly directory.
refer to clauses of Also if it is a very large model
AS1170.4 – 1993. try saving the results files to
your local drive while solving,
1. Equivalent Static reducing the time taken to
For seismic analysis, the write across the network.
dynamic load due to the However while all this will
earthquake can be show some small performance
described as quasi-static improvements in solve time, for
and the structural response large improvements, the
then can be determined by Sparse Solver should be top of
the agenda.
running a linear static
analysis.
The implementation in Strand7 is a general one, which
therefore makes it applicable to virtually any national design
code (e.g. AS1170.4 – Australian Standard, Minimum
design loads on Structures, Part 4: Earthquake loads).
An equivalent static analysis can be applied to buildings
with the following attributes:
● Structural regularity (plan and elevation)
Figure 5.7 – Fatigue life contour.
Conclusion ● First Natural Period smaller than 2.0 seconds.
This article has shown that with the Strand7 API, users can Within Strand7 a seismic load case can be created and
create their own User Defined results, which can be solved by entering the appropriate values from design
subsequently contoured within Strand7 codes.
2. Spectral
This analysis is a more general approach and allows a much
larger number of structural configurations to be dealt with.
However, a few assumptions are still needed:
● Linear material and geometry
● Small structural damping

13

www.strand7.com
G+D Computing
● Spectral results are based on a natural frequency
analysis. These frequencies are linear and
undamped.
3. Transient
This analysis is the most realistic as it considers the structure
subject to a time acceleration history. The previous
hypothesis regarding linear behaviour and small damping
are no longer required. You can also consider material and
geometric nonlinearities. Since the model is solved at every
time step the complete solution time can be an important
consideration.
Benchmark Example
The model (Figure 6.1) used is a simple concrete four-storey
building. Its dimensions are 12m x 6m with a floor height Figure 6.2 – Modified Athens earthquake history.
of 3m. This acceleration time history can be converted to a
response spectrum using Strand7’s inbuilt Convert to
Response Spectrum tool in the table dialog. An
AS1170.4 response spectrum curve is also used for the
comparison (an appropriate acceleration coefficient for
Athens has been chosen as part of the factors, see Table 6.1,
to allow for direct comparison between the two curves).
These are given in Figure 6.3.

Figure 6.1 – Four-storey building model.

An appropriate earthquake time history and response
spectrum has been chosen such that the results given by
each method are
directly comparable.
Did you know? The reference
Strand7 Viewer acceleration time
If you have clients that don’t have a history chosen is a
copy of Strand7 and you want to be modified Athens
able to display your Strand7 model Earthquake shown in Figure 6.3 – Response Spectrum.
and results to them in their office the Figure 6.2. For this example the following parameters have been
Strand7 Viewer can be used.
chosen from AS1170.4
You can download a free copy of the
Strand7 Viewer from our website Site Factor (S) 1.25
www.strand7.com.
Importance Factor (I) 1.00
This allows you to interrogate a
model file and results in the same Acceleration Coefficient (a) 0.30
way as you would in the full version
of Strand7. Structural Response Factor (Rf) 4.00

Table 6.1 – AS 1170.4 Factors.

The natural frequencies of the building are:
Mode 1: 0.659 Hz
Mode 2: 0.745 Hz
These first two natural frequencies are free of torsional
effects and together they have a total mass participation
greater than 90%. Hence in this case all three approaches
can be applied.

14

www.strand7.com
G+D Computing
Equivalent Static Analysis Results
Create a new seismic load case in the Global/Load and Figures 6.4 and 6.5 illustrate the results for the different
Freedom Case dialog. The direction of the earthquake is in approaches.
the X direction (horizontal).
The parameters required are based on the following:
V = βG

where,
V = base shear.
β = a factor, which may be related to the importance of the
structure, the site factor, etc.
G = total gravity load = ag (MT).
where,
ag = the acceleration of gravity. Figure 6.5 – Comparison of Max DX.

MT = (αM + ψQ ) = the effective mass of the structure,

which is a combination of the structural (M) and the non-
structural (Q) masses.
α and ψ = non-dimensional combination factors, which are
usually greater than zero and never negative.
And finally,
k = an exponent related to the period of the structure
h0 = the vertical coordinate of the base.
For more details on exactly how these factors are applied in
an analysis see the Strand7 Online Help.
Figure 6.6 – Comparison of Bending Moment.
Spectral Analysis (Calculated Spectrum)
Click Direction Vectors in the Spectral Response Solver
Conclusion
dialog. Select the calculated response spectrum table. This
table in is units of m/s2 and so a factor of 1 is applied in the As the Calculated spectrum is derived directly from the time
X direction. Damping was included at 5% when the curve response graph we expect these results to be virtually
was created so damping should be set to None in the solver identical to the linear transient results. This is shown in
dialog. figures 6.5 and 6.6. The Australian Code spectrum includes
design factors, i.e. site and importance factors, and will
Spectral Analysis (AS1170.4 Response Spectrum)
therefore give a more conservative result. The equivalent
For the Direction Vectors in the Spectral Solver dialog for static analysis is more conservative again.
this case select the AS1170.4 response spectrum curve.
This curve is in units of g’s so the factor needs to include Exclusively to readers of News.St7: You can download this
gravity. The code also requires the value to be factored benchmark model from the News.St7 readers’ page of our
according to the following: website:www.strand7.com/news.st7.htm

I
Factor = 9.81 m / s 2 × a Announcements - Theoretical Manual
Rf

Apply 5% Rayleigh Damping.

In this example a factor of 0.736 is assigned in the X
W e are pleased to announce the release of the Strand7
direction. Theoretical Manual.

Linear Transient
Click Base Acceleration in the Linear Transient solver
dialog and select the time history table that has already been
created for the building. The table is in units of m/s2 so a
factor of 1.0 should be applied in the X direction. 5%
Rayleigh damping should be applied.
This manual will provide users with the theory beyond the
Strand7 interface.
Already some supported users have discovered the benefits
of this manual having been sent small sections as part of
their support query.
Please contact us for more details on how to obtain a copy
of this manual

15

www.strand7.com
G+D Computing

Training Training Calendar

I n the last issue of News.st7 we discussed the availability of tailored

2004
training courses. Following on from this we have recently completed a very
successful training course in South Africa. For a week in early September a 12-15 October Introducing Strand7
number of engineers from a variety of fields took part in an intensive Strand7 19-20 October Nonlinear Analysis
course, ranging from introductory lessons to the more advanced nonlinear 21-22 October Dynamic Analysis
and dynamics fields. Thanks to course organiser (and Strand7 distributor in 26-27 October Structural Engineering
South Africa) Allyson Lawless and Hatch South Africa for providing the 28 October Automeshing with Strand7
training venue. To find out the thoughts of the course attendees, visit 29 October Heat Transfer Analysis
www.strand7.com/SATraining.htm.
If you are interested in tailored onsite training courses then please
contact us for more information.
Strand7 User Profile In October we are running our full series of courses ranging from
Peter Seid Introductory through to Structural Analysis held in our offices in
Sydney. These are the final courses for this year. But our
APS Plastics Pty Ltd calendar for next year’s courses will be released soon.
You can find full details of the content of the courses, prices and
Q - How would you describe your booking form at: www.strand7.com/training.htm
business?
Plastics Designers and Engineers. Exhibitions
Q - What types of jobs do you use Strand7
for?
Anything plastic basically. This includes new
R ecently Strand7 exhibited at the China International Steel
inventions and concepts in the industrial and Construction Expo/Conference with great success. We were
automotive industries, consumer goods, delighted with the opportunity to meet all those who attended.
medical devices, etc. Recent projects have We plan to exhibit at the following upcoming conferences:
included theatre seats, car grills, a pull-along
cooler and syringe. 15-17 December 2004 FHWA Steel Bridge San Antonio,
2004 Conference Texas, USA
Q - How does Strand7 help you in what you
design and analyse? 10-12 January Pacific Design and Anaheim,
It helps to identify a new product’s weaknesses 2005 Manufacturing Show California, USA
and enables us to iterate to a satisfactory
design. If you get the chance then we would be delighted to meet you
and discuss Strand7 with you. You will have the opportunity to
Q - What features of Strand7 do you find meet representatives from Strand7 and from our US distributor.
most useful?
We will be keeping you updated on the latest developments in
The automesher has reduced modelling time future editions of News.St7
as well as the sparse solver. The animation of
results with amplified deflections also helps
visualise behaviour. Also ranging the stress
levels helps to identify “hot-spots” above the
materials yield limit leading to mesh refinement
Strand7 Head Office
or redesigning.
G+D Computing Pty Ltd
Contact Details Suite 1, Level 7
Unit 41/756 Burwood Highway 541 Kent Street
Ferntree Gully Sydney NSW 2000
3156 VIC Australia AUSTRALIA
03 9752 2144
Tel +61 2 9264 2977
www.apsplastics.com.au
Fax +61 2 9264 2066
aps@apsplastics.com.au
Website: www.strand7.com
Email: info@strand7.com

16

www.strand7.com
Strand7 Pty Limited

News.St7
Newsletter for Strand7 and Straus7 users
Strand7 is marketed as Straus7 in continental Europe

Issue 4, March 2005

Strand7 Moves Into New Year!

Happy New Year to all our Strand7 users as Strand7 Software heads into an
exciting year of developments. Some of you may have noticed that G+D In this issue…
Computing no longer features at the top of this newsletter. In the New Year we
• Feature Article 1
rebranded to Strand7 Pty Ltd and moved to newer and larger premises. Details
and pictures of our new offices can be found at the end of this newsletter. • Plate Alignment 3
The name change is designed to more accurately reflect our principal focus and • API in the Field 5
intended directions, namely the development and support of the Strand7 software.
To read the full press release visit www.strand7.com/PR-Jan05.htm. • Strand7 Functions 6

This year will see another major version of Strand7, Release 2.4, training courses • Modelling Tip 10
interstate and, of course, more News.St7 issues packed with helpful hints and • Training 14
feature stories.
• Exhibitions 14
In this Issue of News.St7, the first for 2005, we showcase how Strand7 is being
used to design compact range reflectors by Beaufort Composites Technologies, • Announcements 15
Inc, a client in the USA. • User Profile 16
It also includes articles on the Grade Plates and Bricks tool, plate orientation and a
practical discussion on how to determine the mechanical properties of plates in structures with repeating patterns.

If you have any feedback or suggestions regarding the content of News.st7 or would like to feature any of your projects then please email
news.st7@strand7.com. If you would like to automatically receive your copy of News.st7 directly by email simply send a blank email to
newsletter-subscribe@strand7.com.

Strand7 for the Design and Analysis of the The compact range’s reflector is the system’s most
Backup Structure of a Compact Range Reflector prominent feature. A feed antenna, located at the
parabolic surface focal point and pointed at the reflector,

A
radiates a controlled spherical wave, which in turn
compact range allows for accurate, indoor antenna reflects from the parabolic surface creating the required
planar wave. The test zone of the reflector accurately
and radar cross section (RCS) electromagnetic
simulates far-field conditions for articles under test.
measurements. During antenna measurements, an
antenna’s characteristics are measured and evaluated such Test/Measurement
as the antenna’s electromagnetic pattern, gain, zone
polarization and mechanical alignment. During RCS
measurements, a target’s radar “signature” is measured Reflector
and evaluated. RCS measurements are of increasing Planar Wave
importance to many industries including the military
where they are used in an effort to reduce the RCS of
airplanes, ships, etc.
Compact range equipment allows measurements, Feed
normally only possible with a large distance between the Spherical Wave Front
microwave source and the antenna or target under test, to Fig 1.1 – Compact Range Diagram.
be realized in a small space.

www.strand7.com
Making Finite Element Analysis Easier

Fig 1.3 – Strand7 model of the reflector’s backup structure.

Fig 1.2 – Compact Range Set Up (Courtesy MI Technologies).

A compact range reflector is composed of a paraboloidal

surface supported by a backup structure. The reflector
surface is machined to provide superior surface accuracy.
Traditionally, the backup structures are fabricated of steel
components. Reflector sizes can vary widely depending
on the size and frequency of operation of the articles
under test. A serrated or rolled edge treatment is used to
minimize electromagnetic edge diffractions and
scattering.
MI Technologies, Suwanee, GA, is the world leader in
compact range technology. MI Technologies has Fig 1.4 – Picture of the completed backup structure.
manufactured and installed over 300 units ranging in size
from 2 feet x 2 feet to exceeding 55 feet x 45 feet. A very important aspect of one of the Strand7 tools,
Working with Beaufort Composite Technologies, Inc. which does not often hold a critical role for most of the
(BCT), MI Technologies has developed a new innovative users, is the geometric accuracy of the automesher. The
“egg crate” design fabricated of carbon fiber material. quality of the reflector is highly dependent on how true
the manufactured surface, affected by the loading
Reflector’s Design conditions (self-weight and thermal stress), is to the
Several key design requirements were defined for the theoretical paraboloidal shape.
reflector structure: stiffness, strength, lightweight, Any inaccuracies inherent to the modeling process,
modular, improved thermal properties (over the starting with the creation of the mesh, participate in the
traditional steel structures) and cost effectiveness. Finding evaluation of the accuracy of the final structure. It is
the optimum backup structure shape to maintain the absolutely necessary to start with a mesh that has as small
extremely tight surface tolerance while minimizing a deviation from the mathematical surface as possible.
manufacturing costs was the big challenge of this project.
A compromise was found by establishing a close The backup structure is fabricated from hundreds of
approximation to the paraboloidal shape to minimize the laminated composite parts. The laminated composite
amount of machinable material (weight with no structural module, combined with a range of tools available in
contribution) for minimum surface weight and, utilizing Strand7 made the modeling of the complex structure an
molded carbon fiber components to form the lightweight, easy and efficient process. A series of linear static
stiff structure. analyses were run to achieve optimization of the laminate
structures of these parts. Optimization was in respect to
minimum displacement of the surface and minimum
Strand7 Modeling weight.
The CAD geometry of the reflector (backup structure and
surface along with the hardware needed to attach the
serrations to the body) was originally created by BCT in
Rhinoceros and then imported into Strand7. The backup
structure was modeled using plate elements; the
reflector’s surface was modeled using brick elements.

www.strand7.com
Making Finite Element Analysis Easier
Fig 1.5 – Displacement results for the static analysis, studying the effects
of gravity on the reflector.
A thermal analysis was performed in order to understand
and minimize the effects of temperature on the reflector
and, more specifically at the interface between the surface
material and the backup structure. Prior to using carbon
materials, the reflectors had to be located in a highly
controlled environment with minimum temperature
variations.
Finally, a natural frequency and response spectrum
analysis were completed to study the behavior of the
structure under loadings associated with seismic events.
Fig 1.6 – Picture of the finished reflector (Courtesy MI Technologies).
Strand7 allowed for the optimization of the design of the
backup structure, now made out of carbon fiber,
maximizing the stiffness of the structure while minimizing
weight and material cost. The design using carbon fiber
(versus earlier steel designs) results in an unprecedented
stiffness of the structure while decreasing its weight by a
factor of approximately five. The use of carbon fibers also
minimizes the impact of temperature variation on the
performance of the reflector.
This article was prepared by the Strand7 distributor in the US: Beaufort
Analysis Inc (www.beaufort-analysis.com) with the assistance of
Beaufort Composite Technologies and with the permission of MI
Technologies
www.strand7.com
Plate Alignment: Part 1
Plate Orientation
A n important, yet often overlooked, part of modeling in
FEA is the alignment of the orientation of plate elements.
This article briefly discusses the impact misaligned plates
can have on results and also ways to identify and fix this.
Overview
Plate orientation refers to the positive and negative
surfaces of plate elements, which in turn refers to the
positive and negative z directions of the plate local
coordinate system.
y
+z
-z x
Figure 2.1 – Plate axes diagram.
Figure 2.1 shows the plate local axes x and y, which also
defines the local z axis direction. +z is the top surface of
the plate and –z is therefore the bottom surface of the
plate.
What is the Default Orientation?
Plate orientation is initially defined when a plate element
is first created. It is dependent on whether the plate is
connected in a clockwise or anticlockwise fashion. To
create a model with consistent plate orientation an
attempt should be made, in the initial construction of the
model, to connect plate elements in a consistent
direction.
Identifying Plate Orientation
There is a simple display option in Strand7 that allows for
the easy identification of plate orientation in a model.
Choose View/Entity Display and click the Plate tab .
Under Draw As select Orientation. Plates will be drawn
in one of two colors, which represent the positive or
negative z surface. The colors can be set by the user in
the Entity Display dialog, however the default displays +z
surfaces in cyan and –z surfaces in red. Figure 2.2 shows
that there is an inconsistent alignment of plates in this
model as we have a small section with the +z surface
facing upwards.
Making Finite Element Analysis Easier

the first ply lies on the -z surface and the last ply lies
on the +z surface. This becomes very important
when the laminate is non-symmetric and a specific
ply should be the outer layer.

How can the Orientation be Changed?

Two tools exist that allow the user to change the
orientation of plate elements. One is useful for aligning
individually selected plates while the other provides a
more global functionality.

Figure 2.2 – Display of plate orientation, (cyan = +z, red = -z). Method 1 – Flip Elements.
Choose Tools/Align/Flip Elements. Select those plates
Why is this important? that require alignment of their orientations and click
Apply. Those elements will have their z direction
There are four main reasons why consistent and known
swapped.
orientation of plates is important.
1. Normal face pressures are applied based on the
orientation of the plate element. A positive pressure
acts in the direction of the positive z axis and
negative pressure acts in the direction of the negative
z axis. If the orientation of adjoining plate elements
is not consistent this can lead to normal pressures
applied in two directions in a model, when only one
direction is required.
Figure 2.3 – Flip Elements tool.
2. The direction of compression only face supports is
Method 2 – Plate Normals
determined by the plate orientation.
Choose Tools/Align/Plate Normals. Select All Plate
3. Surface stresses in plate elements can be shown for
positive and negative z surfaces as well as mid-plane. elements. Using the hotpointer select a Master
If you wish to display the outer surface stress of a Element. Click Apply. All plate elements that were
group of plate elements, you may not know if the selected will be aligned with the same orientation as the
outer surface is the positive or negative surface. By master element.
displaying the plates in Orientation mode, you can
quickly identify the positive and negative surfaces.
Consistency of plate orientation becomes important
in this situation if Did you know?
plates are in
surface view and Edit-Delete-Attributes
the +z or –z If you find the need to delete
results are all attributes in a model
contoured. If the without deleting the elements
orientation of the themselves there is a simple
tool to do this.
plates is not
consistent, and Choose Edit/Delete and Figure 2.4 – Align Plate Normals tool.
select Attributes Only. You
bending is
can then choose whether to Note: To use the Align Plate Normals tool the selected
present, then do so for all cases or simply a plate elements need to be connected together and the
surface contours single case. Master Element should be one connected to the selected
will not be
set. Also note that this tool will align elements in general
continuous.
3D and 2D models, e.g. a sphere modeled with plate
4. Layup of elements. It is particularly useful for folded plates.
laminate
properties is Conclusion
based on the +z The orientation of plate elements should not be
and –z surface of overlooked when creating a model in Strand7. In any QA
plate elements. checks of a model this should form one of the initial
When laying up checks for model accuracy.
a laminate it
should be Look for Plate Alignment Part 2: Plate Axes in the next
remembered that Issue of News.St7
4

www.strand7.com
Making Finite Element Analysis Easier
structure has the following form for the relationships
API in the Field between stress resultants and (generalized) strains.

 N xx   C xx C xy C xg Bxx Bxy Bxg   ε xx 

Mechanical Properties and Strength of Plates  N  C
and Shells C yy C yg Bxy Byy B yg   ε yy 
 yy   xy  
 N xy  C xg C yg C gg Bxg Bxg Bgg   ε xy 
P lates and shells with repeating patterns of stiffeners

 M xx
=
  Bxx Bxy Bxg Dxx Dxy
 
Dxg  κ xx 
and/or perforations are in very common usage. For a few  M yy   Bxy Byy B yg Dxy D yy D yg  κ yy 
common arrangements, say blade stiffeners, there are     
 M xy   Bxg Byg Bgg Dxg D yg Dgg  κ xy 
closed form expressions for the membrane and bending
stiffness. Such closed form solutions are rare, difficult to
The relationships between the generalized strains and the
implement and often ignore three important aspects of
coordinate displacements of the plate are the usual ones
structural mechanics. Firstly, in modeling such structural
for small displacement, small strain theory and the
forms with FEA, the model is nearly always built in the
relationship between stress resultants and internal plate
plane of the base plate, which is not the neutral plane.
stresses are also the usual ones from traditional plate
Secondly, if the pitches and geometry of stiffeners are
theory. Considering an FEA model involving a single
different, then the concept of a neutral plane does not
repeating cell of a large plate, this cell can have six
apply. Finally the coupling between membrane and
separate unit strain vectors applied, such as the one in the
bending is an important aspect of the structural behavior.
equation below for a unit strain in the x direction only.
Having undertaken FEA on such complex stiffened shell Subsequent recovery of the stress resultants required to
structures it is important to calculate accurate stresses generate the unit strain provides the required coefficients
around the detail of any perforations and/or stiffeners, in the cell’s constitutive relationship, as shown also in the
rather than average bulk values, as it is these stress equation below.
concentrations that will lead to failure.  ε xx   1  N xx   C xx C xy C xg Bxx Bxy Bxg  1  C xx 
ε     N  C B yg   0  C xy 
This brief article describes plate homogenization and  yy   0  yy   xy
C yy C yg Bxy B yy
   
inverse homogenization that can be implemented using a  ε xy   0  N xy  C xg C yg Cgg Bxg Bxg Bgg   0 C xg 
 =  ⇒  =  = 
Strand7 API procedure to provide the combined κ
    0  M xx   Bxx Bxy Bxg Dxx Dxy Dxg   0  Bxx 
xx  
membrane/bending stiffness matrix for any architecture of κ yy   0  M yy   Bxy B yy B yg Dxy D yy D yg   0  Bxy 
          
an anisotropic plate such as the one shown in Figures 3.1 κ xy   0  M xy   Bxg B yg Bgg Dxg D yg Dgg   0  Bxg 
and 3.2.
Thus the first column of the stiffness matrix is obtained.
Repeating the process five more times, with the unit strain
moving down one row each time, produces all the other
coefficients in the plate stiffness matrix. To achieve the
stretching in the x unit strain vector given in the above
equation, the following restraints are applied to an FEA
model of the repeating cell of dimensions (Xmax,Ymax).
● Set all translations and rotations free.
● At a single location on an edge, set DZ = 0 (to
prevent free body motion in Z)
● At X = 0 = LH edge, set DX = θX = θY = 0
Fig 3.1 – Part of a plate structure with simple repeating detail.
● At X = Xmax = RH edge, set DX = Xmax,
θX = θY = 0
● At Y = 0 = Bottom edge, set DY = θX = θY = 0
● At Y = Ymax = Top edge, set DY = θX = θY = 0
● Where edges meet there is an OR Boolean
operation of the above fixed DoF
The first bending unit vector is {0 0 0 1 0 0} and this is
achieved by the following restraints in a Strand7 FEA
model
Fig 3.2 - A cell of the repetitious structure.
● Set all translations and rotations free.
How does it work; well, we need a little bit of mechanics. ● At X = 0 = LH edge, set DX = DZ = θY = 0
The standard Kirchoff plate theory for an anisotropic plate
5

www.strand7.com
Making Finite Element Analysis Easier
● At X = Xmax = RH edge, set θY = Xmax, 1.91E7 8.62E6 0 0 0 0 
DX = θY × z (z is the out of plane coordinate) 8.62E6 1.91E7 0 0 0 0 

● At Y = 0 = Bottom edge, set DY = θX = 0  0 0 6.89E6 0 0 0 
 
● At Y = Ymax = Top edge, set DY = θX = 0  0 0 0 3.53E2 8.86E1 0 
 0 0 0 8.86E1 3.53E 2 0 
● Where edges meet there is an OR Boolean  
operation of the fixed DoF  0 0 0 0 0 1.34E2

A Strand7 API code can be written to automate this
procedure and the following is the result of a typical
situation of a plate with blade stiffeners of different pitch
and dimensions in each orthogonal direction. A cell of
the plate subject to the unit twist strain vector
{0 0 0 0 0 1} is shown in Figure 3.3.
In this case, since the stiffeners are symmetrical about the
plane of the base plate, there is no membrane/bending
coupling.
Inverse homogenization is undertaken after the FEA on
the global model has been conducted. The N’s and M’s
for any element in the global model are used to generate
the strains and curvatures. When these actual strain
values are applied to the detailed cell model, the resulting
stresses on the cell model will reveal any locations of
stress concentration that ultimately would be the source
of failure. For example when the strain vector {0 0 0 0 0
1} is applied to the plate of Fig 3.2 the maximum surface
von Mises stress is over 2 times the mean value, see
Figure 3.4

Fig 3.3 - A 1x1m cell of 10mm thick sheet which is bi-directional blade
stiffened plate, X blades are 50mm high and 10mm thick on 200 mm
pitch, Y blades are 25mm high and 6mm thick on 100mm pitch. All
sheet has E=200GPa and Poisson’s Ratio is 0.3. Displaced shape due to
unit twist vector {0 0 0 0 0 1}.

The plate membrane/bending stiffness matrix for this

structure is
2.63E9 5.32E8 0 1.25E7 0 0 
5.32E8 2.43E9 0 0 3. 76 E 6 0 

 0 0 8.01E8 0 0 0  Fig 3.4 - The perforated ribbed baffle plate of Figure 3.1 with a strain
  vector {0 0 0 0 0 1} applied. Surface von Mises stress contours showing
1.25E7 0 0 4.34E5 4.19E3 0 
the stress concentrations.
 0 3.76E6 0 4.19E3 8.05E4 0 
 
 0 0 0 0 0 8.71E3
Grade Plates and Bricks
This result can be compared with standard orthotropic
plate theory and without going into detail, the values are
the expected ones, including the membrane/bending
coupling terms of 1.25E07 and 3.76E06. Full detail of
A n important part of building a finite

this comparison and the methods are available from
Strand7.
Another example is the perforated baffle plate in Figures
3.1 and 3.2 where the pattern repeats every 400mm in X
and Y and all the molded thicknesses are 10mm. The
elongated slots have a 10mm high two sided lip of the
same material and 5mm high ribbing. Note that on the
edges of the cell lie the planes of some of the ribbing. To
properly capture the correct stiffness of this edge, plates
are half thickness. The material is a toughened poly-
propylene with a Young’s Modulus of 4GPa and a
Poisson’s Ratio of 0.4. The membrane/bending stiffness
matrix is:
element model is mesh refinement around
areas of stress concentration or other areas
of interest. Consequently, the mesh must
transition from a coarse mesh to a fine
mesh whilst ensuring that all elements
remain compatible and retain acceptable
aspect ratios. Strand7 includes many
standard meshing tools used to mesh these
transitions, that is, to go from say 4
elements to 2 or from 3 elements to 1. The
Grade tool (shown right) provides a quick
way to apply these standard details in plate
and brick models by mapping the mesh
pattern onto existing elements and
subdividing the elements.

www.strand7.com
Making Finite Element Analysis Easier
The dotted lines on the mesh pattern buttons indicate the
edge/s you need to select to perform the desired grade.
To access the Grade tool, choose Tools/Grade Plates
and Bricks.
Note: Virtually all the grade tool options are equally
applicable to bricks and plates.

