Publication

Abstract : Electric Vehicles (EVs) are transforming the transport sector through efficiency and sustainability. Still, precise calculation of their component lifetimes, particularly battery life, is key to reliability and safety. In this research work, the prediction of the Remaining Useful Life (RUL) of EV components through advanced Machine Learning (ML) methods is emphasized. Predictive modeling of RUL facilitates advanced maintenance and minimizes the likelihood of catastrophic breakdowns. The approach involved the training and testing of various ML algorithms such as Linear Regression (LR), Decision Tree (DT), Random Forest (RF), Extra Trees (ET), Extreme Gradient Boost (XGB), Gradient Boosting (GB), and K-Nearest Neighbors (KNN). Feature engineering and normalization were used for boosting model precision. Models were tested on parameters such as Coefficient of Determination (R2), Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The ET model proved to be the best performer with Test R2 at 0.9999, Test MSE at 11.1679, and Test MAE at 1.7831. Second-best performers were RF and XGB, which also had high accuracy and generalizability. The experiments verify that ensemble-based algorithms considerably outperform linear models for RUL prediction. The proposed method offers a robust and scalable solution for real-time RUL estimation in EV systems, facilitating predictive maintenance policies that enhance operational efficiency, lower costs, and prolong component life.