Fault diagnosis method based on compressed sensing and improved multi-scale network

A fault diagnosis and compressive sensing technology, applied in biological neural network models, testing of mechanical components, testing of machine/structural components, etc., can solve problems such as difficulties in the optimization process of multi-scale network diagnosis, difficulty in applying small samples, etc. The effect of optimizing the process is difficult, improving the accuracy, and improving the efficiency of data analysis

Active Publication Date: 2019-06-07
HUAZHONG UNIV OF SCI & TECH
View PDF11 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects of the prior art, the purpose of the present invention is to solve the technical problems in the prior art that the multi-scale network diagnosis and optimization process is difficult and difficult to apply to small samples

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
  • Fault diagnosis method based on compressed sensing and improved multi-scale network
  • Fault diagnosis method based on compressed sensing and improved multi-scale network
  • Fault diagnosis method based on compressed sensing and improved multi-scale network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0066] Example 1: Validity Verification

[0067] In order to verify the effectiveness of the fault diagnosis method proposed in the present invention, this embodiment uses the public experimental data set provided by the Bearing Data Center of Case Western Reserve University in the United States for verification. The experimental bench includes: a driving motor, bearing at the driving end, a switch encoder, and a dynamometer. Set the speed of the driving motor at 1750 rpm and the load at 2Hp. Collect the vibration signal at the 6 o'clock position at the driving motor end. The sampling rate is 12kHz. The fault types of rolling bearings are inner ring defects, outer ring defects and ball defects. The severity of the faults is simulated by EDM, and the fault diameters are 7, 14, and 21 mils (mil) respectively. The details of the 10 bearing condition types are shown in Table 1. 2400 data samples are collected under each bearing state type, and each data sample contains 4096 data...

Embodiment 2

[0083] Example 2: Hyperparameter Settings

[0084] The present invention uses compressed sensing technology to compress raw sensor data, which improves data efficiency to a certain extent, and studies the performance of data-driven fault diagnosis methods using different hyperparameters in the case of compressed sensing, including calculation time and classification Accuracy; Comparing computation time with and without Compressed Sensing. Its specific implementation steps are as follows figure 1 As shown, the parameter settings of the deep learning process are shown in Table 3. A total of 10 repeated experiments were carried out, and the average results were calculated and used for the data.

[0085] table 3

[0086]

[0087] Experimental results such as Image 6 , Figure 7 As shown, the experimental results show that for a specific number of feature maps C, as the number of neurons in the output layer increases, the classification accuracy of rolling bearings increase...

Embodiment 3

[0094] Example 3: Small sample problem

[0095] Generally, the performance of deep model-based data-driven fault diagnosis methods depends on the size of the sensing dataset, and when more training datasets are used, the accuracy of fault diagnosis increases, however, it is difficult to obtain sufficient monitoring sensor data. The present invention first uses the measurement matrix of compressed sensing to generate enough training samples for subsequent deep learning analysis. Its specific implementation steps are as follows figure 1 As shown, the parameter settings of the small sample training process are shown in Table 6. A total of 10 repeated experiments were carried out, and the average results were calculated and used for the data.

[0096] Different bearing state types (see Table 1) randomly select 200 training samples, so the total number of training samples for 10 bearing state types is 2000. In this example, the Gaussian measurement matrix is ​​used to generate e...

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 fault diagnosis method based on compressed sensing and an improved multi-scale network, which adopts compressed sensing to compress a sample, can retain interested information from an original sample signal, and improves the data analysis efficiency. A compressed sensing Gaussian measurement matrix is multiplied by a signal matrix. Due to the fact that the Gaussian measurement matrix is a random matrix, matrix elements generated each time are different, generated training samples are different, enough training samples are generated, the problem of small samples existing in the training process of the depth model is further solved, and the accuracy rate of small sample fault diagnosis is increased. Constant mapping is introduced between feature mapping with the same size and the same feature mapping number. Through feature extraction of different layers, features of different layers are connected into a whole to achieve feature fusion, a deep fusion feature vector corresponding to an input image is obtained, and therefore the problem that the multi-scale network optimization process is difficult is solved, and the fault type can be effectively recognized under different working conditions.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis of rotating machinery, and more specifically relates to a fault diagnosis method based on compressed sensing and improved multi-scale network. Background technique [0002] Rotating machinery will suffer various damages and failures under severe and extreme working conditions, and its work failure will affect the performance of the entire system and may cause the system to shut down, causing heavy losses. In order to avoid failures, ensure safe operation of equipment, and reduce economic losses, intelligent diagnosis and prediction of rotating machinery systems are required. However, traditional fault diagnosis methods have limitations, and their diagnostic performance largely depends on expert experience and prior knowledge, and the models used for prediction often have shallow network structures. [0003] Convolutional neural network has been widely used in fault diagnosis in recent yea...

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): G06K9/00G06K9/46G06K9/62G06N3/04G01M13/00
Inventor 刘颉葛子月胡中旭葛明峰周奇许颜贺周凯波张立丽张志鑫
Owner HUAZHONG 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