Digital twins for dynamics of a train-slab track-bridge coupled system

Abstract

In consideration of significant uncertainties arising from the degradation of in-service train-slab track-bridge coupled systems and the limitation in accurately reproducing track irregularities using prescribed track spectra, this study presents a novel framework to establish a digital twin for dynamics of an in-service train-slab track-bridge coupled system to best simulate and predict its dynamic behavior and performance during its operation. The train-slab track-bridge coupled system of a railway test line is taken as a physical entity and subjected to field measurements. The design-document-based virtual entity (numerical model) of the train-slab track-bridge coupled system is then established. A model updating procedure is subsequently proposed for the virtual entity based on the dynamic characteristics identified from the physical entity, and a track irregularity spectrum recognition method is developed in terms of the measured dynamic responses and optimization algorithm, thereby leading to the digital twin establishment. The established digital twin is finally used to simulate and predict the coupled vibration of the train-slab track-bridge system and compare with the measurement results. The results demonstrate the feasibility and accuracy of the digital twin and its prediction capability.