State of Charge Estimation of Lithium-Ion Battery using an Extended Kalman Filter optimized by The Octopus Optimization Algorithm

Savaş K. K., DEMİRTAŞ M., ÇETİNBAŞ İ.

14th International Conference on Smart Grid, icSmartGrid 2026, Suceava, Romanya, 6 - 09 Temmuz 2026, ss.667-672, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/icsmartgrid70882.2026.11680889
  • Basıldığı Şehir: Suceava
  • Basıldığı Ülke: Romanya
  • Sayfa Sayıları: ss.667-672
  • Anahtar Kelimeler: extended Kalman filter, forgetting factor recursive least squares, lithium-ion batteries, octopus optimization algorithm, state of charge
  • Gazi Üniversitesi Adresli: Evet

Özet

Reliable state-of-charge (SoC) estimation is a key requirement for the safe and efficient operation of battery management systems (BMSs) in lithium-ion batteries (LiBs). In this study, an extended Kalman filter (EKF)-based SoC estimation approach optimized by the Octopus Optimization Algorithm (OOA) is proposed. The battery dynamics are represented using a first-order RC equivalent circuit model, whose parameters are updated online by the forgetting-factor recursive least squares (FFRLS) method. In addition, the EKF covariance parameters are optimized by OOA to improve estimation performance. The proposed method is evaluated using the dynamic stress test (DST) profile at 25 °C from the University of Maryland battery dataset. For comparison, fixed-ECM EKF, FFRLS-ECM EKF, particle swarm optimization-based EKF (PSO-EKF), and centered collision optimizer-based EKF (CCO-EKF) are implemented under the same conditions. Among the compared methods, OOA-EKF achieves the best SoC estimation performance, with a root mean square error (RMSE) of 0.9727%. Compared with fixed-ECM EKF, OOA-EKF reduces the SoC estimation RMSE by approximately 31.8%. The method also provides accurate terminal voltage tracking, with a voltage RMSE of 0.2996 mV. These findings indicate that combining FFRLS-based time-varying ECM parameters with OOA-optimized EKF covariance parameters improves both SoC estimation accuracy and terminal voltage tracking performance.