Presel UM
Presel UM
USER MANUAL
Presel
Presel
Preprocessor for Assembling
Superelements
All rights reserved. No part of this book may be reproduced, in any form or by any means, without permission in
writing from the publisher.
Published by:
DNV GL
Veritasveien 1
N-1322 Høvik
Norway
Telephone: +47 67 57 99 00
Facsimile: +47 67 57 72 72
E-mail, support: [email protected]
Website: www.dnvgl.com
If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of Det Norske Veritas, then Det Norske Veritas shall pay compensation to such person for his proved
direct loss or damage. However, the compensation shall not exceed an amount equal to ten times the fee charged for the service in question, provided that the maximum compensation shall never exceed USD
2 millions. In this provision “Det Norske Veritas” shall mean the Foundation Det Norske Veritas as well as all its subsidiaries, directors, officers, employees, agents and any other acting on behalf of Det Norske
Veritas.
Table of Contents
1 INTRODUCTION ............................................................................................................1-1
2 FEATURES OF PRESEL................................................................................................2-1
1 INTRODUCTION
Presel is SESAM’s preprocessor for assembling superelements. A superelement is basically a finite element
(FE) model of a part of the complete structure. The SESAM preprocessors Patran-Pre (for general FE mod-
elling), GeniE (for concept modelling), Preframe (for frame modelling) and Prefem (for general FE model-
ling) are used for creating first level superelements, i.e. part models comprised of finite elements like beams,
shells, etc. adjoined in nodes.
Presel is used to put these parts together to form superelement assemblies. First level superelements are typ-
ically assembled to form second level superelement assemblies, these are in turn assembled to form third
level superelement assemblies, and so on until the complete model — the top level superelement assembly
— is formed. Note that a superelement assembly is referred to as a superelement when itself is included in
an assembly. Superelements on levels higher than 1, i.e. superelements created by Presel rather than by the
preprocessors, are often referred to as higher level superelements.
A superelement hierarchy showing the assembly process from bottom to top can be sketched as illustrated
for the two examples in Figure 1.1.
A superelement model may be used for linear structural analysis using Sestra, or hydrodynamic analysis
(wave loading and motion) of frame structures using Wajac and of general structures using Wadam.
First level superelements created by SESAM’s preprocessor for tubular joint modelling, Pretube, may also
be assembled by Presel. The same goes for superelements created by any FE preprocessor as long as they
are defined according to the SESAM Input Interface File format.
To fully appreciate the purpose and use of Presel an understanding of the mathematical foundation of the
superelement technique is required; see Appendix B THEORY. Also a basic understanding of a first level
superelement as modelled by e.g. GeniE and Patran-Pre is required.
Presel SESAM
1-2 15-MAY-2014 Program version 7.5
1.1
1.2
If you already have a basic understanding of SESAM and the superelement technique then proceed as fol-
lows:
• Read Section 3.1 to learn how to start Presel and use the graphical user interface.
• Read Section 3.2.1 to learn about the basic procedure of assembling superelements.
• Read Section 3.2.3, Section 3.2.4 and Section 3.2.5 to learn how to identify superelements and nodes.
Note two important rules mentioned in Section 3.2.6.
• Read Section 3.3.1 and study the tutorial in combining loads in Section 3.3.2.
• Conclude by looking into Section 3.3.3, Section 3.5, Section 3.8 and Section 3.9.
Chapter 4 EXECUTION OF PRESEL contains more special information not intended for the new user
using Manager to control his SESAM analysis. The chapter explains how to start Presel outside Manager
and operate it in line-mode (not using the graphical user interface). The files used by Presel are also
explained. Practical information is provided on how to operate Presel and manipulate its files in various
ways. Built-in and hardware dependent requirements and limitations are also described.
Chapter 5 COMMAND DESCRIPTION explains in detail all commands of Presel. The commands and sub-
commands are sorted alphabetically.
Appendix A TUTORIAL EXAMPLES provides input (to Prefem) for creating the first level superelements
used in the tutorials of Chapter 3 USER’S GUIDE TO PRESEL.
• New features
• Etc.
To look up information in the most updated version of the Status List go to the support page of our website,
click the SESAM Status Lists link and log into this service. Contact us for log-in information.
SESAM Presel
Program version 7.5 15-MAY-2014 2-1
2 FEATURES OF PRESEL
Presel is a program for assembling superelements (part models) to form the complete model to be analysed.
The first level superelements may be displayed and given names but not modified in any way. If a first level
superelement need to be modified you have to revert to the preprocessor that created it, e.g. Patran-Pre or
GeniE. A first level superelement cannot be modified once it has been read into Presel.
Higher level superelements are created in Presel. A higher level superelement contains no nodes and ele-
ments until one or more superelements have been included into it. A higher level superelement is given con-
tents by including first level superelements and/or previously created higher level superelements. There are
no restrictions as to the number of superelements that may be included in an assembly. Superelements at any
level may be included in the same assembly. The assembly will implicitly be assigned a superelement level
equal to the highest level among its included superelements plus one. For example, if the highest level
among the included superelements is three then the assembly will be a level four superelement.
Boundary conditions like supernodes and fixations as well as loads are defined for higher level superele-
ments (assemblies). The loads will normally be combinations of loads previously defined for the included
superelements but nodal forces may also be defined.
Once a higher level superelement is included in yet a higher level superelement it can no longer be modified
in any way, for example by including more superelements or defining boundary conditions and loads.
The top (highest) level superelement is complete when the whole model to be analysed has been formed
including all its loads and boundary conditions. There is no limit to at which level this top level is. The top
level number is simply a result of the hierarchical assembling of superelements from bottom to top.
The final task of Presel is to produce (write) the Input Interface Files (T#.FEM, the T-files) for all higher
level superelements that has been created.
Presel SESAM
2-2 15-MAY-2014 Program version 7.5
LABEL annotates node symbols, node numbers etc. to the displayed su-
perelement.
PLOT generates a plot file of the current display. The plot file may be
printed or imported in a word processor. In a MS Windows en-
vironment the plot may also be directed to an on-line printer.
PRINT prints tables over model data. The tables may be directed to the
screen or to a file by the SET PRINT command.
TASK switches to input mode for Submod; see the Submod User
Manual. (In most cases you will find it more convenient to ac-
cess Submod directly rather than through Presel.)
WRITE writes Input Interface Files for higher level superelements. See
Section 2.3 on this.
EXIT exits from Presel. The model and log files are saved and closed.
The Input Interface File, the T-file, is a sequential ASCII character file with 80 character long records. The
straightforward definition of the file enables external programs to be connected to the SESAM system with
comparative ease.
One interface file will be created for each superelement. The name of the file will be:
prefixT#.FEM
where:
• ‘prefix’ is an optional character string that may and may not include a directory specification, the string
is given when entering Presel and is common for all superelements in the model.
• ‘T’ is a mandatory character identifying this as an Input Interface File, a T-file, as opposed to a Loads
Interface File, L-file, which uses character L and a Results Interface File, R-file, which uses character R.
Normally, the user may find it convenient to leave the prefix void. This is also the default condition.
Note that Presel both reads first level superelements and writes higher level superelements and all must have
the same prefix. If the above file — superelement 5 — is one of several files of a superelement model then
all Input Interface Files should be named ABCT#.FEM, where # is the superelement number.
2.1
PRE-
PROCESSING
ENVIRONMENTAL
Prefem ANALYSIS
INPUT INTERFACE FILE
Presel produces Input Interface Files for all higher level superelements by the command WRITE:
WRITE top#
where top# is the number of the top level superelement. All higher level superelements are implicitly and
automatically written when this command is given. Note that when the SESAM Manager is employed the
writing of the top and higher level superelements is normally controlled by Manager thus making the
WRITE command superfluous.
Note: If you on MS Windows close the Presel window by the X in the upper right corner (or by the
Close (Alt+F4) command of the window menu) then the Input Interface Files will not be writ-
ten even though you have requested this when starting Presel. This feature may be used if you
change your mind and decide not to write the file after having started Presel.
SESAM Presel
Program version 7.5 15-MAY-2014 3-1
The purpose of Presel is to assemble (put together) part models, so-called first level superelements, to form
the complete analysis model. This is done through the following steps:
• Create a second level superelement assembly and include first level superelements into it. The inclusion
process comprises positioning the first level superelements and coupling matching supernodes.
• Combine loads for the assembly by referring to loads of the first level superelements. Possibly also
define some boundary conditions.
• Create higher level superelement assemblies, combine loads and define boundary conditions in the same
way as explained above only now include previously created superelement assemblies. Repeat this proc-
ess until the complete analysis model has been created.
• Store all superlement assemblies (the whole superelement hierarchy) on file and exit Presel (the storing
is normally done automatically when exiting Presel).
• Get started using the graphical user interface. See Section 3.1.
• Establish the loads for the complete model by combining loads on superelements. See Section 3.3.
• Assemble loads, an alternative and advanced way of establishing the loads. See Section 3.4.
• Use the superelement technique in a practical way to efficiently solve problems. See Section 3.9.
• Optimise node numbering (reduce bandwidth of the stiffness matrix) and when to do it. See Section 3.10.
• Click Options | Superelement and within the window appearing set your analysis to be superelement
analysis (as opposed to direct analysis) and specify top level superelement number (any number may be
used).
• Start Presel by clicking Model | Superelement Handling Presel and click OK in the window then
appearing.
The main part of the graphical user interface is the graphic-mode window. There are also a print window
and a message window. Print requested by the user appears in the print window whereas various program
messages appear in the message window. Figure 3.1 illustrates the three Presel windows.
3.1
Graphic-mode window
Print window
Message window
Presel offers two modes of input and both are available in the graphic-mode window:
• Line-mode input, i.e. typing commands and data using the keyboard
• Graphic-mode input, i.e. selecting commands by clicking the left mouse button (LMB)
A sketch of the graphic-mode window is shown in Figure 3.2 together with explanations of the six different
areas. How to use the areas is explained in more detail in the following.
SESAM Presel
Program version 7.5 15-MAY-2014 3-3
3.2
click left mouse button (LMB) to select command or action
Direct
access Shortcut Command
buttons commands menu
<last given input>
<prompt> <echo> Line-mode input Cursor position feedback
You may at this stage decide to go through a Presel tutorial. Go then to Section 3.2.2 and use the explana-
tions below of the areas of the graphic-mode window for reference.
The six different areas of the graphic-mode window are used as follows:
• Command menu
— The at any time allowable commands plus default values for numerical data are listed here as buttons.
— Commands and values are selected by clicking the left mouse button (LMB).
— Slanted text signifies default choices that are accepted by either:
• Hitting the Return key
Presel SESAM
3-4 15-MAY-2014 Program version 7.5
• Clicking either of the Direct access buttons ‘;’ (semicolon) and ‘/ /’ (double slash). The former
accepts all subsequent default values (see Section 4.1.2) while the latter accepts a single default
value, i.e. the one shown in slanted font.
• Shortcut commands
These provide one-click access to commonly used compound commands. A Shortcut command is logged
as its equivalent full standard commands.
• Line-mode input
— The upper line presents the last given input.
SESAM Presel
Program version 7.5 15-MAY-2014 3-5
— The lower line includes the prompt for input and data entered in line-mode.
— You may paste (Ctrl+V) text into the line-mode input area.
Note: While entering a command by the keyboard it is not possible to click buttons or commands
until hitting the Return key or deleting all data typed. This involves that if you (inadvertently)
have entered a ‘space character’ (which you may overlook as you cannot see it) clicking com-
mands as well as selecting nodes and elements by clicking will not work. Use the backspace to
delete the ‘space character(s)’.
1 Store all first level superelements in the Presel database by reading the SESAM Input Interface Files (T-
files named T#.FEM). Use the command READ. This is optionally done automatically when starting
Presel from the SESAM Manager (select default Command input file).
2 Create a new superelement assembly. Use the command ASSEMBLY NEW. This superelement is ini-
tially empty but it will become a second level superelement once a first level superelement has been
included.
a Include a first level superelement in the new superelement assembly by the following process:
2 If necessary, translate, rotate and mirror the first level superelement to its proper position in the
assembly. See the INCLUDE command for the alternative positioning commands.
4 Check that the supernodes of the first level superelement match the nodes of the already included
superelements, if any. Use the command NOPRINT-CHECK-INCLUDE.
b Include other first level superelements by repeating the process above (item a).
c When the assembly, which has now become a second level superelement, is complete then give
appropriate boundary conditions. This may be fixations and/or supernodes for coupling with other
superelements. Use the command BOUNDARY.
Presel SESAM
3-6 15-MAY-2014 Program version 7.5
d At this stage you may define load cases for the assembly. This task may, however, also be postponed
to after the complete model has been assembled. See Section 3.3 on this.
3 Create new second level superelements by repeating the process above (item 2).
4 Create new superelements at level three and higher levels by repeating the process above (item 2). The
only difference from creating second level superelements is that not only first level superelements are
included. Any mixture of first and higher level superelements may be included in the same new superele-
ment. The level of the new superelement will be equal to the level of the highest level of the included
superelements plus one.
5 The assembly process is concluded when the top level superelement representing the complete model
has been created. Supernodes cannot be defined for the top level superelement.
6 You should now define the loads for the various higher level superelements (ref. item 2 d above). See
Section 3.3 on this.
If you only want to read this section, i.e. you do not intend to perform the tutorial, then you may skip the
four-item list of required actions below.
If you want to perform the tutorial and run Presel for this example, you first need to do the following:
• Start the SESAM Manager and open a new project. When doing so click Structure Type and choose
General. (You may also select type of structure after opening the new project by Options | Structure
Type.)
• In Manager click Options | Superelement and within the window appearing set your analysis to be
Superelement analysis (as opposed to Direct analysis) and set Top level superelement number to 7.
• Create the two first level superelements 5 and 6 by running Prefem twice. Remember to set the superele-
ment number before starting Prefem. The input for the two superelements is provided in Appendix A
TUTORIAL EXAMPLES, Section A 1. Rather than clicking/typing the commands you may want to cre-
ate two command input files using an editor. Note that node numbers in a FE model are automatically
generated by Prefem and may with a new program version deviate from the numbers shown in Figure
3.3.
• Start Presel as follows: Click Model | Superelement Handling Presel to open the Superelement han-
dling window. Make sure the Command input file selected is Default and click OK. This default Com-
mand input file reads the T#.FEM files located in the project area. (Optionally, you may read the
superelements 5 and 6 using the READ command.)
SESAM Presel
Program version 7.5 15-MAY-2014 3-7
3.3
Figure 3.3 The two first level superelements and desired complete model
Figure 3.4 and Figure 3.5 take you step-by-step through the procedure and commands for assembling
superelements 5 and 6 to form the complete model, superelement 7.
Presel SESAM
3-8 15-MAY-2014 Program version 7.5
3.4
3.5
The LABEL COUPLED-NODES command used in Figure 3.4 and Figure 3.5 verifies that the appropriate
number of first level superelements are indeed coupled. This is an important check because a minor coordi-
nate difference, for example due to a modelling error, may involve that some nodes are not coupled even
though they appear to be so in the display. A coordinate tolerance determines whether nodes are coupled or
not. You can set this tolerance by the SET COORDINATE-TOLERANCE command. The LABEL NON-
COUPLED-NODES is the complementary command labelling nodes adjoined by a single first level
superelement by the digit 1. The node in the lower right corner of superelement 7 is such a node.
The command PRINT OVERVIEW-OF-SUPER-ELEMENTS produces a table over all first and higher
level superelements. A double plus in the left margin of the table indicates which superelement is the cur-
rent one. The DISPLAY command displays the current superelement, the INCLUDE command includes
superelements into the current superelement, etc. For the example above the table looks like this:
SUPER EL.
