SIMPACK Code Export

Editor 

INTEC GmbH, Argelsrieder Feld 13, D-82234 Wessling 

VOLUME 8, FIRST ISSUE 

SIMPACK Code Export 

SIMPACK Code Export is designed to 

make existing and future multi-body 

models available to non multi-body 

simulation environments. Learn more 

on the basics of SIMPACK Code Export, 

the principle structure and the 

be ne fi ts it offers in a wide range of 

application areas. 

WHY SIMPACK CODE EXPORT 

How can general multi-body models 

be used in a non multi-body environment 

by non multi-body experts? How 

can a multi-body vehicle model be 

used in a control design environment 

such as MATLAB/Simulink? How can 

a 3D-multi-body model of a construction 

machine be used in a 1D-fl uid dynamics 

software for hydraulic system 

design? 

Multi-body models of trucks are subject 

to a huge variety of possible confi gurations 

such as the number and type of 

axles. What process fl ow is required to 

generate real-time models, which can 

be incorporated in a real-time environment? 

How can complex multi-body 

models along with a highly specialised 

multi-body solver be used outside of 

SIMPACK? What is a suitable process 

for transferring simulation models 

from a multi-body division in a company 

to a control design department? 

What is a suitable link from SIMPACK 

models to in-house specifi c simulation 

environments? How can I close the 

gap between off-line simulation and 

on-line simulation? Is it possible for 

an OEM to transfer black-box vehicle 

dynamics models to a supplier without 

revealing structure and parameters 

of the car? How can engineering and 

consulting companies offer ready to 

use multi-body models to their clients 

without the need of running a multibody 

tool at the clients’ site? 

SIMPACK Code Export has been de veloped 

at INTEC to give answers and 

so lu tions to the technical and process 

relevant questions above. 

BASIC PRINCIPLE OF SIMPACK 

CODE EXPORT 

SIMPACK Code Export takes an indi- 

JULY 2004 

» SOFTWARE ................................01 

Dr. Alex Eichberger, Frank Kohlschmied, 

INTEC GmbH 

SIMPACK Code Export 

» INTEC GMBH..............................03 

Dr. Alex Eichberger, INTEC GmbH 

INTEC partners up with IST: EHD- 

Software TOWER now Available in 

SIMPACK 

» UNIVERSITY APPLICATION........04 

Thomas Thümmel, Jospeh Woyke, 

TU München 

SIMPACK FEMBS Simulation of a 

Moulded Pantograph Mechanism 

» SOFTWARE ................................06 

Johannes Gerl, INTEC GmbH 

SIMPACK Virutal Suspension 

» CUSTOMER APPLICATION.........09 

Roger Gansekow, Siemens Transportation 

System, Krefeld 

Approval of the Siemens Desiro Series 

Vehicle Dynamics for Use in Great 

Britain 

» SOFTWARE ............................... 11 

Dr. Lutz Mauer 

Gearwheels in SIMPACK 

vidual model as its input and generates 

the underlying parameterised motion 

equations 

, 

and corresponding sensor- or measurement 

equations 

, 

whereas x denotes the system states, u 

is the system input, y contains the sensor 

outputs and p stands for the free parameters 

of the exported simulation model.

» SOFTWARE 

Dr. Alex Eichberger, Frank Kohlschmied 

INTEC GmbH 

The motion equations can be generated 

as ordinary fi rst order differential 

equations or, depending on the model, 

in addition, as differential algebraic 

equations. The equations are currently 

exported in Fortran and, in the future, 

the option to export the code in standard 

C will also be available. This makes it 

possible to use the code, once exported, 

in any simulation environment. 

The structure of SIMPACK Code Export 

is shown below. The license Code Export 

enables the export of the model equations 

as source code; additional fi les are 

also generated for supplying the exported 

model with data and initial states. 

Independent from SIMPACK, the model 

can be compiled and linked with external 

code to build an executable program. 

At run-time, the model references a 

subset of the licensed (Execution) SIM- 

PACK libraries which are shipped with 

the module SIMPACK Code Export. The 

equations of motion are solved by the 

external solver in the simulation environment 

or, optionally, if a Time Domain 

Solver license is available, the exported 

SIMPACK solver can be used. 

The exported model and optional solver 

can now be used in any external simu- 

lation environment without requiring 

a SIMPACK installation. To generate 

a ready-to-use model for MATLAB/ 

Simulink, the model can be exported 

as a S-function, with the option of including 

the SIMPACK solver. Model and 

solver can now be run in multiple calculation 

cases by control design engineers. 

The often confi dential model structure 

and topology are not transparent to the 

end user; only the data made available 

by the Code Export modeller for parameterisation 

in the code generation are 

visible and therefore can be altered. 

SCOPE OF FUNCTIONALITY 

Approximately 80% of the SIMPACK 

elements can be currently used in the 

exported code, with the range being 

continually expanded. An up-to-date 

list of the functions ready to use within 

Code Export can be requested from our 

Code Export product manager Frank 

Kohlschmied (frank.kohlschmied@simp 

ack.de). 

APPLICATION AREAS 

SIMPACK Code Export has been designed 

to make existing and future 

multi-body models available to any 

2 

non SIMPACK simulation environment. 

Application areas are hardware-in-the 

loop, software-in-the loop, simulators 

and general real-time applications. It is 

also the tool designed for engineering 

and consulting companies who develop 

multi-body models for clients, who want 

to use those models independently from 

a SIMPACK installation. Finally, by means 

of SIMPACK Code export, multi-body 

models can be used as a „black box“ by 

third-parties, without having to reveal 

the model topology and structure. 

INTRODUCTORY DAY 

In summer this year, a one day free of 

charge introduction to SIMPACK Code 

Export will be offered to those interested 

in learning more about SIMPACK Code 

Export. If you are interested in attending 

the introduction course, don’t hesitate in 

contacting our Marketing Manager, Ms 

Ernie Engert (ernie.engert@simpack.de). 

More information avai lable on SIMPACK 

Code Export, www.simpack.com.

INTEC partners up with IST: 

EHD-Software TOWER now 

Available in SIMPACK 

Proposed by the Mercedes Formula 1 

engine division, IST GmbH and INTEC 

GmbH teamed up to develop a powerful 

interface for the precise simulation 

of hydro-dynamic effects in multibody 

applications. The impedance 

method and the Reynold approach 

of the IST software TOWER are now 

available within the new module 

SIMPACK Engine. INTEC GmbH, along 

with its local distributors, operates as 

a worldwide supplier for the TOWER 

software embedded into SIMPACK. 

