针对主动配电网中高比例分布式光伏发电的接入将会造成配电网电压越限的问题,传统的基于模型的电压控制方法在很大程度上依赖于精确的物理参数,这在实际工程中难以达到理想的效果.基于配置储能设备和无功补偿型功率变换器来解决主动配电网的电压越限问题,提出了一种基于数据驱动的光伏—储能主动配电网电压控制方法.为了实现储能设备和无功补偿型功率变换器的优化控制和管理,通过将具有系统功率损耗和电压约束的优化问题建模为马尔可夫决策过程(Markov De-cision Process,MDP),同时利用基于柔性演员—评论家(Soft Actor Critic,SAC)的强化学习算法进行求解.最后,在IEEE 33节点配电网中对所提出的电压控制方法进行了验证,仿真结果表明,所提出的控制方法能够有效地降低主动配电网的电压偏差和电网的功率损耗,具有较好的控制性能.
Research on data-driven voltage control method for photovoltaic and energy storage active distribution network
The integration of high proportion distributed photovoltaic power generation in active distribution networks will cause voltage exceeding limits in the distribution network.Traditional model-based voltage control methods largely rely on accurate physical parame-ters,which is difficult to achieve ideal results in practical engineering.This article proposed a data-driven voltage control method for photovoltaic and energy storage active distribution networks based on configuring energy storage devices and reactive power compensa-tion type power converters to solve the problem of voltage exceeding limits in active distribution networks.In order to achieve optimal control and management of energy storage devices and reactive power compensation type power converters,the optimization problem with system power loss and voltage constraints was modeled as a Markov Decision Process(MDP),and solved using a reinforcement learning algorithm based on Soft Actor Critical(SAC).Finally,the proposed voltage control method was validated in an IEEE 33 node distribu-tion network.Simulation results showed that the proposed control method can effectively reduce the voltage deviation and power loss of the active distribution network,and has good control performance.
active distribution networkvoltage controldata-drivenenergy storage
万方数据
维普
