基于人工智能算法的电网电压稳定性研究  

Research on power grid voltage stability based on artificial intelligence algorithm

作　　者：陶用伟 TAO Yong-wei(Guizhou Grid Corp Kaili Power Supply Company Kaili,Kaili 556000,Guizhou Province,China)

机构地区：[1]贵州电网有限责任公司凯里供电局,贵州凯里556000

出　　处：《信息技术》2025年第4期121-126,134,共7页Information Technology

摘　　要：电网的稳定是电网的关键因素,尤其是在智能电网快速发展的时代。如何更大程度的保证稳定性是目前的热门话题。文中提出一个知识转移系统,用于电网间的长期电压稳定评估。首先,使用更通用的方法将系统行为转换为heatmap,以避免负迁移。然后,引入一个深度域自适应网络,通过添加一个最大平均差异计算器来学习具有强语义分离的域不变表示。基于深度域自适应网络的知识转移系统将知识从IEEE 39总线系统传输到IEEE 14总线系统。实验表明,提出的方法能够满足实际需求。The stability of the power grid is a key factor for the power grid,especially in the era of rapid development of smart grids.How to ensure the stability to a greater extent is a hot topic at present.In this paper,we implement a knowledge transfer system for long-term voltage stability assessment between grids.Firstly,a more general approach is used to convert the system behavior into a heatmap to avoid negative transfer.Then,a deep adaptive network is introduced to learn domain-invariant representations with strong semantic separation by adding a maximum mean difference calculator.A knowledge transfer system based on deep adaptive networks is used to transfer knowledge from the IEEE 39 bus system to the IEEE 14 bus system.Experiments show that the proposed method can meet the actual needs.

关 键 词：智能电网 人工智能 卷积神经网络 电压稳定 知识转移 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]