Fault prediction method and device for energy storage liquid cooling system
Through the dynamic feature fusion layer of adaptive segmentation and neural network model, the problem of insufficient fault feature extraction in energy storage liquid cooling systems is solved, accurate fault diagnosis in high-noise environments is achieved, and the sensitivity and specificity of the system are improved.
Patent Information
- Application Number
- CN202510941183.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-09
- Publication Date
- 2025-08-08
- Estimated Expiration
- 2045-07-09
AI Technical Summary
The prior art is difficult to effectively capture the characteristics of different fault modes in energy storage liquid cooling systems, especially inadequate extraction of weak fault features in high noise environments. In addition, traditional models lack diagnostic accuracy in multivariate coupling and time-varying data processing, and cannot adapt to fault diagnosis in complex environments.
Through dynamic analysis and monitoring of local noise levels and timescale changes of the data, the data is subjected to adaptive segmentation processing, combined with the dynamic fusion layer and adaptive regression unit of the neural network model, the multi-scale convolution kernel and expert selection matrix are used to dynamically adjust the denoising strategy and feature extraction to achieve accurate prediction of different failure modes.
It improves the sensitivity and specificity of the fault diagnosis system, can accurately extract weak fault characteristics in high noise environments, enhances the adaptability and diagnostic accuracy of the model, especially the ability to identify mutant and slow-tempered faults.