IST GmbH is a consulting engineer 

company whose main focus is the 

numerical simulation of structural dynamics 

and the elasto-hydrodynamics 

of combustion engine components. IST 

develops software tools for the simulation 

and analysis of elasto-hydrodynamically 

coupled systems, such as: 

- piston - cylinder 

- piston pin - piston 

- piston pin - con-rod 

- crank shaft - engine block 

- crank pin - connecting rod 

- piston - piston rings - cylinder. 

IST has undertaken the task of maintenance 

and development of the 

software, as well as offering training 

in its use. The software is based upon 

research work carried out at the IMK, 

University of Kassel, under the direction 

of Professor Gunter Knoll. 

Besides software development and 

maintenance, IST offers a wide range 

of engineering services in the fi eld of 

hydrodynamic bearing calculation and 

structural analysis, with a focus on 

combustion engine components. 

SIMPACK now offers an integrated 

interface to the IST software TOWER, 

which is a software tool for the general 

elasto-hydrodynamic analysis of 

bearings. It is based on the hydrodynamic 

lubrication theory of rough 

surfaces, and uses different lubrication 

models, dependent upon tribological 

properties, fl ow factors and 

contact pressure. The interface markedly 

improves the results delivered by 

multi-body programs regarding the 

relevant frequency dependent spring 

3 

and damping characteristics of the 

lubrication fi lm, as well as the interaction 

between the deformations in 

the bearing and tribological relevant 

quantities, which include minimal gap, 

pressure distribution, centre point 

track and friction losses. 

The interface is offered at two different 

levels. The fi rst level, the so 

called impedance method, is a table 

based solution for the pre-design of 

bearings limited to rigid cylindrical 

bearing types. At the second level, the 

underlying Reynolds differential equations 

are directly solved by a fi nite 

element algorithm. This results in a 

more detailed model of the structural 

dynamics, including the tribological 

characteristics, allowing the bearing 

geometry and global bearing elasticity 

to be taken into account. 

The co-operation between IST and 

INTEC provides an EHD solution from 

one source for all SIMPACK users 

and interested parties. The solution 

comes as part of the SIMPACK Engine 

module; the functionality is, however, 

also available separately. The partnership, 

and with it the expansion into 

EHD analysis, further enhances our 

outstanding technical support, training 

and consulting service. For those 

customers who are looking to extend 

their simulation technology to the 

limits, the partnership gives them the 

opportunity to include detailed elastohydrodynamic 

modelling within their 

multi-body environment. 

More information 

www.ist-aachen.com 

www.simpack.com 

» SOFTWARE 

Dr. Alex Eichberger 

INTEC GmbH 

Deformed bearing showing the 

distribution of oil pressure 

Dr. Schönen, Director Software 

Development, iST GmbH

» UNIVERSITY APPLICATION 

Th. Thümmel, J. Woyke 

TU München 

Figure 1 

Figure 2 

Figure 3 

SIMPACK»News, July 2004 

4 

SIMPACK FEMBS Simulation of a 

Moulded Pantograph Mechanism 

A miniature pantograph mechanism 

of uniform material, including largedefl 

ective hinges , has been designed 

to assemble portable electric devices, 

e.g. mobile telephones. High adaptability 

in its functionality and usage, 

as well as a high-speed of up to 20 

cycles per second are required. Positioning 

errors are to be expected due 

to the inertia forces, which cause the 

mechanism to deform elastically. 

The design of the mechanism has 

necessitated the use of an advanced 

tool in its virtual prototyping. With 

the help of the SIMPACK interface to 

FEA programs, FEMBS, selected eigenmodes, 

calculated in ANSYS, were imported 

into SIMPACK, where a surface 

mount job was successfully modelled. 

OBJECT OF INVESTIGATION 

Small, portable electrical and electronic 

devices and computer accessories 

contain minute devices of less than 

a few square-millimetres located on 

electric circuit boards. However, most 

of the mm-sized devices are often assembled 

by huge surface mounting 

systems in the dimension of a metre. 

A new surface mount system composed 

of parallel arranged miniature 

manipulators has been proposed. One 

miniature manipulator consists of a 

moulded two degree of freedom pantograph 

mechanism, which involves 

hinges, providing large defl ections, (A, 

B, C, D, and F) and links; both of these 

components are made of the same 

material, i.e. PP (Polypropylene), as 

shown in Figure 1. 

However this type of construction 

introduces new problems regarding 

positioning and vibrations due to nonlinear 

kinematics and dynamics, caused 

by large defl ections of the hinges, 

elastic bending of the links and varying 

inertia forces. Building a model, 

that includes the pantograph’s exact 

dynamics and kinematics, could create 

the ability to compensate positioning 

errors by input motion control. 

Figure 2 illustrates the dimensions and 

simplest rigid-body linkage model of 

the pantograph mechanism, with the 

motion of the system defi ned via the 

following global transfer functions 

x = - 4 s x 

y = 5 s y 

MODELLING THE PANTOGRAPH 

MECHANISM 

An enhanced model of the pantograph 

is built up in SIMPACK as shown 

in Figure 3. The hinges A to F are 

modelled as revolute joints combined 

with non-linear torsion spring-damper 

elements. 

A static FEM analysis calculated the 

non-linear spring properties using 

elastic ANSYS-models for the hinge. 

The non-linear stress-strain-curve for 

polypropylene provided the elastic 

material properties. By fi xing one side 

of the hinge and displacing the other, 

the reaction torques were be determined. 

The FEM calculation resulted 

in the non-linear torsion spring characteristics 

of the hinges, approximated 

as a polynomial of 5 th order. 

As a fi rst estimation an approximately 

constant damping characteristic was 

available from experiments performed 

at Tokyo Institute of Technology P&I 

Lab, in the Professor Horie Group. In 

these experiments thin beams including 

one hinge capable of large defl ections 

was given an initial deformation. 

The response function was then measured 

allowing the damping ration to 

be calculated. 

