A method, apparatus, and device for ultrasonic inspection of interface bond quality of a multi-layer coating

Abstract

The application relates to an ultrasonic detection method, device and equipment for the interface bonding quality of a multilayer coating, which comprises the following steps: acquiring parameter data and ultrasonic data of a test sample, extracting relevant features, forming a feature set, using a principal component analysis method to reduce the dimension of the feature set to obtain the first m data reaching a contribution rate threshold, inputting training data formed by label data with an actual bonding state and the reduced dimension data into a support vector machine for training, obtaining a classification detection model of the coating interface bonding quality, inputting data of a sample to be detected into the classification detection model to realize quantitative detection of the coating bonding quality, considering multiple features influencing the coating interface bonding quality, and finally quantitatively detecting the failure of a topcoat and the failure of a primer, and solving the technical problems that existing ultrasonic nondestructive detection coating interface quality is represented by a single feature of the coating interface, and the obtained detection is not comprehensive and the result is inaccurate.

Description

Technical Field

[0001] This application relates to the field of quality inspection data processing technology, and in particular to an ultrasonic testing method, apparatus and equipment for the bonding quality of multilayer coating interfaces. Background Technology

[0002] Surface coating technology is widely used in various fields. The interfacial bonding quality of the coating, as a crucial evaluation parameter, affects its lifespan and performance. Therefore, the detection of coating interfacial bonding quality is of great significance. Currently, destructive testing methods are used to detect coating interfacial bonding quality, such as tensile testing, scratch testing, and cross-cut testing. However, destructive testing methods damage the sample, making it impossible to perform 100% inspection.

[0003] With the development of technology, ultrasonic non-destructive testing methods are being used to replace destructive testing methods. For example, the frequency domain imaging method of existing ultrasonic scanning microscopy extracts the time-domain separated coating / substrate interface echoes, converts them into amplitude spectra through fast Fourier transform, and identifies the amplitude spectrum to replace the time-domain amplitude imaging of the interface echoes, thereby evaluating the bonding strength. Another example is the existing ultrasonic volume wave phase spectrum method for non-destructively measuring the coating interface bonding strength. This method, when the coating thickness is unknown, uses a single ultrasonic volume wave immersion probe perpendicularly incident on the coating sample. Through signal analysis, the corresponding coating aliasing signal phase spectrum is obtained, and the phase change, which has a linear relationship with the interface stiffness coefficient, can be extracted to quantitatively measure the coating / substrate interface bonding strength. However, coatings in actual production often exhibit multi-layered structures and non-uniform distributions, making it incomplete to characterize the coating interface bonding quality using a single feature. Summary of the Invention

[0004] This application provides an ultrasonic testing method, apparatus, and device for the bonding quality of multilayer coating interfaces, which solves the technical problem that existing ultrasonic nondestructive testing of coating interface quality uses a single feature to characterize the coating interface, resulting in incomplete testing and inaccurate results.

[0005] To achieve the above objectives, the embodiments of this application provide the following technical solutions:

[0006] On the one hand, an ultrasonic testing method for the interfacial bonding quality of multilayer coatings is provided, comprising the following steps:

[0007] S1. Obtain parameter data and ultrasonic data of the test sample. The parameter data includes the topcoat thickness and total thickness of the multilayer coating. The ultrasonic data includes the ultrasonic signal of the failed topcoat layer, the ultrasonic signal of the failed primer layer, and the ultrasonic signal of the unfailed coating.

[0008] S2. processing the waveforms of all the ultrasonic data to obtain feature data of the multi-layer coating interface bonding quality characterization; taking the parameter data and the feature data as a feature set;

[0009] S3. marking the coating type corresponding to the feature data to obtain label data and performing dimension reduction processing on the feature set by principal component analysis to obtain dimension reduction data, taking the dimension reduction data and the label data as training data;

[0010] S4. training the training data by support vector machine to obtain a classification detection model of coating bonding quality;

[0011] S5. obtaining the to-be-detected scanning waveforms of the to-be-detected multi-layer coating sample, processing all the to-be-detected scanning waveforms by steps S2 and S3 to obtain to-be-detected data; inputting the to-be-detected data into the classification detection model for detection and outputting a detection result.

[0012] Preferably, processing the waveforms of all the ultrasonic data to obtain feature data of the multi-layer coating interface bonding quality characterization comprises:

[0013] obtaining echo peak data of all the scanning waveforms, and performing Fourier transform on all the scanning waveforms to obtain corresponding frequency spectrum waveforms;

[0014] obtaining maximum value data of all the frequency spectrum waveforms except the first maximum value data, and calculating the feature data of the multi-layer coating interface bonding quality characterization according to the echo peak data and the maximum value data.

[0015] Preferably, marking the coating type corresponding to the feature data to obtain label data comprises:

[0016] if the coating type is a topcoat failure coating, the label data is marked as 1;

[0017] if the coating type is a primer failure coating, the label data is marked as 2;

[0018] if the coating type is a non-failure coating, the label data is marked as 3.

