Multi-fault-mode high-performance binary-classifier of large wind power gearbox

A wind power gearbox and two-classifier technology, applied in character and pattern recognition, instruments, computer parts, etc., can solve the problem of reducing the generalization and promotion ability of the classifier, amplifying the negative effects of noise or outliers, and estimating the statistical characteristics of samples. And other issues

Inactive Publication Date: 2018-09-07
ZHEJIANG NORMAL UNIVERSITY
View PDF5 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

According to the existing research results, when there are local noises or outliers in the original observation data, its distribution often presents typical non-Gaussian characteristics, which increases the difficulty of feature extraction.
In addition, there are many negative factors affecting pattern classification, such as information redundancy of sensory observations, too high feature dimension selected in feature extraction, etc.
Information redundancy will directly cause difficulties in subsequent feature extraction, and further amplify the negative effects of noise or outliers; if the feature dimension is too high, it will make it more difficult to estimate the statistical characteristics of the sample, thereby reducing the generalization ability of the classifier [2]

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-fault-mode high-performance binary-classifier of large wind power gearbox
  • Multi-fault-mode high-performance binary-classifier of large wind power gearbox
  • Multi-fault-mode high-performance binary-classifier of large wind power gearbox

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] Multiple Failure Mode Classification

[0047] The spatial distribution of the fault mode characteristics of the three typical wind turbine gearboxes in normal state, gear tooth damage and frame motion is as follows: image 3 shown. Construct the training set using multimodal class feature samples:

[0048] [I w (k,l)] n =W(k,l)I n (x 0 - +k,y 0 - +l),

[0049] [I w (k,l)] g =W(k,l)I g (x 0 - +k,y 0 - +l),

[0050] ]I w (k,l)] l =W(k,l)I l (x 0 - +k,y 0 - +l).

[0051] where [·] n ,[·] g as well as[·] l are the feature training sets of the normal state, gear tooth damage and machine base transmission mode, respectively, I n (·,·), I g (·,·) and I l (·,·) are the size M respectively 1 × M 2 The spatial feature distribution image of .

[0052] First, the FLSA-SVM classification model is initialized, and the kernel function is selected as Radial Basis Function (RBF). The initial model is then tuned using a one-to-...

Embodiment 2

[0060] Multiple failure mode classification

[0061] Combining the training set and test set of the two modes of gear tooth damage and machine base loosening constitute the abnormal mode training set and test set, and use the FastICA algorithm to batch train two ICA feature extraction networks, thus forming [normal-abnormal Normal] feature extractor extracts two-dimensional quantized features of normal and abnormal patterns respectively. The number of training feature vectors is 20 and 40 respectively, and the same is true for testing feature vectors.

[0062] Two FLSA-SVM classifiers are trained by directly using the training samples of the four pattern categories of normal and abnormal, gear tooth damage and machine base looseness, and the kernel function still uses RBF. The classification test process is divided into two levels, the first level is used to identify normal and abnormal conditions, and the second level is used to identify tooth damage and frame loose mod...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a multi-fault-mode high-performance binary-classifier of a large wind power gearbox. The classifier is characterized in that a method comprises the following steps: 1) spatial-distribution characteristic description of multi-fault-mode class features; 2) multi-fault-mode direct classification based on support vector machines; 3) multi-fault-mode direct binary classificationbased on support vector machines; 4) valid-feature-sample recognition based on improved inter-intra-class clustering; 5) multi-fault-mode SVM direct-classification based on II-C feature sample recognition; and 6) multi-fault-mode SVM binary-classification based on II-C feature sample recognition. The classifier has the advantages that the method is reasonable in design, concise to use, and high in classification performance.

Description

technical field [0001] Based on the pattern recognition theory, the present invention proposes a multi-fault mode binary classification method on the basis of data cluster analysis and elimination of wild points and noise points, and utilizes the characteristics of automatic mode division of cluster analysis to eliminate wild points or noise points in feature data. On this basis, a binary classifier with multiple failure modes is developed based on the forward approximation least square support vector machine. This calculation method lays a theoretical foundation for solving the problem of pattern classification involved in the field of fault diagnosis of large wind turbine gearboxes. Background technique [0002] Pattern Classification is the Heart of Fault Diagnosis [1] . Only by reasonably classifying a variety of complex failure modes can the purpose of automatic and intelligent fault diagnosis be realized, and then correct decisions can be made, such as shutdown for m...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62
CPCG06F18/2411G06F18/214
Inventor 焦卫东杨志强
Owner ZHEJIANG NORMAL UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap