An Approach to Estimate Lithium-Ion Battery State of Charge Based on Adaptive Lyapunov Super Twisting Observer

Abstract

To increase lifespan and optimize capacity utilization of lithium-ion (Li-ion) battery, it is imperative to predict its states accurately. This paper primarily focuses on offering an improved solution for estimating one of the most important battery states known as state of charge (SOC). After extensive analysis of the current state-of-the-art methods, we propose a new adaptive observer based approach for precise estimation of battery SOC using super twisting algorithm. The proposed approach avoids overestimation of gains while ensuring finite-time convergence of the states. For choosing the associated adaptive gains, it does not demand any information on upper bounds of the uncertainty and its derivative, except for their existence. Unlike most of the sliding mode methods employed for SOC estimation, the proposed approach provides continuous control injection resulting in reduced chattering and eliminates the need for low pass filter that leads to improved accuracy and response time. Lyapunov stability theory is utilized to prove the robustness and the error convergence of the proposed observer. The presented algorithm is executed on a lithium-polymer battery and the obtained results demonstrate that it outperforms the well-established approaches in terms of accuracy, chattering, robustness, computational complexity, and convergence time.