Grading from 1 → 2 select the following brick edge,

There are two grade options available:

Option 1 1 x 2 Grade

Option 2 1 x 2 Tri Grade

then apply the 1 x 3 Grade.

In the refinement of most quad dominant meshes Option

1 would be a better choice than Option 2 because of the
two quad elements it creates. Option 2 would be a better
choice if you were preparing a plate mesh to be used as a
bounding surface for a tetrahedral automesh.
If you have pairs of adjoining plates that need to be
graded from 1→2, then an even better option would be
Quarter Quad Grade (detailed on the following page). This can be applied to plates as well as bricks.

Example Grading from 2 → 3

If you had the following mesh in your model, There are two grade options available:

Option 1 2 x 3 Grade

select the following edge,

Option 2 2 x 3 Tri Grade

Note: Both grade 2 → 3 but Option 1 additionally divides

the element across its width.
then apply the 1 x 2 Grade (Option 1) Grade.
Example 1
If you had the following mesh in your model (notice that
all four sides of the plate need to be made compatible),

Grading from 1 → 3
There is one grade option available:

Option 1 1 x 3 Grade

select the following edge,

Example
If you had the following brick mesh in your model,

www.strand7.com
Making Finite Element Analysis Easier
then apply the 2 x 3 Grade (Option 1).
Example 2
If you had the following brick mesh in
your model,
select the following edge,
then apply the 2 x 3 Grade (Option 2).
Grading from 2 → 4
There is one grade option available:
Option 1
Example
If you had the following plate mesh in your model,
www.strand7.com
Did you know?
Animations in PowerPoint
When presenting your Strand7
analysis to your client a
PowerPoint presentation can be
very effective. To give depth to
the content a Strand7 Animation of
results can be included.
There are two common methods select the following edges,
that can be used:
1 – Save the animation as an
exe file. Then in PowerPoint
choose to Insert a Hyperlink
and browse to the correct
location.
2 – In the Strand7 Animation
Window choose to save the
animation as a series of
then apply the 2 x 4 Grade.
images . Then create a
gif animation of these in an
appropriate software package.
This can then be inserted in
the PowerPoint presentation
with the animation permanently
running.
Quarter Quad Grade
There is one grade option available:
Option 1 Quarter Quad Grade
Note: The Quarter Quad Grade is essentially half of the
previous 2 × 4 grade. This is a very useful tool where you
have pairs of adjacent elements that need to be divided
1 → 2.
Example
If you had the following brick mesh in your model,
select the following edges,
2 x 4 Grade
then apply the Quarter Quad Grade.
8
Making Finite Element Analysis Easier

Select the two edges of the orange plate,

It can be seen that a good quality Quad/Hexa mesh can

be achieved.

Circle Grade
There are three grade options available:

then apply the Annulus Grade with a ratio of 2/3 (since

Option 1 Quarter Circle Grade
you want to remove 2/3 of the plate).

Option 2 Full Quarter Circle Grade

Select the two edges of the blue plate,

Option 3 Annulus Grade

Example
Using the three grade options above, the following mesh
will be created:

then apply the Quarter Circle Grade with a ratio of 3/5

(since you want to remove 3/5 of the plate).

Start with 3 squares, partially overlaying each other, with

edge lengths of 2 (green), 3 (orange) and 5 (blue) as
shown in the following picture:

The completed model will look like this:

Select the two edges of the green plate,

Brick Corner Cut

There is one grade option available:

Option 1 Brick Corner Cut

then apply the Full Quarter Circle Grade.
9

www.strand7.com
Making Finite Element Analysis Easier
Example select any edge on the triangle,
If you had the following brick mesh in your model,

then apply the Tri Grade 5 (Option 2).

select the following edges,

then apply the Brick Corner Cut.

Construction Tools
There are five grade tools. These are most commonly
utilised as construction tools, where nodes or splitlines
are required in specific locations

Option 1 Quad-Tri Grade 1

Triangular Grade
There are two grade options available, (the other three are Option 2 Quad Cut-out
more frequently used as construction tools):

Option 1 Tri Grade 3 Option 3 Tri Grade 1

Option 2 Tri Grade 5 Option 4 Tri Grade 2

In most situations Option 2 would be a better choice than

Option 1 because of the quad elements it creates. Option Option 5 Tri Grade 4
1 would be a better choice if you were preparing a plate
mesh to be tetrahedral automeshed.
These two grade options are identical to performing a 2
subdivide on a triangular element. Modeling Method Comparison
Example
Modelling Simplifications Using Symmetry
If you had the following plate mesh in your model,

A nalyzing a structure using FEA does not necessarily

mean that every detail needs to be modeled, or in fact
that the model must be 3D. There are a number of
assumptions that can be made when modeling that can
lead to large reductions in analysis time for a structure.
This article outlines four of these methods and their
advantages and limitations.

10

www.strand7.com
Making Finite Element Analysis Easier
Method 1 – 2D Axisymmetric Modelling The sector symmetry link is used to enforce a sector or
cyclic symmetric behavior on a sector of a circular
In an axisymmetric model, the geometry and boundary structure, about one of the global X, Y or Z axes.
conditions are assumed to be revolved 360° about an
axis. Axisymmetric models are 2D plane models which To create a sector model in Strand7 first create a plate or
mathematically represent a slice that is revolved 360°. An brick mesh of a sector. For the model shown in Figure
advantage of this method is that instead of a full 3D 5.2 this could be done by extruding the axisymmetric
model a finer mesh can be used without significant model a distance of say 30° to create a segment.
increases in solution time. Also if contact or other
nonlinear solutions are required the savings in solution
time can be significant.
To create an axisymmetric model in Strand7 you should
first create a plate mesh in the XY plane. This should be a
cross section of the structure.
Under Property/Plate select Axisymmetric from the
Type options.
Assign the appropriate loads and restraints to the model
and solve.
Y
Figure 5.2 – Sector symmetry model.

Sector Symmetry restraints then need to be applied to the

model.
X Choose Tools/Auto Assign/Restraints and select Sector
Symmetry.

Select the appropriate axis of rotation and enter the θ1

and θ2 coordinates (or use the hotpointer), which
Figure 5.1 – Axisymmetric model. describe the two edges of the sector.

Restrictions
● The model must be created in the +X/+Y or
+X/-Y plane with the global Y axis assumed to be
the axis of rotation.
● Loads can only be applied in the plane of the
plate elements (XY plane). Out of plane loads are
ignored. Any in-plane edge shear or edge
pressure assigned is assumed to be acting around
a 360° perimeter. Any point loads applied are
assumed to be force per radian.

Figure 5.3 – Applying sector symmetry restraints.

Method 2 – Sector or Cyclic Symmetry
Sector symmetry is a specialized condition where features
that are repeated about an axis can be modeled by a
single segment or sector of the feature.
This is strictly different from axisymmetry in that with an
axisymmetric analysis, all result quantities at a given
radius are identical over the full 360° circumference.
With sector symmetry, result quantities at a given radius
are enforced to be identical only on the sector
boundaries, i.e. planes θ1 and θ2 described in the
following. In one sense, axisymmetry can be considered
to be a special case of sector symmetry.
This method is advantageous when there is repeating
geometry and loads that cannot be modeled as an
axisymmetric model, e.g. a gear with teeth.
Again assign the appropriate loads and restraints to the
model and solve.
Restrictions
● The sector must be located such that its axis of
symmetry is positioned at the origin of the global
Cartesian system.
● Any loading applied to the structure is assumed
to be repeated in each sector of the full structure,
i.e. the complete structure can be created by
replicating the sector around the complete 360°.
● Every node on the θ1 plane should have a
corresponding node on the θ2 plane.

11

www.strand7.com
Making Finite Element Analysis Easier
Method 3 – Reflective Symmetry
If symmetry exists in a structure and the applied loads are
also symmetric, then the benefits of this should be
utilized. An advantage of using symmetry is it allows a
detailed model to be created without extensive solving
time.
The restraints on a reflective symmetry model define the
symmetry in the model. These restraints should prevent
translation through the plane of symmetry on the entire
cut face. For beam and plate models, rotation in the
directions parallel to the cut plane should also be
restrained.
To create a reflective symmetry model in Strand7 first
create a plate or brick model of the symmetric portion of
the structure. Again the axisymmetric model could be
extruded 90° to form a quarter model.
Figure 5.4 – Reflective (quarter) symmetry model.
Cartesian symmetry restraints can be automatically
assigned in Strand7.
Choose Tools/Auto Assign/Restraints and
Cartesian Symmetry.
This will assign the appropriate restraints assuming the
planes of symmetry are the XY, YZ and ZX planes,
depending on the spatial location of the nodes.
YZ sym XY sym
Y Following is a comparison of the results for all four
X ●
Z vertical direction.
Figure 5.5 – Applying reflective symmetry restraints using the Auto
Assign Restraints tool.
Note: To ensure that symmetry restraints are not placed
along the ZX symmetry planes the base of the model is
located slightly above the global Y origin.
www.strand7.com
Apply the remaining restraints and loads to the model and
solve.
Although this example has used a 90° sector, in Strand7
reflective symmetry can be applied about any angle sector
by the use of a cylindrical UCS. However, the majority of
models that exploit reflective symmetry are usually
modeled with 90°, i.e. quarter models, or 180°, i.e. half
models, segments.
Restrictions
● Symmetry conditions require that the geometry
and boundary conditions are or can be
approximated as being equal across the symmetry
plane.
● Any loading that is applied directly on a
symmetry plane needs to be divided by the
number of symmetry planes at that point.
Method 4 – Full Model
A structure is modeled in full generally because it is not
possible to take advantage of symmetry in the model,
whether due to unsymmetric loading or geometry. A full
model may also be required for buckling and dynamic
analysis.
Advantages of this method are that no assumptions in
terms of symmetry need to be made and presentation in
reports displays the whole model, which can be easier to
interpret for the inexperienced.
Creating a full model in Strand7 is simply a matter of
modeling all relevant aspects of the geometry in full and
applying real life restraints and loads.
select
The axisymmetric model can be extruded 360° to create
the full model.
Figure 5.6 – Full model.
Results Comparison
methods.
The base of each model is restrained in the
● A material property of Structural Steel is assigned.
● A pressure load of 15MPa is applied as a pressure
acting vertically down on the top of the model.
12
Making Finite Element Analysis Easier

Axisymmetric Model Sector Symmetry Model

Reflective Symmetry Model Full Model

Figure 5.7 – Comparison of vertical displacement (DY) for the four methods.

elements available in Strand7. It

As expected the three models Did you know? demonstrates a correlation of
containing hexa elements produce
Summary-Information results between the four that
absolutely identical results. The
allows the user to be confident
axisymmetric model differs At the top of every print from Strand7 a title block is
presented where information on Project Title, that the method they select will be
slightly for the maximum
Company Name, etc. can be given. accurate. However, the method
displacement. This small margin
that you choose is determined by
is due to the differences between
the level of symmetry that exists in
the quad and hexa element
the model with regard to
formulations.
geometry, restraints and loads.

Conclusion The models for the four examples

To fill this in simply choose used in this article can be
This article shows four methods of downloaded from
Summary/Information and complete the
modeling the same structure using required fields. The information here will then be www.strand7.com/news.st7.htm
the axisymmetric and brick displayed on all Print jobs from Strand7 (unless
turned off).

13

www.strand7.com
Making Finite Element Analysis Easier

Training
Also with the recent increase in the number of users
T he Strand7 training calendar for the first half of 2005 internationally we are investigating the possibility of
holding local Strand7 training courses in the USA and
has been released with new topics added to the already other countries. If you are interested in please contact us
popular course schedule. This year we are launching at info@strand7.com.
courses covering the Strand7 API and also seismic
analysis. We are also pleased to announce plans for a
Strand7 training course to be conducted in Melbourne in
April.
2005
18-22 April Melbourne Strand7 Course
17-20 May Introducing Strand7
23 May Automeshing
24-25 May Structural Analysis
26 May Earthquake Analysis Using Strand7
27 May Introduction to the Strand7 API

Fig 6.1 – New Strand7 training room in York St.
The Melbourne course will be modularised with
participants able to attend topics ranging from Strand7
essentials to more advanced topics. This new course is in
direct response to requests from Strand7 Users for training
in their own state. If you would like to see a Strand7
course in your city then please contact us to register your
interest.
Personalised Training
In recent months we have also had increased interest and
activity in our tailored courses. Late 2004 saw two,
separate, week long courses held in Perth for Sinclair
Knight Merz and Calibre Projects and, just recently, a one
day intensive course held for GHD in Melbourne.
Tailored courses allow companies to choose particular
topics from our extensive list and this personalised service
has been well received. If you would like more
information on tailored courses then please contact us.
Details on Strand7 Training Course content can be found
on our website at www.strand7.com/training.htm or by
contacting us

Exhibitions

Did you know?
Length-Area-Volume Loads
The ability to assign element attributes based on their
geometry is a particularly useful one. In Strand7
attributes can be assigned based on the length of a
beam or plate edge (L), the area of a plate or brick
face (A) and the volume of a brick (V). For example, a
beam pre-tension that is proportional to the beam’s
length can be applied by entering an equation such as
2.5*L, where L represents the length of each beam
element.
W e have been delighted in the last few months to have had the
opportunity to meet so many of our users at various exhibitions
around the world, particularly in China and the USA.
Recent exhibitions attended have included the FHWA Steel Bridge
Conference where we had strong interest from the structural market
and the Pacific Design Show in California, which attracted users
from mechanical fields of engineering. Strand7 was well received
at all of these.

Furthermore, each of
these variables can be
prefixed with the letter
T, signifying Total, e.g.
if five plate elements
are selected, to assign
a force of 1500N, enter
1500/TA which will
divide the force by the
total summed area of Fig 7.1 – China International Steel Construction Expo/Conference.
the five plates to give
the correct pressure.

14

www.strand7.com
Making Finite Element Analysis Easier
In the coming months Strand7 will be exhibiting at the following
locations:

International Technology Education

th Kansas City, Missouri,
3-4 April Association – 67 Annual
USA
Conference; Booth 321

North American Steel Construction

6-8 April Montreal, Canada
Conference; Booth 133

SAMPE 2005 – Composites Show;

3-5 May Long Beach, CA, USA
Booth 559

We look forward to meeting more of you when we next exhibit.

Fig 7.2 – FHWA Steel Bridge Conference, Texas. Please come and introduce yourself

Announcements

A s mentioned on the first page, January saw G+D Coinciding with the re-branding is a move to new and
larger offices in the heart of the Sydney CBD
Computing begin to trade as Strand7 Pty Ltd.

15

www.strand7.com
Making Finite Element Analysis Easier

Strand7 User Profile

Phil Herting
Beaufort Composite Technologies, Inc.
Q – How would you describe your business? Q – What features of Strand7 do you find most
useful?
BCT is a composites manufacturing company that
We own various CAD programs, but the CAD program
designs and produces epoxy, prepregnated, carbon
most widely used by far is Rhinoceros, due to its 3D
components and structures for the aerospace,
precision and ability to import and export most other
automotive and medical industries.
CAD programs. Our customers frequently use various
Q – What types of jobs do you use Strand7 for? other CAD programs including ProE, SolidWorks,
AutoCad and Catia and we easily import their work in
BCT uses Strand7 to model composite structures from
Rhino, clean it up, and then export an iges to import into
laser eye surgery tables to very complex systems such
Strand7. The show free edge function in Strand7
as satellite dishes and compact range reflectors.
makes any questionable surface interfaces quickly
Many analyses have been completed ranging from apparent and these can be repaired as necessary.
linear static, to natural frequencies, as well as response The composites module is flat out - great. Developing
spectrums for seismic analyses. Heat transfer solvers the laminate (at times – greater than 60 plys), setting
are also utilized to account for the thermal expansion of
plate orientations and offsets are quick and simple. We
masters and molds undergoing temperature changes
even use the laminate presentations provided in
during curing cycles.
Strand7 on the shop floor.
Q – How does Strand7 help you in what you design Another feature that deserves mentioning is the - free to
and analyze? download - Strand7 Viewer. This has proven
We have been extremely pleased with Strand7’s ease of invaluable, where our customers can review a complete
use, solving speed, results presentation and particularly analysis that we have created on their project.
its ability to address complex laminate structures. Contact Details
The precision Strand7 provides in its meshing is not just Beaufort Composite Technologies, Inc.
helpful, but essential. Because BCT is concerned about 111 Safrit Drive
the deformation of the overall dish shape under various
Beaufort, NC 28516
load conditions, we run 3D best-fit regressions against
the displaced nodes. This means that the undisplaced Phone: +1 252-728-1547
nodes need to be as accurate to the designed shape as Fax: +1 252-728-1755
possible. When running RMS calculations, we Website: www.composites.com
consistently find the undisplaced mesh to be better than
-6
10 inches, with very, very few exceedences.
Strand7 allows for the optimization of the laminate
structure, where the materials used can be very
expensive. For this reason, Strand7 helps BCT’s
competitiveness in the marketplace by facilitating the
design of the composite structures within their required
specifications and requirements, avoiding overbuilding
and hence overpricing.

Head Office
Strand7 Pty Ltd
Suite 1, Level 5
65 York Street
Sydney NSW 2000
AUSTRALIA
Tel +61 2 9264 2977 Website: www.strand7.com
Fax +61 2 9264 2066 Email: info@strand7.com

16

www.strand7.com
Strand7 Pty Ltd

News.St7
Newsletter for Strand7 and Straus7 users
Strand7 is marketed as Straus7 in continental Europe

Issue 5, September 2005

Welcome to another issue of News.St7 for 2005. In this issue we would

In this issue…
like to say a special hello to our Strand7 users in the United Kingdom who
may be reading this for the first time, and to those who are regular • Feature Article 1
subscribers. This issue of News.St7 may be of particular interest to those of
• Strand7 Skull Analysis 2
you in the UK as we feature the launch of Strand7 UK Ltd, an office of
Strand7 dedicated to our users across the ocean. • Plate Alignment 4
Of course this issue is also of interest to all Strand7 users as we feature • Identifying Ill-Conditioning 7
articles on how Strand7 is being used for a wide range of applications from • API in the Field 9
analysing bat skulls to looking at the stresses in a reinforced tunnel.
• User Jobs 10
We also feature articles on plate axes alignment, identifying ill-conditioning
and the usual array of Did You Know items to motivate the mind. • Training 12

If you have any feedback or suggestions regarding the content of News.St7 or would like to feature any • Exhibitions 12
of your projects then please email news.st7@strand7.com. If you would like to automatically receive
your copy of News.St7 directly by email simply send a blank email to newsletter- • User Profile 12
subscribe@strand7.com. All care is taken to ensure that information in News.St7 is accurate and up to
date at the time of publishing. However Strand7 Pty Ltd accepts no responsibility for inaccuracies in, or changes to, such information.

Strand7 UK Launches

E arlier this year the office of Strand7 UK Limited opened

its doors; the first office of Strand7 Pty Ltd to be located
outside of Australia. This office is seen as an investment
into expanding the already healthy relationship that exists
between the developers of Strand7 and UK users.
Located in the bustling town of St Neots, the third largest
town in Cambridgeshire, and just 57 miles from London,
Strand7 UK has become the centre of all contact for users
across Great Britain. Whether it is just information you
require, or more detailed Strand7 support, this office is
the first port of call for all users, and potential users, in the
UK.
Having a direct presence in the UK will give Strand7 users
significantly enhanced technical support and a more
direct link to the Strand7 Software development team
back here in Australia. Figure 1.1 - Location map of Strand7 UK Limited.

The number of Strand7 users in the UK has grown
consistently in recent years and Strand7 is now widely
used in a range of engineering disciplines including
mechanical, aeronautical and automotive applications.
In addition, Strand7 is increasingly being used for
civil/structural applications by UK consultancies looking
to perform more sophisticated analysis than what is
usually possible with traditional frame analysis software.
In the upcoming months the links between Strand7 UK
and Strand7 Pty Ltd will become more and more
apparent, with many joint activities planned for the

www.strand7.com 1
Making Finite Element Analysis Easier

second half of the year. This constant contact will further
enable our developers in Australia to create and expand
features in Strand7 requested by customers.
November is the biggest month with visits, exhibitions
and courses planned across the UK.
From November 22-24, Strand7 will have a booth located
at Civils 2005, the UK’s number one civil engineering
showcase. Civils 2005 is being held in the National Hall,
Olympia, London.
Also in November we plan to hold a five day Strand7
training course in the UK. This course will follow the
format of the recent external courses held in Melbourne
and Brisbane, Australia, with a modularised course over
the five days. Three days will be devoted to Strand7
Essentials and two days to Nonlinear and Dynamic
analysis, subject to feedback from interested parties. For
more information or to download the registration form go
to www.strand7.com/training.htm. This course will be
presented by Strand7 staff from Australia together with
Strand7 UK staff.
If you would like any more information on Strand7 UK
then please contact us at
Strand7 UK Limited
The Studio Office
Church Walk
St Neots
PE19 1JH
England
Tel: +44 (0) 1480 211 011
Fax: +44 (0) 1480 211 020
Email: info@strand7.co.uk
Web: www.strand7.co.uk
We are always very pleased to hear from Strand7 users
gathered from wild animals to load finite element models
of mammal skulls. By varying the loading conditions, we
study how skulls dissipate forces generated by typical and
atypical feeding behaviors. We predict that forces are
dissipated more efficiently via internal stresses in the skull
(i.e., skulls are more resistant to feeding loads) under
typical loading condition.
Bats are an ideal model organism for this work for three
reasons. First, bats exhibit the greatest diversity in skull
shape and broadest range of diet among all the orders of
mammals. This allows us to compare animals that are
closely related but have very different craniofacial
structures. Second, the skulls of bats are more likely to be
optimized to transmit biting forces than are those of other
mammals because the metabolic cost of flight is high.
Thus, any tendencies for the skull to be “overbuilt” should
have been reduced over evolutionary time. Third, bats are
among the most abundant mammals in the world and are
relatively easy to work with in the wild.
Building a 3-D finite element model of a structure as
complicated as a skull poses a significant technical
problem. Because the skulls of the bats we study are
roughly only 15-20mm long, we turned to micro-ct
scanning to capture detailed anatomical structures (Fig
2.1). We built detailed surface models of entire skulls
from stacks of serial ct-scans using AMIRA (Mercury
Computer Systems) (Fig 2.2). The surface models were
saved in STL format and imported into Geomagic
(Raindrop, Inc.) where small, particularly complex regions
of the skull were edited manually. Once we were satisfied
with the geometry, we imported the models to Strand7 as
a 3-D surface triangulation (i.e., an *.stl file). Within
Strand7 a plate element mesh was automatically
constructed from the imported *.stl triangulation.

Strand7 Skull Analysis

Using Strand7 to Study the Mechanics of
Feeding in Mammals

T he earliest mammals were tiny, insect-eating creatures.

