Network-Level Risks of Extreme Temperature Weather on Rail Infrastructure: A Graph-Based Probability-Impact Risk Assessment Framework

Abstract The US rail network is a large and crucial element of the country’s infrastructure and economy, with approximately 225,308 km (140,000 mi) of track. Extreme temperature events can cause rail damage and increase the risk of derailments, putting lives at risk and causing huge economic damage. High rail temperatures during heatwaves can lead to heat-induced buckling, whereas freezing temperatures can create cracking. Previous research has investigated the effects of temperature on rail failure and the probability of their failure using various approaches such as finite-element analysis, lab experiments, machine learning, and others. However, there is a need for research that investigates the higher level impact of extreme weather events on the large rail network and its operations. Accordingly, this paper addresses that knowledge gap and proposes a framework to investigate the vulnerability of the rail network using a weather-informed graph-based probability-impact risk assessment. A graph of the US rail network was created and embedded with weather data simulating historic extreme weather events; and a probability-impact risk assessment was conducted based on probability (i.e., what is the probability that a rail segment will fail), impact (i.e., what is the impact if a rail segment fails), and combined risk severity. The model can quantify the risk severity on all railroads connecting any two nodes, such as any two major cities, on the US rail network. Results of a case study revealed four high-risk regions. This paper proposes a developed framework to support decision making as related to improving the resilience of existing roads and planning new alternative routes to minimize the impact of extreme weather events and improve rail safety.