Rail surface damage occurs due to repeated overstressing of the surface by wheel-rail contact cycles, which is also known as rolling contact fatigue, and causes regular costly maintenance expenses to railways. The preventive strategies to reduce rail damage can be conservative if it is not based on measured data. A validated digital model of the physical system can provide data for maintenance decision. Interaction between a physical system and a digital model has evolved at the beginning of this century into a new process commonly referred to by the term ‘Digital Twin’. There is a lack of a Digital Twin framework in decision-making on railway maintenance. The term Digital Twin is often confused with advanced simulation methods where the communication of data is only one way, i.e. sensor or design data of the physical system input to a model. In this paper, a Digital Twin has been developed to predict rolling contact fatigue rail surface damage, corresponding to traffic passing along a railway network. Differences between the Digital Twin and advanced simulation techniques have been presented. Finally, a predictive assessment tool has been proposed to provide prediction from the Digital Twin to the physical twin, and results are discussed. Abbreviations: AS: Advanced Simulations; DS: Digital Shadow; DT: Digital Twin; HPC: High Performance Computing; LTS: Longitudinal Train Simulation; MBS: Multibody Simulation; MGT: Million Gross Tonnes; NASA: National Aeronautics and Space Administration; RCF: Rolling Contact Fatigue
Read full abstract