新能源有轨电车混合储能系统优化控制OA
Optimal Control of Hybrid Energy Storage System for New Energy Tram
针对间歇式供电新能源有轨电车搭建其储能系统损耗模型,以提高储能系统运行效率为目标,利用顺序动态规划对混合动力源功率分配进行线下优化.为解决实车优化控制问题,提出一种基于反向传播(Back Propagation,BP)神经网络的在线控制器,将线下优化结果作为训练集,完善神经网络控制器.与传统阈值法相比,动态规划策略结果能明显减少混合储能系统的损耗,同时BP神经网络通过学习功能可以较准确地拟合出动态规划的策略结果,解决实车在线控制问题,使车辆在全工况条件行驶下的能量效率明显提高.
To address the needs of a tram powered by intermittent grid-connected renewable energy sources,a loss model for the energy storage system is established.Aiming to improve the operational efficiency of the energy storage system,this study utilizes sequential dynamic programming to achieve the offline optimization of power allocation among hybrid power sources.To solve the problem of real-vehicle optimal control,this paper proposes an online controller based on BP neural network.The offline optimization results are used as a training set to improve the neural network controller.Compared with the conventional threshold method,the dynamic programming strategy can significantly reduce the power loss of the hybrid energy storage system.Meanwhile,the BP neural network can accurately fit the decision making results of the dynamic programming strategy thanks to its learning capability,thereby solving the problem of real-vehicle online control.Thus,it can significantly improve the energy efficiency of the vehicles across diverse operating conditions.
马铭;姚美如
中国铁路上海局集团有限公司杭州供电段,杭州 310009北京全路通信信号研究设计院集团有限公司,北京 100070
交通工程
有轨电车混合储能能量管理策略顺序动态规划算法BP神经网络
tramhybrid energy storageenergy management strategysequential dynamic programming algorithmBP neural network
《铁路通信信号工程技术》 2026 (8)
97-103,114,8
评论