A method of non-destructive testing of a composite material defect

Abstract

The application discloses a kind of composite material defect nondestructive testing methods, it is related to nondestructive testing technical field, the method includes: first acquisition specimen A scanning waveform, based on echo amplitude and time difference determination potential defect area, output sequence containing spatial coordinates, original waveform and focusing parameter;After pretreatment, intercept effective time window, generate sub-band by wavelet packet decomposition, calculate total energy and normalize energy vector again.Based on known sample, sub-band separability is evaluated and sorted using Fisher criterion, the optimal energy dimension is determined by cross-validation, sub-band energy features are combined with phase, time difference features to form a composite vector, input into a dual-channel lightweight deep network, output defect category and confidence and map to C-scan frame image;Finally, trace back three-dimensional coordinates, calculate defect volume and remaining wall thickness, compare material performance database output conclusion;Solve the misjudgment problem caused by echo waveform similarity, realize detection closed loop optimization.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and more specifically, to a nondestructive testing method for defects in composite materials. Background Technology

[0002] In some technologies, ultrasonic testing is an important means of non-destructive testing of defects in composite materials. Among them, phased array ultrasonic technology acquires A-scan waveform sequences by scanning and uses the characteristics of echo signals to achieve defect detection, and it is widely used in the industrial field.

[0003] However, existing methods have significant shortcomings: First, the echo waveforms of different defects are highly similar, and relying solely on single-band energy characteristics for analysis makes it difficult to accurately distinguish defect types, easily leading to misjudgments; Second, the feature extraction dimensions are limited, often ignoring multi-dimensional information such as phase centroid, phase dispersion, and echo time difference, resulting in insufficient feature fusion and affecting the accuracy of defect identification; Third, the scanning strategy for potential defect areas is fixed, lacking dynamic fine scanning optimization, which easily leads to an imbalance between detection efficiency and accuracy; Fourth, defect assessment mostly remains at the level of type identification, failing to effectively combine the three-dimensional parameters of the defect (such as volume and remaining wall thickness) with the material performance database for comprehensive judgment, making it difficult to provide clear quality control guidance.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] The purpose of this invention is to provide a non-destructive testing method for defects in composite materials, which solves the technical problem in existing ultrasonic testing where different echo waveforms are highly similar and it is difficult to accurately distinguish defect types by relying solely on energy from a single frequency band.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A non-destructive testing method for defects in composite materials includes the following steps: In a constant-temperature water bath, a phased array transducer is used to scan the composite material specimen to acquire the A-scan waveform at each measuring point. Based on the echo amplitude and echo time difference between various echoes in the A-scan waveform, potential defect areas are determined, and an A-scan waveform sequence containing spatial coordinates, the original A-scan waveform, and focusing parameters is output. The A-scan waveform is a time-domain echo signal acquired by the phased array transducer. The focusing parameters at the measuring points include delay line parameters, array element aperture, and center frequency.

[0008] After preprocessing the original A-scan waveform, an effective time window containing surface echo, defect echo, and bottom surface echo is extracted. Multiple sub-bands are obtained through wavelet packet decomposition, and the total energy of each sub-band is calculated and normalized to form the corresponding sub-band energy vector.

[0009] Based on the sample of known defect types, the Fisher criterion is used to evaluate the separability index of each sub-band, and the energy vectors of each sub-band are sorted in descending order of the separability index to form an energy vector priority list of each sub-band;

[0010] An incremental dimension cross-validation method is used to adaptively determine the optimal energy dimension, and the selected sub-band energy vector features corresponding to the optimal energy dimension are combined with the phase centroid, phase dispersion and time difference features to form a composite feature vector;

[0011] Based on the obtained composite feature vector, the classification model is input to output the defect category and classification confidence, and the output result is mapped to the C-scan frame image with spatial coordinates as the dimension; Based on the C-scan frame image, the three-dimensional coordinates of the defect are traced back, the defect volume and the remaining wall thickness are calculated combined with the sound velocity, and the defect detection conclusion is output by comparing the material performance database.

[0012] As a further scheme of the application: when the phased array transducer is scanned, an N-element phased array transducer is used, where N is the number of elements, the element spacing of the phased array transducer is P, and the center frequency is .

[0013] As a further scheme of the application: determining the potential defect area includes: calculating the average amplitude of the surface echo, the average amplitude of the bottom echo and the noise standard deviation of the defect-free area in the A-scan waveform;

[0014] When the phased array transducer completes a scanning line along the X direction with an initial scanning interval, the A-scan waveform of each measurement point is analyzed in real time; if any of the following conditions is met, it is marked as a potential defect area and triggers the phased array transducer fine scanning:

[0015] When the absolute deviation of the surface echo amplitude or the bottom echo amplitude from the average surface echo amplitude exceeds the amplitude anomaly threshold, it is marked as a potential defect area and triggers the phased array transducer fine scanning;

[0016] Or when the surface echo appears at the moment and the bottom echo appears at the moment, the A-scan waveform peak value is greater than the defect echo determination threshold, which is marked as a potential defect area and triggers the phased array transducer fine scanning.

[0017] As a further scheme of the application: the phased array transducer fine scanning includes: calculating the defect depth estimation value based on the defect depth estimation formula as follows: ; , is the difference between the bottom echo appearance time and the surface echo appearance time, d is the longitudinal wave speed of the tested composite material; the center frequency is increased to the center frequency initial value plus a preset increase value, and the element number of the phased array transducer is halved.

[0018] As a further aspect of the present invention: the wavelet packet decomposition includes: performing noise reduction processing on the waveform using typical wavelets while retaining the main reflection information; and only extracting the effective time window containing surface echoes, defect echoes, and bottom surface echoes. in , To reserve a margin threshold, the original A-scan waveform within the above effective time window is decomposed using wavelet packet decomposition; the decomposition level L is set to obtain... There are several sub-bands; for each sub-band j, where, Sub-bands representing different center frequencies; their corresponding time-domain waveforms are obtained through wavelet packet reconstruction algorithm, denoted as reconstructed sub-waveforms. Where n is the time-domain measurement point index of the sub-waveform; and calculate the total energy of each sub-band. Sum the squares of the amplitudes at all measurement points corresponding to the reconstructed sub-waveform; sum the total energy of each sub-band. Divide by the sum of the total energy of all sub-bands The normalized relative energy value is obtained. ; and normalized relative energy values ​​for all sub-bands. Arranged in order to form sub-band energy vectors ;in This represents the number of sub-bands.

