Partition-based Unscented Kalman Filter for Reconfigurable Battery Pack State Estimation using an Electrochemical Model

摘要

Accurate state estimation of large-scale lithium-ion battery packs isnecessary for the advanced control of batteries, which could potentiallyincrease their lifetime through e.g. reconfiguration. To tackle this problem,an enhanced reduced-order electrochemical model is used here. This model allowsconsidering a wider operating range and thermal coupling between cells, thelatter turning out to be significant. The resulting nonlinear model isexploited for state estimation through unscented Kalman filters (UKF). A sensornetwork composed of one sensor node per battery cell is deployed. Each sensornode is equipped with a local UKF, which uses available local measurementstogether with additional information coming from neighboring sensor nodes. Suchstate estimation scheme gives rise to a partition-based unscented Kalman filter(PUKF). The method is validated on data from a detailed simulator for a batterypack comprised of six cells, with reconfiguration capabilities. The resultsshow that the distributed approach outperforms the centralized one in terms ofcomputation time at the expense of a very low increase of mean-squareestimation error.