The State of Health Estimation of Lithium-ion Battery based Support Vector Regression-Particle Swarm Optimization Model

The State of Health (SOH) plays a critical role in Li-ion Battery Management Systems (BMS), serving as a key indicator for evaluating battery performance and predicting its lifespan. Accurate SOH estimation ensures operational reliability, safety, and cost-effective maintenance. To enhance SOH prediction accuracy, this study proposes a Support Vector Regression (SVR) model with a radial basis function (RBF) kernel. To address nonlinear effects, such as capacity regeneration observed in the SOH trajectory, and to improve prediction precision and stability, the model incorporates Particle Swarm Optimization (PSO). This hybrid PSO-SVR approach delivers superior accuracy and robust generalization capabilities for SOH estimation. To validate the proposed method, a comparative analysis was conducted between the PSO-SVR model and a standard SVR model using datasets from three Li-ion battery types: B5, B6, and B8. The results demonstrated that the PSO-SVR model consistently outperformed the standard SVR, achieving high R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> values of 0.9340, 0.9648, and 0.9319 and low RMSE values of 0.0165, 0.0147, and 0.0179 for batteries B5, B6, and B8, respectively. These findings highlight the PSO-SVR model's reliability and effectiveness in accurately predicting SOH, making it a promising solution for advanced battery management systems.