[0019] As a further aspect of this invention: when evaluating the sub-band separability index: based on pre-prepared samples of known defect types, calculate the mean and standard deviation of the energy of each sub-band; use Fisher's criterion to define the sub-band separability index, specifically: for each sub-band, calculate the mean energy difference between all pairs of defect types, then divide by the sum of the standard deviations of the energy of these two defect types, and finally add these results for all pairs of defect types to obtain the separability index of the sub-band; subsequently, sort all sub-band energy vectors in descending order of separability index value to form a priority list of sub-band energy vectors.

[0020] As a further aspect of the present invention: adaptively determining the optimal energy dimension includes: setting the maximum candidate dimension. For dimensions from 1 to the maximum candidate dimension For each dimension k, the energy vectors of the first k sub-bands in the sub-band energy vector priority list are selected sequentially as energy vector features. When the k-dimensional energy vector features undergo N-fold cross-validation, the validation accuracy is recorded. If the validation accuracy of the current dimension k is less than the threshold ε compared to the previous dimension k-1 and this occurs twice consecutively, the loop terminates, and the current dimension k is selected as the optimal energy dimension k*. If the current dimension k reaches the set maximum candidate dimension... If the above conditions are not met, then the maximum candidate dimension will be selected. k* is the optimal energy dimension.

[0021] As a further aspect of this invention: the classification model employs a dual-channel lightweight deep network structure, taking a composite feature vector as input and outputting the defect category and classification confidence score, including:

[0022] The energy-time difference channel takes the optimal energy dimension k* plus time difference features as input, and outputs an abstract feature vector through two layers of 1D convolution. Among them, the time difference characteristics include the time difference between surface echo and defect echo, and the time difference between surface echo and bottom echo;

[0023] The phase channel inputs the phase centroid and phase dispersion, which are mapped to an abstract feature vector through a two-layer fully connected network. Same-dimensional abstract feature vector Here, the phase centroid refers to the weighted average of the instantaneous phases within the time window containing the defect echo, with the weight being the signal amplitude at the corresponding moment; the phase dispersion refers to the deviation between the instantaneous phase and the phase centroid within the time window containing the defect echo, obtained by averaging the squares of these deviations by the corresponding signal amplitudes and then taking the square root; the fusion layer abstracts the feature vectors. and abstract feature vectors The elements are added one by one, and the output vector is obtained through two fully connected layers. Finally, the output class probability vector is activated by Softmax. Each defect class corresponds to a class probability vector. The defect class corresponding to the largest class probability vector is selected as the defect class determination result, and the class probability vector is recorded as the confidence level, which is called the classification confidence level.

[0024] As a further aspect of the present invention: after the phased array transducer is finely scanned, the method further includes: after a scan line is acquired, mapping the amplitude of the A-scan waveform of all measurement points of the scan line at the same depth gate to the pixel color, stacking them along the scanning direction according to the pixel color to form a C-scan array, and stitching them together to form a C-scan frame image; wherein, the depth gate includes the gate start time and the gate width, and the gate start time is the time corresponding to the estimated defect depth.

[0025] As a further aspect of the present invention: the defect detection conclusion generation includes: based on the defect category classification result, regenerating the C-scan image to obtain a new superimposed defect category determination result. The coordinates of the C-scan image corresponding to the classification confidence and defect category determination results are obtained. Based on the new C-scan image, the A-scan waveform is traced back to calculate the defect center depth, find its center coordinates, trace back the original A-scan waveform, and extract the time of appearance of surface echo, defect echo, and bottom surface echo.

[0026] The propagation time is obtained by subtracting the time of occurrence of the surface echo from the time of occurrence of the defect echo and multiplying by the ratio of the propagation path in water; and the calculation is split according to the sound wave propagation path:

[0027] The distance of sound wave propagation in water is the product of the sound speed in water medium and the propagation time;

[0028] The distance of sound wave propagation in the tested composite material sample is the product of the longitudinal wave speed in the tested composite material and the propagation time;

[0029] The half of the sum of the distance of sound wave propagation in water and the distance of sound wave propagation in the tested composite material sample is taken as the defect center depth; the defect depth estimate and its corresponding two-dimensional coordinates are combined with the calculated defect center depth to form a three-dimensional point set; for the three-dimensional point set belonging to the same defect type, a spatial clustering algorithm is used for clustering and then three-dimensional surface reconstruction is performed, after the reconstruction is completed, the defect volume is calculated by voxel integration summation, and the maximum radial size in the X-Y plane is calculated by the maximum Euclidean distance between any two points in the region and the remaining wall thickness obtained by subtracting the defect center depth and the maximum defect depth from the material thickness; wherein the maximum defect depth is the maximum value in the three-dimensional point set of the same defect type; the above three-dimensional parameters, i.e., the defect volume, the maximum radial size in the X-Y plane, and the remaining wall thickness, are input into the material performance database, and the defect volume and the remaining wall thickness are compared with the set upper limit of the defect volume and the lower limit of the remaining wall thickness to generate a defect detection conclusion; when the defect volume is not greater than the upper limit of the defect volume and the remaining wall thickness is not less than the lower limit of the remaining wall thickness, it is recorded as safe and acceptable;

[0030] When the defect volume after rounding is the same as the upper limit of the defect volume, or the remaining wall thickness after rounding is the same as the lower limit of the remaining wall thickness, it is recorded as needing repair;

[0031] When the defect volume is greater than the upper limit of the defect volume, or the remaining wall thickness is less than the lower limit of the remaining wall thickness, it is recorded as unqualified.

[0032] The beneficial effects of the present application are:

[0033] (1) The present application obtains an A-scan waveform sequence through a phased array transducer, determines a potential defect area in combination with echo amplitude and time difference and triggers fine scanning, optimizes scanning efficiency and accuracy, obtains multiple sub-bands through wavelet packet decomposition, evaluates separability by using Fisher criterion and determines an optimal energy dimension through incremental dimension cross-validation, breaks through the limitation of a single frequency band, accurately distinguishes defect types corresponding to similar echoes, combines the optimal energy feature and phase centroid, dispersion and time difference into a composite feature vector, inputs the double-channel lightweight deep network, and improves classification accuracy and confidence. BRIEF DESCRIPTION OF DRAWINGS

[0034] The present application will be further described below in conjunction with the accompanying drawings.

