A method for estimating the state of health of a lithium battery for vehicles and predicting future degradation

By introducing a semi-supervised learning method through domain adversarial training and pseudo-labels, the problems of cross-domain adaptability and insufficient data utilization in the estimation of the health status of automotive lithium batteries and the prediction of future degradation are solved, achieving higher estimation accuracy and prediction accuracy.

CN122430728APending Publication Date: 2026-07-21JILIN UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
JILIN UNIVERSITY
Filing Date
2026-06-17
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing methods for estimating the health status of automotive lithium batteries are insufficient in terms of cross-domain adaptability and future degradation prediction, making it difficult to meet the accuracy and precision requirements under complex operating conditions, and they cannot effectively utilize unlabeled data.

Method used

A semi-supervised learning mechanism is adopted to extract aging features through domain adversarial training. By combining labeled and unlabeled data, a feature extractor and a SOH evaluator are constructed. The model is optimized using a gradient reversal layer and the white whale optimization algorithm, and pseudo-labels are introduced to predict future degradation.

Benefits of technology

It improves the cross-domain adaptability of lithium battery health state estimation and the accuracy of future degradation prediction, simplifies the modeling process, and reduces the reliance on manual feature engineering.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122430728A_ABST
    Figure CN122430728A_ABST
Patent Text Reader

Abstract

The application discloses a kind of vehicle lithium battery health state cross-domain estimation and future degradation prediction method, belong to battery management technical field.The method includes: obtaining source domain with label data, target domain unlabelled data and to be predicted data;Each data is preprocessed, and the discrete characteristic sequence of each data is obtained;The characteristic sequence of source domain and target domain is input to evaluation module, and the gradient inversion layer is combined with domain discriminator and SOH evaluator to carry out counter training, so that feature extractor I learns domain invariant aging feature, and outputs health state estimation value;Then the health state estimation value is input as pseudo label into prediction module, to iteratively predict future health state trajectory in autoregressive mode.The overall framework of health state cross-domain estimation and future degradation prediction integration is adopted in the application, and the evaluation module learns domain invariant aging feature based on domain counter training, which significantly improves the accuracy of lithium battery health state estimation and future degradation trajectory prediction.
Need to check novelty before this filing date? Find Prior Art