基于人工智能的电力系统光纤故障检测研究  

Research on Optic Fiber Fault Detection in Power System Based on Artificial Intelligence

作　　者：王佳 张先涛 程洪超 崔国瑞 邹航 杨立 刘雪莹 徐倩 WANG Jia;ZHANG Xiantao;CHENG Hongchao;CUI Guorui;ZOU Hang;YANG Li;LIU Xueying;XU Qian(Chengdu Power Supply Company,State Grid Sichuan Electic Power Company,Chengdu 610004,China)

机构地区：[1]国网四川省电力公司成都供电公司,四川成都610004

出　　处：《自动化仪表》2024年第6期38-43,共6页Process Automation Instrumentation

摘　　要：在电力系统实际运行过程中,光纤通信面临着故障风险。基于此,提出了一种基于人工智能的电力系统光纤故障检测方法。首先,介绍了光纤网络结构以及故障监测的常用方法;其次,描述了神经网络的具体应用;最后,建立了对比试验,验证了模型的可靠性。试验结果表明:通过与实际故障原因的对比,采用神经网络进行故障分析的误判率在四种故障模式下分别低于3%、15%、14%、5%;优化模型的整体识别正确率为86.34%,高于传统模型。该结果验证了模型的合理性。该研究提出的电力系统光纤故障检测方法具有一定的参考意义。In the actual operation of power system,optic fiber communication faces the risk of failure.Based on this,a optic fiber fault detection method in power system based on artificial intelligence is proposed.Firstly,the optic fiber network structure and the common methods of fault monitoring are introduced;secondly,the specific application of neural network is described;finally,a comparison test is established to verify the reliability of the model.The test results show that:by comparing with the actual fault causes,the misjudgment rate of fault analysis using neural network is lower than 3%,15%,14%,and 5%in four fault modes;the overall identification correct rate of the optimization model is 86.34%,which is higher than that of the traditional model.The reasonableness of the model is verified.This research proposed method for optic fiber fault detection in power system has certain reference significance.

关 键 词：电力系统 光纤通信 人工智能 故障检测 光时域反射 神经网络 反向传播 

分 类 号：TH128[机械工程—机械设计及理论]