Synergy of (L1, H2, H∞) Norms for Nonlinear Optimal PEMFC Dynamic MIMO Model Reduction using a Novel ANN-BBO Approach

In this work, we present an optimal reduced order nonlinear dynamic model for proton exchange membrane fuel cell (PEMFC) using the minimization of error between original and reduced order models via (L1, H2, H∞) norms synergy optimized with biogeography-based optimization (BBO) Algorithm. The data necessary to form the autoregressive exogenous (ARX) artificial neural network (ANN) model are generated by the simulation of the dynamic model of the nonlinear PEMFC500w differential equations to extract space state matrices values. This approach is compared with Balanced Truncation (BT) model reduction method and illustrated through simulation results.