A Thermal Management Strategy for Proton Exchange Membrane Fuel Cell via Nonlinear Model Predictive Control

This paper proposes a thermal management strategy for proton exchange membrane fuel cell based on nonlinear model predictive control (NMPC). To handle the highly coupled and nonlinear nature of the thermal management system, we adopt NARMAX model and introduce BP neural network to fit the nonlinear prediction model. Then in rolling optimization stage, an optimization problem consisting of two inputs and single output is established, for which an improved particle swarm optimization method is employed in combination with a mutation mechanism and an improved parameter update mechanism. Simulation results validate the effectiveness of the proposed NMPC-based thermal management strategy.