Multi-scale analysis of the mapping relationship between turnout irregularities and vehicle responses based on cross-wavelet transform

Abstract

Track irregularities significantly impact the dynamic performance of vehicles passing through turnouts. Despite this, the precise mapping relationship between these irregularities and vehicle responses remains unclear. Utilizing measured data, the inherent irregularities of turnouts were extracted using improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN). The mapping relationship between the wavelength, spatial position, and time lag effect of turnout inherent irregularities and vehicle responses was analyzed using cross-wavelet transform (CWT). The first-order and second-order intrinsic modal components within the original irregularities distinctly reflect the inherent characteristics of the switch and crossing panels. The inherent irregularities in these panels respectively correspond to the vertical and lateral accelerations of the vehicle within the wavelength ranges of 2∼4 m and 4∼8 m. Within the switch panel, the longitudinal levels of the right rail and the twist exert the most significant influence on the vehicle's vertical acceleration, accounting for 17.80% and 22.89% respectively. In the crossing panel, the track gauge and the alignments of the right rail have the greatest impact on the vehicle's lateral acceleration, accounting for 24.72% and 35.18% respectively.