[0035] Figure 1 is a method flow diagram of a composite material defect nondestructive testing method of the present application;

[0036] Figure 2 is an ultrasonic detection schematic diagram of a composite material defect nondestructive testing method of the present application;

[0037] Figure 3 is an implementation logic diagram of step two in the composite material defect nondestructive testing method of the present application. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. 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. EMBODIMENT

[0039] This embodiment takes a carbon fiber composite material laminate with three different embedded inclusion defects (polytetrafluoroethylene film, CFRP sheet and plastic paper) as the detection object, the size of the plate is 200 mm x 150 mm x 2.5 mm, and the laying mode is ; The detection system includes a constant-temperature water tank, a transceiving integrated phased array transducer, a mechanical scanning device, etc.

[0040] Please refer to Figure 1 , the present application is a composite material defect nondestructive testing method, comprising the following steps:

[0041] Step one: under the condition of constant temperature water tank, using phased array transducer to scan in X-Y path, host computer sends out pulse and collects A-scan waveform; based on the amplitude / time difference of surface echo, defect echo and bottom echo compared with threshold value to determine potential defect; for potential defect area, call focus table according to defect depth estimation value to implement dynamic focusing, and conduct aperture and center frequency self-adaptation; color the continuous A-scan waveform according to depth gate to generate C-scan frame, and adjust the next row scanning interval according to defect length threshold value; output data set containing coordinates, A-scan waveform and focusing / gate record;

[0042] In this embodiment, the center frequency of the phased array transducer is selected as , and the initial value can be 5MHz, N=32, and the interval P of the phased array transducer is set to 0.5mm; the temperature of the water tank is maintained at 25℃±1℃ to stabilize the sound velocity of the water medium; the automatic scanning device is used to control the movement of the phased array transducer in the X-Y plane, and the initial scanning interval is set to 1mm;

[0043] Please refer to Figure 2 , the dedicated phased array host sends out a voltage excitation signal with a pulse width of 0.2μs and a voltage amplitude of 200V to the phased array transducer, which is incident on the tested composite material sample through the water medium; based on the difference in acoustic impedance between water and the tested composite material sample, in the typical defect-free area, the original A-scan waveform will have two main reflection peaks: the surface echo near zero time, and the appearance time is recorded as , which is the signal reflected by the sound wave through the target defect detection laminate surface; the bottom echo far from zero time, and the appearance time is recorded as , which is the signal reflected by the sound wave through the target defect detection laminate bottom; if there is an interface (such as delamination, porosity or inclusion) in the material, a defect echo will appear between and , recorded as , the amplitude and appearance time of which can preliminarily reflect the position and size of the defect; it should be noted that the A-scan waveform is the time-domain echo signal curve recorded after the phased array transducer transmits and receives once at a single measurement point; it only represents the reflection information along the sound path of the point;

[0044] The A-scan waveform is sampled at a rate of 100MHz, and the total sampling length is set to 4μs; after collection, the signal is band-pass filtered at 1~10MHz to improve the signal-to-noise ratio; the following calibrations need to be completed before detection:

[0045] 30 groups of A-scan waveforms are collected in the defect-free area of the tested composite material sample, and the average amplitude of the surface echo, the average amplitude of the bottom echo and the noise standard deviation of the defect-free area are calculated;

[0046] When the phased array transducer completes a scanning line along the X direction with an initial scanning pitch, the A-scan waveform of each measurement point is analyzed in real time; if any of the following conditions is met, it is marked as a potential defect area and triggers the phased array transducer to perform a fine scan:

[0047] When the absolute deviation of the surface echo amplitude from the average surface echo amplitude exceeds the amplitude anomaly threshold, or the absolute deviation of the bottom echo amplitude from the average surface echo amplitude exceeds the amplitude anomaly threshold, it is marked as a potential defect area;

[0048] Or when the A-scan waveform peak value between the surface echo occurrence time and the bottom echo occurrence time is greater than the defect echo determination threshold, it is marked as a potential defect area;

[0049] Wherein, the amplitude anomaly threshold is obtained by multiplying the set first proportion coefficient, whose initial value can be 4.0, by the noise standard deviation of the defect-free area;

[0050] The defect echo determination threshold is obtained by multiplying the set second proportion coefficient, whose initial value can be 3.0, by the noise standard deviation of the defect-free area, based on the noise signal zero offset correction, which corrects the noise mean value to 0, and the repeatable statistics to obtain the noise standard deviation of the defect-free area;

[0051] When the phased array transducer moves to within a preset neighborhood radius of the potential defect area, which can be 2mm, the following fine scanning operation is performed: calculate the appropriate measurement point focusing parameters based on the current estimated defect depth, and use the dynamic focusing strategy to improve the resolution, specifically:

[0052] Increase the center frequency by a preset increase value, which can be 2MHz, and halve the number of elements, i.e. adjust the element aperture from 32 elements to 16 elements, thereby reducing the beam diameter and increasing the lateral resolution;

[0053] Calculate the defect depth estimate h based on the defect depth estimation formula, i.e. the time difference between the defect echo and the surface echo divided by 2, then multiplied by the longitudinal wave speed of the test composite material, and the result is recorded as the defect depth estimate; the formula is as follows: ; Wherein, d is the longitudinal wave speed of the test composite material, and the initial value of CFRP is set to 5500m / s, and the bottom echo time difference in the defect-free area is used for online self-calibration, i.e. , is the known material thickness;

[0054] At the same time, call the experimentally calibrated focal length lookup table to rewrite the delay line parameters based on the defect depth estimate, so that the sound beam is focused on the center of the defect area; wherein,

[0055] The N-element configuration, i.e., 32 elements, and the initial center frequency are maintained for the region without abnormal areas, i.e., the region without triggering conditions The setting, i.e., 5 MHz setting, is set to improve the scanning efficiency;