From these unremarkable ancestors, the diversification of
mammals is largely a story about exploiting an ever
broader array of food resources. Broad associations
between the shape of mammals’ skulls and their diets are
obvious. For example, the architecture of a cow’s skull
enables it to chew grass efficiently while the skulls of
tigers are well equipped to bite and tear flesh. Despite
these correlations, we don’t understand the mechanistic
link between diet and skull shape.
Fig 2.1 – A single slice through the skull of a bat as seen in a micro-ct
The skulls of mammals serve a number of different scan. Bone is white.
functions; protecting the brain, housing important sensory
organs, and contributing to the first portion of the
digestive system. Our lab investigates the hypothesis that
skull shape has evolved (at least in part) to withstand the
forces that are generated during feeding. We accomplish
this by using data on feeding behavior and bite force

www.strand7.com 2
Making Finite Element Analysis Easier
Fig 2.2 – 3D surface model of the skull of the Jamaican fruit-eating bat
(Artibeus jamaicensis).
It is impossible to build such complicated models without
errors and Stand7’s mesh checking algorithm has proven
exceptionally useful. Nevertheless, there are inevitably
many plate free edges and t-junctions that must be fixed
manually. Once the plate models were free of errors, we
used Strand7’s automatic tetrahedral mesher to create a
volumetric mesh of ten-noded tetrahedrals. After this step,
all plate elements were removed, leaving only a
volumetric mesh that recreated the geometric structure of
the skull in exquisite detail (Fig 2.3).
Fig 2.3 – FE model of the skull containing 251,968 tetrahedral elements.
Applying realistic loads to the models was, of course,
crucial to generating meaningful results (Fig 2.4). We
applied load vectors to three nodes representing each of
the two primary jaw closing muscles: masseter and
temporalis. Constraints were applied at the three places
where the lower jaw contacts and transfers forces to the
skull during feeding; the center of each jaw joint, and the
tip of the tooth where biting occurs. A single node in the
center of each joint was constrained against displacement.
This effectively created an axis of rotation for the skull
due to the application of muscle forces. To prevent this
rigid body motion and induce elastic deformation in the
skull due to biting forces, nodes on the tips of the teeth
involved in biting were constrained against displacement
www.strand7.com
(i.e., displacements in the X, Y, and Z planes were set
equal to zero).
Fig 2.4 – Point loads representing muscle forces (arrows) and constraints
(asterisks) at the biting tooth and jaw joints.
Each analysis of a biting behavior was completed in two
steps. Initially, an arbitrary total amount of muscle force,
FT, was divided between the masseter and temporalis
muscles based on muscle mass proportions. All muscles
were assumed to act simultaneously and all dynamic or
transient effects were neglected. Once the analysis
problem was solved, the reaction forces at the constrained
tooth required for system static equilibrium were
determined. This reaction force, FRn , was then compared
to experimental in vivo bite force measured for the bat
species, Fexp. Since the computed reaction force is in
direct proportion to the total applied muscle load, the
required total amount of muscle force, (FT)new , necessary
to yield the experimentally measured bite force is given
simply by:
 Fexp 
(FT )new =  F
n  T
 FR 
In the second step of the analysis, the computed total
amount of muscle force (FT )new, was distributed among the
masseter and temporalis muscles based on muscle mass
portions. The solution of this second analysis problem
yielded the deformation of the bat skull, strains, and
stresses for a particular feeding behavior that resulted in
reaction force(s) at the constrained tooth (teeth) that
identically matched voluntary bite force values collected
in the field. Essentially, known bite force values were
used to calculate the muscle forces required to maintain
static equilibrium.
Perhaps not surprisingly, there is no data summarizing
Young’s modulus or Poisson’s ratio for the very thin and
highly curved bones of bat skulls. However, comparative
studies of the stiffness and yield strength of cortical bone
suggest that material properties are relatively constant
over a wide range of vertebrates. Based on these
comparative data, we assigned our models average values
of Young’s modulus (E = 2 x 1010 Pa) and Poisson’s ratio
(ν = 0.3) for cortical bone.
3
Making Finite Element Analysis Easier
Feeding behaviour experiments carried out in the field
demonstrate that the Jamaican fruit bat typically bites hard
fruits with the molar teeth on one side of its mouth; it
rarely bites hard fruits with its canines. We found
significant differences in the patterns of stress under these
two loading conditions, especially in the palate (the roof
of the mouth). During unilateral molar loading, stresses in
the palate were relatively low and localized (Fig 2.5). In
contrast, bilateral canine loading resulted in widely
distributed stresses that were high and concentrated near
the pterygoid plates (toward the back of the palate) (Fig
2.6).
Fig 2.5 – von Mises stress in the palate under the common biting
behavior (bite force = 22.5 N).
Fig 2.6 – von Mises stress in the palate under /atypical/ biting behavior
(bite force = 18.8 N).
The ease of extracting quantitative data from Strand7 also
allowed us to demonstrate dramatic differences in the
volume of the skull experiencing stress under the two
loading conditions. After removing the elements affected
by the application of point loads, a constant bite force led
to a much larger proportion of the skull experiencing high
stress under atypical loading compared to the common
loading regime (Fig 2.7 and Fig 2.8).
www.strand7.com
Fig. 2.7 – Plot of stress versus skull volume demonstrates that the skull is
most resistant to loads imposed by the most common biting behavior.
Fig. 2.8 – Plot of stress versus skull volume for Atypical biting behaviour.
The bat skull does not constitute a “fully stressed design”.
Rather, these data support the idea that the skulls of
mammals are not optimized solely for feeding but
represent a compromise between competing functional
demands. Nevertheless, these results demonstrate a clear
link between the loading regimes that routinely occur
during feeding and the structure of the skull. Based on
simple lever mechanics, bite force is expected to increase
as bite points move closer to the jaw joints. However, the
Jamaican fruit bat exhibits a greater than expected
difference in strength between canine and molar biting.
We hypothesize that the shape of this bat’s skull is a result
of natural selection favoring an ever-increasing ability to
apply high bite forces. This led to the shortening of the
face, focusing biting behaviors on the molar teeth, and
strengthening the skull to withstand unilateral molar
loads
Article prepared for Strand7 Pty Ltd by Elizabeth Dumont, University of
Massachusetts. Research is from the labs of Elizabeth Dumont (Biology)
and Ian Grosse (Engineering) at the University of Massachusetts, Amherst
USA
Plate Alignment: Part 2
Plate Axes
F ollowing on from our feature on Plate Orientation in
Issue 4 of News.st7 we discuss the concept of Plate Axes
in Strand7.
4
Making Finite Element Analysis Easier

Like plate orientation, plate axes form an important part of

the quality assurance procedures of finite element
modelling. This article discusses the impact that
misaligned plate axes can have on results and also ways
to identify and correct this.

Overview
Plate axes refers to the plate local coordinate system.
Every plate element has a local coordinate system with
the x and y axes lying in the plane of the plate.

Fig 3.3 – Display of plate axes.

For clarity, this display option does not show the local z
axis. However, the local z axis can be visualised by using
the Orientation plate display option as described in Issue
Fig 3.1 – Plate axes diagram for a Quad4. 4 of News.St7.

Why is this important?

There are three main reasons why consistent and known
local plate axes is important.
1. Some plate element attributes are applied based on
the local coordinate system, including prestress, face
shear stress and normal pressure. If plate axes are not
aligned in a model then applied loads may be
incorrect.

2. Local plate axes are Did you know?

Fig 3.2 – Plate axes diagram for a Tri3.
used for non- Draw Ply Axes
Figures 3.1 and 3.2 show the plate local x and y axes. isotropic materials,
If your model contains laminate
e.g. orthotropic or properties you may wish to
Plate Axes Convention? laminate materials. investigate the local axis
The 1, 2 and 3 orientation of a specific ply.
Plate axes are initially defined by the node connection directions for Choose View/Entity Display
order of each element. The default local axis system is material properties, and when you set Draw Axes
defined as follows: e.g. E1, E2 and E3 for plate elements also enter the
1. Positive local x runs from the mid-side of side N1-N4 for orthotropic ply number you wish to view
plates, refer to the them for in the box on the right.
to the mid-side of side N2-N3 for a Quad or from N1
to the mid-side of side N2-N3 for a Tri. material properties
in the local x, y and Plate axes for that ply will be
2. Positive local y is normal to the local x axis, directed z axes respectively. displayed on the element.
away from side N1-N2 in the plane of the element. Note: For laminate
Implicit in this definition of the local axes is the direction properties you can
of a local z axis, which form a right-hand local coordinate also choose to display the plate axes for a specific
system on the plate. ply, see the "did you know" on this page.
3. Plate force, moment, stress and strain results can be
Viewing Plate Axes extracted based on the local coordinate system. It is
important that plate axes are aligned when doing this
There is a display option in Strand7 that allows for the
to ensure that results are consistent. For plate/shell
easy identification of plate axes in a model. Choose
models, the local z axis is particularly important
View/Entity Display and click the Plate tab . Set when looking at surface stresses and strains.
Draw Axes. The local plate axes will then be drawn for
each individual plate.

www.strand7.com 5
Making Finite Element Analysis Easier

How can the Plate Axes be Changed? Fig 3.6 shows the model in Fig 3.5 with the local x axis of
the plate elements aligned with the positive global X axis.
Two tools exist that allow the user to change the
alignment of the x-y axes on plates. One is more
commonly used to change the alignment of plate axes on
an individual level while the other is very useful in a
global sense.
Method 1 – Local Axis Angle.
• Choose Attributes/Plate/Local Axis Angle. Select the
plates whose local axes you wish to change, enter a
value for the angle and click Apply.

Fig 3.6 – Aligned plate axes model.

What Happens When You Subdivide?

There are two options available for plate axes when you
choose to subdivide an element in Strand7, these are
Parent centroid and Curvilinear. The option required
can be selected by choosing Tools/Options and selecting
one of the radio buttons under Plate axes alignment.

Fig 3.4 – Local Axis Angle attribute.

The local axis angle will re-orient the local axis system
based on the default coordinate system. A positive angle
rotates the local x axis in the right hand positive direction
about the local z axis. To increment or decrement the
current orientation angle, use the Add button instead of
Apply.
Fig 3.7 – Tools/Options Plate Axes alignment dialog.
The difference between the two options becomes more
apparent when an element is highly curved as shown in
Fig 3.8.

Method 2 – Plate Axes

• Choose Tools/Align/Plate Axes. Select the plate
elements you wish to align, choose the coordinate
system and axis to align to, the local axis to align, and
click Apply.

Fig 3.8 – Single Curved Quad8 element with local axis shown.

If Parent centroid is selected then all subdivided

elements inherit the local axes alignment of the parent
and therefore all resulting elements will have their local
coordinate systems parallel to each other, see Fig 3.9.

Fig 3.5 – Align Plate Axes tool.

www.strand7.com 6
Making Finite Element Analysis Easier

Now lets alter one coefficient of the second equation

slightly:
 1.000 − 1.000  x   4.000 
   =  
− 1.000 1.001  y  − 2.000

The solution is now:

x = 2004
y = 2000
Fig 3.9 - Subdivided Quad8 element with Parent centroid plate axes A 0.1% change in one coefficient has caused a 100%
alignment definition.
change in the solution!
If Curvilinear is selected the local x-axis of all subdivided
elements is aligned with the first curvilinear axis of the A Mathematical Representation
natural coordinate system on the element. For a curved
The first set of equations represents two straight lines in xy
(quadratic) element, this results in a series of local axes
space:
(non-parallel) curving around the element, see Fig 3.10.
1.000 x − 1.000 y = 4.000
− 1.000 x + 1.002 y = −2.000

The solution of these equations yields the intersection

point of the straight lines. If plotted, these two lines are
nearly parallel. Changing the coefficient from 1.002 to
1.001 in the second equation, represents rotating the
second line slightly, but doing so changes the intersection
point significantly.

Fig 3.10 – Subdivided Quad8 element with Curvilinear plate axes A Structural Representation
alignment definition.
The two connected springs shown in Fig 4.1, together
with the nodal forces and restraint indicated, will generate
Conclusion
a system of linear equations identical to the first matrix
The alignment of plate axes in a model should not be above.
overlooked, especially if attributes, material properties or
results that are based on local plate axes are used.
Checking the alignment of plate local axes should form
one of the items on a QA check list

Fig 4.1: Two connected springs with restraints and loads shown.
Identifying Ill Conditioning
It follows then that the solution of this structural system
will be very sensitive to the stiffness of the red spring. It
What is Ill Conditioning? turns out (in this case) that the solution is also relatively
insensitive to the stiffness of the blue spring: a 100%
I ll conditioning in Strand7 refers to an undesirable state change in that spring’s stiffness results in only a 0.2%
change in the solution! (This is left as a verification
of the stiffness (or mass) matrix. exercise for the reader.)
The matrix is considered to be ill conditioned if small
changes in the coefficients of the matrix have large effects Why Does it Happen?
on the results.
In this example, ill conditioning has arisen because the
matrices are nearly singular. In mathematical terms, the
A Numerical Example two equations are nearly linearly dependent, i.e. one
Consider the following simple set of equations: equation is almost a linear ratio of the other; in the above
matrices, the second equation is approximately –1.0 times
 1.000 − 1.000  x   4.000  the first equation. In structural terms, the ill conditioning
   =  
− 1.000 1.002   y  − 2.000 has occurred because of the large relative difference in
the stiffness of the two springs; the addition of the second
The solution is: spring makes little difference to the matrix so its
x = 1004 significance is almost lost.
y = 1000

www.strand7.com 7
Making Finite Element Analysis Easier
The term “condition” is actually a mathematical term.
Most books on numerical methods will cover the so-
called condition number of a matrix and show that the
higher the condition number the more ill conditioned the
matrix becomes. The condition number is related to the
ratio of the magnitudes of the maximum and minimum
eigenvalues of the matrix; in a matrix with zero
eigenvalues, the condition number is infinity, that is, the
matrix is singular.
Although a low condition number usually means high
solution accuracy, a high condition number does not
always mean low solution accuracy. The condition
number provides a lower bound of precision, which is
often pessimistic. More information on this can be found
in texts on numerical methods.
Ill Conditioning and Numerical Precision
Although both solutions for the above matrices are exactly
correct for the given coefficients, that is, neither matrix is
actually singular, if we truncate all the coefficients to
three digits, then both matrices will become singular and
cannot be solved.
This is a very important point with reference to FEA
because the equations are stored and solved using a finite
precision on a digital computer. Strand7 uses so-called
double precision, which stores a real number using 8
bytes, providing around 14 digits of precision. Fourteen
digits sounds like a lot of digits, but it is not difficult to
generate matrices with large differences in coefficients
even for physically meaningful structures.
For example, a plate element representing a 5mm thick
steel plate will generate coefficients corresponding to the
in-plane stiffness of the plate of the order of 1.0E9 while
the coefficients corresponding to the bending stiffness will
be of the order of 2.0E3 (using SI units): a relative
difference of six digits.
Did you know?
Hide Zeros
Have you ever found yourself with so many linear or
nonlinear load case combinations that you lose track of
which columns have load and which are set at zero? In
Strand7 there is a quick and easy way to hide all those
zeros; to make viewing loads a breeze. In the Linear load
case combinations (or nonlinear load increment) window,
right click and select Hide Zeros. All zeros are hidden
making it a simple matter to see where all the load has been
applied.
www.strand7.com
Loss of precision due to the number of digits required to
store, i.e. represent, the real numbers is only half the story
however. The other half is the so-called round-off error
that is generated and accumulates during the matrix
decomposition (solution) phase.
In the above, we have assembled two ill conditioned
matrices with sufficient precision to fully represent the
information (4 digits). We have then solved the matrices
with sufficient additional precision to ensure that the
solutions obtained are accurate. If during the elimination
process we operate on numbers of very different
magnitudes, we progressively lose digits of precision in
the result. Consider the following example:
Working with four digits of precision, we add the number
1.234E3 to the number 1.234E0: the result becomes
1.235E3 which means we have effectively lost three digits
of precision.
What Does this Have to do With my Analysis?
Ill conditioning can have detrimental effects on any
analysis, possibly generating unreliable results. Ill
conditioning can also be very problematic in nonlinear
analysis where it could have a significant effect on
convergence rates and levels. For example, it is possible
to devise an FEA model where the convergence tolerance
will never get below a certain value no matter how many
iterations are used, even if the solution has basically
converged.
What Causes Ill Conditioning in FEA Models?
The following are some examples of where ill
conditioning might occur:
• An element or a group of elements respond to
loads with large rigid body motion but little
deformation. This could be something that is not
restrained properly or a very stiff part that is
supported entirely by significantly more flexible
elements.
• Thin shell models where the tensile stiffness is
much higher than the bending stiffness.
• Models with two materials that have Young’s
moduli differing by several orders of magnitude.
• Short beam elements with high Young’s modulus
and large section areas used to represent quasi-
rigid connections.
• Highly distorted or high aspect ratio elements.
• Overly stiff supports.
• Overly stiff contact elements. In most cases it is
very useful to activate the Dynamic Stiffness
option and let Strand7 determine the most
appropriate stiffness value. This may sometimes
increase the number of iterations required, but
usually will produce more accurate results.
8
Making Finite Element Analysis Easier

In most cases where very stiff parts of a structure are
required, replacing the stiff elements with Rigid Links will
reduce the ill conditioning.
Identifying Ill Conditioning in Strand7
The Strand7 Log File Viewer can help you identify some
of the tell tale signs of an ill conditioned matrix. The
following should be checked:
• Warnings about rigid body motion (the possible
modes will be reported).
• The ratio between the maximum and minimum
pivots.
• Any reported negative diagonals. This indicates
an unstable structure and the source of the
negative pivot should be investigated.
• The maximum displacement magnitudes to
ensure that they are are not abnormally large.
Note that if all your Young’s moduli are out by an
order of magnitude, then in a linear static analysis
your displacements will also be out by an order
of magnitude, but your stresses will not be.
Looking at only the stresses will not reveal an
error in the Young’s modulus.
The point made about the definition of the condition
number is also useful because we can look at a natural
frequency analysis to see if there are any zero (or near
zero) frequencies. Although it is generally not possible to
get the highest eigenvalue to calculate an actual condition
number, a zero eigenvalue will usually point to a free
structure (or part of the structure)
a large number of parameters can be dealt with very
efficiently via the API, including:
• General tapered cross section with welded, flanged or
slip-joint connections.
• General piecewise distributed load for wind actions in
all directions.
• Fast handling of non-linear increments with load
multipliers derived from different limit states codes.
While the PALI input files are text files (eventually these
may be generated through a web interface) the output can
be given in different formats such as:
• Text File
• Excel Spreadsheet
• General Database (via SQL)
It is thus possible to keep track of each calculation
performed. If required, the storage of results may include
the original Strand7 model file containing the parametric
FEM model. This model is available for the extraction of
any result which is not directly found within the PALI
output file.
Figure 4.1 illustrates the data flow within the application.
Geometry and design code inputs are text files, with PALI
and VENTO both Windows EXE files (console
applications which run in batch mode). In addition to the
Strand7 API DLL, another DLL provides a library of
common functions used by both PALI and VENTO. The
application is written in Delphi.

API in the Field

Application PALI
This application has been developed by Engin Soft S.p.a.
(Italy) for Tecnopali S.p.a. It is part of a financed regional
research project named “multi-objective optimization of
steel poles”. It consists of two modules:
1. Pole Calculation Module (PALI)
Figure 4.1: Flowchart of application.
2. Load Calculation Module (VENTO)
The main aim of the application is the structural Figure 4.2 shows a typical Strand7 model automatically
calculation and verification of a geometrically nonlinear generated with PALI. This example represents a 12m high
steel pole subject to static, wind and seismic actions tapered steel pole.
according to various international codes.
The main application is divided into two parts:
1. The first part generates all external loads according to
the selected international design code.
2. The second part interacts with the Strand7 API to
create a parametric model of the steel pole, apply the
loads, run the Strand7 solver, and generate all the
results required for post-processing and verification.
The Strand7 API gives access to almost all the functions
available through the graphical user interface. In addition,

www.strand7.com 9
Making Finite Element Analysis Easier
Figure 4.2: Typical Strand7 model of pole, automatically generated.
The Strand7 Finite element analysis environment has thus
been used, within the same company, to address two
different tasks:
1. Using the Strand7 interface, a general FEA software
system is available to perform global or detailed
calculations (e.g. regarding stress concentrations at
openings or base plate design, etc).
2. Through the interaction with the Strand7 API, very
fast global calculations of known products can be
performed with only location and pole geometry as
input parameters.
This application can also be used within a general
optimization environment to reduce cost and/or weight of
new products or even to generate catalogues for different
sites
For additional information:
Daniele Schiavazzi d.schiavazzi@enginsoft.it
User Kiama
North Jobs Bypass TechSpan
North Kiama Bypass
S ay Kiama and people know what you are talking
about: a beautiful township, two hours south of Sydney.
Stunning beaches, lush green mountainous surroundings,
little craft shops, cafés and of course the widely renowned
spectacular “Blow hole”. Say Kiama at the Reinforced
Earth Company’s office and RECO’s successful team of
Designers, Engineers and Draftsmen think of something
completely different: Challenging; Multi Disciplinary;
Technically complicated; Race against the clock, in short:
every possible reason why people can rely on The
Reinforced Earth Company for their projects.
www.strand7.com
Did you Know?
Did you Know Items on the Web
Do you have a vague memory of a Did
you Know from a previous News.st7 that
you think would be useful to your project
but can’t quite remember the details.
Strand7 Did you Know items are now
available on our website
www.strand7.com/didyouknow.htm.
Updated with each issue, this means that
all those snippets of information can be
accessed from one centralised location.
Currently, this popular area is only accessible by a
winding two-lane road through North Kiama and is
subject to major traffic delays. In 1986, it was decided not
to proceed with the widening of the Princess Highway
thus resulting in the “North Kiama Bypass”. The first stage,
a concrete bridge across the Terragong Swamp is already
completed. RECO’s involvement in the second stage of
this $141 million dollar project comprises a 64-metre long
railway tunnel, 2400m² TerraClass wall area and another
1270m² of temporary wire wall to enable the diversion of
the Princess Highway through various stages of the
project.
The railway tunnel replaces the existing level crossing
through the Princess Highway. The design of this tunnel
was a major feat. The tunnel was restricted in every
imaginable way: the radius of the track, road width and
reduced level of the road, yet trains had to fit through,
without hitting the tunnel, determined by a clearance
envelope. 3D-drawings were created in order to
determine whether the train would have a trouble free
passage through the full length of the tunnel. Tunnel
elements had to be modified and truncated. Both ends of
the tunnel were curved inwards, creating more
challenges: How to cast these units? Would they fit and
fall into place during the installation of the tunnel? The
end result was 86 units in total, 70 full size units, 7 half
units, 9 wedges of which none are identical.
The rail overpass tunnel concept was developed by Brian
Bourne Bridge Engineer Pty Ltd - as the structural
consultant to Hughes Trueman Consulting Engineers.
Head contractor John Holland Pty Ltd awarded RECO the
contract for the detailed design and supply of the precast
concrete arch tunnel and the associated reinforced soil
walls. RECO engaged Interactive Design Services for
specialized design assistance for the arch and Cardno
MBK for independent design verification.
10
Making Finite Element Analysis Easier
Fig 5.1: Plan of the splayed end structure showing wheel load positions.
Fig 5.2: Deformed shape under M1600 load, 200 times magnification.
Initial analysis and design of the arch shape was carried
out using RECO’s in-house software. Preliminary design
of the splayed ends of the arch used 2D plain strain
analyses with various assumptions to model the 3D
effects, but for the final design, it was decided that a full
3D finite element analysis of the structure and the
Did you know?
Whiteboard
Determining the distances between two nodes is an
easy process in Strand7. Choose
Summary/Whiteboard and click the two nodes
you wish to determine the distance between.
Information on the node numbers and attributes will
be displayed. Along with this will be a straight line
distance between the two nodes and relative
distances, DX, DY and DZ.
If three nodes are selected then along with the
distance between the nodes, an angle between the
nodes will be given.
Information on other elements can also be shown
in the whiteboard, e.g. plate area, brick volume,
etc.
www.strand7.com
surrounding fill using Strand7 was necessary to give
adequate confidence that the behaviour of the structure
was modelled adequately.
Extensive use was made of the API both for generation of
the model, and for extraction of results. The 3D
coordinates of the centre-line of each arch panel was
generated in an Excel spreadsheet, and a series of beam
elements were created in the model using the API. The
beams were then extruded in two directions to form eight-
noded plate/shell elements, and the plate nodes were
extruded to form successively: offset beams, master/slave
links, and frictional contact elements. Having defined the
position of the soil/structure interface the generated
coordinates were read back into a spreadsheet using the
API, where the remainder of the model was generated.
The backfill was divided into eight layers, connected with
master-slave links, so that the actual backfill sequence
could be modelled in stages. Each fill layer was allocated
its own Freedom Case, with every node restrained against
movement, so that the fill layers could be fixed in space
until they were connected to the model.
The analysis was controlled through a spreadsheet, using
the API. The procedure for each layer was:
• Disable Freedom Case restraining the current layer
nodes
• Add in master/slave links to connect the layer to the
rest of the model
• Run non-linear analysis using the re-start file from the
previous layer
• Apply gravity loads, add compaction loads, remove
compaction loads
Finally design vehicle wheel loads are applied to the top
surface of the fill, and the final arch actions are extracted.
Fig 6.3: View of the North Kiama Tunnel under construction.
For more details of this project contact Paul Quach at Reinforced Earth
Pty Ltd, Tel: + 61 2 9910 9930; pquach@reco.com.au,
www.reco.com.au, and for details of the Strand7 analysis contact Doug
Jenkins at Interactive Design Services, Tel: + 61 2 9940 3414;
dougjenkins@interactiveds.com.au, www.interactiveds.com.au
11
Making Finite Element Analysis Easier

Training Exhibitions

T he Strand7 Training calendar for the second half of S trand7 continues to exhibit at a variety of exhibitions
2005 has been released. This calendar includes the usual across the globe. These are a great opportunity for us to
array of courses as well as dates for courses in Perth and meet not only potential new users but catch up with our
the UK. existing users.

