CN115901954A - Nondestructive detection method for ultrasonic guided wave pipeline defects - Google Patents

Abstract

The method adopts an artificial intelligence detection technology based on deep learning to detect whether the pipeline to be detected has defects or not by comparing the multi-scale characteristic difference of a detection ultrasonic guided wave signal of the pipeline to be detected and a reference ultrasonic guided wave signal of a non-defective pipeline in a high-dimensional space. Thus, the nondestructive inspection of the pipe defect can be accurately performed.

Description

Nondestructive testing method for pipeline defects using ultrasonic guided waves

Technical Field

The present application relates to the technical field of nondestructive testing, and more specifically, to a method for nondestructive testing of pipeline defects using ultrasonic guided waves.

Background Art

Pipeline laying involves all walks of life and undertakes an important production mission. Most pipelines need to face various complex environments, such as bad weather, impact, corrosion of chemical substances inside the pipeline, etc., so pipelines are prone to various defects. In engineering, incidents where pipeline defects cause great damage to people's survival and property safety frequently occur. Therefore, the detection of pipeline health status is particularly important.

At present, pipeline maintenance personnel generally detect pipeline damage through visual inspection, electromagnetic flaw detection and other methods. These methods are not only time-consuming but also have low detection accuracy.

Therefore, an optimized nondestructive detection scheme for pipeline defects is desired.

Summary of the invention

In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an ultrasonic guided wave pipeline defect non-destructive detection method, which adopts artificial intelligence detection technology based on deep learning to detect and judge whether the pipeline to be detected has defects by comparing the multi-scale feature differences of the detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in a high-dimensional space. In this way, non-destructive detection of pipeline defects can be accurately performed.

According to one aspect of the present application, a method for nondestructive detection of pipeline defects using ultrasonic guided waves is provided, which includes: obtaining a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of a pipeline to be detected, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline; performing Fourier transform on the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal respectively to obtain a plurality of detection frequency domain statistical values and a plurality of reference frequency domain statistical values; passing the waveform diagrams of the plurality of detection frequency domain statistical values and the detection ultrasonic guided wave signal through a first Clip model to obtain a detection ultrasonic feature matrix; passing the plurality of reference frequency domain statistical values and the reference ultrasonic guided wave signal through a second Clip model to obtain a reference ultrasonic feature matrix; calculating a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix; correcting the eigenvalues of each position of the differential feature matrix to obtain a corrected differential feature matrix; and passing the corrected differential feature matrix through a classifier to obtain a classification result, wherein the classification result indicates whether the pipeline to be detected has defects.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the multiple detection frequency domain statistical values and the waveform diagram of the detection ultrasonic guided wave signal are passed through the first Clip model to obtain a detection ultrasonic feature matrix, including: inputting the multiple detection frequency domain statistical values into the sequence encoder of the first Clip model to obtain a detection frequency domain statistical feature vector; inputting the waveform diagram of the detection ultrasonic guided wave signal into the image encoder of the first Clip model to obtain a detection ultrasonic guided wave image feature vector; and inputting the detection ultrasonic guided wave image feature vector and the detection frequency domain statistical feature vector into the encoding optimizer of the first Clip model to obtain the detection ultrasonic feature matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive testing method, the inputting the multiple detection frequency domain statistical values into the sequence encoder of the first Clip model to obtain the detection frequency domain statistical feature vector includes: inputting the multiple detection frequency domain statistical values into the first convolution layer of the multi-scale neighborhood feature extraction module to obtain the first-scale detection frequency domain statistical feature vector, wherein the first convolution layer has a first one-dimensional convolution kernel of a first length; inputting the multiple detection frequency domain statistical values into the second convolution layer of the multi-scale neighborhood feature extraction module to obtain the second-scale detection frequency domain statistical feature vector, wherein the second convolution layer has a second one-dimensional convolution kernel of a second length, and the first length is different from the second length; and cascading the first-scale detection frequency domain statistical feature vector and the second-scale detection frequency domain statistical feature vector to obtain the detection frequency domain statistical feature vector.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the waveform diagram of the detection ultrasonic guided wave signal is input into the image encoder of the first Clip model to obtain the detection ultrasonic guided wave image feature vector, including: using each layer of the image encoder of the first Clip model to perform in the forward transfer of the layer: convolution processing on the input data to obtain a convolution feature map; mean pooling based on the local feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the image encoder of the first Clip model is the detection ultrasonic guided wave image feature vector, and the input of the first layer of the image encoder of the first Clip model is the waveform diagram of the detection ultrasonic guided wave signal.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the step of inputting the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detected ultrasonic feature matrix comprises: inputting the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detected ultrasonic feature matrix according to the following formula; wherein the formula is:

表示所述检测超声导波图像特征向量的转置向量，V₂表示所述检测频域统计特征向量，M表示所述检测超声特征矩阵，

表示矩阵相乘。in

represents the transposed vector of the detected ultrasonic guided wave image feature vector, V ₂ represents the detected frequency domain statistical feature vector, M represents the detected ultrasonic feature matrix,

Represents matrix multiplication.

其中M₁表示所述检测超声特征矩阵，M₂表示所述参考超声特征矩阵，M_c表示所述差分特征矩阵。In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the calculation of the differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix includes: calculating the differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix using the following formula; wherein the formula is:

Wherein, _M1 represents the detected ultrasonic feature matrix, _M2 represents the reference ultrasonic feature matrix, and _Mc represents the differential feature matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the eigenvalues of each position of the differential feature matrix are corrected to obtain the corrected differential feature matrix, including: calculating the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by full orthographic projection nonlinear reweighting; and multiplying the differential feature matrix by the weight matrix at each position to obtain the corrected differential feature matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated by using the full orthogonal projection nonlinear reweighting method, including: using the full orthogonal projection nonlinear reweighting method to calculate the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by the following formula; wherein the formula is:

表示矩阵相乘，且分子矩阵和分母矩阵之间的除法为矩阵特征值的按位置相除，exp(·)表示矩阵的指数运算，所述矩阵的指数运算表示计算以矩阵中各个位置的特征值为幂的自然指数函数值。Wherein, _M1 and _M2 are the detection ultrasonic feature matrix and the reference ultrasonic feature matrix respectively, _Mw is the weight matrix, ReLU(·) represents the ReLU activation function,

represents matrix multiplication, and the division between the numerator matrix and the denominator matrix is the positional division of the matrix eigenvalues. exp(·) represents the exponential operation of the matrix, and the exponential operation of the matrix represents the calculation of the natural exponential function value raised to the power of the eigenvalues at each position in the matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection method, the corrected differential feature matrix is passed through a classifier to obtain a classification result, and the classification result indicates whether there is a defect in the pipeline to be detected, including: expanding the corrected differential feature matrix into a classification feature vector according to a row vector or a column vector; using the fully connected layer of the classifier to fully connect the classification feature vector to obtain an encoded classification feature vector; and inputting the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

According to another aspect of the present application, an ultrasonic guided wave pipeline defect nondestructive detection system is provided, comprising: an ultrasonic guided wave acquisition module, used to acquire a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of a pipeline to be detected, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline; a domain change module, used to perform Fourier transform on the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal, respectively, to obtain a plurality of detection frequency domain statistical values and a plurality of reference frequency domain statistical values; a detection ultrasonic encoding module, used to convert the waveform diagrams of the plurality of detection frequency domain statistical values and the detection ultrasonic guided wave signal into a waveform diagram through a first Clip model; A detection ultrasonic feature matrix is obtained; a reference ultrasonic encoding module is used to pass the multiple reference frequency domain statistical values and the reference ultrasonic guided wave signal through a second Clip model to obtain a reference ultrasonic feature matrix; a difference module is used to calculate a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix; an eigenvalue correction module is used to correct the eigenvalues of each position of the differential feature matrix to obtain a corrected differential feature matrix; and a detection result generation module is used to pass the corrected differential feature matrix through a classifier to obtain a classification result, wherein the classification result indicates whether there is a defect in the pipeline to be detected.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system, the detection ultrasonic encoding module includes: a sequence encoding unit, which is used to input the multiple detection frequency domain statistical values into the sequence encoder of the first Clip model to obtain a detection frequency domain statistical feature vector; an image encoding unit, which is used to input the waveform diagram of the detection ultrasonic guided wave signal into the image encoder of the first Clip model to obtain a detection ultrasonic guided wave image feature vector; and a coding optimization unit, which is used to input the detection ultrasonic guided wave image feature vector and the detection frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detection ultrasonic feature matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive testing system, the sequence encoding unit is further used to: input the multiple detection frequency domain statistical values into the first convolution layer of the multi-scale neighborhood feature extraction module to obtain a first-scale detection frequency domain statistical feature vector, wherein the first convolution layer has a first one-dimensional convolution kernel of a first length; input the multiple detection frequency domain statistical values into the second convolution layer of the multi-scale neighborhood feature extraction module to obtain a second-scale detection frequency domain statistical feature vector, wherein the second convolution layer has a second one-dimensional convolution kernel of a second length, and the first length is different from the second length; and cascade the first-scale detection frequency domain statistical feature vector and the second-scale detection frequency domain statistical feature vector to obtain the detection frequency domain statistical feature vector.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system, the image encoding unit is further used to: use each layer of the image encoder of the first Clip model to perform in the forward transfer of the layer: convolution processing on the input data to obtain a convolution feature map; mean pooling based on the local feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the image encoder of the first Clip model is the detected ultrasonic guided wave image feature vector, and the input of the first layer of the image encoder of the first Clip model is the waveform diagram of the detected ultrasonic guided wave signal.

In the above ultrasonic guided wave pipeline defect nondestructive testing system, the coding optimization unit is further used to: input the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detected ultrasonic feature matrix according to the following formula; wherein the formula is:

表示所述检测超声导波图像特征向量的转置向量，V₂表示所述检测频域统计特征向量，M表示所述检测超声特征矩阵，

表示矩阵相乘。in

represents the transposed vector of the detected ultrasonic guided wave image feature vector, V ₂ represents the detected frequency domain statistical feature vector, M represents the detected ultrasonic feature matrix,

Represents matrix multiplication.

其中M₁表示所述检测超声特征矩阵，M₂表示所述参考超声特征矩阵，M_c表示所述差分特征矩阵。In the above ultrasonic guided wave pipeline defect nondestructive detection system, the differential module is further used to calculate the differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix using the following formula; wherein the formula is:

Wherein, _M1 represents the detected ultrasonic feature matrix, _M2 represents the reference ultrasonic feature matrix, and _Mc represents the differential feature matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system, the eigenvalue correction module includes: a weight unit, which calculates the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by full orthographic projection nonlinear reweighting; and an application unit, which multiplies the differential feature matrix by the weight matrix at the position point to obtain the corrected differential feature matrix.

In the above ultrasonic guided wave pipeline defect nondestructive testing system, the weight unit is further used to: calculate the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by the following formula using a full orthographic projection nonlinear reweighting method; wherein the formula is:

表示矩阵相乘，且分子矩阵和分母矩阵之间的除法为矩阵特征值的按位置相除，exp(·)表示矩阵的指数运算，所述矩阵的指数运算表示计算以矩阵中各个位置的特征值为幂的自然指数函数值。Wherein, _M1 and _M2 are the detection ultrasonic feature matrix and the reference ultrasonic feature matrix respectively, _Mw is the weight matrix, ReLU(·) represents the ReLU activation function,

represents matrix multiplication, and the division between the numerator matrix and the denominator matrix is the positional division of the matrix eigenvalues. exp(·) represents the exponential operation of the matrix, and the exponential operation of the matrix represents the calculation of the natural exponential function value raised to the power of the eigenvalues at each position in the matrix.

In the above-mentioned ultrasonic guided wave pipeline defect nondestructive testing system, the detection result generation module is further used to: expand the corrected differential feature matrix into a classification feature vector according to a row vector or a column vector; use the fully connected layer of the classifier to fully connect the classification feature vector to obtain an encoded classification feature vector; and input the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

According to another aspect of the present application, an electronic device is provided, comprising: a processor; and a memory, wherein computer program instructions are stored in the memory, and when the computer program instructions are executed by the processor, the processor executes the ultrasonic guided wave pipeline defect nondestructive detection method as described above.

According to another aspect of the present application, a computer-readable medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the ultrasonic guided wave pipeline defect nondestructive detection method as described above.

Compared with the prior art, the ultrasonic guided wave pipeline defect nondestructive detection method provided by the present application adopts artificial intelligence detection technology based on deep learning to detect and judge whether the pipeline to be detected has defects by comparing the multi-scale feature differences between the detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in a high-dimensional space. In this way, nondestructive detection of pipeline defects can be accurately performed.

BRIEF DESCRIPTION OF THE DRAWINGS

By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

FIG1 is a diagram showing an application scenario of a method for nondestructive detection of pipeline defects using ultrasonic guided waves according to an embodiment of the present application.

FIG2 is a flow chart of a method for nondestructive detection of pipeline defects using ultrasonic guided waves according to an embodiment of the present application.

FIG3 is a schematic diagram of a method for nondestructive detection of pipeline defects using ultrasonic guided waves according to an embodiment of the present application.

