Hybrid energy-based model for Li-ion battery degradation using real residential operating data in PV-battery systems

Battery Energy Storage Systems are playing a central role in the energy transition, thanks to their ability to integrate non-programmable renewable sources and ensure stability and flexibility in electricity grids. However, repeated charge and discharge cycles cause degradation processes that reduce the available capacity and useful life of the system. Understanding and predicting the evolution of the State-of-Health is therefore essential to optimising their use and planning their replacement. Several degradation models have been proposed in the scientific literature, generally based on controlled cyclic tests or electrochemical parameters that are not always accessible during normal operation. Conversely, this study presents an energy-based degradation model derived from experimental data collected from batteries operating under actual residential load conditions. The model uses charge and discharge power, cumulative energy, and State-of-Charge as input variables. This approach allows the evolution of State-of-Health to be accurately estimated, obtaining an R2 value of 96.55% and a Root-Mean-Square-Error of 0.0015 from comparison with real residential data. The model is therefore an effective tool for predictive monitoring and intelligent management of Battery Energy Storage Systems, promoting greater reliability and sustainability in the context of the energy transition.