In order to get the modal parameters 

of the pantograph, a modal analysis 

with a maximum of 300 Hz was 

performed using ANSYS. Fixed drives 

were assumed as boundary conditions, 

which meant the linear actuators were 

not displaced. 

This delivered four eigenfrequencies 

and corresponding eigenforms, two in 

transverse, i.e. normal to the working 

plane, and two bending eigenforms. 

In the mechanism itself it is unlikely 

that there are excitations in the transverse 

direction, and therefore the two 

bending eigenforms at 94 Hz and 205 

Hz were investigated, see fi gure 4. 

The fi rst approach investigated the two 

beams (link 2 and link 5) in ANSYS with 

the boundary conditions matching the

eigenmodes of the pantograph parts. 

The results were then exported into 

SIMPACK via the interface FEMBS. 

An elastic multi-body system model 

shown in fi gure 3 was built, which 

delivered the eigen frequencies of 106 

and 239 Hz. 

VERIFICATION BY A SPECIAL 

MOTION TASK 

Figure 5 shows two chosen trajectories 

of the tool point E when surface 

mounted. Trajectory A is a very simple 

motion, where the mechanism is used 

to pick and place an object. Trajectory 

B is an advanced version of trajectory 

A, designed to reduce the dynamic 

displacement error of output point E. 

Trajectory B provides the required motion, 

but without the corners seen in 

the motion of trajectory A. Nevertheless, 

input functions s x and s y for both 

trajectories include sinusoidal connections 

in the time domain (no jumps in 

velocity and acceleration). The rigid 

linkage model, see fi gure 2, would 

deliver the reference without dynamic 

displacement error and vibrations. 

The frequencies of these jobs are both 

6.67 Hz i.e. the cycle is 150 ms. The input 

functions can be designed in many 

ways, however, the comparison trajectories 

A and B are suffi ciently simple 

to allow the dynamic characteristics of 

the deviations of the tool point E to be 

investigated. 

Figure 6 shows the dynamic displacement 

error of the output point E from 

the reference trajectory shown in Figure 

5. 

In the future, the elastic multi-body 

system will be verifi ed experimentally. 

An iterative improvement of the multibody 

system with FEM and experimental 

methods will also follow. 

The import of mode shapes of elastic 

bodies does increase the accuracy 

obtained when calculating a mechanism’s 

motion. 

5 

More information at 

Thomas Thümmel, Joseph Woyke 

Lehrstuhl für Angewandte Mechanik, 

Technische Universität München 

www.amm.mw.tum.de 

in cooperation with: 

Tokyo Institute of Technology 

P&I Lab, Prof. Horie Group 

» UNIVERSITY APPLICATION 

Th. Thümmel, J. Woyke 

TU München 

Figure 4 

Figure 5 

Figure 6

» SOFTWARE 

Johannes Gerl 

INTEC GmbH 

Virtual and standard suspension 

Integration has been performed 

from 0.0 sec to 8.0 sec. The indicated 

CPU times were output to 

MATLAB Simulink and include initialisation 

time and time for the 

measurement of the outputs from 

the states. 

Calculations have been performed 

on Intel Pentium III, 866MHz , 

261.100 kb RAM. 

SIMPACK Virtual Suspension 

In order to improve vehicle performance 

in terms of handling, stability and ride 

comfort the usage of controlled systems 

has increased tremendously. This trend 

requires the availability of effi cient 

and easy-to-use, three dimensional 

vehicle models within control system 

simulation environments. By replacing 

the detailed kinematic structure of a 

multi-body suspension model with a 

simple look-up table based modelling 

approach, SIMPACK Virtual Suspension 

now offers an integrated process to 

reduce the model complexity essential 

for hardware- and software-in-the-loop 

simulations. 

REQUIREMENTS 

3D vehicle models that are used by 

control system experts do not have to 

fulfi l the same requirements as models 

used for the design of purely mechanical 

system parts. The purely mechanical 

systems need to allow insight into 

the effects that appear at each single 

bushing, whereby the control system 

modeller need only know the behaviour 

of the suspension as a whole, reducing 

the complexity of the required model. 

In addition, the co-operation between 

the vehicle manufacturer and the system 

supplier may restrict the release of confi 

dential suspension data. 

For the model to be used in a control 

system environment, it is necessary that 

equations do not contain differential 

algebraic equations (DAE), as the solvers 

in control environments often do not 

have the functionality to deal with DAEs, 

the case with MATLAB Simulink. One of 

the major features of the virtual suspension 

joint is that the suspension system 

does not require any DAEs. 

KINEMATICS STORED IN LOOK-UP 

TABLES 

The basic idea of SIMPACK’s virtual suspension 

joint is to read-in the position of 

a suspension system from look-up tables 

during an analysis, instead of evaluating 

the system’s kinematic behaviour during 

the calculation. The look-up tables are 

generated from a previous multi-body 

system analysis in which SIMPACK writes 

the suspension kinematics to the tables. 


6 

When using the virtual suspension joint 

to connect the wheel carrier with the 

chassis body within automotive models, 

all the connecting bodies that represent 

the wishbones can be eliminated. The 

‘dependent’ degrees of freedom of the 

wheel carrier such as the toe and camber 

angle are therefore stored as a function 

of the wheel lift and, where applicable, 

of the steering state. SIMPACK automatically 

applies splining algorithms to 

evaluate the relevant function tables 

whilst the analysis is carried out. The 

equations of motion that are required 

to represent the suspension behaviour 

are reduced and, more importantly, the 

algebraic equations are avoided. The 

virtual suspension system modelled 

in SIMPACK is represented entirely by 

ODEs. Rear suspensions therefore are 

modelled with just one degree of freedom 

(wheel lift) and front suspensions 

with two degrees of freedom (wheel lift 

and steering angle). 

ADDITIONAL SUSPENSION ELEMENTS 

Due to the fact that e.g. wishbone 

bodies are removed when using the virtual 

suspension joint, any other element 

which was referenced to these bodies 

(like shock absorbers, stabiliser bar) must 

now be connected to the wheel carrier. 

This means that, for instance, the stiffness 

and damping characteristics of a 

suspension strut must be transformed 

according to the modifi ed connection 

points. The SIMPACK Virtual Suspension 

offers specialised elements for the suspension 

strut and the stabiliser bar that 

support the user in carrying out these 

transformations, avoiding too much manual 

calculation effort. 

