Model reduction in vehicle dynamics using importance analysis

Text written in October 2008 for the mor4ansys group.

After reading the review A Review of Proper Modeling Techniques, I have contacted the authors. They have many other papers on model reduction

http://www-personal.umich.edu/~stein/Publications.pdf

and I have asked for some of them. The first paper that I have read is Model reduction in vehicle dynamics using importance analysis to be published soon in Vehicle System Dynamics. My comments are below. If you are interested in the paper, please contact the authors. Alternatively the related information is in Chapter 6 of Tulga Ersal’s thesis

http://deepblue.lib.umich.edu/handle/2027.42/57603

It is interesting area of applications. One of my goals was to understand how people model vehicle dynamics and the paper gives very good introduction in this subject. So this is the level of rigid body dynamics and the model is nonlinear. After finite elements rigid body dynamics is in a way already a reduced model but it is understandable that for different types of analysis even this model should be further simplified.

The model reduction method is also interesting. The idea has resembled me POD. The difference is that the energy flowing through the connections was monitored. Then after the SVD the connections are compared between each other by importance. If the energy flow was small the connection has been considered as unimportant and removed from the model. This way the model has been simplified (reduced).

The results depend on the trajectory chosen for simulation and the authors have developed different reduced models for different trajectories. This idea seems to be reasonable. It is hard to suggest another approach for a nonlinear model.

It is interesting if this method could be generalized for an arbitrary nonlinear dynamic system. In my view it would be interesting to try to use it for an electrical circuit. Say to take an example from Yong Chen master thesis

http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.25.3742

which seems to be a typical benchmark in this respect. Well, it is unclear however if this model could be simplified similar to the vehicle model.


Comments are closed.