A reliability-based prognostics framework for railway track management

Abstract

Railway track geometry deterioration due to traffic loadingis a complex problem with important implications in costand safety. Without appropriate maintenance, track deteriorationcan lead to severe speed restrictions or disruptions, andin extreme cases, to train derailment. This paper proposesa physics-based reliability-based prognostics framework asa paradigm shift to approach the problem of railway trackmanagement. As key contribution, a geo-mechanical elastoplasticmodel for cyclic ballast settlement is adopted and embeddedinto a particle filtering algorithm for sequential stateestimation and RUL prediction. The suitability of the proposedmethodology is investigated and discussed through acase study using published data taken from a laboratory simulationof train loading and tamping on ballast carried out atthe University of Nottingham (UK).