Additionally, a steering element designed 

for the use with the virtual suspension 

joint is available which makes 

it easy to connect any external steering 

input, provided by MATLAB Simulink for 

example, to the suspension joint. 

ELASTO-KINEMATIC SUSPENSION JOINT 

When preparing a suspension for the 

elasto-kinematic virtual suspension 

joint, additionally to the kinematics, 

the movement of the wheel carrier as 

a reaction to certain forces or torques is

saved. By adding a certain elastic movement 

to the kinematic movement of the 

suspension, the virtual suspension joint 

can then also take into consideration 

the elasto-kinematic behaviour of the 

suspension. The force/torque dependency 

on the wheel carrier movement can 

be both a linear or non-linear function. 

CREATION OF LOOK-UP TABLES 

In order to create the look-up tables 

required by the virtual suspension joint, 

an analysis with the SIMPACK module 

Virtual Testing Lab is carried out. By successive 

time domain or static equilibrium 

simulation of the complex suspension 

model, together with an appropriate 

test bench template and plot fi lter, the 

look-up tables can easily be created. Alternatively, 

the virtual suspension data 

format from MSC.ADAMS (scf-fi les) is 

supported. 

The look-up tables can also be provided 

from measurement data. For the virtual 

suspension joint, SIMPACK relies on the 

function array set data format which is 

a standard modelling element within 

SIMPACK and is in ASCII-format. 

CALCULATION PERFORMANCE 

Due to the signifi cant reduction in the 

degrees of freedom, vehicle models that 

are based on the virtual suspension joint 

show extremely high calculation performance. 

The table shows a comparison 

between a fully detailed SIMPACK 

model and the same model, but using 

look-up table joints. Kinematic and elasto-kinematic 

variants were simulated for 

a double lane change and transition of 

a road bump. Due to the elimination of 

algebraic equations (and therefore an 

ODE representation instead of DAE), a 

simple explicit Euler time integrator can 

be used for the model with virtual suspension 

joints. The simulation performance 

on a Pentium III (specifi cation see 

below) with the Euler integrator clearly 

demonstrates that these kinds of models 

are suited for real-time simulation. 

The models are available for download 

from www.simpack.com, service/support, 

model database. They require SIMPACK 

Version 8618 or higher. Please contact 

your local SIMPACK distributor to re- 

7 » SOFTWARE 

ceive a trial version. A tutorial about 

how to use the virtual suspension joint 

is also available. 

COMPATIBILITY AND AREAS OF APPLI- 

CATION 

All of the modelling elements that are 

delivered with the new package SIM- 

PACK Virtual Suspension are fully compatible 

with all other SIMPACK modules, 

including SIMPACK Code Export. The 

virtual suspension joint, as well as the 

virtual steering, virtual stabiliser bar and 

virtual suspension strut elements can be 

stored in SIMPACK databases. It is therefore 

possible to use one main model 

with combinations of the complex multibody 

suspension model or alternatively 

with the virtual suspensions, read-in as 

substructures. 

As mentioned in the article SIMPACK 

Code Export by Dr. Alex Eichberger in 

this edition of SIMPACK News, the compatibility 

with SIMPACK Code Exports 

allows the user to create SIMPACK models 

that are perfectly suited for SIL and, 

for the case that certain requirements 

regarding the creation of code are fulfi 

lled, also HIL simulation. These models 

can be exported to MATLAB Simulink 

and other control simulation tools very 

effi ciently, both from a technical and a 

cost point of view. Therefore, for the 

fi rst time, a tool is available that offers 

an integrated process to create multibody 

models for both mechanical design 

purposes of suspensions and the design 

of control systems. 

In spite of the fact that the design of 

SIMPACK Virtual Suspension was driven 

mainly by automotive application fi elds, 

the approach is not at all limited to 

automotive models. Users from other areas 

like railway engineering can benefi t 

from its features, too, for the look-up 

table joint can be used as a standard 

SIMPACK joint between arbitrary bodies. 

FURTHER STEPS 

INTEC is working on the implementation 

of optimised bumper elements to be 

used along with the Virtual Suspension 

increasing the performance, particularly 

for real-time simulations. 

Johannes Gerl 

INTEC GmbH 

If you are interested in taking 

part in our free-of-charge introduction 

day to SIMPACK Virtual 

Suspension, please contact Ms. 

Ernie Engert via telephone 0049 

8153 9288-40 or send an e-mail to 

ernie.engert@simpack.de.

» CUSTOMER APPLICATION 

Roger Gansekow 

Siemens Transportation Systems, Krefeld 

Desiro UK for South West Trains 

Simulation model of Desiro UK- 

eigenmode from ANSYS is scaled 

up 


8 

Approval of the Siemens 

Desiro Series Vehicle Dynamics 

for Use in Great Britain 

The approval of railway vehicles requires 

extensive verifi cation, both 

theoretical and practical testing. 

The Siemens Desiro series has been 

approved according to the British 

regulations. The approval of the vehicle 

dynamics was achieved through 

theoretical testing, with the help of 

multi-body simulation, and practically, 

through applied static and dynamic 

testing. The close co-operation of the 

calculation and test divisions is essential 

for the use of the test results for 

further analyses. 

APPROVAL OF SIEMENS DESIRO 

Different versions of the Siemens Desiro 

Family have been developed for 

regional service in Great Britain. The 

Desiro UK is an electrical multiple 

unit (EMU) operating at speeds up to 

160km/h in direct and alternating current 

networks. Siemens was contractually 

required deliver approved trains. 

In Great Britain a „Safety case“ has to 

be produced prior to a new train coming 

into service. The safety case must 

demonstrate that, during the design 

and manufacture of the train, the 

customer requisites and the standards 

and laws in force are observed and the 

best available technology has been 

used. The safety case must prove the 

requirements of the Rail Group Standards 

(RGS) are complied with, with the 

verifi cation provided by theoretical 

and practical testing. Results provided 

from calculations performed with validated 

calculation models are accepted 

for the approval of the vehicle. 

