Non-linear frequency response analysis for assessment of the ageing history of lithium ion batteries: A combined simulation and experimental approach

In this work, we present a novel approach for identifying the ageing history of lithium-ion batteries based on experimental nonlinear frequency response analysis (NFRA) measurements. A regression model, trained on simulated NFRA data, is shown to be capable of quantifying degradation modes such as solid electrolyte interphase (SEI) growth, lithium plating, and loss of active material (LAM) with no a-priori knowledge of the cell’s historical duty. Our analysis, combining experimental and simulation approaches, demonstrates NFRA’s potential as a powerful tool for ageing diagnosis by capturing various degradation modes. Changes in NFRA response through life exhibit strong correlations with ageing paths, particularly in the frequency range of 0.2 to 10 Hz. Observations highlight a strong influence of the state of charge on the resultant NFRA response, emphasizing that measurements at a single open circuit voltage (OCV) and harmonics values from a single frequency are insufficient for comprehensive characterization. This analysis underscores the need for correlating NFRA at multiple OCVs and frequencies for detailed ageing assessment. Evaluation on commercially relevant cells enhanced the models’ reliability for industrial applications. This quantitative, data-driven approach using NFRA holds potential to enhance battery management strategies, extend lifespan and improve confidence in second-life applications of batteries. Future work should focus on improving regression analysis robustness, reducing dimensionality, and broadening testing conditions. •Quantification of lithium-ion battery ageing modes using NFRA•P2D model simulated ageing and NFRA spectra, validated against experimental data•Simulated NFRA trained regression model detecting lithium plating, SEI and LAM•Model identified dominant ageing modes aligned with those expected from experiments