中国电机工程学报2025，Vol.45Issue(12)：4593-4607,中插5,16.DOI:10.13334/j.0258-8013.pcsee.232763
基于图像分析的电能质量扰动边-云协同辨识框架
Edge-cloud Collaborative Identification Framework for Power Quality Disturbances Based on Image Analysis
摘要
Abstract
With the increase in the penetration rate of distributed sources and loads,the monitoring data of sensors is increasing dramatically.Power grid maintenance services propose a rapid response requirement in power quality data analysis.To achieve rapid response and highly accurate identification of power quality disturbance(PQDs),this paper proposes an edge-cloud collaborative identification framework for PQDs based on image analysis.Leveraging the latest advancements in the field of image analysis,the concept of double-phase Lissajous locus(DPLL)is proposed to convert PQDs signals into locus image with unique shapes.A lightweight convolutional neural network(CNN)with the same structure is deployed at the edge and cloud,respectively,to perform rapid response and training tasks.By sharing model weights of edge-cloud,this framework can achieve rapid and high-precision identification of PQDs.To continuously improve model performance,a depth CNN is designed to be deployed to the cloud for data labeling to assist model update.The experimental results demonstrate that the proposed framework can provide more accurate PQDs identification and meet the real-time response requirements in engineering practice.关键词
边-云协同/电能质量扰动/双相Lissajous轨迹/轻量级卷积神经网络/图像识别Key words
edge-cloud collaboration/power quality disturbances/double-phase Lissajous locus/lightweight convolutional neural network/image identification分类
动力与电气工程引用本文复制引用
张玺,郑建勇,梅飞,高昂,缪惠宇..基于图像分析的电能质量扰动边-云协同辨识框架[J].中国电机工程学报,2025,45(12):4593-4607,中插5,16.基金项目
江苏省国际科技合作项目(BZ2021012). Project Supported by the International Science and Technology Cooperation Program of Jiangsu Province(BZ2021012). （BZ2021012）
