Estimation method for safety state of battery pack based on deep learning and consistency detection
US11774505B1 · kind B1 · utility
Assignees
Inventors
Key dates
| Filing date | Dec 18, 2022 |
| Grant date | Oct 3, 2023 |
| Priority date | — |
| Expiry date | Dec 18, 2042 |
Classification
- Technology area (CPC Y)Emerging Cross-Sectional Technologies
- CPC primaryY02T10/70
- WIPO fieldMeasurement
- WIPO sectorInstruments
Abstract
Disclosed is an estimation method for the safety state of a battery pack based on deep learning and consistency detection, including: acquiring battery parameters of each single battery in the battery pack in a charging process to be identified; calculating multiple groups of feature data according to the battery parameters; constituting a first matrix by the multiple groups of feature data, and calculating a covariance matrix of the first matrix; inputting the covariance matrix into a first trained fully connected layer, so as to extract principal components of the first matrix and obtain a second matrix; multiplying the first matrix and the second matrix to obtain a third matrix; and inputting the third matrix into a series-connected and trained multi-head self-attention layer and classification layer, to identify whether single battery consistency safety hazards exist in the charging process. This embodiment improves the accuracy of identification.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.