TYPE LEVEL NODES ELEMENTS LOADCASES
--------- --------- --------- --------- ---------
5 1 16 9 3
6 1 10 4 2
++ 7 2 10 3 0
The command PRINT SUPER-ELEMENT-HIERARCHY produces a table illustrating the hierarchy. For
the current example the table will be as shown below. It is common practice to manually sketch the
superelement hierarchy as shown in Figure 3.4.
SUPER ELEMENT LEVEL
2 1
-----1:7.1-----------1:5.1
!
!-----2:6.1
!
!-----3:6.2
3.6
7
5 6 6
Also see Section 3.8 on display and print commands for verifying the model.
Within Presel there are two ways of uniquely identifying an occurrence of a superelement:
• By referring to its superelement number plus superelement index. The superelement index is the number
in sequence in which the superelement was included; see below for a more complete explanation. In the
LOAD COMBINATION command you need to refer to superelements in this way; see Section 3.3.2.
• By referring to name and location string as used in the LOAD ASSEMBLY command; see Section 3.4.
Yet another way is employed in Sestra for determining selective retracking (the SELID parameters on the
RETR command in the Sestra input):
• By referring to the string of hierarchy branch numbers from the top level superelement and down to the
superelement in question.
The hierarchy branch numbers, superelement numbers and superelement indexes are all given in the
superelement hierarchy printed by Presel as shown in Section 3.2.3. Each superelement is identified by three
numbers:
• Superelement number.
• Superelement index — The first time a superelement number is included in an assembly it is assigned
index 1. The second time it is included it is given index 2, and so on. When a higher level superelement
is repeatedly included in an assembly the indexes of its included superelements will also be incremented.
If, for instance, the second level superelement 7 in the current example were to be included twice in a
third level superelement 9 then a superelement 5 with index 2 would appear, also two new occurrences of
superelement 6 would appear, namely indexes 3 and 4; this is illustrated in Figure 3.7.
3.7
Figure 3.7 Superelement hierarchy — second level superelement 7 included twice in third level 9
Presel SESAM
3-12 15-MAY-2014 Program version 7.5
A specific node in a superelement model is uniquely identified by so-called node number triplets:
i.j.k
Figure 3.5 includes a sketch of a model showing node number triplets for a couple of nodes.
• Nodes of a higher level superelement are the union of the supernodes of the included superelements.
• Nodes of a superelement not defined as super will not exist in a higher level assembly into which the
superelement is included. Therefore, nodes not defined as super for the first level superelements (in
Prefem or Preframe) will not exist as far as Presel is concerned.
• Nodes may be defined as super or given any other boundary condition only for the current higher level
superelement.
• Nodes are identified by node number triplets: ‘superelement.index.node’. See Section 3.2.5.
• A node has as many triplets as there are first level superelements coupled to the node. Any of these tri-
plets are unique identifications of the node.
• Any selection of the six degrees of freedom (d.o.f.) of a node may be defined as super d.o.f. (the solid
and membrane elements have only three d.o.f.).
Note: If a superelement is to be rotated or mirrored then either all three translational d.o.f.s, or all
three rotational d.o.f.s, or all six d.o.f.s must be super. (The reason for this is explained in
Appendix B THEORY, Section B 1.3.)
Note: Nodes to couple must have equal sets of super d.o.f. For example, both may have all 6 d.o.f. as
super, which is the most common case, or both may have the translations in x and y and rota-
tion about z as super, etc.
In Presel commands like BOUNDARY, LOAD, PRINT and TAG you need to select nodes. Section 5.1
explains how this is done.
SESAM Presel
Program version 7.5 15-MAY-2014 3-13
• LOAD ASSEMBLY
This is an even more advanced method for combining loads. Section 3.4 explains this.
Combining loads is in effect assigning the proper load to the proper superelement occurrence. In this proc-
ess the different superelement occurrences need to be identified. This is done by referring to two numbers:
the superelement number and the superelement index. Section 3.2.4 explains the superelement index.
Note: Only the load combinations made for the top level superelement are the loads analysed in Ses-
tra and available for postprocessing. I.e. a load case that not directly or through intermediate
combinations contributes to a top level load combination has no effect.
The loads for the example of Section 3.2.2 are shown in Figure 3.8. Both the local loads for the two first
level superelements (loads defined in e.g. Prefem) and the desired global loads for the complete model are
shown.
Presel SESAM
3-14 15-MAY-2014 Program version 7.5
As you will see this tutorial does not have loads numbered from 1 and up. This will result in zero loads fill-
ing in the gaps in the load numbering thereby consuming some additional computational time and disk
space. This is, therefore, in conflict with normal and advisable practice. The reason for numbering the loads
like this is merely to avoid confusion in the tutorial between loads, superelement numbers and superelement
indexes.
Note: Number the load cases and combinations from 1 and up to avoid zero load cases which would
result in waste of computational time and disk space.
3.8
Figure 3.8 Loads modelled for the first level superelements and desired loads for the complete model
Figure 3.9 takes you step-by-step through the commands for establishing the desired global loads. Notes
explaining the data to enter are also provided.
SESAM Presel
Program version 7.5 15-MAY-2014 3-15
3.9
If you at this stage want to continue the exercise by running the model through a structural analysis in Sestra
you need to fix the model in space. You may for example fix the two nodes along the X-axis (see the sketch
in Figure 3.9) by giving the command:
BOUNDARY FIX FIX FIX FIX FIX FIX GLOBAL
SELECT LINE INFINITE COORDINATE 0 0 0 COORDINATE 1 0 0 END
Then leave Presel by clicking EXIT. (If you do not use Manager you must give the command WRITE 7
before exiting Presel to produce the T7.FEM file for top level superelement 7. If you do use Manager you
must have checked Write top level superelement on exit when starting Presel.) The model is now com-
plete and may be analysed using Sestra and the results may be presented by Xtract.
If you are a new user you may at this stage want to learn about the effect on loads of rotating and mirroring
superelements by reading Section 3.3.3. Other sections of interest to the new user are Section 3.5 Boundary
Presel SESAM
3-16 15-MAY-2014 Program version 7.5
Conditions, Section 3.8 Display and Print and Section 3.9 Practical and Efficient Application of the
Superelement Technique.
Figure 3.10 illustrates this. A, B and C are loads that rotate or mirror with the superelement to new positions
A', B' and C'.
3.10
Figure 3.10 Loads are rotated and mirrored along with the superelement
SESAM Presel
Program version 7.5 15-MAY-2014 3-17
A model is typically comprised of several superelements. Wajac or Wadam are employed for computing a
number of wave loads, e.g. 8 wave directions and 12 wave frequencies for each direction, altogether 96 load
cases for each occurrence of all superelements. Combining all these load cases for all these superelements
through all the superelement levels to the top will require a considerable amount of input. The LOAD COM-
BINATION GROUP command reduces this input to a manageable amount.
Note: Wave loads computed by Wajac and Wadam are stored on Loads Interface Files, the L#.FEM
files. These files are not read by Presel. This means that the load combination — whether this
is done by the one-by-one, group or assembly method — is performed without the program
having knowledge about the existence of the wave loads. Therefore, when referring to a wave
load in the load combination a warning is issued saying that the load case is accepted but it
must be computed by a load program prior to performing the analysis.
The LOAD COMBINATION GROUP command is explained below referring to the model of Section 3.2.2.
We assume that the model is subjected to wave loading from 8 directions and 12 frequencies, altogether 96
global load cases. As there are two occurrences of superelement 6, indexes 1 and 2, there will be a double
set of loads computed by the wave loading program for superelement 6. This means that the L5.FEM file
will contain 96 loads whereas the L6.FEM file will contain 192 loads.
Note: There is a difference between Wajac and Wadam in the way they number the wave load cases
when there are two or more occurrences of a superelement. Their numbering systems are as
follows (in parentheses are given the numbering applicable to the current example). In Wajac
all wave loads for the first occurrence come first (load cases 1, 2, 3, ... 96), thereafter follows all
wave loads for the second occurrence (load cases 97, 98, 99, ... 192) and so on. See Figure 3.11.
In Wadam the first wave load for all occurrences comes first (load cases 1 and 2), thereafter
follows the second wave load for all occurrences (load cases 3 and 4), then the third wave load
for all occurrences (load cases 5 and 6) and so on. See Figure 3.12
3.11
3.12
Figure 3.11 illustrates how the load cases for the superelement occurrences must be combined to establish
the global loads. The one-by-one method for combining loads computed by Wajac will for superelement 7
be:
% glc sup indx llc factor
LOAD COMB 1 5 1 1 1.0
END
6 1 1 1.0
END
6 2 97 1.0
END
END
2 5 1 2 1.0
END
6 1 2 1.0
END
6 2 98 1.0
END
END
3 5 1 3 1.0
END
6 1 3 1.0
END
6 2 99 1.0
END
END
:
END
END
END END
As can be seen the one-by-one method involves a lot of input (24 times the amount given above) even for
such a small superelement hierarchy. For a large superelement hierarchy the amount of input will increase
many times.
This command says that global load cases 1 to 96 with step 1 will be created. The superelement occurrences
(‘sup indx’) 5 1, 6 1 and 6 2 will contribute to these global load cases. For each occurrence only the local
load case (‘lowllc’) corresponding to the first global load case is given, i.e. 1, 1 and 97 for the three occur-
rences, respectively. Implicitly, the command says that the global load cases 2, 3, etc. are defined by incre-
menting the local load cases in parallel. Also see the description of the LOAD COMBINATION command
in Chapter 5.
The LOAD COMBINATION GROUP command must be given for each higher level superelement. Still,
the amount of input is manageable.
Figure 3.12 illustrates how the load cases for the superelement occurrences must be combined to establish
the global loads. The one-by-one method for combining loads computed by Wadam will for superelement 7
be:
% glc sup indx llc factor
LOAD COMB 1 5 1 1 1.0
END
6 1 1 1.0
END
6 2 2 1.0 *
END
END
2 5 1 2 1.0
END
6 1 3 1.0 *
END
6 2 4 1.0 *
END
END
3 5 1 3 1.0
END
6 1 5 1.0 *
END
6 2 6 1.0 *
Presel SESAM
3-20 15-MAY-2014 Program version 7.5
END
END
:
Note: The asterisk above indicates lines of input deviating from the corresponding Wajac input.
As can be seen the one-by-one method involves a lot of input (24 times the amount given above) even for
such a small superelement hierarchy. For a large superelement hierarchy the amount of input will increase
many times.
This command says that global load cases 1 to 96 with step 1 will be created. The superelement occurrences
(‘sup indx’) 5 1, 6 1 and 6 2 will contribute to these global load cases. For each occurrence only the local
load case (‘lowllc’) corresponding to the first global load case is given, i.e. 1, 1 and 2 for the three occur-
rences, respectively. The command STEP given prior to the ‘lowllc’ for occurrences 6 1 and 6 2 means that
the local load case is incremented by the value ‘incr’ = 2 for each increment of 1 of the global load case.
Notice that STEP is not given for occurrence 5 1 and neither is an ‘incr’ value, this is because superelement
5 occurs only once.
Note: The STEP command is used for Wadam and not for Wajac.
The LOAD COMBINATION GROUP command must be given for each higher level superelement. Still,
the amount of input is manageable.
Note: There is a requirement to the way the superelements are assembled to get a proper numbering
of the wave loads computed by Wajac and Wadam. Section 3.3.6 explains this.
superelement 21 is included in 31 the second time. Superelement 12 also appears with indexes 1 and 2 for
the same reason.
3.13
In the load combination, however, there will only be references to index 1 for superelements 11 and 12. This
is because the higher level superelement 21 only recognises index 1 of superelements 11 and 12. It ‘doesn’t
know’ that itself is going to be included twice in superelement 31.
Let us assume that we have 8 wave load cases computed by Wadam. As both 11 and 12 are used twice
Wadam will compute 16 loads for each superelement. When combining loads for superelement 21 we must,
therefore, create 16 load combinations. Using the LOAD COMBINATION GROUP command the input will
be:
% lowglc higlc step sup indx lowllc factor
LOAD COMB GROUP 1 16 1 11 1 1 1.0
12 1 1 1.0
END
END END
Note: Wajac and Wadam are based on the assumption that for all superelements the indexes are
numbered consecutively (1, 2, 3, ...) when counting the individual superelements numbers
from top to bottom in the table printed by Presel.
It is possible to assemble the superelements in such a way that the combination of wave loads computed by
Wajac and Wadam will fail! A description follows below of (1) an awkward superelement hierarchy, (2) a
hierarchy that fails in combination with Wajac and Wadam and (3) a proper hierarchy.
Presel SESAM
3-22 15-MAY-2014 Program version 7.5
Consider the example of Figure 3.14. Superelement 11 is included in superelement 31 prior to superelement
21, i.e. 11 is found in branch 1 while 21 is found in branches 2 (and 3). Superelement 21 also includes 11.
For this superelement hierarchy to be correct superelement 31 must be created (by ASSEMBLY NEW 31)
and 11 included in 31 prior to including 11 in 21. 21 can of course not yet be included in 31 because it does
not yet exist! Superelement 31 is therefore temporarily left incomplete and superelement 21 is created and
completed by including 11 and 12. Then superelement 31 is completed by the commands ASSEMBLY OLD
31 and INCLUDE 21 twice. This ensures that the occurrence of superelement 11 at the top of the table
printed by Presel is assigned index 1.
3.14
If you were to create superelement 21 and include 11 and 12 first and thereafter create 31 and include 11, 21
and 21 in that order the superelement hierarchy tabulated by Presel would be as shown in Figure 3.15. As
seen, index 2 of superelement 11 comes before index 1 when counting superelement 11 from top and down.
Note: If you establish your load combination input based on these indexes then you will get wrong
results for wave loads computed by Wajac and Wadam! More specifically, the load case num-
bering will not correspond to the indexes as described in Section 3.3.4.
SESAM Presel
Program version 7.5 15-MAY-2014 3-23
3.15
Figure 3.16 shows a superelement hierarchy avoiding the problem described above. Superelement 21 is cre-
ated first. Thereafter superelement 31 is created including superelements in the following order: 21, 21 and
11 at the end.
3.16
To be able to use the LOAD ASSEMBLY command, unique identifications must be given for all occur-
rences of first level superelements for which loads exist. Giving such unique identifications makes the task
of assembling the superelements more complex. But, as shown by the example below, defining the loads
requires considerably less input.
There is another advantage of the load assembly method: The LOAD ASSEMBLY command will not be
affected by changes to the superelement hierarchy. This is because the principle of this method is to give
identifications for superelement occurrences depending on their final location in the complete model. The
order in which the superelements are assembled is of no consequence.
• Superelement occurrence represents an actual part of a higher level superelement, it will have an exact
location in that higher level superelement. If the higher level superelement is the top level superelement
— the complete model — then the superelement occurrence also represents an actual part of the real
structure.
It follows that only when the complete model is assembled the superelement occurrences may be given their
final identifications. However, rather than assembling the whole model including the top level superelement
before introducing the identifications a dynamic way of giving identifications is available. By ‘dynamic’ is
meant that preliminary identifications are introduced and thereafter modified during the assembly process
until the final identifications are determined when the complete model — the top level superelement — is
created.
The identification of a superelement occurrence is composed of a name and a location string as follows:
name.location
The name is given to a superelement when it is created and is not changed during the assembly process. The
location string is given when a superelement is included in a superelement assembly. The superelement has
then become a superelement occurrence in that specific assembly. The location string is modified during the
assembly process.
In this tutorial five first level superelements shall be assembled to form a house as illustrated by Figure 3.17.
