Hierarchical energy management for extended-range electric vehicles considering range extender dynamic coordination

Due to the dynamic characteristics of the range extender, the energy management commands of the rangeextender electric vehicle (EREV) cannot be tracked well, which substantially degrades the energy management performance. To tackle this issue, the paper proposes a hierarchical energy management strategy (EMS) which integrates an upper level power distribution strategy and a lower level coordinated control strategy. In the upper layer, the twin delayed deep deterministic policy gradient (TD3) algorithm is introduced to decide the reference power for the range extender based on the current EREV states. Then, the lower layer receives the reference power, and together with the real-time states of the range extender, multi-scale model predictive control (MSMPC) method is developed for optimizing the control command sent to the engine control unit and the generator control unit. Thus, the power is intelligently distributed, and the command tracking performance is effectively improved, which further enhances the energy management results. Hardware-in-the-loop test results show that compared with the benchmark combination of dynamic programming (DP) and proportion-integraldifferential (PID), the proposed hierarchical EMS realizes 129.5 % battery life loss and 98.19 % fuel economy, demonstrating its effectiveness.