State of health estimation in composite electrode lithium-ion cells

Electrochemical models of lithium-ion batteries have been increasingly considered for online state of health estimation. These models can more accurately predict cell performance than traditional circuit models and can better relate physical degradation mechanisms to changes in model parameters. However, examples of state of health estimation algorithms that are validated with experimental data are scarce in the literature, particularly for cells with a composite electrode. The individual electrode active materials in a composite electrode may degrade at different rates and according to different physical mechanisms, and online estimation of this degradation facilitates more robust knowledge of how battery performance changes over its life. In this paper we use a reduced-order electrochemical model for a composite LiMn2O4-LiNi1/3Mn1/3 Co1/3O2 (LMO-NMC) electrode cell for online estimation of active material loss. Experimental data collected from composite electrode half cells that were aged under constant current cycling are used in an extended Kalman filter to estimate model parameters associated with loss of each active material. The capacity loss predicted by the online estimates agrees well with the measured capacity loss. Additionally, a differential capacity analysis demonstrates that active materials lose capacity at a similar rate, the same conclusion obtained from the online estimation algorithm.