Prospective approaches to predictive modeling of degradation processes of track superstructure elements and its application in creating digital twins

Abstract

It’s impossible to use digital technologies without using the amount of information coming from various systems designed to manage the transportation process and plan its work, taking into account modern economic requirements and resource constraints. Digital twins are currently the most promising tool for solving the problems of managing technically rich multi-level assets, which include railway transport. The track facilities are one of the most expensive assets, and the issues of organizing the management of the maintenance of the railway track are very acute, since they are directly related to the safety of train traffic, therefore, the development of a digital twin of the railway track is a priority task for track science. A digital twin of a railway track should contain elements of BigData technology in the form of arrays of diagnostic data coming online from mobile and stationary diagnostic tools, an array of passport data about the track device, as well as a set of models that can convert this data into matrices “state — action”, suitable for making organizational and technical decisions on the management of the track complex, starting from the level of linear enterprises and ending with network tasks. The article presents models that can be taken as a foundation for building digital twins of a railway track. The results of verification and approbation of the proposed models in the “Neyroekspert-Put’” software package are also presented.

Keywords

• digital twin
• railway track
• track superstructure
• railway track degradation forecasting