基于神经网络的电气火灾智慧监控平台研究  

Research on intelligent electrical fire monitoring platform based on Neural Network

作　　者：胡志朝 吴桂初[1,2] 吴自然 舒亮[1] 闫俊涛 HU Zhi-zhao;WU Gui-chu;WU Zi-ran;SHU Liang;YAN Jun-tao(Zhejiang Key Laboratory of Low Voltage Electrical Intelligent Technology,Wenzhou University,Zhejiang Wenzhou 325000,China;Wenzhou Juchuang Electrical Technology Co.,Ltd.,Zhejiang Wenzhou 325000,China)

机构地区：[1]温州大学浙江省低压电器智能技术重点实验室,浙江温州325000 [2]温州市聚创电气科技有限公司,浙江温州325000

出　　处：《消防科学与技术》2019年第11期1563-1566,共4页Fire Science and Technology

基　　金：温州市重大科技专项项目“基于LoRa无线传感的智慧用电系统的研究和应用”（ZG2019017）

摘　　要：为对电气火灾进行早期预测和预警,开发了电气火灾智慧监控系统。介绍该系统的工作原理、功能、硬件组成以及漏电流的BP神经网络模型设计,并从SOM神经网络结构、训练等方面探讨SOM神经网络聚类在判别高危时间段中的应用,分析SOM神经网络聚类结果。结果显示:该智慧系统能够提前发现可能存在的电气火灾隐患并进行预警,避免电气火灾的发生。In order to predict and forewarn the electrical fire,an intelligent monitoring system of electrical fire is developed.This paper introduces the working principle,function,hardware composition of the system and the design of BP neural network model of leakage current,and discusses the application of SOM neural network clustering in identifying high-risk time period from SOM neural network structure and training,and analyzes the clustering results of SOM neural network.The results show that the intelligent system can detect the potential electrical fire hazards in advance and give early warning to avoid the occurrence of electrical fire.

关 键 词：电气火灾 神经网络 火灾预警 安全用电系数 

分 类 号：X924.4[环境科学与工程—安全科学] TP277.2[自动化与计算机技术—检测技术与自动化装置]