A global maximum error controller-based method for linearization point selection in trajectory piecewise-linear model order reduction

We propose a new linearization point selection method based on a global maximum error controller for the trajectory piecewise-linear (TPWL) model order reduction (MOR). This method is based on a simple fact that the simulation cost of the TPWL model is very low, and selects the state at which the responses of the current TPWL model and the full nonlinear model have the maximum difference as a new linearization point. Numerical results show that the proposed method can generate the TPWL model of smaller size and higher accuracy, and can easily be extended to generate the TPWL model for multiple training inputs.