基于粗糙集神经网络的高压设备红外图像自动故障识别方法  

Rough Set Neural Network-based Automatic Fault Identification for Infrared Images of High Voltage Equipment

作　　者：刘磊 LIU Lei(Shigatse Power Supply Company of State Grid Tibet Electric Power Co.,Ltd.,Shigatse 857000,China)

机构地区：[1]国网西藏电力有限公司日喀则供电公司,西藏日喀则857000

出　　处：《电工技术》2024年第16期95-97,共3页Electric Engineering

摘　　要：针对现有故障识别方法在对高压设备红外图像故障识别时存在识别结果损失值过大、准确性差的问题,引入粗糙集神经网络,开展高压设备红外图像自动故障识别方法的设计研究。收集不同类型、型号和运行状态的高压设备红外图像,建立高压设备红外图像样本库;利用粗糙集神经网络,实现红外图像数据挖掘;然后通过计算预警阈值、预警系数,实现故障自动识别与预警。对比实验证明,新的故障识别方法识别结果的损失值得到显著降低,识别准确性较强。In view of the problems of excessive loss of recognition results and poor accuracy of currently prevailing fault recognition methods in infrared image fault recognition of high voltage equipment,this paper introduces rough set neural network to the method of automatic fault recognition of infrared images of high voltage equipment.The method works by collecting infrared images of high voltage equipment of different types,models and operating states,establishing infrared image sample database of high voltage equipment,using rough set neural network to mine the infrared image data,and calculating the warning threshold and warning coefficient to realize automatic fault identification and warning.The experimental results show that the new fault identification method achieves significant reduction in the loss value of equipment results,and higher identification accuracy.

关 键 词：粗糙集神经网络 高压设备 红外图像 故障识别 

分 类 号：R197.324[医药卫生—卫生事业管理]