The Role of Discharge Curve Knee Point Analysis for Early Li-ion Battery End of Life Forecasting

Estimating the End of Life (EOL) of a lithium-ion battery is challenging due to the complex and nonlinear degradation behavior influenced by various factors. This work proposes a data-driven approach to predict both the cycle life and the degradation knee point of lithium-ion batteries, using knee points of voltage-capacity discharge curves. These knee points are used to construct a novel form of the degradation curve. The performance of these models is evaluated and compared against models trained on conventional degradation curves, typically defined by the discharge capacity at the cut-off voltage in each cycle. The study also explores the effect of excluding early-cycle data, simulating scenarios where only limited early-life information is available. Additionally, the influence of incorporating charge protocol characteristics into the model is assessed to understand their contribution to prediction accuracy. The results indicate that knee points in voltage-capacity discharge curves enable more accurate predictions of cycle life and degradation knee points.