FIG4 is a block diagram of an ultrasonic guided wave pipeline defect nondestructive detection system according to an embodiment of the present application.

FIG5 is a block diagram of an electronic device according to an embodiment of the present application.

DETAILED DESCRIPTION

Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

Application Overview

As mentioned in the above background technology, pipeline laying involves all walks of life and undertakes an important production mission. Most pipelines need to face various complex environments, such as bad weather, impact, corrosion of chemical substances inside the pipeline, etc., so pipelines are prone to various defects. In engineering, incidents where pipeline defects cause great damage to people's survival and property safety frequently occur. Therefore, the detection of pipeline health status is particularly important.

At present, pipeline maintenance personnel generally detect pipeline damage through visual inspection, electromagnetic flaw detection and other methods. These methods are not only time-consuming, but also have low detection accuracy. Therefore, an optimized non-destructive detection solution for pipeline defects is desired.

At present, deep learning and neural networks have been widely used in computer vision, natural language processing, speech signal processing and other fields. In addition, deep learning and neural networks have also shown a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields.

In recent years, the development of deep learning and neural networks has provided new solutions and solutions for non-destructive detection of pipeline defects.

Accordingly, since nondestructive testing is to detect defects on the tested piece without destroying the test piece. The advantages of nondestructive testing are non-destructiveness, low testing cost, wide testing range, and high testing accuracy. Ultrasonic waves are mechanical waves with a frequency higher than 20KHz, which are widely used in thickness measurement, nondestructive flaw detection and other fields. Since the elastic medium of the pipeline is actually limited by boundaries, the ultrasonic wave in its propagation process is a guided wave. Therefore, the ultrasonic guided wave signal actually detected in the pipeline and the reference ultrasonic guided wave signal of the defect-free pipeline can be used for defect detection. However, considering that the ultrasonic guided wave signal has a large amount of information, it is difficult to compare and observe the two in actual applications, and since the defects in the pipeline are small-scale characteristic information, this brings difficulties to the nondestructive detection of pipeline defects.

Based on this, in the technical solution of this application, it is expected to adopt artificial intelligence detection technology based on deep learning, so as to detect and judge whether the pipeline to be detected has defects by comparing the multi-scale feature differences of the detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in a high-dimensional space. In this way, non-destructive detection of pipeline defects can be accurately performed, and then the health status of the pipeline can be paid attention to at all times to avoid the loss of people's survival and property safety caused by pipeline defects.

Specifically, in the technical solution of the present application, first, a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of the pipeline to be detected are obtained, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline. Next, considering that the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal are expressed in the time domain as a waveform, an image encoder can be used to extract the time domain features of the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal, and the pipeline defect detection is performed using this feature difference distribution information. However, when the time domain feature difference distribution information of the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal is used to perform pipeline defect detection, the time domain feature will contain more environmental noise interference feature information, which will have a serious impact on the detection result. Therefore, the correlation feature distribution between the frequency domain statistical feature values of the ultrasonic guided wave signal is further combined to improve the detection accuracy. That is, the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal are further Fourier transformed to obtain multiple detection frequency domain statistical values and multiple reference frequency domain statistical values.

Then, for the feature extraction of the detection ultrasonic guided wave signal, the first Clip model including a sequence encoder and an image encoder is used to process the multiple detection frequency domain statistics and the waveform of the detection ultrasonic guided wave signal respectively to obtain the detection ultrasonic feature matrix. That is, specifically, first, considering that the defect characteristics in the pipeline are small-scale feature information, the sequence encoder of the first Clip model is used to perform feature mining on the multiple detection frequency domain statistics of the detection ultrasonic guided wave signal to extract the multi-scale implicit correlation feature distribution information between the multiple detection frequency domain statistics, thereby obtaining the detection frequency domain statistical feature vector. In particular, here, the sequence encoder uses a multi-scale neighborhood feature extraction module to extract the correlation feature of the multiple detection frequency domain statistics. Next, the image encoder of the first Clip model is used to perform feature mining on the waveform of the detection ultrasonic guided wave signal to extract the time domain implicit feature distribution information of the detection ultrasonic guided wave signal, thereby obtaining the detection ultrasonic guided wave image feature vector. Finally, the detection ultrasonic guided wave image feature vector and the detection frequency domain statistical feature vector are input into the encoding optimizer of the first Clip model to obtain the detection ultrasonic feature matrix. That is, based on the multi-scale implicit correlation feature information of the frequency domain statistical eigenvalues of the detection ultrasonic guided wave signal, the time domain implicit features of the detection ultrasonic guided wave signal waveform are optimized by image attribute coding to obtain the detection ultrasonic feature matrix. In this way, the obtained detection ultrasonic feature matrix not only contains the frequency domain feature content of the detection ultrasonic guided wave signal, but also reflects the characteristics of the change of the frequency domain content over time, thereby improving the accuracy of pipeline defect detection.

Similarly, for the sound feature extraction of the reference ultrasonic guided wave signal, considering that the periodic feature information of the reference ultrasonic guided wave signal has similar regularity to the periodic feature of the detection ultrasonic guided wave signal, therefore, in the technical solution of the present application, the Clip model is also used to encode the reference ultrasonic guided wave signal. That is, specifically, the multiple reference frequency domain statistical values and the reference ultrasonic guided wave signal are passed through a second Clip model including a column encoder and an image encoder to obtain a reference ultrasonic feature matrix, and then based on the multi-scale implicit correlation features of the frequency domain statistical feature values of the reference ultrasonic guided wave signal, the time domain implicit features of the reference ultrasonic guided wave signal waveform are optimized by image attribute encoding to obtain the reference ultrasonic feature matrix.

Furthermore, the differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated to represent the difference features of the actual detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in the high-dimensional space, and this is used as a classification feature matrix for classification processing in a classifier to obtain a classification result indicating whether the pipeline to be detected has defects. In this way, non-destructive detection of pipeline defects can be accurately performed, and the health status of the pipeline can be kept under constant attention to avoid losses to people's survival and property safety caused by pipeline defects.

In particular, in the technical solution of the present application, when calculating the feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix to obtain the differential feature matrix, since the calculation of the differential feature matrix is the positional eigenvalue difference calculation between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix, it is expected that the feature distribution between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix maintains the same phase distribution as much as possible, that is, it is expected to avoid the negative correlation between the corresponding positions of the detection ultrasonic feature matrix and the reference ultrasonic feature matrix as much as possible, thereby improving the calculation accuracy of the differential feature matrix.

Therefore, the applicant of the present application uses a full orthographic projection nonlinear reweighting method to calculate the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix, which is expressed as:

_M1 and _M2 are the detection ultrasonic feature matrix and the reference ultrasonic feature matrix respectively, _Mc is the weight matrix, and the division between the numerator matrix and the denominator matrix is the positional division of the matrix eigenvalues.

