Introduction & Aims
The Manchester Benchmarks were agreed at the International Workshop 'COMPUTER SIMULATION OF RAIL VEHICLE DYNAMICS' at Manchester Metropolitan University on June 23rd and 24th 1997. The final version of these benchmarks was published in a special edition of the journal 'COMPUTER SIMULATION OF RAIL VEHICLE DYNAMICS' Vehicle System Dynamics Vol. 30 Numbers 3-4 September 1998, together with the results of simulations carried out with a number of simulation packages.
The aim of producing these benchmarks was to allow railway vehicle suspension designers and researchers investigating vehicle dynamic behaviour to assess the suitability of the various software packages that now exist for simulation of such behaviour. The intention is not to provide accurate validation of the software packages, although comparison of the results should prove useful in assessing the effect of the various techniques and approximations made. There are several existing benchmarks but these have generally been relatively complex with the aim of proving the validity of the various software routines. The results of the existing benchmarks, where published may be useful to the software authors, but they are inconclusive and they do not say very much about the ease or otherwise of carrying out typical simulations.
The benchmark models have been chosen to represent typical vehicles and track cases which provide examples of the modelling tasks often carried out. The vehicles are simplified but still contain examples of more complex elements. The level of detail given in this document is such that all participants will be able to produce a suitable model. The exact modelling methods are not specified but are left to the modellers. Flexibility is also extended to the rail wheel modelling to allow for the various methods used in this hey aspect of the simulation of a railway vehicle.
The statement of methods is a critical part of this benchmark exercise. It relates to the aim to provide information to users of the various packages and allows the participants to explain their approach to the various approximations made during the modeling.
Vehicle Models
The two benchmark vehicles and the various test cases have been carefully chosen to represent typical vehicles and track cases being modelled by designers and researchers. The vehicle models are not intended to represent actual vehicles (although the bogie vehicle is based partly on a simplified version of the ERRI benchmark #1 which is in turn based on an existing coach and bogies). Benchmark vehicle 1 is a general passenger coach with two bogies and a simple primary suspension.
Track Cases
Four track cases have been chosen to run with the benchmark vehicle models to represent typical real situations and to allow common dynamic behaviour to be seen in the simulation.
Statement of Methods
In addition to the Eigenvalue and the simulation results a statement of methods used in setting up the two vehicle models is given. This statement illustrates the steps required for creating the model and the track cases, for running the simulation and for presenting the results in the required format.
References are made to any approximations made during the modeling and the chosen numerical simulation routines should be stated.
The aim of this statement is to show the prospective modeler how to set up a model using the particular package and how to obtain the desired results and is an important part of the submission. Some indication of the time taken to carry out the various setting up stages and also the simulation itself is given.
A full description of the benchmark vehicles and track cases including all the parameters to allow modellers to carry out the required simulations is available free of charge from the RTU. The results of the benchmark exercise, including ADAMS/Rail , MEDYNA, GENSYS, NUCARS, SIMPACK and VAMPIRE are published in a supplement to Vehicle System Dynamics 1999 (vol 31).
Further Results
Some of the participants have updated their models to take into account minor mistakes made in the first exercise and to improve their results by using the new enhancement of their package. Results have been re-produced and are available for some of the participants. You can find their links below to contact them.
- VI-Rail (powered by MSC Adams): results are not available on-line anymore. Please request them from VI-Grade directly.
- SIMPACK: results are not available on-line anymore. Please request them from INTEC directly.
- Vampire: find out more about the product and their comments on the Benchmark results from AEA Technology Rail.
- Nucars: find out more about the product on TTCI's web site.
- Gensys: find out more about the product and their comments on the Benchmark results on GENSYS website.
Other packages that did not initially participated to the benchmark can be found below:
- Voco codes: series of codes (VOCOLIN, VOCODYM, VOCOLINK, etc) developed at the INRETS in France.
- Universal Mechanism Loco: MBD software developed at Bryansk State Technical University, Russia. Information available from Universal Mechanism website.
Downloads
Please read the disclaimer before downloading the files
| Full simulation cases - statement: |
 (551Kb) |
| Wheel & rail profiles data (originals): |
Rail Left Rail Right Wheel Left Wheel Right |
| Wheel & rail profiles data (ADAMS/Rail): |
Rail Left Rail Right Wheel Left Wheel Right |
Future Benchmarks
Following the success of the original Manchester Benchmark and the current interest for a more up to date comparison of existing wheel-rail contact codes, MMU is planning to organise a new exercise focusing on the wheel-rail contact predictions.