Assessment of railway transition zone settlements using InSAR – A comparison between Sentinel-1, TerraSAR-X and track recording car data

Abstract

Satellite-based remote sensing tools such as Synthetic Aperture Radar (SAR) Interferometry (InSAR) has emerged as a potential tool for condition monitoring of railway track and infrastructure. This is due to its frequent and stable collection of data without requiring personnel or capacity occupation on the railway track. SAR is an active radar, meaning that it can collect images during cloud coverage and do not rely on sunlight. It collects scattered reflections as complex images with pixel intensity and phase. Current satellites instrumented with SAR have a passing frequency over a location on the earth at best 6 days. Interferometry exploits the phase difference between two or more SAR images collected at either different positions or times, which effectively can give information about the relative difference in distance between the images. Differential InSAR (D-InSAR) considers SAR images of the same location at different times, resulting in measurements of ground motion in the satellite line of sight. The system design of the radar used for SAR has implications on the spatial resolution of the earth surface measurement, which is generally inversely proportional to spatial coverage. The aim of this paper is to assess the applicability of Sentinel-1 and TerraSAR-X data for track irregularity monitoring considering their different system design of the radar. Sentinel-1 has wider coverage and lower resolution compared to TerraSAR-X. In contrast, Sentinel-1 satellite data is open to global users. Exploring its possibilities and limitations for railway maintenance supports the development of more robust and cost-efficient maintenance decisions. The case study for this assessment is a transition zone between railway bridge and ballasted track in Sweden. Transition zones are interesting objects of study from a condition monitoring point of view as they often exhibit differential settlements and therefore worse track geometry. This is because of differences in settlement resistance for the different types of track structure that meet in the transition zone. The reference geometry data for the transition zone consists of measured track irregularities from chord-based track monitoring vehicles which are used to assess railway longitudinal levels (vertical track irregularities) at three different wavelength ranges, D1 = 3-25 m, D2 = 25-70m, and D3 = 70-150m. Shorter wavelength longitudinal level variations is associated with safety levels and is the basis for track maintenance in Sweden, whereas the longer wavelengths are associated with ride comfort.