A fan blade fault early warning method, system, device and medium
Through the non-contact pickup, the fan blade audio data is collected, combined with preprocessing, feature extraction and dimensionality reduction optimization, a fault classifier model is built, which solves the problems of structural damage, installation and maintenance difficulties and failure degree of blade fault diagnosis in the existing technology, and realizes the objectivity of fan blade fault warning and the degree of failure.
CN115376302BActive Publication Date: 2025-03-28GUANGDONG MINGYANG WIND POWER IND GRP CO LTD
2 Cites 0 Cited by
Patent Information
- Application Number
- CN202210943367.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-08
- Publication Date
- 2025-03-28
- Estimated Expiration
- 2042-08-08
AI Technical Summary
Technical Problem
The existing fan blade fault diagnosis technology has problems such as structural damage, installation and maintenance difficulties, noise interference, long detection cycle and failure degree.
Method used
The non-contact pickup is used to collect audio data when the fan blade rotates, and through preprocessing, feature extraction, dimensionality reduction optimization and construction of a fault classifier model, the objectivity of fault warning and the degree of fault are realized.
Benefits of technology
The objectivity and degree of failure warning of fan blades is realized, which avoids the damage to the blade structure and installation and maintenance difficulties, and improves the detection efficiency and applicability.
✦ Generated by Eureka AI based on patent content.
Abstract
The present invention discloses a method, system, device and medium for early warning of fan blade faults, the method comprising: setting an early warning level table; collecting audio data when the fan blades rotate; preprocessing the collected audio data when the fan blades rotate; extracting features from the preprocessed audio data when the fan blades rotate; optimizing the extracted features by dimensionality reduction; building a fault classifier model; training the fault classifier model by the feature information after dimensionality reduction optimization; collecting the rotating audio data of the fan blades that need to be early-warned for faults, and preprocessing, extracting features and optimizing the rotating audio data by dimensionality reduction, and then inputting the data into the trained fault classifier model, outputting the proportion of the data identified as faults; comparing the proportion of the data identified as faults with the early warning level table, thereby obtaining the fault early warning level of the fan blades that need to be early-warned for faults. The present invention has the advantages of not causing damage to the original structure of the blades and objective fault early warning.
Need to check novelty before this filing date? Find Prior Art