一种基于知识蒸馏与迁移学习的储能电站云端-边缘协同健康管理方法及系统
By employing knowledge distillation and transfer learning methods in energy storage power stations, a deep spatiotemporal network model is constructed, and knowledge distillation and progressive domain adaptation are performed. This solves the data transmission and computational efficiency problems in fault diagnosis during the health management of energy storage power stations, and enables high-precision lightweight deployment and early fault warning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG JIFENG ENERGY TECH CO LTD
- Filing Date
- 2026-03-31
- Publication Date
- 2026-07-17
AI Technical Summary
Existing health management methods for energy storage power stations suffer from problems such as high data transmission overhead, insufficient real-time performance, large model computation volume, weak generalization ability, and difficulty in adapting to heterogeneous hardware environments in fault diagnosis. In particular, it is difficult to achieve high precision and lightweight deployment in battery aging diagnosis.
A cloud-edge collaborative health management method based on knowledge distillation and transfer learning is adopted. Battery data is collected through edge nodes to construct a deep spatiotemporal network model, and knowledge distillation and progressive domain adaptation are performed to achieve the deployment of a lightweight inference model and early warning of faults.
It enables efficient migration and stable deployment of high-precision aging diagnostic models to lightweight edge devices, ensuring the accuracy and real-time nature of early fault warnings.
Smart Images
Figure CN121965711B_ABST