Research On Secondary Equipment Condition Monitoring System Driven By Big Data

With the increasing maturity of substation relay protection technology,a large number of cables connecting various levels of equipment in the secondary system of the substation are replaced by communication optical fiber links,and digital signals replace traditional analog signals,thereby realizing the information integration of the secondary system of the substation.The status monitoring data of the secondary system of the substation has the characteristics of large data size,variable structure and diversified types.The early operation and maintenance methods of the secondary system of the substation relied on manual judgment of the alarm signals parsed by the message,so it was difficult to form an efficient and accurate condition monitoring and fault location decision.In order to solve the problems of difficult extraction of feature quantities,difficult fault source location,and long troubleshooting time in the big data of secondary monitoring of substations,this paper started research on this subject with the support of the State Grid Sichuan Electric Power Academy Relay Protection Project.Firstly,classify the secondary equipment alarm signals of the substation.Based on a large amount of actual operation and maintenance experience and the constraints of the national grid technical specifications,analyze the mapping relationship between the substation secondary system alarm information and fault output types.The purpose is to operate The maintenance personnel can better understand the logical relationship between the alarm signal and the failure mode,better understand the overall structure of the secondary system of the substation,and improve the efficiency of fault test and location;second,the alarm for the failure alarm of the optical fiber communication link in the secondary system of the substation Mechanism,an optical fiber link communication fault location system algorithm based on graph theory is proposed to realize the module optimization of the optical fiber communication link in the secondary system of the substation,reduce the core algorithm input and data load platform load,and improve the secondary state monitoring and failure of the substation.The efficiency of the positioning system.Finally,for the complex alarm information that is difficult to trace directly,a new substation equipment condition monitoring and fault location algorithm based on deep learning theory driven by big data technology is proposed.An RNN(Recurrent Neural Network RNN)recurrent neural network model is used as a deep learning classifier model to automatically extract,classify,and predict data features,and continuously optimize network parameters based on backpropagation errors.This algorithm can achieve a comprehensive fault location accuracy rate of 82.02% on multiple fault types.Compared with the low-efficiency maintenance mode in which the secondary equipment of the substation relies on manual maintenance,the operation and maintenance of the secondary equipment condition monitoring system of the substation is greatly improved.effectiveness.The sample data of the background monitoring data of the first integrated substation in Yixiang,Sichuan proves that the RNN recurrent neural network algorithm can accurately predict the type of fault in the secondary equipment condition monitoring and fault location system of the substation,and validates the design method of this paper Effectiveness and feasibility.