基于PSO-DBN的矿用电缆局部放电特征识别研究  

Research on Local Discharge Characteristic Recognition Based on PSODBN for Mining Cable

作　　者：郭兴超 张红娟[1] 靳宝全 GUO Xingchao;ZHANG Hongjuan;JIN Baoquan(College of Electrical and Power Engineering,Taiyuan University of Technology,Taiyuan 030024,China;Key Laboratory of Advanced Transducers and Intelligent Control System Ministry of Education,Taiyuan University of Technology,Taiyuan 030024,China)

机构地区：[1]太原理工大学电气与动力工程学院,太原030024 [2]太原理工大学新型传感器与智能控制教育部与山西省重点实验室,太原030024

出　　处：《煤炭技术》2024年第9期263-266,共4页Coal Technology

基　　金：中央引导地方科技发展资金项目(YDZJSX20231B004);山西省科技创新团队项目(201805D131003)。

摘　　要：为解决矿用电缆局部放电特征识别的问题,提出利用粒子群算法优化的深度置信网络(PSO-DBN)实现矿用电缆局部放电特征识别。搭建了隔爆兼本安型局部放电监测系统采集放电信号,构造了局部放电相位分布模式(PRPD)图谱并提取特征参数,采用PSO算法自适应地选择DBN的超参数,从而确定最优的网络结构。用PSO-DBN模型对气隙放电、划伤放电、沿面放电3种矿用电缆缺陷局部放电信号进行特征识别,识别准确率分别为93.3%、96.7%、95.0%。研究表明,该方法可以有效识别局部放电类型,为矿用电缆故障检测提供了新的方案。To address the issue of local discharge characteristic recognition in mining cables,a deep belief network optimized by particle swarm optimization(PSO-DBN)is proposed for cable local discharge feature recognition.A flameproof and intrinsically safe partial discharge monitoring system was set up to collect discharge signals,and a partial discharge phase distribution pattern(PRPD)chart was constructed and feature parameters were extracted.The PSO algorithm was used to adaptively select the hyperparameters of DBN,thus determining the optimal network structure.The PSO-DBN model was used for feature recognition of three types of mining cable defects:air gap discharge,scratch discharge,and surface discharge,with recognition accuracies of 93.3%,96.7%,and 95.0%respectively.The study demonstrates that this method can effectively identify the type of local discharges,providing a new approach for fault detection in mining cables.

关 键 词：矿用电缆 局部放电 PSO DBN PRPD 

分 类 号：TM855[电气工程—高电压与绝缘技术]