Method and system for detecting internal defects of composite material part

Abstract

The invention relates to the technical field of composite material detection, and particularly discloses a composite material workpiece internal defect detection method and system.The method comprises the steps that by obtaining defect detection features and calculating the feature overlapping degree, after modal feature degradation is judged, a hierarchical decomposition model is constructed to separate a thermal stress coupling signal, and a defect detection result is obtained; gradient damp and hot loads are applied to an area needing difference analysis to generate a dynamic response matrix, a sudden change direction angle is extracted through time-frequency coupling analysis, anisotropic interference is filtered, a layered hidden boundary and a boundary hidden area are determined based on the sudden change direction angle, and after geometric parameters are extracted through imaging, a model is optimized through an error correction matrix; the system comprises an overlapping extraction module, a confusion analysis module, a difference analysis module, a mutation recognition module and a feedback correction module, the problems of boundary fuzziness and small defect missing detection in traditional detection can be reduced, defect positioning and closed-loop optimization under multi-field coupling are achieved, and the composite material quality control efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of composite material detection, and in particular to a method and system for detecting internal defects of composite material parts. Background Art

[0002] Carbon fiber-epoxy resin laminate composites are widely used in high-end fields such as aerospace and rail transportation due to their excellent properties, such as high specific strength and fatigue resistance. However, during the composite material preparation and service process, the presence of internal defects can significantly reduce the mechanical properties of the component. Therefore, detecting internal defects is key to ensuring the reliability of composite components.

[0003] When existing technologies capture the characteristics of different types of defects in a single mode, it is easy to cause misjudgment of defects in areas of overlapping features; the anisotropic intrinsic properties of composite materials will interfere with the detection signal, resulting in blurred positioning of defect boundaries; traditional static detection cannot simulate the impact of moisture-heat-stress coupling loads on defect evolution in the service environment, and it is difficult to identify potential defect extension under dynamic loads.

[0004] To this end, the present invention provides a method and system for detecting internal defects of composite material parts. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for detecting internal defects of composite material parts to solve the above-mentioned background problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions: A method for detecting internal defects of a composite material part comprises the following steps: In the historical layer micro-defect inspection of carbon fiber and epoxy resin layer composite materials, the defect detection features are obtained and overlap detection is performed to obtain the feature overlap of historical layer micro-defects; Based on the feature overlap, determine whether modal feature degradation occurs in different defects. If so, construct a hierarchical decomposition model to separate signal confusion caused by thermal stress coupling, perform confusion evaluation on defect detection features, obtain confusion improvement rate, and determine whether differential analysis is required. If difference analysis is required, the diagnostic test interface is divided and a gradient moist heat load is applied to analyze whether the difference in defect detection characteristics of the diagnostic test interface is significant. If significant, a dynamic response matrix under dynamic load is generated; Based on the dynamic response matrix, time-frequency coupling analysis is performed to identify the strain gradient in the delaminated fiber direction, extract the sudden direction angle, and identify the interference of the composite material anisotropy on the sudden direction angle. If interference exists, the defective area is filtered.

[0007] As a further technical solution of the present invention: the overlap detection is performed as follows: Obtain overlapping detection vectors, map the overlapping detection vectors into three-dimensional space, and form characteristic distribution areas of the three types of defect detection features; The intersection and union of the distribution areas of different types of defect detection features are calculated, and the overlap is calculated to obtain the feature overlap.

[0008] As a further technical solution of the present invention: the method of obtaining the overlap detection vector is: Obtaining the delamination defect detection features, the pore defect detection features, and the debonding defect detection features of the defect area, and using the three defect detection features of the defect area as defect detection features; The defect area is simplified into sampling points, and an overlapping detection vector containing the defect detection features of each sampling point is constructed.

[0009] As a further technical solution of the present invention: the method of performing confusion evaluation on the defect detection feature is: If modal feature degradation occurs, the three types of defect detection features of the overlapping detection vectors of all historical sampling points and the historical P-level thermal stress load are obtained for correlation coupling analysis to determine whether there is correlation coupling between the three types of defect detection features and the P-level thermal stress load; If there is relevant coupling, construct a third-order tensor and perform low-rank constraint decomposition on the third-order tensor, and construct the defect classification confusion matrix before and after the low-rank constraint decomposition; The confusion improvement rate is calculated based on the defect classification confusion matrix. If the confusion improvement rate is lower than the preset confusion improvement threshold, difference analysis is required.

[0010] As a further technical solution of the present invention: the judgment method of the occurrence of the modal feature degradation is: Construct defect pairs of any two sets of defect detection features, such as delamination defect detection features, porosity defect detection features, and debonding defect detection features; Obtain the coordinates of the two sets of defect sampling points corresponding to the defect pair, calculate the Euclidean distance between the two sets of defect sampling point coordinates, and classify the defect pairs into spatially adjacent pairs and spatially unrelated pairs; Calculate the probability of spatial proximity pairs appearing in the defect detection area, calculate the spatial proximity entropy of spatial proximity pairs using the information entropy algorithm, and divide the feature overlap into fixed overlap and random overlap; Obtain the characteristic overlap of the current level micro-defect inspection, and establish a state characteristic degradation equation based on the characteristic overlap of the current carbon fiber and epoxy resin layer composite material, the fixed overlap mean of all level micro-defects, and the random overlap mean of all level micro-defects; If the modal characteristic degradation equation is satisfied, modal characteristic degradation occurs.