REQUIREMENTS FOR THE VEHICLE 

DYNAMICS 

The following points are defi ned by 

the RGS as criteria for the approval to 

international standards: 

− resistance against derailment 

− wheel/rail-force 

− bogie stability 

− gauging 

− pantograph sway 

A catalogue of calculations and tests 

is then created, from these criteria, to 

be used for the vehicles compliance 

analysis. Desiro UK has been subjected 

to the following investigations: 

− wheel unloading on twisted track 

(calculation and test) 

− bogie rotational resistance (calculation 

and test) 

− wheel/rail-forces (calculation) 

− bogie stability (calculation and test) 

− running safety and ride comfort on 

track (calculation and test) 

− sway test (calculation and test). 

SIMPACK SIMULATION MODEL 

At Siemens TS TR calculations of the 

vehicle dynamics are performed with 

SIMPACK. The Desiro UK is modelled 

as multi body system with concentrated 

masses, stiffnesses and damping. 

Wheelsets and bogie frame are stiff 

bodies, while the car body is elastic 

with additional structural degrees 

of freedom up to 40 Hz; the body is 

imported from ANSYS via the FEMBS 

interface. 

Springs, dampers and bumpstops have 

linear or non-linear characteristics. 

Dependent upon the calculation, linear 

or non-linear contact mechanics 

including theoretical wheel and rail 

profi les, are used. The track model 

has dynamic masses, stiffnesses and 

damping. 

This model is used for the calculation 

of eigenmodes, static equilibrium, linear 

and non-linear stability analyses 

and non-linear time integration. The 

specifi c operating conditions relating 

to track irregularities and wheel/rail 

profi les or conicity are also considered. 

The results generated included the 

coupling of eigenmodes, compliance 

with stability requirements, wheel/ 

rail-forces, suspension travel and accelerations 

of components and of the 

car body. 

SWAY CALCULATION FOR GAUGING 

PURPOSES 

A particularly demanding criterion 

in Great Britain is the compliance 

with the vehicle gauge. As the clearances 

along the track are very tight 

every millimetre counts. The goal is 

to achieve the largest vehicle possible 

with the smallest gap between the

vehicle and platform. The compliance 

with standard gauges in most cases 

does not provide the required space 

for the passengers. Therefore realistic 

movements of the vehicle have to be 

determined by calculation. These consist 

of quasi-static and dynamic sway 

movements and vertical suspension 

travel. 

The sway is calculated by a dedicated 

sway model. Since the required cant 

values and subsequently the suspension 

travel is quite large, particularly 

in the secondary springs (air springs), 

these need to be modelled in detail. 

The offset of the coupling points in 

the force element result in additional 

torques (r x F – terms). By means of 

a weighting factor the torque is distributed 

between the two coupling 

points. It has to be pointed out that 

due to this approach, the applied and 

opposing forces and torques are not 

necessarily the same at the two coupling 

points. This modelling requires 

specialised force elements. 

VERIFICATION OF THE SWAY CALCU- 

LATION 

The results of the calculations alone 

cannot be accepted for the vehicles’ 

approval. Because testing of all vehicle 

types and load cases would be imprac- 

9 » CUSTOMER APPLICATION 

tical, a close interaction between calculation 

and tests is applied. The most 

critical vehicle types and load cases, according 

to the calculation results, are 

selected for the tests. In a special test 

rig, the basis (the track) is canted by 

defi ned cant values. The movements 

of the vehicle are measured by optical 

means; the identifi ed movements are 

then compared with the calculation 

predictions. By virtue of the good correlation 

of the sway test results, the 

extensive calculations, which were carried 

out, could be used for the verifi cation 

of vehicle gauging. 

CONCLUSION 

The vehicles of the Desiro UK family 

have been approved in Great Britain 

from the calculations and tests performed. 

The compliance with the criteria 

for derailment safety, wheel/rail 

forces, bogie stability, gauging and 

pantograph sway has been verifi ed. In 

addition the predictions from the simulations 

performed were confi rmed 

by tests that were carried out. After 

the completion of the tests in Great 

Britain, no further optimisation of the 

vehicles was necessary. 

Sway model Spring modelling: distribution of 

the torques on the two coupling 

points 

Roger Gansekow 

Siemens Transportation Systems, Krefeld 

Sway Test in the Validation Center 

Wildenrath 

Comparison of calculation and test 

results

» SOFTWARE 


INTEC GmbH 

GUI for the defi nition of the geometrical 

gearwheel parameters 

Helical gearwheels 

Gearwheels have now become an 

integral part of modern internal combustion 

engines and gearboxes, and 

therefore the ability to incorporate 

them into a multi-body system has become 

essential for the simulation engineer. 

The module SIMPACK Engine 

offers a Force Element to accurately 

simulate Gearwheel pairs by considering 

the forces and moments generated 

in meshing gears; the resulting overall 

transfer stiffness is calculated by adding 

the individual stiffness at each 

contacting tooth. The Force Element 

models involute gear geometry and 

consider the effect of multi-tooth contact, 

as well as backlash and changes 

in direction of the gear rotation. 

This new Force Element was designed 

to be easily confi gurable whilst offering 

a high level of technical performance, 

incorporating the characteristics 

inherent in the modelling of gears. 

Along with the contact mechanics 

within the meshing gears, the effects 

of changes to the relative position of 

the gear axes are an important feature 

within this element. The Gearwheel 

can be used to model spur tooth and 

helical involute gearing for multitooth 

contact, and to include the 

effects of backlash and gear rotation 

direction changes. 

GRAPHICAL PRIMITIVES 

The Gearwheel geometry is created 

from the respective graphical Primitive 

and is defi ned via the standard Gearwheel 

parameters; number of teeth, 

normal module, pressure angle, helix 

angle, addendum and dedendum 

coeffi cients, along with addendum 

modifi cation coeffi cient and backlash. 

The amount of detail shown in the 

visualisation of the Gearwheels can 

be modifi ed by defi ning the number 

of points used to describe the tooth 

fl ank, tip and root. However, changes 

to the graphical representation do not 

affect the accuracy of the force calculation. 

INPUT PARAMETERS 

The Gearwheel Force Element uses 

10 

Gearwheels in SIMPACK 


various different parameters to describe 

the geometrical relationship of 

the wheels to each other, as well as the 

physical and material properties of the 

Gearwheels. The parameters describing 

the geometrical form include: 

- A reference angle for the tooth position. 

- The face width of the meshing teeth. 

The parameters describing the physical 

and material properties include: 

- The mean Young’s modulus of both 

wheels. 

