A real-time multi-objective predictive energy management for range-extended electric vehicles considering driving scenarios

Energy management strategy (EMS) plays a crucial role in optimizing the power distribution among multiple energy sources in Range Extended Electric Vehicles (REEVs). To achieve an optimal balance among overall efficiency, battery throughput, and real-time performance, the study proposes a real-time multi-objective Predictive Energy Management Strategy (PEMS) for REEV, adapted to diverse driving scenarios. The framework consists of two components: offline weight updates and an online controller. Firstly, to balance the accuracy and efficiency in the prediction module, a Multi-Network Prediction model (MNP) is proposed, which accounts for variations in driving conditions. Secondly, to optimize both overall efficiency and battery throughput in real-time, an online controller based on a PEMS-oriented Model Predictive Control (MPC) approach integrated with the MNP is developed. To ensure algorithmic real-time performance, a variable parameterization-based MPC is design to solve NLP problems efficiently. Finally, the effectiveness, real-time performance, and robustness of the proposed strategy are verified through simulation and comparison with existing methods. The results demonstrate that PEMS achieves 97.47% of the globally optimal fuel economy, reduces the battery life loss rate significantly by 42.0%, and improves the real-time performance of the algorithm with a 95.49% reduction in computation time.