Real-time energy management strategy for fuel cell/battery vehicle based on speed prediction DP solver model predictive control

Fuel cell hybrid electric vehicles (FCHEVs) are widely considered to be an ideal substitute for the traditional vehicles. However, their high costs of operation make them uncompetitive in the current automotive market. To maximize the economy of fuel cell/battery-based hybrid electric vehicles, herein, integrating with speed prediction model, a dynamic programming solver model predictive control (DPS-MPC) strategy is applied for the hybrid vehicle energy management. The main principle of the proposed DPS-MPC is to obtain the manipulated variable that minimizes the vehicle operating cost by using the DP solver according to the information of state variables and disturbance variable. Assisted by the predictive vehicle speed, the future demand power of the vehicle is acquired by vehicle dynamic model. The demand power imports into the energy management system response prediction model for improving the control performance by considering more exact disturbance. Additionally, a comparative study is conducted to verify the improvement of economic performance by proposed strategy. The results indicate that DPS-MPC strategy can respectively reduce the operating cost by 21 % in average versus a normal MPC strategy. Moreover, the average computing time per step is 9.27 ms, less than the sampling time 1 s. In order to verify the real-time performance of the DPS-MPC strategy, a hardware in the loop test is conducted.