- The gearwheel form factor. This is 

used for the calculation of the weakening 

of the tooth, when compared 

to a gearwheel of complete crosssection. 

- The normal damping coeffi cient of 

meshing. 

- The coeffi cient of friction used for 

the calculation of the tangential 

forces induced from the meshing 

gears. 

- The parabolic function used to 

describe the normal stiffness, from 

the minimum to the maximum value, 

as the contact point progresses over 

the tooth. 

The parameter tip relief is used to 

smooth the running of the Gearwheels. 

The tip relief reduces the large 

changes in normal stiffness, which can 

occur for multi-tooth contact when 

gear teeth pairs commence and cease 

meshing. 

INPUT VARIABLES 

The Gearwheel Force Element is connected 

to two Body Fixed Markers, 

which are located on the rotational 

axis of the Gearwheels, and should 

be located at the centre point of the 

Gearwheel’s width. The parameters 

automatically provided by SIMPACK 

are as follows: 

- The angle of the reference Marker 

of both gear Bodies. 

- The angular velocity of both gear 

Bodies.

- The relative axial displacement of 

the gear Bodies to each other. 

- The relative radial displacement of 

the Body Fixed Reference Frames. 

Due to radial or axial displacement of 

the wheels, it is possible for the wheels 

to come out of meshing. If the nominal 

tooth backlash reduces, due to a 

reduction in the axial displacement, it 

is possible for both tooth fl anks of the 

meshing teeth to come into contact. 

As an initial condition, the Gearwheel 

axes should be parallel. If the gear angle 

for all gear pairs are set to zero in 

the initial position, the simulation will 

begin without forces applied as the 

gear pairs are located in the centre of 

the backlash position. 

The force law for the calculation of the 

nominal force is calculated linearly or 

non-linearly as a function of the penetration 

of the involute tooth profi le. 

The tangential force, during meshing, 

is calculated from a Coulombic force 

law, with the friction coeffi cient given 

by the user. 

For the calculation of the resulting 

meshing stiffness, the contact ratio of 

the involute gearing is considered. The 

contact ratio is dependent on the current 

Gearwheel axial displacement. Up 

to fi ve tooth pairs can be in contact at 

any one time. 

OUTPUT VALUES 

The output values for the External 

Gear Force Element are the effective 

forces and moments applied at each 

wheel. The force and moments can 

also be output in Reference Frames, 

which do not rotate with the Gearwheels. 

In addition to other internal 

measurements which include Gearwheel 

angle, stiffness, force and position, 

it is possible to return the current 

meshing condition concerning the 

type of fl ank contact. 

PARTICULAR FEATURES OF THE FORCE 

ELEMENT 

The analytical calculation of the 

Gearwheel contact geometry reduces 

the need to search for the individual 

contact points, and therefore leads to 

11 

very effi cient calculation performance. 

At the start of the time integration, 

the meshing tooth pairs are all located 

at the centre of the tooth backlash 

position, ensuring no transient effects 

are present. The separate input 

of the initial tooth angle allows, for 

each Gearwheel level, the required 

meshing. For chains of Gearwheels, 

the phase relationship of the dependent 

Gearwheels is calculated from the 

initial tooth angle input of one of the 

Gearwheels. The tooth force pulsation, 

resulting from multi-tooth contact 

and the tooth stiffness function, can 

be reset to zero from two Force parameters. 

FUTURE DEVELOPMENTS 

An important aspect considered when 

the Gearwheel Force Element was 

implemented was to easily allow the 

future realisation of new features. 

Further planned developments to the 

Gearwheel are as follows: 

- Internal Gearwheel functionality. 

- Implementation of straight and helical 

bevel gears. 

- Calculation of the nominal tooth 

stiffness, dependent upon the 

number of teeth, the x-gear pair 

shift factors and the shape factor. 

- Implementation of non axially 

aligned gears, for use on fl exible 

shafts. 

SUMMARY 

The SIMPACK Gearwheel is an easily 

confi gurable Force Element, and 

allows the data to be entered easily 

and effi ciently. The calculation of the 

tooth forces considers all of the geometric 

features of the involute tooth. 

The optimised calculation algorithm 

used for the tooth forces means a 

simulation containing gearboxes 

with a large number of gears can be 

performed, whereby the phase difference 

of the meshing tooth pairs is 

accurately modelled. 

» SOFTWARE 


INTEC GmbH 

Spur gearwheels 

Automatically created primitive for 

Gear element

» NEWS 

SIMPACK Chain module 

Gunter Schupp with Prof. Arnold in 

the background 

LATEST NEWS 

NEW SIMPACK FEATURE MARKER 

RELATIVE TO MARKER 

SIMPACK Version 8614 introduced the 

feature, which allowed the position 

and orientation of one marker to be 

defi ned relative to another marker. 

The relationship between the two 

markers can be defi ned by simply 

selecting, in the ‘MBS Defi ne Marker’ 

window, one marker (B) as the reference 

marker for another marker (A). 

The fi nal position of the marker A relative 

to its Body Fixed Reference Frame 

(BFRF) is calculated in a pre-processing 

step and does not change during the 

simulation. When the fi nal position 

and orientation of marker A on a body 

is calculated, fi rst the position and 

orientation of the reference marker 

B relative to its BFRF is calculated; 

this transformation is then applied to 

marker A in addition to the relative 

transformation of marker A that was 

entered by the user. The user entered 

relative transformation, i.e. Built-in Position 

and Orientation, is given in the 

co-ordinates of the reference marker B, 

whereby the two markers can belong 

to different bodies. The orientation 

and position of marker A can be fully 

parameterised. This feature is only 

available for Body Fixed Markers. 

NEW SIMPACK FEATURE: CENTRE OF 

MASS RELATIVE TO MARKER 

Since SIMPACK Version 8614 the centre 

of mass of a body can be defi ned at 

the position of, or relative to another 

marker. The marker that is referenced 

for the centre of mass must be a Body 

Fixed Marker and must be on the body 

to which the centre of mass is defi ned. 

NEW HOW-TO-MODEL SIMPACK 

WHEEL/RAIL AT WWW.SIMPACK.COM 

A new How-To-Model guide is available 

from the SIMPACK website. The 

guide describes how to model and 

simulate a quasi-static curving analysis 

in SIMPACK Wheel/Rail for both standard 

and Virtual Testing Lab analyses. 

