Application of Local State Decomposition in Fault Feature Detection of Rotating Machinery
- Resource Type
- Authors
- Yuxue Jiang
- Source
- 2019 12th International Conference on Intelligent Computation Technology and Automation (ICICTA).
- Subject
- Artificial neural network
Computer science
business.industry
Feature recognition
Pattern recognition
Hardware_PERFORMANCEANDRELIABILITY
Fault (power engineering)
Fault detection and isolation
Support vector machine
Computer Science::Hardware Architecture
Decomposition (computer science)
Artificial intelligence
State (computer science)
business
Computer Science::Operating Systems
Computer Science::Distributed, Parallel, and Cluster Computing
Feature detection (computer vision)
- Language
In order to solve the problem of low detection integrity of traditional fault feature detection methods, local state decomposition is applied to detect fault features of rotating machinery. The fault signals of rotating machinery collected by sensors are decomposed and processed, and the decomposed signals are de-noised. The fault features are extracted from the processed fault signals according to the linearity. Combining SOM neural network and vector machine model, the fault feature recognition is completed. The fault feature recognition results are analyzed and the fault feature detection of rotating machinery is completed. Finally, a comparative experiment with traditional fault feature detection methods is carried out. The experimental results show that the application of local state decomposition can improve the integrity of mechanical fault detection.