[0011] As a further technical solution of the present invention: the method for analyzing whether the difference in the defect detection characteristics of the diagnostic detection interface is significant is: Divide the defect detection area where modal feature degradation occurs into a diagnostic detection interface; Apply a gradient moist heat load to the defect detection area to obtain the characteristic sensitivity of different types of defect detection features in the defect detection area. If the defect detection features of the diagnostic detection interface are significantly different, the characteristic sensitivity of different types of defects is ranked. If the characteristic sensitivity of the delamination defect is the highest, the diagnostic detection interface is divided into regions again according to the characteristic sensitivity, the number of divided interface partitions is obtained, and a dynamic response matrix is ​​constructed.

[0012] As a further technical solution of the present invention: the method of filtering the defective area is: Obtain the strain gradient and characteristic frequency band, map the strain gradient value of the characteristic frequency band to the spatial position, and extract the mutation point of the strain gradient; Obtain the angle between the main direction of the strain gradient point and the fiber direction to obtain the sudden change direction angle; Decompose the strain gradient into components in different directions, calculate the frequency domain energy proportion of the strain gradient components in each direction at the characteristic frequency, and calculate the anisotropy correlation factor; Based on the anisotropic correlation factor, the interference mutation direction angle is filtered out, and the defect area to which the delamination defect corresponding to the interference belongs is filtered out.

[0013] As a further technical solution of the present invention, the following steps are also included: the strain gradient is obtained by: The time-frequency domain is converted based on the dynamic response matrix, and the time-frequency matrix is ​​constructed. The spatial gradient of the time-frequency matrix is ​​calculated to obtain the spatial derivative in the fiber direction as the strain gradient.

[0014] As a further technical solution of the present invention, the following steps are also included: The hidden boundary of the delamination area is determined based on the mutation direction angle, and the boundary hidden area is determined. The boundary hidden area is re-imaged, and the geometric parameters of the hidden detection features are extracted. After comparison with the detection model output, an error correction matrix is ​​generated and fed back to the hierarchical decomposition model. The corrected defect identification results are output and embedded in the composite material quality control system.

[0015] A composite material part internal defect detection system includes the following modules: Overlap extraction module: used to obtain defect detection features and perform overlap detection in the historical layer micro-defect inspection of carbon fiber and epoxy resin layer composite materials, and obtain the feature overlap of historical layer micro-defects; Confusion Analysis Module: Based on feature overlap, it determines whether modal feature degradation occurs in different defects. If so, a hierarchical decomposition model is constructed to separate signal confusion caused by thermal stress coupling. Confusion evaluation is performed on defect detection features to obtain the confusion improvement rate and determine whether differential analysis is required. Difference analysis module: If difference analysis is required, the diagnostic test interface is divided and a gradient moist heat load is applied to analyze whether the difference in defect detection characteristics of the diagnostic test interface is significant. If significant, a dynamic response matrix under dynamic load is generated; Mutation identification module: Based on the dynamic response matrix, it is used to perform time-frequency coupling analysis to identify the strain gradient in the fiber direction of the layer, extract the mutation direction angle and identify the interference of the anisotropy of the composite material on the mutation direction angle. If interference exists, the defective area is filtered.

[0016] Beneficial effects of the present invention: By fusing three types of defect detection features, namely ultrasonic echo amplitude attenuation rate, thermal diffusion time constant and grayscale threshold porosity, and standardizing them to construct overlapping detection vectors and perform three-dimensional spatial mapping, it is beneficial to achieve multimodal feature cross-validation of delamination, debonding and porosity defects, reduce the one-sidedness of a single detection method, improve the quantitative fusion accuracy of complex defect detection features, and provide a visual spatial distribution basis for subsequent modal degradation judgment; based on spatial proximity entropy, fixed overlap and random overlap are distinguished, modal degradation is determined by threshold, and the third-order tensor low-rank constraint decomposition model is used to separate thermal stress coupling interference, which is beneficial to diagnosing feature confusion of different defects, reducing misjudgment caused by signal aliasing, and effectively alleviating the interference of thermal stress on defect detection features, thereby improving the reliability of classification and recognition of multiple defect types.

[0017] Analyzing the dynamic response differences of defect detection characteristics through characteristic sensitivity calculation is conducive to capturing the evolution characteristics of defects under multi-field coupling, solving the problem that traditional static detection cannot reflect the dynamic expansion of defects, and providing data support for defect evolution path analysis and multi-field response matrix construction; through time-frequency coupling analysis, the dynamic response matrix is ​​wavelet transformed, the strain gradient spatial derivative is calculated along the fiber direction, and the anisotropic correlation factor is used to filter the interference of the intrinsic properties of the composite material, which is conducive to locating the strain gradient mutation point at the edge of the delamination defect, suppressing the false signal interference caused by fiber laying, and improving the authenticity and accuracy of the hidden boundary positioning of the defect.

