Disturbance prediction-based enhanced stochastic model predictive control for hydrogen supply and circulating of vehicular fuel cells

Abstract Hydrogen supply and circulating in vehicular fuel cells is crucial for their output capability and lifetime. In this article, an enhanced multiple-input multiple-output (MIMO) model predictive control (MPC) scheme is proposed for hydrogen regulation based on vehicle speed-induced fuel cell current disturbance stochastic prediction. The Markov exponential smoothing law is first developed for the vehicle speed prediction. The forecasted fuel cell power demand is obtained through vehicle dynamics model and rule-based energy management to release the predictive stack current regarding as the disturbance of hydrogen control system. The discrete predicted current sequence is with stochastic features and typed into the predictive model of MPC which is on longer the length of control horizon. Two case studies are presented to discuss the influence of different speed sampling times on the hydrogen regulation result under the proposed enhanced MPC. The enhanced MPC has a better performance than the traditional MPC, and the control RMSE of which can be reduced by 44.09% in case 1 and 69.78% in case 2 during automotive driving cycles. A dSPACE MicroAutoBox hardware in loop (HIL) experiment was conducted and the results well matched with the simulation which has verified the real-time performance of the enhanced MPC scheme.