Unlock instant, AI-driven research and patent intelligence for your innovation.

Fatigue monitoring method and system for metal parts based on hollow convolutional network

A technology of metal parts and convolutional network, applied in the field of fatigue monitoring of metal parts based on hollow convolutional network, can solve the problems of relying on expert knowledge, low monitoring accuracy, long monitoring cycle, etc., and achieve high degree of automation, high precision, good accuracy effect

Active Publication Date: 2020-09-22
BEIJING REALAI TECH CO LTD
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing technical schemes are based on feature extraction methods, and the effect of feature extraction is often not high. Feature extraction methods often rely heavily on expert knowledge. For a new machine and different piezoelectric signals, it is often necessary to perform feature extraction again, thereby bringing To monitor problems such as low accuracy and long monitoring cycle

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
  • Fatigue monitoring method and system for metal parts based on hollow convolutional network
  • Fatigue monitoring method and system for metal parts based on hollow convolutional network
  • Fatigue monitoring method and system for metal parts based on hollow convolutional network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] A method for fatigue monitoring of metal parts based on a hollow convolutional network is provided, comprising the following steps: see figure 1 and figure 2 , to perform non-destructive detection on the metal object. During the non-destructive detection process, a piezoelectric signal is applied to one end of the metal object and a piezoelectric signal is received at the other end of the metal object. The piezoelectric signal propagates in the metal object to form Lambo wave with time attribute and crack length attribute. As the cracks on the metal surface continue to increase, the received piezoelectric signal will change. see image 3 , as the crack length increases, the Lambert wave shows a general trend of increasing phase and decreasing amplitude.

[0056] Specifically, a convolution kernel of size k and a hole rate of size l are used to perform hole convolution on the time series signal f of the Lambert wave:

[0057]

[0058] Among them, τ represents the...

Embodiment 2

[0076] see Figure 4 , providing a metal parts fatigue monitoring system based on a hollow convolutional network, including:

[0077] The Lambert wave acquisition unit 1 is used for non-destructive detection of metal objects. During the non-destructive detection process, a piezoelectric signal is applied to one end of the metal object and a piezoelectric signal is received at the other end of the metal object. The piezoelectric signal Propagating in the metal object to form a Lambert wave with a time attribute and a crack length attribute; resampling the piezoelectric signal through the Lambert wave acquisition unit;

[0078] Hole convolution unit 2, for adopting the convolution kernel of k size and the hole rate of l size to carry out hole convolution to the time series signal f of described Lambert wave:

[0079]

[0080] Among them, τ represents the convolution extraction time, and t represents the dimension of the features extracted by the hollow convolution machine; ...

Embodiment 3

[0095] A computer-readable storage medium is provided, wherein the computer-readable storage medium stores program codes for fatigue monitoring of metal parts based on hollow convolutional networks, and the program codes include implementation of Embodiment 1 or any possible Instructions for implementing a method for fatigue monitoring of metal parts based on dilated convolutional networks.

[0096] The computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server, a data center, etc. integrated with one or more available media. The available medium may be a magnetic medium (for example, a floppy disk, a hard disk, or a magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, a solid state disk (SolidStateDisk, SSD)) and the like.

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 metal part fatigue monitoring method and system based on a hollow convolution network, and performs non-destructive detection on a metal object. During the non-destructive detection, a piezoelectric signal is applied to one end of the metal object and a pressure is received on the other end of the metal object. An electrical signal, the piezoelectric signal propagates in the metal object to form a Lambo wave with a time attribute and a crack length attribute; a convolution kernel of k size and a hole rate of l size are used to analyze the time series of the Lambo wave Atrous convolution is performed on the signal f, the piezoelectric signal is resampled and task regression is performed to obtain a neural network model for fatigue monitoring of metal parts. The invention establishes a machine learning model between the characteristics of the piezoelectric signal and the accurate metal crack length obtained by in-situ observation with an electron microscope. Through the machine learning model, it can realize the reflection of the metal fatigue degree only by means of simple and easy-to-operate piezoelectric signal transmission and reception. crack length prediction.

Description

technical field [0001] The invention relates to the technical field of metal monitoring, in particular to a method and system for fatigue monitoring of metal parts based on a hollow convolution network. Background technique [0002] Lamb wave (Lamb wave) is a kind of elastic wave, which is formed by the mutual coupling of transverse wave and longitudinal wave in a structure with two parallel surfaces. During the propagation of Lambert waves, the vibration displacement of the medium can be decomposed into two directions, that is, the direction along the wave propagation and the direction perpendicular to the wave propagation. [0003] At present, due to the inclusions, segregation or defects of the metal material for manufacturing parts, or due to unreasonable design, or due to unreasonable processing and manufacturing processes, etc., stress concentration often occurs in some parts of metal parts, and repeated stress alternating The lower metal parts will initiate cracks, 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): G01N29/04G01N29/44G06K9/00G06K9/62G06N3/04G06N3/08
CPCG01N29/041G01N29/4481G06N3/08G01N2291/0234G01N2291/0423G06N3/048G06N3/045G06F2218/02G06F2218/08G06F2218/12G06F18/241
Inventor 胡文波高嘉欣陈云天田天
Owner BEIJING REALAI TECH CO LTD