Equipment Linkage And Fault Early Warning Of Power Environment Monitoring System

In the power system,power communication is an important part to ensure the safe production of power.If there is a problem in the power communication room,the power communication will be affected,which may further affect the power supply reliability of the power system.The power environment monitoring system of the communication room can monitor the power environment status of the communication room in real time.At present,there are still many problems in the power environment monitoring system,such as a lot of manual inspection,too many useless logs and redundant alarms in the system.It is very necessary to reduce the amount of redundant alarms according to the equipment linkage relationship,and it can greatly improve the efficiency of user.Deep learning has the advantages of actively extracting features from a large number of data and reducing the impact of human factors on the results.It can better realize the research on fault early warning.In this paper,the components of the power supply system and environmental factors in the communication room are taken as the research object.By analyzing the related alarm information in a large number of faults,and with the help of the expert experience of a large number of engineers,the alarm information is divided into corresponding fault information,and the equipment linkage scheme is designed according to the fault type.Then,a one-dimensional convolutional neural network model is designed to research on fault early warning,and the correlation data is fused into a sample set for training.The effectiveness is verified by experiments.The main work of this paper is as follows:(1)After consulting the relevant national and enterprise standards and literature,this paper deeply analyzes the data range of the power environment monitoring system,and determines the Manhattan distance as the method of alarm correlation analysis.(2)By sorting and analyzing a large number of actual data,through the previously determined alarm correlation analysis method,the alarm information is sorted into corresponding fault information according to the AC common circuit part,AC power supply part,rectifier module part,DC power distribution part,battery pack part and computer room environment part.(3)The functional requirements of equipment linkage are determined.With the help of the expert experience of a large number of engineers,the linkage scheme in case of different faults is designed.(4)A one-dimensional convolutional neural network model is designed.According to the needs of the research,the parameters of convolution layer,pooling layer,global average pooling layer,full connection layer and dropout layer are determined.During the training,a sample fusion method is proposed.The data collected at multiple different points are divided into five parts according to the common AC circuit,AC distribution unit,rectifier module,battery pack and DC distribution unit.Different types of related data in each part are fused to form a related sample set,and the effectiveness of each group is verified by experiments.