Toward Smart Railway Infrastructure Predictive and Optimised Maintenance through Digital Twin (DT) System

Railway systems are a critical infrastructure that requires continuous maintenance to ensure safety, reliability, and cost efficiency. Traditional maintenance strategies, such as scheduled or corrective maintenance, often lead to unnecessary costs, unplanned failures, and limited use of real-time condition information. With advances in digitalisation and smart sensing, Digital Twins (DTs) have emerged as a promising approach for predictive and optimised railway maintenance. A DT provides a virtual representation of physical railway assets, supported by live sensor data, data-driven analytics, and cyber-physical integration to monitor asset conditions, predict faults, and support maintenance decision-making. This study presents a systematic review of DT applications in railway engineering, based on a structured and transparent screening process that narrowed a broad body of literature to 34 representative peer-reviewed studies published between 2020 and 2025. These studies were selected because they explicitly address DT-enabled monitoring, analysis, and maintenance decision support in railway systems. The reviewed works span track, rolling stock, bridge, and communication domains, highlighting current technologies, implementation trends, and research gaps. Overall, the literature demonstrates that DT-based methods have strong potential to improve maintenance efficiency and system reliability. Reported benefits include improved fault detection, better maintenance planning, and reduced manual inspection effort. However, several critical challenges remain. Common gaps include a lack of standardisation and interoperability, limited real-time operation and scalability, weak data governance and cybersecurity, and shortages of skilled personnel for effective DT deployment. In addition, most studies remain conceptual or pilot-scale, providing limited quantitative or long-term validation under real operating conditions.