[0019] Preferably, performing dimension reduction processing on the feature set by principal component analysis to obtain dimension reduction data comprises:

[0020] standardizing the data of the feature set to obtain a standardized feature set;

[0021] calculating the correlation coefficient of the data of the standardized feature set to obtain correlation coefficient data and constructing a P×P correlation coefficient matrix according to the correlation coefficient data;

[0022] According to the correlation coefficient matrix calculation, P matrix eigenvalues and P matrix eigenvectors of the matrix are obtained;

[0023] According to the P matrix eigenvalues, a contribution rate corresponding to each matrix eigenvalue is calculated;

[0024] From all the contribution rates, m matrix eigenvalues corresponding to a cumulative contribution rate greater than a contribution rate threshold are extracted;

[0025] According to the normalized feature set and the P matrix eigenvectors, corresponding reduced dimension data is calculated;

[0026] Wherein, m is a natural number less than P and greater than 0.

[0027] Preferably, the ultrasonic detection method of the interface bonding quality of the multi-layer coating comprises: according to the P matrix eigenvalues, a contribution rate corresponding to each matrix eigenvalue is calculated by using a contribution rate calculation formula; the contribution rate calculation formula is:

[0028]

[0029] In the formula, λ is the matrix eigenvalue of the matrix, e j is the contribution rate corresponding to the jth matrix eigenvalue.

[0030] Preferably, the ultrasonic detection method of the interface bonding quality of the multi-layer coating comprises: according to the normalized feature set and the P matrix eigenvectors, corresponding reduced dimension data is calculated by using a principal component score calculation formula; the principal component score calculation formula is:

[0031]

[0032] In the formula, x i * is the i th data in the normalized feature set, a ji is the i th matrix eigenvector corresponding to the j th matrix eigenvalue, F j is the reduced dimension data corresponding to the j th matrix eigenvalue.

[0033] Preferably, the detection result output by inputting the to-be-detected data into the classification detection model comprises:

[0034] If the number of sampling points of the to-be-detected scanning waveform is K×L, the number of label data recorded as 1 is b1, the number of label data recorded as 2 is b2, and the number of label data recorded as 3 is b3; the detection result output is:

[0035] The proportion of coating topcoat failure is α=b1 / (K×L);

[0036] The ratio of the coating primer failure is β = b2 / (K x L);

[0037] The ratio of the coating not failure is χ = b3 / (K x L).

[0038] In another aspect, an ultrasonic detection device for multi-layer coating interface bonding quality is provided, comprising a data acquisition module, a waveform processing module, a data processing module, a model training module and a detection output module;

[0039] The data acquisition module is configured to acquire parameter data and ultrasonic data of a test sample, wherein the parameter data includes the topcoat thickness and the total thickness of the multi-layer coating, and the ultrasonic data includes ultrasonic signals of the topcoat failure coating, the primer failure coating and the non-failure coating.

[0040] The waveform processing module is configured to process the waveforms of all the ultrasonic data to obtain feature data representing the multi-layer coating interface bonding quality, and to use the parameter data and the feature data as a feature set.

[0041] The data processing module is configured to label the coating types corresponding to the feature data to obtain label data, and to use a principal component analysis method to reduce the dimension of the feature set to obtain reduced dimension data, and to use the reduced dimension data and the label data as training data.

[0042] The model training module is configured to use a support vector machine to train the training data to obtain a classification detection model for the coating bonding quality.

[0043] The detection output module is configured to acquire to-be-detected scanning waveforms of a to-be-detected multi-layer coating sample, to process all the to-be-detected scanning waveforms using the waveform processing module and the data processing module to obtain to-be-detected data, and to input the to-be-detected data into the classification detection model for detection and output a detection result.

[0044] Preferably, the data processing module comprises a standardization submodule, a matrix construction submodule, a first calculation submodule, a second calculation submodule, an extraction submodule and a third calculation submodule.

[0045] The standardization submodule is configured to perform standardization processing on the data of the feature set to obtain a standardized feature set.

[0046] The matrix construction submodule is configured to perform correlation coefficient calculation on the data of the standardized feature set to obtain correlation coefficient data and construct a P x P correlation coefficient matrix according to the correlation coefficient data.

[0047] The first calculation submodule is configured to calculate the P matrix eigenvalues and P matrix eigenvectors of the matrix according to the correlation coefficient matrix.

[0048] The second calculation sub-module is configured to calculate a contribution rate corresponding to each matrix eigenvalue according to a contribution rate calculation formula of P matrix eigenvalues;

[0049] The extraction sub-module is configured to extract m matrix eigenvalues corresponding to a cumulative contribution rate greater than a contribution rate threshold from all the contribution rates;

[0050] The third calculation sub-module is configured to calculate corresponding dimension reduction data according to the normalized feature set and P matrix characteristic vectors according to a principal component score calculation formula;

[0051] wherein m is a natural number less than P and greater than 0; the contribution rate calculation formula is:

[0052]

[0053] The principal component score calculation formula is:

[0054]

[0055] wherein xi is the ith data in the normalized feature set, aj is the ith matrix characteristic vector corresponding to the jth matrix eigenvalue, Fj is the dimension reduction data corresponding to the jth matrix eigenvalue, λ is the matrix eigenvalue of the matrix, and ej is the contribution rate corresponding to the jth matrix eigenvalue. i * ji j j

[0056] In another aspect, a terminal device is provided, comprising a processor and a memory;

[0057] The memory is configured to store program code and transmit the program code to the processor;

[0058] The processor is configured to execute the above-mentioned ultrasonic detection method for multi-layer coating interface bonding quality according to instructions in the program code.