SESAM Presel
Program version 7.5 15-MAY-2014 3-25
3.17
After reading the first level superelements into Presel the first task is to assign names by the commands:
ASSEMBLY OLD 1
NAME CREATE FNDT
ASSEMBLY OLD 2
NAME CREATE LWALL
ASSEMBLY OLD 3
NAME CREATE SWALL
Presel SESAM
3-26 15-MAY-2014 Program version 7.5
ASSEMBLY OLD 4
NAME CREATE FLOOR
ASSEMBLY OLD 5
NAME CREATE ROOF
Then the second level superelement 10 is created and given its name (names for higher level superelements
are normally not required but included here for completeness):
ASSEMBLY NEW 10
NAME CREATE STOREY
When including superelements 2 (long wall) and 3 (short wall), twice each, and superelement 4 (floor) these
are assigned location strings. The LOCATION CREATE command is used:
INCLUDE 2 NOPRINT-CHECK-INCLUDE
LOCATION CREATE SOUTH
PERFORM-INCLUDE
%
3 NOPRINT-CHECK-INCLUDE
LOCATION CREATE SOUTH
PERFORM-INCLUDE
%
2 TRANSLATE 0 4 0
NOPRINT-CHECK-INCLUDE
LOCATION CREATE NORTH
PERFORM-INCLUDE
%
3 TRANSLATE 7 0 0
NOPRINT-CHECK-INCLUDE
LOCATION CREATE NORTH
PERFORM-INCLUDE
%
4 TRANSLATE 0 0 0.8
NOPRINT-CHECK-INCLUDE
LOCATION CREATE BASIC
PERFORM-INCLUDE
END
Figure 3.18 shows the current identifications of the superelement occurrences included in superelement 10,
STOREY. The short and long walls are given location strings corresponding to their southerly and northerly
positions. The four wall superelements now have unique identifications in terms of names plus location
strings. The floor is given the location string BASIC; it is neither southerly nor northerly.
Note: The LOAD ASSEMBLY command can only refer to superelement occurrences having loca-
tion strings. Even a superelement not repeated and therefore having a unique name must be
given a location string. The location string may be skipped for a superelement without loads.
SESAM Presel
Program version 7.5 15-MAY-2014 3-27
3.18
Supernodes are then defined at the bottom and top planes for superelement 10, STOREY:
BOUNDARY SUPER SUPER SUPER SUPER SUPER SUPER GLOBAL
SELECT PLANE 2-PLANE COORDINATE 0 0 0.8 COORDINATE 0 0 0
PLANE 2-PLANE COORDINATE 0 0 3.8 COORDINATE 0 0 0
END
Then the top level superelement 100 is created and given its name:
ASSEMBLY NEW 100
NAME CREATE HOUSE
Now, when including superelement 10 twice to make the two storeys of the house the location strings of the
first level superelements included in 10 are assigned new location strings. The current strings are modified
by the LOCATION USE command. The new location strings are more specific reflecting that the positions
of the superelements in the complete model are more specific. The first inclusion of 10 is the ground floor,
GRD is therefore added to the location strings. The second inclusion is the first floor, 1ST is therefore added
to the location strings.
INCLUDE 1 NOPRINT-CHECK-INCLUDE
LOCATION CREATE BASIC
PERFORM-INCLUDE
%
10 NOPRINT-CHECK-INCLUDE
% Modify location for the south walls:
LOCATION USE SOUTH SOUTHGRD
% Modify location for the north walls:
LOCATION USE NORTH NORTHGRD
% Modify location for the floor:
LOCATION USE BASIC BASICGRD
PERFORM-INCLUDE
%
10 TRANSLATE 0 0 3
NOPRINT-CHECK-INCLUDE
% Modify locations for all using wild-card:
LOCATION USE * *1ST
Presel SESAM
3-28 15-MAY-2014 Program version 7.5
PERFORM-INCLUDE
%
5 TRANSLATE 0 0 6.8
NOPRINT-CHECK-INCLUDE
LOCATION CREATE BASIC
PERFORM-INCLUDE
END
For the first inclusion of superelement 10 the location strings are modified one-by-one: The two superele-
ment occurrences (LWALL and SWALL) having location string SOUTH get the new string SOUTHGRD,
the two occurrences having location string NORTH get the new string NORTHGRD and the single occur-
rence having location string BASIC gets the new string BASICGRD.
But when all superelement occurrences shall have the same string added to their locations, as in this case, a
wild-card notation can be used as shown for the second inclusion of superelement 10.
The foundation (1) and roof (5) superelements are included for the first time, the LOCATION CREATE
command is used for these. They are given location strings BASIC as no specific identifications are
required.
Note: If you give location strings also for higher level superelements then a wild-card reference in
the LOAD ASSEMBLY command matching the higher level superelements will pick up loads
from this superelement directly in addition to loads picked up from first level.
Therefore, LOCATION CREATE is not used when including 10 (STOREY) in the HOUSE below or else
the wild-card *.* when defining load 1 (gravity) — see Section 3.4.3 — will pick up gravity loads from 10
in addition to gravity loads directly from first level. Gravity for parts of the model will erroneously contrib-
ute twice!
Figure 3.19 illustrates the final identifications of the superelement occurrences of superelement 100,
HOUSE.
SESAM Presel
Program version 7.5 15-MAY-2014 3-29
3.19
Now the assembling of superelements including assigning of occurrence identifications is complete. Their
identifications have become more specific in parallel with knowing more about their final positions. Figure
3.20 shows the superelement hierarchy with superelement numbers, indexes, names and location strings.
Presel SESAM
3-30 15-MAY-2014 Program version 7.5
3.20
3.21
In the input for assembling the loads you refer to a single first level superelement occurrence by giving its
name and location string and to several occurrences by using wild-card notation (*):
ASSEMBLY OLD 100
LOAD ASSEMBLY
% --- Load 1 is self weight (gravity)
1 INCLUDE-LOAD *.* 1 1.0
END
% --- Load 2 is wind pressure on south walls (both long and short walls)
2 INCLUDE-LOAD *WALL.SOUTH* 2 1.0
END
% --- Load 3 is wind pressure on south long wall and north short wall
3 INCLUDE-LOAD LWALL.SOUTH* 2 1.0
INCLUDE-LOAD SWALL.NORTH* 2 -1.0
END
% --- Load 4 is weight of snow on roof
4 INCLUDE-LOAD ROOF.BASIC 2 1.0
END
% --- Load 5 is weight of people and furniture on floors, there are
% 50% more people and furniture in ground floor than in first floor
5 INCLUDE-LOAD FLOOR.BASICGRD 2 1.5
INCLUDE-LOAD FLOOR.BASIC1ST 2 1.0
END
END END
• Global load case 1 (gravity) will include local load case 1 from all (*.*) first level superelements.
• Global load case 2 (wind pressure on south walls) will include local load case 2 from the superelements
long and short walls (*WALL), located to the south and including both storeys (SOUTH*).
• Etc.
The command will establish the load combinations required on all levels from the second level and to the
top. This can be verified by printing the load combinations for the various higher level superelements.
Note: The LOAD ASSEMBLY command will typically create more load combinations for interme-
diate level superelements (between first and top level) than you normally will create by the
more manual one-by-one and group methods. Such extra load combinations for intermediate
level superelements will not make any difference except for spending somewhat more compu-
ter time and disk space.
Note: In the above example location strings are only given for first level superelements and loads are
assembled from first level directly to the top level. You may also assemble loads to intermedi-
ate level superelements. You then need to combine loads from these intermediate level
superelements to the top by either the assembly or one-by-one method.
SESAM Presel
Program version 7.5 15-MAY-2014 3-33
You need to determine a convention for location strings for your model. You also have to device a scheme
for how to arrive at this convention, i.e. how to employ the LOCATION CREATE and LOCATION USE
commands during the assembly process. The scheme will depend on how the superelement hierarchy is
organised. To be able to determine a practical convention and a scheme you should be familiar with the pos-
sibilities and limitations of the LOCATION USE command; see Chapter 5 for this. You may find that
assigning name and location strings to your superelement occurrences will influence the organisation of the
superelement hierarchy. In the house example of Section 3.4.2 and Section 3.4.3 we were able to merely
append characters to the location strings assigned the first time. For example, when the long wall (LWALL)
was included in the second level superelement STOREY the first time it was assigned location string
SOUTH. Later in the assembly process the location string was merely appended by GRD and 1ST because
these occurrences were all located in a southerly direction. This may not always be possible.
Consider the example of Figure 3.22. A single superelement 10 is used nine times to establish the complete
model 40 at fourth level. When being included into the second level superelement 20, we may decide to give
superelement 10 the location strings A and B. Thereafter we may choose to merely append a character to the
location strings each time superelement 10 is repeated by repeating higher level superelements. We will end
up with location strings as shown. However, this may not be a very logical convention for location strings.
3.22
Alternatively, we may want location strings reflecting the positions of the superelement occurrences relative
to the north, west, south and east directions. EN means east-north, NE means north-east, NW means north-
west, etc. Figure 3.23 shows a scheme arriving at this more logical convention for location strings.
3.23
Note: If you are unable to device an easy scheme for arriving at the chosen convention for naming
location strings you may always resort to the scheme shown in Figure 3.22 up to the top level
and for the top level substitute the location strings with the proper ones. See Figure 3.24.
SESAM Presel
Program version 7.5 15-MAY-2014 3-35
3.24
• Free — the d.o.f. is free to displace, this is the default boundary condition for all d.o.f.s of all nodes
• Prescribed — the d.o.f. is fixed at a given displacement given by the LOAD NODE command
• Super
A superelement can only be included in a superelement assembly if one or more of its nodes (or d.o.f.s) are
defined as super. If one or more of the nodes (d.o.f.s) of a superelement are super the superelement cannot
be the top level superelement (the complete model). It must be included in a higher level assembly.
In the BOUNDARY command you need to select nodes. This may be done inside the BOUNDARY com-
mand by the SELECT option. Or you may pre-select the relevant nodes by the TAG command in which case
you refer to these pre-selected nodes by the TAGGED option inside the BOUNDARY command.
(TAGGED is in effect a set with pre-defined name, you may also use the SET command to define sets.) In
either case you need to select nodes using the node select features described in Section 5.1.
As explained in Section 5.1, alternatively to selecting nodes directly by giving their triplets you may select a
line, plane or volume through or enclosing the desired nodes. And these lines, planes and volumes may be
defined referring to nodes or to coordinates. Moreover, the coordinates may be given in the cartesian system
or in a pre-defined cylindrical system.
For example, if you want to select all nodes in a plane you may find that the so-called 2-PLANE option and
giving two sets of cartesian coordinates is the quickest alternative. First give a point in the plane and then a
point defining together with the first point a normal to the plane. All nodes on a cylindrical surface are easily
Presel SESAM
3-36 15-MAY-2014 Program version 7.5
selected by the 2-PLANE option referring to a cylindrical system; the first point is positioned in the desired
cylindrical plane and the second point on a line through the first point and normal to the cylindrical axis.
By default the boundary conditions given refer to the global cartesian system. Alternatively, you may give
the boundary conditions in a pre-defined cylindrical system (defined by the COORDINATE-SYSTEM com-
mand) or with pre-defined transformations (a cartesian system with rotation compared to the global system
defined by the TRANSFORMATION command). This may for example be used to introduce a sloping slip
surface: rotate to a system having one of its axes normal to the slip surface and fix the translational d.o.f.
corresponding to this axis while letting the other two translational d.o.f.s be free.
Using the ‘general node dependency’ any d.o.f. of a node may be made linearly dependent on any other
d.o.f.s of any other nodes. The user explicitly specifies the linear dependency factor for all the independent
d.o.f.s. The displacement of the dependent d.o.f. will then be:
r d = r i1 ⋅ β 1 + r i2 ⋅ β 2 + r i3 ⋅ β 3 + …
where r represents the displacements, subscripts d and i represent the dependent and independent d.o.f.s
respectively, and β is the given dependency factors.
With the ‘two node dependency’ all d.o.f.s of a given node are made linearly dependent on the correspond-
ing d.o.f.s of two other nodes. The displacement of the dependent d.o.f.s will be:
r d = r i1 ⋅ β + r i2 ⋅ ( 1 – β )
where β is a dependency factor given by the user. Presel will compute a default value for β as explained in
Figure 3.25. β is computed based on the projection of the dependent node onto the line between the two
independent nodes.
Normally, the ‘two node dependency’ has physical meaning only when the dependent and the two independ-
ent nodes all lie on a straight line.
3.25
Note: Dependent nodes are indicated by blue triangles in the model display.
Note: Independent nodes = supernodes are indicated by blue octagons (will look like circles).
All independent d.o.f.s must be super d.o.f.s, i.e. they are defined with boundary condition code ‘super’
using the BOUNDARY command prior to giving the LINEAR-DEPENDENCY command. The ‘two node
dependency’ alternative also allows the user to define the independent nodes as super within the command
by the FORCE-INTO-SUPER alternative. This implies that introducing linear dependency for a superele-
ment involves that the superelement must be included in yet a higher level superelement.
The requirement that the independent d.o.f.s must be super only concerns second and higher level superele-
ments as created by Presel. In the case of first level superelements created by Preframe and Prefem the inde-
pendent d.o.f.s need not be super provided that the Multifront equation solver is used in Sestra.
Linear dependencies in a transformed coordinate system may be specified by first assigning a transforma-
tion to the dependent and independent nodes using the BOUNDARY command.
3.7 Sets
The command TAG may be used to define a set of nodes that may be referred to by the option TAGGED in,
for example, the BOUNDARY command; see Section 3.5. In effect, this is therefore a set of nodes with the
pre-defined name TAGGED. When you create a new higher level superelement (by the ASSEMBLY NEW
command) the set TAGGED is empty until you put nodes into it. The set will not be changed by moving
between the superelements (by the ASSEMBLY OLD command) or by exiting and re-entering Presel. You
may refer to the complementary set of nodes by the pre-defined name UNTAGGED. The command
UNTAG is used to remove nodes from the set TAGGED.
In addition to the set TAGGED you may define any number of named sets of nodes by the DEFINE SET
command. The standard set operators UNION-WITH, SUBTRACT-BY and INTERSECTION-WITH are
used to define the sets.
A set that contains supernodes will be available also within an assembly into which the superelement is
included. This is provided that the set was defined prior to including the superelement. A set containing
supernodes defined for a first level superelement (in Prefem and Preframe) will also be available in assem-
blies. Note that if a set contains both supernodes and other (free or fixed) nodes then the set at assembly
level will only contain the supernodes (and these will be free nodes at assembly level).
Moreover, if any of the nodes of the set are given the boundary condition super (re-defined as supernodes)
for the superelement assembly then the set, containing these nodes, will be available for yet a higher level
superelement assembly.
Finally, if two or more superelements have sets of supernodes with the same name then these will be merged
at assembly level.
Note: Sets defined in Prefem or Preframe containing only elements or geometry (Prefem only) will
not be available in Presel as only nodes are relevant here.
Note: The transfer and merging of sets of supernodes from superelements to superelement assem-
blies is only avaialble for named sets and not the predefined set TAGGED (and UNTAGGED).
Presel SESAM
3-38 15-MAY-2014 Program version 7.5
In addition to displaying the current superelement any superelement may be displayed without changing
current superelement by the DISPLAY SPECIFIED-SUPERELEMENT command. This command is practi-
cal when you are working with assembling superelements in an assembly and only want to make a quick
check of the appearance of a certain superelement.
Furthermore, the DISPLAY LOCATE-SUPERELEMENT command allows highlighting (by a different col-
our) a given first level superelement in an assembly.
Note: There are shortcut command buttons under the headings ‘DispMod’ and ‘Display’ for all dis-
play alternatives explained above.
3.26
The LABEL command may be used to add information to the displayed superelement.
You may also use the DISPLAY option during the inclusion of a superelement to see the current position of
the superelement being included, i.e. to check its position before you complete the inclusion. The tutorial of
Section 3.2.2 illustrates this and the description of the INCLUDE supno DISPLAY command in Chapter 5
shows an example. This display cannot be annotated by the LABEL command.
