一种基于多特征集成极限学习机的冷水机组故障诊断方法
By combining multi-feature ensemble extreme learning machine with weighted fusion evidence reasoning theory, the problems of insufficient feature extraction and insufficient model generalization in traditional chiller unit fault diagnosis methods are solved, achieving high-precision fault diagnosis, especially showing excellent performance in the identification of minor faults.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANTONG UNIV
- Filing Date
- 2025-08-26
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional methods for diagnosing chiller unit faults rely on single feature extraction, resulting in low diagnostic accuracy. Furthermore, deep learning models are prone to overfitting in small sample scenarios, leading to insufficient generalization performance.
We employ a multi-feature ensemble extreme learning machine to extract basic statistical features, probability distribution features, and independent features through a sliding window. Combined with the features of the original data, we reduce the randomness of individual models using the ensemble extreme learning machine and improve diagnostic accuracy through weighted fusion evidence reasoning theory.
It improves the comprehensiveness and accuracy of fault diagnosis, especially in the ability to identify minor faults, while maintaining fast training efficiency, reducing computational costs, and improving the robustness and generalization ability of the model.
Smart Images
Figure CN121350442B_ABST