Here, the full orthographic projection nonlinear reweighting uses the ReLU function to ensure that the projection is fully positive to avoid aggregating negatively correlated information, and at the same time introduces a nonlinear reweighting mechanism to aggregate the eigenvalue distributions of the detection ultrasonic feature matrix and the reference ultrasonic feature matrix relative to each other, so that the intrinsic structure of the weight matrix can punish long-distance connections and strengthen local coupling. In this way, by performing point multiplication on the differential feature matrix with the weight matrix to perform positional weighting, the synergistic effect of the spatial feature transformation (feature transform) corresponding to the full orthographic projection reweighting of the detection ultrasonic feature matrix and the reference ultrasonic feature matrix in the high-dimensional feature space is achieved, which improves the accuracy of the calculation of the differential feature matrix, thereby improving the accuracy of classification. In this way, non-destructive detection of pipeline defects can be accurately performed, and the health status of the pipeline can be always paid attention to to avoid pipeline defects causing losses to people's survival and property safety.

Based on this, the present application proposes a nondestructive detection method for ultrasonic guided wave pipeline defects, which includes: obtaining a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of a pipeline to be detected, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline; performing Fourier transform on the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal respectively to obtain a plurality of detection frequency domain statistical values and a plurality of reference frequency domain statistical values; passing the waveform diagrams of the plurality of detection frequency domain statistical values and the detection ultrasonic guided wave signal through a first Clip model to obtain a detection ultrasonic feature matrix; passing the plurality of reference frequency domain statistical values and the reference ultrasonic guided wave signal through a second Clip model to obtain a reference ultrasonic feature matrix; calculating a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix; correcting the eigenvalues of each position of the differential feature matrix to obtain a corrected differential feature matrix; and passing the corrected differential feature matrix through a classifier to obtain a classification result, wherein the classification result indicates whether there is a defect in the pipeline to be detected.

FIG1 is an application scenario diagram of an ultrasonic guided wave pipeline defect nondestructive detection method according to an embodiment of the present application. As shown in FIG1, in this application scenario, an ultrasonic guided wave instrument (for example, Se as shown in FIG1) is first used to obtain a detection ultrasonic guided wave signal of a pipeline to be detected (for example, P as shown in FIG1). Then, the detection ultrasonic guided wave signal of the pipeline to be detected is input into a server (for example, S as shown in FIG1) deployed with an ultrasonic guided wave pipeline defect nondestructive detection algorithm, wherein the server can process the detection ultrasonic guided wave signal of the pipeline to be detected based on the ultrasonic guided wave pipeline defect nondestructive detection algorithm to obtain a classification result for indicating whether the pipeline to be detected has defects.

After introducing the basic principles of the present application, various non-limiting embodiments of the present application will be described in detail with reference to the accompanying drawings.

Exemplary Methods

FIG2 is a flow chart of the ultrasonic guided wave pipeline defect nondestructive detection method according to an embodiment of the present application. As shown in FIG2, the ultrasonic guided wave pipeline defect nondestructive detection method according to an embodiment of the present application includes: S110, obtaining a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of the pipeline to be detected, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline; S120, performing Fourier transform on the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal respectively to obtain a plurality of detection frequency domain statistics and a plurality of reference frequency domain statistics; S130, transforming the waveform diagrams of the plurality of detection frequency domain statistics and the detection ultrasonic guided wave signal through a first Clip model To obtain a detection ultrasonic feature matrix; S140, the multiple reference frequency domain statistical values and the reference ultrasonic guided wave signal are passed through a second Clip model to obtain a reference ultrasonic feature matrix; S150, a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated; S160, the eigenvalues of each position of the differential feature matrix are corrected to obtain a corrected differential feature matrix; and, S170, the corrected differential feature matrix is passed through a classifier to obtain a classification result, wherein the classification result indicates whether there is a defect in the pipeline to be detected.

FIG3 is an architecture diagram of a method for nondestructive detection of defects of ultrasonic guided wave pipelines according to an embodiment of the present application. As shown in FIG3, in the architecture, first, a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of the pipeline to be detected are obtained, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline. Then, the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal are respectively subjected to Fourier transform to obtain a plurality of detection frequency domain statistics and a plurality of reference frequency domain statistics. Then, the waveform diagrams of the plurality of detection frequency domain statistics and the detection ultrasonic guided wave signal are passed through a first Clip model to obtain a detection ultrasonic feature matrix, and at the same time, the plurality of reference frequency domain statistics and the reference ultrasonic guided wave signal are passed through a second Clip model to obtain a reference ultrasonic feature matrix. Then, a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated. Then, the eigenvalues of each position of the differential feature matrix are corrected to obtain a corrected differential feature matrix. Furthermore, the corrected differential feature matrix is passed through a classifier to obtain a classification result, and the classification result indicates whether there is a defect in the pipeline to be detected.

In step S110, the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal of the pipeline to be detected are obtained, wherein the reference ultrasonic guided wave signal is the ultrasonic guided wave signal of a defect-free pipeline. As mentioned in the above background technology, pipeline laying involves all walks of life and undertakes an important production mission. Most pipelines need to face various complex environments, such as bad weather, impact, corrosion of chemical substances inside the pipeline, etc., so the pipeline is prone to various defects. In engineering, because pipeline defects cause great damage to people's survival and property safety, incidents occur frequently, therefore, the detection of pipeline health status is particularly important. At present, pipeline maintenance personnel generally detect pipeline damage by visual inspection, electromagnetic flaw detection and other methods. These methods are not only time-consuming, but also have low detection accuracy. Therefore, an optimized non-destructive detection scheme for pipeline defects is desired.

Accordingly, since nondestructive testing is to detect defects on the tested piece without destroying the test piece. The advantages of nondestructive testing are non-destructiveness, low testing cost, wide testing range, and high testing accuracy. Ultrasonic waves are mechanical waves with a frequency higher than 20KHz, which are widely used in thickness measurement, nondestructive flaw detection and other fields. Since the elastic medium of the pipeline is actually limited by boundaries, the ultrasonic wave in its propagation process is a guided wave. Therefore, the ultrasonic guided wave signal actually detected in the pipeline and the reference ultrasonic guided wave signal of the defect-free pipeline can be used for defect detection. However, considering that the ultrasonic guided wave signal has a large amount of information, it is difficult to compare and observe the two in actual applications, and since the defects in the pipeline are small-scale characteristic information, this brings difficulties to the nondestructive detection of pipeline defects.