You may also verify the load combinations for an assembly using the DISPLAY LOAD command. There
are two ways of doing this:
SESAM Presel
Program version 7.5 15-MAY-2014 3-39
• You may determine how many load cases each included superelement contributes with to a given load
combination. This is available through the DISPLAY LOAD LOADED-SUPERELEMENT command.
Colour coding of the superelements indicates which of them contribute with no load cases, which con-
tribute with one, which contribute with two, and so on.
• You may determine which load cases with what factor each included superelement contributes with to a
given load combination. This is available through the two commands DISPLAY LOAD FIRST-CON-
TRIBUTING-LOAD and DISPLAY LOAD NEXT-CONTRIBUTING-LOAD. The FIRST-CONTRIB-
UTING-LOAD option colour codes superelements contributing with their load case number ‘i’ where ‘i’
is the lowest contributing load case number over all superelements. Thereafter you should use the
NEXT-CONTRIBUTING-LOAD option which colour codes superelements contributing with their load
case number ‘j’ where ‘j’ is the second lowest contributing load case number. Repeating the NEXT-
CONTRIBUTING-LOAD option will loop through all contributing load cases. The load factors are for
each display printed on top of the colour coded superelements.
Note: The loads values defined in Prefem/Preframe and computed in Wajac/Wadam cannot be dis-
played in Presel as it has no knowledge of the contents of these loads.
In addition to displaying the model the PRINT command is useful for verification purposes. Various PRINT
commands are exemplified in Figure 3.27, Figure 3.28, Figure 3.29, Figure 3.30 and Figure 3.31.
3.27
3.28
3.29
3.30
3.31
There are several aspects of how to apply the superelement technique in an optimal way and an in-depth dis-
cussion cannot be provided here. Efficient application of the technique rely on a combination of an under-
standing of the theoretical foundation, consideration of the hardware being used and practical experience.
Discussing with experienced users and taking heed of the following advice will enable you to take advan-
tage of the superelement technique in an efficient way while building up your own practical experience.
Some of the items below contradict each other to some extent, this only underlines the fact that you often
have to balance between conflicting considerations. Also see Figure 3.32 for illustrations of some of the
items below.
• How to assemble the superelements to form the complete model (some of the advice below will have
bearing on how to split the structure into superelements)?
— Limit the number of higher level superelements and the number of levels. Few and moderately large
higher level superelements are better than many smaller ones. For a large model this implies that the
second level superelements will include many first level superelements, the third level superelements
will include many second level superelements, and so on.
— Be aware of that reduction of higher level superelements is time consuming. A higher level superele-
ment will normally be much more time consuming than a first level superelement with the same
number of internal and super nodes. This is because the stiffness matrix of a higher level superele-
ment has few zeros and a large bandwidth.
— Assemble superelements in an order corresponding to their topological sequence and not in a haphaz-
ard way. (Assemble superelements in the way Lego bricks are put together.)
— Avoid coupling one superelement with many other superelements.
— Do not assemble superelements not geometrically coupled. This concerns superelements belonging to
the same structure, i.e. their stiffnesses are coupled through other superelements. On the other hand, if
you want to analyse several non-coupled models — for example to test various meshes or designs for
the same structure — you may find it convenient to make a single model comprised of non-coupled
superelements and analyse all these in one operation (one Sestra run). (This approach should not be
employed for large models.)
• You also need to consider which load cases to create for the first level superelements and how to com-
bine these through the higher level superelements to create the final loads for the top level superelement.
SESAM Presel
Program version 7.5 15-MAY-2014 3-43
3.32
For the traditional Supermatrix solver, however, it is absolutely essential to minimise the bandwidth of the
first level superelements’ stiffness matrices by optimising or re-numbering their internal node numbering.
The auxiliary program Bpopt is used for this purpose (in the case of Preframe the optimization is normally
done inside the program). When using the SESAM Manager optimization of first level superelements is
controlled by Manager.
Note: The optimization should be performed prior to reading the superelements into Presel.
Optimising higher level superelements may be done inside Presel using the command OPTIMIZE. Note that
a higher level superelement can only be optimised when it is complete, i.e. when all relevant superelements
have been included into it and before it is included in new higher level superelements. The top level
superelement may, however, be optimised after concluding the Presel session using the auxiliary program
Bpopt.
Presel SESAM
3-44 15-MAY-2014 Program version 7.5
Provided the superelements are assembled into a higher level superelement (an assembly) in a logical brick-
on-brick way as advised in Section 3.9 there is very little, if anything, to gain in terms of CPU reduction by
optimising higher level superelements. Using the OPTIMIZE command is therefore normally not necessary.
See Section 3.9 on practical and efficient use of the superelement technique.
On the other hand, if a higher level superelement is created by assembling superelements in a haphazard
way (in effect, this is the way the automatic assembling of superelements in Pretube works) optimising the
higher level superelements may be important.
SESAM Presel
Program version 7.5 15-MAY-2014 4-1
4 EXECUTION OF PRESEL
• Files used
• Program requirements
• Program limitations
• The parameters (commands, sub-commands and data) are separated by one or more blank characters (or
a comma) and may be entered one by one or with two or more entries on a single line of input. For exam-
ple:
COMMAND
SUB-COMMAND
SUB-SUB-COMMAND
data ...
is equivalent to:
COMMAND SUB-COMMAND SUB-SUB-COMMAND data ...
Note, however, that data belonging to different data sets cannot be entered on a single line.
• UPPER CASE = lower case (all commands will be logged on a ‘command log file’ in UPPER case).
• Commands and sub-commands may be abbreviated as long as they are unique. In a command consisting
of words separated by hyphens, each word may be abbreviated or completely left out. Examples:
NODE-NUMBERS = N-N
COMMAND-INPUT-FILE = C-I
• Default values are provided between slashes, ‘/default/’. The defaults are accepted by hitting Return.
• Real or integer input may be entered irrespective of type of numerical data, use ‘E’ for exponent.
• ‘?’ will list all legal commands and data options. (This command is irrelevant for the graphical user inter-
face where all legal commands and data options are at any time given in the command column of the
graphic-mode window.)
• ‘..’ (two dots) will execute the input data before ‘..’ and subsequently abort the current command. The
program is thereafter ready for more commands. If the data before the ‘..’ is incomplete it will be dis-
carded.
• ‘,,’ (two commas) will cause one default parameter to be accepted. (May be useful when editing a ‘com-
mand input file’.)
• ‘;’ (semicolon) will cause default parameters to be accepted until the end of the parameter group or until
there is no default provided.
• Text containing blank characters has to be enclosed within single quotes: 'this is a text'.
• ‘%’ (percentage sign) at the beginning of a line is used for entering a comment. Comments will be
logged together with commands on the ‘command log file’ (see Section 4.1.3). Note that the program
will occasionally log information on the ‘command log file’, this will appear as comments in between
data and comments entered by the user. The program information is preceded by ‘%%’ (two percentage
signs) to distinguish it from the user’s own comments. This makes it easy to strip a ‘command log file’
for program information in connection with creating a ‘command input file’ (any fairly good editor will
have a macro-functionality or similar enabling you to locate and remove all lines with ‘%%’). Moreover,
comments preceded by ‘%%’ will not be logged on the ‘command log file’ to avoid irrelevant logging of
program information when using an unedited ‘command log file’ as a ‘command input file’.
SESAM Presel
Program version 7.5 15-MAY-2014 4-3
4.1
• The ‘command log (journal) file’ (.JNL) is an ASCII file on which all commands and data given to the
program are logged. This means that both data typed (or clicked) by the user and data read by the pro-
gram from a ‘command input file’ will be logged. However, commands not changing the model (and
data base), e.g. a command displaying data, will not be logged. The time of opening and closing the
‘model file’ is also logged. The file is very useful as a backup file both for verification purposes and for
later use as a ‘command input file’. The ‘command log file’ can be read and modified by a text editor.
• The ‘command input file’ (.JNL) is an ASCII file which may be read into the program. The commands
contained on this file will have the same effect as if they where given by the user directly. The file is
processed by using the command ‘SET COMMAND-INPUT-FILE ...’ followed by ‘# ALL’ (the
latter command means: read all commands found on the file). Alternatively, you may specify a ‘com-
mand input file’ when starting Presel from Manager.
• The ‘model file’ (.MOD) is the binary data base containing all model data. The file cannot be read by a
text editor.
• The ‘print file’ (.LIS) is an ASCII file which contains tables over data requested for printing by the
PRINT command.
Presel SESAM
4-4 15-MAY-2014 Program version 7.5
• The ‘plot file’ contains graphic information produced by the PLOT command. The file extension will
depend on the plot format chosen (see the SET PLOT FORMAT command). See Section 4.1.4 for advice
on using the CGM format to include plots in reports.
• The ‘Input Interface File’ (.FEM) — termed T-files for short — is comprised of:
— First level superelements (created by Prefem or Preframe) which are read by Presel, and
— Higher level superelements which are written by Presel.
Presel has been designed to protect the user against loss of valuable data. However, accidental loss of data
may occur. This may be caused by the user by for example inadvertently deleting the ‘model file’ or it may
be due to an inconsistency in the data model. Such inconsistency may occur for several reasons:
• The disk is full, the disk quota is exhausted or user privileges are inadequate.
If Presel discovers an inconsistency in the data model the program will normally close all files opened and
abort the execution. Presel may then be restarted using the ‘model file’. In some cases, however, it will not
be possible to resume normal execution due to an irrecoverable inconsistency.
If the ‘model file’ is lost it can be reconstructed by re-executing the program and reading input from the
‘command log file’, i.e. using it as a ‘command input file’.
Note: The ‘model file’ will normally not be compatible between different versions of Presel. The
‘command log file’ may, however, be used as input to a new version.
Depending on the capabilities of your word processor the PostScript plot format may also be used for the
purpose of importing SESAM plots into reports. Contrary to CGM, PostScript is an ASCII formatted file
and is therefore more easily transferred from one computer make to another.
Note that a word processor will normally recognise only one picture (display) on each file. You should,
therefore, specify a new file name for each plot command using the SET PLOT FILE command.
/WRITE-SUPERELEMENT=number Write an Input Interface File with the given (top level) su-
perelement number (plus all lower level superelements) when
exiting the program.
/COMMAND-FILE=filename Read the specified command input file after the model/journal
file has been accepted.
/FORCED-EXIT Force EXIT after initialisation and after processing of the file
defined by the /COMMAND-FILE argument.
Note the following about how to enter the command line arguments:
• Each argument name must begin with a slash (/) and each argument value must be preceded by an equal
sign (=). Spaces can freely be distributed around the equal sign and before each slash.
• Texts with blank spaces and special characters (e.g. file names) must be enclosed in quotes. Note that
some operating systems change the case of the input text if it is not enclosed in quotes.
• If at least one of the arguments /PREFIX, /NAME and /STATUS is specified then the prompt for data
base and journal file name is skipped and defaults are used for any unspecified values.
• The values given to the /EYEDIR are real values. The default is the Presel default values. If one of the
three are given the other two are set to 0.0 unless specified.
• In some cases a virtual screen larger than the real screen is used. In such cases reduce the /WINDOW-
SIZE argument value.
Graphics Devices
The graphical user interface is implemented on Microsoft Windows. Other devices are currently not used.
Memory
Presel allocates memory buffers for access to data of the data base file. When using the graphical user inter-
face Presel will allocate memory for the display.
You may limit the size of the working array used for optimising the node numbering of a higher level su-
perelement by defining the parameter MSIZE-PRESEL-OPTIMIZE in the configuration (password) file.
If you do not give any value for this parameter then as much memory as needed will be allocated.
Typing
While typing a command using the keyboard you cannot click commands in menus or select nodes by click-
ing or use the mouse in any other way until the Return key has been hit or until the typed text has been
deleted by backspace.
Presel SESAM
4-8 15-MAY-2014 Program version 7.5
SESAM Presel
Program version 7.5 15-MAY-2014 5-1
5 COMMAND DESCRIPTION
The hierarchical structure of the commands and numerical data is documented in this chapter by use of
tables. How to interpret these tables is explained below. Examples are used to illustrate how the command
structure may diverge into multiple choices and converge to a single choice.
In the example below command A is followed by either of the commands B and C. Thereafter command D
is given. Legal alternatives are, therefore, A B D and A C D.
B
A D
C
In the example below command A is followed by three selections of either of commands B and C as indi-
cated by *3. For example: A B B B, or: A B B C, or A C B C, etc.
B
A *3
C
In the example below the three dots in the left-most column indicate that the command sequence is a contin-
uation of a preceding command sequence. The single asterisk indicate that B and C may be given any
number of times. Conclude this sequence by the command END. The three dots in the right-most column
indicate that the command sequence is to be continued by another command sequence.
B
*
... A C ...
END
In the example below command A is followed by any number of repetitions of either of the sequences B D
and C D. Note that a pair of braces ({ }) is used here merely to define a sequence that may be repeated. The
braces are not commands themselves.
B
A { D }*
C
The characters A, B, C and D in the examples above represent parameters being COMMANDS (written in
upper case) and numbers (written in lower case). All numbers may be entered as real or integer values.
Brackets ([ ]) are used to enclose optional parameters.
Presel SESAM
5-2 15-MAY-2014 Program version 7.5
Note: The command END is generally used to end repetitive entering of data. Using double dot (..)
rather than END to terminate a command will, depending on at which level in the command it
is given, save or discard the data entered. Generally, if the data entered up to the double dot is
complete and self-contained the double dot will save the data. If in doubt, it is always safest to
leave a command by entering the required number of END commands.
• Several nodes by referring to their node number triplets and enclosing them in parentheses.
• A SINGLE node by referring to its node number triplet. This option is less relevant after introduction of
the previous option of enclosing node number triplets in parentheses.
• A GROUP of nodes by referring to the triplet of the first node plus the last node number and the step
(increment) in node numbering.
• All nodes inside a VOLUME by referring to two points being the diagonally opposite corners of a box
with side surfaces parallel with the coordinate system planes XY, YZ and ZX (or parallel interpreted in
the space of the named coordinate system if the USE-COORDINATE option has been chosen).
• Several nodes by repeatedly using any of the above selection methods (stop selection by END).
• ALL nodes.
Whether a certain node lies on the given straight line, or lies in the given plane, or is located within the
given volume is decided by a coordinate tolerance; see the SET COORDINATE-TOLERANCE command.
Also, whenever a single node is to be selected, either after the line-mode option SINGLE or inside any of
the other options the node may be clicked as an alternative to typing in the node number triplet.
Where ‘point’ in the command syntax above represents specifying a point as follows:
NODE supno index nodeno
COORDINATE x y z
LOCAL-COORDINATE r phi z
The LINE, PLANE and VOLUME alternatives offer selection by referring to points in space. These lines,
planes and volumes may be interpreted in the cartesian coordinate space of the superelement — the GLO-
BAL-COORDINATES option — or in a cylindrical coordinate space — the USE-COORDINATE-SYS-
TEM option. Figure 5.1 illustrates this. Such a cylindrical coordinate system must previously have been
defined by the COORDINATE-SYSTEM command. You may switch back and forth between these two
spaces within the same selection sequence, i.e. before giving END. The space chosen last is valid for the
subsequent LINE, PLANE and VOLUME commands. If neither the GLOBAL-COORDINATES nor the
USE-COORDINATE-SYSTEM space is given the former is valid.
The points defining the lines, planes and volumes may, as shown by the table explaining ‘point’ above,
either be nodes (selected by giving node number triplets) or specified by coordinates.
When a cylindrical coordinate space has been referred to (by USE-COORDINATE-SYSTEM) in a selection
sequence the point coordinates may optionally be given in this cylindrical coordinate system (the LOCAL-
COORDINATE option). Note that the choice between COORDINATE and LOCAL-COORDINATE is
merely for giving the point coordinates. The line, plane or volume is still interpreted in the global cartesian
or given cylindrical space according to the choice between GLOBAL-COORDINATES and USE-COORDI-
Presel SESAM
5-4 15-MAY-2014 Program version 7.5
NATE-SYSTEM. Normally though, you want to use the LOCAL-COORDINATE option when the line,
plane or volume is interpreted in a cylindrical coordinate space and the COORDINATE option when they
are interpreted in the cartesian space.
