Algorithm development to classify the deterioration of insulated rail joints by means of onboard measurements

Insulated rail joints are fundamental components of some railway lines due to their role in the signalling systems. On the other hand, being highly stressed components, their monitoring is a fundamental activity to guarantee the safety of running trains. Currently, the monitoring of these elements is carried out directly by trained personnel. It follows that the control of a railway line takes long time and there are no objective parameters that guarantee the health condition of a joint. Considering this context, this work shows the preliminary results of the analysis of experimental data through an algorithm developed to detect and classify the glued insulated rail joints by means of accelerometers placed on two wheelsets of a train.