A Calculation-Efficient Predictive Energy Management Strategy for Fuel Cell Hybrid Electric Bus based on Improved Optimization Solver

This paper proposes a calculation-efficient predictive energy management strategy (EMS) for a fuel cell hybrid electric bus (FCHEB) based on an improved optimization solver. A deep neural network (DNN) based speed predictor is trained using real bus driving dataset and then adopted to predict the bus speed. After that, the direct collocation method is applied to transform the energy management problem to a nonlinear optimization one. Moreover, a sequential quadratic programming (SQP) based high-precision solver combined with a low-precision dynamic programming (DP) solver is used to optimize the power allocation. Simulation results indicate that the proposed EMS can improve both the calculation accuracy and efficiency, decreasing the equivalent hydrogen consumption by 0.92% compared to the predictive EMS with a traditional DP solver.