• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

电力建设 ›› 2022, Vol. 43 ›› Issue (8): 22-32.doi: 10.12204/j.issn.1000-7229.2022.08.003

• 新型电力系统背景下储能系统规划配置与运行控制·栏目主持 李相俊教授级高工、帅智康教授、颜宁副教授· • 上一篇    下一篇

基于近端策略优化算法的电化学/氢混合储能系统双层配置及运行优化

闫庆友1,2(), 史超凡1,2(), 秦光宇1,2,3(), 许传博1,2()   

  1. 1.华北电力大学经济与管理学院,北京市 102206
    2.新能源电力与低碳发展研究北京市重点实验室(华北电力大学),北京市 102206
    3.加利福尼亚大学伯克利分校,可再生能源重点实验室,美国旧金山伯克利市 94709
  • 收稿日期:2022-04-07 出版日期:2022-08-01 发布日期:2022-07-27
  • 通讯作者: 史超凡 E-mail:yanqingyou@263.net;shichaofan@ncepu.edu.cn;qinguangyu@berkeley.edu;90102479@ncepu.edu.cn
  • 作者简介:闫庆友(1963),男,博士,教授,博士生导师,主要研究方向为电力经济等,E-mail: yanqingyou@263.net
    秦光宇(1994),男,博士研究生,主要研究方向为综合能源系统运行优化,E-mail: qinguangyu@berkeley.edu
    许传博(1993),男,博士,讲师,主要研究方向为能源经济、氢储能系统配置及调度优化,E-mail: 90102479@ncepu.edu.cn
  • 基金资助:
    国家留学基金委资助项目(202006730045)

Research on Two-Layer Configuration and Operation Optimization Based on Proximal Policy Optimization for Electrochemical/Hydrogen Hybrid Energy Storage System

YAN Qingyou1,2(), SHI Chaofan1,2(), QIN Guangyu1,2,3(), XU Chuanbo1,2()   

  1. 1. School of Economics and Management, North China Electric Power University, Beijing 102206, China
    2. Beijing Key Laboratory of New Energy and Low-Carbon Development (North China Electric Power University), Beijing 102206, China
    3. University of California, Berkeley, Renewable and Appropriate Laboratory, Berkeley 94709, San Francisco, US
  • Received:2022-04-07 Online:2022-08-01 Published:2022-07-27
  • Contact: SHI Chaofan E-mail:yanqingyou@263.net;shichaofan@ncepu.edu.cn;qinguangyu@berkeley.edu;90102479@ncepu.edu.cn
  • Supported by:
    China Scholarship Council Program(202006730045)

摘要:

针对电化学储能和氢储能的互补特性,提出了一种包含电化学和氢储能的混合储能系统配置和运行的综合优化模型,并提出了智能算法进行求解。该模型基于双层决策优化问题,将混合储能系统配置及运行2个不同时间维度的问题分上下层进行综合求解,并考虑了两者间的相互影响,采用强化学习近端策略优化(proximal policy optimization,PPO)算法求解该双层优化模型。以甘肃省某地区的风光数据,通过对比应用多种传统算法求解结果,验证了所用算法在复杂环境下适应度最高且收敛速度最快。研究结果表明,应用该模型最大可降低24%的弃风、弃光率,有效提升系统综合效益。氢储能作为容量型储能配置不受地形因素限制,适用于多样的应用场景,从而为氢储能这一新型储能形态在全国的广泛配置提供了应用示范。

关键词: 风光消纳, 储能配置, 双层优化, 氢储能, 近端策略优化(PPO)算法

Abstract:

According to the complementary characteristics of electrochemical energy storage and hydrogen storage, an integrated optimization model for the configuration and operation of a hybrid energy storage system is given, including electrochemical energy storage, hydrogen storage proposed and an intelligent algorithm. The model is based on a two-layer decision optimization problem, in which two different time dimensions of the hybrid energy storage system configuration and operation are solved in upper and lower layers, and the interaction between them is considered. A reinforcement learning proximal policy optimization (PPO) algorithm is used to solve the two-layer optimization model. By comparing the results of applying various traditional algorithms to solve the scenery data of a region in Gansu Province, it is verified that the used algorithm has the highest adaptability and the fastest convergence speed in a complex environment. The results show that the application of this model can reduce the abandoning rate of wind and solar power by 24% and effectively improve the comprehensive benefit of the system, and that hydrogen storage as a capacity-based energy storage configuration is not limited by topographical factors and is suitable for diverse application scenarios, thus providing an application demonstration for the widespread deployment of hydrogen storage, a new form of energy storage, in the whole country.

Key words: wind-solar consumption, energy storage configuration, two-level optimization, hydrogen energy storage, proximal policy optimization (PPO) algorithm

中图分类号: