CN-122020349-A - Crack identification method, medium and terminal based on acoustic emission

Abstract

The invention is suitable for the technical field of crack monitoring, and relates to a crack identification method, medium and terminal based on acoustic emission, which comprises the steps of S10, extracting RA parameters and AF parameters of AE signals, constructing a two-dimensional feature set, S20, training a GMM model by adopting an EM algorithm, outputting a mean value, a covariance matrix and weights of each Gaussian distribution component, S30, determining posterior probability of each AE event in all Gaussian distributions, S40, comparing posterior probability of each component, taking a category corresponding to the maximum value as a classification result of the event crack type, S50, introducing a crack identification confidence index C i , selecting a threshold C th to judge the result, S60, introducing a crack mechanism energy dominant index to judge the energy dominant degree of a crack mechanism in the damage process, and assisting engineering analysis. The method has the advantages of simple flow and convenient operation, and effectively improves the accuracy of crack identification.

Inventors

• AN ZHIWEN
• GUO CHENGCHAO
• CAO DINGFENG
• GAO XING
• QIN LEI

Assignees

• 中山大学

Dates

Publication Date

20260512

Application Date

20250929

Claims (9)

1. 1. A crack identification method based on acoustic emission, characterized by comprising the steps of: s10, extracting RA parameters and AF parameters of AE signals to construct a two-dimensional feature set; S20, training a GMM model by adopting an EM algorithm, and outputting the mean value, covariance matrix and weight of each Gaussian distribution component; s30, determining posterior probability of each AE event in all Gaussian distributions; S40, comparing the posterior probabilities of the components, and taking the category corresponding to the maximum value as the classification result of the event crack type; S50, introducing a crack identification confidence index C i , selecting a threshold C th to judge a result, if C i ≥C th is the result of the event classification is reliable, and if Ci is less than C th , the result of the event classification needs to be carefully used; s60, introducing a crack mechanism energy dominance index for judging the energy dominance degree of the shearing mechanism in the damage process and assisting engineering analysis.
2. 2. The crack recognition method based on acoustic emission of claim 1, wherein the GMM model expression in step S20 is as follows: ; ; Wherein M is the number of single Gaussian models, ωk is the weight of each single Gaussian model, , As a probability density function of the kth SGM, In the form of a covariance matrix, Is desirable.
3. 3. The crack recognition method based on acoustic emission of claim 2, wherein the parameter estimation is performed using EM algorithm to construct an objective function To make it reach the maximum value, initializing the mean, covariance matrix and weight coefficient: ; ; 。
4. 4. A crack recognition method based on acoustic emission as claimed in claim 3, characterized in that the EM algorithm is divided into a desired step and a maximization step, in which the posterior probability of each gaussian component is calculated as follows: ; in the maximizing step, the model parameters are updated according to the calculated posterior probability as follows: ; ; ; Repeating the expectation step and the maximization step until convergence conditions are satisfied, wherein the difference between likelihood estimates of two times before and after satisfies a preset value, namely 。
5. 5. The method for crack recognition based on acoustic emission as set forth in claim 4, wherein, in the step S50, Threshold C th is 0.0002.
6. 6. The crack recognition method based on acoustic emission as claimed in claim 5, wherein in the step S40, the posterior probability Pk (x) after the output of the GMM model is used, where k=1, 2 is two gaussian components, and for each AE event, The probability that it belongs to two types of distributions is calculated: ; ; If it is The mixture was classified into component 1, component 2, and the mixture was judged to be a mixed crack when the mixture was equal.
7. 7. The crack recognition method based on acoustic emission of claim 6, wherein the expression in the step S60 is as follows: ; Wherein E i is the energy of the ith AE event, the posterior probabilities of which belong to shear cracks and tensile cracks are P shear, i and P tensile, i, the classification confidence of which is Ci= |P shear, i-P tensile, i|, cth is a confidence threshold, and only events with Ci being greater than or equal to Cth are included in statistics.
8. 8. A computer readable storage medium, characterized in that the storage medium stores a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 7.
9. 9. An electronic terminal is characterized by comprising a processor and a memory; The memory is configured to store a computer program, and the processor is configured to execute the computer program stored in the memory, to cause the terminal to perform the method according to any one of claims 1 to 7.

