Research On Local Intelligent Monitoring System For Integrated Traction Substation

With the rapid development of urbanization and economy in China,high-speed railway and urban rail transit have entered a new historical stage of development.The safe and reliable operation of traction power supply system is the prerequisite for the safety and punctuality of rail transportation.As an important component of traction power supply system,traction substation attracts more and more attention.In the face of high requirements of China's high speed railway and railway safety,the existing traction substation monitoring system have become increasingly demanding,especially the real-time monitoring of the monitoring system acquisition accuracy and substation environment has been unable to meet the requirements of unattended substation.To this end,this paper presents an integrated traction variable in design of intelligent monitoring system.To improve the efficiency of the system,integrated design of traction substation on-site monitoring system is adopted.The artificial intelligence algorithm based on BP neural network is used to filter the collected signals to eliminate the power consumption,so as to ensure the reliability of the information and he accuracy of the local monitoring system.Through the intelligent recognition of the surveillance video image,it plays a very good auxiliary role in monitoring the effective environment of substation..First of all,according to the model of local monitoring system in existing substations,the design principles and functions of the system are studied in detail,and the design points of local monitoring system for traction substation are given.By comparing and analyzing the network model of the local monitoring system of traction substation,the Ethernet based network model is determined,and the hierarchical management structure is established based on the concept of hierarchy and modularity.Next,data sharing and linkage operation needs for monitoring system is given.Logical linkage between each subsystem and the data server for data fusion of large quantities are designed,software structure is established in substation monitoring system for traction integration,and gives the system can realize the function of the application.The software structure of the integrated monitoring system is completed.Then,on the basis of data sharing between sensors and data acquisition devices in traction substation,sample training is carried out to complete parameter selection based on BP neural network,so as to realize data filtering in data server.At last,the original video surveillance image provided by the local monitoring system,combined with machine vision and image processing technology,the personnel activities in the video are pre warning.In order to remove image noise,median filtering for enhancement image is adopted.Based on Canny operator on the target region preliminary discrimination,improved edge detection method.is used.According to the template matching based on similarity principle for personnel identification,it can effectively identify the number of people within the field of view,and provide effective support for environmental monitoring.Finally,,,the simulation experiment of signal acquisition based on BP neural network and KingView configuration software,is carried out.After 105 iterations,the optimal value is obtained,and the correct rate of existing data screening is 85%.Using the existing image samples to verify the intelligent recognition of the video surveillance system,the correct rate of the number recognition is 86%.The simulation experiment shows that the system meets the design requirements for the local intelligent monitoring of the integrated traction substation.