A Novel Approach for Predicting Remaining Useful Life and Capacity Fade in Lithium-Ion Batteries Using Hybrid Machine Learning
Sadiqa Jafari, Yung-Cheol Byun, Seokjun Ko
Abstract
Lithium-ion batteries (LIBs) Remaining Useful Life (RUL) prediction is vital for Battery Management Systems (BMS). It is crucial for providing the optimum performance and longevity of batteries used in different industries. In this study, we propose an innovative approach that combines machine learning techniques and hybrid modeling strategies to enhance the accuracy and robustness of battery analysis. By leveraging the power of k-Nearest Neighbors (kNN), Random Forest (RF), and XGBoost algorithms, our proposed model effectively captures complex relationships and patterns in battery data. We meticulously curate a comprehensive dataset comprising essential battery parameters: capacity, voltage, cycle, and temperature. Through rigorous experimentation and evaluation, our proposed method outperforms existing approaches in predicting RUL and capacity fade. The insights gained from our analysis offer valuable guidance for optimizing battery performance, enabling informed maintenance planning, and improving energy storage system efficiency.