State of Charge Estimation for Lithium Batteries Based on the Square Root Cubature Kalman Filter Algorithm

The State of Charge of lithium batteries is a crucial parameter for the battery management system in electric vehicles and energy storage applications. Accurate estimation of the State of Charge is of significant importance for the design of the battery management system. To address the problem of accurately estimating the State of Charge, this paper employs the square root cubature Kalman filter algorithm in conjunction with a second-order resistance-capacitance equivalent circuit model of lithium batteries for precise State of Charge estimation. This study utilizes the open-source experimental dataset of the Panasonic NCR18650PF lithium battery and identifies the parameters of the second-order resistance-capacitance equivalent circuit model through the forgetting factor recursive least squares method. To verify the estimation accuracy of the Square Root Cubature Kalman Filter algorithm, the algorithm is simulated and tested using the Urban Dynamometer Driving Schedule conditions within the experimental dataset. The estimation accuracy of the Square Root Cubature Kalman Filter algorithm is compared with that of the Extended Kalman Filter algorithm and the Unscented Kalman Filter algorithm under the same conditions. Simulation test results indicate that the Square Root Cubature Kalman Filter algorithm achieves higher accuracy in estimating the State of Charge of lithium batteries.