Note: You may also use the TAG command to pre-select nodes and refer to these TAGGED nodes
rather than selecting nodes directly within the command in question.
5.1
PARAMETERS:
GLOBAL-COORDINATES The lines, planes and volumes subsequently given are to be in-
terpreted in the cartesian coordinate space of the superelement
in question.
USE-COORDINATE-SYSTEM The lines, planes and volumes subsequently given are to be in-
terpreted in the subsequently named cylindrical coordinate
space.
SEGMENT Only the nodes on the line between the two points are selected.
PLANE All nodes in an infinite plane defined by two or three points are
selected. The tolerance or ‘thickness’ of the plane is defined by
the SET COORDINATE-TOLERANCE command.
VOLUME All nodes inside a box-shaped volume are selected; see Figure
5.1.
Note: Use the command END to conclude a node selection sequence. Do not use the ‘..’ command as
that will involve termination of the current command with no nodes selected.
ASSEMBLY
NEW
ASSEMBLY supno
OLD
PURPOSE:
This command creates a new superelement and makes it the current one — the NEW option — or an exist-
ing superelement is made the current one — the OLD option.
A NEW superelement will be empty until one or more superelements are included into it. It will then
become a second or higher level superelement.
• A first level superelement, typically created by Preframe or Prefem, and already read into Presel.
• For which boundary conditions are given when issuing the BOUNDARY command.
• Into which superelements are included when issuing the INCLUDE command.
• For which load combinations are given when issuing the LOAD command.
• Etc.
PARAMETERS:
supno The superelement number. A vacant superelement number for the NEW option, an
existing one for the OLD option.
SESAM Presel
Program version 7.5 15-MAY-2014 5-9
BOUNDARY
FREE GLOBAL
FIXED LOCAL-COORDINATE-SYSTEM coord-name
BOUNDARY PRESCRIBED *6 [ ] ...
SUPER TRANSFORMATION trano
fixcode
SELECT select-nodes
TAGGED
...
UNTAGGED
END
PURPOSE:
The command defines boundary conditions for the current superelement which must be a higher level
superelement, i.e. boundary conditions cannot be defined within Presel for first level superelements. The
following boundary conditions may be defined for each individual degree of freedom (d.o.f.) of the nodes:
FREE = Free to move
FIXED = Fixed at zero displacement
PRESCRIBED = Prescribed displacement or acceleration (value is given by the LOAD command)
SUPER = Super d.o.f.
In addition, the boundary conditions LINEAR and SUPERL are defined using the LINEAR-DEPEND-
ENCY command. SUPERL has exactly the same effect as SUPER, only that it was defined within the LIN-
EAR-DEPENDENCY command.
Note that only supernodes (or super d.o.f.s) will appear as nodes (or d.o.f.s) in the higher level superele-
ments into which this superelement is included. All other boundary conditions involve that the node (or
d.o.f.) will not exist at higher level superelements. See Section 3.2.6 for more information on this.
Nodes for which no boundary conditions are given will by default have all its d.o.f.s as FREE.
A node may have less that six d.o.f.s. This will for example be the case for membrane models in which the
nodes have three d.o.f.s. Another example is when not all six d.o.f.s were defined as super for the included
lower level superelement(s). When boundary conditions are given for such nodes the boundary conditions
for non-existent d.o.f.s will be ignored. If for example only the translations in X and Y and the rotation
about Z exist the node is fixed by the sequence:
FIX FIX * * * FIX
where * means that any legal boundary condition may be given, it will be ignored anyway.
Presel SESAM
5-10 15-MAY-2014 Program version 7.5
A node for which boundary conditions previously has been given will not be affected by a new boundary
command (that includes the node in question). I.e. boundary conditions are neither overwritten nor
‘accummulated’. The only way to given a node new boundary conditions is to delete the current conditions
(using the DELETE command) and then redefine them.
The boundary conditions may be given both in a rotated cartesian coordinate system (the TRANSFORMA-
TION option) and a cylindrical coordinate system (the LOCAL-COORDINATE-SYSTEM option). Both
these systems must previously have been defined by the TRANSFORMATION and COORDINATE-SYS-
TEM commands respectively.
The boundary conditions may be verified by the LABEL BOUNDARY-CONDITION-SYMBOL and the
PRINT NODE BOUNDARY-CONDITIONS commands. Boundary conditions are deleted by the DELETE
BOUNDARY command.
PARAMETERS:
UNTAGGED Refers to all but the previously selected (tagged) nodes; see the
TAG/UNTAG commands.
NOTES:
The brackets denote optional parameters. I.e. if you do not specify any coordinate system then the global
system is used by default.
SESAM Presel
Program version 7.5 15-MAY-2014 5-11
CHANGE
LINEAR-DEPENDENCY
LOAD
CHANGE ...
SET
TRANSFORMATION
PURPOSE:
The CHANGE LINEAR-DEPENDENCY and CHANGE LOAD commands are described in more detail in
the following.
The CHANGE SET and CHANGE TRANSFORMATION commands, however, have identical syntax with
the commands defining the data. Refer to the DEFINE SET and TRANSFORMATION commands for
details.
Presel SESAM
5-12 15-MAY-2014 Program version 7.5
CHANGE LINEAR-DEPENDENCY
PURPOSE:
The command changes linear dependencies between nodes. Only the linear dependency factor beta may be
changed. The linear dependency may originally have been defined by either the GENERAL-NODE-
DEPENDENCY or the TWO-NODE-DEPENDENCY option; see the LINEAR-DEPENDENCY com-
mand.
PARAMETERS:
dep-node-triplet Node number triplet (supno index nodeno) previously defined as dependent.
indep-node-triplet Node number triplet (supno index nodeno) previously defined as independent, the
one being supernode.
CHANGE LOAD
NODE lc
... LOAD ...
END
PURPOSE:
The command changes nodal loads previously defined by the LOAD NODE command.
Changing a nodal load is done in the same way as it was defined with one exception: a load index has to be
given. The load index is used to distinguish between different nodal loads for the same node for the same
load case. For example, a nodal force defined for the second time for the same node for the same load case
is given index 2.
The PRINT LOAD command gives an overview of the loads including the automatically assigned load
indexes, refer to this table when a nodal load is to be changed.
Rather than describing the CHANGE LOAD NODE command in detail reference is made to the LOAD
NODE command.
Load combinations cannot be changed, rather they must be deleted (DELETE LOAD COMBINATION) and
redefined.
Presel SESAM
5-14 15-MAY-2014 Program version 7.5
COORDINATE-SYSTEM
COORDINATE x y z
... *3
NODE supno index nodeno
PURPOSE:
The command defines a cylindrical coordinate system. This coordinate system may conveniently be used
for selecting nodes (see Section 5.1) and for defining boundary conditions.
A cylindrical coordinate system is defined by three points: its origin, a point defining its z-axis and a point
defining its φ=0 plane (which determines the r-axis). See Figure 5.2. The three points may be defined by
giving coordinates in the cartesian system of the superelement or nodes may be referred to.
PARAMETERS:
5.2
NOTES:
Note that the coordinate system is only defined for, and therefore available to, the current superelement.
Coordinate systems defined are printed by the PRINT TRANSFORMATION command. The coordinate
system name (coord-name) is not given in the printout but rather an internally assigned number.
Presel SESAM
5-16 15-MAY-2014 Program version 7.5
DEFINE SET
intersection-with
SUBTRACT-BY NODE select-nodes
DEFINE SET setname
UNION-WITH
END
PURPOSE:
The command defines a set of nodes that may be referred to in commands where selecting nodes is required.
DEFINE SET creates a new set while CHANGE SET changes an existing set. The command syntaxes of
these two commands are identical and based on standard set operators.
Initially, after giving the command DEFINE SET and entering a name the set is empty. The first operation to
do will therefore be to add to the set by the UNION-WITH command. Thereafter, repetitive set operations
may be performed until the content of the set is as desired. The operations are executed consecutively, the
order of the operations are therefore of consequence. Conclude the definition (or changing) of the set by
entering END.
PARAMETERS:
SUBTRACT-BY The subsequently selected nodes will be removed from the set.
DELETE
ASSEMBLY supno
SELECT select-nodes
TAGGED
BOUNDARY
UNTAGGED
END
supno index
INCLUDED
END
DELETE SELECT select-nodes
TAGGED
LINEAR-DEPENDENCY
UNTAGGED
END
COMBINATION glc
LOAD NODE ...
END
TRANSFORMATION trano
PURPOSE:
The command deletes data previously defined. Only the DELETE LOAD NODE command is described in
detail in the following. Notes are given below for the other alternatives. Otherwise see the commands defin-
ing the data.
PARAMETERS:
UNTAGGED Refers to all but the previously selected (tagged) nodes; see the
TAG/UNTAG commands.
INCLUDED Delete a lower level superelement that previously has been in-
cluded in the current higher level superelement. The current su-
perelement cannot itself be part of an assembly.
NODE Delete a nodal load for a higher level superelement. See a spe-
cific description for this alternative below.
SELECT select-nodes
TAGGED
YES
ALL UNTAGGED
END
NO
...
FORCE SELECT select-nodes
PRESCRIBED-DISPLACEMENT TAGGED
... index
UNTAGGED
PRESCRIBED-ACCELERATION
END
END
PURPOSE:
The command deletes loads. See the LOAD command for a more detailed explanation of the load types.
This command differs from the command defining the load in that a load index must be given. The load
index is used to distinguish between individual loads of the same type for the same node for the same load
case. For example, a nodal force defined for the second time for the same node for the same load case is
given index 2. Note that load indexes may change after deleting a load as the index always goes from 1 to N
where N is the number of loads of the same type for that particular node.
PARAMETERS:
UNTAGGED Refers to all but the previously selected (tagged) nodes; see the
TAG/UNTAG commands.
index Load index, either select one index or all by entering the text
ALL.
SESAM Presel
Program version 7.5 15-MAY-2014 5-21
DISPLAY
CURRENT-SUPERELEMENT
LOADED-SUPERELEMENT glc
FIRST-CONTRIBUTING-LOAD glc
LOAD
DISPLAY NEXT-CONTRIBUTING-LOAD
END
LOCATE-SUPERELEMENT supno
SPECIFIED-SUPERELEMENT supno
PURPOSE:
The command displays superelements, optionally with information about loads. The nodes of the superele-
ment are shown by small coloured dots (yellow for free nodes and blue for supernodes). These node sym-
bols may be switched off (and on again) by SET GRAPHICS NODE-SELECTION.
Node symbols, node numbers and boundary conditions may be added by the LABEL command.
Note that there is also a DISPLAY command within the command sequence for including a superelement in
an assembly. The purposes of these two DISPLAY commands are different and should not be confused. See
the INCLUDE supno DISPLAY command.
PARAMETERS:
SPECIFIED-SUPERELEMENT Display the specified superelement alone. Note that the current
superelement is not changed by this option, i.e. DISPLAY
CURRENT-SUPERELEMENT will revert to displaying the
current superelement.
This option differs from LOCATE-SUPERELEMENT in that
only the specified superelement is displayed and it need not be
a first level superelement.
SESAM Presel
Program version 7.5 15-MAY-2014 5-23
EXIT
EXIT
PURPOSE:
The command interrupts the program execution. All files opened are properly saved and closed. The user
may resume the superelement assembling at a later stage by referring to the model file and command log file
as ‘old’ when re-entering Presel.
Presel SESAM
5-24 15-MAY-2014 Program version 7.5
HELP
GENERAL-SYNTAX
SPECIAL-KEYS
HELP
STATUS-LIST
SUPPORT
PURPOSE:
The command provides information on various subjects. Except for the STATUS-LIST option the informa-
tion is printed in the message window.
PARAMETERS:
STATUS-LIST This command is obsolete. See Section 1.4 for looking up information in the Status
List.
SUPPORT The telephone and facsimile numbers and the Internet address for requesting sup-
port is printed together with detailed information on the program version used. This
information is of importance in connection with support requests.
SESAM Presel
Program version 7.5 15-MAY-2014 5-25
INCLUDE
CHECK-INCLUDE
DECODE-T-MATRIX
DISPLAY
DISTANCE-CHECK
END-DO-NOT-INCLUDE
LOCATION
MIRROR
INCLUDE supno NOPRINT-CHECK-INCLUDE ...
PERFORM-INCLUDE
POSITION
PRINT-INV-T-MATRIX
PRINT-T-MATRIX
RESET-T-MATRIX
ROTATE
TRANSLATE
PURPOSE:
The command starts the process of including a superelement in the current superelement assembly. See Sec-
tion 3.2.2 for a tutorial in how to use the INCLUDE command.
Initially, the superelement being included will be position with its coordinate system overlapping the coordi-
nate system of the superelement assembly. It may then be moved by repetitive use of the sub-commands
TRANSLATE, ROTATE, MIRROR and POSITION until it is properly positioned. The sub-command DIS-
PLAY will at any time show the current position of the superelement being included.
NOPRINT-CHECK-INCLUDE checks and tabulates the match of the nodes of the superelement being
included and the current superelement. (CHECK-INCLUDE does the same but produces more output.) It is
mandatory to perform this check.
PERFORM-INCLUDE performs and concludes the inclusion of the superelement in the current superele-
ment assembly.
The various INCLUDE sub-commands listed above are explained in the following.
PARAMETERS:
supno Superelement number being included in the current higher level superelement.
Presel SESAM
5-26 15-MAY-2014 Program version 7.5
... CHECK-INCLUDE
PURPOSE:
The sub-command compares and tabulates the match between supernodes of the superelement being
included and the nodes of the current superelement assembly. The typical appearance of the table is shown
in Figure 5.3.
5.3
Figure 5.3 CHECK-INCLUDE tabulates match between nodes, here for first level superelements
The table header contains information about including a superelement in the second level superelement
assembly 21. The coordinate tolerance is 0.1 (see the SET COORDINATE-TOLERANCE command). The
table contains the following information:
• The first line, marked A, says that the node 11 1 101 (a triplet) of the superelement being included
matches node 11 2 103 (a triplet) of the superelement assembly. And the coordinates of these nodes are
(20,20,0).
• The fourth line, marked B, says that node 11 1 303 of the superelement being included does not match
any node of the assembly. Its coordinates are (2,18,20).
Higher level superelements will have nodes with more than one triplet (see Section 3.2.6). When higher
level superelements are assembled into even higher level superelements all these triplets will appear in the
table of matching nodes. The typical appearance of the table will then be as shown in Figure 5.4
5.4
Figure 5.4 CHECK-INCLUDE tabulates match between nodes, here for higher level superelements
The table header now informs that a superelement is being included in the third level superelement assem-
bly 31. The table contains the following information:
• The two first lines, marked A, says that the node 11 2 103 (a triplet) which is the same node as 11 3 101
and both belonging to the superelement being included matches node 12 1 703 of the superelement
assembly. The coordinates of these nodes are (20,20,0).
• The third and fourth lines, marked B, says that node 11 2 303 which is the same node as 11 3 303 and
both belonging to the superelement being included matches node 12 1 803 which is the same node as 12
2 801 both belonging to the assembly. The coordinates are (18,18,20).
• The asterisk signals that there is a discrepancy in boundary condition of the matching pair of nodes
marked B. This means that the superelement cannot be included in the assembly. See Section 3.2.6 about
requirements to nodes.
NOTES:
The DISTANCE-CHECK sub-command can be used to find the distance between the two nodes. If this dis-
tance is greater than the coordinate tolerance printed in the heading of the tables the nodes will not match.
Assuming that the superelement being included has been positioned correctly there are only two ways to get
non-matching nodes to match: you must either correct the first level superelements (leave Presel and re-run
after the correction) or increase the coordinate tolerance (see the SET COORDINATE-TOLERANCE com-
mand).
