一种电池容量预测方法、系统以及计算机程序产品
By combining a bidirectional long short-term memory network and a particle filter model with a dynamic time warping strategy, the nonlinear time shift and capacity jump problems in lithium metal battery capacity prediction are solved, achieving higher accuracy and more stable battery capacity prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TIANFU JIANGXI LAB
- Filing Date
- 2025-10-27
- Publication Date
- 2026-07-17
AI Technical Summary
Existing battery life prediction methods for lithium metal batteries suffer from problems such as rapid capacity decay, short cycle life, and drastic capacity jumps. Furthermore, traditional methods exhibit nonlinear time shifts and unstable prediction accuracy during model fusion.
A method combining bidirectional long short-term memory networks and particle filter models is adopted. The mapping relationship between the initial prediction sequence and the capacity degradation sequence is constructed through dynamic time warping strategy. The particle filter model is used to determine the target prediction residual to correct the initial prediction sequence, forming a closed loop of data-driven and physical model integration.
It improves the accuracy and stability of lithium metal battery capacity prediction, can accurately capture nonlinear degradation trends, adapt to prediction under capacity jump scenarios, reduce dependence on a large amount of training data, and improve the long-term reliability and stability of prediction.
Smart Images
Figure CN121410573B_ABST