基于组合在线学习的有源配电网空调负荷电压控制系统设计  

Design of voltage control system for air-conditioning loads in active distribution networks based on combinatorial online learning

作　　者：刘学 周昊 万程 周涛 王颖 LIU Xue;ZHOU Hao;WAN Cheng;ZHOU Tao;WANG Ying(State Grid Hunan Electric Power Co.,Ltd.,Energy Conservation Management Branch,Changsha 410004,China)

机构地区：[1]国网湖南省电力有限公司节能管理分公司,长沙410004

出　　处：《自动化与仪器仪表》2025年第3期299-304,共6页Automation & Instrumentation

基　　金：国网湖南省电力公司科技项目(B216AS20007)。

摘　　要：分布式资源的大量接入使得配电网潮流由传统单向流动向双向流动化转变,导致并网点电压频率波动异常,给配电网优化调控带来了巨大挑战。空调负荷作为负荷侧重要的资源之一,可以灵活参与电力系统相关调控,从而促进电力系统安全稳定运行。提出一种基于组合在线学习的配电网空调负荷聚合优化调度方法,通过探索与利用不断优化反馈研究用户空调负荷在线聚合最优控制。首先建立基于等效热参数的分散式空调典型模型,进一步引入空调负荷同时率,分析空调负荷有功、无功功率等效聚合模型;在此基础上,考虑有源配电网运行经济性与安全性,以全系统网损、节点电压偏差和调节成本最小为目标函数,通过组合在线学习方法求解含空调负荷的有源配电网优化调度模型。最后,采用实际负荷数据在IEEE33节点系统进行仿真,验证文章提出的方法和所建模型的有效性,实现通过空调负荷调度达到降损调压的目标。With the massive access of distributed resources,the distribution network trend is changing from the traditional one-way flow to two-way flow,resulting in abnormal voltage and frequency fluctuations at the grid connection point,which poses great challenges to the optimization and regulation of the distribution network.As one of the important resources on the load side,air-conditioning loads can flexibly participate in power system-related regulation and control,thus promoting the safe and stable operation of the power system.The article proposes a distribution network air-conditioning load aggregation and optimal scheduling method based on combinatorial online learning,through the exploration and use of continuous optimization feedback to study the user air-conditioning load online aggregation control.Firstly,a typical model of decentralized air-conditioning based on equivalent thermal parameters is established,and the simultaneous rate of air-conditioning load is further introduced to analyze the equivalent aggregation model of active and reactive power of air-conditioning loads;on this basis,considering the operational economy and security of active distribution networks,the active distribution network optimization scheduling model containing air-conditioning loads is solved through the combinatorial online learning method by taking the minimization of the whole-system network loss,node voltage deviation and adjustment cost as the objective function.Finally,simulation is carried out in the IEEE33 node system using actual load data to verify the validity of the proposed method and the constructed model,and to realize the goal of loss reduction and voltage regulation through air conditioning load dispatch.

关 键 词：空调负荷 组合在线学习 负荷聚合模型 有源配电网 

分 类 号：TM73[电气工程—电力系统及自动化]