[0059] ​​​​It can be seen from the above technical solutions that the multi-layer coating interface bonding quality ultrasonic detection method, device and equipment have the following advantages: the method comprises the following steps: acquiring parameter data and ultrasonic data of a test sample; processing waveforms of all the ultrasonic data to obtain feature data representing the multi-layer coating interface bonding quality; taking the parameter data and the feature data as a feature set; marking coating types corresponding to the feature data to obtain label data, and performing dimension reduction processing on the feature set by using a principal component analysis method to obtain dimension reduction data, wherein the dimension reduction data and the label data are taken as training data; training the training data by using a support vector machine to obtain a classification detection model of the coating bonding quality; processing all the to-be-detected scan waveforms of a to-be-detected multi-layer coating sample by using steps S2 and S3 to obtain to-be-detected data; and inputting the to-be-detected data into the classification detection model for detection, and outputting a detection result. The multi-layer coating interface bonding quality ultrasonic detection method comprises the following steps: acquiring parameter data and ultrasonic data of a test sample, extracting relevant features, constructing a feature set, performing dimension reduction on the feature set by using a principal component analysis method to obtain the first m data that cumulatively reach a contribution rate threshold, inputting training data composed of label data and dimension reduction data into a support vector machine for training, obtaining a classification detection model of coating interface bonding quality, inputting data of a to-be-detected sample into the classification detection model to realize quantitative detection of coating bonding quality, and considering multiple features that affect coating interface bonding quality, finally quantitatively detecting the results of topcoat failure and primer failure, and solving the technical problems that existing ultrasonic nondestructive detection of coating interface quality uses a single feature to represent the interface of the coating, and the detection is not comprehensive and the result is inaccurate. BRIEF DESCRIPTION OF DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0061] Figure 1 The step flow chart of the multi-layer coating interface bonding quality ultrasonic detection method described in the embodiments of the present application;

[0062] Figure 2 The finite element simulation model in the multi-layer coating interface bonding quality ultrasonic detection method described in the embodiments of the present application;

[0063] Figure 3 The time domain graph of ultrasonic signals of different coating types in the multi-layer coating interface bonding quality ultrasonic detection method described in the embodiments of the present application;

[0064] Figure 4 The amplitude spectrum diagram of different coating types in the ultrasonic detection method of the interface bonding quality of the multilayer coating described in the embodiments of the present application;

[0065] Figure 5 The frame diagram of the ultrasonic detection device of the interface bonding quality of the multilayer coating in the embodiments of the present application. DETAILED DESCRIPTION

[0066] In order to make the objectives, characteristics and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the following described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0067] The embodiments of the present application provide an ultrasonic detection method, device and equipment for the interface bonding quality of a multilayer coating, which are used to solve the technical problems that the existing ultrasonic nondestructive detection coating interface quality is characterized by a single feature to represent the interface of the coating, and the detection is not comprehensive and the result is inaccurate.

[0068] Embodiment one:

[0069] Figure 1 The step flow chart of the ultrasonic detection method of the interface bonding quality of the multilayer coating described in the embodiments of the present application, Figure 2 The finite element simulation model diagram in the ultrasonic detection method of the interface bonding quality of the multilayer coating described in the embodiments of the present application, Figure 3 The time domain diagram of the ultrasonic signal of different coating types in the ultrasonic detection method of the interface bonding quality of the multilayer coating described in the embodiments of the present application, Figure 4 The amplitude spectrum diagram of different coating types in the ultrasonic detection method of the interface bonding quality of the multilayer coating described in the embodiments of the present application.

[0070] As Figure 1 shown, the embodiments of the present application provide an ultrasonic detection method for the interface bonding quality of a multilayer coating, which includes the following steps:

[0071] S1. Obtain the parameter data and ultrasonic data of the test sample, the parameter data including the topcoat thickness T1 and the total thickness T2 of the multilayer coating, and the ultrasonic data including the ultrasonic signal of the topcoat failure coating, the ultrasonic signal of the primer failure coating and the ultrasonic signal of the non-failure coating.

[0072] It should be noted that the data of the test sample of the multi-layer coating structure is obtained in step S1. In the embodiment, the ultrasonic data of the test sample is obtained by moving the ultrasonic transducer to the area of the test sample to be detected, adjusting the ultrasonic transducer to be perpendicular to the surface of the test sample, and then adjusting the Z-axis of the ultrasonic transducer so that the test sample is located in the focal point area of the ultrasonic transducer. The ultrasonic wave emitted by the ultrasonic transducer is incident to the water / topcoat interface, the topcoat / primer interface and the primer / substrate interface, and the ultrasonic wave is reflected and transmitted at the interfaces. The reflected echo signal of the test sample is collected as the detected ultrasonic signal.

[0073] In the embodiment of the present application, as shown in Figure 2 , the multi-layer coating structure includes four layers of water, topcoat, primer and substrate. An elastic thin layer is arranged at the bonding interface between the substrate and the primer and between the primer and the topcoat. The interface stiffness coefficient K t of the elastic thin layer is adjusted to simulate the good and bad bonding quality of the coating interface. An ultrasonic excitation source is arranged above the topcoat. The ultrasonic excitation adopts a Gaussian echo model.