[0018] Ultrasonic phased array and infrared thermal imaging are used to re-image the hidden boundary areas, extract the defect geometric parameters and compare them with the model prediction values ​​to generate an error correction matrix. This is beneficial for capturing tiny defects and fuzzy boundary areas that are missed by traditional detection. Error feedback is used to achieve iterative optimization of the hierarchical decomposition model, thereby improving the detection system's adaptability and recognition reliability for different materials and defect types. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The present invention will be further described below with reference to the accompanying drawings.

[0020] Figure 1 This is a flow chart of a method for detecting internal defects of a composite material part according to the present invention; Figure 2 Is a flow chart showing whether the present invention requires differential analysis; Figure 3 This is a module diagram of a composite material part internal defect detection system of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0022] Example 1:

[0023] See also Figure 1 As shown, the present invention is a method for detecting internal defects of composite material parts, comprising the following steps: S1. In the inspection of historical layer micro-defects of carbon fiber and epoxy resin layer composite materials, defect detection features are obtained and overlap detection is performed to obtain the feature overlap of historical layer micro-defects; Among them, the method of obtaining the defect detection features of delamination, debonding, and porosity is: In the historical layer-level micro-defect inspection of the carbon fiber and epoxy resin layer composite material, the micro-defect inspection is divided into different defect detection areas to obtain the defect area and defect-free area of ​​the carbon fiber and epoxy resin layer composite material; It can be understood that in the historical layered micro-defect inspection, the ultrasonic echo amplitude of the defective area is obtained by ultrasonic testing, compared with the amplitude of the defect-free area, and the percentage of the difference to the baseline value is calculated to obtain the ultrasonic echo amplitude attenuation rate, which is used as the detection feature of the layered defect; Using infrared thermal imaging technology, the time required for the surface temperature of the defect area to decay from the peak value to the preset decay value 1-1 / e is recorded. The thermal diffusion time constant is obtained by exponentially fitting the thermal decay curve, which is used as the detection feature of the debonding defect. With the help of X-ray CT imaging, the grayscale threshold segmentation method is used to extract the defect area from the CT image, and the percentage of the pore volume in the detection volume is calculated to obtain the CT grayscale threshold porosity, which is used as the pore defect detection feature; Based on the three types of defect detection features, if the detection feature is higher than the preset corresponding detection feature threshold, a comparison is made to determine the defect boundary where the detection feature is higher than the preset corresponding detection feature threshold, thereby determining the defect area and the corresponding non-defect area; The defects contained in the defective area may be delamination defects, pore defects, and debonding defects; It is understandable that a defect detection area may contain multiple delamination defect areas, pore defect areas, and debonding defect areas; Delamination defect detection features, pore defect detection features, and debonding defect detection features are used as defect detection features; Among them, the method of performing overlap detection on defect detection features and obtaining the feature overlap of historical level micro defects is as follows: The defect area is simplified into sampling points, and the defect detection features of delamination, debonding, and porosity at each sampling point are normalized respectively, and an overlapping detection vector containing the defect detection features of delamination, debonding, and porosity at each sampling point is constructed; Map the overlapping detection vectors into three-dimensional space to form the characteristic distribution areas V of the three types of defect detection features; The overlap degree is calculated based on the overlap detection vector to obtain the feature overlap degree of the historical level; Preferably, by formula Obtain the feature overlap O of the defect detection area; in, represents the intersection of the characteristic distribution areas of the i-th defect and the j-th defect, Represents the union of the characteristic distribution areas of the i-th defect and the j-th defect; Vo represents the volume of the intersection or union of feature distribution areas.

[0024] It can be understood that feature overlap quantifies the degree of aliasing of different types of defect detection features in spatial distribution, reflecting the risk of confusion of defect detection features due to intrinsic material properties or external load coupling, and provides a basis for multi-defect classification and detection model optimization: Among them, the function of calculating feature overlap is: Function 1: Improve the accuracy and reliability of defect detection. By fusing three types of features, namely ultrasonic echo amplitude attenuation rate, thermal diffusion time constant, and CT grayscale threshold porosity, and standardizing them, an overlapping detection vector is constructed and mapped to three-dimensional space. This realizes multi-modal feature cross-validation of delamination, debonding, and porosity defects, reduces the one-sidedness of single-modal detection and misjudgment caused by signal aliasing, and improves the quantitative fusion accuracy of complex defect detection features.

[0025] Function 2: Dynamic load analysis captures defect evolution characteristics: Applying gradient moisture and heat loads to the feature confusion area, through feature sensitivity calculation and dynamic response matrix construction, is conducive to identifying the dynamic response differences of defects under multi-field coupling, and reducing the problem that traditional static detection cannot reflect defect expansion.

