Modelling and control of PEMFC based on support vector machine

Abstract

When current is drawn from a proton exchange membrane fuel cell (PEMFC), it is critical that the reacted oxygen is replenished rapidly by the air supply system to avoid oxygen starvation and damage. This paper proposes a support vector machine (SVM) model based model predictive control (MPC) strategy to maintain a necessary level of the oxygen excess ratio during abrupt changes in the stack current. Due to its excellent performance in function regression, SVM is used to establish PEMFC model by mapping PEMFC performance (for the oxygen excess ratio in this paper) as a function of various operation conditions. Based on the SVM model, a model predictive controller is designed using Model Predictive Control Toolbox. The SVM model and the model predictive controller have been implemented in the MATLAB/SIMULINK environment. Simulation results demonstrate the effectiveness of the model and the controller. The optimum oxygen excess ratio is able to be maintained during abrupt changes in the stack current.