[0074] It should be noted that the ultrasonic signal g(t) emitted by the ultrasonic excitation source has the following expression:

[0075]

[0076] In the formula, b is the bandwidth coefficient of the ultrasonic signal, t is the time of the ultrasonic signal, f c is the center frequency of the ultrasonic signal.

[0077] In the embodiment of the present application, the ultrasonic signals P1(t), P2(t) and P3(t) of the three types of coatings of the topcoat failure, the primer failure and the non-failure are obtained in step S1, as shown in Figure 3 , t is the time, and the unit is us.

[0078] S2. Process the waveforms of all the ultrasonic data to obtain the characteristic data of the multi-layer coating interface bonding quality characterization, and take the parameter data and the characteristic data as the feature set.

[0079] It should be noted that in step S2, firstly, the waveforms of all the ultrasonic data obtained in step S1 are processed to obtain the characteristic data related to the multi-layer coating interface bonding quality characterization; and secondly, the feature set is constructed. In the embodiment, the characteristic data is F=[R 21 ,R 31 ,S 23 ,S 34 ,S 45 ]. Meanwhile, the thickness of the multi-layer coating also has an influence on the interface bonding quality of the multi-layer coating, so the topcoat thickness T1 and the total thickness T2 are added to the characteristic data to obtain the feature set F*=[T1, T2, R21 R 31 S 23 S 34 S 45 ],wherein R is a waveform echo peak to peak ratio, and S is a maximum value ratio.

[0080] In the embodiment of the present application, before processing the waveform of all the ultrasonic data, the ultrasonic detection method for the interface bonding quality of the multilayer coating first constructs a finite element simulation model of the multilayer coating structure by using COMSOL software, and determines the characteristic data representing the bonding quality of the multilayer coating according to the finite element simulation model, so as to provide the required characteristic data for processing the ultrasonic data in step S2.

[0081] S3. Labeling the coating type corresponding to the characteristic data to obtain label data, and performing dimension reduction processing on the feature set by using a principal component analysis method to obtain dimension reduction data, wherein the dimension reduction data and the label data are used as training data.

[0082] It should be noted that in step S3, firstly, the coating type corresponding to the characteristic data in the feature set is labeled to obtain label data; secondly, the dimension reduction data is obtained by performing dimension reduction processing on the data of the feature set by using a principal component analysis method, and the dimension reduction data and the label data are used as training data. In the embodiment, the coating types include a topcoat failure coating, a primer failure coating, and an unfailed coating.

[0083] S4. Training the training data by using a support vector machine to obtain a classification detection model of the coating bonding quality.

[0084] It should be noted that the classification detection model of the coating bonding quality is obtained by training the training data by using a support vector machine (SVM). In the embodiment, the dimension reduction data obtained after the dimension reduction and the actual label data corresponding thereto are used as the input of the support vector machine (SVM). In the training process, the optimal parameters of the support vector machine (SVM) are obtained by using a grid search method, and then the classification detection model for detecting the coating bonding quality is obtained after a large amount of data is trained. The optimization of the parameters of the support vector machine (SVM) can also be performed by using a simulated annealing method, a genetic algorithm, etc. The algorithm for classification detection in the classification detection model can also be a random forest, a BP neural network, a convolutional neural network, etc.

[0085] S5. Obtaining the to-be-detected scanning waveform of the to-be-detected multilayer coating sample, processing all the to-be-detected scanning waveforms by using steps S2 and S3 to obtain to-be-detected data, inputting the to-be-detected data into the classification detection model for detection, and outputting a detection result.

[0086] It should be noted that in step S5, firstly, the to-be-tested scanning waveform of the to-be-tested multi-layer coating sample is acquired, and the acquired to-be-tested scanning waveform is processed through steps S2 and S3 to obtain to-be-tested data; secondly, the to-be-tested data is classified and detected by using the classification detection model obtained in step S4 to obtain the detection result of the interface bonding quality of the to-be-tested multi-layer coating sample. In the embodiment, in the process of acquiring the to-be-tested scanning waveform of the to-be-tested multi-layer coating sample, the to-be-tested multi-layer coating sample is scanned by using the step interval x and the scanning interval y, and the number of points collected is K*L (K and L are positive integers), K and L are respectively the number of sampling points of the scanning axis and the number of sampling points of the stepping axis; for different coatings, the unit scanning and stepping distance is selected by the GB / T9286-1998 grid test. The acquired to-be-tested scanning waveform is processed through steps S2 and S3, and to-be-tested data containing a three-dimensional matrix Q m×K×L of the first m principal component scores can be obtained, and the matrix Q m×K×L of the to-be-tested data is input into the classification detection model to obtain a detection result of a classification result matrix W K×L . The content of GB / T9286-1998 is an industry rule in the field, which is not limited in detail here.

[0087] In the embodiment of the application, the to-be-tested data is input into the classification detection model for detection, and the output detection result includes:

[0088] If the number of sampling points of the to-be-tested scanning waveform is K*L, the number of label data recorded as 1 is b1, the number of label data recorded as 2 is b2, and the number of label data recorded as 3 is b3; the output detection result is:

[0089] The proportion of coating topcoat failure is α = b1 / (K*L);

[0090] The proportion of coating primer failure is β = b2 / (K*L);

[0091] The proportion of coating non-failure is χ = b3 / (K*L).

