Battery State of Health Estimation using a Novel Regression Framework

Abstract

Electrochemical cells and their capacity to retain charge is fundamental in electric transportation. As cells undergoes use, they can hold lesser amount of charge as they age and degrade slowly. In this paper we present a novel machine learning/regression framework to estimate the state of health of a cell and remaining capacity at any time. We compute a partial capacity value for a standard battery dataset, and then build a machine learning based regression model.