Composite structure damage identification method based on acoustic emission wavelet packet energy features

Abstract

The present application relates to the technical field of composite material structure monitoring, in particular to a composite material structure damage identification method based on acoustic emission wavelet packet energy characteristics, which comprises: S1, obtaining acoustic emission signal data of the composite material and performing pretreatment; S2, using a wavelet base function and a discrete wavelet inverse transform to perform wavelet packet decomposition on the denoised acoustic emission signal data of the composite material; S3, obtaining the processed acoustic emission signal data of the composite material, and extracting the characteristic energy of each acoustic emission signal data; S4, using a clustering method to perform clustering analysis and completing the composite material structure damage identification. The present application uses acoustic emission signals as the basis for analysis, uses wavelet transform and k-means clustering to accurately realize the structure damage identification and judgment of the composite material, can realize dynamic and continuous monitoring, is not affected by the complexity of the structure, has wide applicability, and is suitable for various composite materials.

Description

Technical Field

[0001] This invention belongs to the field of composite material structure monitoring technology, and specifically relates to a method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics. Background Technology

[0002] Composite materials, due to their lightweight, corrosion resistance, superior mechanical properties, and high design flexibility, are now widely used in the aerospace industry, and their usage has become an important indicator of the structural advancement and market competitiveness of aircraft. However, composite material connection structures are characterized by complex failure mechanisms, coupled failure modes, diverse damage propagation pathways, and large-scale adhesive layer damage zones. These characteristics make strength and failure analysis of composite material connection structures difficult, posing a significant obstacle to the reliability analysis of composite materials. As an indispensable key component of composite material structures, connection structure damage currently accounts for nearly 80% of composite material failures. Therefore, to avoid sudden structural failure of composite material connection structures, the development of intelligent health monitoring technologies and non-destructive damage detection technologies is urgently needed.

[0003] Acoustic emission (AE) refers to the phenomenon where transient elastic waves are generated in a material due to minute changes in a certain internal region when the material is subjected to external loading. As a non-destructive testing technique, AE is used to monitor the minute sounds released by materials and structures during loading, caused by processes such as microcracks and crack propagation within the material. These sounds are captured by placing sensors on the material surface, and their characteristics are analyzed to identify and assess the material's damage state. This technology is of great significance in structural health monitoring, including damage diagnosis, location, identification, and prediction in composite adhesive bonding structures.

[0004] Damage identification in composite material joint structures currently relies primarily on extracting traditional univariate statistical signal features from acoustic emission signals, such as rise time, peak frequency, centroid frequency, weighted frequency, ring count, and threshold voltage, to analyze the acoustic emission signal clusters. However, due to the limited understanding of existing univariate statistical features, effectively distinguishing and identifying various damage modes based on arbitrary univariate statistical acoustic emission time-domain and spectral characteristics remains extremely difficult. Furthermore, clustering analysis based on traditional acoustic emission features still suffers from numerous problems. First, peak frequency significantly impacts clustering; however, peak frequency extraction is highly susceptible to noise interference, which greatly affects the accuracy of clustering acoustic emission events with different damage modes. Therefore, a novel acoustic emission signal feature extraction method is needed to obtain more interpretable clustering results and optimize composite material damage identification methods.

[0005] Against this backdrop, this patent uses a composite material adhesive bonding structure with screw holes as the test object. It employs acoustic emission technology, combined with wavelet packet analysis, to extract features from the acoustic emission signal, in order to explore the failure mechanism and failure mode of this structure under tensile conditions. Furthermore, cluster analysis is introduced to classify and identify the damage types of the test specimen structure. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method for identifying structural damage in composite materials based on acoustic emission wavelet packet energy characteristics. Using acoustic emission signals as the basis for analysis, wavelet transform and k-means clustering analysis are used to accurately identify and judge structural damage in composite materials. This method has high sensitivity, enables dynamic and continuous monitoring, is unaffected by structural complexity, and has wide applicability to various composite materials.

[0007] To achieve the above objectives, the present invention discloses the following technical solution:

[0008] A method for damage identification of composite material structures based on acoustic emission wavelet packet energy characteristics, comprising:

[0009] S1: Acquire acoustic emission signal data of composite materials and preprocess the acoustic emission signal data;

[0010] Acoustic emission sensors are installed on the surface of the composite material structure, and a force loading test is performed on the composite material using a universal testing machine to collect acoustic emission signal data. The acoustic emission signal data is preprocessed, and the power spectrum of the acoustic emission signal data is calculated using the Welch power spectrum method. The peak frequency of the acoustic emission signal data is extracted, a peak frequency threshold is set, and acoustic emission signal data with a peak frequency lower than the threshold is filtered out as noise to obtain the denoised acoustic emission signal data f(t) of the composite material.