Based on this, in the technical solution of the present application, it is expected to adopt artificial intelligence detection technology based on deep learning, so as to detect and judge whether the pipeline to be detected has defects by comparing the detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in a high-dimensional space. In this way, non-destructive detection of pipeline defects can be accurately performed, and then attention can be paid to the health status of the pipeline at all times to avoid losses to people's survival and property safety caused by pipeline defects. Specifically, in the technical solution of the present application, first, the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal of the pipeline to be detected are obtained, wherein the reference ultrasonic guided wave signal is the ultrasonic guided wave signal of the defect-free pipeline. Here, the detection ultrasonic guided wave signal of the pipeline to be detected can be obtained by an ultrasonic guided wave instrument, and the reference ultrasonic guided wave signal is existing data.

In step S120, the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal are respectively subjected to Fourier transformation to obtain a plurality of detection frequency domain statistics and a plurality of reference frequency domain statistics. Considering that the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal are expressed in the time domain in the form of a waveform, an image encoder can be used to extract the time domain features of the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal, and the pipeline defect detection is performed based on the feature difference distribution information. However, when the pipeline defect detection is performed using the time domain feature difference distribution information of the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal, the time domain features contain a lot of environmental noise interference feature information, which will have a serious impact on the detection result. Therefore, the correlation feature distribution between the frequency domain statistical feature values of the ultrasonic guided wave signal is further combined to improve the detection accuracy. That is, the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal are further subjected to Fourier transformation to obtain a plurality of detection frequency domain statistics and a plurality of reference frequency domain statistics.

In step S130, the plurality of detection frequency domain statistics and the waveform of the detection ultrasonic guided wave signal are passed through a first Clip model to obtain a detection ultrasonic feature matrix. For feature extraction of the detection ultrasonic guided wave signal, the first Clip model including a sequence encoder and an image encoder is used to process the plurality of detection frequency domain statistics and the waveform of the detection ultrasonic guided wave signal to obtain a detection ultrasonic feature matrix.

That is, specifically, first, considering that the defect characteristics in the pipeline are small-scale feature information, the sequence encoder of the first Clip model performs feature mining on multiple detection frequency domain statistical values of the detection ultrasonic guided wave signal to extract multi-scale implicit correlation feature distribution information between the multiple detection frequency domain statistical values, thereby obtaining the detection frequency domain statistical feature vector. In particular, here, the sequence encoder uses a multi-scale neighborhood feature extraction module to extract the correlation features of the multiple detection frequency domain statistical values. In one embodiment of the present application, the step of inputting the multiple detection frequency domain statistical values into the sequence encoder of the first Clip model to obtain the detection frequency domain statistical feature vector includes: inputting the multiple detection frequency domain statistical values into the first convolution layer of the multi-scale neighborhood feature extraction module to obtain the first-scale detection frequency domain statistical feature vector, wherein the first convolution layer has a first one-dimensional convolution kernel of a first length; inputting the multiple detection frequency domain statistical values into the second convolution layer of the multi-scale neighborhood feature extraction module to obtain the second-scale detection frequency domain statistical feature vector, wherein the second convolution layer has a second one-dimensional convolution kernel of a second length, and the first length is different from the second length; and, cascading the first-scale detection frequency domain statistical feature vector and the second-scale detection frequency domain statistical feature vector to obtain the detection frequency domain statistical feature vector. Here, the multi-scale neighborhood feature extraction module can extract the multi-scale neighborhood correlation features of the multiple detection frequency domain statistical values under different time spans to characterize the multi-scale neighborhood dynamic change feature information of the detection frequency domain statistical values in the time dimension, and at the same time, the output features include both the smoothed features and the original input features to avoid information loss, thereby improving the accuracy of subsequent classification.

Next, the image encoder of the first Clip model is used to perform feature mining on the waveform of the detection ultrasonic guided wave signal to extract the time domain implicit feature distribution information of the detection ultrasonic guided wave signal, thereby obtaining the detection ultrasonic guided wave image feature vector. Specifically, in an embodiment of the present application, the waveform of the detection ultrasonic guided wave signal is input into the image encoder of the first Clip model to obtain the detection ultrasonic guided wave image feature vector, including: using each layer of the image encoder of the first Clip model to perform in the forward transfer of the layer: convolution processing on the input data to obtain a convolution feature map; mean pooling based on the local feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the image encoder of the first Clip model is the detection ultrasonic guided wave image feature vector, and the input of the first layer of the image encoder of the first Clip model is the waveform of the detection ultrasonic guided wave signal.

Finally, the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector are input into the coding optimizer of the first Clip model to obtain the detected ultrasonic feature matrix. That is, based on the multi-scale implicit correlation feature information of the frequency domain statistical eigenvalues of the detected ultrasonic guided wave signal, the time domain implicit features of the detected ultrasonic guided wave signal waveform are optimized for image attribute coding to obtain the detected ultrasonic feature matrix. Specifically, in the technical solution of the present application, the input of the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detected ultrasonic feature matrix includes: inputting the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into the coding optimizer of the first Clip model according to the following formula to obtain the detected ultrasonic feature matrix; wherein, the formula is:

表示所述检测超声导波图像特征向量的转置向量，V₂表示所述检测频域统计特征向量，M表示所述检测超声特征矩阵，

表示矩阵相乘。这样，所得到的所述检测超声特征矩阵既包含了所述检测超声导波信号的频域特征内容，又反应了频域内容随时间的变化规律特征，提高了管道缺陷检测的精准度。in

represents the transposed vector of the detected ultrasonic guided wave image feature vector, V ₂ represents the detected frequency domain statistical feature vector, M represents the detected ultrasonic feature matrix,

In this way, the obtained detection ultrasonic feature matrix not only contains the frequency domain feature content of the detection ultrasonic guided wave signal, but also reflects the characteristics of the change rule of the frequency domain content over time, thereby improving the accuracy of pipeline defect detection.

In step S140, the multiple reference frequency domain statistics and the reference ultrasonic guided wave signal are passed through a second Clip model to obtain a reference ultrasonic feature matrix. Similarly, for the sound feature extraction of the reference ultrasonic guided wave signal, considering that the periodic feature information of the reference ultrasonic guided wave signal has similar regularity to the periodic feature of the detection ultrasonic guided wave signal, therefore, in the technical solution of the present application, the Clip model is also used to encode the reference ultrasonic guided wave signal. That is, specifically, the multiple reference frequency domain statistics and the reference ultrasonic guided wave signal are passed through a second Clip model including a column encoder and an image encoder to obtain a reference ultrasonic feature matrix, and then based on the multi-scale implicit correlation features of the frequency domain statistical feature values of the reference ultrasonic guided wave signal, the time domain implicit features of the reference ultrasonic guided wave signal waveform are optimized by image attribute encoding to obtain the reference ultrasonic feature matrix.

