一种车用储能锂电池剩余使用寿命预测方法和系统
By combining feature clustering and attention mechanism deep neural networks, the problem of unstable prediction accuracy of lithium batteries under different driving scenarios is solved, and high-precision and stable prediction of the remaining lifespan of lithium batteries is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANTONG UNIV
- Filing Date
- 2025-02-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing data-driven methods for predicting the remaining lifespan of lithium batteries have inconsistent prediction accuracy under different driving scenarios and lack a systematic information fusion strategy, making it difficult to achieve consistently excellent prediction performance across multiple scenarios.
We use a feature clustering method to divide lithium battery data into multiple sub-databases, construct a prediction model based on a deep neural network with an attention mechanism, and select the optimal sub-prediction model for prediction using a similarity evaluation algorithm. We then combine a denoising autoencoder and an improved loss function to optimize the model's convergence and accuracy.
It improves the accuracy and stability of lithium battery remaining life prediction, adapts to data changes under different driving scenarios, and achieves high-precision prediction.
Smart Images
Figure CN120103189B_ABST