Detection Technology Of Rail Squats By Bogie Acceleration Based On Operating Vehicles

ABSTRACT:As one of the most dangerous short track irregularities, a squat is a type of rolling contact fatigue (RCF) damage. At the early stage squats can be treated by grinding of a thin layer from the rail surface. At the late stage squats always lead to rail replacement. If not treated in time, the cracks may grow in the subsurface and cause rail break. Therefore, the existence of squats directly impacts the operating condition of urban rail vehicle. For cost effective maintenance policy squats should be detected at an early stage. The current researches on the rail squats damage are based on track inspection car and rail flaw detector, but this method belongs to the artificial detection method. It has some disadvantages, for example, long detection cycle, nonreal-time detecting, large investment, narrow scope, limit of the form of detected squats and the squats degree etc. To compensate for the lacking of current research methods, this thesis provides a new detection method of squats which based on the operating vehicles. Specifically, collecting bogie vibration acceleration (BA) as the original detection signal, then the signature tunes of squats are extracted from the time-frequency analysis results of test signal, finally the early detection of rail squats damages are done on the basis of signature tunes of squats.In order to extract the signature tunes of squats, firstly, an operating vehicle-track dynamic model is built based on the multi-body dynamics software-SIMPACK, and the vertical-longitudinal profile of squats that to be simulated on the rail are variously. The simulation signals from the normal rail and squats are studied by the time-frequency domain analysis. The result can proof that the time domain waveform and the signature frequency of the power spectrum power spectral density based a squat rail are quite different from the common signals. Then wavelet packet analysis and HHT of simulation signals are discussed. And these two analysis methods are found to be the most suitable for investigation of information in spatial domain and local frequency characteristics of BA, both0-500Hz and300Hz are considered as the unique vibration frequency caused by squats damages. In addition, this thesis also proposes a W-shaped natural vibration mode through HHT. Finally, the above analysis methods are applied to the actual rail irregularity data of Shanghai metro line1, the result can proof the feasibility and reliability of the method which based on the operating vehicles to do detection for the squats damages.As follows is idiographic research of this thesis. (1) Introduce the main characteristics of squats, including the initiation and growth of squats, as well as a squat fitting contour model.(2) Establish an operating vehicle-track dynamic model based on the multi-body dynamics software-SIMPACK. The complete simulation model is consist of the following parts, operating vehicle dynamics model, wheel/rail contact model, elastic rail model, and squats with varied vertical-longitudinal simulation model.(3) Represent the theoretical knowledge of time and frequency domain analysis method. And apply this method to rail squats analysis, analyzing and handling the various squats signals based on simulation. Conclude the time and frequency domain features of each failure mode, and the results have proved the feasibility and affectivity of this method.(4) Describe the theoretical knowledge of wavelet packet as well as HHT, and research its application in detection of early squats. Analyze and compare the normal and fault signal, and then get the characteristic of fault signal.(5) Analyze the actual rail irregularity data of Shanghai metro line1with the above analysis methods, the result can proof the feasibility and reliability of the method applied in detection for the squats damages based on the operating vehicles.