首页|基于RL-GSK算法的光储联合电站运行策略研究

基于RL-GSK算法的光储联合电站运行策略研究

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
光伏电站配置储能可以促进光伏的最大消纳,减少"弃光".此外,针对光伏出力和负荷情况合理分配储能充放电功率不仅能满足系统运行要求,还能利用峰谷套利提高光伏并网收益.该文以光储联合电站为研究对象,以电力市场环境下光储联合电站利润最大化为目标,充分考虑光伏系统与储能系统运行时的各项约束条件,建立了光储联合电站优化运行模型.同时,针对原始知识获取共享算法收敛性能较弱的问题,采用强化学习构建一套系统的个体更新选择机制,提出了强化学习知识获取共享算法,并用来求解光储联合电站的优化运行策略.通过对南方某地区实际数据进行仿真研究,分析了联合电站的优化策略和经济效益模式,并验证了所提方法的有效性.
A Study on the Operation Strategy of the Photovoltaic Storage Joint Power Plant Based on the RL-GSK Algorithm
The energy storage configuration in the photovoltaic(PV)power plant can promote the maximum consumption of PV and reduce the abandonment of solar energy.In addition,reasonable distribution of energy storage charge and discharge power according to the PV power and load conditions can not only meet the system operating requirements,but can also use peak and valley arbitrage to improve the revenue.Taking the profit maximization of the joint power plant as the goal in in the context of the power market,fully considering different operation constraints,this paper establishes an optimization operation model for the photovoltaic storage joint power plant.Furthermore,to address the weak convergence of the original gaining-sharing knowledge algorithm,the reinforcement learning is adopted to form a systematic individual updating and selection mechanism to design an enhanced method for solving the operation model.Through simulation and analysis of the actual data in a certain region in South China,we analyze the optimization strategy and economic benefit mode of the joint power plant,and verify the effectiveness of the proposed method.

photovoltaic storage joint power plantoptimal operation modelreinforcement learninggaining-sharing knowledge algorithmoperation strategy

杨莎、熊国江

展开 >

贵州大学电气工程学院,贵州贵阳 550025

贵州大学勘察设计研究院有限责任公司,贵州贵阳 550025

光储联合电站 优化运行模型 强化学习 知识获取共享算法 运行策略

国家自然科学基金贵州省科技计划贵州大学勘察设计研究院有限责任公司创新基金

52167007黔科合基础-ZK[2022]一般121贵大勘察[2022]03号

2024

电网与清洁能源
西北电网有限公司　西安理工大学水电土木建筑研究设计院

电网与清洁能源

CSTPCD北大核心
影响因子：1.122
ISSN：1674-3814
年,卷(期)：2024.40(7)