Optical sensing and self-learning approach to estimate the state condition of railway infrastructure sublayers

Abstract

The objective pursued is the implementation of a technique for intensive information capture of the state of the deep layers of the infrastructure of transport linear works, in order to (a) detect faults, (b) predict the evolution of the state as a function of time, (c) estimate the necessary maintenance operations, and (d) plan the required interventions. All this is focused on achieving greater efficiency in the management of the conservation of this kind of infrastructure. The data obtained in real time correspond to the tests carried out on a substructure section model housed in a scaled sublayer’s test-rig with installed Fiber Bragg Grating (FBG) optical sensors. Finally, the methodology is described, based on data analytics and Machine Learning techniques, in order to infer the severity of measured deformations and failures.