[0092] It should be noted that according to the grade range specified by the GB / T9286-1998 grid test, the grading results 0B-5B of each layer of the coating of the test sample are obtained from the proportion data of the output result, and finally the quantitative detection of the coating interface bonding strength can be realized.

[0093] The application provides an ultrasonic detection method for interface bonding quality of a multilayer coating. The method comprises the following steps: acquiring parameter data and ultrasonic data of a test sample; processing waveforms of all ultrasonic data to obtain characteristic data for characterizing interface bonding quality of the multilayer coating; taking the parameter data and the characteristic data as a feature set; marking a coating type corresponding to the characteristic data to obtain label data, and performing dimension reduction processing on the feature set by using a principal component analysis method to obtain dimension reduction data, taking the dimension reduction data and the label data as training data; training the training data by using a support vector machine to obtain a classification detection model for coating bonding quality; acquiring to-be-detected scanning waveforms of a to-be-detected multilayer coating sample, processing all to-be-detected scanning waveforms by using steps S2 and S3 to obtain to-be-detected data; and inputting the to-be-detected data into the classification detection model for detection and outputting a detection result. The ultrasonic detection method for interface bonding quality of the multilayer coating comprises the following steps: acquiring parameter data and ultrasonic data of a test sample and extracting relevant features to form a feature set, performing dimension reduction on the feature set by using a principal component analysis method to obtain the first m data that cumulatively reach a contribution rate threshold to obtain dimension reduction data, inputting training data composed of label data with an actual bonding state and the dimension reduction data into a support vector machine for training to obtain a classification detection model for coating interface bonding quality, and inputting data of a to-be-detected sample into the classification detection model to realize quantitative detection of coating bonding quality. The method considers multiple features that affect interface bonding quality of the coating, and finally quantitatively detects the results of topcoat failure and primer failure, thereby solving the technical problems that existing ultrasonic nondestructive detection of coating interface quality uses a single feature to represent the interface of the coating, and the detection is not comprehensive and the result is inaccurate.

[0094] As shown in Figure 3 and Figure 4 In one embodiment of the application, processing waveforms of all ultrasonic data to obtain characteristic data for characterizing interface bonding quality of the multilayer coating comprises the following steps:

[0095] Acquiring echo peak data of all scanning waveforms, and performing Fourier transform on all scanning waveforms to obtain corresponding frequency spectrum waveforms;

[0096] Acquiring maximum value data of all frequency spectrum waveforms except the first maximum value data, and calculating the echo peak data and the maximum value data to obtain characteristic data for characterizing interface bonding quality of the multilayer coating.

[0097] It should be noted that, as shown in Figure 3As shown, the data corresponding to the first echo peak peak value PP1, the second echo peak peak value PP2 and the third echo peak peak value PP3 in the ultrasonic data scanning waveform P1(t), P2(t), P3(t) are obtained, and since the second echo peak peak value PP2 and the third echo peak peak value PP3 are more sensitive to the interface bonding quality of the multi-layer coating, the characteristic data representing the interface bonding quality of the multi-layer coating is calculated according to the echo peak data: R 21 = PP2 / PP1 and R 31 = PP3 / PP1. The spectrum waveform of the ultrasonic detection method of the interface bonding quality of the multi-layer coating is A1(f), A2(f), A3(f), and f is the frequency, unit: MHz. It can be seen from Figure 4 that the maximum data includes B2, B3, B4 and B5, and the characteristic data representing the interface bonding quality of the multi-layer coating is calculated according to the maximum data: S 23 = B2 / B3, S 34 = B3 / B4 and S 45 = B4 / B5.

[0098] In an embodiment of the present application, marking the coating type corresponding to the characteristic data to obtain the label data includes:

[0099] If the coating type is a topcoat failure coating, the label data is marked as 1;

[0100] If the coating type is a primer failure coating, the label data is marked as 2;

[0101] If the coating type is a non-failed coating, the label data is marked as 3.

[0102] In an embodiment of the present application, the feature set is reduced in dimension by principal component analysis to obtain reduced dimension data, which includes:

[0103] The data of the feature set is standardized to obtain a standardized feature set;

[0104] The data of the standardized feature set is calculated to obtain correlation coefficient data and a P×P correlation coefficient matrix is constructed according to the correlation coefficient data;

[0105] According to the correlation coefficient matrix calculation, P matrix eigenvalues and P matrix characteristic vectors of the matrix are obtained;

[0106] According to the P matrix eigenvalues, the contribution rate corresponding to each matrix eigenvalue is calculated;

[0107] From all the contribution rates, m matrix eigenvalues corresponding to the cumulative contribution rate greater than the contribution rate threshold are extracted;

[0108] According to the standardized feature set and P matrix characteristic vectors, corresponding dimension-reduced data is obtained;

[0109] wherein m is a natural number less than P and greater than 0.

