Prediction of compression force evolution over degradation for a lithium-ion battery

This study proposes a method to predict the evolution of compression force during degradation lithium-ion battery under packed conditions. The total comprises irreversible and reversible forces. former is estimated using multivariate machine learning method, whereas latter by combining phenomenological modeling. For predicting force, impedance-related features are extracted their correlations with quantitatively analyzed Grey relational analysis. Subsequently, high grades employed as representative health indicators for inputs Gaussian process regression. charge/discharge period predicted model. equivalent stiffness used in this model separately depending on state charge (SOC) account inherent characteristics phase transition different behaviors. SOC shows nonlinearity but weak characteristics, those low medium SOCs show linearity strong characteristics. Finally, proposed enable control design two potential applications: estimations health-dependent separator compression.