Evaluating Rosenbrock Methods for Usage in Real-Time Simulation of Multibody Models

The upward trend in the usage of sensors and intelligent systems in mechanical systems is requiring the continuing growth in hardware-in-the-loop (HIL) testing in theautomotive industry. This trend in turn is requiring that more sophisticated multibody system models (MBS) be incorporated into hardware-in-the-loop design and testing phases. Running sophisticated multibody models consisting of a large number of bodies and a variety of compliant elements is a challenging on a HIL system. The models must be detailed enough to give realistic results yet run fast enough to meet strict turnaround times of the HIL system. This paper will explore the use of a Rosenbrock method in a real-time environment and compare its performance and accuracy to other integrators in particular to a Backward Difference Formula (BDF) that is restricted in order and number of iterations.