Probabilistic State of Health and Remaining Useful Life Prediction for Li-ion Batteries

Abstract

Lithium-ion batteries are often operated to reach excellence in technical and economical performance, but they rapidly degrade as a consequence of charge/discharge profiles. Maintaining the knowledge of the actual capacity of the battery is mandatory to pursue the objectives without incurring into unexpected constraints. This paper addresses battery prognostic from the viewpoint of probabilistic prediction of the State of Health (SoH) and of the Remaining Useful Life (RUL) of the batteries. Two probabilistic models based on time series and quantile regression, each developed in a different framework, are developed and compared for this purpose. They are specifically suited up to exploit data coming from Accelerated Degradation Tests (ADTs). Moreover, a dedicated procedure to extract a single, point value from the probabilistic predictions is presented to let the models work also in deterministic scenarios. Numerical experiments conducted on actual public data confirm the validity of the proposal, within a rigorous comparison with relevant benchmarks taken from the literature on the topic.