Predictive Maintenance Based On Machine Learning In Public Transportation Vehicles

Predictive maintenance is an approach to prevent failure in a system by estimating the time of failure before a mechanical component fails, so that the maintenance decision can be properly planned. Maintenance forecasting models are used to increase the productivity and efficiency of a hardware. In the public transport sector, whose efficiency is heavily dependent on equipment, forecasting of failures is vital. In this study, predictive maintenance work was carried out in order to minimize the problems such as stopping the voyage, delaying the journey and having an accident caused by unplanned breakdowns in public transport vehicles. Based on instant vehicle health data obtained from IoT sensors, classification techniques were run in machine learning. For maintenance planning, the probability of vehicles being normal and malfunctioning was examined with fuzzy logic and fuzzy outputs were obtained at maintenance speed. As a result of the study, almost all of the faults in the vehicles could be detected with the predictive maintenance approach applied.