Machine Learning Applied to Battery Prognostics based on Advanced State of Health Estimation

Lithium Ion batteries are very attractive technology in the field of electric energy-storage devices with high energy and power densities. Knowledge of lithium-ion lifetime and aging acceleration factors is a crucial aspect for fruitful market introduction, including automotive and embedded applications. The degradation of batteries through various mechanisms over time is due to several factors, such as: temperature, voltage, time for calendar aging, Depth Of Discharge (DOD), State Of Charge (SOC), range and current rate. That’s why it is extremely important to get an accurate prediction of the factors that cause battery degradation. This paper analyzes the extent to which these factors cause battery degradation. Our approach has been tested on the Massachusetts Institute of Technology (MIT) datasets, and the experimental results show the effectiveness of the proposed approach.