面向快速自愈的分布式智能配电网管理系统设计  

Design of a Distributed Intelligent Distribution Grid Management System Oriented for Rapid Self-Healing

作　　者：周俊卿 ZHOU Junqing(State Grid Huanggang Power Supply Company,Huanggang 438000,China)

机构地区：[1]国网黄冈供电公司,湖北黄冈438000

出　　处：《通信电源技术》2024年第17期44-46,共3页Telecom Power Technology

摘　　要：文章旨在设计一种面向快速自愈的分布式智能配电网管理系统,采用实时监测、机器学习故障检测与定位、深度强化学习自愈控制等技术,构建一套完整的系统架构。通过在某市郊区10 kV配电网进行为期10 d的实证研究,结果表明该系统能在16.8 ms内检测故障,178.3 ms内完成自愈,故障定位准确率达96.5%,证明该系统能够提升配电网供电可靠性和自愈能力。The purpose of this paper is to design a distributed intelligent distribution grid management system for rapid self-healing,and to construct a complete system architecture by using real-time monitoring,machine learning fault detection and location,deep reinforcement learning self-healing control and other technologies.A 10-day empirical study on a 10 kV distribution network in a suburb shows that the system can detect faults within 16.8 ms and complete self-healing within 178.3 ms,and the accuracy of fault location reaches 96.5%,which proves that the system can improve the power supply reliability and self-healing ability of the distribution grid.

关 键 词：分布式智能配电网 快速自愈 机器学习 

分 类 号：TM76[电气工程—电力系统及自动化]