Description

Crack identification method, medium and terminal based on acoustic emission Technical Field The invention belongs to the technical field of crack monitoring, and particularly relates to a crack identification method, medium and terminal based on acoustic emission. Background The acoustic emission technology is used as a nondestructive detection means and is widely applied to health monitoring of materials and structures. AE events often reflect crack mechanisms by characteristic parameters (e.g., RA values, AF values) and distinguish shear cracks from tensile cracks accordingly. The Gaussian Mixture Model (GMM) is a soft clustering algorithm which can model complex distribution and provide posterior probability and is commonly used for AE parameter clustering analysis, but the GMM only provides a clustering structure, crack classification cannot be directly completed, and a discrimination boundary is required to be additionally constructed or a new method is required to be introduced for discrimination. In the existing crack classification method based on acoustic emission (Acoustic Emission, AE) characteristic parameters, the classification flow relies on manual boundary discrimination or additionally introduces a discrimination method (such as a support vector machine, SVM; bayesian information criterion, BIC), and the problems of complex classification steps, low accuracy, poor generalization capability and the like exist. Meanwhile, the existing method lacks credibility evaluation and mechanism dominant quantization means of classification results, and cannot provide clear and interpretable auxiliary judgment indexes for engineering practice. The patent application with publication number of CN116821726A provides a UHPC tensile shear crack judging method based on acoustic emission characteristic parameter cluster analysis, firstly, acoustic emission is utilized to synchronously monitor the UHPC damage fracture process to obtain acoustic emission characteristic parameters, RA values and AF values are calculated according to the characteristic parameters, and a label-free data set is constructed. And secondly, calculating the probability of each sample point generated by different clusters by using a Gaussian mixture model clustering algorithm, repeatedly obtaining the probability of updating the clusters until Gaussian convergence, dividing each sample into the cluster with the maximum probability to obtain two clusters, marking sample point data, displaying the marked sample data set in a mode that the RA value is taken as an x axis and the AF value is taken as a y axis, and obtaining an optimal linear boundary for dividing the two clusters by using a support vector machine algorithm, wherein the optimal linear boundary is used for dividing the tensile and shearing cracks. In the patent application, an SVM judging method is used, the SVM depends on a pseudo tag output by the GMM, the complexity of a model structure is increased, the demarcation mode does not have definite physical meaning, and the demarcation mode lacks stability and automatic classification capability, is not beneficial to rapid deployment and explanation in actual engineering, and has the same defects as the prior art. Therefore, how to provide a crack identification method with convenient operation and high identification accuracy is a problem to be solved by the person skilled in the art. Disclosure of Invention Aiming at the defects of the prior art, the invention aims to provide a crack identification method based on acoustic emission, so as to solve the problems of complex operation and low accuracy of the crack identification process in the prior art, and further provides a crack identification medium and a terminal based on acoustic emission. In order to solve the technical problems, the invention adopts the following technical scheme: in a first aspect, the present invention provides a crack identification method based on acoustic emission, comprising the steps of: s10, extracting RA parameters and AF parameters of AE signals to construct a two-dimensional feature set; S20, training a GMM model by adopting an EM algorithm, and outputting the mean value, covariance matrix and weight of each Gaussian distribution component; s30, determining posterior probability of each AE event in all Gaussian distributions; S40, comparing the posterior probabilities of the components, and taking the category corresponding to the maximum value as the classification result of the event crack type; S50, introducing a crack identification confidence index C i, selecting a threshold C th to judge a result, if C i≥Cth is the result of the event classification is reliable, and if Ci is less than C th, the result of the event classification needs to be carefully used; s60, introducing a crack mechanism energy dominance index for judging the energy dominance degree of the shearing mechanism in the damage process and assisting engineering analysis. Furt