The invention provides an energy management method for a hybrid vehicle. The method comprises the following steps: firstly, synchronously acquiring traffic road condition information and vehicle state information and generating a database; then taking the obtained information as a state space, taking a reference SOC track as an action space, combining a lower-layer energy management algorithm, and training an upper-layer deep reinforcement learning algorithm agent to be converged; and finally, extracting a neural network of the trained deep reinforcement learning agent as a battery SOC optimal reference model, and performing energy consumption distribution of the whole vehicle according to the neural network model and in combination with a lower-layer energy management algorithm. According to the method, the problem that the SOC track is difficult to determine in the energy management strategy can be solved, the effectiveness, generalization and stability of the energy management strategy of the new energy vehicle under the actual road can be improved, and therefore the method has many beneficial effects which are not achieved in the prior art.
本发明提供了一种混合动力车辆能量管理方法,首先同步获取交通路况信息和车辆状态信息并生成数据库;然后以获取的信息作为状态空间,参考SOC轨迹为动作空间,结合下层能量管理算法,将上层深度强化学习算法智能体训练至收敛;最后提取训练完毕的深度强化学习智能体的神经网络作为电池SOC最优参考模型,根据神经网络模型并结合下层能量管理算法进行整车的能耗分配。本发明不仅能够解决能量管理策略中SOC轨迹难以确定的问题,而且可以提升实际道路下的新能源车辆能量管理策略的有效性、泛化性和稳定性,从而具有了现有技术所不具备的诸多有益效果。
Hybrid vehicle energy management method
一种混合动力车辆能量管理方法
2023-08-01
Patent
Electronic Resource
Chinese
IPC: | B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion |
Intelligent hybrid electric vehicle energy management method
European Patent Office | 2022
|