Prognosis of engine failure based on modelling by using live parameter data from vehicle.

Predicting engine failure upfront to prevent major damages is a challenging task. But with the data available from the ECU (engine control unit) transmitted live via the telematics unit provides a data base to predict the health of engine and its major sub system. Through proper data analytics using the logics arrived based on previous / tested data the possibility of failure occurrence and the corrective action required to avoid the same can be indicated. This paper aims at an application with Modelled data analytics to analyze the live parameter data and apply predefined logics to predict engine / sub system level health and possible occurrence of a failure. Vehicle data was used to predict the engine / sub system failures. Application using PHYTON was developed to evaluate the severity of application duty cycle on engine and also to evaluate the health of major sub system. The logic and modelling was verified on field vehicle data and the application was able to predict the failure accurately. The severity was mapped in I-Alert application to enable fleet owners understand the engine running criticality. In future, more field cases to be evaluated and accuracy of prediction to be improved.