Modelling, Simulation and Analysis of Multiphysical Systems

Overall Modeling Challenges

A problem is encountered in the simulation of large mechanical systems, like test benches or complete vehicles: a useful model must not only deal with the interactions of very many movable parts, but also the behaviors of complex force elements or active feedback control elements.

In practice, the results are limited by time and hardware resources and, mainly, from parameterization difficulties. A major challenge is the reduction of physically "correct" behavior of the individual component down to the "fundamental" behavior, in a controlled manner without diminishing the predictive ability of the entire system.

The framework of the widely familiar multi-body system modeling (MBS modeling) neglects, for example, the elastic bending ability of many components and replaces this in the model with rigid, inert masses, which interact across ideal joints and force elements. The resulting errors must necessarily be compensated for by a clever choice of model parameters and non-linear attributes for friction, flexible bearings, buffers, etc. Using the modal MBS-FEM coupling method, key components can be considered with elastic degrees of freedom in the overall model.

Complex drive and control elements usually also have pronounced proportions of physical domains beyond the mechanical (hydraulic, pneumatic, electronic...). Special programs for modular or one dimensional system simulation enable not only a highly flexible and efficient implementation for the level of detail for the sub-model, but also support integration in the overall mechatronic model (co-simulation).

Quantitative modeling requires a comparison with the appropriate direct or indirect measurements at the system and component levels (parameterization of the model). Generally, the predictive ability of the simulation depends not just on the details of the modeling, but equally on the quality of the input data (measurement method, sampling rate, filtering, etc.).