2005 22-24 Nov 2005 Civils 2005

National Hall, Olympia,
London, UK – Stand J18
10-14 Oct Perth Strand7 Course
From November 22-24, Strand7 will have a booth at
8-11 Nov Introducing Strand7
Civils 2005, the UK’s number one civil engineering
15-16 Nov Structural Analysis showcase being held in the National Hall, Olympia,
17 Nov Automeshing London.
18 Nov Introduction to the Strand7 API
We are also planning to attend other exhibitions in the
14-18 Nov UK Strand7 Course
USA, so keep watching our website news pages to check
Following from the success of our Strand7 Melbourne if we are planning to visit a city near you
course, and based on feedback from users in Brisbane and
Perth, dedicated courses in these capital cities by our Strand7 User Profile
trainers was deemed a primary goal for the second half of
2005. The Brisbane course has just been presented with a Doug Jenkins
full capacity of participants attending on each day. The
Perth course is now beginning to fill up, so if you are Interactive Design Services
interested in attending one or more of the modules
(Strand7 Essentials, Nonlinear Analysis or Dynamic Q – How would you describe your business?
Analysis) please contact us.
A small civil engineering consultancy, specialising in the
In late June we also presented a Strand7 training course in analysis and design of bridges, buried structures,
Boston, USA for a group of local Strand7 users. As the retaining walls and related structures.
number of Strand7 users in the USA continues to grow,
Q – What types of jobs do you use Strand7 for?
we have more courses planned in various cities of the
USA in the near future. Analysis and design of structures, particularly soil
structure interaction and seismic analysis.
We would like to take this opportunity to thank all those
who attended not only our Melbourne, Brisbane and Q – How does Strand7 help you in what you design
Boston training courses, but also our in house training and analyse?
courses. Your positive and enthusiastic response during The use of plate and brick elements together with non-
classes makes teaching an easy and enjoyable linear materials models and re-start files allows soil-
experience. structure interaction problems to be analysed much
If anyone would like to register their interest in any of our more realistically than simplified approaches.
external courses please contact us at info@strand7.com. Q – What features of Strand7 do you find most
Details on Strand7 Training Course content can be found useful?
on our website at www.strand7.com/training.htm
I find the API very useful in setting up complex models
efficiently, and the post-processing and animation
facilities are very useful for visualising analysis results.
Head Office Contact Details
Strand7 Pty Ltd Doug Jenkins
Suite 1, Level 5 Interactive Design Services Pty Ltd
65 York Street Tel +61 (0)2 9940 3414
Sydney NSW 2000 Fax + 61 (0)2 9940 6330
AUSTRALIA Mobile +61 (0)414 854 402
Tel +61 2 9264 2977 Web: www.interactiveds.com.au
e-mail: dougjenkins@interactiveds.com.au
Fax +61 2 9264 2066
Email info@strand7.com
Web www.strand7.com

www.strand7.com 12
Strand7 Pty Limited

News.St7 Newsletter for Strand7 and Straus7 users

Strand7 is marketed as Straus7 in continental Europe

Issue 6, December 2008

Strand7 R2.4 Beta Available

2008 has been a big development year for Strand7, and as the year In this Issue
draws to a close we are very pleased to announce the availability of
Strand7 Release 2.4 Beta. This issue of News.St7 explains what you • Strand7 R2.4 Beta Availability 1
need to do to obtain a copy. We also highlight some of the new • Using Strand7 R2.4 Beta 2
features, such as new nonlinear options and the new attachment tools.
Other major features will be covered in upcoming issues of News.St7. • Strand7 R2.4 Features 2
As always you will find information on training and exhibitions together • R2.4 Nonlinear Solver 2
with new Did You Know items.
• Nonlinear Load Increments 10
If you have any feedback or suggestions regarding the content of News.St7 or would like to feature • Attachment Link 13
any of your projects then please email news.st7@strand7.com. If you would like to automatically
receive your copy of News.St7 directly by email simply send a blank email to newsletter- • Training 16
subscribe@strand7.com. All care is taken to ensure that information in News.St7 is accurate and up
to date at the time of publishing. However Strand7 Pty Ltd accepts no responsibility for inaccuracies • Exhibitions 16
in, or changes to, such information.

useful feedback and comments. The release is now ready

Strand7 R2.4 Beta Availability
S trand7 R2.4 is a major new release with over 140 new
features plus many enhancements to existing
functionality. Many of the new features are in direct
response to feedback and requests from users.
Figure 1.1: Strand7 R2.4 Beta Splashscreen.
Over the past few months, Strand7 R2.4 has been in beta
test mode with a growing list of beta testers providing
to enter the next stage of beta testing, and this involves
making it available to a much larger user base. If you are
currently up to date with your support and maintenance
you may be eligible to download this beta release.
Users will need to visit our website, www.strand7.com
and click on Strand7 Release 2.4 Beta in the latest news
section. You will need to enter your name, company
name and hardware lock name. You will then receive an
email from us with the download site location, user ID
and password.
If you are not sure what your hardware lock name is, go
to Strand7 and click Help/Licence Information. The
hardware lock name is given there.
We hope this next beta testing period will allow us to
obtain a wider range of feedback and therefore help us to
iron out any remaining issues before making a final
release.
Your feedback is welcomed and encouraged.
We request that you direct any questions, suggestions,
comments or problems that you encounter while using
the beta release, to support on +61 (2) 9264 2977 or
support@strand7.com

www.strand7.com 1
Making Finite Element Analysis Easier
Using Strand7 R2.4 Beta
R unning the setup.exe for Strand7 R2.4 Beta will
launch the installation wizard. If you are presently
running R2.3, you do not have to uninstall that release as
R2.4 can run in parallel on the same system.
Strand7 R2.4 Beta can be run as either a standalone or
network installation from either a parallel or USB
hardware lock. You will need to run St7Config.exe if you
are running in network mode.
Once the setup is complete you can run Strand7, and on
initial activation a new keycode may be required,
depending on your installation. Use email to send an
automatic email request or click txt to create a text file
and attach this to an email to key@strand7.com. Both
methods will return an automatic keycode.
After installing the beta, be sure to look through the
Strand7 R24 Book.pdf, which is located in the What’s
New folder of your installation. The pdf is a 300+ page
book with many detailed examples of the new features.
Strand7 R2.4 Beta is time limited, but regular updates will
be posted to the download site. For supported users
interested in continuing to use R2.4 Beta after expiry,
please check the download site for updates. At the time
of publishing this newsletter, the latest version is Strand7
R2.4 Beta9
Strand7 R2.4 Features
S trand7 R2.4 sees a number of significant changes and
additions to functionality. There are too many to list in
this newsletter, but a full list is available at
www.strand7.com - check out Whats New in Strand7
R24.pdf, or look for it in your What’s New folder.
Some of the new features include:
• Moving Load and Influence Analysis, with provision
for major codes such as AS5100, BS5400, AASHTO
LRFD and EN1992-2003.
• Construction Sequence Simulation, which allows for
the addition, removal and re-addition of elements at
any stage of the analysis, taking full account of
equilibrium changes as elements are added/removed,
with full support for material, geometric and contact
nonlinearity.
• Creep of metallic and non-metallic materials,
including concrete creep and shrinkage to most
national codes such as ACI and CEB-FIP.
• A new Quasi-Static solver that can be used to solve
problems in the time domain, while ignoring inertia
effects (e.g. creep problems).
• Modified Cam-Clay material model for modelling 2D
and 3D nonlinear drained and undrained soils.
www.strand7.com
• New fluid element for 2D and 3D analysis of fluid-
structure interaction.
• New element types including tapered beams, a new
significantly improved Tri3 element, which supports
elasto-plastic nonlinearity and geometric nonlinearity,
and new link types.
• New nonlinear fibre stress beam element of arbitrary
section shape, which may be used for material
nonlinear analysis (with or without geometric
nonlinearity) taking into account the interaction
between axial force and moments.
• New attribute types for all elements, including new
load attributes plus the ID attribute, which allows
elements to have a user-specified identification
number that is independent of the internal element
numbering.
• A new iterative solver based on the Preconditioned
Conjugate Gradient method (PCG), which is ideal for
the large brick models that increasingly arise from
automeshing CAD solid geometry.
• Improved convergence on most nonlinear models,
including material, geometric and contact
nonlinearity.
• New nonlinear substepping options, including the
very powerful arc-length method.
• STEP import/export, IGES export, binary STL import,
STL export, ANSYS import/export, SAP2000 import,
enhanced NASTRAN import/export.
• New pre-processing tools such as element cutting
tools.
• New selection tools.
• New geometry manipulation tools, including the
ability to create 2D and 3D geometry directly within
Strand7 for automeshing or exporting to other CAD
systems.
• Automatic attachment of dissimilar meshes, such as
assemblies from CAD.
• Many additional post-processing functions such as
user-defined contours, deviatoric and mean stresses,
SRSS and CQC combinations on all result types,
graphs in cylindrical systems and cutting planes.
• Concrete reinforcement post-processing on plate/shell
elements, including Wood-Armer moments and
allowance for membrane forces.
• Element Node Force results for facilitating the display
of free-body-diagrams on any section or subsection of
a model.
• Many new productivity enhancing features, such as
the ability to solve for multiple linear static freedom
cases in the one run and store all the results in the
same result file.
2
Making Finite Element Analysis Easier
• Enhanced configuration options that allow you to set
your own default values for virtually every option, and
the ability to easily return to the factory defaults.
• Significantly expanded API functionality, with over
900 new functions, including very powerful functions
that allow you to incorporate the Strand7 model
window directly inside your own application.
• Plus many more…
R2.4 Nonlinear Solver
T here are a number of significant changes in the
nonlinear solvers of Strand7 R2.4, not only to
accommodate new features but also to improve existing
functionality. These changes will not only expand the
types of analyses that can be performed in Strand7, but
also improve the convergence of many nonlinear models.
In this issue of News.St7 we will look at changes to the
automatic sub incrementation methods and the new
nonlinear beam element.
Nonlinear Beam
In previous releases of Strand7, modelling of nonlinear
material beams was achieved by assigning an axial stress
vs strain table and/or moment vs curvature tables in the
two principal axes. For these analyses the axial stress vs
strain and moment vs curvature were not coupled.
While this is still available, R2.4 offers a beam element
that simply requires the material stress-strain table and
cross-section definition, and Strand7 handles the rest. The
interaction between axial and bending is automatically
accommodated.
The following tutorial uses a pushover analysis on a shed
frame to explain the different settings and their influence
on a nonlinear material beam analysis.
Figure 2.1: Shed frame.
• Open the file Nonlinear Beam – Initial Model.st7.
www.strand7.com
The frame is 10m long by 5 m wide and has a maximum
height of 6m. The members are structural steel I beams,
with dimensions of 200mm×200mm×10mm.
Two load patches are placed on the roof of the shed to
allow a pressure load to be applied to all the beams. This
means that only the stiffness of the frame is considered
and that any stiffness provided by cladding, etc, is
ignored.
A vertical pressure of Y = -5kPa is assigned to each of the
load patch plates.
• Solve the model using the Linear Static solver,
ensuring that Node Reactions are calculated.
Figure 2.2: Displacement of frame under vertical pressure load.
• Using the peek tool, find the vertical reaction at the
base of the legs.
You may have noticed that when the results were
displayed for the frame the patch plates were no longer
visible. This is because of settings in the new
Results/Options. Here you can control how some
elements are shown during post-processing: e.g.
show/hide patch plates, hide inactive contact elements,
hide inactive construction sequence elements or show
them as wireframe, etc.
Figure 2.3: Results/Options, Show/Hide tab.
An assessment of the stability of the frame structure can
be made by undertaking a pushover analysis. This method
determines the amount of horizontal load the frame can
withstand before collapse, as a ratio of the vertical load.
3
Making Finite Element Analysis Easier

• Create a second load case in the model called
Horizontal Load.
• Select the four nodes at the top of the columns and
apply a horizontal load equal to the vertical reaction,
i.e. X = 72886.9N.
• Select the Nonlinear tab.
• Select Elastic-Plastic from the Material dropdown
and Isotropic from the Hardening dropdown.
• Select the stress vs strain table from the Stress vs
Strain dropdown.
If instead you wish to perform an uncoupled nonlinear
material analysis, e.g. you wish to assign a specific
moment vs curvature table to one of the principal axes
you can do so by setting Use Moment vs Curvature
Tables. This will activate the axial stress vs strain table
option as well the moment vs curvature options in both
principal axis directions and effectively sets the beam to
behave as in previous Strand7 releases.

Figure 2.4: Horizontal nodal loads.

For this example we are interested in the collapse of the

structure due to plasticity of the material. To determine
this we need to include a stress vs strain table for the
steel. This steel has a yield strength of 300MPa and an
ultimate strength of 350 MPa.
• Choose Tables/Stress vs Strain and create a trilinear
stress vs strain table with the following points.

Strain Stress (MPa)

0.0 0

0.0015 300
Figure 2.6: Beam property, Nonlinear tab.

0.0022 350
• Select the Section tab (Figure 2.7).
0.006 350 Integration Points may be set between 4 and 10. This
defines the number of points along the length of the beam
where the bending stiffness is fully integrated from the
stress-strain relationship over the cross section. By default
the setting is 5 and this should be sufficient for the
majority of analyses.

Figure 2.5: Steel stress vs strain table. Figure 2.7: Beam property, section tab.

• Choose Property/Beam.

www.strand7.com 4
Making Finite Element Analysis Easier

• From the Geometry tab select Discretisation • Select the Results tab and ensure Beam Fibre Stress
(Nonlinear) (Figure 2.8). is calculated.
• Click Solve.
• Open the results and create a graph Vs Result Case
and name this Lateral Displacement.
• Plot a graph of nodal DX displacement.
• The Position selected should be the node at the top of
one column, relative to the node at the top of another
column. These two nodes should be on columns with
the same Z and Y coordinates.
• The X axis should be vs the Horizontal Load applied
to the structure (Figure 2.9).

Figure 2.8: Beam property, Geometry tab.

The discretisation specifies the number of cells the beam

cross section is divided into for the analysis. Each cell is
monitored independently with respect to its position on
the stress-strain table and the integration of all the cells,
over the length, determines the effective beam properties.
For a standard cross section such as this I Beam, Nominal
Divisions by default is 10, which sets the number of
divisions in the both the x and y directions. The value is
taken as the number of divisions along the longest edge
with the number of divisions in the other direction set to
make the cells roughly square. Alternatively, if Specified
Divisions is set, the number of divisions in x and y can be
directly entered.
Solid and hollow circles as well as BXS sections use a
variation on this Nominal Divisions. For further Figure 2.9: Graph vs Result Case, Position tab.
information on these section types see the Help file.
• Add to the graph the total displacement at node 8.
When the number of integration points and/or divisions
are increased the solution time will also increase.
The nonlinear solver is used for this analysis due to the
nonlinear material considerations.
• Choose Solver/Nonlinear Static.
• Set Nonlinear Material but clear Nonlinear
Geometry.

• Click Load increments and set them up as follows,

Vertical Load – Constant factor of 1.0 for 25
increments.
Horizontal Load – Factor varies from 0.0 to 1.2 over
the 25 increments.
• Under the Solver/Defaults tab select Nonlinear and
Figure 2.10: Graph of nodal DX displacement (relative and total).
set the maximum number of iterations to 100.
• Under the Sub-Steps option set Displacement Control You can see from figure 2.10 that initially the relative
(Arc Length) with a Maximum Displacement displacement between the two legs is constant but that at
Increment (Ratio) of 1.0×10-3. Also set Save Sub- around 60% of the horizontal load the distance between
Increments. the columns begins to increase.
• Contour beam fibre stress for the frame.

www.strand7.com 5
Making Finite Element Analysis Easier
Figure 2.11: Beam fibre stress contour – Increment 22.
Figure 2.11 shows the plastic hinges that are eventually
formed at the top and bottom of the columns; stresses
have reached the maximum ultimate values of +350 MPa
and –350 MPa.
The peek tool can also be used to investigate the yielding
of the beam.
Figure 2.12: Peek tool showing beam fibre stress in the yielded beam.
Prior to yielding, all beam stress types on an elasto-plastic
nonlinear beam can be extracted. However once a beam
yields, it is no longer possible to calculate fibre stress as
simply a linear summation of axial stress and bending
stress. Consequently certain stress contours will no
longer be available for yielded elements. For example, if
an attempt is made to plot Bending Stress (as opposed to
Fibre Stress) yielded beams will be displayed as
wireframe, as shown in figure 2.13. Similar explanation
applies to the axial stress plot for a nonlinear beam.
However, if axial stress contours are requested on a beam
that has yielded, Strand7 simply contours the total axial
www.strand7.com
force divided by the cross section area, as an indicative
value.
Figure 2.13: Bending stress Plane 2 contour .
Sub-incrementation
I n Strand7 R2.3 sub-incrementation was handled via
load scaling, i.e. if the solution didn’t converge within the
specified number of iterations, or the displacement
increments were too large, the amount of load was
reduced and the solution proceeded. This is still available
in R2.4, but now, two more sub-incrementation
procedures have been introduced, Displacement Scaling
and Displacement Control (Arc Length). The different
sub-incrementation techniques can be set via the solver
dialog Defaults, under Sub-Steps. This is shown in
Figure 2.14.
Figure 2.14: Nonlinear static solver Defaults – Sub-Steps.
Load Scaling
As was the case in previous releases of Strand7, load
scaling scales the amount of applied load should the
solution fail to converge. The amount of load applied in
the increment is reduced and the iteration process starts
again. The load continues to be reduced until either the
6
Making Finite Element Analysis Easier
solution converges or the Minimum Reduction Factor is
reached, before increasing it again towards the value
requested by the user.
Load Scaling can be triggered by one of the following,
• The solution does not converge in the set number of
iterations.
• The matrix is singular.
• The Maximum Rotation Increment or Maximum
Displacement Increment are exceeded.
Displacement Scaling
Displacement scaling is similar to load scaling except
when this is triggered, instead of scaling the externally
applied loads, the initially calculated displacements are
scaled. The solver scales back the calculated
displacements when determining the internal state of
stress and updated model stiffness. This allows the
structure to develop the necessary stiffness and stability
(e.g. due to membrane loads) to eventually equilibrate
with the externally applied loads. The displacement is
scaled such that at any iteration it is no larger than the
Maximum Rotation Increment or the Maximum
Displacement Increment.
Did you Know?
Nonlinear Convergence Graph
Convergence of the nonlinear solvers is
based on the displacement and force norms
for the model for each iteration. Within R2.4
you can now click in the solver panel to
display a graph of the displacement and force
norms for each iteration.
This means that it is easier to track the path
of convergence of the model.
Displacement Control (Arc Length)
The addition of arc length is the most significant addition
to the sub-stepping options. This will enable models that
previously diverged after a certain amount of load, to
proceed further down the solution path. In particular this
www.strand7.com
sub-stepping procedure is useful for buckling and post
buckling analysis.
The procedure searches along the equilibrium path
allowing total movement, in any increment, up to a
specified amount; this amount is the nominal arc length
value. The Maximum Rotation Increment and the
Maximum Displacement Increment (Ratio) settings
directly control the nominal arc length value, which is
calculated as a type of norm of the displacement vector.
Arc length sub-incrementation is then triggered when the
displacement/rotation increment generates an
instantaneous arc length value that exceeds the nominal
arc length value.
In a single degree of freedom system, for example, a
single spring along the X axis being extended, it is easy to
visualise the nominal arc length: in this case it is simply
the X displacement of the end of the spring. The arc
length procedure therefore scales the effective load on the
spring such that each increment produces an X
displacement no bigger than the nominated arc length. In
a general case, it is not trivial to determine which degree
of freedom is controlling the response (it can change as
the structure is loaded and deforms), therefore the
nominal arc length is calculated based on the complete
displacement increment vector.
As mentioned above, the Maximum Rotation Increment
and the Maximum Displacement Increment (Ratio)
control the nominal arc length. The rotation is simply a
nodal rotation, whereas the displacement increment,
being a ratio, is converted to an actual displacement value
by multiplying the ratio by the largest dimension of the
model, parallel to the global axes. The nominal arc
length is then the norm of the displacement vector scaled
such that the maximum nodal rotation component
increment equals the Maximum Rotation Increment, or
the maximum displacement component increment equals
the Maximum Displacement Increment, whichever occurs
first.
Sub-Incrementation – Example 1
This first sub-incrementation example will look at a cable
net structure. The structure is 30m long and 10m high. It
consists of relatively stiff I beams on the sides and top and
very flexible circular solid sections forming the cables.
Figure 2.15: Cable net model.
7
Making Finite Element Analysis Easier

The nodes at the base are pinned and the structure is

subjected to self weight.
• Open the model Cable Net.st7.
• Choose Solver/Nonlinear Static.
• Set Nonlinear Geometry.
There is a single nonlinear load increment in this model
with a factor of 1 for self weight.
• Select the Defaults tab and click Sub-Steps.
• Select Load Scaling and click Solve.

Figure 2.17: Cable net displaced shape.

Figure 2.17 shows an exaggerated plot of displacement

for the cable net. You can see from this that the cables
have tended to droop under the gravity load.
Displacement scaling is useful for structures that will
develop stiffness under loading but initially may be
unstable.

Did you Know?

View Load Factors
With the release of Strand7 R2.4 comes the
ability to view the applied loads for a
specific load case combination, nonlinear
load increment, quasi-static or transient
dynamic time step. Results/View Load
Figure 2.16: Nonlinear solver Defaults tab.
Factors will show the load factors applied
for each load case for the displayed result
The solver terminates due to divergence. This is because case.
the load is causing large deflections in the structure which
cannot be resolved. In reality, as the cable begins to Sub-Incrementation – Example 2
deflect, some lateral stiffness would develop reducing the
overall amount of deflection. For this example we will look at a
• Choose Solver/Nonlinear Static and select the rectangular hollow section
Defaults tab again.
modelled using plate/shell
elements.
• Click Sub-Steps.
The structure is 197mm ×
• Select Displacement Scaling. 236.4mm, 1000mm tall and has a
thickness of 6mm.
• Enter a Maximum Rotation Increment of 30
degrees, and a Maximum Displacement Increment Only half the structure is modelled
of 0.01. with the appropriate restraints
applied along the symmetry plane
This means that the structure can only move 0.01 ×
and the base fully fixed.
maximum model size in an iteration. The displacement
scaling routine will scale back the displacements when A moment of 1 × 108 Nmm is
this is exceeded, and determine the load required for this applied across the top of the
amount of displacement and then proceed to iterate to column.
balance the internal and external loads. This allows the
• Open the model Rectangular
structure to develop some lateral stiffness due to
Hollow Section.st7.
membrane effects and to stabilise under the load.
• Set Save Sub-Increments and click Solve. • Choose Solver/Nonlinear Static.
This time the displacement scaling routine allows the • Set Nonlinear Geometry.
solver to scale back the displacements and the solution to
proceed until 100% of the load is applied. • Click Load Increments.
There are 10 increments factoring from 1 to 10.

www.strand7.com 8
Making Finite Element Analysis Easier

Figure 2.18: Nonlinear load increment table.

• Select the Defaults tab.

• Click Sub-steps.
In this first section we will look at using Load Scaling as
the sub-stepping option.
Figure 2.20: RHS lateral deflection (absolute 1).
• Enter a Minimum Reduction Factor of 0.025.
A graph of the lateral (DX) deflection vs load halfway up
• Set Save Sub-Increments.
the column, shown in figure 2.21, also shows this
• Click Solve. buckling occurring. This is indicated by the sharp change
in gradient of the slope.
Note: The iteration limit has been changed to 75 for this
model.

Figure 2.21: RHS load stepping lateral deflection graph.

However it could be useful to be able to track with more

Figure 2.19: Sub-Steps dialog.
The model begins to sub-step at increment 5 and
terminates due to divergence of the solution at the start of
increment 6.
Figure 2.20 shows a displacement plot of absolute 1 for
approximately 5% of increment 6. This clearly shows the
structure is beginning to buckle under the current load.
detail exactly how the structure is buckling as well as to
look further into the post-buckling behaviour. The
nonlinear displacement control (Arc Length) sub-stepping
procedure is ideally suited to this type of problem.
• Choose Solver/Nonlinear Static and select the
Defaults tab.
• Click Sub-Steps.
• Select Displacement Control (Arc Length).
• Click Solve.

www.strand7.com 9
Making Finite Element Analysis Easier

Figure 2.24: RHS lateral deflection, Increment 6 (Sub-increment 38).

Figure 2.22: Sub-Steps dialog.

Again the solver begins to sub-step but now the sub-

stepping routine is arc length. The solver no longer
terminates due to divergence but continues to update the
load. Figures 2.23 to 2.25 show different stages of the
buckling of the column. Initially the shape is very similar
to the shape found using the load stepping method,
however 2.24 and 2.25 show that the post buckling
behaviour is now being captured. Figure 2.25 shows the
column completely failing. Figure 2.25: RHS lateral deflection, Increment 6 (Sub increment 100)

The same graph created for the load stepping procedure

can now be viewed for the arc length results. This clearly
shows both the ultimate moment carrying capacity of the
column and its response beyond that point.

Figure 2.23: RHS lateral deflection, Increment 6 (Sub-increment 1).

Figure 2.26: RHS lateral deflection vs load graph.

The availability of displacement control (arc length) opens

up new possibilities in the analysis of nonlinear models
allowing the buckling and post buckling behaviour of
models to be tracked in great detail

www.strand7.com 10
Making Finite Element Analysis Easier
Nonlinear Load Increments
W hile the application of nonlinear load increments
remains the same in Strand7 R2.4, creating the increment
table has become a little bit different, now incorporating
stages of a construction sequence and an automatic
creation tool. This article will revisit the application of
loads from the load increment table and then discuss how
the new dialog differs from previous releases.
Revisiting Nonlinear Load Increments
The factors applied to a load case in the nonlinear load
increment table are factors applied to that load. It is
important to remember when setting up your nonlinear
load increments that the factors applied to each increment
are independent of the previous increments, not
cumulative. This has not changed from previous release
versions of Strand7.
Figure 3.1: Nonlinear load increment table with two increments.
For example, Figure 3.1 shows Increment 1 as applying a
factor of 1.3 to X = 130N (load case 1). This means that
for this increment 1.3 × (X=130N) is applied giving a
total applied load of X=169N. In Increment 2, the factors
are 1.0 for X = 130N (LC1) and 1.0 to Y = 25N (LC2).
This means that the load applied to the structure would be
a combination of X = 130N and Y = 25N at the same
time. This type of nonlinear load increment table is
common for modeling different load combinations within
the nonlinear solver.
Figure 3.2: Nonlinear load increment table with five increments.
www.strand7.com
Fig 3.2 shows a different nonlinear load increment table.
In this example gravity is included and kept constant
while a second load is gradually added. So in Increment
1, only gravity is included at 100% (a factor of 1 is
assigned to Gravity). In the second increment gravity
remains at 100% but now 25% or 0.25 times load case 2
is applied. So we have gravity and X = 15N applied to
the model. The increments then continue to keep gravity
constant while steadily increasing the second load. This
method of setting up the nonlinear load increment table is
usually used for one of two scenarios. The first is when
you are looking at situations such as buckling when one
load is kept constant but another load is steadily increased
until reaching the buckling load factor. The other time
this setup might be used is when a model is very
nonlinear and only small amounts of load are applied in a
single step to aid solution convergence.
Did you Know?
Anchoring Node Coordinates
If you wish to create nodes in a specific plane or line, you can
anchor one or more nodal coordinates and then enter the
remaining values.
In the Create/Node dialog enter
the coordinates that are to stay
the same, e.g Z = 250. Then
clear the Z checkbox. Now every
node that is created when an X
and or Y coordinate is entered
will be at Z = 250.
Both Fig 3.1 and 3.2 have the freedom case factors set to
zero. In this case both freedom cases include only fully
fixed restraints. This means that regardless of the factor
assigned to the freedom case the restraint will remain
fixed: 0 × fully fixed = 1 × fully fixed. The factor assigned
to a freedom case in the increment table only becomes
important when either an enforced displacement or shrink
link is present in a model.
Figure 3.3: Nonlinear load increment table with three increments.
Fig 3.3 shows an increment table for a model containing
an enforced displacement of X = 5mm. For the first
11
Making Finite Element Analysis Easier

increment the factor is 0, so this will simulate a fixed X

restraint in the model, or 0 × 5mm = 0mm. For the
second increment the factor is 0.5, giving a total X
enforced displacement of 2.5mm in the model. It is not
until the final increment when 1 is applied as the factor
that 5mm is enforced in the model. If instead of enforced
displacements in the model, shrink links existed, the
freedom case factors would still apply. In this situation
increment 1 would create no change in the length of the
shrink link. Increment 2 however would cause the shrink
link to close by 50%, or half its length. At increment 3,
Figure 3.5: Nonlinear load increment dialog with five increments.
both ends of the shrink link come together.
Automatic Load Increments The automatic load increments can also be used for a
more complicated loading scenario. If your model had
In Strand7 R2.4 while load increments can still be created
two load cases but you wanted one to be at 100% for all
manually it is now possible to automatically create load
increments but the second load case to be applied
gradually this can be handled in two steps using the
increments. By clicking the dialog box in Fig 3.4 is
automatic load increment generator.
displayed.

Figure 3.6: Automatic load increment dialog.

Figure 3.4: Automatic load increment dialog.

To setup the increments so that the first load case is at

Here you can specify the Start Factor, End Factor, Start
100% for all increments, only load case 1 should be
Increment and End Increment for the load and/or freedom
checked and the Start and End Factors should be 1.0. The
cases. For example, If you require a load case to linearly
Start Increment should be 1 and the End Increment, 5.
increase the loading from 0 to 100% in five increments
This will produce the nonlinear load increment table
you would set up the automatic load increments as
shown in Fig 3.7.
follows:
• Start Factor = 0
• End Factor = 1
• Start Increment = 1
• End Increment = 5
This is shown in Fig 3.4 with the resulting nonlinear load
increment table shown in Fig 3.5.

