Reliability analysis of railway assets considering the impact of geographical and climatic properties

Abstract

Various factors, including climate change and geographical features, contribute to the deterioration of railway infrastructures over time. The impacts of climate change have caused significant damage to critical components, particularly switch and crossing (S&C) elements in the railway network. These components are sensitive to abnormal temperatures, snow and ice, and flooding, making them susceptible to failures. The consequences of S&C failures can have a detrimental effect on the reliability and safety of the entire railway network.It is crucial to have a reliable clustering of railway infrastructure assets based on various climate zones to make informed decisions for railway network operation and maintenance in the face of current and future climate scenarios. This study employs machine learning models to categorize S&Cs; therefore, historical maintenance data, asset registry information, inspection data, and weather data are leveraged to identify patterns and cluster failures. The analysis reveals four distinct clusters based on climatic patterns. The effectiveness of the proposed model is validated using S&C data from the Swedish railway network.By utilizing this clustering approach, the whole of Sweden railway network divided into 4 various groups. Utilizing this groups the development of model can associated with enhancing certainty of decision-making in railway operation and maintenance management. It provides a means to reduce uncertainty in model building, supporting robust and reliable decision-making. Additionally, this categorization supports infrastructure managers in implementing climate adaptation actions and maintenance activities management, ultimately contributing to developing a more resilient transport infrastructure.