考虑信息-物理耦合与资源调度智能性的灾后配电网重构方法

Post-Disaster Distribution Network Reconfiguration Method Considering Cyber-Physical Coupling and Intelligent Resource Dispatching‌

中文关键词:多层感知机  信息物理  灾害  配电网

keywords:multi-layer perceptron  cyber-physical  disaster  distribution network

中文摘要:为提升配电网在极端事件下快速恢复的能力，增强系统韧性，提出一种融合信息物理系统建模与人工智能评估机制的智能化灾后恢复优化策略。首先，构建基于供电中断区域的恢复效能指标体系，涵盖负荷恢复率、重要节点供电优先度、调度响应时间、通信修复占比等6项关键指标，从物理子网与信息子网耦合角度刻画恢复能力；其次，构建融合信息-物理-智能3层结构的配电网建模框架，提出一种改进的恢复效能评估方法，融合多层感知机神经网络模型实现对候选恢复策略的非线性评分与排序，引导多目标优化求解器以实现资源调度与网络重构协同优化；最后，基于IEEE 33节点配电系统构建典型灾害场景，分别从不同策略、多场景与不同优化权重角度开展仿真验证。

Abstract:To improve the rapid power restoration capacity and resilience of distribution networks under extreme disasters，this paper proposes an intelligent post-disaster restoration optimization strategy integrated with cyber-physical system modeling and artificial intelligence evaluation mechanism. First，a restoration efficiency index system targeting blackout areas is established，covering six core indicators: load restoration rate，power supply priority of critical nodes，dispatch response time，and communication restoration ratio. The system quantifies the post-disaster restoration capacity from the coupling perspective of physical subnetworks and cyber subnetworks. Second，a three-layer distribution network modeling framework integrating cyber layer，physical layer and intelligent decision layer is constructed. An improved restoration efficiency evaluation method is proposed，which adopts the MLP neural network to realize nonlinear scoring and ranking of candidate restoration strategies. The ranking results guide the multi-objective optimization solver to coordinate emergency resource dispatching? and distribution network reconfiguration. Finally，a typical disaster case is built based on the IEEE 33-bus distribution test system. Simulations are performed from three perspectives: comparative strategies，multiple disaster scenarios and different optimization weight coefficients.