基于神经网络技术的电气火灾预警系统研究  

Research on Electrical Fire Early Warning System Based on Neural Network Technology

作　　者：于兰[1,2] 贾振国 YU Lan;JIA Zhen-guo(School of Energy and Power,Changchun Institute of Technology,Changchun 130012,China;Jilin Construction Energy Supply and Indoor Environment Control Engineering Research Center,Changchun 130012,China)

机构地区：[1]长春工程学院能源动力工程学院,长春130012 [2]吉林省建筑能源供应及室内环境控制工程研究中心,长春130012

出　　处：《自动化与仪表》2022年第8期19-23,35,共6页Automation & Instrumentation

基　　金：吉林省教育厅科学技术研究项目(JJKH20210677KJ);吉林省科技发展计划项目(20190303131SF);吉林省科技厅项目(20160204019SF)。

摘　　要：为了提高电气火灾预警能力,构建了一套电气火灾预警系统。该系统通过主控制单元可编程逻辑控制器(PLC)实现数据的分析与处理。采用探测器对电气设备运行状态进行探测。采用监控主机对电气运行设备进行现场监控。采用窄带物联网(NB-IoT)技术进行对监控数据的长距离传输。利用改进反向传播(BP)神经网络算法,提高了电气火灾故障的识别能力,增强了电气火灾预警能力,实验表明,该研究系统电气火灾预警响应最短时间为0.5 s,电气火灾故障识别的准确率高达96%,故障识别准确率高,大大提高了电气火灾预警能力。In order to improve the electrical fire early warning capability,a set of electrical fire early warning system is constructed.The system realizes data analysis and processing through the main control unit programmable logic controller(PLC).Detectors are used to detect the operating status of electrical equipment.The monitoring host is used for on-site monitoring of electrical operating equipment.The narrowband internet of things(NB-IoT)technology is used for long-distance transmission of monitoring data.Using the improved back-propagation(BP)neural network algorithm,the results have improved the identification ability of electrical fire faults and enhanced the ability of electrical fire early warning.Experimental result demonstrates the minimum response of the research system is 0.5 s,the accuracy rate of electrical fire fault identification is as high as 96%,and the fault identification accuracy rate is high.The electrical fire early warning capability is greatly improved.

关 键 词：电气火灾预警 BP神经网络 监控 无线网络 故障识别 

分 类 号：TM933[电气工程—电力电子与电力传动]