基于机器学习的新型电力系统下水电运行规则提取研究  

Research on extraction of hydroelectric operation rules in novel electric power systems based on machine learning

作　　者：周开喜 王靖 赵羽西 闫孟婷 黄炜斌[2] ZHOU Kaixi;WANG Jing;ZHAO Yuxi;YAN Mengting;HUANG Weibin(Southwest Branch,State Grid Corporation of China,Chengdu 610041,Sichuan,China;College of Water Resource and Hydropower,Sichuan University,Chengdu 610065,Jiangsu,China)

机构地区：[1]国家电网有限公司西南分部,四川成都610041 [2]四川大学水利水电学院,四川成都610065

出　　处：《水利水电技术(中英文)》2024年第S2期523-530,共8页Water Resources and Hydropower Engineering

基　　金：国家电网有限公司西南分部科技项目资助(SGSW0000DKJS2310037)

摘　　要：随着新能源装机稳步提高,新型电力系统不确定属性进一步加强,亟需水电等调节性电源平抑风光等不确定能源的季节性波动、随机性波动等。本研究旨在探讨新型电力系统下水电站最优运行规则的提取,以提高清洁电力生产效率和电网稳定性。首先,研究构建了以多能系统发电量最大和枯期系统保证出力最大的多目标优化模型,采用逐步优化算法求解,得到水电站最优运行过程;其次利用支持向量回归模拟水电站运行,最后以某典型水电站及其周边风光电站为例进行水电站规则提取。结果表明本文构建的优化模型能够有效改善系统丰枯电量结构,相较各电源单独优化出力极差降低13.90%,枯期最小出力提升12.43%。此外,使用支持向量机模拟水电站运行取得较高精度,与多元线性回归相比决定性系数最高提升2.18%,均方误差最高减少29.84%,能更好地学习新型电力系统下水电参与风光补偿后的运行规律。With the steady increase in the installation of new energy sources,the uncertain nature of the new power system is further strengthened,urgently requiring hydroelectric and other regulating power sources to dampen the seasonal and stochastic fluctuations of uncertain energy sources such as wind and solar power.This study aims to explore the extraction of optimal operating rules for hydroelectric power stations in the context of the new power system,to enhance the efficiency of clean electricity production and grid stability.Firstly,a multi-objective optimization model is constructed,focusing on maximizing the generation capacity of a multi-energy system and maximum guaranteed output during the dry season,which is solved using a stepwise optimization algorithm to obtain the optimal operating process of hydroelectric power stations.Secondly,support vector regression is employed to simulate the operation of hydroelectric power stations.Finally,using a typical hydroelectric power station and its surrounding wind and solar power stations as an example,the extraction of operating rules for hydroelectric power stations is conducted.The result indicate that the optimization model constructed in this study effectively improves the structure of the hydroelectric power system,reducing the extreme differences in output by 13.90%compared to optimizing each power source individually,and increasing the minimum output during the dry season by 12.43%.Furthermore,using support vector regression to simulate the operation of hydroelectric power stations achieves higher accuracy,with the coefficient of determination increasing by 2.18%compared to multiple linear regression,and the mean square error decreasing by up to 29.84%,enabling better learning of the operational patterns of hydroelectric power stations participating in wind and solar compensation in the new power system.

关 键 词：新型电力系统 水风光互补调度 支持向量回归 运行策略 逐步优化算法 

分 类 号：TM715[电气工程—电力系统及自动化] TV737[水利工程—水利水电工程]