In step S150, a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated. That is, a differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated to represent the difference features between the actual detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in a high-dimensional space.

其中M₁表示所述检测超声特征矩阵，M₂表示所述参考超声特征矩阵，M_c表示所述差分特征矩阵。Specifically, in the embodiment of the present application, the calculation of the differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix includes: calculating the differential feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix using the following formula; wherein the formula is:

Wherein, _M1 represents the detected ultrasonic feature matrix, _M2 represents the reference ultrasonic feature matrix, and _Mc represents the differential feature matrix.

In step S160, the eigenvalues of each position of the differential feature matrix are corrected to obtain a corrected differential feature matrix. In particular, in the technical solution of the present application, when calculating the feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix to obtain the differential feature matrix, since the calculation of the differential feature matrix is the positional eigenvalue difference calculation between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix, it is expected that the feature distribution between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix should be kept in phase as much as possible, that is, it is expected to avoid the negative correlation between the corresponding positions of the detection ultrasonic feature matrix and the reference ultrasonic feature matrix as much as possible, thereby improving the calculation accuracy of the differential feature matrix. Therefore, the applicant of the present application adopts the method of full orthogonal projection nonlinear reweighting to calculate the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix.

Specifically, in an embodiment of the present application, the eigenvalues of each position of the differential feature matrix are corrected to obtain a corrected differential feature matrix, including: calculating a weight matrix between the detected ultrasonic feature matrix and the reference ultrasonic feature matrix by full orthographic projection nonlinear reweighting; and multiplying the differential feature matrix by the weight matrix at each position to obtain the corrected differential feature matrix.

More specifically, in an embodiment of the present application, the weight matrix between the detected ultrasonic feature matrix and the reference ultrasonic feature matrix is calculated by using a full orthogonal projection nonlinear reweighting method, including: using a full orthogonal projection nonlinear reweighting method to calculate the weight matrix between the detected ultrasonic feature matrix and the reference ultrasonic feature matrix using the following formula; wherein the formula is:

表示矩阵相乘，且分子矩阵和分母矩阵之间的除法为矩阵特征值的按位置相除，exp(·)表示矩阵的指数运算，所述矩阵的指数运算表示计算以矩阵中各个位置的特征值为幂的自然指数函数值。Wherein, _M1 and _M2 are the detection ultrasonic feature matrix and the reference ultrasonic feature matrix respectively, _Mw is the weight matrix, ReLU(·) represents the ReLU activation function,

represents matrix multiplication, and the division between the numerator matrix and the denominator matrix is the positional division of the matrix eigenvalues. exp(·) represents the exponential operation of the matrix, and the exponential operation of the matrix represents the calculation of the natural exponential function value raised to the power of the eigenvalues at each position in the matrix.

Here, the full orthographic projection nonlinear reweighting uses the ReLU function to ensure that the projection is fully positive to avoid aggregating negatively correlated information, and at the same time introduces a nonlinear reweighting mechanism to aggregate the eigenvalue distributions of the detection ultrasonic feature matrix and the reference ultrasonic feature matrix relative to each other, so that the intrinsic structure of the weight matrix can punish long-distance connections and strengthen local coupling. In this way, by performing point multiplication on the differential feature matrix with the weight matrix to perform positional weighting, the synergistic effect of the spatial feature transformation (feature transform) corresponding to the full orthographic projection reweighting of the detection ultrasonic feature matrix and the reference ultrasonic feature matrix in the high-dimensional feature space is achieved, which improves the accuracy of the calculation of the differential feature matrix, thereby improving the accuracy of classification. In this way, non-destructive detection of pipeline defects can be accurately performed, and the health status of the pipeline can be always paid attention to to avoid pipeline defects causing losses to people's survival and property safety.

In step S170, the corrected differential feature matrix is passed through a classifier to obtain a classification result, which indicates whether the pipeline to be inspected has defects. In this way, non-destructive detection of pipeline defects can be accurately performed, and the health status of the pipeline can be always paid attention to to avoid losses to people's survival and property safety caused by pipeline defects.

Specifically, in an embodiment of the present application, the corrected differential feature matrix is passed through a classifier to obtain a classification result, and the classification result indicates whether there is a defect in the pipeline to be inspected, including: expanding the corrected differential feature matrix into a classification feature vector according to a row vector or a column vector; using the fully connected layer of the classifier to fully connect encode the classification feature vector to obtain an encoded classification feature vector; and inputting the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

In summary, the ultrasonic guided wave pipeline defect nondestructive detection method based on the embodiment of the present application is explained, which adopts artificial intelligence detection technology based on deep learning to detect and judge whether the pipeline to be detected has defects by comparing the multi-scale feature differences between the detection ultrasonic guided wave signal of the pipeline to be detected and the reference ultrasonic guided wave signal of the defect-free pipeline in a high-dimensional space. In this way, nondestructive detection of pipeline defects can be accurately performed.

Exemplary Systems

FIG4 is a block diagram of an ultrasonic guided wave pipeline defect nondestructive detection system according to an embodiment of the present application. As shown in FIG4, an ultrasonic guided wave pipeline defect nondestructive detection system 100 according to an embodiment of the present application includes: an ultrasonic guided wave acquisition module 110, which is used to acquire a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of a pipeline to be detected, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline; a domain change module 120, which is used to perform Fourier transform on the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal respectively to obtain a plurality of detection frequency domain statistics and a plurality of reference frequency domain statistics; a detection ultrasonic encoding module 130, which is used to transform the waveform diagram of the plurality of detection frequency domain statistics and the detection ultrasonic guided wave signal through a first Clip model to obtain to detect ultrasonic feature matrix; a reference ultrasonic encoding module 140, used to pass the multiple reference frequency domain statistical values and the reference ultrasonic guided wave signal through a second Clip model to obtain a reference ultrasonic feature matrix; a difference module 150, used to calculate a differential feature matrix between the detected ultrasonic feature matrix and the reference ultrasonic feature matrix; an eigenvalue correction module 160, used to correct the eigenvalues of each position of the differential feature matrix to obtain a corrected differential feature matrix; and a detection result generation module 170, used to pass the corrected differential feature matrix through a classifier to obtain a classification result, wherein the classification result indicates whether there is a defect in the pipeline to be detected.

In one example, in the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system 100, the detection ultrasonic encoding module 130 includes: a sequence encoding unit, which is used to input the multiple detection frequency domain statistical values into the sequence encoder of the first Clip model to obtain a detection frequency domain statistical feature vector; an image encoding unit, which is used to input the waveform diagram of the detection ultrasonic guided wave signal into the image encoder of the first Clip model to obtain a detection ultrasonic guided wave image feature vector; and a coding optimization unit, which is used to input the detection ultrasonic guided wave image feature vector and the detection frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detection ultrasonic feature matrix.

