State-of-health estimation for lithium-ion batteries using differential thermal voltammetry and Gaussian process regression

The state of health (SOH) of lithium-ion batteries is a key factor to ensure the safe and reliable operation of electric vehicles (EVs). Differential thermal voltammetry (DTV) has been shown to be an effective in situ diagnosis method. In this paper, a SOH estimation method for lithium-ion batteries by a fusion of DTV and Gaussian process regression (GPR) is proposed. First, a wavelet transform (WT) is used to smooth DTV curves. Then the health features (HFs) are extracted, and their correlation with capacity loss is quantitatively evaluated through a Pearson correlation coefficient. To compress the data, canonical correlation analysis (CCA) is applied to obtain the canonical health feature (CHF). Subsequently, the GPR algorithm is used to establish the nonlinear mapping relationship between CHF and SOH. Experiments are performed on two datasets at different temperatures. Results show that the proposed method has high estimation accuracy and robustness.