Efficient multi-cell SOC estimation for electrified vehicle battery packs

Abstract

Reporting pack-level state-of-charge (SOC) requires identification of the maximum and minimum SOC cells in the pack. Applying separate estimators for every cell greatly increases the computational burden of battery state estimation. Efficient approaches based on either an average cell estimator and m-top bottom estimators are benchmarked on their ability to report maximum minimum SOC. A simulation case study on a 14-cell battery with cell balancing and typical aged-pack cell-to-cell variations is performed. The average estimator approach with voltage-based correction is shown to induce unacceptably large errors. The m-top/bottom method with voltage based selection criteria is shown to be a viable computational efficient approach in estimating maximum minimum SOC cells.