[0026] S2. Based on the feature overlap, determine whether modal feature degradation occurs for different defects. If so, construct a hierarchical decomposition model to separate signal confusion caused by thermal stress coupling, perform confusion evaluation on the defect detection features, obtain the confusion improvement rate, and determine whether differential analysis is required. Among them, the method of judging whether modal feature degradation occurs in different defects is: Defect spatial correlation analysis of carbon fiber and epoxy resin layer composites to distinguish fixed overlap and random overlap ; Preferably, the method of performing defect space correlation analysis on the carbon fiber and epoxy resin layer composite material is: Construct defect pairs of any two sets of defect detection features, such as delamination defect detection features, pore defect detection features, and debonding defect detection features, from the defect areas in the historical level micro-defect inspection; Obtain the coordinates of the two sets of sampling points corresponding to the defect pair, and calculate the Euclidean distance between the two sets of defect sampling point coordinates; Compare the Euclidean distance between the two groups of defect sampling point coordinates with the preset critical distance for each defect pair, and classify the defect pairs into spatially adjacent pairs and spatially unrelated pairs; For example, in the inspection of carbon fiber and epoxy resin layer composite materials, delamination and debonding defect pairs are constructed, and two sets of defect sampling point coordinates are obtained (for example, the coordinates of delamination defect point A are (2, 3) and the coordinates of debonding defect point B are (6, 6)). The Euclidean distance between point A and point B is calculated to be 5 mm. If the preset critical distance is 5 mm, the distance is equal to the critical value and is determined to be a spatially adjacent pair. Then, the delamination defect point C (1, 1) and the pore defect point D (8, 9) are taken, and the calculated distance is 10.63 mm. Because it exceeds the critical value, it is determined to be a spatially unrelated pair. The Euclidean distance can be used to quickly distinguish the spatial correlation of the defect pairs, providing a basis for subsequent modal degradation analysis. Calculate the probability of spatial proximity pairs appearing in the defect detection area, and calculate the spatial proximity entropy of the spatial proximity pairs using the information entropy algorithm; Based on spatial proximity entropy, the inspection feature overlap of all historical micro-defects is divided into fixed overlaps. and random overlap ; It should be explained that spatial proximity entropy measures the uncertainty of the spatial distribution of defects; based on spatial proximity entropy, the historical feature overlap is divided into fixed overlap and random overlap ; A high entropy value (H=0.97) is set for the spatial proximity entropy. A high entropy value indicates that the distribution of spatial proximity pairs is highly random, and feature overlap may be caused by external interference, corresponding to the dominant random overlap. If the spatial proximity entropy is lower than 0.2, it indicates that the defect pairs are concentrated, and the overlap is mostly caused by the intrinsic properties of the material (such as fiber layup direction and resin curing shrinkage), corresponding to the dominant fixed overlap. Different from the historical level micro-defect inspection, the characteristic overlap of the current level micro-defect inspection of carbon fiber and epoxy resin layer composite materials is obtained ; If the modal characteristic degradation equation is satisfied: , that is, to determine whether modal feature degradation occurs in different defects; It should be noted that the role of determining whether modal feature degradation occurs in different defects is: Function 1: Identify the source of feature confusion: By calculating the probability of occurrence of spatial proximity pairs and spatial proximity entropy, the feature overlap is divided into fixed overlap (stable confusion caused by intrinsic material properties) and random overlap (abnormal confusion caused by external interference), reducing the misjudgment of inherent material characteristics as defects or ignoring true abnormal confusion; Function 2: Triggering the hierarchical decomposition model to optimize classification. When modal feature degradation is determined, the third-order tensor low-rank constraint decomposition model is forcibly activated to separate the thermal stress coupling signal. By constructing the confusion matrix before and after decomposition to calculate the improvement rate, the reliability of multi-defect type classification is improved and signal aliasing misjudgment is reduced. Function 3: Determine the start conditions of dynamic load analysis: If the modal characteristics are degraded and the confusion improvement rate is lower than the threshold, it indicates that static detection cannot effectively distinguish defects. It is necessary to trigger dynamic response analysis under gradient moist heat load to capture the dynamic differences in defect detection characteristics under multi-field coupling, reducing the problem that traditional static detection cannot reflect defect evolution; in, is the characteristic overlap of the current carbon fiber and epoxy resin layer composite material, is the fixed overlap mean of all levels of micro-defects, is the mean random overlap of all levels of micro-defects, and k is the preset random overlap coefficient, which is used to adjust the threshold sensitivity of modal degradation judgment and is set according to the detection accuracy requirements; Among them, the hierarchical decomposition model is constructed to separate the signal confusion caused by thermal stress coupling, and the confusion evaluation method of the defect detection features is as follows: If modal feature degradation occurs, the three types of defect detection features of the overlapping detection vectors of all historical sampling points and the historical P-level thermal stress load are obtained for correlation coupling analysis to determine whether there is correlation coupling between the three types of defect detection features and the P-level thermal stress load; It can be understood by those skilled in the art that by calculating the Pearson correlation coefficient r between the three types of defect detection features and the P-level thermal stress load, if , determine the existence of thermal coupling to determine the correlation coupling between the three types of defect detection features and the P-level thermal stress load; If there is a related coupling, the three types of defect detection features based on the overlapping detection vectors of all current sampling points are combined with the P-level thermal stress load to construct a third-order tensor X, , and perform standardization; Where M is the total number of sampling points, N is the number of categories of defect detection features, P is the thermal stress load level, and R is the real number space; Preferably, a low-rank constraint is established: Perform low-rank constraint decomposition on the third-order tensor and construct the defect classification confusion matrix before and after the low-rank constraint decomposition; Among them, G, A, B, and C are the factor matrix after third-order tensor decomposition, sampling points, defect detection features, and potential characteristic components of thermal stress load, which are used to extract defect detection features independent of thermal stress interference. 、 is a regularization parameter used to balance the low-rank constraint of tensor decomposition and the fitting error to reduce overfitting; F is the norm, which quantifies the error between the original data and the decomposition result during the tensor decomposition process, achieving the separation accuracy of the thermal stress coupling signal. At the same time, it is coordinated with the low-rank constraint term to improve the reliability of defect detection feature extraction. Calculate the confusion improvement rate based on the defect classification confusion matrix and compare it with the preset confusion improvement threshold. If the confusion improvement rate is lower than the preset confusion improvement threshold, differential analysis is required. Otherwise, the changes in the confusion improvement rate are continuously monitored. Those skilled in the art will understand that the accuracy rate is obtained by ratioing the number of correctly classified samples to the total number of samples, and the defect classification accuracy rates before and after decomposition are calculated respectively. The deviation ratio of the accuracy rate after decomposition to the accuracy rate before decomposition is calculated to obtain the confusion improvement rate. The purpose of obfuscation assessment is: Objective 1: To evaluate the effectiveness of the hierarchical decomposition model in separating thermal stress coupling signals. By constructing the defect classification confusion matrix before and after low-rank constraint decomposition, the confusion improvement rate is calculated to determine the model's effectiveness in separating defect detection features from thermal stress coupling signals, thereby improving the reliability of classification and identification of multiple defect types. Purpose 2: Determine whether it is necessary to enter the dynamic load difference analysis process. If the confusion improvement rate is lower than the preset threshold, it indicates that static detection cannot effectively distinguish the confusion of defect detection features. It is necessary to trigger the dynamic response analysis under gradient wet and hot load to capture the difference in defect detection features under multi-field coupling.

