State of energy estimation for a series-connected lithium-ion battery pack based on an adaptive weighted strategy

Due to the inconsistency among battery cells, it is very difficult to estimate the state of energy (SOE) of a battery pack online. In this paper, an adaptive SOE estimation method for a series-connected lithium-ion battery pack based on representative cells is proposed. The dynamic characteristics of a battery are modeled by a first-order resistor-capacitor model. The key parameters and the SOEs of the representative cells are estimated by the recursive least squares algorithm and an adaptive cubature Kalman filter, respectively. The SOE of the series-connected battery pack is obtained by weighting the SOEs of the representative cells based on an adaptive strategy. Experimental results indicate that the SOE estimation result of the series-connected battery pack is close to the SOE of the strongest representative cell at the fully charged state, while it is close to the SOE of the weakest representative cell at the ending point of discharging. Even with a large initial error, the estimated SOE can quickly track the reference value. The root-mean square errors of the SOE estimation results at 25 degrees C, 50 degrees C and 0 degrees C are 1.3%, 2.2% and 1.7%, respectively. (C) 2020 Elsevier Ltd. All rights reserved.

Automated summary

A method for adaptive SOE estimation of a series-connected lithium-ion battery pack based on representative cells has been proposed, showing promising results with low root-mean square errors under various temperature conditions.