[0011] S2: Use wavelet basis functions and discrete wavelet inverse transform equations to perform wavelet packet decomposition on the denoised composite acoustic emission signal data in step S1;

[0012] S21: Decompose the denoised composite acoustic emission signal data from step S1 using wavelet basis functions:

[0013]

[0014] Where f(t) is the processed acoustic emission signal data; F(ω) is the frequency domain function of the acoustic emission signal data after Fourier transform of f(t); ω is the imaginary unit; t is the angular frequency of the acoustic emission signal data; WT(a,τ) is the time-frequency function of the acoustic emission signal data after wavelet transform; τ is the time shift of the acoustic emission signal data; a is the time scaling of the acoustic emission signal data; ψ is the mother wavelet function.

[0015] S22: Perform inverse discrete wavelet transform on the acoustic emission signal data f(t) of the composite material to obtain the first and second components of the acoustic emission signal data of the composite material. The inverse discrete wavelet transform equation is:

[0016]

[0017] Where DWT(i,k) are the wavelet transform coefficients of the acoustic emission signal data; i is the decomposition level of the acoustic emission signal data; and k is the discrete time-domain shift amount.

[0018] S3: Obtain the acoustic emission signal data of the composite material after processing in step S2, and extract the characteristic energy of each acoustic emission signal data;

[0019] The energy of the wavelet packet coefficients and the energy proportion of each frequency band in the acoustic emission signal data of the composite material are calculated, and the relative energy distribution of each energy level of the composite material is extracted as follows:

[0020]

[0021] in, Let represent the relative energy distribution of the i-th layer energy level of the composite material; E represents the frequency band energy of the acoustic emission signal data. Total (t) represents the energy of the wavelet packet coefficients of the acoustic emission signal data;

[0022] Each acoustic emission signal data is decomposed to the i-th level using wavelet packet decomposition. The acoustic emission signal data is decomposed to the i-th level using 2... i The energy ratio characteristics of each wavelet packet form a 2 i The characteristic energies of each energy level of the composite material are:

[0023]

[0024] Among them, X i Let i be the characteristic energy of the i-th layer energy level of the composite material;

[0025] S4: Use the k-means clustering method to perform cluster analysis on the characteristic energies of each energy level of the composite material in step S3, identify the types of structural damage in the composite material, and complete the full life cycle health monitoring of the composite material.

[0026] Preferably, step S1 uses the Welch power spectrum method to calculate the power spectrum of the acoustic emission signal data, specifically including the following steps:

[0027] S11: Segment the acoustic emission signal data. Divide the original acoustic emission signal data into N segments. In order to obtain better frequency resolution in the frequency domain and make the power spectrum smoother, set a 20% overlap between each window.

[0028] S12: Window the acoustic emission signal data and perform window function processing on each data segment to reduce the leakage effect of spectrum estimation and reduce the frequency resolution loss introduced by the window function;

[0029] S13: Perform Fourier transform on the signal within each acoustic emission signal data window;

[0030] S14: Calculate the power spectral density estimate for each frequency band of the acoustic emission signal data.

[0031] Preferably, the first component of the acoustic emission signal data of the composite material in step S22 is calculated using the inverse discrete wavelet transform equation, specifically as follows:

[0032]

[0033] Where AVC is the bandwidth of the first component after wavelet packet decomposition of the acoustic emission signal data; f s The sampling rate of the acoustic emission signal data; 2 i The number of components in the i-th layer after wavelet packet decomposition of acoustic emission signal data.

[0034] Preferably, the second component of the acoustic emission signal data transmitted by the composite material in step S22 is calculated using the inverse discrete wavelet transform equation, specifically as follows:

[0035]

[0036] Where DQ is the bandwidth of the second component of the acoustic emission signal data after wavelet packet decomposition.

[0037] Preferably, the discrete wavelet inverse transform equation in step S22 is determined based on the discrete wavelet transform mathematical model, specifically as follows:

[0038]

[0039] Where c represents the discrete wavelet transform coefficients.

