Method for predicting dynamic loads for a health monitoring system for subway tracks

Abstract

This paper presents a method for processing acceleration data registered on a train and determining the health condition of a subway’s substructure. The acceleration data was converted into a dynamic deformation by applying a transfer function defined using the Empirical Mode Decomposition Method. The transfer function was constructed using data produced on an experimental rig, and it was scaled to an existing subway system. The equivalent deformation improved the analysis of the dynamic loads that affect the substructure of the subway tracks because it is considered the primary load that acts on the track and substructure. The acceleration data and the estimated deformations were analyzed with the Continues Wavelet Transform. The equivalent deformation data facilitated the application of a health monitoring system and simplified the development of predictive maintenance programs for the subway or railroad operators. This method better identified cracks in the substructure than using the acceleration data.