[0056] When the acquisition of one scan line is completed, the A-scan waveform amplitude of all measurement points of the scan line at the same depth gate is mapped to the pixel color, and the C-scan column array is stacked in the pixel color along the scanning direction to form a C-scan frame image; wherein the depth gate includes a gate starting time and a gate width, and the gate starting time is the time corresponding to the estimated depth of the defect, and the initial value of the gate width is 0.3 μs, which can be automatically adjusted according to the material, sound speed and resolution requirement;

[0057] The C-scan frame image is generated and temporarily stored in the cache area in real time every time one scan line is completed, and then the defect length of the current defect is obtained by multiplying the number of pixels of the defect area in the C-scan frame image by the scanning pitch; and the scanning pitch of the next row is dynamically adjusted according to the current defect length:

[0058] If the defect length is less than or equal to a preset first length threshold; wherein the initial value of the preset first length threshold is set to 3 mm, the scanning pitch of the next row is maintained at the initial pitch, i.e., 1 mm;

[0059] If the defect length is greater than the preset first length threshold and less than a preset second length threshold; wherein the initial value of the preset second length threshold is set to 5 mm; the scanning pitch of the next row is linearly reduced from the initial pitch to a specified minimum scanning pitch by linear interpolation; wherein the specified minimum scanning pitch is set according to actual requirements and can be 0.2 mm;

[0060] If the defect length is greater than or equal to the preset second length threshold; the scanning pitch of the next row is the initial pitch multiplied by a preset scaling factor; the initial value of the preset scaling factor is 0.5;

[0061] Before the next row is scanned, the focusing depth table is updated according to the current defect depth interval to ensure that the sound beam is focused on the target depth range;

[0062] At each measurement point, in addition to storing the original A-scan waveform, a gate range is set according to the focusing depth of the phased array transducer; the average amplitude within the gate is calculated and mapped to a color table to obtain the color value of the C-scan column array; the change of the value is used to indicate the probability of potential defects of the next measurement point, which further guides the dynamic focusing and pitch adjustment; for example, the gate width is set to 0.3 μs, the gate starting time is the time corresponding to the estimated depth of the defect, the average amplitude within the gate is calculated and mapped to a color table to obtain the color value of the C-scan column array; the change of the value is used to indicate the probability of potential defects of the next measurement point, which further guides the dynamic focusing and pitch adjustment.

[0063] Thus, the final output of step one is: a sequence of A-scan waveforms after dynamic focusing correction. This sequence contains the spatial coordinates of the A-scan waveform, the original A-scan waveform, and the focusing parameters for each measurement point, including delay line parameters, array element aperture, and center frequency, providing complete contextual information for subsequent analysis.

[0064] Please see Figure 3 As shown, step two: perform zero-offset correction and normalization on the A-scan waveform sequence from step one, and extract the effective time window containing surface echo, defect echo, and bottom surface echo; perform wavelet packet decomposition and calculate the sub-band normalized energy; evaluate separability and sort it according to the Fisher criterion, and adaptively determine the optimal energy dimension using incremental dimension cross-validation; combine the selected optimal energy dimension features with the phase centroid, phase dispersion, and time difference features of the corresponding waveforms in the A-scan waveform sequence to form a composite feature vector set;

[0065] The original A-scan waveform obtained from the A-scan waveform sequence output in step one has problems such as noise and baseline drift. First, it is subjected to zero-offset correction and amplitude normalization. Then, typical wavelets such as Daubechiesdb8 or bior5.5 are used to denoise the waveform, preserving the main reflection information. To reduce interference from the incident wave and reverberation, only the effective time window containing surface echoes, defect echoes, and bottom surface echoes is extracted. ,in , To allow for a margin in the threshold values, the initial values ​​are all 0.2 μs, which can be automatically adjusted based on material thickness and sound velocity. This processing ensures the quality of the input data, making subsequent feature extraction more stable.

[0066] Furthermore, wavelet packet decomposition is used on the original A-scan waveform within the aforementioned effective time window to represent it on different frequency sub-bands, allowing for a more detailed analysis of the energy distribution in each frequency band. With an initial decomposition level L of 5, the following results are obtained: There are several sub-bands; for each sub-band j, where, Sub-bands representing different center frequencies; their corresponding time-domain waveforms are obtained through wavelet packet reconstruction algorithm, denoted as reconstructed sub-waveforms. , where n is the time-domain measurement point index of the sub-waveform;

[0067] And calculate the total energy of each sub-band. Sum the squares of the amplitudes of all measurement points of the corresponding reconstructed sub-waveform;

[0068] To eliminate the difference in absolute energy between sub-bands, the total energy of each sub-band is... Divide by the sum of the total energy of all sub-bands The normalized relative energy value is obtained. The sub-band energy vector formed thereby reflects the relative distribution of energy in different frequency bands;

[0069] The normalized relative energy values of all sub-bands are combined to form a sub-band energy vector in sequence ; wherein is the number of sub-bands;

[0070] On this basis, in order to objectively evaluate the contribution of different sub-bands to the classification of defect types, based on pre-prepared known defect type samples such as inclusions, delamination, holes, etc., these samples are constructed by a training data set through standard composite material test pieces or historical detection which have been verified by destructive testing to confirm the type, the mean and standard deviation of the energy of each sub-band are calculated for each type of defect type sample in the training data set; the Fisher criterion is extended to define the separability index of the sub-band, specifically, for each sub-band, the energy mean difference between all pairs of defect types is calculated, and then divided by the sum of the energy standard deviations of the two defect types, and finally the results of all pairs of defect types are added to obtain the separability index of the sub-band. The larger the separability index, the better the frequency band can distinguish different defect types; then all sub-band energy vectors are sorted in descending order according to the separability index value to form a sub-band energy vector priority list;

[0071] Based on the above sub-band energy vector priority list, the optimal energy dimension k* is further adaptively selected through incremental dimension cross-validation, and the maximum candidate dimension is set The initial value can be 10, and for each dimension k from 1 to the maximum candidate dimension , the energy vectors of the first k sub-bands in the sub-band energy vector priority list are sequentially selected as energy vector features; after combining these energy vector features with phase centroid (phase unwrapping after Hilbert transform, weight is signal amplitude), surface echo-defect echo time difference feature, surface echo-bottom echo time difference feature, etc. Fixed feature combination, a classification model (such as support vector machine or simple neural network) is used to perform N-fold cross-validation on the training set, and the validation accuracy is recorded; if the validation accuracy of the current dimension k is insufficient to improve the threshold value ε of the previous dimension k-1, which can be 0.5%, and appears twice in a row, the loop is terminated, and the current dimension k is selected as the optimal energy dimension k*; if the current dimension k does not meet the above conditions before the set maximum candidate dimension , the set maximum candidate dimension is taken as the optimal energy dimension k*; this process ensures that the selection of energy dimension comes from data, and can adaptively output the best frequency band combination for different detection tasks;

