Lithium ion battery health state estimation method and system
By extracting multi-dimensional operating features of lithium-ion batteries under full charge and discharge conditions, combining causal convolution and sparse attention mechanisms for cross-condition fusion, and introducing high-value calibration fragments for physical model calibration, the problem of incomplete information and insufficient robustness in the data-driven method for lithium-ion battery health state estimation is solved, achieving high-precision and robust health state estimation.
Patent Information
- Application Number
- CN202610329226.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-18
- Publication Date
- 2026-05-15
AI Technical Summary
Existing data-driven methods for estimating the health status of lithium-ion batteries fail to capture all the underlying details and lack robustness under complex operating conditions. They are also unable to sensitively detect early minor losses and are susceptible to interference from random loads and sensor noise.
By acquiring real-time operating data of lithium-ion batteries during the full charge and discharge cycle, multi-dimensional operating features are extracted, and cross-condition fusion is performed using causal convolution and sparse attention mechanisms. The data is then corrected by combining the battery's historical degradation trajectory, and high-value calibration fragments are introduced to calibrate the physical model, ultimately determining the estimated health status.
It improves the accuracy and robustness of lithium-ion battery state of health estimation, reduces the drift rate of assessment results, achieves high-precision online estimation of lithium-ion battery state of health (SOH), and ensures the robustness and accuracy of the estimation.