A Novel Framework for Estimating Battery Health Indicator Under Different Operating Conditions

With the increasing demand for reliable battery systems, accurate State-of-Health (SOH) estimation has become a key challenge. Traditional direct HIs based on Capacity and Internal Resistance (IR) measurements face limitations in terms of applicability in real-case scenarios. Thus, in existing literature, alternative approaches have been proposed, including the usage of impedance and electrical and thermal features extracted during charge and discharge processes. However, the quality of the HIs obtained through these methods still relies on fixed and controlled operating conditions, lacking the flexibility required for realistic applications. Trying to fill this gap, in this paper, a novel framework for the extraction of a robust HI is discussed. In particular, this work utilizes 12 features derived from simple electrical, temperature and time measurements directly taken during the cycle life of the battery, including alternative metrics less influenced by the operating conditions. Furthermore, the paper introduces a low computation and innovative framework for the feature selection and the dimensional reduction necessary for obtaining a one-dimensional HI feasible for SOH assessment and prognostic estimation. The approach proposed has been validated on a dataset consisting of several different operating conditions. The dataset has been collected by planning a split-plot experimental design. The suggested approach is a robust process across the considered working conditions showing promising performances compared to traditional HI extraction techniques.