A Data-Driven Approach for Predicting Vehicle Component Failures

The automotive industry has faced several issues with respect to sudden failures of vehicle components, which results in reduced efficiency of the vehicles, breakdowns, and increased maintenance costs. To this end, this research suggests a new approach to predictive maintenance based on machine learning to predict the failures of critical vehicle components. The proposed system is developed using historical data containing operational parameters like engine RPM, lubricant oil pressure, coolant temperature, and fuel pressure to develop prediction models. The analysis of the correlation between these parameters and their impact on the component was made in detail. Several machine learning algorithms, including Logistic Regression, Decision Tree, Random Forest, Support Vector Machine (SVM), and K-Nearest Neighbours (KNN), were also compared. Among the above mentioned, the best result was achieved by Logistic Regression with the accuracy of 66%, and the second best result was obtained by using SVM with the accuracy of 92<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>. The predictive framework is also supported by a web-based dashboard that gives the user real time information on the state of the vehicle, the probability of failure and sending of maintenance reminders at the right time. This dashboard helps the fleet managers in making informed decisions as it provides suggestions based on the output of the model. Thus, the developed system brings down the occurrence of unplanned downtime, improves the maintenance schedule, and increases the dependability of vehicles and thus can have been an effectively large used number in of industries vehicles. that the use of machine learning in predictive maintenance is also introduced and the way for future also development suggested. such This as scalable real and time adaptable IoT, solution advanced is deep a learning valuable and contribution edge to computing the for growing failure field prediction of is automotive predictive maintenance and the potential to revolutionize fleet management.