[0072] After that, the processed A-scan waveform is subjected to Hilbert transform to obtain its analytic signal, and then the instantaneous phase is obtained; define the time window containing the defect echo as W, calculate the phase centroid and phase dispersion in the time window W:

[0073] The phase centroid refers to the weighted average of the instantaneous phase in the time window W, and the weight is the signal amplitude at the corresponding time; the phase dispersion refers to the deviation of the instantaneous phase in the time window W from the phase centroid, which is obtained by multiplying the square of the deviation by the corresponding signal amplitude and then taking the square root; wherein the phase centroid reflects the average position of the echo phase, and the phase dispersion reflects the diffusion degree of the phase; the phase information is very effective for distinguishing different defect types with the same energy but different phases;

[0074] At the same time, the peak occurrence time of the surface echo, the defect echo and the bottom echo in the A-scan waveform is automatically detected, and the time difference between the surface echo and the defect echo, and the time difference between the surface echo and the bottom echo are calculated; for the defect-free measurement point, the time difference between the surface echo and the defect echo is filled with 0 or distinguished from the time difference between the surface echo and the bottom echo to avoid information loss; these time differences reflect the relative position of the echo on the time axis, which can assist the energy and phase characteristics to more accurately estimate the defect depth and type;

[0075] Next, the selected k* optimal energy dimension features are combined with the phase centroid, the phase dispersion, and the above two time difference features to form a composite feature vector set; wherein the vector dimension is k*+4, and if only one time difference is used, the vector dimension is k*+3; the vector dimension is determined by the algorithm and written into the system parameter table to maintain the consistency of subsequent model input; after training, if it is found that some energy bands or phase features have limited contribution to classification, they can be further removed or replaced according to the verification results.

[0076] Finally, the outputs of this step include: the optimal energy dimension k*, the composite feature vector set, and the sub-band separability indicators for evaluation; these outputs are used for the classification model training and online identification in step three; through the above process, the adaptive selection of energy dimension and the fusion of phase and time difference into composite features can effectively amplify the subtle differences between different defects, laying a solid foundation for solving the echo waveform similarity problem.

[0077] Step three: based on the composite feature vector of step two, a lightweight deep network of energy-time difference channel and phase channel is constructed, and after training / validation / testing, the defect category and classification confidence of each measurement point are output in real time; according to high / low threshold, reliable, low confidence or unknown can be determined, and unknown samples are put into the small sample pool; the pool is full to trigger active learning, merge samples to recalculate the optimal energy dimension and fine-tune the network; the results are superimposed on the C-scan frame image for two-dimensional visualization, and the confidence distribution is fed back to step one to adjust the interval and frequency adaptively;

[0078] With the composite feature vector constructed in step two, real-time defect classification is performed for each scanning measurement point, while relying on the deep network structure and online learning strategy to continuously improve the recognition accuracy, and combining the classification results with the C-scan frame image, both two-dimensional plane visualization and accurate plane positioning for subsequent depth calculation are achieved;

[0079] The training data set constructed in step two is called, which contains pre-prepared known defect type samples and has generated corresponding composite feature vectors by step two processing. The composite feature vectors of these samples are bound with their known defect types to form (composite feature vector, defect type) sample pairs, which are divided into training set, validation set and test set in the ratio of, for example, 7:1:2, wherein the training set is used for model training, the validation set is used for monitoring the training process, and the test set is used for evaluating the final performance; the defect types include, for example, no defect, polytetrafluoroethylene inclusion, CFRP sheet inclusion, plastic paper inclusion, etc.

[0080] Based on the composite feature vector, a classification model is constructed, which adopts a dual-channel lightweight deep network structure and takes the composite feature vector as input, and outputs the defect category and classification confidence, including:

[0081] The energy-time difference channel input dimension is the optimal energy dimension k* plus the time difference feature quantity, which depends on the actual feature quantity, and is extracted through two layers of 1D convolution; wherein the convolution kernel number Nc1, Nc2, kernel length Lc1, Lc2 are parameters, ReLU activation and pooling layer extract local patterns, and output abstract feature vector ;

[0082] The phase channel inputs low-dimensional features such as phase centroid and dispersion, which are mapped to abstract feature vector of the same dimension through one to two layers of fully connected network ;

[0083] The fusion layer adds abstract feature vector and abstract feature vector element by element, and obtains the output vector through two layers of fully connected network, and finally outputs the category probability vector ; wherein i represents the defect category index in the defect category task, which is used to distinguish different defect categories; each defect category corresponds to a category probability vector;

[0084] The defect category corresponding to the maximum category probability vector is selected as the defect category determination result , and the category probability vector is recorded as the confidence, which is called the classification confidence;

[0085] and compare the classification confidence with a preset high confidence threshold and a preset low confidence threshold; wherein the preset high confidence threshold can be 0.9 and the preset low confidence threshold can be 0.5;

[0086] When the classification confidence is greater than the preset high confidence threshold, it is considered that the classification result is reliable;

[0087] When the classification confidence is between the preset low confidence threshold and the preset high confidence threshold, it is marked as low classification confidence;

[0088] When the classification confidence is less than the preset low confidence threshold, it is marked as an unknown sample and stored in a small sample pool for subsequent learning;

[0089] It should be noted that the number of layers, the size of the convolution kernel, and the number of fully connected neurons of the network are all set as adjustable initial values in the parameter table. This structure can simultaneously process multi-dimensional energy / time difference information and a small amount of phase information, balancing model complexity and real-time performance;

