State of Charge Estimation of Lithium Battery Energy Storage Systems Based on Adaptive Correntropy Unscented Kalman Filter

State of Charge (SOC) estimation plays a crucial role in lithium battery energy storage systems (LBESS). In this paper, we describes a Correntropy Adaptive Unscented Kalman Filter (UKF) algorithms that is able to solve the problem of non-Gaussian noise interference in SOC estimation of LBESS. We believe that the problem of non-Gaussian noise interference will appear widely in power system because Equipment failure or serious electromagnetic interference are increasing common and can cause negative effect for grid. Whereas previous algorithms used mean Square Error (MSE) as the cost function of the UKF too inefficient to deal with non-Gaussian in practice, the algorithm described in this practical: a robust Correntropy Unscented Kalman Filter (UKF) is introduced to improve SOC estimation accuracy in the existence of non-Gaussian noise. Furthermore, this paper combines the advantages of Correntropy and Adaptive filter to improve UKF algorithm, and derives a new robust Kalman Filter, which is called Adaptive C-UKF. The experiment results show that the proposed Adaptive C-UKF method can efficiently improve the estimation accuracy in the non-Gaussian noise environment.