Figure 3.7: Nonlinear load increment table.

The second stage requires us to incrementally increase

load case 2 from 0.0 to 1.0 over the five increments. The
automatic load increment tool is again used but this time
load case 2 is set instead of load case 1. The Start Factor is
now 0.0 but the End Factor is still 1.0. Start and End
Increments remain the same. This setup is shown in Fig

www.strand7.com 12
Making Finite Element Analysis Easier
3.8 with the resulting nonlinear load increment table
shown in Fig 3.9.
Figure 3.8: Automatic load increment dialog.
Figure 3.9: Nonlinear load increment table.
Increments do not necessarily need to be increasing when
using the automatic increment tool. Start and End Factors
can be entered such that the loading is decreasing. Not
only can it be one or the other, the functionality of the
automatic increment tool allows for a combination of
loading and unloading to be accomplished.
Staged Increments
As part of the new features of R2.4, tools are available for
easy staged construction sequence analysis. This means
that the nonlinear load increment table now
accommodates stages of an analysis. By clicking staged
analysis nonlinear load increments can be assigned
for each of the different stages. This can consist of one to
multiple increments in the one stage.
Figure 3.13: Staged nonlinear load increment table.
www.strand7.com
More details on staged construction sequence will be
covered in a future issue of News.st7, or for further
information see Strand7 R24 Book.pdf.
Other Options
Two other buttons have been introduced to the nonlinear
load increment table that will improve efficiency.
The first is Clear All Values . When this is used all
values in the load increment table become zero.
The second is Reset Increment Totals . This allows
for the number of increments to be reset to a user
specified number, e.g. 20 increments can be reset back to
one in a single operation. This tool can be used on single
or multi-stage analyses
Attachment Link
T he attachment link was developed to allow
incompatible meshes to be connected together and
deform as if they are compatible. This will make
modelling contact scenarios between two dissimilar
meshes or joining the separate components of an
assembly a more streamlined process. This article is not
written to discuss how to use the attachment link, for
information on this please see the Strand7 R24 Book.pdf.
Instead this article will focus on the different types of
attachment available and how the results can vary due to
this as will as give a brief example of using the new link.
Figure 4.1:Attachment attribute dialog.
The three types available are Rigid, Flexible and Direct.
While each of these types will include the creation of an
attachment link, what this link will attach to varies.
13
Making Finite Element Analysis Easier

• Rigid
Did you Know?
Background Image
A background image can be added to your Strand7
model window to allow you to include your
company logo or perhaps a theme that matches
your structure to “jazz up” the presentation of your
model to your client.
2.0m
1.25m
Figure 4.2: Rigid attachment.

When rigid is selected, a rigid link 1.55m

is created from the node to the
point of attachment. The
attachment link is then connected Figure 4.5: Model configuration.
between the rigid link and the
other element. With the model configuration
Choose View/Options and select the jpeg or bmp shown in Figure 4.5 we would
• Flexible image you wish to use as the background. This will expect a moment at the fixed node
be displayed in the model window. The model can of
still be manipulated within Strand7 as per usual.
MZ = 5 N × (1.25m + 0.3m ) = 7.75 Nm

Three models with the configuration shown in Fig 4.5

have been created. The first uses a rigid attachment, the
second a flexible attachment, and the third a direct
attachment. Moment results at the fixed node are:
Rigid: MZ = 7.75Nm

Figure 4.3: Flexible attachment. Flexible: MZ = 7.75Nm

When flexible is selected, a new beam element (of the
property type of your choice) is created from the node to
the point of attachment. The attachment link is then
connected between the new beam and the other element.
This is basically the same as rigid except that the rigid link
becomes a beam element.
• Direct
Figure 4.4: Direct attachment.
When direct is selected, the attachment is assumed to be
directly between the node and the point of attachment
regardless of any distance between them.
Moment Handling
The main difference between the three attachment types
can be seen when looking at the results.
Direct: MZ = 6.25Nm
The results show that while the rigid and flexible
attachments give the expected moment, direct does not.
In fact the moment value calculated for direct ignores the
gap between the two components being attached
(MZ = 5N×1.25m = 6.25Nm). This is as you would
expect as the definition of the direct attachment is that the
two components are directly attached regardless of
distance. This result is completely analogous to the
Master-Slave link results reported in News.St7 Issue 3.
Conclusion
The above discussion is important to remember when
choosing which attachment type is appropriate for your
model. If the gap between the two attached components
is relatively small, the difference between the three types
is minimal, however as the gap grows so does the
difference. Whether any gap truly exists and therefore
impacts the moment calculation should be considered
carefully.
Attachment Link Application
Figure 4.6 shows the geometry for a mechanism that was
imported into Strand7. This geometry will be used to give
a brief outline on how the new attachment link works.

www.strand7.com 14
Making Finite Element Analysis Easier
Figure 4.6: Assembly geometry with attributes.
The geometry consists of two links, a shaft and a bearing
block. Loads and restraints are assigned directly to the
geometry; fixed restraints at the base of the bearing block
and inside the small hole of one of the links, and a
pressure load normal to a link side. We can also assign
the attachment attribute directly to the geometry. As the
connection between the shaft and the links is to be a
welded connection, a rigid connection for the attachment
is assigned here. For the connection between the shaft
and bearing block we need to simulate a sliding
connection and so this is assigned as flexible.
Figure 4.7: Section of incompatible mesh in assembly.
Fig 4.7 shows the assembly after it has been solid
meshed. It can easily be seen that there are areas of mesh
incompatibility.
Figure 4.8: Brick face attachment attribute on areas to be attached.
Once the meshing is completed and the attachment
attribute is assigned to the elements the attachment links
are created via Tools/Attach Parts.
www.strand7.com
Figure 4.8: Flexible attachment links (beam elements created).
Figure 4.8 shows the automatic creation of the attachment
links in the model. For rigid attachments, rigid links are
created, while for flexible attachments, beam elements are
created. For this example, the beam elements are set to
normal point contacts.
Welded
connection Sliding
connection
Figure 4.9: Theta displacement of the assembly.
Solving the model in the nonlinear solver produces the
theta displacement shown in figure 4.9. This shows that
the displacement across the incompatible meshes is
continuous, as would occur in a welded joint. At the
interface between the shaft and the bearing block, the
theta displacement is not continuous as the shaft is free to
rotate relative to the bearing block
15
Making Finite Element Analysis Easier

In October we exhibited at SNAME – Society of Naval

Training
Architects and Marine Engineers conference and expo in

T his year we have conducted a number of training

Houston, Texas, together with Beaufort Analysis Inc, who
are the Strand7 distributors in the USA.
courses in the use of Strand7 here in Sydney, throughout In November we also exhibited at Civils UK 2008, the
Australia and internationally. Both our introductory and premier Civil Engineering exhibition in the UK. Now in its
our advanced courses, covering nonlinear, dynamics and tenth year, it brings together the whole industry to
structures, continue to be popular with new and existing educate, learn, promote and network, covering all aspects
users alike. of civil engineering.
Our final scheduled course for the year has just been
completed in the UK. This course was a five day course
similar to the courses we run interstate and in the USA.
Although our course schedule for 2009 is yet to be
finalised, we plan to run our usual suite of courses as well
as more training in the UK, USA and interstate. When the
training schedule becomes available it will be posted on
our website, however if you wish to register to have the
schedule sent to you please send an email to
info@strand7.com.
If our standard courses don’t quite match your
requirements, you may also like to consider customised
training courses. These can be conducted at our offices in
Sydney or in your own offices. They can be for one or for
Figure 5.2: Civils 2008 Exhibition November 2008 - UK
a group of engineers, and from a few hours to five days in
duration. They are tailored to your needs. If you would
like more information about customised Strand7 training In 2009 we plan to exhibit at a range of exhibitions (a full
list will be available on our website). So far we can
for your company, please contact us confirm our attendance at NASCC 2009 (North American
Steel Construction Conference) and OTC 2009 (Offshore
Exhibitions Technology Conference).

S trand7 participated in a number of exhibitions and 1–4 April 2009 NASCC Booth 728 Phoenix, USA

conferences in 2008, in the USA, the UK and Korea. 4-7 May 2009 OTC Booth 8377 Houston, USA

These exhibitions are a great opportunity for us to not

only meet prospective new Strand7 users, but also to
meet our existing users in a relaxed environment, and to
find out about interesting engineering projects and the
application of FEA to these.
F inally all the staff here at
Strand7 would like to take this
In May we participated at the 8th World Congress on opportunity to wish all our Strand7
Railway Research in Seoul Korea, together with CNG- users a very merry Christmas and a
Softek, who are the Strand7 distributors in Korea. safe and happy 2009. We thank
you for your continued support
and hope that you will enjoy using
Strand7 R2.4.

Head Office
Strand7 Pty Ltd
Suite 1, Level 5
65 York Street
Sydney NSW 2000
AUSTRALIA
Tel +61 2 9264 2977
Fax +61 2 9264 2066
Email info@strand7.com
Figure 5.1: SNAME exhibition October 2008 – USA. Web www.strand7.com

www.strand7.com 16
Strand7

News.St7 Newsletter for Strand7 and Straus7 users

Strand7 is marketed as Straus7 in continental Europe

Issue 7, February 2010

Strand7 R2.4 Released

As most of our users are aware, Strand7 R2.4 was released on December In this Issue
7th 2009. We are very grateful to our users who participated in the
Strand7 R2.4 Beta Program during 2009; their invaluable feedback  Strand7 R2.4 Features 1
helped us achieve this goal.  Multi-Point Link 2
In this issue of News.St7 you will find detailed articles on the new  Concrete Creep and Shrinkage 6
concrete creep material model, the multi-point link and the new
geometry tools. We also illustrate the new friction options of the Point-  Modelling Tip - Concrete
Contact element via a benchmark. Future editions of News.St7 will Reinforcement 11
cover more of the new features in Strand7 R2.4, but if you have a  API Functions 12
suggestion about what you would like to see next, please let us know.
 Geometry Tools 12
As always, this issue of News.St7 contains information on training and
exhibitions together with new Did You Know items.  Benchmark 15

If you have any feedback or suggestions regarding the content of News.St7 or you would like to  Training and Exhibitions 16
contribute a case study showing how you use Strand7 then please email news.st7@strand7.com. If
you would like to receive your copy of News.St7 automatically by email, simply send a blank email to newsletter-subscribe@strand7.com. All care is taken
to ensure that information in News.St7 is accurate and up to date at the time of publishing. However Strand7 Pty Ltd accepts no responsibility for
inaccuracies in, or changes to, such information.

and to provide results as effective rather than total stress

Strand7 R2.4 Features
(effective stress = total stress minus pore pressure). The

S
ome of the new features in Strand7 were mentioned in
Issue 6 of News.st7. A few more features have been
added in the past year and here are some of the
highlights.
One important development is the new set of geometry
tools including the ability to create geometry within
Strand7. More about these later.
We have also added new material types including the
ability to use soil materials in brick elements. There are
now several new soil material models including Modified
Cam-Clay Soil, Mohr-Coulomb Soil, Drucker-Prager Soil,
and Linear Elastic Soil. The Cam-Clay Soil material is a
totally new development available for plane strain,
axisymmetric and 3D analyses. The Mohr-Coulomb and
Drucker-Prager Soils are extensions to the previously
available Mohr-Coulomb and Drucker-Prager elasto-
plastic material models, but have been enhanced with the
ability to consider fluid (pore) pressure and in-situ stress,
Linear Elastic soil material also considers pore pressure
and in-situ stress, and produces results as effective stress,
but it remains a linear elastic material.
The new Fluid material model is available for 3D brick
and 2D plane strain/axisymmetric elements. It can be
used for applications such as fluid-filled tanks and other
containers.
Creep of metallic and non-metallic materials can now be
simulated. A special feature on this topic, namely creep
and shrinkage of concrete, is included in this issue of
News.St7.
Another major development has been the enhancement of
the Strand7 Application Programming Interface (API). We
have added hundreds of new functions allowing users to
customise and automate modelling and post-processing
procedures, including the ability to embed a Strand7
graphical model window directly inside another
application. A summary of the improvements can be
found in this issue of News.St7.

www.strand7.com 1
Making Finite Element Analysis Easier

New Units have been added to Strand7, including the Strand7 R24 for R23 Users.pdf and What’s New in
calorie energy unit, stress in ksi, and time units for creep Strand7 R24.pdf (they are also available for download at
material data and for use in the transient solvers www.strand7.com)
(milliseconds, seconds, minutes, hours and days). Scaling
of units now works on all data including Tables and
anisotropic material properties. In addition, changes to
the units no longer reset the Undo list, which means that Multi-Point Link
you can now Undo beyond a units change.
New display options
including Ignore
O ne of the new link types introduced in Strand7 R2.4
Cavity Loop on Face is the multi-point link. There are two versions of this link:
free edge display 1. User-Defined Multi-Point Link;
helps you identify
your model’s 2. Interpolated Multi-Point Link.
unintentional free The User-Defined version can be used to generate
edges. arbitrary equations that constrain any number of degrees
The behaviour of of freedom to behave in a user-specified way. For
Solid View Dynamic Rotation has been improved to example, the equation DX(2)=0.5*DX(4) + 0.5*DX(6)
maintain frame rate whilst moving the display. constrains DX at node 2 to be the average of DX at nodes
4 and 6. Here the user is free to specify all the degrees of
There have also been many improvements to the way freedom and constants that make up the equation.
results are displayed. Free body diagrams on any part of a The Interpolated version requires a slightly different set of
model can now be plotted using Element Node Reaction inputs. These are:
Forces; you can also contour the reaction forces to help
you diagnose the 1. The nodes to be constrained - one of these nodes
cause of non- Did You Know? is designated as the slave node whilst all the
convergence in others in the link are designated as master nodes,
Nonlinear Numbered Backups or the cluster of master nodes.
analyses. Contact In R2.4 you can set the file
preferences to allow Strand7 to save 2. Whether the link is to act on the translational
pressure at Face degrees of freedom, the rotational degrees of
your model with a different numbered
Supports can be backup every time you click save. freedom or both translational and rotational
plotted directly, degrees of freedom.
without the need This should give users peace of
mind, knowing that model changes With these inputs, the equations that relate the master
to extrapolate can be retraced and, more
element stresses. importantly, recovered. degrees of freedom (DoF) to the slave DoF are
This is particularly automatically generated by Strand7. To generate these
useful in equations, the link seeks to apply the average motion of
nonlinear analysis the cluster of master nodes to the slave node; the average
when the face translation and average rigid body rotation of the cluster
supports are of are then applied at the slave node. To achieve this, a
the Compression- least squares formulation is used based on the following
Only type: a characteristics:
contour of the 1. The translational DoF at the slave node are a
Face Support function of only the translational DoF at the
stress clearly master nodes.
shows the section
of the model that 2. The rotational DoF at the slave node are a
is in contact. function of both the translational DoF at the
master nodes and the geometry of the link (i.e.
New element attributes include the ability to define the coordinates of the master nodes); there is no
individual node initial velocities for transient dynamic direct coupling of the rotational DoF at the slave
analyses. Accelerations can now also be applied to the node with the rotational DoF at the master nodes.
nodes to generate inertial loads that are completely
arbitrary; this augments the previously available linear The general form for the system of equations is given by:
and angular accelerations, which generate the same
accelerations over the whole mesh. Non-structural
masses can be offset from the node or element to which  LHS1   x1 
they have been applied, taking into account the moments  LHS  x 
generated by the acceleration field without the need for  2
x  RHS1 RHS 2  RHS N  2 
   i   
rigid links.    
LHS N  x N 
Full documentation of the new features and changes to
old ones are included in the Strand7 R2.4 installation. i.e. Lx i  Ry
The files can be found in the What’s New directory:

www.strand7.com 2
Making Finite Element Analysis Easier

where x i is the vector of slave DoF and x1, 2, N are the freedom case type is set to Free Body Inertia Relief, so it is
vectors of master DoF. This over-specified system can be not necessary to apply restraints to the model.
solved via a least squares approach to give the slave DoF: Each of the three applied nodal masses model different
payload behaviour, so the link type we choose is
L Lxi  L Ry
T T
important to model these differences realistically. The
100kg mass in the instrument bay represents a loaded
xi  LTL  
1 T
L Ry instrument tray, bolted to 8 lugs which are welded to the
walls of the rocket. This tray would add some stiffness to
Depending on the link option Translations, Rotations or the structure; thus, using rigid links to distribute the
Both, the xi vector will contain translational and/or inertial effects of the instruments is appropriate but comes
rotational degrees of freedom. with the assumption that the tray is much stiffer than the
rocket walls and lugs. Alternatively the tray could be
This link can be used to distribute load to multiple nodes
explicitly modelled using shell elements with multi-point
without influencing the stiffness of the structure.
links distributing the mass to the tray. The rigid links
Essentially it couples multiple nodes to a single slave
attach to multi-point links in the lug holes to model some
node without tying them all to each other – only to the
bolted joint compliance.
slave node. Attributes applied to the single slave node are
distributed to the master nodes. This is useful if you want
to smear a mass onto a system, or apply load to many
locations from a single point. Using rigid links to do this
is not always desirable as it adds rigid bodies to your
model with infinite stiffness that can significantly
influence results, particularly local results. You can think
of the interpolated multi-point link as a “soft” version of
the rigid link.

Example: Mass Distribution in a Rocket

The simple rocket model shown

in Figure 2.1 uses links in this
way. We will explore the use
of rigid links and multi-point
links in various sections of the Figure 2.2: Rocket re-entry pressure profile.
rocket:
The fuel tank has a 200kg mass distributed to the walls
1) Nosecone with multi-point links. This allows the walls to flex and
2) Instrument bay distort while still transferring the inertial loading from the
3) Fuel tank fuel. Had we used rigid links, the entire fuel tank would
behave as a rigid body, giving no stress results and
4) Burner
causing stress concentrations at the interface with the
5) Tail and fins other sections of the rocket. Figure 2.3 shows the links
The instrument bay has a 100kg representative of the instrument tray. Figure 2.4 shows
mass representing experiments, the fuel tank mass and links.
electronics and payload. The
effect of this mass is distributed
with rigid links to 8 lugs.
The fuel tank has a 200kg mass
representing fuel. The inertia of
the fuel is distributed to the
walls of the tank by multi-point
links.
A 20kg mass in the tail
representing fin actuator
hardware, is also distributed Figure 2.1: Simple rocket.
using multi-point links.
Figure 2.3: Instrument bay cutaway - mass and links.
The rocket is 0.4m diameter and 3m long. It weights
48kg and carries a payload of 320kg. It is subject to a 6g The rocket’s tail contains some fin actuator hardware.
lateral acceleration in a tumbling re-entry as it slams back Each fin is independent and has a 5kg actuator. This is
through the upper atmosphere. A normal pressure modelled as a 20kg mass distributed to all four fins with
sufficient to produce this acceleration is applied to the multi-point links. Had we used rigid links instead, the fins
underside of the rocket, shown in Figure 2.2. The would behave as though they were all connected by a
rigid body within the tail, which would not be reasonable.

www.strand7.com 3
Making Finite Element Analysis Easier

You could also choose to model the four actuators
independently, but having a single point mass to modify
in the future can be more efficient from a modelling
standpoint. Figure 2.5 shows the fin actuator mass
connected to the fins.
The six lugs attached to the bottom skin of the rocket
distort the skin surface causing local stress concentrations.
Although the tray is modelled with rigid links, sufficient
compliance is modelled in the lugs to maintain realism.

Figure 2.7: Instrument bay cutaway showing stress.

Figure 2.4: Fuel tank cutaway mass and multi-point links.

Figure 2.8: External view instrument bay deflection (100x).

The fuel tank section has the highest moment due to high
inertial loading there. Figure 2.9 shows the tank’s internal
structure stress at 6g.

Figure 2.5: Rocket tail cutaway mass and multi-point links.

Examining results, the rocket bends about the 200kg fuel

load, which is the primary source of inertial loading
(11.6kN at 6g).
Figure 2.9: Fuel tank cutaway showing internal walls.

Load Direction

Figure 2.10: Fuel tank stress using rigid links.

Figure 2.6: Rocket lateral 6g deflected shape and stress.

www.strand7.com 4
Making Finite Element Analysis Easier
If rigid links were used to distribute the inertial load, the
entire tank would behave as a rigid body and there would
be no stress or distortion, as shown in Figure 2.10.
Interpolated multi-point links allow the natural distortion
and stress field to develop.
No significant stress or deflection occurred in the tail, and
the small 20kg mass was distributed to the four fins
without adding stiffness.
Figure 2.11: Tail cutaway stresses.
Example: Load Transfer at a Sliding Strap
The interpolated multi-point link can also be used to
model interaction between two structures, such as a
floating pier with a pile shown in Figure 2.12.
Pile
Strap
Tides
Floating Pier
Surge
Figure 2.12: Floating pier showing tidal motion.
Figure 2.13: Pier model layout.
www.strand7.com
In this scenario, the pier is attached to the pile with a
loose fitting sliding strap. This strap allows the hollow
steel pile to deform within its grasp. The interpolated
multi-point link is well suited to this type of soft coupling,
where the strap can load the pile but doesn’t add any
stiffness to it.
The pier and pile will be modelled using shell elements.
The strap is modelled as a beam element attached to a
multi-point link on the pile.
Figure 2.14: Strap modelling details.
The piles are fixed at the base, with two
of the three piles also restrained above
the floating pier to represent their
attachment to a non-floating pier. A
critical load case is an incident wave or
surge hitting the side of the pier. This is
modelled as a shell edge normal
pressure, and loads the pile through the
strap.
The pier was analysed using the linear
static solver. One column was critical
in a three-point bend mode with
ovalisation shown in Figure 2.15.
This ovalisation would not be allowed had we used rigid
links to represent the strap/pile interaction. The
interpolated multi-point link allows load transfer from the
pier to the pile and into the upper support without adding
stiffness to the pile, giving a more realistic response.
Figure 2.15: Pier stress and deflection (100x).
5
Making Finite Element Analysis Easier

Interpolated multi-point links are a good tool for smearing
load or mass and for coupling structures in which one
member transfers load but not stiffness to the other.
The rocket and pier examples demonstrate the general use
of the interpolated multi-point links. This new capability
in Strand7 R2.4 expands the possibility for model
simplification and efficient model construction.
Interpolated multi-point links can be generated in a
number of ways in Strand7. One approach is to use
Create/Link from the main menu: select Multi-Point from
the dropdown list and then set the parameters (such as the
number of points to be linked and the degrees of freedom
to be coupled). Then click all the nodes in the cluster,
including the slave node. This approach is useful when a
relatively small number of nodes needs to be linked. For
linking larger clusters, use Tools/Auto Assign/Restraints.
With this tool you can select all the nodes in the cluster
by any convenient selection tool and then automatically
create a link to any desired slave node. There is even the
option of automatically creating a new node at the
average coordinates of the cluster and assigning it as the
slave node
Did You Know?
Select Brick or Plate Faces
In R2.4 you can select brick or plate elements based upon
the relative angles between free faces. This is useful for
situations where a surface pressure needs to be applied,
but the shape does not lend itself to easy selection using
Select by Region or any other convenient method.
In this article we discuss:
1. extracting time-dependent material data from
these two particular codes; and
2. making the appropriate entries into Strand7 R2.4
in order to realise the various recommendations
within these design codes.
We will also examine the long term structural behaviour
of a reinforced concrete beam to further highlight the new
concrete creep and shrinkage modelling capabilities
within Strand7 R2.4.
Entering Design Code Data for Time-Dependent
Concrete Properties
The concrete design codes, CEB-FIP Model Code 1990
and ACI 209R-92, provide recommendations on the
following time-dependent concrete properties:
1. The development of modulus of elasticity with
time;
2. Creep; and
3. Shrinkage.
In this section we will look at some of the details of these
recommendations and at the same time highlight the
corresponding inputs required in the Strand7 R2.4
interface in order to comply with these concrete design
codes. (See the Online Help for a more complete
discussion on this subject.)
Development of Modulus of Elasticity with Time
It is well known that as concrete ages it increases in
strength and stiffness. The equations proposed by CEB-
FIP Model Code 1990 and ACI 209R-92 to model the
increasing modulus of elasticity with time are given in
Table 3.1.
To input these suggested equations for the time variation
of the elastic modulus in Strand7, choose Tables/Factor vs
Time and use the Equation Editor . The 28 day
modulus E 28 needs to be entered in the Structural tab of
the Element Property dialog box and the generated Factor
vs Time table needs to be assigned to the Modulus vs
Time dropdown in the Table/Time tabs of the Element
Property dialog box.

Concrete Creep and Shrinkage CEB-FIP Model Code 1990 ACI 209R-92

W ith the release of Strand7 R2.4 comes a new   

E t   E28  exp s1 
  
 
28   
t   
0.5

(1) E t  
t
  t
E 28 (2)
capability that allows users to model the creep and  
shrinkage of concrete. These time-dependent and E 28   1.5 0.043 f c 28 (3)
where E28 is the modulus at
inelastic deformations play a significant role in the long
28 days and s is a constant
term performance of concrete structures, and can be where E28 is the modulus at
based on the cement type.
factors of economic importance.
28 days, f c 28 is the 28 day
In practice when design engineers are investigating the compressive strength; α and β
long term behaviour of concrete, they must generally refer are constants based on curing
to one of many proposed design codes to establish conditions and cement type.
appropriate time-dependent properties. Among the most
commonly used concrete design codes are the European Table 3.1: Concrete Design Code expressions for Modulus vs Time.
code CEB-FIP Model Code 1990 [1] and the American
code ACI 209R-92 [2].

www.strand7.com 6
Making Finite Element Analysis Easier

Creep
Creep is time dependent deformation that occurs due to
loading. The majority of design codes, including CEB-FIP
Model Code 1990 and ACI 209R-92, characterise creep
behaviour through the use of a dimensionless quantity
known as the creep coefficient  (t , ) at time, t , and age
of first loading,  . This quantity is defined as the ratio of
creep strain to elastic strain:
creep strain  c t ,  
 t ,     (4)
elastic strain  e  

As t approaches infinity, it is generally assumed that the

creep coefficient reaches an ultimate limit value u  
where

 cu  
u     ,   (5)
 e  
This final or ultimate creep coefficient is a useful and
commonly used measure of the capacity of concrete to Figure 3.1: Creep Law Definitions dialog, Creep tab.
creep. The CEB-FIP Model Code 1990 and ACI 209R-92 So how are the actual numerical values for the
expressions for the concrete creep coefficient are given in coefficients in the generalised hyperbolic expression
Table 3.2. determined? Well, the various parameters in the design
code equations (i.e.  H in Equation 6;  RH and   fcm  in
CEB-FIP Model Code 1990 ACI 209R-92
Equation 7;  2 to  6 in Equation 10 etc.) are empirical
Creep Coefficient: Creep Coefficient: parameters that account for a variety of physical factors
including the composition of the concrete material, curing
 t    
0.3
 t ,  
t   0.60 
 t ,     u (9) methods, the humidity and the size of the concrete

  H  t     10  t   0.60
u
(6) member. Once you have quantified these physical factors
Ultimate Creep Coefficient: Ultimate Creep Coefficient: as they relate to the actual structure being analysed, it is a
simple matter of consulting the design codes to extract the
u       RH   f cm  (7) u   1   2.35 2 3 4 5 6 (10) parameter values. While the majority of the parameters
Creep Strain: extracted from the design codes are then easily matched
Creep Strain:
to the coefficients of the generalised hyperbolic law
  (t , )    (t , )  offered in Strand7, it is worthwhile highlighting two subtle
 creep      creep     (11)
 E28 
 (8)
 E    aspects of this discussion.
Firstly, in order to account for the time variation of the
Table 3.2: Concrete Design Code expressions for the development of ultimate concrete creep coefficient as suggested by the
design codes - see the first terms in Equation 7 (i.e.    )
concrete creep strain.