SESAM Presel
Program version 7.5 15-MAY-2014 5-29
... DECODE-T-MATRIX
PURPOSE:
The sub-command prints the transformation matrix containing the accumulated translations, rotations and
mirrorings of the superelement being included. Also see the PRINT-T-MATRIX sub-command. The table
looks like this:
MIRROR ABOUT XY-PLANE: NO
ROTATION ABOUT GLOBAL X-AXIS: 0.0000
ROTATION ABOUT GLOBAL Y-AXIS: 0.0000
ROTATION ABOUT GLOBAL Z-AXIS: 90.0000
TRANSLATION IN X-DIRECTION: 20.0000
TRANSLATION IN Y-DIRECTION: 0.0000
TRANSLATION IN Z-DIRECTION: 0.0000
NOTES:
The translations and rotations refer to the coordinate system of the superelement assembly.
Starting from the initial position of the superelement being included (for example after RESET-T-MATRIX)
the translations and rotations must be made in the given sequence (first mirror about the XY-plane if rele-
vant, then rotate about X, etc.) in order to yield the position resulting from the accumulated transformations.
Presel SESAM
5-30 15-MAY-2014 Program version 7.5
... DISPLAY
PURPOSE:
The sub-command displays the current position of the superelement being included on top of the superele-
ment assembly.
5.5
PURPOSE:
The sub-command computes and prints the distance between a node or point in the superelement assembly
and the corresponding node or point in the superelement being included. The node/point of the superelement
assembly is given first.
If the COORDINATE alternative is chosen the point does not have to correspond to a node. The coordinates
are given in the coordinate systems of the assembly and superelement being included, respectively.
PARAMETERS:
... END-DO-NOT-INCLUDE
PURPOSE:
The sub-command aborts the inclusion process of the superelement and prepares for including another
superelement.
To merely discard the given transformations of the superelement currently being included use the RESET-T-
MATRIX command.
SESAM Presel
Program version 7.5 15-MAY-2014 5-33
PURPOSE:
The sub-command is used in connection with the LOAD ASSEMBLY command. It creates and modifies
location strings for superelements being included. See Section 3.4 for an explanation of assembling loads.
PARAMETERS:
CHANGE Changes a previously created location string for the superelement being included.
Not to be confused with the USE command; see below.
USE Modifies location strings for superelements forming the superelement being in-
cluded. For example, if the current superelement assembly is a fourth level su-
perelement and the superelement being included is a third level superelement, then
the command modifies location strings for the second and first level superelements
forming the third level superelement.
refloc Location string(s) to be modified. Wild-cards may be used; see notes below.
modloc The modified location string(s). Wild-cards may be used; see notes below.
NOTES:
The LOCATION USE command allows modifying location strings by use of wild-cards. The following
modifications are allowed (interpret the single characters in the examples as several characters):
refloc modloc
* X replaces all strings by X
*A X replaces all strings ending with A by X
A* X replaces all strings beginning with A by X
Presel SESAM
5-34 15-MAY-2014 Program version 7.5
YZ-PLANE
... MIRROR ZX-PLANE
XY-PLANE
PURPOSE:
The sub-command mirrors the superelement being included about one of the three planes defined by the
axes of the superelement assembly.
5.6
NOTES:
Mirroring a superelement involves that the superelement occurrence gets a left-handed coordinate system.
The results must then be interpreted in such a coordinate system.
Mirroring a superelement twice (e.g. by also mirroring the superelement assembly when this in turn is
included in yet a higher level superelement) involves that the superelement occurrence regains a right-
handed coordinate system.
Presel SESAM
5-36 15-MAY-2014 Program version 7.5
... NOPRINT-CHECK-INCLUDE
PURPOSE:
The sub-command compares the supernodes of the superelement being included and the nodes of the current
superelement assembly. It has the same purpose and functionality as the NOPRINT-CHECK-INCLUDE
sub-command except for that the table over matching nodes is omitted. Only the summary information is
given; see Figure 5.3.
SESAM Presel
Program version 7.5 15-MAY-2014 5-37
... PERFORM-INCLUDE
PURPOSE:
The sub-command concludes the process of including a superelement in an assembly. It is given subse-
quently to the sub-command NOPRINT-CHECK-INCLUDE (or CHECK-INCLUDE).
NOTES:
PURPOSE:
The sub-command positions the superelement by referring to three points or nodes of the superelement
assembly and the corresponding three points or nodes of the superelement being included. The three points/
nodes of the assembly are given first. The two sets of three points/nodes must form a triangle and they must
be congruent.
If the COORDINATE alternative is chosen the point does not have to correspond to a node. The coordinates
are given in the coordinate systems of the assembly and superelement being included, respectively.
The POSITION sub-command cannot be used if transformations have already been specified for the
superelement being included. Use the RESET-T-MATRIX sub-command in such case.
5.7
PARAMETERS:
PRINT-T-MATRIX
...
PRINT-INV-T-MATRIX
PURPOSE:
The sub-commands print the transformation matrix of the superelement being included. This 3 by 4 matrix
is an accumulation of all translations, rotations and mirrorings given for the superelement.
The first three columns of the transformation matrix (a 3 by 3 matrix) constitute the cosine matrix (rota-
tions) between the coordinate systems of the superelement being included and the superelement assembly.
The fourth column describe the translations between the systems. The conversion of coordinates from one
system to the other is done by adding a row of zeros to the transformation matrix to make it a 4 by 4 matrix
and adding a fourth term being equal to 1 to the coordinate vectors as follows:
x T 11 T 12 T 13 T 14 x'
y = T 21 T 22 T 23 T 24 y'
z T 31 T 32 T 33 T 34 z'
1 0 0 0 0 1
PARAMETERS:
C = T × C´
Where:
... RESET-T-MATRIX
PURPOSE:
The sub-command resets the transformation matrix containing the accumulated translations, rotations and
mirrorings of the superelement being included. This means that the superelement being included is brought
back to its original position with its coordinate system overlapping the coordinate system of the superele-
ment assembly. It is then ready for new transformations.
SESAM Presel
Program version 7.5 15-MAY-2014 5-41
X-AXIS
GLOBAL-AXIS Y-AXIS degrees
Z-AXIS
X-AXIS
... ROTATE
OBJECT-AXIS Y-AXIS degrees
Z-AXIS
NODE supno index nodeno
ARBITRARY-AXIS *2 degrees
COORDINATES x y z
PURPOSE:
The sub-command rotates the superelement being included an angle about a specified axis.
5.8
PARAMETERS:
GLOBAL-AXIS The superelement being included rotates about one of the coordinate axes of the su-
perelement assembly.
OBJECT-AXIS The superelement being included rotates about one of the coordinate axes of the su-
perelement itself.
Presel SESAM
5-42 15-MAY-2014 Program version 7.5
ARBITRARY-AXIS The superelement being included rotates about an axis defined by two points. The
axis points from the first to the second point which then determines the positive di-
rection of rotation (the right hand rule). The points are given by referring to nodes
or by giving coordinates in the coordinate system of the superelement assembly.
supno index nodeno The node number triplet. This node must be a part of the current assembly, i.e. be-
long to a superelement that has previously been included.
... TRANSLATE dx dy dz
PURPOSE:
The sub-command translates the superelement being included along the axes of the superelement assembly.
PARAMETERS:
LABEL
BOUNDARY-CONDITION-SYMBOLS
COUPLED-NODES
EXTERNAL-NODE-NUMBER supno index
INTERNAL-NODE-NUMBER
NODE-NUMBERS supno index
LABEL NODE-NUMBER-TRIPLET ONE-NODE-NUMBER-TRIPLET
ALL-NODE-NUMBER-TRIPLETS
ALL-NODES
NODE SYMBOLS
SUPER-NODES-ONLY
NON-COUPLED-NODES
ORIGIN-SYMBOL
PURPOSE:
The command adds (labels) boundary condition symbols, node numbers, etc. to the display. The labels are
shown until a new display is made; the label command may then be re-entered. The size of the symbols are
adjusted by the SET GRAPHICS SIZE-SYMBOLS command. The symbols used are shown in Figure 5.9.
PARAMETERS:
COUPLED-NODES Add numbers to the display telling how many first level su-
perelements there are coupled to each node. The numbers are
only given for nodes where two or more first level superele-
ments are coupled, i.e. the number will always be ≥2. Also see
the NON-COUPLED-NODES alternative.
NODE-NUMBERS Add node numbers. A choice must be made between various al-
ternatives. To understand these alternatives; see Section 3.2.6
about node numbers.
EXTERNAL-NODE-NUMBER Add only the node number of a given first level superelement.
INTERNAL-NODE-NUMBER Add only the internal number of the current superelement as-
sembly (normally, this is of little interest to the user).
NODE-NUMBER-TRIPLET Add one or all node number triplets — the full and unique ref-
erences to nodes (see Section 3.2.5 about triplets).
SESAM Presel
Program version 7.5 15-MAY-2014 5-45
supno index Select one of the possibly several triplets by giving superele-
ment number and index.
ONE-NODE-NUMBER-TRIPLET Only one of the possibly several triplets is shown. The program
automatically selects which of the triplets to show.
ALL-NODE-NUMBER-TRIPLETS All triplets are shown reflecting the fact that the nodes have one
triplet for each first level superelement connected.
NON-COUPLED-NODES Add the number 1 to the display for all nodes to which only a
single first level superelement is coupled. There will be no label
for nodes where two or more first level superelements are cou-
pled. This alternative is the complement to COUPLED-
NODES.
5.9
LINEAR-DEPENDENCY
GENERAL-NODE-DEPENDENCY
LINEAR-DEPENDENCY TWO-NODE-DEPENDENCY ...
END
PURPOSE:
The command defines the displacements of selected nodes to be linearly dependent of displacements of
other selected nodes. See also Section 3.6.
The GENERAL-NODE-DEPENDENCY option couples any d.o.f. of a node (the dependent d.o.f.) to any
other d.o.f.s of any other nodes (the independent d.o.f.s). The TWO-NODE-DEPENDENCY option couples
all d.o.f.s of a given node to the corresponding d.o.f.s of two other nodes.
Linear dependencies involves that the dependent d.o.f.s get the boundary condition LINEAR and the inde-
pendent d.o.f.s get the boundary condition SUPERL (super due to linear dependency). (A SUPERL d.o.f.
will appear in the next level superelement in the same way as a SUPER d.o.f.) There are certain rules as con-
cerns the boundary condition of a d.o.f. before and after the definition of a linear dependency and whether
the linear dependency can at all be defined. Table 5.1 describes these rules for the dependent d.o.f. and Table
5.2 for the independent d.o.f. A violation of the rules involves that the linear dependency is not accepted.
SUPER SUPERL OK
SUPERL SUPERL OK
Presel SESAM
5-48 15-MAY-2014 Program version 7.5
LINEAR-DEPENDENCY GENERAL-NODE-DEPENDENCY
PURPOSE:
The command defines general linear dependency between nodes. See also Section 3.6.
The dependency is defined by selecting a single d.o.f. of a node to be dependent of any other d.o.f.s of any
other nodes. You may keep defining several d.o.f.s of a given node to be dependent. And for each of these
dependent d.o.f. you may keep selecting independent (governing) nodes. And, finally, for each of these
independent nodes you may keep selecting d.o.f.s to govern, with a factor, the displacement of the depend-
ent d.o.f.
Note: Alternatively to defining the independent d.o.f. as super prior to this command it may be made
super within this command by the FORCE-...-INTO-SUPER option. Using this option for a
d.o.f. that is already super has no consequence.
PARAMETERS:
dep-node-triplet Node number triplet (supno index nodeno) of the dependent node.
R-X, R-Y and R-Z: Rotations about the X-, Y- and Z-directions
R-X, R-Y and R-Z: Rotations about the X-, Y- and Z-directions
If the independent d.o.f. has not previously been defined as SUPERL (or SUPER)
then use the appropriate of the FORCE-...-INTO-SUPER alternatives to force it
into being SUPERL.
LINEAR-DEPENDENCY TWO-NODE-DEPENDENCY
indep-node1-triplet
... ...
FORCE-INTO-SUPER indep-node1-triplet
indep-node2-triplet
... beta }*
FORCE-INTO-SUPER indep-node2-triplet
PURPOSE:
The command defines linear dependency for a node on two other nodes. All d.o.f.s of the dependent node
are dependent of the corresponding d.o.f.s of the first independent node by the factor beta and the second
independent node by the factor (1 - beta). See also Section 3.6.
Note that alternatively to defining the independent node as super prior to this command it may be made
super within this command by the FORCE-INTO-SUPER option. Using this option for a node already being
super has no consequence.
PARAMETERS:
dep-node-triplet Node number triplet (supno index nodeno) of the dependent node.
indep-node1-triplet Node number triplet (supno index nodeno) of the first independent node.
FORCE-INTO-SUPER Using this option implies that the d.o.f.s of the independent nodes are forced into
SUPERL if they are not SUPER or SUPERL already.
indep-node2-triplet Node number triplet (supno index nodeno) of the second independent node.
LOAD
ASSEMBLY
LOAD COMBINATION ...
NODE
PURPOSE:
The command defines loads for the current superelement which must be a second or higher level superele-
ment.
PARAMETERS:
ASSEMBLY This option allows assembling loads directly to the top level. See Section 3.4 for
more information on this.
COMBINATION This option creates a load for a superelement assembly by combining loads of its
included superelements.
LOAD ASSEMBLY
PURPOSE:
The command defines loads for the current superelement assembly (normally the top level superelement) by
combining loads of lower level superelements (normally 1st level superelements). The lower level superele-
ments are referred to by names and location strings. Load combinations are automatically created for all
intermediate level superelements. The use of the command is explained in Section 3.4.
The load combinations created for the current superelement assembly and all intermediate level superele-
ments may be verified by the PRINT LOAD command.
PARAMETERS:
INCLUDE-LOAD Include in the global load case the following lower level superelement loads.
refname.refloc Refers to superelement occurrences. refname is the name of the lower level su-
perelement. refloc is the location string of the superelement occurrence. refname
and refloc are given as indicated: separated by a punctuation mark only. The aster-
isk (*) may be used as wild-card.
*.* are all lower level superelements and all location strings of these. I.e. absolutely
all superelement occurrences at any level for which names and location strings
have been defined.
*A.*X are all lower level superelements with names ending with A and all location
strings of these ending with X.
A*.X* are all lower level superelements with names beginning with A and all lo-
cation strings of these beginning with X.
A*B*.*X*Y are all lower level superelements with names beginning with A and
containing B and all location strings of these containing X and ending with Y.
llc (Local) load case number of the lower level superelements being referred to.
LOAD COMBINATION
{ llc factor }*
{ supno index }*
glc END
END
...
{ supno index [STEP] lowllc [incr] factor }*
GROUP lowglc higlc step
END
END
PURPOSE:
The command defines loads for the current superelement by combining loads of the included superelements.
The current superelement must be a second or higher level superelement. The loads may be combined one-
by-one (explicitly) or a group of loads may be combined. See Section 3.3 for an explanation of the princi-
ples of combining loads.
Several superelement occurrences (superelement number and index) and several (local) load cases belong-
ing to the superelement occurrences may contribute to a single (global) load case of the current superele-
ment.
The (local) load cases (llc) of included first level superelements need not exist prior to giving this command,
warnings are then given saying that the input is accepted even though the local load cases are unknown.
Prior to running the analysis, however, the local load cases must have been created. Read the note in Section
3.3.4 on this.
PARAMETERS:
glc A single (global) load case number to be defined for the current superelement.
llc (Local) load case number of the included superelement contributing to glc.
lowglc higlc step The lowest and highest (global) load case numbers and the step (increment) in
numbering. An example: 1 5 2 will define loads 1, 3 and 5.
