Rolling bearing fault identification method based on the combination of spa-map and resnet

A rolling bearing and fault identification technology, which is applied in neural learning methods, character and pattern recognition, machine/structural component testing, etc., can solve problems such as large gradients, gradient dispersion, difficult network training, etc., to improve efficiency and accuracy, Simplify the preprocessing process and reduce the effect of component items

Active Publication Date: 2021-12-03
SOUTHWEST JIAOTONG UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although CNN has a better ability to extract image features, and as the number of network layers deepens, the network can extract deeper feature information and obtain good generalization ability, but when the number of network layers deepens, the network will change. It is becoming more and more difficult to train, because in a network with a deeper number of layers, when the gradient information is transmitted from the last layer to the first layer of the network, there will be a phenomenon that the gradient is close to zero or the gradient is very large, which is called gradient dispersion or Gradient explosion, the deeper the number of layers in the network, the more serious the phenomenon of gradient dispersion or gradient explosion will be, which will make it more difficult to extract deep-level features

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
  • Rolling bearing fault identification method based on the combination of spa-map and resnet
  • Rolling bearing fault identification method based on the combination of spa-map and resnet
  • Rolling bearing fault identification method based on the combination of spa-map and resnet

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0081] The experimental data used in the present embodiment is the 6205-2RS JME SKF deep groove ball bearing data set in the rolling bearing fault vibration signal opened by case Western Reserve University (CWRU) (SMITH W, RANDALL R. Rolling elementbearing diagnostics using case western reserve university data: a benchmarkstudy[ J]. Mechanical Systems and Signal Processing, 2015, 64-65(3): 100-131), its fault types are divided into inner ring fault, rolling element fault and outer ring fault, each fault is introduced by the EDM method single point fault, the fault diameter is 0.007inches, the load on each fault is 0, 1HP, 2HP and 3HP, and the data are used to construct a comparison with normal bearings under the same circumstances. See Table 1 for details.

[0082] Table 1 16 operating conditions under normal and 0.007 fault

[0083]

[0084]

[0085] When dividing the data, according to the general training set, verification set and test set division principle, the data is di...

Embodiment 2

[0211] In order to further observe the failure identification method of the rolling bearing provided by the present invention in the classification of the fault site, and additionally select the degree of damage of 0.014inches and 0.021inches of data in accordance with the method given in Example 1 2 sets of 16 classification experiments (the specific working conditions are shown in Table 4). First of all, the data is divided into training sets, validation sets and test sets, and then the data in the training set and the validation set are used to train the ResNet network model (the same structure as in Example 1) in Steps A1-A6, and then the data in the test set are used to test the resNet network model trained according to steps S1-S4, and the experimental results and the experimental results under 0.007inches are recorded in Table 3. As can be seen from Table 3, the accuracy rate is more than 99% under all three types of damage, indicating that the method can effectively identi...

Embodiment 3

[0215] In order to further observe the performance of the rolling bearing fault identification method provided by the present invention in the classification of the degree of failure, all normal data and all the inner ring under load, the outer ring (6 o'clock direction) and the degree of damage of the rolling element, each set of 16 kinds of data to carry out 3 groups of experiments, wherein the inner ring condition as shown in Table 4, the outer ring (6 o'clock direction) and the rolling body working condition is similar to Table 4. First of all, the data is divided into training sets, validation sets and test sets, and then the data in the training set and the validation set are used to train the ResNet network model (the same structure as in Example 1) in Steps A1-A6, and then the data in the test set are used to test the resNet network model trained in steps S1-S4, and the experimental results are recorded in Table 5.

[0216] Table 4 16 operating conditions of normal and inn...

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 rolling bearing fault identification method based on the combination of SPA-atlas and ResNet. First, the original signal is decomposed into trend items and de-trend items with large differences by using the SPA method, and then the obtained trend items and de-trend items are Combining the original signal items into a color map, and finally using the ResNet network model to realize fault type identification, while extracting as much bearing fault information as possible, greatly reducing the component items, and at the same time extracting deep bearing fault information through deep ResNet to improve rolling bearing faults Recognition efficiency and accuracy.

Description

Technical field [0001] The present invention belongs to the field of rotating machinery fault identification technology, relates to rolling bearing fault identification, specifically to a rolling bearing fault identification method based on SPA- map combined with ResNet. Background [0002] Rolling bearings are one of the most commonly used components in mechanical systems, which play an extremely important role in maintaining the stability of mechanical systems, but are also one of the parts that are easily damaged. According to the statistics of mechanical system failures, the proportion of rolling bearing failures exceeds 40%. When the bearing is damaged, it will lead to abnormal mechanical system work, serious will lead to major accidents, and then cause economic losses and casualties, so to achieve accurate bearing fault diagnosis is crucial to ensure the normal operation of the mechanical system. Because rolling bearings mostly work under complex working conditions, once a ...

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): G01M13/045G06K9/00G06N3/04G06N3/08
CPCG01M13/045G06N3/08G06N3/045G06F2218/12
Inventor 张敏李贤均程文明
Owner SOUTHWEST JIAOTONG UNIV
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