By examining the proposed equations for the concrete and Equation 10 (i.e.  1   ) - an appropriate Factor vs
coefficient in Table 3.2 (see Equations 6 and 9), it can be
seen that both the CEB-FIP Model Code 1990 and ACI Time table may be assigned. Equations 13 and 14 in
209R-92 code subscribe to the same type of hyperbolic Table 3.3 show the actual design code recommendations
expression. In order to support both design codes, for the influence of time on the ultimate creep coefficient.
Strand7 offers the following generalised hyperbolic law*: We suggest using the Equation Editor within the Factor vs
Time table to generate data according to these equations.

 t     The second issue that prompts further consideration
 t ,      u (12)
   t     relates to the code definitions of creep strain given in
 
Equations 8 and 11. According to the CEB-FIP Model
To access this creep law in the Strand7 interface, you will Code 1990, calculating the creep strain from creep
need to bring up the Creep Law Definitions dialog under coefficient involves a division by a constant value of the
the Property menu (choose Property/Creep). By clicking elastic modulus E28 (see Equation 8). To specify this
on the Creep Law dropdown list within this dialog, you aspect of the concrete creep model within Strand7, the
can access all of the various creep laws supported by Constant Modulus option within the Creep Law
Strand7 including the Concrete Creep and Shrinkage - Definitions dialog (see Figure 3.1) needs to be set.
Hyperbolic Law option. Conversely, the ACI 209R-92 model requires that the
creep strain be determined by dividing by an elastic
modulus that varies with time (see Equation 11). For this
* situation, the Constant Modulus option needs to be
The User Defined option can be used for any design codes which do
not subscribe to a hyperbolic law. cleared.

www.strand7.com 7
Making Finite Element Analysis Easier

Strand7 CEB-FIP Model ACI 209R-92

Parameter Code 1990

 1.0 0.6

 Determined from 10
code based on the
relative humidity
and size of member

 0.3 1.0

u  RH   f cm  u  2.35 2 3 4 5 6
determined from the determined from relative
relative humidity, humidity, size of
size of member and member, curing
compressive conditions and concrete
strength composition

u – Factor    
1 (13)  1  1.25 0.118
0.1  τ 0.2 Figure 3.2: Creep Law Definitions dialog, Shrinkage tab.
vs Time
Table (moist cured) (14a)
The process of determining the numerical values for the
 1  1.13 0.094 various parameters for shrinkage hyperbolic law follows
the same steps discussed for the creep parameters. Again,
(steam cured) (14b) you will need to consult the design codes to determine
which physical factors influence the shrinkage properties
Constant Checked Unchecked (e.g. the relative humidity, the size of the member, the
Modulus cement type, the compressive strength) and decide on the
appropriate values of these factors as they apply to your
Table 3.3: Summary of Strand7 inputs for matching design code
analysis. Following this, it is a straightforward process of
recommendations for creep.
obtaining the required parameter values for the
Shrinkage hyperbolic expression and matching the inputs into
Strand7. A summary of the entries required in Strand7 to
Shrinkage is the time-dependent deformation that occurs
match the design code recommendations for shrinkage is
independently of any applied loading. The recommended
provided in Table 3.5.
expressions for the development of shrinkage strain over
time provided by the CEB-FIP Model Code 1990 and ACI
209R-92 codes are given in Table 3.4. Strand7 CEB-FIP Model ACI 209R-92
Parameter Code 1990
CEB-FIP Model Code 1990 ACI 209R-92
s 1.0 1.0

 t   
0.5

 sh 
t     (17)
 sh     sh,u (15)
 350h h 2  t       t    s 350h ho 2   35 (moist cured ) or
sh ,u
 0 

 sh,u   s  fcm  RH (16) where   35 (moist cured) where h and   55 (steam cured)
or   55 (steam cured) ho are measures
of the size of the
 sh,u  780  106  concrete
(18)
 sh1 sh 2 sh3 sh 4 sh5 sh 6 sh 7 member.

Table 3.4: Concrete Design Code expressions for the development of s 0.5 1.0
concrete shrinkage strain.

Once again, in order to accommodate both design  s0  s  f cm  RH 780  106 

recommendations, a generalised hyperbolic expression
determined from  sh1 sh 2 sh3 sh 4 sh5 sh 6 sh 7
for the Shrinkage Strain is offered in Strand7: concrete determined from relative
 strength, cement humidity, size of member,
 t    s  s
 s t ,      s0 (19) type and relative curing conditions and
   t    s  humidity concrete composition.
 s 
To access this hyperbolic law for shrinkage in the Strand7 Table 3.5: Summary of Strand7 inputs for matching design code
interface, click the Shrinkage tab of the Creep Law recommendations for shrinkage.
Definitions dialog and choose the Shrinkage Formula
option.

www.strand7.com 8
Making Finite Element Analysis Easier

Example: Creep and Shrinkage of a Reinforced Concrete

Beam
Now that you know how to extract and input code based
concrete creep and shrinkage properties into Strand7, let’s
have a look at an actual analysis. This example
investigates the effects of concrete creep and shrinkage on
a pre-tensioned reinforced concrete beam over a duration
of 50 years. The concrete beam was modelled using
brick elements and the reinforcement modelled using
truss elements assigned to various string groups. The
concrete beam was fixed at both ends and the structural
response was compared under two different loading
scenarios:
1. with no loading; and
2. with an applied pre-tension in the top layer of Figure 3.4: Import Creep Data dialog, showing the available creep data.
reinforcement within the flange.
Before investigating the results of the analysis, we will
have a look at some specific details of the Strand7 model
setup that are particularly relevant to performing a
concrete creep and shrinkage analysis in Strand7.

Brick mesh
Pre-tensioned bars

Reinforcement bar

Figure 3.3: Sketch of the reinforced concrete beam. The beam outline is
drawn in black and is meshed using brick elements. The blue lines
represent the pre-tensioned bars, while the red line represents the
reinforcement bar. These are meshed with beam elements.

Firstly, let’s review the process of assigning concrete Figure 3.5: Creep Law Definitions dialog after the ACI 209 data is
creep and shrinkage properties in the model. From the imported.
main menu Property/Creep was selected and from the
Creep Law dropdown list, the Concrete Creep and
Shrinkage – Hyperbolic Law was accessed. Defining the
required coefficients could be performed by following the
process described previously, but in this case Strand7
R2.4 includes creep and shrinkage data for concrete in the
library. By clicking the Import Data button and
selecting ACI-209 R-92 Concrete Creep +
Shrinkage/Steam Cured Concrete the appropriate
coefficients were filled in. Users have the option of
Figure 3.6: Selecting creep data in material property dialog.
adding to the library (click Export Data ), so if
concrete creep and shrinkage is a consideration that you As well as defining the creep and shrinkage properties, we
will need to tackle often, we suggest that you make use of needed to assign the elastic properties for the concrete
this facility. material. To comply with the ACI 209R-92 code for the
development of the modulus of elasticity with time (see
Next we assigned the concrete creep and shrinkage data Table 3.1), the following steps were undertaken:
set to the appropriate brick property. This was done by
clicking Property/Brick, choosing the Nonlinear tab and 1. From the main menu Table/Factor vs Time was
then selecting from the dropdown list under Creep Data. selected and the Equation Editor was used to
All the creep sets that are generated under enter the data in Figure 3.7 (this is Equation 2
Property/Creep are made available here to be selected as without E 28 where  =1.0 and  =0.95);
part of the Element Property definition (Figure 3.6).

www.strand7.com 9
Making Finite Element Analysis Easier

2. The 28-day modulus ( E 28 ) was assigned by
clicking Property/Brick and entering
appropriate value in the Structural tab;
3. Within the Brick Element Property dialog, the
Tables tab and Time tab were selected and the
Factor vs Time table was assigned in the Modulus
vs Time dropdown.
experienced creep and shrinkage deformations. As
mentioned previously, analyses were performed with and
the
without pre-tension. A comparison of these two cases is
shown in Figure 3.9 where the evolution of the concrete
fibre stress is plotted.
Brick Stress vs Time
2

1.5

0.5

Stress (MPa)
0
0 5000 10000 15000 20000
-0.5

-1
No Preload
-1.5
Preload
-2

-2.5
Loading Time (Days)

Figure 3.9: Brick ZZ stress vs time plots for the beam without and with
pre-tension.

Did You Know?

Figure 3.7: Strand7 Equation Editor for Factor vs Time tables.
Once the mesh setup was complete and all the material
properties were assigned, we solved the model using the
new quasi static solver (Solver/Quasi Static) with the
Creep option set. This solver allows you to calculate the
time-dependent response of a structure subject to any
loading conditions and initial conditions, but ignores any
dynamic (inertial) effects. It is, therefore, the most
suitable solver for performing a creep analysis. Also, by
using the time step editor in the quasi static solver (Time
Steps…) we were able to specify an appropriate time
stepping scheme over the 50 year period of interest.
Plate and Brick User Contour
For plate and brick elements a User Contour based on stress
and strain result quantities can be defined in Results Settings.
This is useful in plotting the contours of derived quantities
based on the elemental stress and/or strain, for example a
material failure criterion. This is also available for the Peek
tool and Results Listings function.
In this example, we derive and contour a concrete cracking
f t t 
index according to I cr  , where f t t  is the tensile
 1 t 
strength at age t obtained from concrete design codes and

 1 t is the maximum principal stress calculated by Strand7.
Since only the tensile stress is required, the function IFPOS
can be used in this case, as shown below (IFPOS(x) returns x
when x is positive and zero otherwise). The contour below,
obtained with the equation 0.35*SQRT(32)/IFPOS([S11]),
indicates that some cracking can be expected in the concrete
beam due to the creep and shrinkage – the results shown are
for a case where there is no pre-tensioning.

Figure 3.8: Quasi Static Analysis setup. Note that the Nonlinear Material
option is enabled to account for the Modulus vs Time table.

As the model is relatively large and the mesh is made up

of predominantly brick elements, the newly available
iterative PCG scheme was used (see Scheme/Iterative
(PCG) in the Quasi Static Solver dialog). This scheme is
most useful for meshes that have a large stiffness matrix
with predominantly brick elements, and in general has
shorter solver running time compared to the Sparse
scheme.
After solving the model we were able to investigate the
structural response of the reinforced beam over time as it

www.strand7.com 10
Making Finite Element Analysis Easier

For the first case where no loading is applied, it is no
surprise that the instantaneous response (at time = 0) is a
stress state of zero. However as time continues, the
concrete material undergoes compressive shrinkage
deformation. A consequence of this deformation, along
with the beam’s end fixity, is that the concrete beam is
placed in tension. The creep deformations within the
concrete would perhaps have a relieving effect – however
the resulting tensile stresses are not desirable and may
lead to cracking.
To mitigate the possibility of long term cracking, pre-
tension can be applied to the reinforcement. From Figure
3.9, it can be seen that pre-tensioning the reinforcement
bars places the concrete beam into compression. While
the creep and shrinkage mechanisms influence the long-
term stress field within the concrete material, the desired
compressive state is maintained.
The model used in this article can be downloaded from
www.strand7.com/news.st7.htm
References
[1] CEB-FIP Model Code 1990, Comité Euro-International
du Béton, 1990.
[2] ACI Committee 209, Prediction of Creep, Shrinkage
and Temperature Effects in Concrete Structures, ACI
209R-92.

Modelling Tip – Concrete Reinforcement

One of the common questions we receive on concrete structures is related to the modelling of concrete reinforcement. As with other
structural modelling it depends on the level of accuracy and local details required by the user. In this issue, some of the common
approaches are presented in the order of increasing complexity.

1. Beams Model 2. Plates and Beams 3. Simple Bricks and 4. Detailed Bricks and
Model Beams Model Beams Model

This option is used if the global
response of a beam or column-
like structure is of interest. The
concrete and reinforcement are
modelled as coincident beams
of different properties and the
reinforcement is positioned in
the correct location using the
Offset Attribute.
This option is used if the
structure of interest has slabs
or walls in addition to beams
and columns. The
reinforcement can be modelled
either as beams located at the
correct locations, or as plates
using an equivalent smeared
representation.
This option is used if a
relatively detailed model is
required. The concrete is
modelled as brick elements
and the reinforcement as
beams. The beams are located
on the brick nodes.
This option is a more detailed
version of Option 3. The
location and diameter of each
slot in concrete is modelled.
The reinforcement bars are
joined to the concrete via rigid
links and possibly contact
elements.

Pros: Pros: Pros: Pros:

- Very simple to model.
- Good global representation.
- Solution time is very short.
- Can generate bending
moment and other quantities
conveniently.
- Good representation of a
general structure.
- Solution time is relatively
short.
- Good representation for a
detailed model.
- Can model the interaction
between concrete and
reinforcement.
- Most accurate
representation of all options
presented.
- Singularities are avoided.
- Can model the interaction
between concrete and
reinforcement.

Cons: Cons: Cons: Cons:

- Limited to beam and column
type structures.
- Can be difficult to model for
complex reinforcement.
- The interaction between
concrete and reinforcement
is difficult to model.
- Can be difficult to model for
complex reinforcement.
- Solution time can be long. - Solution time is the longest
- Stress singularity at the ends of all representations.
of the reinforcement bars. - The model can be time
consuming to create.

As you can see, the modelling options depend on the type of structure to be analysed and the results of interest. If you are interested
in global responses, Options 1 and 2 are more appropriate. Options 3 and 4 are appropriate for detail analysis. It is also possible to
create a model which features more than one option introduced above. For example, a complete building model can be constructed
using Option 2, with all columns modelled using Option 1. We also encourage you to seek out modelling methods that are not
covered here, but can give a good representation of the structure

www.strand7.com 11
Making Finite Element Analysis Easier

experiments or optimisation approach to design; even

API Functions with a single Strand7 licence you can launch multiple

T
he Strand7 Application Programming Interface (API)
received major upgrades for R2.4, including access to
hundreds of new functions. New capabilities found in
Strand7 R2.4 are included in the API, allowing
construction sequences, moving loads, import and export
of geometry files and other formats, geometry cleaning
and tools, and even automeshing.
This suite of tools gives you access to the full flow of
model creation and analysis, from geometry import,
through automeshing and application of attributes,
solving, and on
to results post-
processing – all
from within the
Strand7 API
environment.
New API
functionality allows
you to open interactive
Strand7 model view windows,
giving instant graphical feedback
including results. These windows can also be embedded
directly within your Strand7 API program for an integrated
appearance. You can create and display animated results,
and custom result files can be built up within the API, for
example fatigue life or combined damage contours.
solvers on the same PC.
The Strand7 API interfaces with many different
programming environments and languages, including
C++, Delphi, Intel Fortran, MS Visual Basic, MS Excel
VBA, MS Visual C++ and, new for R2.4, MS Visual
C++.NET and Matlab 7.3
Geometry Tools
W orking with geometry entities has become easier in
Strand7 R2.4 with the new Geometry Tools. For
example, a geometry face can now be split by vertices or
by a plane entirely within Strand7, making the process of
generating a compatible mesh much simpler than before.
Some other great additions to the Geometry Tools include
grafting edges to faces, generating faces from beam and
plate elements, intersecting and morphing edges; all
accessible under the Tools/Geometry Tools submenu.
The following section highlights some of these modelling
techniques when working with geometry entities.
Face from Plate and Split Face by Vertices

Plate elements turned into

geometry faces.

Geometry faces split and

common edges created by
vertices.

The ability to launch multiple Strand7 solvers at once

allows for parallelisation of large batch jobs in today’s
multiprocessor desktop environments. This can Geometry faces can be generated within Strand7 by
dramatically decrease your time to results for a design-of- converting plate elements into faces. For simple

www.strand7.com 12
Making Finite Element Analysis Easier
geometries, this can be a quicker and easier method than
creating geometric entities and importing them from a
CAD model.
Should it be necessary, vertices can be created to define a
line for geometry faces to be split. A split line on a planar
face will remain straight while the split line on a curved
face will follow its face definition.
Rebuild Faces
Bad geometry definition
rebuilt to improve mesh
quality.
Sometimes geometry imported from CAD has poor
underlying B-spline surface definitions. Try rebuilding it
to refit a new, better quality surface over the existing
surface.
Graft Edges to Faces and Morph Edges
Curved edge grafted to
curved face.
Blue edge morphed
towards the red edge.
A mesh generated from a grafted face will inherit its
grafted feature, thus creating interfaces between the
geometric entities.
Edges at the interface can then be morphed together to
create proper intersections for compatible automeshing.
www.strand7.com
Did You Know?
Skip Transitioning
Skip Transitioning is an automeshing option that aims to
produce a mesh as regular as possible by ignoring small
features in the geometry. This option can be used in
conjunction with the Vertex Mesh Size attributes to create a
mesh with specific boundaries such that it can be directly
“stitched” onto the existing mesh.
Initial Mesh
A modification is required to include
an additional hole to the existing
mesh.
Using the new geometry tools such
as Face from Beam Polygon and
Graft Edges to Faces, a geometry
face with a cavity can be created.
Additional vertices with matching
Mesh Size attributes are created.
Modified Mesh
The geometry face is surface meshed
with the Skip Transitioning option
and the mesh will be modified with
the required feature.
Skip Transitioning can also be useful for skipping over small
features, such as small mismatched edges producing small
element edge lengths.
With
Transitioning
Skip
Transitioning
13
Making Finite Element Analysis Easier

In the example below, the magenta beam polygon is not

Split Face by Plane
planar. If the Face from Beam Polygon tool is used
without choosing a Target Face, Strand7 calculates the
least-squares best fit plane and creates the face on that
plane (shown in green) with the vertex locations projected
perpendicular to the new plane. If the Target Face is
selected (here we have used the blue face), the resulting
face (the red face) is created using projection
perpendicular to that target face.

Intersect Edges

Geometry faces can also be split by plane. The cutting

plane is defined by a user cartesian coordinate system of
choice.
In addition to the splitting, geometry faces at the cutting
plane can be optionally generated. This is particularly
useful to separate the volumes for solid meshing later on.

Face from Beam Polygon

A closed polygon formed by beam elements can also be

used as a definition to create geometry faces. Edges can be easily intersected in a few clicks.
An angle tolerance is available that allows you to control Intersections are detected if the distance between the
whether curves should be smooth or faceted. edges falls within the specified tolerance.

You can generate planar faces in this way, or you can The faces will remain unconnected but an option to split
generate curved faces by projecting the polygon onto the faces while intersecting the edges is provided.
another existing face. Splitting the faces will achieve mesh compatibility.

www.strand7.com 14
Making Finite Element Analysis Easier
Face from Cavity
Geometry face created
from cavity.
Filling in a cavity is easily achieved by generating a
geometric face from a cavity loop.
Faces created from cavities will inherit the surface
definition from the face with the cavity; in this instance it
is cylindrical
Did You Know?
Browse…
If you are trying to find a model and cannot remember its
name you can try using File/Browse in Strand7.
Navigate to a likely model location in the Browse window
and click Browse. This will display an image of the last
saved model view with the file name of all Strand7 files
located in this folder. A little contour bar is displayed in the
bottom left hand corner of the image if the model has
results. Also, if you right-click a model image you can
open, rename, delete or see information about the file.
www.strand7.com
Benchmark
I n this edition we consider a problem based upon one of
Strand7’s Verification models, documented in the
Verification Manual (the Verification Manual is a 300+
page book, shipped in PDF format, and is located in the
Verification folder of your Strand7 R2.4 installation).
This problem makes use of the Plastic Friction Model,
another new addition to R2.4. The previously available
Elastic Friction Model assumes that the friction force due
to contact is equivalent to a nonlinear spring. This means
that if the movement direction reverses, the friction force
does not reverse until zero relative displacement is
reached. As a result, this friction model is only valid in
situations where the movement does not stop or reverse
direction.
The Plastic Friction Model accounts for changes in
direction, so if the direction of movement reverses the
frictional force reverses accordingly. This allows the
friction model to demonstrate hysteresis and should give
more realistic friction forces when contact is lost and
regained between entities.
Contact
Movement
Plastic
Friction
Force
Another improvement in the friction capabilities of the
new contact element is the addition of an elliptical yield
surface, in addition to the previously available rectangular
yield surface. In the rectangular option, equilibrium
satisfies the following equations independently:
F1  1 N and F2   2 N ,
where F1 and F2 are the frictional forces in the element's
1 and 2 principal directions respectively, 1 and  2 are
the corresponding coefficients of friction, and N is the
normal force in the element. With the elliptical option,
the equilibrium equation is
2 2
 F1   
    F2   1 .
 1 N   2 N 
   
The model uses 20 2D plane stress elements to define the
block that is supported by 21 Normal Point Contact
elements. To improve the accuracy of the model it is
necessary to introduce damping locally to the contact in
the vertical direction. This has been achieved by using
15
Making Finite Element Analysis Easier

small Spring-Damper elements between the contact

elements and the block.
Training

T he training calendar for the first half of 2010 is still

being finalised, but the following courses have been
scheduled:

Date Course Title Location

22-24 February Strand7 Essentials Perth, Australia
25 February Nonlinear Analysis Perth, Australia
with Strand7
The nonlinear transient dynamic solver is used with the 26 February Dynamic Analysis Perth, Australia
nonlinear geometry option set. Viscous damping does with Strand7
not need to be added in the solver because the viscous 1-3 March Strand7 Essentials Cambridgeshire,
damping properties of Spring-Damper elements are UK
always included in nonlinear transient analyses.
4 March Nonlinear Analysis Cambridgeshire,
The initial velocity is applied as an initial condition within with Strand7 UK
the solver dialog (using VX=1) and is applied to all the
5 March Dynamic Analysis Cambridgeshire,
nodes in the model. It is only used on those nodes which with Strand7 UK
are not restrained and are therefore free to move in the
direction of the initial velocity. To register, please go to www.strand7.com and follow the
The box slides with initial velocity of V0=1m/s along a links.
horizontal surface with friction coefficient =0.3. The We are also planning a Strand7 Workshop for our UK
distance the box will take to slide to rest depends on the users on the 8th and 9th of March. This will be held in
energy balance of the system. By energy methods the Cambridgeshire - details on content will be released on
analytical result for the distance is: the Strand7 website shortly
V02
X rest   0.16995m
2 g Exhibitions
A graph of the horizontal displacement with respect to
time is shown in Figure 4.2. It takes just over 0.3 seconds
for the box to stop moving and reaches a distance
0.17045m at the end of the analysis, 0.3% above the
B etween now and end of May, we will be exhibiting
Strand7 at the following events: OTC 2010 (Offshore
analytical solution. The slightly higher value is caused by Technology Conference) and NASCC 2010 (North
the initial settling period where the contact has not yet American Steel Construction Conference).
established itself fully giving a lower friction force for a
short time. 3-6 May 2010 OTC Houston, USA
12-14 May 2010 NASCC Orlando, USA

Other exhibitions will be announced soon on the Strand7

website and in the next edition of News.St7

Head Office
Strand7 Pty Ltd
Suite 1, Level 5
65 York Street
Sydney NSW 2000
AUSTRALIA
Tel +61 2 9264 2977
Figure 4.2: Horizontal displacement vs. time (s).
Fax +61 2 9264 2066
Email info@strand7.com
This model is available to download from the News.St7 Web www.strand7.com
page of our website www.strand7.com/news.st7.htm

www.strand7.com 16
Strand7

News.St7 Newsletter for Strand7 and Straus7 users

Strand7 is marketed as Straus7 in continental Europe

Issue 8, September 2010

Welcome to another issue of News.St7. The articles in this issue focus on some of In this Issue
the many new features in Strand7 R2.4. We present an overview and example of
the Soil In-Situ Stress tool, which calculates the steady-state (initial) stress state in  Soil In-Situ Tools 1
soil models taking into account fluid content and pore pressure in the soil; it is an
important step in soil analysis. We also cover in more detail the entities, attributes  API Windowing Tools 5
and tools available for attaching dissimilar meshes together, in particular the edge
 Attaching Dissimilar Meshes 6
attachment options for plate elements. New functionality in the Strand7 API is also
presented - namely, the ability to launch the Strand7 Model Window through the  Free Body Diagrams 7
API and to make it part of your own application.
 Factor vs Position Table 9
As usual, there are new Did You Know? items plus information about Strand7
Training, and our participation at upcoming exhibitions around the world. We are  API Troubleshooting 11
also considering to run a Strand7 course in Abu Dhabi in late November or early  Training and Exhibitions 12
December. The course is expected to run for five days and to cater for both new
and experienced Strand7 users. The course can only go ahead subject to there being sufficient participation, so please
contact us as soon as possible if you are interested in attending.
Since Strand7 R2.4 was released last December, we have issued two point updates - the latest release is now Strand7 R2.4.3.
If you did not receive our emails announcing the updates, please send us your contact details so that we can amend our
mailing list. This will ensure that you receive a timely notification for future updates.
We hope you enjoy reading this issue of News.St7 and look forward to hearing from you with comments on its contents, or
with suggestions for topics you would like us to cover in future issues

If you have any feedback or suggestions regarding the content of News.St7 or you would like to contribute a case study showing how you use Strand7,
please email us at news.st7@strand7.com. If you would like to receive your copy of News.St7 automatically by email, simply send a blank email to
newsletter-subscribe@strand7.com. All care is taken to ensure that information in News.St7 is accurate and up to date at the time of publishing. However
Strand7 Pty Ltd accepts no responsibility for inaccuracies in, or changes to, such information.