[0027] The technical solution of this embodiment is: in the historical hierarchical micro-defect inspection of carbon fiber and epoxy resin layer composite materials, defect detection features are obtained and overlap detection is performed to obtain the feature overlap of historical hierarchical micro-defects; based on the feature overlap, it is used to determine whether modal feature degradation occurs in different defects. If so, a hierarchical decomposition model is constructed to separate the signal confusion caused by thermal stress coupling, and the confusion evaluation is performed on the defect detection features to obtain the confusion improvement rate and determine whether difference analysis is required; this is conducive to diagnosing the feature confusion phenomenon of different defects, reducing misjudgments caused by signal aliasing, and effectively alleviating the interference of thermal stress on defect detection features, thereby improving the reliability of classification and identification of multiple defect types.

[0028] Example 2:

[0029] See also Figure 1 As shown, the present invention is a method for detecting internal defects of composite material parts, comprising the following steps: S3. If difference analysis is required, the diagnostic test interface is divided and a gradient moist heat load is applied to analyze whether the difference in defect detection characteristics of the diagnostic test interface is significant. If significant, a dynamic response matrix under dynamic load is generated; The method of dividing the diagnostic detection interface and applying gradient heat load is as follows: Divide the defect detection area where modal feature degradation occurs into a diagnostic detection interface; Apply gradient moist heat load to the defect detection area according to the load parameter design principle; It will be understood by those skilled in the art that, based on historical thermal stress analysis, the temperature range is set to the actual service temperature of the cover material and extended by 20% (e.g., -10°C to 150°C), with a step size of 10°C to capture the characteristic mutation near the glass transition point of the resin; The humidity gradient was set from 30% RH to 95% RH in 10% RH steps, with each load level maintained for 4 hours to allow the resin to reach moisture equilibrium. Simultaneously, a static compressive stress of 0.1 times the interlaminar shear strength was applied to the highly overlapped defect detection area to simulate service loads. Multi-field parameters were synchronously controlled using a temperature and humidity chamber and a built-in servo loading system. After each load level stabilized, dynamic detection of defect detection features was performed using ultrasonic phased array, infrared thermal imaging, and CT. This data on the evolution of delamination, debonding, and porosity characteristics under the effects of moisture-heat-stress coupling was obtained, providing a foundation for constructing a dynamic response matrix. The method for analyzing whether the difference in defect detection characteristics of the diagnostic detection interface is significant is as follows: On the basis of applying gradient moisture and heat load, the formula is: Get the characteristic sensitivity S of the defect detection feature of the i-th type in the defect detection area i ; in, They represent the maximum and minimum values ​​of the i-th type defect detection feature within the applied gradient moist heat load range, L is the applied gradient moist heat load, and L max and L min Represent the maximum and minimum values ​​within the range of the applied gradient heat load, is the symbol of partial derivative; For example, Z1 represents a delamination defect detection feature; Z2 represents a debonding defect detection feature; and Z3 represents a porosity defect detection feature. It can be understood that the characteristic sensitivity is used to measure the dynamic response ability of the defect detection feature under gradient moisture and heat loads. It can quantify the change amplitude of the i-th type of defect detection feature (such as delamination, debonding, and porosity characteristics) within the gradient moisture and heat load range. The response slope of the defect detection feature to multi-field coupled loads (temperature, humidity, and stress) reflects the evolutionary sensitivity of the defect under dynamic loads. Among them, the function of obtaining feature sensitivity is: Function 1: Quantify dynamic response. By calculating the variation of defect characteristics under gradient moisture and heat load, quantify its dynamic response capability to multi-field coupling load and reflect the evolution sensitivity of defect detection characteristics. Function 2: Verify the effectiveness of the load, determine whether the applied gradient moisture and heat load effectively stimulates the difference in defect characteristics, enable the dynamic response matrix to truly reflect the defect behavior, and provide a data basis for subsequent analysis; Function 3: Use sensitivity data as feedback parameters to correct detection model errors and improve the system's adaptability to complex defects; Perform a significance test on the characteristic sensitivity of all defect detection features. If the defect detection features of the diagnostic detection interface are significantly different, rank the characteristic sensitivity of different types of defects. If the characteristic sensitivity of the delamination defect is the highest, the diagnostic detection interface is divided into regions again according to the characteristic sensitivity, and the number of divided interface partitions is obtained; Dynamic response matrix based on the number of regional partitions, characteristic dimensions, and number of load combinations.