[0090] After classification, according to the gate setting of step one, the amplitudes within the gate corresponding to the A-scan waveform are normalized, mapped to pixel values according to the color table, and these pixel values are arranged in the order of the measurement points to form a C-scan column array of the current scan line. The new C-scan frame image is generated by vertically splicing the C-scan column array of the previous frame, and the defect classification result is superimposed on the image , the coordinates corresponding to the classification confidence and the defect classification result, and different colors or symbols are used to distinguish different types of defects, so that the operator can directly see the location, range, and type of the defect on the plane;

[0091] When low classification confidence or unknown samples are detected, their corresponding abstract feature vectors, A-scan waveforms, and tentative defect classification results are placed in the small sample pool. The administrator or expert can periodically review the small sample pool and manually label the unknown samples;

[0092] Whenever the small sample pool accumulates a certain number of samples, for example, 50 samples, the active learning process is automatically triggered: the small sample pool is combined with the original training set in step two, the corresponding composite feature vector is generated for the new sample, the composite feature vector is updated, and the manually labeled label is bound, that is, the defect classification result is updated. The energy dimension analysis of step two is re-executed to update the optimal energy dimension k* and the abstract feature vector. If the new sample introduces a new frequency mode, adjust the input layer size of the network according to the new optimal energy dimension, retrain or fine-tune using the merged (composite feature vector, defect type) sample pair, test the new model on the validation set, and if the performance improves, replace the old model. If the performance decreases, it can be rolled back. Active learning ensures that the model continuously adapts to new defect types or material changes, improving the long-term stability and generalization ability of the system;

[0093] The classification result and the confidence are also used to feedback the scanning strategy of step one. For example, when a region with unknown samples appears in a plurality of continuous lines of scanning, the scanning interval can be automatically reduced, the excitation frequency can be increased, and more detailed scanning can be performed in the region. For a region without defects or with a high recognition confidence, the interval can be appropriately increased to improve the detection efficiency. In addition, the classification result can also be used as early warning information. If a certain type of defect (such as delamination) frequently appears, the system can prompt the maintenance personnel to check the manufacturing process or the material batch.

[0094] It should be noted that the network is trained using a cross-entropy loss function and an optimization algorithm such as Adam. Batch normalization, dropout, and other techniques are used to prevent overfitting during the training process. The model performance is monitored using a validation set, and an early stopping strategy is used to stop training when the validation loss no longer decreases. For imbalanced sample cases, class weights or data augmentation strategies can be used to balance the influence of each type of sample during training. After training is complete, the recognition accuracy, recall rate, and other indicators are evaluated on the test set, and the best weights are saved. The above training process is a mature technology, and the specific implementation process will not be described again.

[0095] Step four: For each new C-scan frame image, the center coordinates are taken, the surface echo occurrence time, the defect echo occurrence time, and the bottom echo occurrence time are extracted by backtracking the A-scan waveform, the path length is decomposed according to the sound wave propagation path, the defect center depth is calculated by combining the water / material sound speed, the three-dimensional point set is formed, the defects are clustered by DBSCAN according to the defect type, the volume is calculated by voxel integration after α shape reconstruction, and the X-Y maximum radial size and the remaining wall thickness are calculated. The parameters are compared with the defect volume upper limit and the remaining wall thickness lower limit of the material performance database to output qualified, repair, and unqualified conclusions, and the coordinates and threshold adjustment instructions are fed back to step one to realize a closed loop.

[0096] For each new C-scan frame image in step three, the center coordinates are found, and the surface echo occurrence time, the defect echo occurrence time, and the bottom echo occurrence time are extracted by backtracking the original A-scan waveform recorded in step one.

[0097] The water medium sound speed is about 1480 m / s at 25°C, and the propagation time is real-time calibrated with the water tank temperature and the defect echo occurrence time minus the surface echo occurrence time multiplied by the water path ratio; the sound wave propagation path is split for calculation:

[0098] The distance of sound wave propagation in water is the product of water medium sound speed and propagation time;

[0099] The distance of sound wave propagation in the tested composite specimen is the product of the tested composite longitudinal wave sound speed and the propagation time;

[0100] And half of the sum of the distance of sound wave propagation in water and the distance of sound wave propagation in the tested composite specimen is taken as the defect center depth;

[0101] Based on the defect depth estimation value obtained in step one, combine its two-dimensional coordinates (X, Y) with the calculated defect center depth to form a three-dimensional point set. For the point set belonging to the same defect type (step three classification result)

[0102] At the same time, optimize the depth gate setting according to the defect depth range: set the gate width to 1.5 times the defect echo pulse width, ensure that the complete echo is included and no extra noise is introduced, and the gate starting time corresponds to the defect depth estimation time, which is consistent with the gate setting logic in step one;

[0103] Combine the two-dimensional coordinates (X, Y) of each measurement point with the calculated depth value to form a three-dimensional point set Pᵢ=(Xᵢ,Yᵢ,Zᵢ); For the three-dimensional point set belonging to the same defect type, use spatial clustering algorithm such as DBSCAN to segment independent defect area, and then perform three-dimensional surface reconstruction through α shape reconstruction, where α value is the average distance of point cloud multiplied by 1.5; The reconstruction parameters such as radius and α value are automatically adjusted according to the point cloud density; After reconstruction, calculate the defect volume by voxel integration summation, and calculate the maximum radial size in X-Y plane by calculating the maximum Euclidean distance between any two points in the region and the remaining wall thickness obtained by subtracting the defect center depth and the maximum defect depth from the material thickness; Where the maximum defect depth is the maximum value of the depth in the three-dimensional point set of the same defect type;

[0104] For complex defects, additional shape features such as ellipticity and flatness are calculated as supplementary parameters for strength evaluation;

[0105] Input the above three-dimensional parameters, i.e. defect volume, maximum radial size in X-Y plane, and remaining wall thickness, into the material performance database, and compare the defect volume and remaining wall thickness with the set upper limit of defect volume and lower limit of remaining wall thickness to generate defect detection conclusion. The upper limit of defect volume and the lower limit of remaining wall thickness are adjusted according to the material category and working condition; For example, the lower limit of remaining wall thickness for aviation composite materials is set to 60% of the design thickness, and for automotive materials, it is set to 50%. According to the comparison results of defect volume and remaining wall thickness with the upper limit of defect volume and the lower limit of remaining wall thickness, the evaluation conclusion is given:

[0106] It should be noted that the material performance database at least includes {material brand, design thickness, working condition level, working condition level, upper limit of defect volume, lower limit of remaining wall thickness};

