Robust and adaptive online estimation of Li-ion battery cell capacity - CEA - Commissariat à l’énergie atomique et aux énergies alternatives Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Robust and adaptive online estimation of Li-ion battery cell capacity

Résumé

The proposed algorithm aims to estimate the instantaneous capacity of an electrical battery. The battery capacity is a crucial parameter of the battery management system and the knowledge of its value is necessary for the state of charge, state of function and remaining useful life estimators. Moreover, operation of lithium batteries raises safety issues, and the very accurate estimation of the battery capacity makes it possible to always guarantee its use in its Safe Operating Area (SOA). As a starting point, the classical Coulomb counting relation is taken as an observation equation in a Kalman framework based on a state equation of the capacity evolution. The observation equation is a biased linear relation with the variables of interest being affected by errors and corrupted by outlier samples. A random consensus resampling is thus applied to reject these outliers and a pre-estimation of the ordinate at the origin of the affine function is performed. The bias reduction is performed thanks to a modified hough- transform. This hough-transform uses the a-priori prediction of the capacity and a multi-resolution recursive approach is used to optimize the bias reduction while limiting the computational complexity. The algorithm is robust to the presence of a high level of outlier and the Kalman approach makes the estimation adaptive. Moreover a statistical test is realized to determine whether the estimated capacity should be directly injected back in the SOC estimator or not. Our algorithm is tested on a database delivered by the NASA DashLink plateform. Statistical simulations confirm the robustness and the adaptive nature of the method. The obtained RMS of the relative errors over the entire ageing cycles are below 3%.
Fichier principal
Vignette du fichier
Paper_VPPC_Michel_Heiries_2023.pdf (468.31 Ko) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

cea-04520715 , version 1 (25-03-2024)

Identifiants

Citer

Paul-Henri Michel, Vincent Heiries. Robust and adaptive online estimation of Li-ion battery cell capacity. VPPC 2023 - 2023 IEEE Vehicle Power and Propulsion Conference, Oct 2023, milan, Italy. ⟨10.1109/VPPC60535.2023.10403135⟩. ⟨cea-04520715⟩
0 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More