[0030] S4. Based on the dynamic response matrix, perform time-frequency coupling analysis to identify the strain gradient in the fiber direction of the delamination, extract the sudden change direction angle, and identify the interference of the composite material anisotropy on the sudden change direction angle. If interference exists, filter the defective area; The method for performing time-frequency coupling analysis to identify the strain gradient in the fiber direction of the delamination is as follows: Perform time-frequency domain conversion based on the dynamic response matrix and construct the time-frequency matrix W(a,b); Preferably, the dynamic response matrix is ​​subjected to data preprocessing, the dynamic response matrix after data preprocessing is subjected to wavelet transform, and the axial strain signal is subjected to time-frequency decomposition to obtain the time-frequency matrix W(a, b) of the strain signal under different loads; Among them, a is the scale factor of frequency, and b is the translation factor; In the time-frequency domain, the spatial gradient of the time-frequency matrix is ​​calculated along the fiber direction of the carbon fiber and epoxy resin layer composite material to obtain the spatial derivative in the fiber direction. , as the strain gradient; in, is the coordinate of the fiber direction; Extract and analyze the strain gradient to obtain the characteristic frequency band, map the strain gradient value of the characteristic frequency band to the spatial position, and extract the mutation point of the strain gradient; Those skilled in the art will appreciate that, through spectrum analysis, characteristic frequency bands with concentrated energy in the strain gradient are determined (e.g., by filtering by frequency domain energy ratio), and the strain gradient values ​​within this frequency band are mapped to corresponding spatial position coordinates to form a position-strain gradient distribution map. Finally, an edge detection algorithm is used to extract points in the distribution map where the strain gradient mutation amplitude exceeds a preset threshold. These points are the strain gradient mutation points at the defect edge, which are used to locate the hidden boundaries of the delamination defect. Obtain the angle between the main direction of the strain gradient point and the fiber direction to obtain the sudden change direction angle; Among them, the method of identifying the interference of composite material anisotropy on sudden direction angle is: Decompose the strain gradient into components in different directions and calculate the frequency domain energy proportion of the strain gradient components in each direction at the characteristic frequency; By formula: Get the anisotropic correlation factor ; in, Fiber direction The frequency domain energy of is the angle between the mutation direction angle and the fiber direction, It is the sum of the frequency domain energy proportions of the components in each direction at the characteristic frequency; Based on anisotropic correlation factor Filter out the sudden change direction angles with interference, and filter out the defect areas corresponding to the delamination defects with interference; It should be explained that the purpose of filtering out interfering sudden direction angles is to eliminate the interference of the anisotropic intrinsic properties of the composite material (such as the fiber layup direction) on the detection signal and avoid the misjudgment of defect boundaries due to false signals of strain gradients in the fiber direction. By calculating the anisotropy correlation factor (η) to quantify the degree of interference, the sudden direction angles interfered by the inherent characteristics of the material are filtered out to ensure that the extracted sudden direction angles truly reflect the strain gradient mutation at the edge of the delamination defect, thereby locating the hidden delamination boundary, improving the accuracy of spatial position identification of the defect area, and reducing the problem of missed detection of small defects due to signal confusion. S5. Determine the hidden boundary of the delamination area based on the sudden change direction angle, determine the boundary hidden area, re-image the boundary hidden area, extract the geometric parameters of the hidden detection feature, compare it with the detection model output, generate an error correction matrix, feed it back to the hierarchical decomposition model, output the corrected defect recognition result and embed it into the composite material quality control system; The hidden boundary of the layered region is determined based on the mutation direction angle, and the boundary hidden region is determined based on the hidden boundary as follows: Preferably, based on the defect area corresponding to the filtered delamination defect, the coordinates of the mutation point of the strain gradient and the mutation direction angle corresponding to the mutation point are obtained, the coordinates of the mutation point and the mutation direction angle are spatially mapped, and a mutation direction angle-position range map is established; Based on the sudden change direction angle-position range map, an edge detection algorithm is used to identify the area in the distribution map where the direction angle change rate exceeds the preset threshold, which is the hidden boundary of the defect area corresponding to the delamination defect; Taking the hidden boundary as the basis, the distance is expanded to both sides to form a strip area as the boundary hidden area; The method for re-imaging the suspected edge defect area and extracting the geometric parameters of the defect detection features is as follows: Ultrasonic phased array and infrared thermal imaging are used to perform high-resolution re-imaging scans on hidden boundary areas. Edge extraction algorithms are used to extract the geometric parameters of defects from the scanned images, including the length, width, depth, shape, contour, and boundary location of the defects. The distribution of the defects in the composite material layer is also recorded. The extracted geometric parameters are compared with the predicted parameters output by the hierarchical decomposition model, the deviation values ​​of the same parameters are calculated, and an error correction matrix containing the location and size dimensions of the delamination defect is generated; The error correction matrix is ​​input into the hierarchical decomposition model, and the corrected defect identification results are output and embedded into the composite material quality control system; It will be understood by those skilled in the art that the error correction matrix is ​​input as a feedback signal into the hierarchical decomposition model, and by adjusting the factor matrix (such as G, A, B, C) and low-rank constraint parameters of the third-order tensor decomposition, the fitting accuracy of the model for complex boundaries and tiny defects is iteratively optimized, thereby outputting the corrected defect identification results; finally, the correction results are integrated into the composite material quality control system database to realize the dynamic update of the detection model and the closed-loop management of the quality control process.

[0031] The technical solution of this embodiment is: a difference analysis module: if difference analysis is required, the diagnostic detection interface is divided and a gradient wet-heat load is applied to analyze whether the defect detection feature differences of the diagnostic detection interface are significant. If significant, a dynamic response matrix under dynamic load is generated; based on the dynamic response matrix, a time-frequency coupling analysis is performed to identify the strain gradient in the fiber direction of the layer, the mutation direction angle is extracted and the interference of the anisotropy of the composite material on the mutation direction angle is identified. If interference exists, the defect area is filtered; based on the mutation direction angle, the hidden boundary of the layered area is determined, and the boundary hidden area is determined, the boundary hidden area is re-imaged, the geometric parameters of the hidden detection features are extracted, and an error correction matrix is ​​generated after comparison with the detection model output, which is fed back to the hierarchical decomposition model, and the corrected defect recognition results are output and embedded in the composite material quality control system. The hierarchical decomposition model is iteratively optimized through error feedback, thereby improving the adaptability and recognition reliability of the detection system to different materials and defect types.

[0032] Example 3:

[0033] like Figure 3 As shown, the present invention is a composite material part internal defect detection system, including the following modules Overlap extraction module: used to obtain defect detection features and perform overlap detection in the historical layer micro-defect inspection of carbon fiber and epoxy resin layer composite materials, and obtain the feature overlap of historical layer micro-defects; Confusion Analysis Module: Based on feature overlap, it determines whether modal feature degradation occurs in different defects. If so, a hierarchical decomposition model is constructed to separate signal confusion caused by thermal stress coupling. Confusion evaluation is performed on defect detection features to obtain the confusion improvement rate and determine whether differential analysis is required. Difference analysis module: If difference analysis is required, the diagnostic test interface is divided and a gradient moist heat load is applied to analyze whether the difference in defect detection characteristics of the diagnostic test interface is significant. If significant, a dynamic response matrix under dynamic load is generated; Mutation Identification Module: Based on the dynamic response matrix, it is used to perform time-frequency coupling analysis to identify the strain gradient in the fiber direction of the delamination, extract the mutation direction angle, and identify the interference of the composite material anisotropy on the mutation direction angle. If interference exists, the defective area is filtered out. Feedback correction module: Determine the hidden boundary of the delamination area based on the mutation direction angle, determine the boundary hidden area, re-image the boundary hidden area, extract the geometric parameters of the hidden detection features, generate an error correction matrix after comparing with the detection model output, feed back to the hierarchical decomposition model, output the corrected defect identification results and embed them into the composite material quality control system.

[0034] The above is a detailed description of an embodiment of the present invention. However, the content is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for detecting internal defects of composite material parts, characterized by: In the historical layer micro-defect inspection of carbon fiber and epoxy resin layer composite materials, the defect detection features are obtained and overlap detection is performed to obtain the feature overlap of historical layer micro-defects; Based on the feature overlap, determine whether modal feature degradation occurs in different defects. If so, construct a hierarchical decomposition model to separate signal confusion caused by thermal stress coupling, perform confusion evaluation on defect detection features, obtain confusion improvement rate, and determine whether differential analysis is required. If difference analysis is required, the diagnostic test interface is divided and a gradient moist heat load is applied to analyze whether the difference in defect detection characteristics of the diagnostic test interface is significant. If significant, a dynamic response matrix under dynamic load is generated; Based on the dynamic response matrix, time-frequency coupling analysis is performed to identify the strain gradient in the delaminated fiber direction, extract the sudden direction angle, and identify the interference of the composite material anisotropy on the sudden direction angle. If interference exists, the defective area is filtered.

2. The method for detecting internal defects of a composite material part according to claim 1, wherein: The overlap detection is performed as follows: Obtain overlapping detection vectors, map the overlapping detection vectors into three-dimensional space, and form characteristic distribution areas of the three types of defect detection features; Calculate the intersection and union of the distribution areas of different types of defect detection features, and calculate the overlap. Get the feature overlap.

3. The method for detecting internal defects of a composite material part according to claim 2, wherein: The method for obtaining the overlap detection vector is: Obtaining the delamination defect detection features, the pore defect detection features, and the debonding defect detection features of the defect area, and using the three defect detection features of the defect area as defect detection features; The defect area is simplified into sampling points, and an overlapping detection vector containing the defect detection features of each sampling point is constructed.