[0107] Obtain the upper limit of defect volume and the lower limit of remaining wall thickness with {the same brand and the same working condition level} as the retrieval condition; And compare them with the defect volume and the remaining wall thickness;

[0108] When the defect volume is not greater than the upper limit of defect volume and the remaining wall thickness is not less than the lower limit of remaining wall thickness, it is recorded as safe and acceptable;

[0109] When the defect volume after rounding is the same as the upper limit of the defect volume, or the remaining wall thickness after rounding is the same as the lower limit of the remaining wall thickness, it is recorded as needing repair;

[0110] When the defect volume is greater than the upper limit of the defect volume, or the remaining wall thickness is less than the lower limit of the remaining wall thickness, it is recorded as unqualified;

[0111] The defect detection conclusion is fed back to the previous step to form a closed loop: the scanning interval of the unknown sample dense area is reduced to 0.5mm, the center frequency is increased to 7MHz, and the interval of the high confidence area is increased to 2mm; for the defect position corresponding to the safety acceptable, the frequency and position of its occurrence are recorded, and the manufacturing execution system is fed back to the production line to prompt the adjustment of the corresponding process, such as strengthening the cleaning control of the layer where inclusions frequently occur; the mold temperature is checked in the layered concentrated area.

[0112] In this embodiment, the A-scan waveform of the test piece is collected, the potential defect area is determined based on the echo amplitude and time difference, and the A-scan waveform sequence containing spatial coordinates, original waveform and measurement point focusing parameters is output; after waveform preprocessing, the effective time window is intercepted, a plurality of sub-bands are generated through wavelet packet decomposition, the total energy of each sub-band is calculated and normalized to form a sub-band energy vector; based on the known defect sample, the Fisher criterion is used to evaluate the sub-band separability index and sort; the optimal energy dimension is adaptively determined through dimension cross-validation, the selected sub-band energy features are combined with phase centroid, phase dispersion and time difference features to form a composite feature vector; the double-channel lightweight deep network classification model is input, and the defect category and confidence are output and mapped to the C-scan frame image; based on the image backtracking three-dimensional coordinates, the defect volume and the remaining wall thickness are calculated by combining the sound velocity, and the detection conclusion is output by comparing the material performance database; the present application solves the problem of defect misjudgment caused by echo waveform similarity, and realizes detection closed loop optimization.

[0113] The above formulas are dimensionless numerical calculations, the formulas are obtained by software simulation of a large amount of data to obtain a formula of the nearest true situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0114] The above embodiments can be realized wholly or partially by software, hardware, firmware or any other combination. When realized by software, the above embodiments can be realized in the form of a computer program product wholly or partially.

[0115] Those skilled in the art can understand that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software manner depends on the specific application of the technical solution and the constraints of the invention. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0116] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0117] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0118] Finally: the above is only the preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A non-destructive testing method for defects in composite materials, characterized in that, The process includes the following steps: In a constant temperature water bath, a phased array transducer is used to scan the composite material specimen along the X–Y path to obtain the A-scan waveform at each measuring point. Based on the echo amplitude and echo time difference between various echoes in the A-scan waveform, potential defect areas are determined, and an A-scan waveform sequence containing spatial coordinates, the original A-scan waveform, and the focusing parameters of the measuring points is output. The A-scan waveform is a time-domain echo signal acquired by scanning with a phased array transducer. The focusing parameters of the measuring points include delay line parameters, array element aperture, and center frequency. After preprocessing the original A-scan waveform, an effective time window containing surface echo, defect echo, and bottom surface echo is extracted. Multiple sub-bands are obtained through wavelet packet decomposition, and the total energy of each sub-band is calculated and normalized to form the corresponding sub-band energy vector. Based on samples with known defect types, the Fisher criterion is used to evaluate the separability index of each sub-band, and the energy vectors of each sub-band are sorted in descending order of separability index to form a priority list of energy vectors for each sub-band. An incremental dimensional cross-validation method is adopted to adaptively determine the optimal energy dimension. The sub-band energy vector features corresponding to the selected optimal energy dimension are combined with the phase centroid, phase dispersion, and time difference features to form a composite feature vector. Based on the obtained composite feature vector, the classification model is input to output the defect category and classification confidence. The output results are then mapped to a C-scan frame image with spatial coordinates as the dimension. Based on the C-scan frame image, the three-dimensional coordinates of the defect are traced back, and the defect volume and remaining wall thickness are calculated by combining the sound velocity. The defect detection conclusion is then compared with the material property database. The determination of potential defect areas includes: When the phased array transducer completes a scan line along the X direction with the initial scan interval, the A-scan waveform of each measurement point is analyzed in real time. If any of the following conditions are met, it is marked as a potential defect region and the phased array transducer fine scan is triggered: When the absolute deviation between the surface echo amplitude or bottom echo amplitude and the average surface echo amplitude exceeds the amplitude anomaly threshold, it is marked as a potential defect area and a fine scan of the phased array transducer is triggered. Or, if the peak value of the A-scan waveform exceeds the defect echo judgment threshold between the time when the surface echo appears and the time when the bottom echo appears, it is marked as a potential defect area and the phased array transducer fine scan is triggered. The phased array transducer fine scanning includes: calculating the defect depth estimate based on the defect depth prediction formula, as follows: In the formula, d is the difference between the time when the bottom echo appears and the time when the surface echo appears, and d is the longitudinal wave velocity of the composite material being tested; the center frequency is increased to the initial value of the center frequency plus a preset increment, and the number of array elements of the phased array transducer is halved.

2. The non-destructive testing method for defects in composite materials according to claim 1, characterized in that, When scanning via a phased array transducer: an N-element phased array transducer is used, where N is the number of elements, the element spacing is P, and the center frequency is... .

