基于差分阈值算法的水电站厂房漏电火灾预警研究  

Research on leakage fire warning of hydropower plant based on differential threshold algorithm

作　　者：查海涛 杨峰 任雪峰 谢青 帅挽澜 ZHA Hai-tao;YANG Feng;REN Xue-feng;XIE Qing;SHUAI Wan-lan(Zhelin Hydropower Plant,State Grid Jiangxi ElectricPower Co.,Ltd.,Jiangxi Jiujiang 332000,China;School ofInformation Engineering,Nanchang University,Jiangxi Nan-chang 330031,China)

机构地区：[1]国网江西省电力有限公司柘林水电厂,江西九江332000 [2]南昌大学信息工程学院,江西南昌330031

出　　处：《消防科学与技术》2021年第3期378-380,401,共4页Fire Science and Technology

基　　金：国家自然科学基金项目(51867017)。

摘　　要：面对水电站厂房电气设备复杂的漏电情况,传统的漏电检测技术难以有效检测所有异常情况,据此提出基于差分阈值的漏电检测技术。首先采用平滑滤波技术对信号进行滤波处理。其次计算信号差分值,设立时间窗口,并根据窗口设立初始阈值,由此检测相对异常信号点。最后结合环境因素更新阈值,对异常漏电情况进行识别和预警,进行下一次的检测循环。通过相应数据分析,验证了该技术的正确性和有效性,能够有效检测出相对异常剩余电流,预防漏电火灾。In the face of complex leakage situation of electrical equipment in hydropower station,the traditional leakage detection technology is difficult to effectively detect all abnormal situations.In view of this situation,a leakage detection technology based on differential threshold is proposed.First,the smooth filtering technology is used to filter the signal.Then the signal difference value is calculated,the time window is set up,and the initial threshold is set up according to the window,so as to detect the relative abnormal signal points.Finally,the threshold value is updated based on environmental factors,and the abnormal leakage situation is identified and warned,and the next detection cycle is carried out.The correctness and effectiveness of the technology are verified by analyzing the corresponding data,which can effectively detect the relatively abnormal residual current and prevent the leakage fire.

关 键 词：水电厂 漏电流 阈值 火灾预警 

分 类 号：X928.7[环境科学与工程—安全科学] TV737[水利工程—水利水电工程]