Fuzzy adaptive singular value decomposition cubature Kalman filtering algorithm for lithium-ion battery state-of-charge estimation (vol 50, pg 614, 2022)

To solve the problem of the slow convergence speed for the battery state-of-charge estimation of cubature Kalman filter algorithm, the ternary lithium-ion battery is taken as the research object, and an algorithm combining the fuzzy self-adaptation and singular value decomposition cubature Kalman filtering is proposed. The algorithm takes the system innovation and its change rate as the fuzzy input and the output as the adjustment factor, which is used to adjust the process noise covariance matrix R. The Kalman gain is adjusted through the fuzzy control of R. To ensure the stability of the algorithm in the calculation process, the singular value decomposition is applied to cubature Kalman algorithm. Then, a second-order RC equivalent circuit model with double internal resistance is built and tested under different conditions to verify the rationality of the improved algorithm. The verification results show that under the simple condition, the convergence speed of the proposed algorithm in the different initial state-of-charge values increased by 40.00% and 25.00%, the maximum estimation error of the state-of-charge is 2.52% and 2.51%, the Mean Absolute Error is 0.816% and 0.880%, and the Root Mean Square Error is 1.276% and 1.380%. When the initial state-of-charge value is 0.8, the convergence speed in the complex condition is increased by about 30.00%; the maximum estimation result error, Mean Absolute Error, and Root Mean Square Error are 2.21%, 0.222%, and 1.327%, respectively. When the initial state-of-charge value is 0.6, the convergence speed in the complex condition is increased by about 10.00%; the maximum estimation result error, Mean Absolute Error, and Root Mean Square Error are 2.72%, 0.941%, and 1.327%, respectively. Without reducing the estimation accuracy, the improved algorithm can significantly increase the convergence speed of predictive value tracking, which provides a theoretical basis for the wide application of lithium-ion batteries.