Surface Defects Detection Of Turnouts Based On In-service Vehicle

Turnout is a necessary part of railway, it is also one of the most weak parts of railway, the defects of it will directly affect the urban rail transit operation safety. To ensure the normality of turnout, it is of great significance to transfer the periodic detection to online detection. At present, the mostly common used of turnout defect detection methods are hand detection and small instruments detection, such as use specific instrument to detect the poor fit of switch rail and detect the curve opposite super-elevation by hand. The detection methods mentioned above need plenty of human and material resources, besides, the methods cannot detect the sudden defect effetely because of their bad real time capability. In recent years the vehicle-turnout dynamics has become the research focus in the field of railway and the theory results can offer support for turnout defect detection based on in-service vehicle, but the existing research are mainly focus on the interaction between vehicle and turnout, rarely take the turnout defect into consideration. According to the phenomenon, in this paper, the vehicle-turnout model is established in SIMPACK to simulate turnout defects, the dynamics response when the vehicle pass by a normal turnout and a defect turnout are compared in time domain and frequency domain, the simulation model is validated by test-bed. In addition, the feasibility analysis of the detection method is given. The research of this paper can offer a new thought of turnout defect detection.As follows is idiographic research of this thesis:(1) Establish the vehicle-turnout model in SIMPACK and simulate four kinds of defects on the model.(2) Analyze the vehicle acceleration when it pass by a turnout in time domain and frequency domain and summarize the effect of vehicle vibration acceleration by turnout defects.(3) Introduce the theoretical of SVM algorithm and apply it to realize the defect isolation of turnout.(4) Simulate the turnout defect in test-bed to validate the model established by SIMPACK. (5) Study the location of the turnout in railway by eigenvalue and Hilbert, give a preliminary analysis of vehicle vibration acceleration in Shanghai metro line1to validate the feasibility of the detection method proposed in this paper.