Spatial–Temporal Model to Identify the Deformation of Underlying High-Speed Railway Infrastructure

Abstract

Railway track geometry is generally understood to be influenced by the deformation of the track infrastructure. This study developed a spatial–temporal identification model for the deformation of the underlying high-speed railway infrastructure, including simply supported beams and track slabs based upon track geometry data collected between 2016 and 2019. To achieve this, we first preprocessed the data, including data collection and cleaning. Next, we developed a track irregularity degradation indicator (TIDI) for different track infrastructures using wavelet coefficients. Then, we combined the TIDIs of 40 inspection runs over 3 years to obtain the TIDI distribution matrixes in the spatial and temporal domains for different track infrastructures. In the spatial domain, we extracted the most probable abnormal position of the track slab using kernel density estimation. In the temporal domain, we developed a logarithmic–linear regression model and noted that the growth trend of the TIDI of the abnormal bridge gradually slows with time.