Performance Analysis of ASR-UKFs for Supercapacitor SoC Estimation in Hybrid Energy Storage Systems

Due to their high specific power and durability, supercapacitors are promising candidates to be formed with batteries in hybrid energy storage systems (HESSs). Accurately determining the supercapacitor's state of charge (SoC) represents a crucial task and improves system performance and energy management. In this article, state-of-the-art algorithms to assess the SoC of supercapacitors are initially described, and their performances have been compared in simulation and validated experimentally. To improve the estimation performance, an adaptive square-root unscented Kalman filter (ASR-UKF) method has been finally proposed in this article. Two electric circuit models have been defined for implementing the Kalman filtering method and analyzing its performance. Numerical results demonstrated a 7% estimation error reduction in terms of absolute value with respect to the conventional methods. Moreover, an estimation error lower than 1.5% has been achieved by the proposed method in experimental tests under realistic grid power profile, validating the numerical results and demonstrating the applicability of the developed estimator for supercapacitor SoC estimation.