State-of-Charge Estimation Method for Lithium Batteries Based on Adaptive Fusion Factors

Abstract

Abstract Accurate estimation of the state of charge is significant for battery safety. To improve robustness, computational efficiency, and noise stability in state of charge estimation for lithium batteries, we propose a method based on adaptive fusion factors. Based on the methodological properties of the open circuit voltage method and the ampere hour method, we designed a fusion factor function to combine the strong correction ability of the open circuit voltage method with the smoothing advantages of the ampere hour method. The proposed method utilizes an adaptive forgetting factor recursive least squares approach to address the limitation of the traditional open circuit voltage method, which cannot estimate state of charge online. It corrects the battery capacity using historical data to achieve accurate state of charge estimation. The effectiveness and robustness of the proposed method are validated using self-tests and a public dataset. The results demonstrate that the mean absolute error in state of charge estimation is approximately 1%, even when the initial state of charge value deviates from the actual value and the dataset contains noise.