A Big Data Architecture for Digital Twin Creation of Railway Signals Based on Synthetic Data

Abstract

Industry 5.0 has introduced new possibilities for defining key features of the factories of the future. This trend has transformed traditional industrial production by exploiting Digital Twin (DT) models as virtual representations of physical manufacturing assets. In the railway industry, Digital Twin models offer significant benefits by enabling anticipation of developments in rail systems and subsystems, providing insight into the future performance of physical assets, and allowing testing and prototyping solutions prior to implementation. This paper presents our approach for creating a Digital Twin model in the railway domain. We particularly emphasize the critical role of Big Data in supporting decision-making for railway companies and the importance of data in creating virtual representations of physical objects in railway systems. Our results show that the Digital Twin model of railway switch points, based on synthetic data, accurately represents the behavior of physical railway switches in terms of data points.

Keywords

• Big data
• digital twin
• machine learning
• synthetic data
• railway industry
• artificial intelligence