Novel Statistical Analysis Approach for Remaining Useful Life Prediction of Lithium-Ion Battery

Abstract

In modern times, every device runs on electricity. Batteries are fulfilling electricity needs for portable devices. In all the latest portable electronic devices, lithium-ion batteries (LIBs) are the primary source of power: the reason being their high charge storage density. The primary issue with the LIB is the lack of information on its remaining useful life (RUL). Thus, we try to analyze and predict the RUL of LIB in this paper. We present a novel statistical approach for extracting various critical points from the discharge cycle by inspecting the LIB dataset. The inspection of LIB assisted in finding the pattern and relationship with the number of cycles. Thus, two feature sets are proposed in this paper based on critical points. The feature set is examined with various machine learning regression models using the LIB dataset. Also, the comparative analysis is performed for the proposed merging parameter to find the decent performance.