Research And Implementation Of Intelligent Detection System For Pantograph Of EMU

In recent years,with the continuous increase of the railway operating mileage in China,the number of EMUs increased rapidly and the detection of the dynamic performance parameters of the pantograph has received an increasing attention as well.The safe running of train is inseparable from the normal operation of the pantograph.However,in the real railway environment,the pantograph contact status may be affected by many uncertain factors,such as the change in the shape and structure of the pantograph caused by the impact of flying objects,and pantograph deviation from catenary caused by the large arcing between the pantographs,contact obstacles between the pantographs caused by the suspension of foreign objects,etc.In some cases,the train may even stop running.In the real environment,it is difficult to continuously detect the pantograph status artificially.Therefore,it is urgent to conduct the research on the intelligent detection system of the pantograph of the EMU,which has important research value and broad application prospects.In this paper,the important indicators of the pantograph operation quality are analysed,and the method for detecting the dynamic performance of the pantograph is investiged.Based on this,a complete intelligent detection system for the dynamic performance of the pantograph is designed.This system realized the real-time monitoring of the pantograph status and intelligently identify whether the pantograph status is normal during the high-speed operation of the EMU,which could provide strong protection for the safe and stable operation of the train.The overall goal of this paper is to research on the abnormal state detection of the pantograph.The key issues and challenges of this work are descrbied through the detailed introduction of the hardware and software system composition of the system.Then the arc detection,pantograph abnormal detection and wear measurement are investigated.Finally,based on the deep learning method,experiments and analysis are carried out on arc detection,pantograph anomaly detection and wear measurement.The main work of this paper has been made in the following areas:1.The current railway construction and future development plans are investigated.The importance of pantograph state detection is realized through the current research status at home and abroad.The status of the pantograph is directly related to the safety of railway operations.The pantograph intelligent detection system of the EMU uses the data collection module installed on the roof and the intelligent analysis host installed in the car to jointly monitor the status of the pantograph,and output the real-time alarm results to the monitoring screen in the EMU.2.The working principle of the pantograph intelligent detection system is described,and the hardware and software composition of the intelligent analysis system composed of the pantograph intelligent detection system are introduced.Based on previous research,some traditional algorithms are analyzed to detect the abnormal arc and pantograph,and measure the wear of the carbon sliding plate.3.An algorithm based on deep learning is studied for intelligent detection of pantograph to overcome the shortcomings of some traditional algorithms.Classic algorithms are used for some common problems.Res-DFR is developed for the pantograph abnormality detection,which can realize the abnormality judgment of the pantograph and determine the abnormal area through the heat map.The data sets of arc detection,pantograph anomaly detection and wear measurement are constructed.The experiment uses the Py Torch deep learning framework and a large amount of data that verified by comparison and video,and uses deep learning-based methods for arc detection,pantograph anomaly detection and wear measurement,which can realize real-time pantograph status monitoring and detect abnormal conditions.The experiment has high experimental accuracy and low false alarm rate,which have laid the foundation for the improvement and daily application of the pantograph intelligent detection system for EMUs.