3. The non-destructive testing method for defects in composite materials according to claim 1, characterized in that, Wavelet packet decomposition includes: denoising the waveform using typical wavelets while preserving the main reflection information; and extracting only the effective time window containing surface echoes, defect echoes, and bottom echoes. ,in , To reserve a margin threshold, the original A-scan waveform within the above effective time window is decomposed using wavelet packet decomposition; the decomposition level L is set to obtain... There are several sub-bands; for each sub-band j, where, Sub-bands representing different center frequencies; their corresponding time-domain waveforms are obtained through wavelet packet reconstruction algorithm, denoted as reconstructed sub-waveforms. , where n is the time-domain measurement point index of the sub-waveform; And calculate the total energy of each sub-band. Sum the squares of the amplitudes at all measurement points corresponding to the reconstructed sub-waveform; sum the total energy of each sub-band. Divide by the sum of the total energy of all sub-bands The normalized relative energy value is obtained. ; and normalized relative energy values ​​for all sub-bands. Arranged in order, forming sub-band energy vectors ;in This represents the number of sub-bands.

4. The non-destructive testing method for defects in composite materials according to claim 1, characterized in that, When evaluating the separability index of a sub-band: Based on pre-prepared samples of known defect types, the mean and standard deviation of the energy of each sub-band are calculated; the Fisher criterion is used to define the separability index of the sub-band, specifically: for each sub-band, the mean energy difference between all pairs of defect types is calculated, then divided by the sum of the standard deviations of the energy of these two defect types, and finally these results for all pairs of defect types are added together to obtain the separability index of the sub-band; subsequently, all sub-band energy vectors are sorted from largest to smallest according to the separability index value to form a priority list of sub-band energy vectors.

5. The non-destructive testing method for defects in composite materials according to claim 4, characterized in that, Adaptively determining the optimal energy dimension includes: setting the maximum candidate dimension. For dimensions from 1 to the maximum candidate dimension For each dimension k, the energy vectors of the first k sub-bands in the sub-band energy vector priority list are selected sequentially as energy vector features. When the k-dimensional energy vector features undergo N-fold cross-validation, the validation accuracy is recorded. If the validation accuracy of the current dimension k is less than the threshold ε compared to the previous dimension k-1 and this occurs twice consecutively, the loop terminates, and the current dimension k is selected as the optimal energy dimension k*. If the current dimension k reaches the set maximum candidate dimension... If the above conditions are not met, then the maximum candidate dimension will be selected. k* is the optimal energy dimension.

6. The non-destructive testing method for defects in composite materials according to claim 1, characterized in that, The classification model employs a dual-channel lightweight deep network structure, taking composite feature vectors as input and outputting defect categories and classification confidence scores, including: The energy-time difference channel takes the optimal energy dimension k* plus time difference features as input, and outputs an abstract feature vector through two layers of 1D convolution. Among them, the time difference characteristics include the time difference between surface echo and defect echo, and the time difference between surface echo and bottom echo; The phase channel inputs the phase centroid and phase dispersion, which are mapped to an abstract feature vector through a two-layer fully connected network. Same-dimensional abstract feature vector Here, the phase centroid refers to the weighted average of the instantaneous phase within the time window containing the defect echo, with the weight being the signal amplitude at the corresponding moment; the phase dispersion refers to the deviation between the instantaneous phase and the phase centroid within the time window containing the defect echo, which is obtained by multiplying the squares of these deviations by the corresponding signal amplitude, taking the average, and then taking the square root. The fusion layer will abstract the feature vectors and abstract feature vectors The elements are added one by one, and the output vector is obtained through two fully connected layers. Finally, the output class probability vector is activated by Softmax. Each defect class corresponds to a class probability vector. The defect class corresponding to the largest class probability vector is selected as the defect class determination result, and the class probability vector is recorded as the confidence level, which is called the classification confidence level.

7. The non-destructive testing method for defects in composite materials according to claim 1, characterized in that, After the phased array transducer is finely scanned, the process also includes: after a scan line is acquired, mapping the amplitude of the A-scan waveform of all measurement points of the scan line at the same depth gate to the pixel color, stacking them along the scanning direction according to the pixel color to form a C-scan array, and stitching them together to form a C-scan frame image; wherein, the depth gate includes the gate start time and the gate width, and the gate start time is the time corresponding to the estimated defect depth.

8. The non-destructive testing method for defects in composite materials according to claim 7, characterized in that, The defect detection conclusion generation includes: based on the defect category classification results, regenerating the C-scan image to obtain a new superimposed defect category determination result. The coordinates of the C-scan image corresponding to the classification confidence and defect category determination results are obtained. Based on the new C-scan image, the A-scan waveform is traced back to calculate the defect center depth, find its center coordinates, trace back the original A-scan waveform, and extract the time of appearance of surface echo, defect echo, and bottom surface echo. Combining the sound velocity and propagation time in the water medium, the propagation time is obtained by subtracting the surface echo occurrence time from the defect echo occurrence time and multiplying by the proportion of the propagation path in the water; the calculation is performed by breaking down the sound wave propagation path: The distance a sound wave travels in water is the product of the speed of sound in the water and the propagation time. The propagation distance of sound waves in the tested composite material specimen is the product of the longitudinal wave velocity of the tested composite material and the propagation time; The defect center depth is defined as half the sum of the distance the sound wave travels in water and the distance it travels in the tested composite material specimen. The estimated defect depth and its corresponding two-dimensional coordinates are combined with the calculated defect center depth to form a three-dimensional point set. For three-dimensional point sets belonging to the same defect type, a spatial clustering algorithm is used for clustering, followed by three-dimensional surface reconstruction. After reconstruction, the defect volume is calculated by summing voxel integrals, and the maximum Euclidean distance between any two points within the region is calculated to obtain the maximum radial dimension in the XY plane. The remaining wall thickness is obtained by subtracting the defect center depth and the maximum defect depth from the material thickness. The maximum defect depth is the maximum depth value in the three-dimensional point set of the same defect type. These three-dimensional parameters—defect volume, maximum radial dimension in the XY plane, and remaining wall thickness—are input into a material performance database. The defect volume and remaining wall thickness are compared with the set upper limit for defect volume and lower limit for remaining wall thickness to generate a defect detection conclusion. When the defect volume is not greater than the upper limit and the remaining wall thickness is not less than the lower limit, it is considered safe and acceptable. If the rounded defect volume is the same as the upper limit of the defect volume, or the rounded remaining wall thickness is the same as the lower limit of the remaining wall thickness, it is recorded as needing repair. If the defect volume is greater than the upper limit of the defect volume, or the remaining wall thickness is less than the lower limit of the remaining wall thickness, it is considered unqualified.

Images