In addition, a How-To-Model guide 

for the calculation of Nominal Forces 

for Wheel/Rail models is also in the 

pipeline; the guide explains the math- 


12 

ematical fundamentals and goes into 

detail about the modelling approaches 

and possible sources of error. 

SIMPACK CHAIN MODULE 

As a development within SIMPACK 

Engine, INTEC is currently implementing 

a new chain mechanism simulation 

module. The prototype of this software 

has demonstrated extraordinary 

calculation performance and stability, 

benefi ting strongly from SIMPACK‘s 

relative co-ordinate recursive algorithm. 

An automatic chain model 

editor has been implemented to allow 

the effi cient creation of chain models 

based on geometric inputs from the 

user. The release date for the chain 

module is late summer 04. 

INTEC CONGRATULATES GUNTER 

SCHUPP ON COMPLETING HIS 

DOCTORAL EXAMINATION 

On the 12 th of March this year Gunter 

Schupp celebrated his doctoral ceremony 

at the institute B for Mechanics 

at the Technical University Stuttgart. A 

large contingent from his department 

at the DLR, as well as the MBS hardcore 

from INTEC travelled to Stuttgart 

for his doctoral presentation, to give 

Gunter their support. The topic was 

spot on for the limit cycle enthusiasts 

from the Wheel/Rail fi eld: 

“Numerical bifurcation analysis of nonlinear 

mechanical systems for use with 

railway vehicles.” 

For the long standing SIMPACK developer 

Gunter Schupp, it goes without 

saying that the Hopf and saddle graphs 

presented were from calculations performed 

within SIMPACK, whereby, for 

the analyses, the software PATH was 

coupled with SIMPACK. The examiners 

Professor Eberhart, Arnold and Schiehlen, 

after a worrying wait for Gunter, 

announced that he had passed with 

fl ying colours. As you can imagine, not 

only Professor True was overjoyed....

Hello, 

my name is Mario Baumann. In 

January 2001 I joined the SIMPACK 

development team. While studying 

Mathematics at the FH (University of 

Applied Sciences) Regensburg, I was 

introduced to MBS simulation. I then 

went on to use the MBS tool SIMPACK 

for my diploma thesis „optimisation 

algorithm for parameter identifi cation 

in a MBS fi eld“ at the Institute 

of Aeroelasticity, DLR Research Centre 

Oberpfaffenhofen. My programming 

tasks include: 

- Parameterisation (substitution variables, 

expressions) 

- Contact functionality (geometric contact, 

fl exible moved marker) 

- Virtual Testing Module 

- Inter Process Communication (cosim, 

simat) 

- LinuX port 

I spend my free time with my girlfriend 

Adele and playing table-tennis 

in my hometown of Boebrach. 

Hi there, 

my name is Gerhard Hippmann and I 

was born in 1968. After growing up in 

the charming Allgaeu region in Bavaria, 

I studied mechanical engineering 

at the Technical University of Munich. 

I gained my fi rst practical experience 

as a project engineer at Intec in 1997. 

One and a half years later I decided to 

move to the Vehicle System Dynamics 

Group of the German Aerospace 

Centre DLR to study for my doctorate 

in the fi eld of contact mechanics. In 

June 2003 I returned to Intec and am 

now working on the development and 

implementation of MBS methods. Visit 

my homepage www.hippie.org to fi nd 

out more about my private interests. 

13 » STAFF 

Hello, 

my name is Gerry Hofmann. I work 

on projects and support in Intec’s automotive 

department. After working 

for Intec as a freelancer for a couple of 

months, I fi nally joined the company 

in September 2003. 

I discovered my interest in simulating 

reality during my studies of „Mikro- 

und Feinwerktechnik“ (Mechatronics) 

at the Munich FH (University of Applied 

Sciences), and so consequently 

chose multi-body simulation for my diploma 

thesis and later in my work. My 

strong personal interest in automotive 

technology fi ts in perfectly with my 

day-to-day work. 

I spend a lot of my free time with classic 

black/white photography and as a 

drummer in a rock band. 

Mario Baumann 

Gerhard Hippmann 

Gerhard Hofmann

» MARKET REVIEW 

Johannes Gerl 

INTEC GmbH 

Johannes Gerl, Director Sales and 

Marketing 

MARKET REVIEW 

The current development of controlled 

vehicle dynamics systems, 

even for the high-end user, is to 

use different models, created in different 

programs, for each of the 

separate modelling tasks. Why is that? 

A suspension system incorporates a diverse 

band of relevant vibration effects, 

which cover a wide frequency range. 

Due to calculation performance and 

also the degree of specialisation of the 

engineers involved, it is not possible to 

use a single model for a NVH analysis, 

with a frequency content of 100 Hz or 

more, in a SIL project where (for instance 

in MATLAB Simulink) only handling 

dynamics are required. This model 

would also not be appropriate in a HIL 


14 

analysis that must run in real-time. 

However with SIMPACK, it is possible 

to create the models in the same program, 

whereby the HIL and SIL simulation 

models can be created as a simplifi 

ed version of the complex multi-body 

models, and thereby meeting the 

specifi ed demands of the respective 

simulation environments. With the 

new product SIMPACK Virtual Suspension, 

the user can create look-up table 

joints that enable a simplifi ed and 

high performance representation of 

more complex models, which can then 

be exported to Simulink, via SIMPACK 

Code Export, and thereby reducing 

the overall licensing costs. 

