Implementation of the approximate weighted total least squares method for capacity estimation of aging LFP cells

Details

Supervisors

De Jaeger, Emmanuel ; Vanheusden, Wannes

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en génie de l'énergie , à finalité spécialisée

Abstract

enBattery Energy Storage Systems are becoming increasingly common due to the energy shift driven by the energy transition. As the capacity of LFP cells decreases over their lifetime its estimation for long term operation is crucial to improve the State of Charge and to estimate the State of Health. The cell capacity degradation depends on many factors and is not trivial to estimate. This thesis proposes a capacity estimation algorithm based on the Approximate Weighted Total Least Squares (AWTLS) method to estimate the capacity of LFP cells at different aging states, the algorithm is designed to be integrated into a dual estimation framework. The AWTLS method uses multiple pairs of SoC and charge variations between two times to estimate the capacity. The SoC is estimated using an Unscented Kalman Filter (UKF) with a first-order Equivalent Circuit Model (ECM). The algorithm was validated on laboratory data from two CATL 280Ah LFP cells at different aging states. The algorithm achieves an error of approximately 2%. This error was found to depend on the cell temperature and the aging state.