Towards predictive maintenance and management in rail sector: A clustering approach

Abstract

It is accepted beyond doubt that the back bone of mass transportation in India is Railways and it is regarded as the lifeline of the nation. Indian Railways is using various information technology based systems to maintain and manage its infrastructural resources, leading to generation and accumulation of voluminous data. The data captured is expected to possess hidden meaning and trend which, if properly explored and modelled, can aid in implementing the new paradigm called Predictive Maintenance and Management. The thrust to usher in the new paradigm amply manifests in its Vision 2020 statement, too. The paper aims to showcase a methodology using Clustering, a Data Mining Technique that can bring in a maintenance plan built upon the predictive behaviour exhibited by failure data. The model outlined here has been tested on actual failure data pertaining to passenger carrying vehicles of trains.