[0040] Preferably, in step S3, the energy of the wavelet packet coefficients of each acoustic emission signal data in the composite material acoustic emission signal data and the energy proportion of each frequency band are calculated as follows:

[0041] The energy distribution of each composite material is distinguished using an energy criterion, and the energy level of each acoustic emission signal data is set as follows:

[0042]

[0043] Where j = 1, ..., 2 i Number the j-th component of the wavelet packet decomposed to the i-th level; To decompose the components of the i-th layer of the signal; t0 represents the frequency band energy of the acoustic emission signal data; t0 represents the start time of the acoustic emission signal data.

[0044] The energy of the wavelet packet coefficients of the acoustic emission signal data is set as follows:

[0045]

[0046] Among them, E Total (t) represents the energy of the wavelet packet coefficients of the acoustic emission signal data;

[0047] Based on the energy level of each acoustic emission signal data and the total energy of the acoustic emission signal data, the relative energy distribution of each energy level of the composite material is determined.

[0048] Preferably, in step S4, the k-means clustering method is used to perform cluster analysis on the characteristic energies of each energy level of the composite material in step S3, specifically as follows:

[0049] Let the characteristic energy of each energy level of the n sets of composite materials be X = {X1, X2, ..., X} i …X n The input is k-means clustering, which divides the data into k clusters such that the distance between each feature energy data sample and the cluster center is minimized. The k cluster centers are initialized as follows: The Euclidean distance from each acoustic emission signal data point to each cluster center is calculated as follows:

[0050]

[0051] Among them, Dis(X) i C p X is the Euclidean distance from the i-th acoustic emission signal data point to the cluster center of the p-th cluster; p is the cluster center number; X i C represents the characteristic energy of the i-th energy level of the composite material; p Let {C1, ..., C2 be the cluster center of the p-th class.} p …,C k}, 1 < p ≤ n; X it C represents the coordinate value of the t-th feature energy of the i-th acoustic emission feature energy data sample.pt Let t be the feature energy coordinate value of the cluster center of the p-th cluster; t is the feature energy coordinate value number; m is the feature energy dimension.

[0052] Preferably, the types of composite material structural damage in step S4 include matrix cracking (MC), adhesive debonding (AD), and fiber breakage (FB); the acoustic emission signal energy generated by the three types of damage is concentrated in different frequency bands, and the type of composite material structural damage is identified by the acoustic emission damage information contained in the wavelet packet energy.

[0053] The matrix cracking (MC) damage refers to the presence of bonded portions in the composite material structure, where fibers need to withstand pressure. The fibers at the bonded portions first experience a certain range of fracture, leading to a gradual decline in the mechanical properties of the composite material structure.

[0054] The adhesive debonding AD damage manifests as the composite material structure's mechanical properties becoming insufficient as the pressure increases, causing the adhesive to begin to debond. The continuous debonding of the adhesive leads to the gradual pull-out and breakage of the fibers bonded to the surface of the composite material structure.

[0055] The fiber breakage FB damage manifests as complete rupture and detachment of the adhesive, with a large number of fibers being pulled out and broken, leading to large-scale cracking of the resin matrix that is connected to the fibers and has weaker mechanical properties.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] (1) This invention uses acoustic emission signals as the basis for analyzing the structure of composite materials. It has high sensitivity, can realize dynamic monitoring and continuous monitoring, is not affected by the complexity of composite material structure, and the obtained acoustic emission signal data has real-time performance and can be applied to a variety of composite materials.

[0058] (2) The present invention uses the Welch method to calculate the power spectrum of acoustic emission signals, which can reduce the variance of acoustic emission signal data, reduce leakage effect, improve frequency resolution, and better reflect the characteristics of non-stationary signals changing over time.

[0059] (3) The present invention uses the Meyer wavelet, which is compactly supported, smooth, asymmetric, and has the characteristics of multi-resolution analysis, as the wavelet basis of wavelet transform. Its compact support enables it to effectively capture the local details of acoustic emission signals, while its smoothness and asymmetry endow it with sensitivity to the overall characteristics and transient properties of the signal.

[0060] (4) The present invention uses an unsupervised k-means machine learning algorithm to perform cluster analysis on acoustic emission signals. It is simple, convenient, efficient, and highly scalable. It does not require prior specification of the cluster prototype, has wide applicability, is easy to parallelize, has rigorous mathematical theory, and the results are interpretable. Attached Figure Description

[0061] Figure 1 This is a flowchart of the composite material structure damage identification method based on acoustic emission wavelet packet energy characteristics according to the present invention;

[0062] Figure 2 This is a photograph of a laboratory composite material joint specimen used in a tensile mechanical test, as an embodiment of the present invention.