NEW UNIVERSITY AND RESEARCH LICENCES SINCE JANUARY 2004 

FH Rosenheim, Germany 

FH Heilbronn, Germany 

Fraunhofer Gesellschaft, Institut für Techno- und Wirtschaftsmathematik, 

Kaiserslautern, Germany 

Fraunhofer Institut für Betriebsfestigkeit, Darmstadt, Germany 

Nanjing University of Aeronautics and Astronautics (NUAA), P.R. China 

Neimenggu Technical University, Transportation Dept., China 

PLA University, Transportation Dept., Tianjin P.R. China 

TU Berlin, Institut für Land- und Seeverkehr, Germany 

TU Berlin, Institut für Konstruktion, Mikro- und Medizintechnik, Germany 

TU Graz, Austria 

TU Hanoi, Vietnam 

Universität Hohenheim, Institut für Agrartechnik, Hohenheim, Germany 

Uni Siegen, Germany 

NEW COMMERCIAL LICENCES SINCE JANUARY 2004 

AEA Technology Rail BV, Utrecht, The Netherlands 

AMST Systemtechnik Ges.m.b.H., Austria 

Baotou First Mechanism Works, P.R. China 

Das Kompetenzzentrum - Das Virtuelle Fahrzeug Forschungsges. mbH, Graz, Austria 

DaimlerChrysler Research, Stuttgart, Germany 

Mercedes Motorsport, Stuttgart, Germany 

MeSH Engineering, Dettenhausen, Germany 

R. Bosch GmbH, Stuttgart, Germany 

Voith Turbo, Crailsheim, Germany

SIMPACK TRAINING COURSES 

July 2004 12.07. – 13.07.2004 SIMPACK BASIC Training 

15 

14.07. - 15.07.2004 SIMPACK Wheel/Rail Training 

September 2004 13.09. – 14.09.2004 SIMPACK BASIC Training 

15.09. – 16.09.2004 SIMPACK Wheel/Rail Training 

15.09. – 15.09.2004 SIMPACK Automotive + 

17.09. – 17.09.2004 SIMPACK Contact Mechanics Training 

October 2004 11.10. – 12.10.2004 SIMPACK BASIC Training 

13.10. – 14.10.2004 SIMPACK FEMBS Training 

15.10. – 15.10.2004 SIMULINK Interface Training 

November 2004 15.11. – 16.11.2004 SIMPACK BASIC Training 

SIMPACK ACADEMY 

17.11. – 18.11.2004 SIMPACK Wheel/Rail Training 

17.11. – 17.11.2004 SIMPACK Automotive + 

05.10. - 08.10.2004 SIMPACK Academy: Low / High Frequency Tire Models 

27.01. - 28.01.2005 SIMPACK Academy: FE-MBS-Interfacing 

SIMPACK AT CONFERENCES AND EXHIBITIONS 

23.08. - 27.08.2004 AVEC 04, Arnhem, The Netherlands 

30.08. - 31.08.2004 3rd International Colloquium on „Tyre Models for Vehicle 

Dynamics Analysis“, TU Wien 

29.09. - 30.09.2004 VDI-Tagung Berechnung und Simulation im Fahrzeugbau, 

Würzburg 

04.10. - 06.10.2004 Aachen Colloquium, Automobile and Engine Technology, 

Aachen 

27.10. - 28.10.2004 NAFEMS Seminar: Analysis of Multi-Body Systems Using 

FEM and MBS 

09.11. - 10.11.2004 SIMPACK User Meeting 

at the Wartburg in Eisenach 

» TRAININGS AND CONFERENCES 

We would be glad to welcome you 

to our SIMPACK training courses 

or to a SIMPACK Academy course. 

To register for one of this courses 

or to discuss your training requirements 

please contact 

Ernie Sabine Engert by e-mail: 

ernie.engert@SIMPACK.de 

or by telephone: 

0049 8153 9288-40 

or by fax: 

0049 8153 9288-11 

Ms Engert looks forward to hearing 

from you. 

The Wartburg is where the SIMPACK 

User Meeting 2004 will be held, on 

the 09th and 10th November

CONTACT 

» CHINA 

ESP 

Intelli-Center B Room 1709 

Zhongguancun East Road No.18 

Haidian District, Beijing 100080 

P.R. China 

Tel.: +86 - 10 - 826 015 56 

www.bj-esp.com 

» KOREA 

ADVANCED TECHNOLOGY ENGINEER- 

ING SERVICE LTD (ATES) 

8F Fine Bldg., 673-5 Deungchon-Dong, 

Kangseo-Gu, Seoul, KOREA 

Tel.: +82 - 2 - 2657 - 3544 

www.ates.co.kr 

PLEASE SIGN ME UP FOR THE FREE DELIVERY OF THE SIMPACK»NEWS 

(copy, fi ll in, fax to +49-8153-9288-11) 

Name 

First Name 

Firm 

Street 

Zip Code 

Location 

Country 

Telephone 

Telefax 

E-Mail 

I WOULD BE INTERESTED IN THE FOLLOWING APPLICATIONS 

Automotive 

Wheel/Rail 

Aerospace 

General Machinery 

other: 

» GREAT BRITAIN 

INTEC Dynamics Ltd. 

Cambridge Road Industrial Estate 

Whetstone, Leicester LE8 6LH, UK 

Tel.: +44 116 275 1313 

Fax: +44 116 275 1333 

info@intecdynamics.co.uk 


» UNITED STATES 

Altair Engeneering 

1820 E Big Beaver 

Troy, MI 48083-2031 

Tel.: +1 - 248 614 - 2400 

Fax: +1 - 248 614 - 2411 

www.altair.com 

» JAPAN 

Altair Engineering, Ltd 

Tact No. 4, Bldg. 9F 

2 - 32 - 12 Minami Ikebukuro 

Toshima-Ku, Tokyo 171 - 0022, Japan 

Tel.: +81 - 3 - 5396 - 1341 

Fax: +81 - 3 - 5396 - 1851 

www.altairjp.co.jp 

» WORLDWIDE 

INTEC GmbH (Headquarter) 

Argelsrieder Feld 13 

82234 Wessling 

Tel.: +49 - 8153 - 9288 - 0 

Fax: +49 - 8153 - 9288 - 11 


SIMPACK»NEWS 

1996 – 2004 

Circulation: 3200 

SIMPACK Version 5, 6, 7, 8, 8.5, 8.6, 8.7 

FEMBS, Loads, ProSIM, CatSIM, 

IdeSIM, MATSIM, SIMAT 

(INTEC GmbH) 

REGISTERED TRADEMARKS 

ABAQUS: 

Abaqus, Inc 

ANSYS: 

Swanson Analysis Systems, Inc. 

CATIA: 

Dassault Systems 

MATLAB: 

The MathWorks, Inc. 

MSC.MARC: 

MSC.Software Corporation 

MSC.NASTRAN: 

MSC.Software Corporation 

Pro/ENGINEER: 

Parametric Technology Corporation

SIMPACK Code Export

Create successful ePaper yourself

Delete template?

Save as template?