Nonlinear Parameter Estimation for Capacity Fade in Lithium-Ion Cells Based on a Reduced-Order Electrochemical Model

Lithium-ion batteries are central to the powertrain transformation taking place in the automotive industry, but the duration, cost, and complexity of experimental work for the characterization of aging mechanisms drive the need for models and model-based estimation approaches. This paper presents a model-based nonlinear parameter estimation method for the characterization of long-term capacity fade of Lithium-ion cells. The proposed approach relies on a reduced-order model of a LiC6/LiFePO4 cell, describing the mass and charge transfer in the solid and liquid phase, and the governing electrochemical principles. The model, validated with experimental data from a battery cell at beginning of life, is used to conduct a sensitivity analysis of the capacity to a subset of physicochemical parameters that are hypothesized to evolve throughout the battery's life. After isolating the most significant model parameters characterizing the long-term capacity degradation, experimental data from battery aging studies were used to solve a system identification problem to identify the degradation trend for the aging-related parameters. The developed tool is applicable to model-based diagnostic algorithms for ascertaining battery state-of-health and predicting the remaining useful life for Li-ion cells subjected to relevant usage and environmental conditions for automotive applications.