Virtual Axle Detector: Train Axle Localization based on Bridge Vibrations

Abstract

Abstract Infrastructure worldwide is facing the challenge of aging bridges and increasing traffic loads. Prolonged serviceability and safety of these structures can be enabled by Structural Health Monitoring (SHM) methods. Knowledge of the actual operating loads is critical for evaluation of the remaining service life. However, direct measurement of the loads is challenging and requires a significant financial investment. Bridge Weigh‐In‐Motion (BWIM) methods use the structural response of bridge structures to determine loads, but generally rely on accurate knowledge of the position of loads as a function of time. Positions can be determined using conventional axle detectors, but their lifetime is limited, and their installation is expensive. To avoid these problems, we propose an improved Virtual Axle Detector (VAD) with Enhanced Receptive field (VADER) that can detect axles for all bridge types using accelerometers that can be placed anywhere along the bridge. The same data set with 3787 train passages recorded on a steel trough railway bridge under real operating conditions was used. Our results show that, in comparison with VAD, VADER reduces the number of undetected axles by over 79% and detects 99.5% of axles with an average spatial accuracy of 4.6 cm.