基于自学习的直流系统绝缘故障研究  

A Study on the Insulation Failure of the DC System Based on Self-Learning

作　　者：张开迪 李世勉 高凌 冯宗琮 张葛 ZHANG Kaidi;LI Shimian;GAO Ling;FENG Zongcong;ZHANG Ge(Beibei Power Supply Branch of State Grid Chongqing Electric Power Company,Chongqing 400700,P.R.China;Chongqing Engineering Research Center of Energy Internet of Chongqing University of Technology,Chongqing 400054,P.R.China)

机构地区：[1]国网重庆市电力公司北碚供电分公司,重庆400700 [2]重庆理工大学重庆市能源互联网工程技术研究中心,重庆400054

出　　处：《重庆电力高等专科学校学报》2020年第5期1-4,共4页Journal of Chongqing Electric Power College

基　　金：国网重庆市电力公司北碚供电分公司2020研究开发项目(SGCQBP00YJJS2000229);重庆市技术创新与应用发展专项cstc2019jscx-msxmX0003。

摘　　要：针对常用的直流系统绝缘检测方法的不足,在直流漏电流检测法基础之上,提出将漏电流转换为漏电压,运用运放检测的方法,克服灵敏度低的不足。首先,在MATLAB上建立等效电路模型,模拟其工作状态,确定拓扑可行性,然后分析拓扑的工作方式以及工作原理。其次,提出卡尔曼滤波分析法,收集直流系统的历史数据,运用卡尔曼滤波纠正和修改历史数据。再次,系统分析数据,得到规律,建立数学模型,通过不断地进行自学习,预测直流系统绝缘故障,保证电力系统安全稳定的运行。最后,在电厂模拟故障,比较系统测试值与人工测试值,验证系统的可行性。In order to improve the deficiencies of commonly used insulation detection methods for the DC system,based on the DC leakage current detection method,this paper suggests converting leakage current to leakage voltage and overcoming the inaccurate detection and non-isolation through the distribution detection.The equivalent circuit model can be established by means of the MATLAB to simulate its operating state and ensure its topology feasibility.Then,the operating mode and principle of the topology can be analyzed.The Kalman filtering analytical method can be used to collect,collect and modify historical data of the system.After the data analysis,the law can be gained to establish the mathematical model.Continuous self-learning is helpful in predicting insulation failures of the DC system and ensuring safe and stable operation of the power system.Failure simulations can be implemented in the power plant to compare the values measured by the system and manual work to verify the feasibility of the system.

关 键 词：直流系统 绝缘检测 卡尔曼滤波 自学习 

分 类 号：TM721.1[电气工程—电力系统及自动化]