| CPC B60L 58/16 (2019.02) [B60L 58/40 (2019.02)] | 20 Claims |
|
1. A method comprising:
receiving battery node diagnostic data from a battery system via a communication interface, the battery system including a plurality of battery nodes that each include a respective one or more battery cells, the battery system also including a plurality of battery node controllers, each battery node controller selectively coupling a power bus with a respective battery node of the plurality of battery nodes;
selecting one or more outlier detection machine learning models based on profile information included in the battery node diagnostic data, the profile information identifying a respective battery node operation profile associated with a respective subset of the battery node diagnostic data;
identifying a designated subset of the battery nodes as outliers by applying the one or more outlier detection machine learning models to identify one or more differences between first diagnostic data for the designated subset of the battery nodes and a population-level representation of the battery node diagnostic data;
determining a plurality of outcome values by applying one or more predetermined rules to the battery node diagnostic data;
identifying a designated battery node as exhibiting a fault based on the designated subset of the battery nodes and the plurality of outcome values; and
transmitting via the communication interface an instruction to update a designated operation profile for the designated battery node to mitigate or revert the fault.
|