Data-optimization based SOC-SOH estimation for lithium-ion batteries with current bias compensation

To achieve high-precision state estimation for lithium-ion batteries under current bias interference, this paper proposes a data-optimized state of charge-state of health (SOC-SOH) estimation method with current bias compensation. First, we detailedly analyze the impact of current bias on the state estimation problem, and demonstrate why Kalman filter fails to operate effectively. Next, an improved parameter identification framework is designed, iteratively refining battery model parameters using high-quality data segments, which are selected via Fisher Information Matrix-based sensitivity analysis, ensuring they provide proper information for parameter estimation. Then, a current bias compensation term is incorporated into the system state equations, and SOH estimation is performed using short-term constant current discharge data. Furthermore, the acquired SOH result are subsequently used for high-precision SOC estimation under complex conditions. Meanwhile, the effectiveness and robustness of the proposed method are validated with relevant experimental data. With different current bias interferences, the final SOH estimation errors for all data segments remain within 1 %, and the Root Mean Square Error and Mean Absolute Error of SOC estimation results under all conditions also remain below 1.5 %. Finally, based on an advanced scalable battery management system, the potential application scheme of the proposed method in real vehicles is discussed.