State-of-charge estimation of lithium-ion battery based on improved equivalent circuit model considering hysteresis combined with adaptive iterative unscented Kalman filtering

The state of charge is an essential index of the operation of a lithium-ion battery. In most cases, precise values of the state of charge can already be obtained by current estimation methods. However, by cause of the hysteresis of lithium-ion batteries, the accurate open circuit voltage value cannot be transmitted to the battery management system, resulting in a biased appraisal of the state of charge. To address this issue, this article analyzes the characteristics of the hysteresis in lithium-ion batteries and defines an improved equivalent circuit model that takes hysteresis into account. A state of charge evaluation scheme based upon the improved equivalent circuit model combined with adaptive iterative unscented Kalman filtering is proffered, which is verified through experimental data under different operating conditions. The findings demonstrate that compared with state of charge estimation methods predicated on the principles of extended Kalman filtering and unscented Kalman filtering, the approach proposed in this essay shows greater accuracy, showcasing robustness and enormous potential for practical applications.