[0110] It should be noted that the contribution rate threshold can be set according to requirements. The standardized feature set is obtained by data standardization processing Then, the correlation coefficient between two data in the feature set is calculated by using a correlation coefficient calculation formula, and a P*P correlation coefficient matrix is constructed. The correlation coefficient calculation formula is:

[0111]

[0112] In the formula, x i * , x j * are the i th and j th data in the standardized feature set, cov(x i * , x j * is the covariance of x i * and x j * , var[x i * ] is the variance of x i * , and var[x j * ] is the variance of x j * . Secondly, the matrix eigenvalues and matrix eigenvectors a j = [a j1 , a j2 …, a jp ] of the matrix are calculated according to the correlation coefficient matrix. The data standardization processing, the calculation of the eigenvalues and eigenvectors of the matrix are relatively mature technologies in the art, and will not be described in detail here. In this embodiment, the contribution rate threshold is 0.85%.

[0113] In the embodiment of the application, the ultrasonic detection method for the interface bonding quality of the multilayer coating includes: according to P matrix eigenvalues, a contribution rate corresponding to each matrix eigenvalue is obtained by using a contribution rate calculation formula:

[0114]

[0115] In the formula, λ is the matrix eigenvalue, and e j is the contribution rate corresponding to the j th matrix eigenvalue.

[0116] In the embodiment of the present application, the ultrasonic detection method for the interface bonding quality of the multi-layer coating comprises: calculating corresponding dimension reduction data according to the normalized feature set and P matrix characteristic vectors by using a principal component score calculation formula:

[0117]

[0118] In the formula, x i * is the i th data in the normalized feature set, a ji is the i th matrix characteristic vector corresponding to the j th matrix characteristic value, F j is the dimension reduction data corresponding to the j th matrix characteristic value.

[0119] It should be noted that the ultrasonic detection method for the interface bonding quality of the multi-layer coating can also use kernel principal component analysis (KPCA), linear discriminant analysis (LDA), multidimensional scaling transformation (MDS) and the like to obtain dimension reduction data by dimension reduction processing of the feature set.

[0120] Embodiment two:

[0121] Figure 5 is a frame flowchart of the ultrasonic detection device for the interface bonding quality of the multi-layer coating described in the embodiment of the present application.

[0122] As Figure 5 shown, the embodiment of the present application provides an ultrasonic detection device for the interface bonding quality of the multi-layer coating, which comprises a data acquisition module 10, a waveform processing module 20, a data processing module 30, a model training module 40 and a detection output module 50.

[0123] The data acquisition module 10 is used to acquire parameter data and ultrasonic data of a test sample, wherein the parameter data comprises the thickness of the multi-layer coating and the total thickness, and the ultrasonic data comprises the ultrasonic signal of the topcoat failure coating, the ultrasonic signal of the primer failure coating and the ultrasonic signal of the non-failure coating.

[0124] The waveform processing module 20 is used to process the waveforms of all the ultrasonic data to obtain feature data representing the interface bonding quality of the multi-layer coating; and the parameter data and the feature data are taken as a feature set.

[0125] The data processing module 30 is used to mark the coating type corresponding to the feature data to obtain label data, and to perform dimension reduction processing on the feature set by using a principal component analysis method to obtain dimension reduction data; and the dimension reduction data and the label data are taken as training data.

[0126] The model training module 40 is used to train the training data by using a support vector machine to obtain a classification detection model for the coating bonding quality.

[0127] The detection output module 50 is configured to acquire to-be-detected scanning waveforms of a to-be-detected multilayer coating sample, process all the to-be-detected scanning waveforms by using the waveform processing module and the data processing module to obtain to-be-detected data, input the to-be-detected data into a classification detection model for detection, and output a detection result.

[0128] In the embodiment of the present application, the data processing module 30 includes a standardization submodule, a matrix construction submodule, a first calculation submodule, a second calculation submodule, an extraction submodule, and a third calculation submodule.

[0129] The standardization submodule is configured to perform standardization processing on data of the feature set to obtain a standardized feature set.

[0130] The matrix construction submodule is configured to perform correlation coefficient calculation on data of the standardized feature set to obtain correlation coefficient data and construct a P*P correlation coefficient matrix according to the correlation coefficient data.

[0131] The first calculation submodule is configured to calculate the P matrix eigenvalues and P matrix eigenvectors of the matrix according to the correlation coefficient matrix.

[0132] The second calculation submodule is configured to calculate the contribution rate corresponding to each matrix eigenvalue according to the P matrix eigenvalues by using a contribution rate calculation formula.

[0133] The extraction submodule is configured to extract m matrix eigenvalues corresponding to a cumulative contribution rate greater than a contribution rate threshold from all the contribution rates.

[0134] The third calculation submodule is configured to calculate corresponding dimension reduction data according to the standardized feature set and the P matrix eigenvectors by using a principal component score calculation formula.

[0135] Wherein, m is a natural number less than P and greater than 0; the contribution rate calculation formula is:

[0136]

[0137] The principal component score calculation formula is:

[0138]

[0139] In the formula, x i * is the i th data in the standardized feature set, a ji is the i th matrix eigenvector corresponding to the j th matrix eigenvalue, F j is the dimension reduction data corresponding to the j th matrix eigenvalue, λ is the matrix eigenvalue of the matrix, and e j is the contribution rate corresponding to the j th matrix eigenvalue.

[0140] It should be noted that the modules in the device of Embodiment Two correspond to the steps in the method of Embodiment One, and the content of the ultrasonic detection method for the interface bonding quality of the multi-layer coating has been described in detail in Embodiment One, and the content of the modules in the device will not be described in detail in Embodiment Two.

