Predicting remaining useful life of lithium-ion batteries: A review of degradation mechanisms and open-source data availability

Lithium-ion batteries are vital for large-scale industries, especially in transport and renewable energy applications, due to their high energy density, extended cycle life, and low self-discharge rate as compared to other battery types. With increasing demand for sustainable and energy-efficient solutions, it is critical to study, understand, and improve the performance of batteries. Monitoring degradation is important, but acquiring real-time sensor data over lengthy periods is challenging due to the longer life cycles of batteries. This makes data collection costly and time-consuming. Cell-based open-source datasets provide a viable alternative, allowing researchers to estimate the degradation of battery cells without the requirement for constant, real-time testing. Furthermore, estimating degradation factors is crucial for forecasting Remaining Useful Life and extending battery lifespan. Methods such as adaptive filtering techniques, machine learning approaches, etc., have demonstrated reliable solutions in simulating battery degradation. This paper reviews the battery cell degradation mechanisms, followed by the prediction of battery health parameters and relevant degradation modelling approaches for individual cells. The purpose of this review is to provide a structured analysis of how different modelling methods capture degradation behavior, to identify their strengths and limitations, and to clarify how they can be applied for battery health prediction. It also highlights the importance of datasets required for developing predictive models and summarizes open-source datasets based on the chemistry, cycling process, and their key features. • The research examines how degradation mechanisms affect Lithium-ion Battery performance. • Various battery health states, such as State of Health, State of Charge, and Remaining Useful Life, are discussed. • The paper also reviews approaches like machine learning, time series models, and statistical techniques for predicting battery health states. • Open-source datasets are discussed and summarized based on cell type, chemistry, and cycling process.