Buckled-track derailments: Long-term trends and implications for rail neutral temperature management

Train derailments are the most common type of mainline railroad accident in the United States and often result in substantial property damage and operational disruptions. Among track-related derailment causes, buckled track and broken rails or welds present the highest risk. Buckled-track derailments result from compressive thermal stresses, whereas broken-rail derailments are most often associated with metallurgical defects or fatigue cycles, and are not necessarily attributed to tensile stress. This study focuses on buckled-track derailments due to their direct relation to rail thermal behavior and their challenging detection, as they rarely interrupt signal circuits. Buckled-track derailment risk was analyzed using the Federal Railroad Administration's (FRA) accident database and track classification system that is based on maximum allowable train speed and minimum geometric conditions. Buckled-track derailment risk was studied from 2000 to 2024 across FRA Track Classes. Additionally, ambient and rail temperatures at and near the time of buckled-track derailments were analyzed. Results indicated that buckled-track derailment risk has decreased in recent years. However, the proportion of buckled-track derailments relative to all track-caused derailments has increased. Analysis of temperature trends revealed that 62 % of buckled-track derailments on Class I mainlines and sidings between 2011 and 2024 occurred at rail temperatures higher than 90 % of all days that year, demonstrating a strong association between extreme heat and derailments. Furthermore, comparison of rail temperature with design rail neutral temperature (RNT) revealed that derailments likely did not occur at the railroad's desired RNT. This finding suggests that RNT may decrease over time or may not be properly set when rail was installed or adjusted, underscoring the need for improved continuous welded rail (CWR) stress management practices or monitoring methods.