[0141] Embodiment Three

[0142] The embodiment of the present application provides a terminal device, comprising a processor and a memory.

[0143] The memory is configured to store program code and transmit the program code to the processor.

[0144] The processor is configured to execute the ultrasonic detection method for the interface bonding quality of the multi-layer coating according to the instructions in the program code.

[0145] It should be noted that the processor is configured to execute the steps in the ultrasonic detection method for the interface bonding quality of the multi-layer coating according to the instructions in the program code. Alternatively, the processor executes the computer program to realize the functions of the modules / units in each system / device embodiment.

[0146] For example, the computer program can be divided into one or more modules / units, one or more modules / units are stored in the memory and executed by the processor to complete the present application. One or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which is used to describe the execution process of the computer program in the terminal device.

[0147] The terminal device can be a desktop computer, a notebook computer, a palm computer, a cloud server and the like. The terminal device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that it does not constitute a limitation on the terminal device, and can include more or fewer components than the illustration, or combine certain components, or different components, for example, the terminal device can also include an input / output device, a network access device, a bus and the like.

[0148] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0149] The memory can be an internal storage unit of the terminal device, such as a hard disk or a memory of the terminal device. The memory can also be an external storage device of the terminal device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device. Further, the memory can include both the internal storage unit and the external storage device of the terminal device. The memory is used to store computer programs and other programs and data required by the terminal device. The memory can also be used to temporarily store data that has been output or will be output.

[0150] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0151] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0152] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0153] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0154] The integrated unit, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in the form of a contribution to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk.

[0155] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method of ultrasonic inspection of the interface bond quality of a multilayer coating, characterized in that, The method comprises the following steps: S1. obtaining parameter data and ultrasonic data of a test sample, wherein the parameter data comprises a topcoat thickness and a total thickness of a multi-layer coating, and the ultrasonic data comprises an ultrasonic signal of a topcoat failure coating, an ultrasonic signal of a primer failure coating, and an ultrasonic signal of an unfailed coating; S2. processing waveforms of all the ultrasonic data to obtain feature data representing the interface bonding quality of the multi-layer coating; taking the parameter data and the feature data as a feature set; S3. marking a coating type corresponding to the feature data to obtain label data, and performing dimension reduction processing on the feature set by using a principal component analysis method to obtain reduced dimension data, taking the reduced dimension data and the label data as training data; S4. training the training data by using a support vector machine to obtain a classification detection model of the coating bonding quality; S5. obtaining to-be-detected scanning waveforms of a to-be-detected multi-layer coating sample, processing all the to-be-detected scanning waveforms by using steps S2 and S3 to obtain to-be-detected data, inputting the to-be-detected data into the classification detection model for detection, and outputting a detection result; processing waveforms of all the ultrasonic data to obtain feature data representing the interface bonding quality of the multi-layer coating comprises: obtaining echo peak data of all the scanning waveforms, and obtaining corresponding frequency spectrum waveforms of all the scanning waveforms by using a Fourier transform; obtaining maximum value data of all the frequency spectrum waveforms except the first maximum value data, and calculating the feature data representing the interface bonding quality of the multi-layer coating according to the echo peak data and the maximum value data; The characteristic data of the interface bonding quality of the multi-layer coating calculated according to the echo peak data is: and The characteristic data of the interface bonding quality of the multi-layer coating calculated according to the maximum data is: , and In the formula, PP1, PP2 and PP3 are respectively the first echo peak value, the second echo peak value and the third echo peak value in the sweep waveform of the ultrasonic data, R 21 is the ratio of the second echo peak value to the first echo peak value, R 31 is the ratio of the third echo peak value to the first echo peak value; B2, B3, B4 and B5 are respectively the second maximum data, the third maximum data, the fourth maximum data and the fifth maximum data of the frequency spectrum waveform, S 23 is the ratio of the second maximum data to the third maximum data, S 34 is the ratio of the third maximum data to the fourth maximum data, and S 45 is the ratio of the fourth maximum data to the fifth maximum data.

2. The method of ultrasonic inspection of multi-layer coating interface bond quality according to claim 1, characterized in that, marking a coating type corresponding to the feature data to obtain label data comprises: if the coating type is a topcoat failure coating, marking the label data as 1; if the coating type is a primer failure coating, marking the label data as 2; if the coating type is an unfailed coating, marking the label data as 3.

3. The method of claim 1, wherein, performing dimension reduction processing on the feature set by using a principal component analysis method to obtain reduced dimension data comprises: performing standardization processing on data of the feature set to obtain a standardized feature set; performing correlation coefficient calculation on data of the standardized feature set to obtain correlation coefficient data and constructing a P×P correlation coefficient matrix according to the correlation coefficient data; calculating P matrix characteristic values and P matrix characteristic vectors of the correlation coefficient matrix; calculating a contribution rate corresponding to each matrix characteristic value according to the P matrix characteristic values; extracting m matrix characteristic values corresponding to a cumulative contribution rate greater than a contribution rate threshold from all the contribution rates; calculating corresponding reduced dimension data according to the standardized feature set and the P matrix characteristic vectors; wherein m is a natural number less than P and greater than 0.

4. The method of claim 3, wherein, comprises: calculating a contribution rate corresponding to each matrix characteristic value according to the P matrix characteristic values by using a contribution rate calculation formula; the contribution rate calculation formula is: ； wherein is the matrix eigenvalue of the matrix, e j is the contribution rate corresponding to the jth matrix eigenvalue.

