[关键词]
[摘要]
为了支撑轨道交通系统绿色环保的可持续发展需求,从系统设计的角度出发,在光伏发电及储能方面展开研究,从新能源技术的角度为城轨系统整体能效的提升提供有力的支撑。研究城轨交通系统用分布式光伏-储能供电系统方案,研究多系统、高可靠度的供电模式和源储荷多能源耦合下的能量管理策略。面向光伏储能接入单个牵引变电所,提出一种基于深度强化学习的能量管理与优化方法。该方法使用深度Q网络对列车负荷、光伏单元和储能单元功率输出等状态信息进行训练学习,通过训练好的代理对直流牵引网进行能量管理,解决光伏发电系统难以适应城轨列车启停频繁、工况多变,以及多能源系统引入后带来的供电稳定性、容量配置和能量管理等问题,有效提升城市轨道交通系统的绿色能源利用率、降低变电所输出能耗。
[Key word]
[Abstract]
In order to support the sustainable development for green and environmental protection of rail transit system, this paper is started from the perspective of system design, studied from photovoltaic power generation, hydrogen fuel cell and hybrid energy storage, and provides strong support for the improvement of overall energy efficiency of urban rail transit system from the perspective of new energy technology. Distributed photovoltaic - energy storage power supply system for urban rail transit system is researched. Multi - system, high reliability power supply mode and coordinated control strategy is developed. The energy management strategy under the coupling of source storage and multi energy is studied. For photovoltaic energy storage connected to a single traction substation, an energy management and optimization method based on deep reinforcement learning is proposed. This method uses deep Q-network to train and learn the state information such as train load, power output of photovoltaic unit and energy storage unit, and manages the energy of DC traction network through trained agents. The photovoltaic power generation system is difficult to adapt to the frequent start and stop of urban rail trains, the changeable working conditions, and the problems of power supply stability, capacity configuration and energy management caused by the introduction of multi-energy system, so as to effectively improve the green energy utilization rate of urban rail transit system and reduce the output energy consumption of substation.
[中图分类号]
U231
[基金项目]
国家重点研发计划《超大城市轨道交通系统高效运输与安全服务关键技术》项目(2020YFB1600700)