Simulation of fluid power systems has become a tool widely used for testing, designing and virtual prototyping. The choice of a numerical integration method for solving stiff systems of ordinary differential equations is a key factor for achieving proper stability and computational efficiency during simulation. Whereas widely used A-stable methods require small integration step sizes in order to avoid numerical oscillations when solving numerically stiff problems, the L-stable Rosenbrock method presented in this paper can take large steps. The method is implemented with an estimator of the local truncation error and a predictor of the step size. Simulations results show the good performance of the integrator in terms of both stability and efficiency.

This content is only available via PDF.
You do not currently have access to this content.