STEP A step (increment) in numbering of the (local) load cases of the included superele-
ments is to be given; see incr below. If this command is omitted then incr is omitted
as well.
SESAM Presel
Program version 7.5 15-MAY-2014 5-53
lowllc The lowest (local) load case number of the included superelement contributing to
glc.
LOAD NODE
FORCE
... NODE lc PRESCRIBED-ACCELLERATION ...
PRESCRIBED-DISPLACEMENT
SELECT select-nodes
TAGGED
... ...
UNTAGGED
END
GLOBAL
... fx fy fz mx my mz ...
TRANSFORMTAION trano
PURPOSE:
The command defines nodal loads for the current superelement. The current superelement must be a second
or higher level superelement.
• Prescribed displacements
Prior to giving prescribed displacements and accelerations the corresponding nodes must previously have
been given the PRESCRIBED boundary condition; see the BOUNDARY command. Alternatively to all six
only selected d.o.f.s may be given prescribed displacements/accelerations. The PRESCRIBED boundary
condition must then have been defined only for the relevant d.o.f.s. Note that even if only selected d.o.f.s
have prescribed boundary condition values must be entered in the LOAD NODE command for all six
d.o.f.s. The values given for the non-prescribed d.o.f.s are discarded.
PARAMETERS:
UNTAGGED Refers to all but the previously selected (tagged) nodes; see the
TAG/UNTAG commands.
ifx ify ifz imx imy imz Imaginary components of the forces/moments or prescribed
displacements/accelerations. Prescribed rotations are given in
radians.
PHASE-COMPLEX The load is complex and the phase angle components are to be
given.
pfx pfy pfz pmx pmy pmz Phase angle components in degrees of the forces/moments or
prescribed displacements/accelerations.
NAME
PURPOSE:
The command creates a name for a superelement. It is used in connection with the LOAD ASSEMBLY
command. See Section 3.4 for an explanation of assembling loads.
PARAMETERS:
name Name given to the current first or higher level superelement. It is a string of maxi-
mum 8 characters.
SESAM Presel
Program version 7.5 15-MAY-2014 5-57
NODE INEQUALITY
node-a-degree ---
node-a-triplet node-b-triplet {
NODE INEQUALITY { END
END
GREATER-OR-EQUAL
... LESS-OR-EQUAL ...
EQUAL
NODE-DOF node-b-degree
... scale }* }*
CONSTANT
PURPOSE:
Nodal inequalities is used to solve the small displacement elasto-static node-to-node contact problem.
The nodal inequalities may be verified by the PRINT NODE INEQUALITY <select-nodes> command.
PARAMETERS:
NOTES
For documentation on using the inequalities see the documentation for Sestra.
Presel SESAM
5-58 15-MAY-2014 Program version 7.5
OPTIMIZE
BANDWIDTH
OPTIMIZE supno OPTIONS
PROFILE
PURPOSE:
The command optimises (rearranges) the internal node numbering of a higher level superelement. The
objective of this is to reduce the bandwidth or profile of the stiffness matrix of the superelement thereby
reducing the time required to solve the equation system of the superelement (this is done in the analysis pro-
gram, e.g. Sestra).
Note: The node numbers seen by the user, the node number triplets, are not affected by this opera-
tion.
The superelement cannot be a first level superelement. First level superelements must be optimised prior to
being read into Presel. See Prefem or Preframe on this.
If a superelement is to be optimised it must be done prior to being included in a higher level superelement
assembly.
PARAMETERS:
BANDWIDTH The internal node numbering is optimised to reduce the bandwidth of the stiffness
matrix. This is the preferred option for Sestra.
PROFILE The internal node numbering is optimised to reduce the profile of the stiffness ma-
trix. This alternative is currently irrelevant.
SESAM Presel
Program version 7.5 15-MAY-2014 5-59
PLOT
AS-LAST-DISPLAY
PLOT text *4 page-size
choices
Where ‘choices’ in the command syntax above represents all of the following:
ALL-NODES
Mesh on plot? Origin symbol? Boundary conditions?
... SUPER-NODES-ONLY ...
YES or NO YES or NO YES or NO
NONE
COUPLED-NODES
NON-COUPLED-NODES
EXTERNAL-NODE-NUMBER supno index
INTERNAL-NODE-NUMBER
... ...
supno index
NODE-NUMBER-TRIPLET ONE-NODE-NUMBER-TRIPLET
ALL-NODE-NUMBER-TRIPLETS
NONE
PURPOSE:
The command reproduces the display on a plot file (or sends it directly to the printer in case of WINDOWS-
PRINTER format of the plot file).
Unless the AS-LAST-DISPLAY option is chosen the command poses the following questions:
• Include a symbol for the origin of the superelement coordinate system? Answer YES or NO.
Whether the AS-LAST-DISPLAY option is chosen or not the PLOT command is concluded by:
• Enter four lines of text. Each text line is limited to 24 characters and must be enclosed in apostrophes if
containing blanks. For example: 'THIS IS A TEXT'. These lines are reproduced on the plot.
Presel SESAM
5-60 15-MAY-2014 Program version 7.5
• Finally the page size is given. This is only relevant for the SESAM-NEUTRAL plot format; see the SET
PLOT command. For other plot formats give any parameter, e.g. the default A4.
The date and time is also reproduced on the plot together with scale, superelement number and the superele-
ment level. The scale is based on the assumption that metres are used as unit for the coordinates.
PARAMETERS:
AS-LAST-DISPLAY The screen display with current labelling and other display in-
formation is plotted.
text The screen display with current labelling and other display in-
formation is plotted.
page-size The page size. Choose between A1, A2, A3, A4 and A5. A4 is
the default choice.
COUPLED-NODES Add numbers to the plot telling how many first level superele-
ments there are coupled to each node. The numbers are only
given for nodes where to or more first level superelements are
coupled, i.e. the number will always be ≥2. Also see the NON-
COUPLED-NODES alternative.
NON-COUPLED-NODES Add the number 1 to the plot for all nodes to which only a single
first level superelement is coupled. There will be no label for
nodes where two or more first level superelements are coupled.
This alternative is the complement to COUPLED-NODES.
EXTERNAL-NODE-NUMBER Add only the node number of a given first level superelement.
INTERNAL-NODE-NUMBER Add only the internal number of the current superelement as-
sembly (normally, this is of little interest to the user).
NODE-NUMBER-TRIPLET Add one or all node number triplets — the full and unique ref-
erences to nodes (see Section 3.2.5 about triplets).
ONE-NODE-NUMBER-TRIPLET Only one of possibly several triplets is shown. The program au-
tomatically selects which of the triplets to show.
ALL-NODE-NUMBER-TRIPLETS All triplets are shown reflecting the fact that the nodes have one
triplet for each first level superelement connected.
SESAM Presel
Program version 7.5 15-MAY-2014 5-61
supno
ALL
supno
CPU-TIME-ESTIMATES-IN-REDUCTION
ALL
ELEMENT
LOAD ...
NODE ...
PRINT
OVERVIEW-OF-SUPER-ELEMENTS
STATUS
SUPER-ELEMENT-HIERARCHY supno index
trano
TRANSFORMATION ALL
END
END
PURPOSE:
The command prints data in tables on screen and to file. The destination depends on what to print: e.g.
PRINT ALL goes to file whereas PRINT STATUS goes to the screen. The SET PRINT DESTINATION
command overrules these default destinations. Long prints are broken into several sub-tables, each limited
to a certain number of lines. When printing to screen in interactive mode enter CONTINUE to print the next
sub-table (or END to stop). The SET PRINT PAGESIZE command changes the number of lines contained
in each sub-table.
PARAMETERS:
STATUS Print on screen model and log file names and cur-
rent tolerances.
PRINT LOAD
ALL-TYPES-NODES-COMBINATIONS
COMBINATIONS
lc NODE-FORCE
...
NODE-PRESCRIBED
... LOAD END
ALL-LOADCASES
lc
OVERVIEW ALL-LOADCASES
END
SELECT select-nodes
TAGGED
...
UNTAGGED
END
PURPOSE:
PARAMETERS:
ALL-TYPES-NODES-COMBINATIONS Print all types of loads for the selected load case number.
COMBINATIONS Print only the load combinations belonging to the selected load
case number.
NODE-FORCE Print only the nodal forces, for selected nodes, belonging to the
selected load case number.
UNTAGGED Refers to all but the previously selected (tagged) nodes; see the
TAG/UNTAG commands.
OVERVIEW Print an overview for a selected load case or for all load cases.
SESAM Presel
Program version 7.5 15-MAY-2014 5-65
PRINT NODE
PURPOSE:
Note that you may change the print format for real numbers (FORTRAN E, F or G formats) by the SET
PRINT FORMAT command. E-format is the default choice.
PARAMETERS:
COORDINATES Print coordinates for selected nodes. See example print in Sec-
tion 3.8.
NUMBER Print the node numbers of selected nodes. The table shows the
node number triplets along with the internal node numbers.
UNTAGGED Refers to all but the previously selected (tagged) nodes; see the
TAG/UNTAG commands.
SESAM Presel
Program version 7.5 15-MAY-2014 5-67
READ
supno
READ
SHOW-PROGRESS supno
PURPOSE:
The command reads a first level superelement into Presel’s database. That is, the Input Interface File of the
superelement is read. The Input Interface File must have the following name, also see Section 2.3:
prefixTsupno.FEM
Note: When using the SESAM Manager to control your analysis the prefix will normally be void.
Note: The prefix is given when starting Presel. This means that if the Input Interface File has a pre-
fix then this prefix must be given at start-up. Also note that the Input Interface File names of
all first level superelements must have the same prefix.
PARAMETERS:
SHOW-PROGRESS This option causes feedback to be given on how many cards (records) are read. The
feedback looks like this:
By default no such feedback is given. (In previous versions of Presel such feedback
was by default given.)
Presel SESAM
5-68 15-MAY-2014 Program version 7.5
ROTATE
X-AXIS
ROTATE Y-AXIS degrees
Z-AXIS
PURPOSE:
The command rotates the display about the coordinate axes of the current superelement.
The SET GRAPHICS EYE-DIRECTION and the interactive rotations provide alternative ways of rotating
the display. See the ‘direct access buttons’ described in Section 3.1 about interactive rotation.
PARAMETERS:
SET
ANGLE-TOLERANCE angtol
COMMAND-INPUT-FILE comfilnam
COORDINATE-TOLERANCE cotol
GRAPHICS ...
GRAPHICS ON
JOURNALLING ...
SET PRINT OFF
NEW
MODEL-FILE prefix filnam
OLD
PLOT ...
PRINT ...
UNIT-VECTOR-TOLERANCE uvtol
PURPOSE:
The command sets different parameters for controlling the execution of other commands.
PARAMETERS:
COMMAND-INPUT-FILE Specify a command input file. The file is opened and is ready
for reading using the # command. The command input file can-
not have the same name as the Presel command log file.
comfilnam Name of the command input file. The file extension must be
JNL and shall not be given.
COORDINATE-TOLERANCE Specify the coordinate tolerance used for deciding whether two
points (nodes) have the same geometrical position and for de-
ciding whether a node lies in a plane or on a line.
cotol The coordinate tolerance given in the same unit as the coordi-
nates. The default value is 0.1.
GRAPHICS Set various parameters controlling the display and plot. The
command is described in detail in the following.
Presel SESAM
5-70 15-MAY-2014 Program version 7.5
MODEL-FILE Close the current model file and open another without exiting
and re-entering the program.
NEW OLD NEW means that the model file will be created (starting a new
session from scratch), OLD means that it exists already (contin-
uing an earlier session).
UNIT-VECTOR-TOLERANCE Specify a unit vector tolerance used for deciding whether two
vectors span a plane and whether a matrix is orthonormal.
uvtol The unit vector tolerance (without unit) has default value 0.001.
SESAM Presel
Program version 7.5 15-MAY-2014 5-71
SET GRAPHICS
ALTERNATIVE-SCREEN-DEVICE
AUTO
BASIC-ELEMENT-MODE
SOFTWARE
CHARACTER-TYPE
HARDWARE
BOUNDARY-CONDITION
ELEMENT-LINES
INCLUDED-SUPERELEM
COLOUR colour tone
NODE-NUMBER
NODE-SYMBOL
SUPER-NODES
DEVICE device-name
EYE-DIRECTION eyex eyey eyez
ON
HIDDEN
OFF
ON
INPUT
... GRAPHICS OFF
ON
NODE-SELECTION
OFF
PLOT-FILE prefix filnam
ON
COLOUR-SUPERELEMENTS
OFF
PRESENTATION
ON
FILLED-ELEMENT
OFF
ON
SCALING-AUTOMATIC
OFF
SHRINK-FACTOR shrinkfac
BOUNDARY-CONDITION-SYMBOLS
LOAD-NUMBERS
NODE-NUMBERS
SIZE-SYMBOLS size
NODE-SYMBOLS
ONE-NODED-ELEMENT-SYMBOLS
ORIGIN-SYMBOL
Presel SESAM
5-72 15-MAY-2014 Program version 7.5
PURPOSE:
The command sets different parameters for controlling the DISPLAY, PLOT and LABEL commands.
PARAMETERS:
EYE-DIRECTION Set the viewpoint for the display. Note that the default view-
point is set through command line arguments, see Section 4.1.5,
and these may in turn be set by Manager. You can also use the
ROTATE command or the interactive rotations using the ‘direct
access buttons’ described in Section 3.1.
SESAM Presel
Program version 7.5 15-MAY-2014 5-73
eyex eyey eyez X-, Y- and Z-coordinates of the viewpoint (the eye).
HIDDEN Switch hidden display mode ON and OFF. The default is OFF.
PLOT-FILE Set the name of the plot file. By default it is the same as the
model and command log files. The extension of the plot file de-
pends on the plot format; see the SET PLOT FORMAT com-
mand. This command has the same functionality as the SET
PLOT FILE command. You may want to use the latter as it is
more consistent with the other SET PLOT commands. Note
that the command closes the current plot file (if such exists) en-
abling this to be sent to a laser printer without having to exit
Presel.
NODE-SYMBOLS Symbols for the nodes will be re-sized to the given value.
ONE-NODED-ELEMENT-SYMBOLS Symbols for one node elements will be re-sized to the given
value. These are the elements connected to only one node, i.e.
SPRING-TO-GROUND and DAMPER-TO-GROUND.
ORIGIN-SYMBOLS The symbol for the origin will be re-sized to the given value.
SET PLOT
ON
COLOUR
OFF
FILE prefix filnam
number
CGM-BINARY
HPGL-2
FORMAT HPGL-7550
POSTSCRIPT
... PLOT
SESAM-NEUTRAL
WINDOWS-PRINTER
ORIENTATION PORTRAIT
A1
A2
PAGE-SIZE A3
A4
A5
PURPOSE:
The command sets parameters for plotting. The settings must be done prior to giving the PLOT command.
PARAMETERS:
COLOUR Switch ON or OFF colours. The default is OFF. Colours are supported by the for-
mats PostScript, HPGL-2 and CGM. Give this command after the SET PLOT
FILE/FORMAT commands and prior to the PLOT command.
FILE Set the name of the plot file. By default it is the same as the model and command
log files. The extension of the plot file depends on the plot format; see the SET
PLOT FORMAT command. Note that the command closes the current plot file (if
such exists) enabling this to be sent to a laser printer without having to exit Presel.
number The plotter number may alternatively be given. However, you will normally not
know this.
CGM-BINARY The ISO 8632-3 Computer Graphics Metafile (CGM) plot format (binary encod-
ing). File extension is .CGM. This format is convenient for including plots in re-
ports; see more information on this in Section 4.1.4.
SESAM-NEUTRAL A plot format of the SESAM system. This is the default format on most computers.
File extension is .PLO. The SESAM auxiliary program Pltcnv is required to repro-
duce the plot on paper.