module (Tools/Auto Assign/Soil In-situ Stress) is designed

Soil In-Situ Tools
I n Strand7 Release 2.4 the soil modelling capabilities
have been greatly expanded from the previous releases.
This release has seen the addition of new soil material
models as well as several modules and element attributes
that help the user to define the soil’s initial stress state.
This article takes a look at these tools.
The behaviour of a soil model in a geotechnical analysis
is highly dependent on the in-situ (initial or in position)
conditions of the soil, i.e. its stress state. The soil in-situ
conditions represent the stress history of the sample, and
are generally specified in the soil material models by
input parameters to define the influence of the stress
history upon the soil behaviour under the current loading
situation.
In Strand7, the soil in-situ state can be defined using
dedicated tools. The Auto Assign Soil In-situ Stress
to automatically calculate and assign the soil in-situ stress
state that gives the initial equilibrium condition. Two
Strand7 element attributes are also available for the
definition of soil model-specific parameters, the In-situ
Stress and the In-situ Ratios.
Soil In-Situ Stress State
The soil in-situ stress state is the three-dimensional stress
distribution within a body of soil that maintains the
equilibrium condition. This can be considered as two
separate components: the vertical, or the direction in
which the gravity is acting, and the horizontal, or the
direction that is perpendicular to the gravity. In theory,
the vertical component can be more easily calculated
when compared to the horizontal component, as the
mechanics in the vertical direction are relatively simple to
describe. Therefore the horizontal stress component is
generally expressed as a factor of the vertical component.

www.strand7.com 1
Making Finite Element Analysis Easier

Trapped Fluid and Soil Material Parameters defined in the Fluid tab of the soil property dialog, with
the default parameters set via the Soil/Fluid option of the
Real life soil samples often contain micro-pores or cavities solver dialog.
in the soil skeleton structure, which are mostly filled with
air and/or fluid. Under loading, and depending on the In addition to the Did You Know?
permeability of the soil, the fluid may flow through the drained/undrained soil state,
soil structure or may be partially retained in the pores. real life soil samples can
Middle Button for
The two limiting cases of this behaviour result in the so- also be classified by the
Dynamic Rotate
called drained and the undrained states of the soil. degree of soil saturation, or As of R242
The undrained state is where the fluid within the voids of the percentage of the pores you can now
filled with fluid. This is a quickly switch
the soil is not allowed to displace out of the soil structure
significant factor in the to dynamic
under loading. This means that any change in load will rotate mode by
change both the stress in the soil skeleton and the transient analysis, where the
simply clicking
pressure in the fluid. The drained state, on the other fluid displaces from the soil the middle
hand, is where the fluid is completely free to displace out skeleton due to mouse button.
of the structure under loading. This means that any consolidation. Currently The equivalent
change in load will change only the stress in the soil Strand7 considers only two keyboard
skeleton, with the fluid pressure remaining constant. In states of soil saturation: the shortcut is the
fully saturated state, where F4 key. Refer
reality, the response of soil to changes in load is a
all pores are filled by fluid, to the Online
transient one, where the soil state is somewhere between Help for other
the drained and undrained states. Nevertheless, the and the unsaturated state, shortcuts.
assumption of undrained behaviour is usually a good where all pores are filled by
approximation to the short term effects of loading changes air.
on soil, whilst the drained behaviour is usually considered
to be representative of the long term response. Soil In-Situ Stress Calculations
The soil mass therefore experiences both the stress carried Assuming the drained condition, the vertical stress
by the soil skeleton (known as the effective stress), and component of the in-situ stress state, S v , can be
the stress in the fluid (known as the pore pressure). The
calculated from the equilibrium condition between the
combination of these two stresses is known as the total
weight of the soil and the buoyancy force due to any fluid
stress of the soil. In Strand7, these two components are
within the voids of the soil structure. This gives the
clearly separated when using soil elements; stress results
following equilibrium form.
for soil elements are given as effective stresses, with the
pore pressure also available as an output option. Effective Sv    soil ghsoil   fluid ghsoil _ sub (1)
stresses are also used for the in-situ stress attributes
applied to elements. where

 soil = density of the soil skeleton;

g = gravitational acceleration constant;

h soil = height of the soil element measured from

the ground level;

 fluid = effective density of the trapped fluid,

taking into account that the volume of
fluid is a function of the voids in the soil;
Strand7 offers the void ratio, e , to define
this; and

hsoil _ sub = height of the soil element that contains

trapped fluid, measured from the fluid
reference level.
Figure 1.1: Strand7 Soil Material dialog, Fluid tab.
In the unsaturated case, as the voids are filled with air,
It is important to know that in laboratory tests of the shear only the soil skeleton is contributing to the stress state of
strength, soil samples can be tested under drained or the soil. This gives the following equilibrium form.
undrained conditions. This can result in different material
properties, which will affect the subsequent calculation of
Sv    soil ghsoil (2)
results. In Strand7, the element property definition The horizontal stress component is calculated from the
requires the drained material properties; the effect of the following relationship.
fluid for the undrained case is added to the element
behaviour based on the defined fluid characteristics (i.e. S h  K 0 Sv (3)
bulk modulus, density, level). Fluid properties may be

www.strand7.com 2
Making Finite Element Analysis Easier
where
K0 = ratio of horizontal to vertical stress.
The stress ratio K 0 is generally obtained from the
laboratory testing of the soil sample. A common
assumption made about this parameter is based on the
drained friction angle   of the sample.
K0  1  sin  
This relationship can be used if no laboratory tests have
been conducted. (See Bowles, J.E., Foundation Analysis
and Design, McGraw-Hill, 4th Edition, 1988.)
Soil In-Situ Stress Attribute
The plate and brick Soil In-situ Stress attribute is
applicable to all soil models to describe the soil in-situ
conditions discussed above. This attribute defines the
effective (not total) vertical stress (i.e. the stress
experienced by the element in the direction of gravity)
and the horizontal stress ratio, which is an estimate of the
stress in the direction perpendicular to the applied gravity.
Figure 1.2: Strand7 Plate In-situ Stress attribute dialog.
If applied, the parameter K 0 in this dialog overwrites the
value of the same parameter defined in the Element
Property dialog.
Auto Assign Soil In-Situ Stress Module
The Auto Assign Soil In-situ Stress module (Tools/Auto
Assign/Soil In-situ Stress) can perform the stress
calculation and assignment automatically. This module
can be applied to any of the Strand7 soil models and
considers both soil and non-soil elements contained in the
model.
To perform the stress calculation, the module requires the
following parameters:
1. The type of fluid trapped within the soil structure –
The calculation requires the fluid density to be
entered. The default option is water, which can be
entered automatically by clicking the Water button.
2. The reference level for the trapped fluid – This
specifies the global reference level for the fluid. If the
fluid reference level is set at the element property
level, then the property values will be used instead.
www.strand7.com
This allows for different parts of the mesh to have
different fluid reference values.
3. The load case that contains the gravity definition –
The Load case defining gravity dropdown list in the
In-situ Stress Generation dialog is used to select a
load case that specifies the direction and magnitude
of gravity used for the in-situ stress state calculation.
4. The boundary conditions – The Freedom case
(4) dropdown list in the In-situ Stress Generation dialog
specifies the restraint conditions of the mesh. This
represents the restraints on the structure at the in-situ
equilibrium state (i.e. before any additional loads are
added or any changes are made to boundary
conditions or staged construction changes occur).
5. The stage for which the in-situ calculation is to be
performed – This option is used in construction
sequence analysis, where a stage representing the
initial situation is used to calculate the in-situ
conditions.
Figure 1.3: Strand7 Soil In-situ Stress Generation dialog.
The module calculates the stress distribution according to
the defined settings, assuming drained conditions, and
automatically applies the element In-situ Stress attribute
to all elements within the selected stage. It aims to
determine the in-situ stresses that equilibrate the applied
gravity loads (soil weight and fluid pressure), such that the
net vertical displacement is approximately zero. It is
possible that, in some meshes, the stress calculation will
return a warning message about the convergence of the
calculation. This indicates that the applied In-situ Stress
attribute has to be checked to ensure that the correct
magnitude is being applied to the mesh.
Because the soil elements generally require the
application of initial in-situ stress, when these are not
applied, Strand7 automatically calculates an approximate
in-situ stress distribution before the solver is run - this
distribution is approximate because it considers only the
geometry and not the stiffness of the elements. Strand7
will issue a warning message to notify the user about this
automatic, approximate stress calculation.
3
Making Finite Element Analysis Easier

Soil In-Situ Ratios Attribute

The Soil In-situ Ratios attribute allows for the
specification of the initial void ratio and the Over-
Consolidation Ratio (OCR). OCR is the ratio of the
maximum stress in the history of the soil to the current Figure 1.6: Soil In-situ parameters for saturated soil defined in the
stress and is applicable only to the Modified Cam-Clay Strand7 Plate Property Definition dialog. We can apply a Void Ratio of
soil model. This attribute, if applied, takes precedence 0.0 for unsaturated soil.
over the default values specified in the Element Property
dialog. Now we apply the in-situ stress using the Auto Assign tool
described above.

Figure 1.4: Strand7 dialog for plate In-situ Ratios attribute.

Example – Fluid State in a Soil Sample

The following example is a simple demonstration of the
Strand7 soil in-situ calculations. We will consider three
cases of soil condition.
Figure 1.7: The In-situ Stress Generation dialog, using the settings
 The soil is saturated through the entire depth. described above.

 The top half of the soil is unsaturated while the bottom From the soil in-situ stress contours shown below we can
half is saturated. observe the difference between the saturated and
 The entire depth of soil is unsaturated. unsaturated conditions.

The soil sample is treated as a 2D Plane Strain case 10

metres deep and 2 metres wide. The mesh of each soil
sample is shown in the figure below.

Figure 1.8: Vertical In-situ stress contour for the three cases.

When the model is solved under the action of gravity

Figure 1.5: The mesh representing the column of soil. The blue elements
represent saturated soil, and the red elements represent unsaturated soil.
only, including the automatically calculated in-situ stress
values, the displacement contours below show that the
vertical displacement in all three cases is zero or very
small. This agrees with the in-situ condition definition
discussed above.

The soil void ratio for the saturated soil is taken to be 0.2,
while for the unsaturated soil the void ratio will not be
relevant. A horizontal stress ratio of 0.5 is assumed. The
bulk density of the soil is taken to be 1500 kg/m3 (this is
the mass density of the dry soil), and the fluid filling the
void is taken to be water.

www.strand7.com 4
Making Finite Element Analysis Easier

Did You Know?

User Contour of Principal Stress with Largest
Magnitude
For plate and brick elements a User Contour based on stress
and strain result quantities can be defined in Results Settings.
This is useful for plotting contours of derived quantities based
on the elemental stress and/or strain, such as a contour of
principal stress with largest magnitude according to:
 max  MAX (11, 33 )
This allows you to display both maximum principal stresses in
the same contour. Enter the following formula into the user
defined contour for plates: Figure 1.9: Displacement contours of the soil under gravity only.
IFPOSB(ABS([S11])-ABS([S22]))*[S11]
+ IFNEGB([S11]-ABS([S22]))*[S22] In a typical soil analysis, this would be the starting
condition for the simulation. Normally the first increment
This produces the signed principal stress with largest would be the gravity + in-situ stress conditions, with
magnitude.
subsequent increments applied to simulate the effects of
IFPOSB(X) evaluates to +1.0 if X is positive and 0.0 if X is construction, tunnelling, etc.
negative or zero.
IFNEGB(X) evaluates to +1.0 if X is negative and 0.0 if X is
positive or zero.
To plot the absolute value of the principal stress with largest
magnitude enter:
IFPOSB(ABS([S11])-ABS([S22]))*ABS([S11])
+ IFNEGB([S11]-ABS([S22]))*ABS([S22])
Substitute [S33] for [S22] when contouring for bricks.
The signed principal stress with largest magnitude for bricks
is shown below.
Figure 1.10: Relative Displacements dialog.

A feature of the Results Settings in Strand7 that is

particularly useful in the analysis of soil problems is the
Reference Displacement option (Results/Reference
Displacement). This option can be used to select a
reference result case, whose displacements are subtracted
from all subsequent increments so that relative
displacements can be visualised easily

API Windowing Tools

S trand7 R24 introduces new API functionality to allow

you to open interactive Strand7 model view windows,
giving instant graphical feedback on your models and
results.
To display a model window in your application, it is
necessary to first open a new or existing Strand7 model.
A model window can then be created by calling
St7CreateModelWindow. The window is not displayed
at first. To show the window call St7ShowModelWindow
and to hide the window call St7HideModelWindow or
simply close the window. The window is destroyed when
the model is closed. You can explicitly destroy a model
window by calling St7DestroyModelWindow prior to
closing a model.
The model window is displayed as a resizable window
that moves independently of your application. It is
possible to embed the model window within a window in
your application for an integrated appearance. This can

www.strand7.com 5
Making Finite Element Analysis Easier

be achieved by passing a window handle to the function

St7SetModelWindowParent or by getting the model
window handle using the function
St7GetModelWindowHandle and assigning it as a child
to an existing window. It is even possible to customise
the appearance of the model view window. Functions
exist to show and hide the top panel, toolbar, status bar
and right-click popup menu.
When changes are made to a model, the API does not
automatically update the model window. It is necessary
to call St7RedrawModel to update the display. Some
model changes such as changes to beam and plate
properties are not detected on redraw and a call to
St7UpdateElementPropertyData is first necessary to
synchronise the display database.
Although it is possible to generate compatible meshes to
connect the parts together conventionally, one quick and
easy solution is to use the Plate Edge Attachment
attributes to stitch the parts together.
The Plate Edge Attachment attribute is accessed via
Attributes/Plate/Attachment/Edge and can also be
applied to geometry edges via
Attribute/Edge/Attachment. The edge attachment can be
defined in three different directions: Planar, +z or –z.
The distance that the attachment should use to search for
elements to attach to is defined as part of this attribute.
planar attachment
+z attachment

-z attachment

The Planar attachment is used to attach the lug to the

round panel whereas the +z or –z attachment (depending
on the orientation of the panel) is used to define the
attachment between the edges of the round panel and the
Entity selection can be performed using the model view base plate.
window. The selection tools are available from the
Selecting menu option on the right-click popup menu.
From your program code you can then use the function
St7GetEntitySelectState to check whether a given entity
has been selected.
An example has been created to demonstrate the
windowing tools available in the API. This example is an
extension of the plate optimisation demo that ships with
Strand7. The demo is available as a Microsoft Excel
spreadsheet written in VBA and can be downloaded from
www.strand7.com/news.st7.htm
The attachments should cover the distance between the
mid-planes of the attached parts to account for their
thicknesses. The attachments can be made Flexible to
Attaching Dissimilar Meshes approximate the weld stiffness. Rigid and Direct
connections are also available if suitable for the type of
Using Edge Attachment Attributes to Attach analyses.
Dissimilar Meshes Together
After the attributes have been applied it is then necessary

S
to activate the attachment process. This is done using
ometimes the geometry of a problem dictates the Tools/Attach Parts. For a Flexible attachment, this tool
creates beam elements up to the maximum distance
layout of the finite element mesh. For instance, the
defined in the attribute (it creates rigid links if using Rigid
following model consists of independent meshes from a
type) to coincide with the plane of the elements being
base plate, a round panel and a lug which are quite
attached to. A special link is also created at the end of
regular but are not compatible at the interfaces. The lug is
each beam element (an Attachment Link) which directly
supposed to be welded onto the round panel while the
couples the displacements at the node on the beam to the
round panel is welded around its edges onto the base
shape functions on the appropriate plate element. The
plate.
link ensures that the connection is a fully compatible one

www.strand7.com 6
Making Finite Element Analysis Easier

that takes into account all the active degrees of freedom

on the beam and on the connecting plate. The link
couples the two together and transfers loads into the plate
element from the beam element.
With the attachments created, a lifting load applied to the
lug hole not only lifts the lug, but also deforms the
attached part underneath.
In summary, attaching parts together is a two-step process:
1. Apply the appropriate type of attachment attributes to
the elements,
2. Attach the parts together by using Tools/Attach Parts.
Note that these attachments act to make the motion
between the two bodies compatible and represent the
stiffness of the joint. They should generally be used in
cases where you need to transfer load. For a detailed
analysis of the connection itself, a compatible mesh is
preferable
Free Body Diagrams
How to get Free Body Diagrams without a
Submodel
E xtracting forces and moments from simple models
made from beam elements has always been
straightforward with Strand7. All that needs to be done is
draw a force/moment diagram and take the values at the
required locations.
In previous versions of Strand7, to extract a total of the
forces and moments carried by multiple beam, plate and
brick elements in a cross-section it was necessary to
create a submodel (creating equivalent prescribed
displacements), analyse the submodel and then use the
Peek tool to calculate the force and moment total at the
restrained nodes. If several locations needed to be
investigated, then a submodel was required for each one.
A new calculated result is available for all element types
in Strand7 R24. This is the Element Node Force quantity,
which can be displayed as a Vector plot. Before
analysing the model, ensure that there is a node or a line
of nodes at the location where the forces and moments
are to be calculated.

Did You Know?

Perspective in 3D Drawing
Strand7 allows you to change the perspective setting in a 3D
model (View/Options/Drawing) so that you can see the 3D
model on a plane view (say YZ plane view) more clearly. The
setting is simply a number that ranges from 0 (no perspective)
to 200 (very high perspective, like a fish-eye lens). It can be 1. In the solver, calculate the Element Node Force
very useful when viewing/editing a complex 3D model. results.

Isometric view with YZ plane view with YZ plane view with

Perspective set to Perspective set to Perspective set to 2. Hide the unwanted elements. Groups are often used
25 (default) 25 (default) 0 (parallel view -
for this.
no perspective)
3. Show a Vector plot of Node Force.

www.strand7.com 7
Making Finite Element Analysis Easier
4. Select the nodes to be used for the calculation.
5. Use the Peek tool to calculate the force and moment
sum.
The moment summation needs to be about a specific
point, the default being the Global origin. If there is a
node in the model which should be used, click the Node
box to activate the Hotpointer and select the relevant
node. If there is no node then the co-ordinates can be
entered manually. In this example, the node in the
middle of the section was used.
The Peek tool allows you to sum the selected nodes, the
unselected nodes or all the nodes. The option to include
all nodes is a useful way to check that a group of
connected elements is in equilibrium (zero total reactions
should be calculated for a static problem).
In a second example, a brick mesh was used to model a
square hollow section beam defined as follows:
L  10m
A  0.36m2
  7870kg / m3
www.strand7.com
The beam was simply supported at each end at the centre
of the hollow square section.
The applied loads were gravity, g  9.80665 m/s2 and an
axial pressure at the end, P  100 kPa putting the beam
in tension.
Force/Moment Calculations
Load per unit length due to gravity
w  gA  9.80665  0.36  7870  27784.2 N / m2
Maximum bending moment expected in the middle of the
simply supported beam
wL2 27784.2  102
M   347302.5 Nm
8 8
Axial force expected in the beam
F  PA  100000  0.36  36000N
The model’s Element Node Force results were calculated
and the axial force Vectors plotted for half of the model.
The nodes in the middle of the beam were selected and
the Peek tool opened. There was no node in the middle
of the beam cross-section, so the About… location was
entered manually in the Global XYZ system.
The force and moments summation shows a total axial
force of 36000N and the bending moment 347302.5Nm,
both agreeing exactly with the expected results.
This method can also be used in models where there are
beams, plates and bricks at the cut. It is a dynamic tool in
that as you hide elements, the nodal forces and moments
contributed by the hidden elements are automatically
subtracted from the remaining elements shown in the
model window. Conversely, as elements are added to the
display, the resulting vectors automatically update to add
back the element's nodal force contribution.
When the whole model is shown, reaction vectors are
expected only at restrained nodes. For nonlinear analysis,
the reaction vectors are also a good way of visualising any
unbalanced force remaining in the model due to the
iteration process
8
Making Finite Element Analysis Easier
Factor vs Position Table
I n Strand7 R24, a new type of table, Factor vs Position
table, is available. There are two typical applications of
this table. The first is for applying a complex pattern of
attributes, such as distributed loads on beams, edge
pressures on plates, etc. The second is for modelling
situations where forces depend on position, such as
buoyancy. The following examples demonstrate these
two applications.
Applying Attributes of a Complex Pattern onto
a Model
Often one needs to apply loads which are of a complex
pattern such as surcharge loads on soil or wind pressure
on a building. If the complex pattern does not conform to
a relatively simple equation so that the functions in
Strand7 can be used to define it, the Factor vs Position
table can be a very helpful option.
Suppose that we want to apply a surcharge pressure load
onto the soil modelled as plates as shown in Figure 5.1.
The distribution of the surcharge load is difficult to
describe with a simple equation and so we use the Factor
vs Position table to perform the task.
Figure 5.1: Soil model with surcharge load.
Firstly, we define a Factor vs Position table by using
Tables/Factor vs Position to represent the pattern to be
applied as shown in Figure 5.2.
Figure 5.2: Defining a Factor vs Position table to represent the loading
pattern.
www.strand7.com
Next we assign the surcharge pressure described by the
table curve as plate edge pressure by referring to the
defined table with an expression of the form PT[1](x)
where PT represents Position Table, 1 is the table
identification number, and x is the X coordinate at the
point being interpolated (note here that UCS1 is the
current coordinate system and so the table refers to this
system as it interpolates x from the horizontal axis of the
table - see Figure 5.1). This expression, PT[1](x), is
entered directly into the attribute dialog edit box.
The Factor vs Position table follows the extrapolation
conventions of the other table types. Therefore, when the
to-be-applied length is outside the range of the defined
table, extrapolated data will be used.
Used for Modelling Buoyancy
Suppose we want to find the steady state position and the
internal stress distribution within the structure of a loaded
boat when it is stationary in water. How can we do this
using Strand7?
The boat is in a stable (equilibrium) state under the action
of gravity (self weight of the boat plus cargo) and
buoyancy force.
The challenging part is the modelling of the buoyancy
force. According to Archimedes' principle, the buoyancy
force depends on the submerged volume (or depth) of the
boat. For a curved hull, it is not easy to work out this
submerged depth (i.e. where the water line is on the hull).
By using the Factor vs Position table together with the
nonlinear transient dynamic solver, this problem can be
solved without much difficulty.
For the sake of simplicity and clarity, a simple example is
used to demonstrate the modelling of buoyancy. More
information and examples can be found in the Strand7
Verification Manual (VNT11) and in the Online Help
(Special Topics\Factor vs Position Table\Example 2).
Let’s consider the 2D plate model shown in Figure 5.3,
which represents the cross-section of a long block being
lowered into the water.
Figure 5.3: 2D plate model at initial position representing the cross-
section of the block.
9
Making Finite Element Analysis Easier

The whole block consists of two types of material as

indicated by blue and red properties. The blue part has a
negligible density and the whole weight of the block (set
to be half of the total volume of the block multiplied by
the density of the water and acceleration due to gravity) is
concentrated in the red part. Initially the bottom face of
the block is in contact with the surface of the water
marked by the X axis of the UCS, i.e. the water level is at
Y=0 in the UCS reference system. Our task is to find the
final position of the block in the water.
As mentioned earlier, floating forces depend on the
submerged depth. This position dependence can be
modelled by using the Factor vs Position table which
scales the element load attributes according to the
Figure 5.5: 2D plate model with unit edge pressure.
position of the element with respect to the reference level.
We first define a Factor vs Position table as shown in
Figure 5.4. This table represents the relationship between
the hydrostatic pressure and the depth, i.e. p  gh
where p is the pressure,  is the density of water (1000
kg/m3 in the example), g is the acceleration due to
gravity (10 m/s2 in the example) and h is the depth (the
position of a point of the plate in the Y direction of the
UCS reference). We then apply (Attributes/Plate/Edge Figure 5.6: Assign the table to the unit pressure load case.
Load/Pressure) a unit edge pressure on the four edges of
the plate (the plate may be submerged totally into the Solving the model using the nonlinear transient dynamic
water therefore the pressure is applied to all four edges, solver, the final position of the block in the water and its
see Figure 5.5). By associating this unit pressure with the stress distribution are calculated. Since the block
defined table on the nonlinear transient dynamic solver undergoes large displacements, the Nonlinear Geometry
(Solvers/Nonlinear Transient Dynamic/Load Tables), option on the solver panel should be set. Also, damping
buoyancy force will be automatically handled by the is needed to allow the model to eventually come to rest
transient solver (see Figure 5.6). Note that in the table in so that the block reaches a stable state. Rayleigh damping
Figure 5.4, when the position is greater than zero, the is used for this example (Solvers/Nonlinear Transient
factor is zero. That is, if any part of the edges of the plate Dynamic/Added Damping/Rayleigh).
is above the water level, the pressure on it will be zero.
Also note that in Figure 5.6, the Factor vs Position table is
assigned to axis 2 (the Y axis) of UCS1 system. The table
can only be assigned to one axis of a reference system in
one load case, but it can be used by more than one load
case.

Figure 5.7: Displacement curves at node 3.

Figure 5.7 shows the vertical and horizontal

Figure 5.4: Factor vs Position table representing buoyancy force.
displacements at node 3 (at the bottom right corner of the
red part). Initially, gravity load is greater than the
buoyancy force and so the block quickly moves
downwards. Inertial force will make the block move
beyond the static equilibrium line. With the buoyancy
force increasing, the block starts to move back up. After a
period of oscillation, the block finally reaches its static
equilibrium position as indicated by the level part of the
displacement curves. At this position the weight is
balanced by the buoyancy force. Note that due to the
uneven distribution of the mass, the block also rotates and

www.strand7.com 10
Making Finite Element Analysis Easier

swings as it sinks. Figure 5.8 shows the final stable
position of the block. Click Figure 5.9 for an animation
showing the whole process.
How does your program find ST7API.DLL?
Windows searches for a DLL in the following order:
1. The directory where the application executable is
located.
2. The Windows system directory.
3. The Windows directory.
4. The current directory.
5. The directories listed under the Path environment
variable.
For computers with multiple installations of Strand7, it
may not be obvious as to which ST7API.DLL is being
used by your program. Two API functions are available to
help identify the DLL in use. The function St7GetPath
returns the path to the DLL and the function
St7APIVersion returns the Strand7 version of the DLL.

Figure 5.8: Final stable position of the block.

Did You Know?
Beam Local and Principal Results
Cross sections in beam elements
contain both local and principal
axes and they do not necessarily
coincide (e.g. in the case of
unsymmetrical or user defined
sections). For example, the
principal axes of an equal angle
section are oriented at a 45
degree angle from its local axes.

