A novel approach for accurate SOC estimation in Li-ion batteries in view of temperature variations

Abstract

• A new polynomial fitting-based algorithm to enhance SOC estimation (FPSOC) accuracy in Li-ion batteries under temperature variations. • Shows lower RMSE errors (0.72 and 0.84) compared to conventional methods like the CC method (7.98 and 7.81). • Performs exceptionally well across various temperature profiles, making it suitable for battery management systems in electric vehicles and solar energy storage. • MATLAB/Simulink simulations validate FPSOC's reliability and accuracy, outperforming existing techniques. Lithium-ion (Li-ion) batteries are extensively utilized in electric vehicles and solar energy storage systems owing to their remarkable energy and power density, capacity, and performance The main problems faced in these systems are the errors that occurred in estimating the battery's state of charge (SOC). For that reason, the current study introduces a novel approach based on a fitting polynomial algorithm for state of charge (FPSOC) estimation. The proposed FPSOC method enhances the accuracy and efficiency of SOC estimation by mitigating temperature-induced inaccuracies. Unlike other methods, this technique eliminates the need for a thermal model, offering a solution to rectify temperature-related errors and overcome limitations found in existing battery SOC estimation methods. Through the utilization of this approach, these challenges can be effectively addressed. Simulation results utilizing the MATLAB/Simulink tool demonstrate that the proposed FPSOC method achieves highly precise SOC determination with significantly lower RMSE errors of 0.72 and 0.84, compared to 7.98 and 7.81 for the CC method, respectively, under real and simulated temperature profiles (from -20°C to 80°C). Contrasting with established techniques such as TBCC-APF, TBCC-AEKF, APF, AEKF, HMM, TBCC, Multi-Time, Scale DEKF, FF-LSTM, EKF-CCD, EKF-FUDS, OCV-SOC, and ASVDUKF, the proposed FPSOC method showcases exceptional performance, yielding a remarkably low average error of 0.72 in battery SOC estimation. This demonstrates its reliability and accuracy, positioning it as a highly promising algorithm for battery management systems (BMS).