Physics-Informed Markov Chain Model to Identify Degradation Pathways of Lithium-Ion Cells

This study proposes a physics-informed Markov chain model to identify the degradation pathways of lithium-ion batteries (LIBs) using capacity measurements only. It comprises six states in three phases, two sleeping states, three active states, and a dead state, and accounts for the complex and nonlinear degradation phenomena of LIBs, including the loss of active material (LAM) and loss of lithium inventory (LLI). The two sleeping states account for the regeneration effect on the degradation phenomena. Specifically, the nonhomogeneous active state accounts for nonlinear degradation. The direct correlation between the phenomena and states, in the proposed model, makes it feasible to accurately account for the complex degradation pathways under different operational conditions. A systematic analysis using two experimental datasets showed that the model is accurate and robust, demonstrating that the physical interpretation of degradation pathways is feasible with the conditional probabilities estimated from the proposed model, including the temperature and compressive pressure dependencies. A feasibility study of the proposed method reveals that it accurately predicts the remaining useful life (RUL) with only the initial 5% of capacity measurements when only capacity measurement is available, suggesting that the model would be effective in estimating the state of health (SOH) and predicting the RUL.