基于BP神经网络的继电保护状态预警研究

Research on State Warning of Relay Protection Device Based on BP Neural Network

继电保护装置是智能变电站的重要设备之一,正确评估其状态是确保电力系统安全运行的重要保证。本文从继电保护设备的运行数据出发,探索设备告警与设备故障的关系。采用BP(Back Propagation)网络方法实现继电保护状态评价和设备隐患预警;针对BP网络的难以处理离散状态问题,提出了基于统计阈值判定的BP网络。经过安徽电网继电保护设备的运行数据检验,本文提出的算法能形成对设备的状态评价,基本与运行设备的数据统计信息相符。

Relay protection is an important secondary equipment in smart substation,The correct evaluation of its state is an important guarantee to ensure the safe operation of power system.Based on the operating data of relay protection,this paper explores the relationship between equipment alarm and equipment failure.The Back Propagation(BP)neural network method was used to realize the relaying protection state evaluation and equipment hidden danger warning.In order to solve the problem of discrete state of BP network,a BP network based on statistical threshold is proposed.Through the operating data test of the relay protection in Anhui Electric Power CO.LTD,the algorithm proposed in this paper can form the state evaluation of equipment,which is basically consistent with the data statistics of running equipment.