A robust modified adaptive extended Kalman filter for state- of-charge estimation of rechargeable battery under dynamic operating condition

Satyaprakash Rout,Satyajit Das
DOI: https://doi.org/10.1007/s00202-024-02523-4
IF: 1.63
2024-06-20
Electrical Engineering
Abstract:In an electric automotive, accurate estimation of state of charge (SOC) ensures the safety and reliability of the battery. Due to ease of implementation, an adaptive extended Kalman filter (AEKF)-based state-of-charge (SOC) estimation is widely used in the battery management system of electric vehicles. However, the performance of AEKF in SOC estimation degrades in high-speed, high-acceleration drive cycles due to transient disturbances in load current. Moreover, the presence of a biased noise estimator in the design of AEKF may generate an anomalous filter gain value, which may lead to filter divergence or reduce the filter convergence speed during SOC estimation. To mitigate the aforementioned issues of AEKF, a robust modified adaptive extended Kalman filter (MAEKF) is proposed that incorporates an adaptive fading factor to counter the adverse effect of transient disturbances in load current. The proposed estimator combines biased and unbiased estimators to cancel out the anomalous Kalman filter gain. The robustness of the proposed MAEKF is studied by applying Gaussian white noise to the measured terminal voltage, bias voltage and offset current in sensor measurement, and transient disturbances in the load current. The comparative results depict preeminent performance and strong robustness in SOC estimation of the proposed MAEKF over other AEKF.
engineering, electrical & electronic
What problem does this paper attempt to address?