US 11,921,163 B2
Modular machine learning structure for electric vehicle battery state of health estimation
Gianina Alina Negoita, Belmont, CA (US); Wesley Teskey, Foster City, CA (US); Jean-Baptiste Renn, San Francisco, CA (US); and William Arthur Paxton, Redwood City, CA (US)
Assigned to VOLKSWAGEN AKTIENGESELLSCHAFT, Wolfsburg (DE)
Filed by VOLKSWAGEN AKTIENGESELLSCHAFT, Wolfsburg (DE)
Filed on Dec. 16, 2021, as Appl. No. 17/553,585.
Prior Publication US 2023/0194614 A1, Jun. 22, 2023
Int. Cl. G01R 31/367 (2019.01); B60L 58/12 (2019.01); G01R 31/382 (2019.01); G01R 31/392 (2019.01); G01R 31/396 (2019.01); G06N 3/02 (2006.01); G06N 3/04 (2023.01); G06N 3/045 (2023.01)
CPC G01R 31/367 (2019.01) [B60L 58/12 (2019.02); G01R 31/382 (2019.01); G01R 31/392 (2019.01); G01R 31/396 (2019.01); G06N 3/02 (2013.01); G06N 3/04 (2013.01); G06N 3/045 (2023.01); G06F 2218/08 (2023.01)] 21 Claims
OG exemplary drawing
 
1. A method comprising:
collecting raw sensor data from a battery module, wherein the battery module includes a plurality of battery cells;
extracting input battery features from the raw sensor data collected from the battery module;
using the input battery features to update node states of a graph neural network (GNN), wherein the GNN includes a plurality of GNN nodes each of which representing a respective battery cell in the plurality of battery cells;
generating, based at least in part on individual output states of individual GNN nodes in the plurality of GNN nodes, estimation of one or more battery state of health (SoH) indicators, wherein the individual output states of individual GNN nodes in the plurality of GNN nodes are determined based at least in part on the updated node states of the GNN; and
performing, based at least in part on the estimation of the one or more battery SoH indicators, one of replacing the battery module, balancing the plurality of battery cells or the battery module, and distributing loads to the battery module.