Bearing fault detection and classification integrated method based on representation learning

A fault detection and fault classification technology, applied in the interdisciplinary field, can solve the problems of difficult bearing fault samples and label limitations, and achieve the effect of improving the reconstruction effect and the accuracy rate.

Active Publication Date: 2022-07-08
HARBIN INST OF TECH
View PDF5 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem that it is very difficult to obtain a large number of bearing fault samples in the existing data-driven method for bearing fault diagnosis, and

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
  • Bearing fault detection and classification integrated method based on representation learning
  • Bearing fault detection and classification integrated method based on representation learning
  • Bearing fault detection and classification integrated method based on representation learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0038] Embodiment 1: The specific process of an integrated method for bearing fault detection and classification based on representation learning in this embodiment is as follows:

[0039] In the present invention, the neural network is used to represent the learning method;

[0040] Step 1. Extract the vibration data of the bearing:

[0041] Through the sensor equipment on the bearing housing, such as wireless sensors, etc., the vibration data of the bearing during the working process is collected as the input of the integrated method of bearing fault detection and classification based on representation learning;

[0042] The vibration data includes health data and fault data (10% of the fault data is used for offline training);

[0043] Step 2. Feature extraction:

[0044] In order to better reflect the healthy state of the bearing during the working process, it is considered to extract the characteristics of the vibration data of the bearing during the working process fro...

specific Embodiment approach 2

[0059] Embodiment 2: The difference between this embodiment and Embodiment 1 is that in the feature extraction in step 2: in order to better reflect the health status of the bearing during the working process, it is considered to extract the bearing from the perspectives of the time domain and the frequency domain. The characteristics of the vibration data during the working process, thus serving as the input of the neural network;

[0060] The characteristics of the vibration data include the characteristics of the health data and the characteristics of the fault data;

[0061] The specific process is:

[0062] In order to better reflect the healthy state of the bearing during the working process, the present invention considers extracting the characteristics of the vibration data generated by the bearing during the working phase from the time domain, frequency domain and these two perspectives, and thus serves as the input of the neural network. As a commonly used signal fe...

specific Embodiment approach 3

[0084] Embodiment 3: The difference between this embodiment and Embodiment 1 or 2 is that a fault detection neural network is built in step 3: the fault detection neural network includes an encoder, a noise introduction, a self-attention mechanism layer, and a decoder;

[0085] In order to reduce the influence of noise signal on sample reconstruction, and assign different weights to different bottleneck layer neurons, so as to effectively realize fault detection, the present invention proposes a self-attention mechanism based on bottleneck layer neurons to correct and denoise Neural network of Modified denoising auto-encoder with Self-attention mechanism for bottleneck layer, MDAE-SAMB. The MDAE-SAMB network is trained by using only the features obtained by feature extraction from the vibration data of the bearing in a healthy state. The MDAE-SAMB structure is mainly composed of four parts: encoder, noise introduction, self-attention mechanism layer and decoder.

[0086] The ...

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 relates to a bearing fault detection and classification integrated method, in particular to a bearing fault detection and classification integrated method based on representation learning. The objective of the invention is to solve the problems that a large number of bearing fault samples are very difficult to obtain in the bearing fault diagnosis process by adopting a data driving method, and the design of accurate labels for data of different fault types is greatly limited. The method comprises the following steps: 1, extracting vibration data of a bearing; 2, feature extraction: extracting the features of the data from two angles of time domain and frequency domain; 3, building a fault detection neural network, wherein the fault detection neural network comprises an encoder, a noise introduction layer, a self-attention mechanism layer and a decoder; 4, training a fault detection neural network; 5, building a fault classification neural network; 6, training the fault classification neural network; and 7, carrying out online fault detection and fault classification. The method is used in the subject crossing field of bearing fault diagnosis and artificial intelligence combination.

Description

technical field [0001] The invention belongs to the interdisciplinary field of the combination of bearing fault diagnosis and artificial intelligence, and specifically relates to an integrated method of bearing fault detection and classification based on representation learning. Background technique [0002] As an important component in precision mechanical equipment, the health of bearings plays a vital role in the normal operation of the machine. How to diagnose bearing faults in time and reduce the losses caused by their faults is of critical significance to the safety and reliability of mechanical equipment. [0003] Fault diagnosis mainly includes two aspects of fault detection and classification. The current fault diagnosis methods for bearings can be generally divided into model-based methods and data-driven methods. Model-based fault diagnosis methods need to rely on rich prior knowledge to construct bearing fault models. However, considering the changing working ...

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
IPC IPC(8): G06K9/62G06F30/27G06N3/04G06N3/08
CPCG06F30/27G06N3/08G06N3/048G06N3/045G06F18/2415G06F18/241Y02T90/00
Inventor 张九思李翔罗浩张可安翼尧田纪伦尹珅
Owner HARBIN INST OF 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