Research On The Rail Defect Detection And Identification Of The Urban Rail Transit

As an important part of the track, rail contacts with train wheels directly and provides a smooth, continuous surface for the rolling wheels. With the accumulation of train operation and the increase of load, rail defects is increasing such as rail corrugation, squat, shell, crushing, dipped joint, humped weld, hollow weld and so on. The detection of these defects is important for improving the safety of operating vehicles.This paper focuses on the detection of rail corrugation, dipped joint, humped weld and hollow weld. To solve these problems, a vehicle-track dynamic model is built based on the SIMPACK and the vertical-longitudinal profiles of corrugation, dipped joint, humped weld and hollow weld are simulated. Then all the profiles which have been simulated are introduced into the vehicle-track dynamic model to obtain bogie acceleration data based on defects. In this paper, the time domain analysis, frequency domain analysis, wavelet analysis and Hilbert-Huang transform are used to analyze bogie acceleration data and extract the frequency characteristics of the defects above, then the location and the type of the defects is detected and identified, so it’s effective to detect and identify the rail defects based on bogie acceleration signal. Finally, the operational data of Shanghai Metro Line 1 are analyzed and the rail corrugation, dipped joint, squat is detected and identified. It verifies the feasibility and reliability of this method, which is of great significance to achieve online monitoring and identify the type of defects in rail transit.The idiographic research of this paper is as follows.(1) Explore the mechanism and serious harm of rail corrugation, dipped joint, humped weld, hollow weld and other defects, establish fitting models of each defect geometry profile.(2) Research on existing detection methods and the corresponding data processing techniques of rail corrugation, dipped joint, humped weld, hollow weld and other defects at home and abroad.(3) Establish an operating vehicle-track dynamic model based on SIMPACK, the multi-body dynamics software. The complete simulation model consists the following parts, operating vehicle dynamics model, wheel/rail contact model, and the vertical-longitudinal profile simulation models of each defect. Collect their bogie vibration acceleration data respectively.(4) Apply the classical time domain analysis, frequency domain analysis, wavelet analysis (wavelet packet decomposition and continuous wavelet transform) and Hilbert-Huang transform respectively, to analyze and compare the bogie vibration acceleration data based on the normal state, corrugation state, dipped joint state, humped weld state and hollow weld state. Then extract the frequency characteristics and natural vibration waveform of the defects above. It’s effective to detect and identify the rail defects based on bogie acceleration signal.(5) Analyze the twice operational data of Shanghai South Station to Jinjiang Action Park Station 0117 car in Shanghai Metro Line 1, and the rail corrugation, dipped joint, squat can be detected and identified from it. It verifies the feasibility and practicability of the simulation model Established by SIMPACK once again.