Real time fault detection in railway tracks using Fast Fourier Transformation and Discrete Wavelet Transformation

Abstract

Railway transportation system is a safety critical system, a small fault in a train, track or train operation control system may cause serious hazards. Therefore, each of them should be carefully investigated for safe running of the trains. Railway tracks are deployed over a long span in diverse geographical conditions hence more vulnerable to the faults. Therefore, real time automatic track monitoring system is needed to timely investigate the tracks with precision without hurdling the normal train traffic. The main goal of this paper is to compare the performance of two mathematical methods (Fast Fourier Transformation and Discrete Wavelet Transformation) widely used to detect faults on railway tracks. The accelerator sensors are deployed on the axle-box of service trains to measure the acceleration of the vibrations produced by the running train. Both methods are used to process the acceleration measurements to estimate the track faults specifically, cracks and corrugations. The effectiveness of both methods is compared to determine the best implementation in a real time scenario. In real time scenario, the data is processed, as soon as it is collected from the accelerator sensors while train is running, rather than using previously collected data of the track. It was observed from the simulation results that using Fast Fourier Transformation, 100% of corrugations and 90.53% of cracks were detected, while using Discrete Wavelet Transformation, 99.33% of corrugations and 99.85% of cracks were detected. © 2021, Bharati Vidyapeeth's Institute of Computer Applications and Management.