MACHINE LEARNING APPROACHES FOR PREDICTING REMAININGUSEFUL LIFE OF ELECTRIC VEHICLE BATTERIES
Keywords:
electric vehicles, lithium-ion battery, remaining useful life, battery degradation, machine learning, predictive diagnostics, battery management system.Abstract
The rapid growth of electric vehicles (EVs) has increased the
importance of battery reliability, safety, and operational efficiency. Battery degradation directly affects driving range, charging performance, and the overall economic effectiveness of EV systems. Therefore, predicting the Remaining Useful Life (RUL) of lithium-ion batteries has become one of the key challenges in battery management systems (BMS). This paper investigates machine learning approaches for forecasting the remaining service life of electric vehicle batteries using operational and diagnostic data. Different predictive techniques, including Artificial Neural Networks (ANN), Support Vector Machines (SVM), Random Forest algorithms, and Long Short-Term Memory (LSTM) networks, are analyzed and compared. The study also examines the impact of temperature, charging cycles, depth of discharge, and current load on battery aging. In addition, the paper discusses the role of intelligent software solutions in improving prediction accuracy and supporting preventive maintenance strategies.