In one example, in the above-mentioned ultrasonic guided wave pipeline defect nondestructive testing system 100, the sequence encoding unit is further used to: input the multiple detection frequency domain statistical values into the first convolution layer of the multi-scale neighborhood feature extraction module to obtain a first-scale detection frequency domain statistical feature vector, wherein the first convolution layer has a first one-dimensional convolution kernel of a first length; input the multiple detection frequency domain statistical values into the second convolution layer of the multi-scale neighborhood feature extraction module to obtain a second-scale detection frequency domain statistical feature vector, wherein the second convolution layer has a second one-dimensional convolution kernel of a second length, and the first length is different from the second length; and, cascade the first-scale detection frequency domain statistical feature vector and the second-scale detection frequency domain statistical feature vector to obtain the detection frequency domain statistical feature vector.

In one example, in the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system 100, the image encoding unit is further used to: use each layer of the image encoder of the first Clip model to perform in the forward transfer of the layer: convolution processing on the input data to obtain a convolution feature map; mean pooling based on the local feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the image encoder of the first Clip model is the detected ultrasonic guided wave image feature vector, and the input of the first layer of the image encoder of the first Clip model is the waveform diagram of the detected ultrasonic guided wave signal.

In one example, in the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system 100, the coding optimization unit is further used to: input the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into the coding optimizer of the first Clip model to obtain the detected ultrasonic feature matrix according to the following formula; wherein the formula is:

表示所述检测超声导波图像特征向量的转置向量，V₂表示所述检测频域统计特征向量，M表示所述检测超声特征矩阵，

表示矩阵相乘。in

represents the transposed vector of the detected ultrasonic guided wave image feature vector, V ₂ represents the detected frequency domain statistical feature vector, M represents the detected ultrasonic feature matrix,

Represents matrix multiplication.

其中M₁表示所述检测超声特征矩阵，M₂表示所述参考超声特征矩阵，M_c表示所述差分特征矩阵。In one example, in the ultrasonic guided wave pipeline defect nondestructive testing system 100, the difference module 150 is further used to calculate the difference feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix using the following formula; wherein the formula is:

Wherein, _M1 represents the detected ultrasonic feature matrix, _M2 represents the reference ultrasonic feature matrix, and _Mc represents the differential feature matrix.

In one example, in the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system 100, the eigenvalue correction module 160 includes: a weight unit, which calculates the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by full orthographic projection nonlinear reweighting; and an application unit, which multiplies the differential feature matrix by the weight matrix at the position point to obtain the corrected differential feature matrix.

In one example, in the ultrasonic guided wave pipeline defect nondestructive testing system 100, the weight unit is further used to: calculate the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by the following formula using full orthographic projection nonlinear reweighting; wherein the formula is:

表示矩阵相乘，且分子矩阵和分母矩阵之间的除法为矩阵特征值的按位置相除，exp(·)表示矩阵的指数运算，所述矩阵的指数运算表示计算以矩阵中各个位置的特征值为幂的自然指数函数值。Wherein, _M1 and _M2 are the detection ultrasonic feature matrix and the reference ultrasonic feature matrix respectively, _Mw is the weight matrix, ReLU(·) represents the ReLU activation function,

represents matrix multiplication, and the division between the numerator matrix and the denominator matrix is the positional division of the matrix eigenvalues. exp(·) represents the exponential operation of the matrix, and the exponential operation of the matrix represents the calculation of the natural exponential function value raised to the power of the eigenvalues at each position in the matrix.

In one example, in the above-mentioned ultrasonic guided wave pipeline defect nondestructive testing system 100, the detection result generation module 170 is further used to: expand the corrected differential feature matrix into a classification feature vector according to a row vector or a column vector; use the fully connected layer of the classifier to fully connect the classification feature vector to obtain an encoded classification feature vector; and input the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

Here, those skilled in the art can understand that the specific functions and operations of the various units and modules in the above-mentioned ultrasonic guided wave pipeline defect nondestructive detection system 100 have been described in detail in the description of the ultrasonic guided wave pipeline defect nondestructive detection method with reference to Figures 1 to 3 above, and therefore, its repeated description will be omitted.

As described above, the ultrasonic guided wave pipeline defect nondestructive detection system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for ultrasonic guided wave pipeline defect nondestructive detection. In one example, the ultrasonic guided wave pipeline defect nondestructive detection system 100 according to the embodiment of the present application can be integrated into the terminal device as a software module and/or a hardware module. For example, the ultrasonic guided wave pipeline defect nondestructive detection system 100 can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the ultrasonic guided wave pipeline defect nondestructive detection system 100 can also be one of the many hardware modules of the terminal device.

Alternatively, in another example, the ultrasonic guided wave nondestructive detection system for pipeline defects 100 and the terminal device may also be separate devices, and the ultrasonic guided wave nondestructive detection system for pipeline defects 100 may be connected to the terminal device via a wired and/or wireless network and transmit interactive information in accordance with an agreed data format.

Exemplary Electronic Devices

Next, an electronic device according to an embodiment of the present application is described with reference to FIG5 . FIG5 is a block diagram of an electronic device according to an embodiment of the present application. As shown in FIG5 , the electronic device 10 includes one or more processors 11 and a memory 12 .

The processor 11 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

The memory 12 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and/or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may run the program instructions to implement the functions of the ultrasonic guided wave pipeline defect non-destructive detection method of each embodiment of the present application described above and/or other desired functions. Various contents such as detection ultrasonic guided wave signals of the pipeline to be detected and reference ultrasonic guided wave signals may also be stored in the computer-readable storage medium.

In one example, the electronic device 10 may further include: an input device 13 and an output device 14, and these components are interconnected via a bus system and/or other forms of connection mechanisms (not shown).

The input device 13 may include, for example, a keyboard, a mouse, etc.

The output device 14 can output various information to the outside, including classification results, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

Of course, for simplicity, FIG5 only shows some of the components in the electronic device 10 related to the present application, omitting components such as a bus, an input/output interface, etc. In addition, the electronic device 10 may further include any other appropriate components according to specific application scenarios.

Exemplary computer program products and computer-readable storage media

In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to perform the steps in the functions of the ultrasonic guided wave pipeline defect nondestructive detection method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

The computer program product may be written in any combination of one or more programming languages to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps in the functions of the ultrasonic guided wave pipeline defect nondestructive detection method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

The block diagrams of the devices, devices, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and/or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and/or recombined. Such decomposition and/or recombination should be regarded as equivalent solutions of the present application.

