Data-driven track geometry defects localization and strategies for preventive maintenance: a case study

Traffic and operational factors, such as axle loads, train speed, and cumulative tonnage, cause deviations in the track from the original smooth geometry and deterioration of the functional conditions of its components—rail, fasteners, sleepers, and ballast. Therefore, railway track maintenance and renewal are indispensable tasks for ensuring railway safety, travel comfort, and train punctuality. Carrying out cost-effective maintenance is pivotal in managing railway infrastructure assets. To this aim, predicting track degradation is extremely important due to its beneficial effects in scheduling preventive maintenance. This study proposes the development of a data-driven methodology for predicting isolated geometry defects based on the data obtained from periodic measurements on the track. Assuming geometric parameters such as gauge, longitudinal level, transversal level, alignment, and twist as explanatory variables, the influence of traffic, speed, line conditions, and previous maintenance interventions on the degradation process is assessed through statistical elaboration. The proposed methodology is applied to a section of the Salerno-Reggio Calabria railway line. From data elaboration, decision support models for improving the application of railway track maintenance planning and scheduling can be developed. This approach also makes it possible to intervene at well-identified critical points before the formation of major defects.