To help make the

extraction of beam
results more versatile,
Strand7 reports beam
forces and moments in
Figure 5.9: Animation of process. Click to play. both the local (x,y) and
principal (1,2) axes.
It can be seen that the heavier red part is at the lowest For example, the highest
level and half of the block is submerged into the water bending moment at the
root of the cantilever
beam shown is reported
in plane y if the load is
API Troubleshooting applied in the Global Z
direction.

W hen an API application is launched, Windows

searches for ST7API.DLL so that the entry points of the
API functions can be resolved. When Windows is unable
to locate the DLL, an error message such as the following
is usually displayed:

www.strand7.com 11
Making Finite Element Analysis Easier

How can I make my program find ST7API.DLL? Training

Do not copy the DLL. This approach does not work as the
DLL relies on other Strand7 installation files. By copying
the DLL, the API is no longer able to locate the other files
T he training calendar for the remainder of 2010 is still
that it depends on. being finalised, but the following courses have so far been
scheduled:
The recommended approach is to update your Path
environment variable so that the Strand7 installation is Date Course Title Location
found regardless of where you run your program from.
11-13 October Strand7 Essentials Cambridgeshire,
To update your Path environment variable, complete the UK
following steps:
14-October Nonlinear Analysis Cambridgeshire,
1. Click Start and select Control Panel. Double click with Strand7 UK
System.
15-October Dynamic Analysis Cambridgeshire,
2. If you are using Windows Vista or Windows 7, click with Strand7 UK
the Advanced system settings link. If you are using an
older version of Windows, simply choose the To register, please go to www.strand7.com and follow the
Advanced tab. links.

3. Click Environment Variables. Abu Dhabi Training

4. Under System Variables, scroll down to Path and
click Edit.
5. Identify the Strand7\bin directory path. For a local
installation this might be C:\Program Files\Strand7
R24\bin. For a network installation this might be
\\Server\Apps\Strand7 R24\bin or Z:\Apps\Strand7
R24\bin.
6. Modify the Path environment variable value string by
adding the Strand7\bin directory path identified in the
previous step. Please note that values in this string
need to be separated by semi-colon (;) characters.
Take care not to remove existing paths from this
string.
We are also planning to run a Strand7 Introductory and
Advanced Training course in Abu Dhabi, UAE, in late
November or early December. This course will only go
ahead if there is sufficient participation. If you are
interested in attending, please contact us via
info@strand7.com
Exhibitions
B etween now and end of 2010, we will be exhibiting
Strand7 at the following events:

18-20 Oct The Infrastructure Show NEC, UK

21-22 Oct Civil Expo 2010 Seoul, Korea
3-5 Nov SNAME Seattle, USA

Links to these exhibitions are available on our website.

Other exhibitions will be announced soon on the Strand7
website and in the next edition of News.St7

Head Office
Strand7 Pty Ltd
7. Click OK and close the System Properties dialog box.
Suite 1, Level 5
8. Restart the computer.
65 York Street
Sydney NSW 2000
Still having problems? AUSTRALIA
If you still receive an error message relating to a missing Tel +61 2 9264 2977
ST7API.DLL after completing the above procedure then Fax +61 2 9264 2066
you can download a Strand7 API diagnostic tool from Email info@strand7.com
www.strand7.com/news.st7.htm Web www.strand7.com

www.strand7.com 12
Strand7

News.St7 Newsletter for Strand7 and Straus7 users

Strand7 is marketed as Straus7 in Italy

Issue 9, July 2013

Welcome to another issue of News.St7. In this issue we highlight some of the new In this Issue
features in the recently released Strand7 R2.4.5. We also report on the very
successful launch of the Strand7 Webnotes – a comprehensive online library of  Strand7 Release 2.4.5 1
resource material available to supported Strand7 users. As usual, we include
interesting technical content and Did You Know? items. We also profile a new  Strand7 Webnotes 2
Strand7 API user development for the design of steel members to AS 4100.
 Analysis of an Exhaust System 3
We hope you enjoy reading this issue of News.St7, and look forward to hearing
 Import and Export 4
your comments. We also welcome suggestions for topics you would like to see in
future issues.  Glass Façade Blast Analysis 6

If you have any feedback or suggestions regarding the content of News.St7 or you would like to  Strand7 API Developer 7
contribute a case study showing how you use Strand7, please email us at news.st7@strand7.com. If you
would like to receive your copy of News.St7 automatically by email, simply send a blank email to  Training and Exhibitions 8
newsletter-subscribe@strand7.com. All care is taken to ensure that information in News.St7 is accurate
and up to date at the time of publishing. However Strand7 Pty Limited accepts no responsibility for  Website Update 8
inaccuracies in, or changes to, such information.

Strand7 Release 2.4.5

Strand7 R2.4.5 is now available for download from our
update site, www.strand7.com/st7update. It enhances
R2.4.4 with greater stability, improved nonlinear
convergence, and a few subtle user-interface
enhancements such as better support for multiple monitor
environments.
R2.4.5 also addresses a number of issues in Strand7
R2.4.4, which have been brought to our attention since its
release; so please do take advantage of your release
entitlement and download this worthwhile update.
Did You Know?
Selecting All Entities Behind the Selection Box
One new feature in R245 is the ability to select both visible
and obscured entities by using Ctrl+Alt when dragging a
selection box.
Instead of selecting only those entities which are visible from
the current view point, all entities behind those shown are
also selected.
This alternative selection mode is graphically shown during
selection on the screen as a solid double line.

Changes in R2.4.5
The following is a summary of the changes. For a
complete list, please see the document What’s New in
Strand7 R24.pdf, located in your Strand7 installation
folder.
1. New Selection Mode – Selection All Behind
Holding the Alt + Ctrl keys when selecting now selects all entities
behind the selection box, even if you can’t see them. This selection
mode is indicated by a solid double line when selecting.
Note that the previously available selection mode, Line Selection, can
still be used. By pressing the Alt key (without the Ctrl key) when
dragging, the selection box becomes a line, which is very useful for
selecting entities inclined in the current view.
See Main Tool Bar/Select/Individual Select in the Strand7 Online Help
for more information and examples.

www.strand7.com 1
Making finite element analysis easier.

2. New Groups Right-click Options 11. New API Functionality

In the Groups dialog, right-clicking on a group gives four options for The following new API functions have been added in R2.4.5:
expanding the sub-groups within: Expand All, Collapse All, Expand
Selected and Collapse Selected.  St7SetPropertyCreepID

 St7GetPropertyCreepID
3. Brick Mesh Internal Attachment Links
 St7EnableSaveLastRestartStep
When attaching elements with Tools/Attach Parts, the new option All
Faces for brick attachments allows the attachments to form between  St7DisableSaveLastRestartStep
elements embedded within a brick mesh and the interior faces of the
brick elements, not just the outer (or free) faces.  St7ShowSelectionToolBar

 St7HideSelectionToolBar
4. Beam Fibre Strain Increment Sub-Stepping Control
 St7SetSelectionToolBarPosition
For material nonlinear solutions involving nonlinear fibre stress beam
elements, a sub-stepping trigger has been added which can be used to  St7GetSelectionToolBarPosition
limit the incremental plastic strain a beam can experience in a single
load increment or time step. This helps convergence stability in MNL  St7DeleteGroup
problems.  St7GetInertiaReliefResults

5. Save Solver Restart Files Only Last Step Option  St7SetPlateLayers

When saving restart files, a new option Save only last step (Minimum  St7GetPlateLayers
size) is now available. This option is intended to allow a restart from the  St7EnableAutoAssignPathDivisions
last saved increment only, instead of at every saved increment. This can
save a significant amount of disk space in large models.  St7DisableAutoAssignPathDivisions

 St7EnableModelStrainUnit
6. Surface Automesher Enhancements
 St7DisableModelStrainUnit
The Skip Transitioning option has been renamed to Apply Auto
Transitioning. A new control Consider Nearby Vertices is added to  St7EnableModelRotationUnit
allow the user to consider or ignore adjacent vertices which are not
actually connected to the current face. Thus if there is a gap between  St7DisableModelRotationUnit
two parts, that gap would determine the local mesh size if the Consider  St7SetQuasiStaticRestartFile
Nearby Vertices option is set.
 St7GetQuasiStaticRestartFile
7. Automatic Load Path Divisions  St7SetNodeHistoryFile
This new option is found in the transient solver Moving Load… dialog.  St7GetNodeHistoryFile
When set, it overrides the Divisions attribute of the load paths, instead
making sure that there is at least one division where each vehicle may Additionally, API headers and examples for C# and
land at any time step. In other words, the load path divisions are
controlled by the time step. This enhances the accuracy of transient Python have been added to the R2.4.5 distribution.
moving load solutions by ensuring that enough divisions are taken along
a load path.

Load paths default to 20 divisions, with the option to manually change

Strand7 Webnotes
this with Attributes/Load Path/Divisions. This is still available, and if the
number of user-defined divisions is adequate for a particular solution, Strand7 Webnotes is a repository of documents and
the Automatic Load Path Divisions option will not make a difference. Strand7 model files, available online to supported Strand7
users. Many topics are covered, with cross-linked content
8. Seismic Load Cases Consider NSM Dynamic Factors between Webnotes. A Webnote is a PDF document or a
A new option Apply Dynamic Factors in seismic load cases specifies
ZIP file, if there are associated model files.
that the dynamic factors applied to the included non-structural mass
Webnotes are found on our Website at
should be considered in the equivalent static seismic analysis.
Previously, you could scale all non-structural mass via www.strand7.com/webnotes. The index of
the phi parameter. This new option allows NSM in all Webnote descriptions can be browsed
different load cases to be scaled independently. without entering a password, and is
searchable. Once a Webnote link is
9. Enhanced Import Functionality selected, the user is asked to enter a
SAP2000 imports now support SAP2000 TABLE: "AREA username and password to download the
LOADS - UNIFORM TO FRAME" loading. Additionally, content. This username and password is the
a new option, Merge Duplicate Freedom Cases is same as the one you would use to
available. Nastran imports now support SPCADD,
download Strand7 updates, and can be
MPCADD, PCOMPG, BAROR and BEAMOR bulk data
entries. Ansys imports now support the *GET command. found in your installation CD jacket, by
contacting your IT support, or by emailing
10. Updated Material and Section Libraries webnotes@strand7.com.
Composites libraries for Aramid, Carbon and Glass Fibre Each Webnote is prefixed by a unique ID
have been updated. A new library for Glass has been number which indicates the section and
added. Several South African section libraries have also sub-section. For example, ST7-1.10.20.2
been updated.
Harmonic Response Analysis is the second
document in major section Linear (10) and

www.strand7.com 2
Making finite element analysis easier.
minor section Dynamics (20). There are presently nine
major sections, each with multiple sub-sections.
Since our Webnotes launch on 16-May, we have been
receiving regular positive feedback about this initiative.
We are expanding the Webnotes library with new
releases posted regularly, so please check the website for
the latest additions, or email us to get on the mailing list
for Webnotes announcements.
We hope that providing online access to the Webnotes
will enhance the skills of our user base, as well as the
quality of the Strand7 support experience. We welcome
any feedback about this new system. Requests for topics
which are not yet covered can be directed to
webnotes@strand7.com. Contact info@strand7.com if
you are unsure of the support status of your company, or
if you wish to purchase support.
Analysis of an Exhaust System
This article presents a summary of the Webnote ST7-1.40.70.8 Tuning
an Exhaust System with Harmonic Analysis.
Automotive exhaust systems play a large role in the
comfort of a vehicle. How the exhaust system interacts
with the engine and car body can be determined using
harmonic response analysis. This type of analysis
determines the steady-state vibration response of a system
to a sinusoidally varying input, such as that from a car
engine. To provide maximum sound attenuation, the
exhaust system must be “tuned” such that the range of
forcing vibration of the car does not cause it to resonate.
Additional considerations include avoiding resonance
between the floor of the car body and the exhaust system.
Half symmetric exhaust system CAD geometry.
www.strand7.com
The example geometry is based on a V8 exhaust system.
It has two symmetric, centrally linked exhaust pipes. Each
has a catalytic converter, a resonator and a muffler.
The exhaust system is forced by pressure pulses and by
direct engine vibration at the front. The acoustic tuning of
the exhaust system (i.e. dimensions of the internal cavities
etc.) is assumed to have been previously performed to
attenuate acoustic energy within the exhaust gas.
The goal of the tuning procedure is to avoid known car
body floor resonance points of 65 Hz, 96-100 Hz, 106 Hz
and 127 Hz by at least 3 Hz, and to attenuate direct
engine vibration by at least -20 dB from idle to
3500 RPM, and by -10 dB above 3500 RPM as measured
at the rubber mounts. Additionally, the peak maximum
principal stress should be below the endurance limit of
the steel (assumed to be 250 MPa) to avoid fatigue
cracking.
Geometry is imported from CAD (using IGES format) and
mirrored to form the whole system. This is necessary
because vibrational modes which are not symmetric will
be present.
Strand7 geometry after import and mirror - note that the mirror operation
on the geometry has been performed in Strand7.
Normal pressure is applied as a function of distance from
the engine to model the drop in pressure from the header
to the muffler outlet.
Pressure load applied with variation along the length. In Strand7 it is
easy to apply attributes that vary with spatial position - simply enter the
required equation in any of the attribute dialogs ("0.02/2580*X" has
been used above to apply pressure that varies with X).
3
Making finite element analysis easier.

Catalytic converters are included as lumped masses

connected to the housings with interpolated multi-point
links.

Catalytic converter mass added using multi-point links. The advantage

of these links over rigid links is that they distribute the mass over the Harmonic response of the system to exhaust pulses.
connected nodes without affecting the structural stiffness. See Main
Tool Bar/Create/Link/Multi-Point Link in the Strand7 Online Help for a
simple example that compares the difference between these and the
rigid links.

Natural frequency results are obtained to determine the

natural vibrational response of the structure.

Harmonic response displacement vs frequency results.

Exchanging Geometry and Model Data

Strand7 imports and exports a variety of formats to
facilitate the sharing of data and the building of models
based on CAD geometry.
This functionality is found in File/Import and File/Export.
Two natural frequency mode shapes of the system.
The file type can be selected at the bottom of the dialogs,
as shown below:
The natural frequency results are used to produce a
Harmonic Response analysis. This gives peak results
including displacement, velocity, acceleration and stress
across the operational range of the structure.
This output is graphed and compared to the frequencies
which are to be avoided.

Strand7 can import and export the following file types:

www.strand7.com 4
Making finite element analysis easier.

1. IGES (*.igs, *.iges) 11. SAP2000 (*.s2k)

IGES files hold 2D and 3D CAD geometry. It is typically exported from This is the text-file model format for SAP2000, a structural design and
CAD and imported into Strand7 to use with the automesher. You can analysis program similar to STAAD-Pro. Use this format to import
also export Strand7 geometry into IGES format after modifying (or SAP2000 models. Available for import only.
generating) it in Strand7. The IGES format supported by Strand7 does
not hold finite element information – only CAD geometry. Available for
import and export. Did You Know?
2. ACIS (*.sat) Geometry Creation in Strand7
ACIS files are roughly equivalent to IGES files. Strand7 supports most of As an alternative to importing CAD geometry for automeshing,
the ACIS types, with a few exceptions such as rolling ball blends. you can create geometry in Strand7. This is done by creating
Available for import only. plate elements, and using Tools/Geometry Tools/Face from
Plate to convert them into geometry faces.
3. STEP (*.stp, *.step)
Curved plates (i.e. Quad8,
STEP is also equivalent to IGES, but is a newer ISO standards Quad9 and Tri6) can be used to
specification. If IGES isn’t importing your geometry well, try STEP or create curved geometry faces.
ACIS. STEP can also be used to export geometry from Strand7 into
Set the option Circular Face
CAD. Available for import and export.
Edges to create geometry faces
on underlying cylindrical
4. Strand7 Text File (*.txt) geometry. Otherwise, the
The Strand7 text file format is a Strand7 model in readable ASCII (text) quadratic shape of the original
format. This can be used to pass newer Strand7 models into older plate is retained for the
versions of Strand7 (e.g. opening an R2.4.x model in R2.3.x). It may geometry and the face becomes
also be useful if you want to manually modify specific parts of the model a b-spline.
using a text editor. This format only holds element information, so no
CAD geometry is transferred using this format. Available for import and
export. This functionality can be used to patch holes in pre-existing
geometry, or to build geometry to automesh in Strand7,
5. Strand6 Binary File and Text File *.dir, *.txt, *.asc independently of CAD software.

These import types can be used to translate Strand6 models into Strand7
format. Strand6 was the precursor to Strand7, and was DOS based.
Available for import only.

6. AutoCAD DXF (*.dxf)

The DXF format is used by Strand7 to import simple geometry which is
represented by plate and beam elements. This format is not appropriate
for importing geometry for automeshing but can be very useful for
transferring wireframe models such as beam frameworks. Exporting a
model in this format will produce a file which AutoCAD can read as a
3D model. If you plan to automesh, don’t use this format – use IGES,
ACIS or STEP. Available for import and export.
These faces can then be automeshed, or exported to CAD
7. Stereo-Lithography (*.stl, *.stb) using File/Export and selecting IGES or STEP formats.

This file format is a triangulated surface representation of an object. The

triangles (represented by Tri3 elements in Strand7) are generally not
constrained by quality concerns if sourced from a typical CAD package,
and so may have very small angles, which makes them undesirable from
a finite element perspective. This format has been used to exchange
data from biomedical applications, such as MRI scanned solids.
Available for import and export.

8. MSC/Nastran (*.nas, *.dat, *.bdf)

This is the text-file model format for NASTRAN, a finite element code
which has been around for a very long time. The format is
correspondingly widely used, and can be used to exchange model
information with many FEA packages. Available for import and export.

9. ANSYS (*.dat, *.cdb)

This is the text-file model format for ANSYS, another finite element code.
Use this format to exchange models with ANSYS, or other programs
which can import ANSYS models. The import functionality includes the
option to import IGES entities embedded within the ANSYS file.
Available for import and export.
Note that it is also possible to convert planar polygons into
faces by using Tools/Geometry Tools/Face from Beam
Polygon.

10. STAAD-Pro (*.std)

This is the text-file model format for STAAD-Pro, a structural analysis
based finite element and design program. Use this format to import
STAAD-Pro models. Available for import only.

www.strand7.com 5
Making finite element analysis easier.

Blast Analysis of a Glass Façade

This article presents a summary of the Webnote ST7-1.40.35.1 Glass

Façade Blast Loading.

Glass Façade
construction is subject
to blast loading
considerations. These
unique structures
come in a variety of
custom configurations
suited to the particular
needs of each
installation. Due to
the variability of each Pre-blast static condition.
design, Finite Element
Analysis can be a very Small 1 ms time steps are taken to adequately resolve the
powerful and cost- blast load, which occurs over the first 14 ms of the
effective tool in the analysis. Only every second calculated time step is saved
qualification of these during this phase.
structures.
The deformed structure at 90 ms demonstrates the
Four load cases represent the loads on the structure: whipping action of the glass as it reverses direction due to
1. Self-weight (1g load) loading from the supporting horizontal frames.

2. Blast Pressure (6.894 kPa)

3. Cable Pre-Tensioning (20 kN)
4. Wind Load (0.478 kPa)
A load table is defined to scale the blast pressure through
time in the nonlinear transient dynamic analysis. The
actual blast pressure is factored by 5.28, and decreased to
zero over a period of 14 ms.

Deformed shape and stress at 90 ms.

Blast pressure factor vs time table.

The nonlinear static solver is run before the blast load is

applied. The results for the Self-Weight + Wind case
(pre-blast static condition) are shown below for the glass
and horizontal frames. Deflection has been exaggerated
for clarity. Graph of glass stress time history.
The nonlinear transient dynamic solver requires the
definition of loads to apply, the tables that factor the An envelope of the glass stresses through time is easily
loads, and a time stepping scheme. The time stepping created (using Results/Envelopes), and shows the regions
depends on the nature of the problem, and in this case we of the façade which exceed the 65 MPa limit (shown in
will choose to take 100 x 1 ms steps, for a total simulation grey in the figure below).
time of 100 ms. Additionally, we must specify any initial
static conditions for the façade.

www.strand7.com 6
Making finite element analysis easier.
Envelope of maximum blast stresses through time.
Strand7 API Developer Profile
The Strand7 API (Application Programming Interface)
offers Strand7 users the ability to interact with Strand7
through a computer program instead of the Strand7
Graphical User Interface (GUI). The computer program
can be a compiled program created from a traditional
programming language such as C, Fortran, Pascal, etc., or
an interpreted program within environments such as
Matlab, Python, Excel (with VBA), amongst others.
The Strand7 API can be used for a range of tasks such as
automating repetitive operations, building Strand7 models
in a parametric way, extracting results in customised
formats, performing automated design checks, or for
developing complete applications where FEA forms part
of the process.
Many Strand7 users have discovered the benefits of the
Strand7 API in recent years for their day-to-day analysis
tasks, and some users have also developed commercial
applications based on the Strand7 API.
One such development is the program marketed as
BCDsteel, which offers steel design capabilities of Strand7
models to AS4100. BCDsteel has recently become
commercially available. The following article has been
written by Paul Cain, the developer of BCDsteel. As
BCDsteel is not a product developed by Strand7 Pty Ltd,
anyone interested in the product should visit the official
product website www.BCDsteel.com.
If you have developed a Strand7 API product and you
would like Strand7 users to know about it, please send an
email and a short article to news.st7@strand7.com and
we will consider publishing it in a future edition of
News.St7.
Steel Design to AS 4100
by Paul Cain, developer of BCDsteel.
BCDsteel has been developed independently of Strand7
and is a good example of the professional quality software
that can be created using the Strand7 API.
www.strand7.com
BCDsteel has been in use in a number of consulting
offices throughout Australia since 2010. Earlier this year,
with the release of Strand7 R2.4.5, BCDsteel became
commercially available. It is capable of analysing all
types of steel structures from simple building frames to
complex models of machines like the shiploader shown
below.
BCDsteel uses the API to display a Strand7 model
window very similar to Strand7, including most of the
Strand7 tools used for displaying a model. A separate
panel on the left of the display contains steel design tools.
This panel changes to define members, select load
cases/combinations to analyse, and for displaying results.
The program is simple to use - it is laid out in a logical
order from input through to results. Most controls display
a Windows hint when the mouse pointer rests on the
control, with F1 displaying a context sensitive Help file
for additional information.
BCDsteel analyses steel members that have been
modelled using beam elements. However, models like
the shiploader above typically also include string groups,
links, plates and brick elements, often have multiple
freedom cases, and sometimes staged analyses.
In general, each steel design “member” includes a
number of beam elements in a row, with different
restraints at the nodes along the member. The program
will automatically define all members, restraints and
effective lengths in a model. This is based on the
geometry of the structure taking into account node
restraints, beam end release, orientation/alignment, etc.
However, it is important for the user to check that the
members are correctly defined. Tools are provided to
modify members, restraints and/or effective lengths where
necessary.
Some features are listed here, some of which are possibly
unique to BCDsteel:
 Members may be tapered, and may even have a totally
different section type along their length.
7
Making finite element analysis easier.

 Section capacities are calculated for slender sections Date Course Title
by omitting the ineffective width from each "flat plate 20-23 August 2013 Introducing Strand7
element". These are calculated for bending about
both the major and minor principal axes plus axial 26 August 2013 Automeshing with Strand7
compression. 27-28 August 2013 Structural Analysis with Strand7

 Star equal angle sections used for bracing are analysed
to Clause 6.4.3.
 Equal angle web members in trusses can be analysed
in accordance with Clause 8.4.6. This can result in
significantly higher capacity than Clause 8.4.5.
The following figure shows the range of section types
presently supported. Fabricated RHS can have up to 3
longitudinal stiffeners per face.
29 August 2013 Nonlinear Analysis with Strand7
30 August 2013 Dynamic Analysis with Strand7
To register, please go to www.strand7.com and follow the
links.
We are also planning to run Strand7 courses in the UK
later in the year. If you are interested in attending, please
contact us via info@strand7.co.uk

Exhibitions
Between now and the end of 2013, we will be exhibiting
Strand7 at the following events:
Date Exhibition Location
18-21 Sep 2013 NCSEA Atlanta, GA,
Booth 14 USA
The program uses the Strand7 API to get axial forces,
bending moments and shear forces from the Strand7 25-26 Sep 2013 TCT Live Birmingham,
Stand F36
model. The analyses generally use the highest tier UK
equations in AS 4100, with some exceptions to simplify 2-3 Oct 2013 Engineering Design Show Coventry,
user input. The program will analyse linear and nonlinear Stand C75 UK
results. The members are coloured to indicate the Load to
6-8 Nov 2013 SNAME Bellevue, WA,
Capacity Ratio, which must be < 1.00 to comply with AS Booth 314 USA
4100.
BCDsteel has been verified where possible by comparison Links to these exhibitions are on our website. If you are
with the Design Capacity Tables published by the able to attend, please come to our booth and say hello -
Australian Steel Institute and with Limsteel. Many aspects we would welcome the opportunity to meet you.
such as slender section capacity, the equal angle solvers,
flexural torsional buckling of struts, and tapered member
capacity have had to be verified against data generated Website Update
specifically for this purpose.
A solver for the Eurocode EC3 is presently under Some updates to the Strand7 website have recently been
development. The section capacity calculator has been implemented, coinciding with the Webnotes launch.
completed and this also includes Class 4 (slender) Each page now has a unique URL, which allows you to
sections. bookmark locations within the page, and share specific
links.
To run BCDsteel, Strand7 users will need to have a copy
of the Strand7 API module. Further details including Additionally, a custom Google search has been installed
pricing and trial software are available at for quick searching of all Strand7 content.
www.BCDsteel.com.

Strand7 Training
Head Office UK Office
The Strand7 training calendar is regularly updated and Strand7 Pty Ltd Strand7 UK Ltd
published at www.strand7.com, so please check our site Suite 1, Level 5 Suite 3, 1st Floor
periodically. If you are on our mailing list you will also 65 York Street The Stables
be kept informed by email whenever new courses are Sydney NSW 2000 6 Church Street, St. Neots
scheduled. If you are not on our mailing list but would Australia Cambridgeshire PE19 2BU
like to be, please email info@strand7.com. United Kingdom

At the time of publishing this newsletter, the following Tel +61 2 9264 2977 Tel + 44 1480 211 011
courses were scheduled to be presented in our office in Fax +61 2 9264 2066 Fax + 44 1480 211 020
Sydney: Email info@strand7.com Email info@strand7.co.uk

www.strand7.com 8