5. The method of claim 3, wherein, comprises: calculating corresponding reduced dimension data according to the standardized feature set and the P matrix characteristic vectors by using a principal component score calculation formula; the principal component score calculation formula is: ； In the formula, x i * is the i-th data in the standardized feature set, a ji is the i-th matrix characteristic vector corresponding to the j-th matrix eigenvalue, F j is the reduced dimension data corresponding to the j-th matrix eigenvalue.

6. The method of claim 1, wherein, The detection result outputted by inputting the to-be-tested data into the classification detection model includes: If the sampling point number of the to-be-tested scanning waveform is K*L, the number of label data recorded as 1 is b1, the number of label data recorded as 2 is b2, and the number of label data recorded as 3 is b3, the detection result outputted is: Percentage of coating topcoat failures To ; Percentage of coating primer failure To ; Percentage of coating that did not fail To .

7. An apparatus for ultrasonic detection of interface bond quality of a multilayer coating, characterized by It comprises: A data acquisition module, a waveform processing module, a data processing module, a model training module and a detection output module; The data acquisition module is used for acquiring parameter data and ultrasonic data of a test sample, wherein the parameter data includes the thickness and total thickness of a multi-layer coating surface paint, and the ultrasonic data includes the ultrasonic signals of a surface paint failure coating, a primer failure coating and an unfailed coating; The waveform processing module is used for processing the waveforms of all the ultrasonic data to obtain feature data representing the interface bonding quality of the multi-layer coating; The data processing module is used for marking the coating type corresponding to the feature data to obtain label data, and performing dimension reduction processing on the feature set by using a principal component analysis method to obtain reduced dimension data, wherein the reduced dimension data and the label data are used as training data; The model training module is used for training the training data by using a support vector machine to obtain a classification detection model of coating bonding quality; The detection output module is used for acquiring to-be-tested scanning waveforms of a to-be-tested multi-layer coating sample, processing all the to-be-tested scanning waveforms by using the waveform processing module and the data processing module to obtain to-be-tested data, inputting the to-be-tested data into the classification detection model for detection, and outputting a detection result; Processing the waveforms of all the ultrasonic data to obtain feature data representing the interface bonding quality of the multi-layer coating includes: Acquiring echo peak data of all the scanning waveforms, and obtaining corresponding frequency spectrum waveforms by using Fourier transform on all the scanning waveforms; Acquiring maximum value data of all the frequency spectrum waveforms except the first maximum value data, and calculating the feature data representing the interface bonding quality of the multi-layer coating according to the echo peak data and the maximum value data; The characteristic data of the interface bonding quality of the multi-layer coating calculated according to the echo peak data is: and The characteristic data of the interface bonding quality of the multi-layer coating calculated according to the maximum data is: , and In the formula, PP1, PP2 and PP3 are respectively the first echo peak value, the second echo peak value and the third echo peak value in the sweep waveform of the ultrasonic data, R 21 is the ratio of the second echo peak value to the first echo peak value, R 31 is the ratio of the third echo peak value to the first echo peak value; B2, B3, B4 and B5 are respectively the second maximum data, the third maximum data, the fourth maximum data and the fifth maximum data of the frequency spectrum waveform, S 23 is the ratio of the second maximum data to the third maximum data, S 34 is the ratio of the third maximum data to the fourth maximum data, S 45 is the ratio of the fourth maximum data to the fifth maximum data.

8. The apparatus of claim 7, wherein, The data processing module comprises a standardization submodule, a matrix construction submodule, a first calculation submodule, a second calculation submodule, an extraction submodule and a third calculation submodule; The standardization submodule is used for performing standardization processing on the data of the feature set to obtain a standardized feature set; The matrix construction submodule is used for performing correlation coefficient calculation on the data of the standardized feature set to obtain correlation coefficient data and constructing a P*P correlation coefficient matrix according to the correlation coefficient data; The first calculation submodule is used for calculating the P matrix eigenvalues and P matrix characteristic vectors of the matrix according to the correlation coefficient matrix; The second calculation submodule is used for calculating the contribution rate corresponding to each matrix eigenvalue according to the P matrix eigenvalues by using a contribution rate calculation formula; The extraction submodule is used for extracting m matrix eigenvalues corresponding to the cumulative contribution rate greater than a contribution rate threshold from all the contribution rates. The third calculation sub-module is configured to calculate corresponding dimension-reduced data according to the standardized feature set and P matrix characteristic vectors by using a principal component score calculation formula. Wherein, m is a natural number less than P and greater than 0; the contribution rate calculation formula is: ； The principal component score calculation formula is: ； In the formula, x i * is the i-th data in the standardized feature set, a ji is the i-th matrix characteristic vector corresponding to the j-th matrix eigenvalue, F j is the reduced dimension data corresponding to the j-th matrix eigenvalue, is the matrix eigenvalue of the matrix, e j is the contribution rate corresponding to the j-th matrix eigenvalue.

9. A terminal device, comprising: The processor and the memory are included. The memory is configured to store program code and transmit the program code to the processor. The processor is configured to execute the method for ultrasonic detection of the interface bonding quality of a multi-layer coating according to instructions in the program code.

Images