4. The method for detecting internal defects of a composite material part according to claim 1, wherein: The method of performing confusion evaluation on the defect detection feature is as follows: If modal feature degradation occurs, the three types of defect detection features of the overlapping detection vectors of all historical sampling points and the historical P-level thermal stress load are obtained for correlation coupling analysis to determine whether there is correlation coupling between the three types of defect detection features and the P-level thermal stress load; Wherein, P is the thermal stress load level; If there is relevant coupling, construct a third-order tensor and perform low-rank constraint decomposition, and construct the defect classification confusion matrix before and after low-rank constraint decomposition; The confusion improvement rate is calculated based on the defect classification confusion matrix. If the confusion improvement rate is lower than the preset confusion improvement threshold, difference analysis is required.

5. The method for detecting internal defects of a composite material part according to claim 4, characterized in that: The judgment method for the occurrence of the modal feature degradation is: Construct defect pairs of any two sets of defect detection features, such as delamination defect detection features, porosity defect detection features, and debonding defect detection features; Obtain the coordinates of the two sets of defect sampling points corresponding to the defect pair, calculate the Euclidean distance between the two sets of defect sampling point coordinates, and classify the defect pairs into spatially adjacent pairs and spatially unrelated pairs; Calculate the probability of spatial proximity pairs appearing in the defect detection area, calculate the spatial proximity entropy of spatial proximity pairs using the information entropy algorithm, and divide the feature overlap into fixed overlap and random overlap; Obtain the characteristic overlap of the current level micro-defect inspection, and establish a state characteristic degradation equation based on the characteristic overlap of the current carbon fiber and epoxy resin layer composite material, the fixed overlap mean of all level micro-defects, and the random overlap mean of all level micro-defects; If the modal characteristic degradation equation is satisfied, modal characteristic degradation occurs.

6. The method for detecting internal defects of a composite material part according to claim 1, wherein: The method for analyzing whether the difference in the defect detection characteristics of the diagnostic detection interface is significant is as follows: Divide the defect detection area where modal feature degradation occurs into a diagnostic detection interface; Apply a gradient moist heat load to the defect detection area to obtain the characteristic sensitivity of different types of defect detection features in the defect detection area. If the defect detection features of the diagnostic detection interface are significantly different, the characteristic sensitivity of different types of defects is ranked. If the characteristic sensitivity of the delamination defect is the highest, the diagnostic detection interface is divided into regions again according to the characteristic sensitivity, the number of divided interface partitions is obtained, and a dynamic response matrix is ​​constructed.

7. The method for detecting internal defects of a composite material part according to claim 1, wherein: The method of performing the filtering process on the defective area is as follows: Obtain the strain gradient and characteristic frequency band, map the strain gradient value of the characteristic frequency band to the spatial position, and extract the mutation point of the strain gradient; Obtain the angle between the main direction of the strain gradient point and the fiber direction to obtain the sudden change direction angle; Decompose the strain gradient into components in different directions, calculate the frequency domain energy proportion of the strain gradient components in each direction at the characteristic frequency, and calculate the anisotropy correlation factor; Based on the anisotropic correlation factor, the interference mutation direction angle is filtered out, and the defect area to which the delamination defect corresponding to the interference belongs is filtered out.

8. The method for detecting internal defects of a composite material part according to claim 7, wherein: The method of obtaining the strain gradient is as follows: The time-frequency domain is converted based on the dynamic response matrix, and the time-frequency matrix is ​​constructed. The spatial gradient of the time-frequency matrix is ​​calculated to obtain the spatial derivative in the fiber direction as the strain gradient.

9. The method for detecting internal defects of a composite material part according to claim 1, wherein: The following steps are also included: The hidden boundary of the delamination area is determined based on the mutation direction angle, and the boundary hidden area is determined. The boundary hidden area is re-imaged, and the geometric parameters of the hidden detection features are extracted. After comparison with the detection model output, an error correction matrix is ​​generated and fed back to the hierarchical decomposition model. The corrected defect identification results are output and embedded in the composite material quality control system.

10. A composite material part internal defect detection system, characterized by: A method for detecting internal defects of a composite material part according to any one of claims 1 to 9, comprising the following modules: Overlap extraction module: used to obtain defect detection features and perform overlap detection in the historical layer micro-defect inspection of carbon fiber and epoxy resin layer composite materials, and obtain the feature overlap of historical layer micro-defects; Confusion Analysis Module: Based on feature overlap, it determines whether modal feature degradation occurs in different defects. If so, a hierarchical decomposition model is constructed to separate signal confusion caused by thermal stress coupling. Confusion evaluation is performed on defect detection features to obtain the confusion improvement rate and determine whether differential analysis is required. Difference analysis module: If difference analysis is required, the diagnostic test interface is divided and a gradient moist heat load is applied to analyze whether the difference in defect detection characteristics of the diagnostic test interface is significant. If significant, a dynamic response matrix under dynamic load is generated; Mutation identification module: Based on the dynamic response matrix, it is used to perform time-frequency coupling analysis to identify the strain gradient in the fiber direction of the layer, extract the mutation direction angle and identify the interference of the anisotropy of the composite material on the mutation direction angle. If interference exists, the defective area is filtered.

Images