Correlating Ballast Volume Deficit with the Development of Track Geometry Exceptions Utilizing Data Science Algorithm

This paper introduces the application of a data science algorithm in analyzing Big Data related to the volume of missing ballast in a track section and the development of track geometry defects. The data intensive algorithms are necessary in order to effectively analyze over 100,000 segments of track representing nearly 1000 miles (1700 km) of data containing over 23,000 track geometry exceptions or defects. The specific analysis tool used was logistic regression, which allowed for modeling a dichotomous output, and this tool was implemented with encouraging results. Missing ballast data was obtained from a hy-rail-mounted LIDAR-based ballast profile measurement system and correlated to track geometry defects that developed along the inspected track locations on a major US class I railroad. Analyses were conducted to provide a determination of the proportion of segments that develop geometry exceptions as a function of missing ballast volume. Parameter studies were performed for curvature and annual millions of gross tons (MGT) as well. The logistic regression analysis was then performed and the resulting model was used to calculate the probability that a given track segment will contain or develop track geometry exceptions. The relationship between an increase in missing or deficient ballast and increased probability of developing a track geometry defect was developed and confirmed.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01655334
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Dec 27 2017 10:30AM