In the pursuit of enhancing the reliability of battery health management methods,
accurate estimation of state of charge (SOC) and state of health (SOH) remains a
critical challenge. This article presents a novel fusion estimation algorithm,
combining a dual extended Kalman filter (EKF) with a particle filter (PF), based
on a fractional-order 2-RC battery model (FOEKPF–EKF). The 2-RC fractional-order
model (FOM) is first implemented to accurately depict the battery’s discharge
behavior, outperforming traditional integer-order models (IOM) due to its
ability to capture the cell’s intrinsic diffusion and dispersion
characteristics. An adaptive genetic algorithm (AGA) is then employed for
optimal parameter identification of the FOM, ensuring precise modeling.
Following this, the FOEKPF–EKF algorithm is developed, leveraging the strengths
of FOM, EKF, and PF to effectively handle uncertain, time-varying noise, thereby
improving SOC estimation accuracy. The reliability of the proposed algorithm is
validated through both simulations and experiments. The results demonstrate that
the mean error, maximum error, and RMSE of the terminal voltage are 4.4 mV, 37
mV, and 7.1 mV, respectively—significantly lower than those of the IOM.
Additionally, the AGA shows higher accuracy compared to the recursive least
squares (RLS) method. The FOEKPF–EKF algorithm also achieves more reliable SOC
estimation compared to FOEKPF, IOEKPF, and FOPF, with a mean error of 0.48%, a
maximum error of 1.23%, and an RMSE of 0.55%. Finally, this article outlines
potential directions for future research.