The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims (9)

1. A nondestructive detection method for defects of ultrasonic guided wave pipelines is characterized by comprising the following steps: acquiring a detection ultrasonic guided wave signal and a reference ultrasonic guided wave signal of a to-be-detected pipeline, wherein the reference ultrasonic guided wave signal is an ultrasonic guided wave signal of a defect-free pipeline; respectively carrying out Fourier transform on the detection ultrasonic guided wave signal and the reference ultrasonic guided wave signal to obtain a plurality of detection frequency domain statistics values and a plurality of reference frequency domain statistics values; enabling the plurality of detection frequency domain statistics values and the oscillogram of the detection ultrasonic guided wave signal to pass through a first Clip model to obtain a detection ultrasonic characteristic matrix; enabling the plurality of reference frequency domain statistics and the reference ultrasonic guided wave signals to pass through a second Clip model to obtain a reference ultrasonic characteristic matrix; calculating a difference feature matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix; correcting the characteristic value of each position of the differential characteristic matrix to obtain a corrected differential characteristic matrix; and passing the corrected differential feature matrix through a classifier to obtain a classification result, wherein the classification result indicates whether the pipeline to be detected has defects or not.

2. The method for nondestructive testing of the defect of the ultrasonic guided wave pipeline according to claim 1, wherein the step of passing the plurality of detection frequency domain statistics and the oscillogram of the detected ultrasonic guided wave signal through a first Clip model to obtain a detection ultrasonic feature matrix comprises the steps of: inputting the plurality of detection frequency domain statistical values into a sequence encoder of the first Clip model to obtain detection frequency domain statistical feature vectors; inputting the oscillogram of the detected ultrasonic guided wave signal into an image encoder of the first Clip model to obtain a characteristic vector of the detected ultrasonic guided wave image; and inputting the detection ultrasonic guided wave image feature vectors and the detection frequency domain statistical feature vectors into a coding optimizer of the first Clip model to obtain the detection ultrasonic feature matrix.

3. The method according to claim 2, wherein the inputting the plurality of detection frequency domain statistics into the sequence encoder of the first Clip model to obtain the detection frequency domain statistical feature vector comprises: inputting the plurality of detection frequency domain statistics into a first convolution layer of the multi-scale neighborhood feature extraction module to obtain a first scale detection frequency domain statistics feature vector, wherein the first convolution layer has a first one-dimensional convolution kernel with a first length; inputting the plurality of detection frequency domain statistics into a second convolution layer of the multi-scale neighborhood feature extraction module to obtain a second scale detection frequency domain statistics feature vector, wherein the second convolution layer has a second one-dimensional convolution kernel with a second length, and the first length is different from the second length; and cascading the first scale detection frequency domain statistical characteristic vector and the second scale detection frequency domain statistical characteristic vector to obtain the detection frequency domain statistical characteristic vector.

4. The method for nondestructive detection of the defect of the ultrasonic guided wave pipeline according to claim 3, wherein the step of inputting the oscillogram of the ultrasonic guided wave signal to be detected into the image encoder of the first Clip model to obtain the characteristic vector of the ultrasonic guided wave image comprises the following steps: the layers of the image encoder using the first Clip model are respectively performed in the forward pass of the layers: performing convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling based on a local feature matrix on the convolution feature map to obtain a pooled feature map; and carrying out nonlinear activation on the pooling feature map to obtain an activation feature map; the output of the last layer of the image encoder of the first Clip model is the detected ultrasonic guided wave image characteristic vector, and the input of the first layer of the image encoder of the first Clip model is the oscillogram of the detected ultrasonic guided wave signal.

5. The method for nondestructive testing of the defect of the ultrasonic guided wave pipeline according to claim 4, wherein the inputting the feature vector of the ultrasonic guided wave image to be tested and the statistical feature vector of the frequency domain to be tested into a coding optimizer of the first Clip model to obtain the ultrasonic feature matrix to be tested comprises: inputting the detected ultrasonic guided wave image feature vector and the detected frequency domain statistical feature vector into a coding optimizer of the first Clip model according to the following formula to obtain a detected ultrasonic feature matrix; wherein the formula is:

wherein->

A transposed vector, V, representing the characteristic vector of the detected ultrasound guided wave image ₂ Representing the detected frequency domain statistical feature vector, M denotes the detection ultrasound feature matrix, < > or >>

Representing a matrix multiplication. />

6. The method for nondestructive testing of defects of ultrasonic guided wave pipelines according to claim 5, wherein the calculating of the differential feature matrix between the test ultrasonic feature matrix and the reference ultrasonic feature matrix comprises: calculating a difference feature matrix between the detection ultrasound feature matrix and the reference ultrasound feature matrix in the following formula; wherein the formula is:

wherein M is ₁ Representing the detected ultrasound signature matrix, M ₂ Representing the reference ultrasound feature matrix, M _c Representing the difference feature matrix.

7. The method for nondestructive testing of defects of ultrasonic guided wave pipelines according to claim 6, wherein the step of correcting the eigenvalue of each position of the differential characteristic matrix to obtain a corrected differential characteristic matrix comprises the following steps: calculating a weight matrix between the detection ultrasonic characteristic matrix and the reference ultrasonic characteristic matrix in a full orthographic projection nonlinear weight weighting mode; and multiplying the difference characteristic matrix by the weight matrix according to position points to obtain the corrected difference characteristic matrix.

8. The method for nondestructive testing of the defect of the ultrasonic guided wave pipeline according to claim 7, wherein the calculating the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by adopting a full orthographic nonlinear re-weighting method comprises: calculating the weight matrix between the detection ultrasonic feature matrix and the reference ultrasonic feature matrix by adopting a full orthographic nonlinear weight weighting mode according to the following formula; wherein the formula is:

wherein M is ₁ And M ₂ Respectively being the detection ultrasonic feature matrix and the reference ultrasonic feature matrix, M _w Is the weight matrix, reLU (·) denotes a ReLU activation function,

representing the multiplication of matrices and the division between the numerator matrix and denominator matrix as a division by location of the eigenvalues of the matrices, exp (-) represents the exponential operation of the matrices, which represents the calculation of a natural exponential function value raised to the power of the eigenvalues of each location in the matrix.

9. The method for nondestructive testing of the defect of the ultrasonic guided wave pipeline according to claim 8, wherein the step of passing the corrected differential feature matrix through a classifier to obtain a classification result, wherein the classification result indicates whether the pipeline to be tested has the defect comprises the steps of: expanding the corrected differential feature matrix into classified feature vectors according to row vectors or column vectors; performing full-concatenation coding on the classification feature vectors by using a full-concatenation layer of the classifier to obtain coded classification feature vectors; and inputting the encoding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

Images