摘要
针对直流配电网传统控制策略在分布式电源出力波动情况下,存在电压频繁越限、网损增大的问题,提出了一种基于强化学习近端策略优化(Proximal Policy Optimization,PPO)算法的直流配电网网损/电压优化控制策略。构建了面向电压和网损优化的马尔可夫决策过程(Markov Decision Process,MDP)模型,分别定义了直流配电网网损/电压优化控制的状态空间和动作空间,利用电压偏差和网损优化目标划分设计奖励函数,确保策略在满足电压约束条件的同时优化网损,采用PPO算法进行策略优化训练,得到最优控制策略。以修改的IEEE16直流配电网为算例,通过DIgSILENT和Python进行所提方法的程序设计和仿真,结果表明,所提方法能够在分布式电源波动时,优化系统网损,并提升电压安全运行水平。
The challenges posed by traditional control strategies in DC distribution networks is addresses,particularly in the presence of fluctuations in distributed energy generation,which lead to frequent voltage violations and increased network losses.A novel optimization control strategy based on Proximal Policy Optimization(PPO)algorithm is proposed to optimize both voltage and network losses.A Markov Decision Process(MDP)model is developed for voltage and network loss optimization,with defined state and action spaces for control.A reward function is designed based on the partitioning of voltage deviation and network loss optimization objectives,ensuring voltage constraints are met while minimizing network losses.The PPO algorithm,is employed for policy optimization training to obtain the optimal control strategy.The proposed approach is validated through simulation on a modified IEEE 16-bus DC distribution network,using DIgSILENT/PowerFactory and Python for program design and simulation.The results demonstrate that the proposed method effectively optimizes network losses and enhances voltage stability under fluctuating distributed energy outputs.
作者
张锋
张培超
杨浩
ZHANG Feng;ZHANG Peichao;YANG Hao(Key Laboratory of Modern Power System Simulation and Control&Renewable Energy Technology,Ministry of Education(Northeast Electric Power University),Jilin 132012,China;Shandong Electric Power Engineering Consulting Institute Co.,Ltd.,Jinan 250013,China)
出处
《电气应用》
2025年第4期75-84,共10页
Electrotechnical Application
基金
吉林省自然科学基金面上项目(20240101108JC)。
