高压电器2025，Vol.61Issue(5)：258-267,290,11.DOI:10.13296/j.1001-1609.hva.2025.05.027
基于改进生成对抗网络场景生成的配电网多目标随机规划
Multi-objective Stochastic Planning of Distribution Network Based on Improved Generative Adversarial Network Scenario Generation
摘要
Abstract
The high dimensionality and uncertainty of new energy output restrict the absorption capacity of distribu-tion network to new energy.Therefore,it is necessary to fully consider the daily and periodic new energy outputs in distribution network planning.For this purpose,a multi-objective stochastic programming method of distribution net-work considering the uncertainties in source and load outputs is designed.Firstly,the K-means clustering algorithm is used to cluster the historical scenes into K daily state types and the corresponding probability distribution is ob-tained.Then,the scene set considering high-dimensional correlation between source and load is generated with the daily state types as condition and in combination with scenario generation method based on SN-CWGAN.Further-more,a multi-objective two-stage stochastic programming model is constructed for the distribution network,in which the upper layer determines the location and capacity of energy storage and plans for new energy,while the lower layer optimizes the operational plans of the distribution network for various typical scenarios.Finally,the RVEA algorithm is used to solve the proposed model.The CRITIC weight method is used to obtain objective weight and the to obtain the compromise solution by using a multi-criteria compromise solution sorting method.The planning method pro-posed in this paper is verified by using the IEEE33 node example system.The results show that the method not only accurately describes the probability distribution and high-dimensional correlation of source and load,but also get the planning result concurrently considering both economic factors and life cycle carbon emissions.关键词
SN-CWGAN/多目标随机优化/RVEA/CRITIC权重法/多准则妥协解排序法Key words
SN-CWGAN/multi-objective stochastic optimization/RVEA/CRITIC/VIKOR引用本文复制引用
吴金木,李剑,徐旭,幸进,朱轶群,杨玺..基于改进生成对抗网络场景生成的配电网多目标随机规划[J].高压电器,2025,61(5):258-267,290,11.基金项目
国家电网浙江省电力有限公司科技项目(B711JZ230012).Project Supported by Technology Project of State Grid Zhejiang Electric Power Co.,Ltd.(B711JZ230012). （B711JZ230012）
