The components of railway rolling stock are integral parts of the railway "system". The major challenge for operators and manufacturers is the availability of trains to increase user satisfaction without increasing hardware costs and meet the requirements of regularity of organizing authorities. By developing innovative tools for the diagnosis and maintenance of embedded subsystems, the operator is able to plan maintenance operations (and therefore to optimize the availability and quality of network service), the logistics and parts supply.
The objective of this project was to develop diagnostic and prognostic tools of three important parts of railway rolling stock (air conditioning, brakes and doors) that will give their real potential for use in real time before failure, thus introducing improved conditions for carrying out maintenance and traffic regularity. It will provide more detailed knowledge of the time behaviour and aging of material as well as some answers for optimizing maintenance standards.
The first component of this project concerned the dynamic diagnosis by pattern recognition.
The second part was devoted to forecast preventive maintenance (or prognosis).
Specific algorithms and a tailored model obtained provided an initial estimate of the residual life span and updates it with each new observation of the system operating state, taking into account the possible existence of several dynamics in the degradation process.