A fault identification method of rolling bearing under variable working conditions based on att-cnn

A technology of ATT-CNN and rolling bearings, applied in mechanical bearing testing, geometric CAD, biological neural network models, etc., can solve problems such as generalization ability constraints

Active Publication Date: 2022-07-19
HARBIN UNIV OF SCI & TECH
View PDF11 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] In order to solve the problem that the feature extraction in the existing rolling bearing fault identification method under variable working conditions relies too much on prior knowledge and expert experience due to the use of time-domain features, frequency-domain features or time-frequency domain features, and the feature extraction and model establishment are carried out in isolation , so that its generalization ability is limited for complex classification problems, and then provides a rolling bearing fault identification method based on ATT-CNN (attention mechanism CNN) under variable working conditions

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
  • A fault identification method of rolling bearing under variable working conditions based on att-cnn
  • A fault identification method of rolling bearing under variable working conditions based on att-cnn
  • A fault identification method of rolling bearing under variable working conditions based on att-cnn

Examples

Experimental program
Comparison scheme
Effect test

experiment approach 2

[0140] In order to verify the generalization ability of the algorithm proposed in this paper, experiments are carried out on data of all load types. All state data sets under two loads are used as the training set, and the data under the other two different loads are used as the test set. The data settings and experimental results are shown in Table 4. The structure of multi-condition type data is more complex and diverse, and it is more difficult to identify the fault state when testing and diagnosing single-condition data.

[0141] Table 4 Experimental scheme 2 data set settings and accuracy comparison

[0142]

[0143] Through the analysis of multiple sets of experiments, the average test accuracy of the ATT-CNN model is 96.48%, which is 3.3% higher than that of CNN. It shows that under more complex working conditions, the data structure and distribution of the test set and the training set are quite different. The ATT-CNN model can well adapt to changes in data distrib...

experiment approach 3

[0146] In order to further verify the robustness of the model, the data under one load is used as the training set, and the data under the other three loads is used as the test set. The data settings and experimental results are shown in Table 5. As we all know, for deep neural networks, a more robust model can be obtained by using massive training data. However, in practical applications, some working condition data are difficult to obtain or unknown working condition data may appear. Therefore, it is hoped that a small data set can be used to realize fault diagnosis under variable working conditions.

[0147] Table 5 Experimental scheme three data set settings and accuracy comparison

[0148]

[0149]

[0150] The average test accuracy of the ATT-CNN-based model is 83.40%, while the unimproved CNN model is only 77.98%, which is 5.5% higher. The overall performance of the proposed algorithm on the dataset set by scheme three is not as good as that of the first two sche...

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

An ATT-CNN-based rolling bearing fault identification method under variable working conditions relates to a rolling bearing fault identification technology. In order to solve the problem that the generalization ability of the existing rolling bearing fault identification method under variable working conditions is restricted to a certain extent for complex classification problems. First, the vibration data is mapped to the nonlinear spatial domain through a convolutional neural network (CNN), and the faults of rolling bearings under variable working conditions are adaptively extracted by using its invariance to the small displacement, scaling and other forms of distortion of the input signal. Secondly, it is proposed to integrate the attention mechanism (ATT) idea into the CNN structure to further improve the sensitivity of bearing vibration characteristics under variable working conditions; at the same time, more abundant and diverse training samples are obtained through data enhancement methods, so that the network can be more fully developed. learning, improving robustness. The proposed fault diagnosis model based on attention mechanism CNN (ATT‑CNN) can realize multi-state identification and classification of rolling bearings under variable working conditions, and compared with other methods, it can obtain higher accuracy.

Description

technical field [0001] The invention relates to a rolling bearing fault identification method under variable working conditions, and relates to a rolling bearing fault identification technology. Background technique [0002] The role of rolling bearings in industrial production is light, and they often work in the state of load and speed changes. Therefore, it is important to effectively diagnose the performance state of rolling bearings during the evolution of rolling bearing failures under variable working conditions and improve the reliability of mechanical equipment. significance [1] . In the field of fault diagnosis, the use of condition monitoring data combined with artificial intelligence methods has gradually become a research hotspot in recent years. In particular, the fault signals of rolling bearings under variable working conditions contain richer vibration characteristic information. [2] , the dynamic characteristics of its fault development also pose new cha...

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 Patents(China)
IPC IPC(8): G06F30/17G01M13/04G06N3/04
Inventor 梁欣涛康守强李艺伟王玉静王庆岩
Owner HARBIN UNIV OF SCI & TECH
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