基于AI智能算法的设备故障诊断技术优化与仿真实验  

Equipment fault diagnosis technology optimization and simulation experiment based on AI intelligent algorithm

作　　者：梁志飞 王子石[1] 邓淑斌 牟春风 LIANG Zhifei;WANG Zishi;DENG Shubin;MOU Chunfeng(Guangzhou Electric Power Trading Center Co.,Ltd.,Guangzhou 510180,China;Tsinghua University,Beijing 100084,China)

机构地区：[1]广州电力交易中心有限责任公司,广东广州510180 [2]清华大学,北京100084

出　　处：《粘接》2024年第5期137-140,共4页Adhesion

摘　　要：为提高电力设备在线监测效率,设计了一种基于AI智能算法的智慧电力资源库。利用人群搜索算法(SOA)优化深度置信网络(DBN)的隐含层节点数,并用于电力变压器设备的故障诊断;基于C/S(Client/Server)框架的电力变压器设备故障分析模块,实现电力变压器设备故障在内的实时监测。结果表明,SOA优化DBN的故障诊断模型可有效诊断电力变压器的高温过热和高能放电故障。相较于标准DBN网络、SVM模型和BPNN网络,SOA优化DBN网络具有更高的准确率和查全率,分别达到95.38%和94.78%;所设计的电力设备资源数据库可对电力变压器数据进行实时诊断操作,从而能更好的辅助电力工程人员运维。To improve the efficiency of online monitoring and management of power equipment,a smart power resource library based on AI intelligent algorithm was designed.The crowd search algorithm(SOA)was used to optimize the number of hidden layer nodes of the Deep Belief Network(DBN),which was used for fault diagnosis of power transformer equipment.The power transformer equipment fault analysis module based on the C/S(Client/Server)framework realized real-time monitoring of power transformer equipment faults.The results showed that the fault diagnosis model of SOA optimized DBN can effectively diagnose high-temperature overheating and high-energy discharge faults of power transformer equipment.Compared to the standard DBN model,SVM model and BPNN model,the proposed improved DBN model had higher accuracy and recall,with an average accuracy and recall of 95.38%and 94.78%,respectively.The designed power equipment data management model can carry out real-time diagnosis and operation of power transformer data,so as to better assist the operation and maintenance of power engineering personnel.

关 键 词：AI智能算法 智慧电力资源库 C/S架构 SOA算法 DBN网络 

分 类 号：TP393.02[自动化与计算机技术—计算机应用技术]