State of charge estimation for “LiFePO4 - LiCoxNiyMn1-x-yO2” hybrid battery pack

Abstract

In addition to reconciling the cost and energy density of the battery pack, the “LiFePO4 - LiCoxNiyMn1-x-yO2” (LFP - NMC) hybrid battery pack can also effectively reduce the risk of thermal spread, so it has been initially applied in the industry. However, the electrochemical properties of the two batteries are very different, so realizing the state estimation of the single cell is the key to the effective management of the hybrid battery pack. Therefore, a novel state estimation method for hybrid battery packs is proposed in this paper. First, the easy-to-estimate NMC as the “ruler” to estimate the SOC of the LFP based on the principle of the same amount of electricity in the series cells. Second, through the design of a “difference state observer (DSO)” with fuzzy logic, continuous feedback correction of electricity is realized. Results from the experiments have verified that when the LFP is connected to a 200 Ω resistor to simulate the self-discharge difference along with sensor errors, the algorithm can control the SOC estimation error of the LFP to about 1 %, which has excellent estimation accuracy. In addition, the algorithm has good robustness. In the scenario with current and voltage sensor errors, the SOC estimation error of LFP is within 2 %. Finally, the capacity and SOC of the hybrid battery pack are solved by using the “Electric quantity - Capacity Scatter Diagram” (ECSD), and the estimation errors are within 1.5 Ah and 1 %, respectively.