Principles of the framework development of a physics-based rail vehicle digital twin in innovative decision support and prediction complex systems

While historical data-driven approaches and machine learning models are prevalent in the railway industry, a physics-based modelling approach offers advantages for digital twins (DTs) with well-defined functions and purposes. This paper focuses on the introduction of a systematic approach to developing such DTs, encompassing their entire lifecycle and reflecting the development process within the Internet of Things (IoT) context. The main purpose of this paper is to present a technique for the development of a railway vehicle multibody model with appropriate interfaces that allow the achievement of real-time performance, while also maintaining acceptable accuracy, and further implementation in innovative decision support and prediction systems. Furthermore, this paper describes in detail the design of the framework for the proposed DT concept and focuses on a verification case for a DT real-time component built based on numerical simulation techniques. Finally, the implementation challenges and limitations are summarised and discussed.