基于深度强化学习的光伏主动配电网暂态电压控制  

Transient Voltage Control of Photovoltaic Active Distribution Network Based on Deep Reinforcement Learning

作　　者：宋晓喆 孙剑英 付小标 王长江[2] 王鼎 SONG Xiaozhe;SUN Jianying;FU Xiaobiao;WANG Changjiang;WANG Ding(State Grid Jilin Electric Power Supply Company,Changchun 130021,China;Key Laboratory of Modern Power System Simulation and Control&Renewable Energy Technology,Ministry of Education(Northeast Electric Power University),Jilin 132012,China)

机构地区：[1]国网吉林省电力有限公司,长春130021 [2]现代电力系统仿真控制与绿色电能新技术教育部重点实验室(东北电力大学),吉林132012

出　　处：《电力系统及其自动化学报》2025年第2期78-88,共11页Proceedings of the CSU-EPSA

基　　金：国网吉林省电力有限公司科技项目(SGJL0000DKJS2200369)。

摘　　要：为解决大量电力电子设备接入导致电网故障后暂态电压跌落问题,提出一种基于深度强化学习的光伏主动配电网暂态电压控制策略。首先,分析光伏逆变器在暂态电压跌落情况下的无功输出能力与交流侧电压、有功功率的内在关系,建立光伏逆变器短时无功功率输出的显示表达式。然后,进一步采用深度确定性策略梯度算法优化光伏功率输出的内环控制参数,使光伏逆变器根据电网电压跌落情况调节自身的有功和无功功率输出。为提高暂态无功响应速度,设计暂态/稳态控制策略的灵活切换机制,使光伏逆变器在电压跌落期间采用单闭环控制快速注入无功功率支撑电网电压。最后,通过算例仿真对比,验证所提控制策略的有效性。To solve the problem of transient voltage drop after grid fault which is caused by the access of a large quantity of power electronic equipment,a transient voltage control strategy for photovoltaic(PV)active distribution network based on deep reinforcement learning is proposed.First,the inherent relationship between the reactive power output ca-pacity of a PV inverter and the AC-side voltage and active power under transient voltage drop is analyzed,and an explic-it expression for the short-term reactive power output from the PV inverter is established.Then,a deep deterministic policy gradient algorithm is further used to optimize the inner-loop control parameters of the PV power output,so that the PV inverter can adjust its active and reactive power output according to the grid voltage drop.To improve the re-sponse speed of transient reactive power,a flexible switching mechanism of transient/steady-state control strategy is de-signed,so that the PV inverter can quickly inject reactive power to support the grid voltage during the voltage drop by means of single closed-loop control.Finally,the effectiveness of the proposed control strategy is verified by the compari-son of numerical examples through simulations.

关 键 词：光伏发电 主动配电网 暂态电压 深度强化学习 电压支撑 

分 类 号：TM615[电气工程—电力系统及自动化]