A Dual Kalman Filtering Algorithm for Estimating the SOC of Lithium-Ion Batteries with LiMn0.6Fe0.4PO4/LiNi0.5Co0.2Mn0.3O2 Cathode Based on Multi-Innovation and Schmidt Orthogonal Transformation

Abstract For the use of batteries to be secure and dependable, monitoring the state of charge (SOC) is essential. It is challenging to accurately calculate the SOC of LiMn0.6Fe0.4PO4/LiNi0.5Co0.2Mn0.3O2 batteries because they have two flat voltage plateaus and their performance is significantly influenced by temperature. To improve SOC estimation accuracy, a battery-based SOC research approach based on dual Kalman filtering is put forth. The improved adaptive unscented Kalman filter (MIASOUKF) is utilized for SOC estimation, while the extended Kalman filter (EKF) is employed for online identification of battery model parameters based on the second-order RC equivalent circuit model. By combining adaptive unscented Kalman filtering (AUKF) and multi-innovation, the estimation accuracy of the algorithm is improved by repeating the use of historical information, and the sampling points are subjected to Schmidt orthogonal transformation during the sampling point selection in AUKF to decrease the computational complexity of the AUKF, making the algorithm easy to implement in real time. Under complex working conditions, the average error of the MIASOUKF-EKF joint estimation algorithm is less than 0.53%. Through the comparison of experiments and simulations, MIASOUKF-EKF algorithm can effectively improve the estimation accuracy of SOC.