WINDOWS-PRINTER A plot file will not be created rather the plot will be sent directly to an on-line print-
er.
ORIENTATION Set the page orientation. This command is presently redundant as portrait is the
only choice.
PAGE-SIZE Set the plot page size. All sizes are not available for all plot formats. For SESAM-
NEUTRAL this setting is irrelevant as the page size is set within the PLOT com-
mand. Give this command after the SET PLOT FILE/FORMAT commands and
prior to the PLOT command.
A1 A2 A3 A4 A5 European standard page sizes (paper formats). See explanation for the PLOT com-
mand. A4 is the default choice.
NOTES:
For PostScript and HPGL-2 the size specification in the PLOT command is dummy. It will not change the
plot size. The specification has not been removed to ensure compatibility with old input files.
For SESAM-NEUTRAL format the SET PLOT PAGE-SIZE has no effect as the size specification within
the PLOT command is used.
SESAM Presel
Program version 7.5 15-MAY-2014 5-77
SET PRINT
FILE
DESTINATION
SCREEN
LINEPRINTER
FILE
NAME filnam
E
FORMAT F
... PRINT
G
FILE
PAGESIZE nlines
SCREEN
DIGITS
NODE-BOUNDARY-TABLE
TABLE TEXT
SUPER-ELEMENT-HIERARCHY width
PURPOSE:
The command sets different parameters controlling the execution of the PRINT command.
PARAMETERS:
FILE Decide the name of the print file. The default name is the same
as the model (and command log) file name. The file extension
is .LIS.
PAGESIZE Decide number of lines printed for each ‘page’. The table head-
ing is repeated for each page.
Presel SESAM
5-78 15-MAY-2014 Program version 7.5
SCREEN Decide number of lines printed for each ‘page’ on screen. After
each page give CONTINUE to print the next page and END to
quit printing. Note that the default command at this stage is
CONTINUE. Therefore, giving the command ‘;’ (semicolon)
— which accepts all subsequent defaults — will print all pages.
SUPER-ELEMENT-HIERARCHY Set the number of character positions used for each level in the
table showing the superelement hierarchy which is printed by
the PRINT SUPER-ELEMENT-HIERARCHY command. See
the example print in Figure 3.3.
TAG
TAG select-nodes
PURPOSE:
The command tags (in effect: puts into a set) some or all nodes for the purpose of referring to these nodes
(the TAGGED), or to all other nodes (the UNTAGGED), in subsequent commands (e.g. for defining bound-
ary conditions). Initially, no nodes are tagged.
PARAMETERS:
TASK
ASSEMBLY
TASK
SUB-MODELLING
PURPOSE:
The command switches between the modes ASSEMBLY which is the normal way of using Presel as
explained in this manual and SUB-MODELLING which is for sub-modelling as explained in the Submod
User Manual.
As it is normally more convenient to access Submod directly rather than through Presel the TASK command
has little relevance.
SESAM Presel
Program version 7.5 15-MAY-2014 5-81
TRANSFORMATION
PURPOSE:
The command defines a rotated coordinate system. A transformation matrix is established that transforms
coordinates from a rotated coordinate system to the global coordinate system. The purpose of the command
is:
• For specifying a fixation or a prescribed displacement in a rotated (transformed) coordinate system. See
the BOUNDARY command.
• For specifying nodal loads in a rotated (transformed) coordinate system. See the LOAD NODE com-
mand.
The transformation matrix is defined by giving the global coordinates of a second point (SP) and a guiding
point (GP). The x-axis of the transformed coordinate system, XT, goes from the origin to SP. The trans-
formed z-axis, ZT, is perpendicular to XT so that GP lies in the XT-ZT plane on the positive ZT side. YT is
perpendicular to XT and ZT. See Figure 5.10.
PARAMETERS:
5.10
UNTAG
UNTAG select-nodes
PURPOSE:
The command untags (in effect: removes from a set) some or all nodes for the purpose of referring to these
nodes (the UNTAGGED), or to all other nodes (the TAGGED), in subsequent commands (e.g. for defining
boundary conditions). Initially, all nodes are untagged.
PARAMETERS:
WRITE
WRITE supno
PURPOSE:
The command writes Input Interface Files for the given superelement supno and all superelements in the
hierarchy below supno except for the first level superelements (which exist before the execution of Presel).
ZOOM
FRAME
ZOOM IN
...
OUT
PURPOSE:
The zoom buttons of the ‘direct access button area’ have the same effect and are quicker in use; see Section
3.1. However, the ZOOM command differs from the ‘direct access buttons’ in that they are for command
logging and input purposes. Logging is subject to use of the SET JOURNALLING GRAPHICS command.
The FRAME option makes the displayed model fill the graphic area.
For the IN and OUT options use the mouse (left mouse button) to give the zoom area. You may either:
• press and hold while dragging to the opposite corner of a rectangle and then release, or
• click first one corner of a rectangle and then click the diagonally opposite corner.
The actual zoom area is the smallest square containing the rectangle.
ZOOM IN will magnify the part of the picture that is inside the zoom area.
ZOOM OUT will fit the picture into the zoom area.
SESAM Presel
Program version 7.5 15-MAY-2014 5-85
ncomnd
#
ALL
PURPOSE:
The command reads commands from the command input file. The command input file is opened by the
command SET COMMAND-INPUT-FILE. The command input file may be a command log file from a pre-
vious run or a file prepared by a text editor.
The program will execute commands from the command input file until:
• an end-of-file is detected,
PARAMETERS:
ALL All commands are read from the command input file.
Presel SESAM
5-86 15-MAY-2014 Program version 7.5
SESAM Presel
Program version 7.5 15-MAY-2014 APPENDIX A-1
In order to perform the tutorial examples of Chapter 3 some superelements created by Prefem are required.
The inputs for these are presented in the following.
A.1
Figure A.1 The two first level superelements 5 and 6 created by Prefem
A 1.1 Superelement 5
%=============================================================================
%
% Prefem input for creating superelement number 5.
Presel SESAM
APPENDIX A-2 15-MAY-2014 Program version 7.5
A 1.2 Superelement 6
%=============================================================================
%
% Prefem input for creating superelement number 6.
% Semicolons (;) are used to accept default values.
%
%=============================================================================
% --- First create geometry:
GENERATE SURFACE A 1 2 1 4 1 2 1 1 END
CARTESIAN 0 0 0
4 0 0 END 0 1 0 END
% --- Then select 4 node shell element to be used:
SET ELEMENT-TYPE SURFACE ALL-SURFACES-INCLUDED SHELL-4NODES
..
% --- Define supernodes on two lines:
PROPERTY BOUNDARY-CONDITION ( AI11 AJ21 )
SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE
GLOBAL
SESAM Presel
Program version 7.5 15-MAY-2014 APPENDIX A-3
..
% --- Define thickness:
PROPERTY THICKNESS ALL-SURFACES-INCLUDED .01
..
% --- Define and connect material:
PROPERTY MATERIAL STEEL ELASTIC ;
..
CONNECT MATERIAL STEEL ALL-SURFACES-INCLUDED END
% --- Define line loads:
PROPERTY LOAD 3 LINE-LOAD AJ11 GLOBAL 1.5 0 0 ;
;
PROPERTY LOAD 4 LINE-LOAD AI12 GLOBAL 0 -3.0 0 ;
;
% --- Create mesh
MESH ALL
% --- The model is now complete. Exit Prefem.
%=============================================================================
A.2
A 2.1 Superelement 1
%=============================================================================
%
% Prefem input for creating superelement number 1, the foundation.
%
%=============================================================================
% --- First create geometry:
GENERATE SURFACE A 1 2 1 4 1 2 1 2 1 2 1 1
CARTESIAN 0 0 0
7 0 0 END 0 4 0 END 0 0 0.8 END
DELETE GEOMETRY AU*
..
% --- Then select 8 node shell element to be used:
SET ELEMENT-TYPE SURFACE ALL-SURFACES-INCLUDED SHELL-8NODES
..
% --- Define fixations and supernodes:
PROPERTY BOUNDARY-CONDITION
( AJ&&1 AI&&1 )
FIX FIX FIX FIX FIX FIX
GLOBAL
( AJ&&2 AI&&2 )
SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE
GLOBAL
..
% --- Define thickness:
PROPERTY THICKNESS ALL-SURFACES-INCLUDED 0.1
..
% --- Define and connect material:
PROPERTY MATERIAL CONCR ELASTIC 0.3E11 0.25 2500 0 0
..
CONNECT MATERIAL CONCR ALL-SURFACES-INCLUDED END
% --- Define line load:
PROPERTY LOAD 1 GRAVITY GLOBAL FLEXIBLE-PART-CONTRIBUTION 0 0 -9.81
END
..
% --- Create mesh
MESH ALL
% --- The model is now complete. Exit Prefem.
%=============================================================================
A 2.2 Superelement 2
%=============================================================================
%
% Prefem input for creating superelement number 2, the long wall.
%
%=============================================================================
% --- First create geometry:
GENERATE SURFACE A 1 2 1 4 1 2 1 2 END
CARTESIAN 0 0 0.8
7 0 0 END 0 0 3 END
SESAM Presel
Program version 7.5 15-MAY-2014 APPENDIX A-5
A 2.3 Superelement 3
%=============================================================================
%
% Prefem input for creating superelement number 3, the short wall.
%
%=============================================================================
% --- First create geometry:
GENERATE SURFACE A 1 2 1 2 1 2 1 2 END
CARTESIAN 0 0 0.8
0 4 0 END 0 0 3 END
% --- Then select 8 node shell element to be used:
SET ELEMENT-TYPE SURFACE ALL-SURFACES-INCLUDED SHELL-8NODES
..
% --- Define supernodes:
PROPERTY BOUNDARY-CONDITION ALL-LINES-INCLUDED
SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE
GLOBAL
..
% --- Define thickness:
PROPERTY THICKNESS ALL-SURFACES-INCLUDED 0.04
..
% --- Define and connect material:
PROPERTY MATERIAL CONCR ELASTIC 0.3E11 0.25 2500 0 0
..
CONNECT MATERIAL CONCR ALL-SURFACES-INCLUDED END
Presel SESAM
APPENDIX A-6 15-MAY-2014 Program version 7.5
A 2.4 Superelement 4
%=============================================================================
%
% Prefem input for creating superelement number 4, the floor.
%
%=============================================================================
% --- First create geometry:
GENERATE SURFACE A 1 2 1 4 1 2 1 2 END
CARTESIAN 0 0 0
7 0 0 END 0 4 0 END
% --- Then select 8 node shell element to be used:
SET ELEMENT-TYPE SURFACE ALL-SURFACES-INCLUDED SHELL-8NODES
..
% --- Define supernodes:
PROPERTY BOUNDARY-CONDITION ALL-LINES-INCLUDED
SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE SUPERNODE
GLOBAL
..
% --- Define thickness:
PROPERTY THICKNESS ALL-SURFACES-INCLUDED 0.05
..
% --- Define and connect material:
PROPERTY MATERIAL CONCR ELASTIC 0.3E11 0.25 2500 0 0
..
CONNECT MATERIAL CONCR ALL-SURFACES-INCLUDED END
% --- Define loads:
PROPERTY LOAD 1 GRAVITY GLOBAL FLEXIBLE-PART-CONTRIBUTION 0 0 -9.81
END
LOAD 2 NORMAL-PRESSURE ALL-SURFACES-INCLUDED 3000 END MIDDLE-SURFACE
END
..
% --- Create mesh
MESH ALL
% --- The model is now complete. Exit Prefem.
%=============================================================================
A 2.5 Superelement 5
%=============================================================================
%
SESAM Presel
Program version 7.5 15-MAY-2014 APPENDIX A-7
APPENDIX B THEORY
Figure B.1 shows a very simple plate model. The equation of equilibrium for the single superelement, which
constitute the whole model, is:
Kr = R (B.1)
where K is the stiffness matrix, r is the displacement (d.o.f.) vector and R is the load vector. Solving this
equation system in one operation yields the displacements for the whole model:
–1 (B.2)
r = K R
Presel SESAM
APPENDIX B-2 15-MAY-2014 Program version 7.5
Figure B.2 shows the same model as in the previous section but now it is divided into two superelements
(part models) A and B which assembled constitute the complete model C.
Stiffness matrix, displacement vector and load vector for each superelement are established:
KA, rA, RA for superelement A
KB, rB, RB for superelement B
We now want to perform a reduction to eliminate the internal d.o.f.s for both superelements (rAi and rBi).
Let the following equation system be the equation of equilibrium of one of superelements A and B (sub-
scripts A/B are skipped):
Kr = R (B.3)
This can be written in a partitioned form by sorting the d.o.f.s of r so that all internal d.o.f.s come first, ‘i’
denotes internal d.o.f.s and ‘s’ denotes super d.o.f.s:
K ii K is r i Ri
= (B.4)
T
K is K ss r s Rs
K ii r i + K is r s = R i (B.5)
T
K is r i + K ss r s = R s (B.6)
–1 –1
r i = – K ii K is r s + K ii R i (B.7)
T –1 T –1
( K ss – K is K ii K is )r s = R s – K is K ii R i (B.8)
or:
kr s = F (B.9)
where:
T –1
k = K ss – K is K ii K is (B.10)
T –1
F = R s – K is K ii R i (B.11)
When this operation is performed for both superelements A and B their matrices are reduced as follows:
KA → kA
r A → r As
RA → FA
KB → kB
r B → r Bs
RB → FB
r As = r Bs = r s (B.12)
What this means is that both superelements A and B contribute with stiffness and loads to the supernodes.
The reduced stiffness and loads may consequently be added:
( k A + k B )r s = ( F A + F B ) (B.13)
This is the equation of equilibrium for the complete model expressed in the d.o.f.s of the top level superele-
ment. Solving this equation yields rs.
When rs is known the internal displacements of A and B, rAi and rBi respectively, are found by inserting rs
in Equation (B.7) (the matrices of this equation are saved during reduction of each superelement). This
operation is termed retracking:
–1 –1
r Ai = – K Aii K Ais r s + K Aii R Ai (B.14)
–1 –1
r Bi = – K Bii K Bis r s + K Bii R Bi (B.15)
Presel SESAM
APPENDIX B-4 15-MAY-2014 Program version 7.5
The terms reduction, solving and retracking are highlighted above as these are three major phases of a
superelement analysis (sequential processes automatically executed by the analysis program Sestra).
When translating, rotating and mirroring a superelement to include it in a superelement assembly a transfor-
mation matrix is established; see the ‘INCLUDE supno PRINT-T-MATRIX’ command. The three by three
cosine matrix part of this three by four transformation matrix constitute the required transformation for the
reduced matrices. If T is the name of this three by three cosine matrix the d.o.f. vector of the reduced
superelement after the transformation is:
T
r Ts = T r s (B.16)
The stiffness and load matrices are transformed as follows (note that for the orthogonal three by three trans-
formation matrix T the inverse matrix is equal to the transformed matrix):
T
k T = T kT (B.17)
T (B.18)
FT = T F
It follows from the last equation that the loads of a superelement are rotated and mirrored along with the
superelement when this is being included in a superelement assembly. Translations of the superelement has
no effect on the loads.
This is exemplified in Section 3.3.2 where load 3 on superelement 6, see Figure 3.8, is rotated (and multi-
plied by 2) when combined into global load 2 of top level superelement 7; see note 6 of Figure 3.9.
Also note that to allow the stiffness, k, and load, F, matrices to be multiplied by a three by three transforma-
tion matrix these matrices must for all nodes include either all three translational d.o.f.s, or all three rota-
tional d.o.f.s, or all six d.o.f.s. In other words, to rotate or mirror a superelement only these three alternative
selections of super d.o.f.s for a node are allowed. Any other selection of the six d.o.f.s of a node to be super
will only allow the superelement to be translated when being included in an assembly.
For more information on the superelement technique see the Sestra documentation.