The invention relates to a fuel cell vehicle energy management method based on multi-target reinforcement learning, and the method comprises the steps: collecting standard working condition speed data, and building a vehicle speed prediction model through training an LSTM; constructing a whole vehicle power system model of the vehicle; inputting the current speed of the vehicle into the vehicle speed prediction model, outputting to obtain a predicted speed, and inputting the predicted speed into the whole vehicle power system model to obtain corresponding required power; based on the required power, determining an energy management strategy by adopting a deep reinforcement learning architecture which comprehensively considers the durability and economy of the fuel cell; and according to the energy management strategy, the vehicle hybrid power system controller executes and completes energy distribution. Compared with the prior art, the method has the advantages that the economy of the fuel cell automobile and the durability of the fuel cell storage battery are comprehensively considered, multi-target optimization of the energy management strategy can be realized, and meanwhile, the real-time performance and the self-adaptability of the strategy can be improved by combining the automobile speed prediction model and the whole automobile model and adopting the deep reinforcement learning algorithm.
本发明涉及一种基于多目标强化学习的燃料电池汽车能量管理方法,包括:收集标准工况速度数据,通过训练LSTM,构建得到车速预测模型;构建车辆的整车动力系统模型;将车辆当前速度输入车速预测模型,输出得到预测速度,将预测速度输入整车动力系统模型,得到对应的需求功率;基于需求功率,采用综合考虑燃料电池耐久性和经济性的深度强化学习架构,确定出能量管理策略;根据能量管理策略,由车辆混合动力系统控制器执行完成能量分配。与现有技术相比,本发明综合考虑燃料电池汽车的经济性、燃料电池蓄电池的耐久性,能实现对能量管理策略的多目标寻优,同时结合车速预测模型和整车模型,并采用深度强化学习算法,能够提高策略的实时性和自适应性。
Fuel cell automobile energy management method based on multi-target reinforcement learning
一种基于多目标强化学习的燃料电池汽车能量管理方法
2024-11-15
Patent
Electronic Resource
Chinese
Fuel cell automobile energy management method and device
European Patent Office | 2024
|Fuel cell automobile energy management method and system
European Patent Office | 2024
|Fuel cell automobile energy management method based on convex optimization
European Patent Office | 2021
|Fuel cell automobile energy management method considering dynamic response of fuel cell
European Patent Office | 2024
|Fuel cell hybrid electric vehicle energy management method based on deep reinforcement learning
European Patent Office | 2024
|