Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Bearing fault detection method based on variable learning rate multilayer perceptron

A multi-layer perceptron, fault detection technology, applied in neural learning methods, testing of mechanical components, testing of machine/structural components, etc., can solve problems such as economic loss, high accuracy, and inability to achieve classification accuracy High, fast feature speed, good robustness

Inactive Publication Date: 2022-05-06
TANGSHAN IND VOCATIONAL TECHN COLLEGE
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Rolling bearings are not only widely used in production, they are also vulnerable parts in mechanical equipment. According to statistics, about 30% of the faults in rotating machinery are caused by the failure of rolling bearings; after using fault diagnosis technology, the probability of mechanical equipment failure can be extremely high The maintenance cost is 25%-50% lower than before; due to the long life of the rolling bearing, according to the national standard, there is a 90% reliability threshold as a standard part, even if the same material uses the same process on the same machine The same batch of bearings is produced, and its service life is also different; usually in reality, regular inspection and repairs are carried out based on the designed fatigue life, which has a great chance of damaging the equipment and causing economic losses;
[0004] At present, in the practice of enterprises, if it is a rolling bearing in an unimportant occasion, it mainly depends on the experience of engineers and technicians to judge whether there is a fault in the operation process of the bearing through touch and hearing; in more important occasions, it is necessary to regularly disassemble and check the operation of the bearing , but it is impossible to predict the fault state in a timely and effective manner; at the same time, the current detection algorithms for bearing faults mainly focus on the traditional use of time-domain signals as signal input features, and then use machine learning or deep learning methods for classification, with high accuracy However, it has not yet reached the conditions that meet the industrial testing applications of core components;

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 method based on variable learning rate multilayer perceptron
  • Bearing fault detection method based on variable learning rate multilayer perceptron
  • Bearing fault detection method based on variable learning rate multilayer perceptron

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0042] Embodiment 1: with reference to attached figure 1 Shown is a bearing fault detection method based on variable learning rate multi-layer perceptron, including steps

[0043] Step1. Use the sensor to collect and extract the vibration signal of the bearing

[0044] Use the acceleration sensor to measure the vibration signal of the experimental bearing under the normal state and the four conditions of inner ring fault, outer ring fault, and rolling element fault, and obtain the sample point data of the vibration signal including the normal state and fault state;

[0045] Step2. Input the bearing vibration signal collected by the sensor into the variable learning rate multilayer perceptron, and perform feature iteration on the vibration signal in the fully connected layer network;

[0046] Step201. Input the vibration signal extracted in Step 1 into the fully connected layer 1 for feature extraction;

[0047] Wherein: there are 1024 input channels and 256 output channels i...

Embodiment 2

[0068] Embodiment 2: as Figure 2-3 As shown, in order to prove the effectiveness of this method, the vibration data in the United States is used from the open data set of Case Western Reserve University in the United States:

[0069] (1) In this experiment, the bearings at the drive end of the motor and the fan end are used as the diagnostic objects, and single-point damage is introduced on the inner ring, rolling element, and outer ring of the test bearings by EDM to simulate three types of bearing failures. The fault damage scales are 0.007inch, 0.014inch and 0.021inch respectively, and then under similar working conditions (equal load, close speed), the fault signal is collected by the acceleration sensor on the upper side of the motor drive end, and the sampling frequency is 12kHz. According to the location and size of the bearing fault, the bearing categories are divided into ten categories, with 1024 data points as a sample, and 1600 samples are selected for each catego...

Embodiment 3

[0075] Embodiment 3: use faulty bearing to carry out comparative test to the fault detection method of different rolling bearings below, further the detection method of the present invention is verified:

[0076] Step 1: Select a set of rolling bearing data, and the failure conditions of the bearings are shown in Table 1:

[0077] Table 1: Fault information of rolling bearings

[0078]

[0079] Step 2: use time-frequency feature+SV, ED+SV, wavelet+SV, CNN-SV and diagnostic methods based on variable learning rate multi-layer perceptron of the present invention to detect the faulty bearing described in step 1 respectively, and The detected results are compared with the data in Table 1 above, and the comparison results of the time-consuming and accuracy of each detection method are shown in Table 2:

[0080] Table 2: Comparison of accuracy rates of different methods

[0081]

[0082] As can be seen from the results in Table 2, under the same conditions, the diagnostic met...

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 bearing fault detection method based on a variable learning rate multi-layer perceptron. The method comprises the following steps: Step 1, collecting and extracting a vibration signal of a bearing by using a sensor; step 2, carrying out feature iteration on the vibration signal by using a multilayer perceptron; 3, a back propagation process is used for automatically optimizing the weight coefficient of the multi-layer perceptron, the learning rate of automatic optimization is changed according to the number of iterations, time domain signals obtained after iteration enter a support vector machine classifier, and a diagnosis model is obtained; step 4, inputting test data into the diagnosis model for diagnosis, and completing the detection of the bearing fault; according to the method, the bearing signal faults are extracted in a crossed mode through the multi-layer perceptron based on the variable learning rate, the loss function value can be rapidly reduced to complete fault diagnosis, the feature extraction speed is high, data do not need to be preprocessed, diagnosis precision is high, and the method has the advantages of being rapid in feature extraction, high in precision and good in robustness.

Description

technical field [0001] The invention relates to the technical field of fault detection, in particular to a bearing fault detection method based on a variable learning rate multilayer perceptron. Background technique [0002] The development of modern industry has greatly improved the automation of equipment. The upgrading of intelligent manufacturing industry needs to continuously strengthen the reliability and efficiency of mechanical equipment. High-efficiency health monitoring systems are widely used in rotating mechanical equipment typified by rolling bearings. Bearing condition monitoring is increasingly important in industry, as improved bearing reliability can significantly reduce machine breakdowns, which can lead to major production losses and safety incidents. [0003] Rolling bearings are not only widely used in production, they are also vulnerable parts in mechanical equipment. According to statistics, about 30% of the faults in rotating machinery are caused by...

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): G01M13/045G06K9/00G06K9/62G06N3/04G06N3/08
CPCG01M13/045G06N3/084G06N3/045G06F2218/08G06F2218/12G06F18/2411
Inventor 王萌
Owner TANGSHAN IND VOCATIONAL TECHN COLLEGE
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products