Abstract
The railroad track system and its components are a critical transportation asset that is responsible for transmitting rolling stock wheel loads to the roadbed. To ensure safe and efficient operations, railroads perform frequent track inspections, some of which generate substantial amounts of track health data. Furthermore, with the rise of data science tools and methods, the potential for these data to move maintenance and safety towards more robust analysis is fostered. Recently, railroads around the world have increased their use of data trending for geometry prediction but they do not always cover how the boundary conditions are also changing over time. This study presents an evaluation of the relationship between the change of track geometry condition and ballast profile on both curved and tangent track segments. The stochastic approach proved to be a valid comparative method for the existing and emerging datasets. Track geometry profile degradation was shown to have significant correlation with both the Ballast Health Index (BHI) and initial value of profile. Additionally, profile degradation was found to be more accelerated in regions with poorer initial geometry profiles and higher initial BHI values. Findings have the potential to improve maintenance effectiveness and prioritization and provide a method for quantifying track degradation rates under different operating and maintenance conditions.