[0063] Figure 3 The Welch spectrum of the acoustic emission signal of this invention;

[0064] Figure 4 This is a schematic diagram of the wavelet transform process of the signal in this invention;

[0065] Figure 5 This is a schematic diagram of the wavelet packet transform process of the signal in this invention;

[0066] Figure 6 This is a logic diagram of the wavelet packet decomposition process of the signal according to the present invention;

[0067] Figure 7 This is a time-domain signal diagram of the acoustic emission signal of the present invention;

[0068] Figures 8a-8h This is a time-domain signal diagram of wavelet packet reconstruction according to the present invention;

[0069] Figures 9a-9h The spectrum of the wavelet packet reconstructed signal of this invention;

[0070] Figure 10 This is a bar chart showing the energy ratio of the extracted wavelet packets from this invention.

[0071] Figure 11 This is a flowchart of the k-means clustering algorithm of the present invention;

[0072] Figure 12 A three-dimensional visualization of the acoustic emission signal k-means clustering analysis process of the present invention;

[0073] Figure 13a and Figure 13b The first and second projection diagrams are visualizations of the k-means clustering analysis process of acoustic emission signals in this invention.

[0074] Figure 14 This is a diagram illustrating the damage pattern evolution based on k-means clustering analysis using wavelet packet energy ratio characteristics, as presented in this invention.

[0075] Figures 15a-15c This is a bar chart showing the wavelet packet energy ratios at the centers of different damage types in this invention.

[0076] Figure 16aand Figure 16b This is a diagram showing the cracking and damage results of the CFRP matrix composite material output by the present invention;

[0077] Figure 17a and Figure 17b This is a diagram showing the debonding damage results of the CFRP adhesive in the composite material, as output by this invention.

[0078] Figure 18a and Figure 18b The diagram shows the fracture damage results of the CFRP fiber composite material output by this invention. Detailed Implementation

[0079] Exemplary embodiments, features, and aspects of the present invention will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0080] This invention provides a method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics, using composite material connectors for verification and analysis. Figure 1 As shown, acoustic emission signal data of composite materials is acquired and preprocessed; wavelet packet decomposition is performed on the denoised acoustic emission signal data of composite materials using wavelet basis functions and inverse discrete wavelet transform; the processed acoustic emission signal data of composite materials is acquired, and the feature energy of each acoustic emission signal data is extracted; clustering analysis is performed using clustering methods to complete the identification of structural damage in composite materials; this includes:

[0081] Step S1: Acquire acoustic emission signal data of composite material connectors and preprocess the acoustic emission signal data.

[0082] When composite material connectors are subjected to external forces or undergo minute deformations, micro-cracks or damage can occur, accompanied by the release of energy. A portion of this released energy propagates as elastic waves and reaches the material surface, causing minute vibrations. Acoustic emission sensors are then mounted on the surface of the composite material structure.

[0083] These weak vibration signals are converted into electrical signals, and after being amplified and digitally processed by the acquisition system, they are collected as acoustic emission signal data. A universal testing machine is used to perform force loading tests on the composite material connectors to collect the acoustic emission signal data. Taking a laboratory test as an example, the entire testing system includes an acoustic emission sensor 1, a preamplifier 2, an acoustic emission receiver 3, a composite material connector specimen 4, and a universal testing machine 5. Figure 2As shown, the acoustic emission sensor is arranged on the surface of the composite material joint specimen through a coupling agent. When the universal testing machine applies force to the specimen for tension, the acoustic emission sensor will continuously transmit the collected acoustic emission signal to the acoustic emission receiver after being amplified by the preamplifier. The signal is recorded after being preprocessed and digitized.

[0084] Preprocessing of acoustic emission signal data is performed because friction between the surface of composite material connectors and force-loading equipment or external particles also generates acoustic emission signals, which, as noise, can be converted to the frequency domain. Frequency domain thresholds are set to filter out these internal damage signals from non-composite material connectors. The Welch power spectrum method is used to calculate the power spectrum of the acoustic emission signal data, extracting the peak frequency. Noise signals with excessively low or high peak frequencies are filtered out. The specific steps include:

[0085] Step S11: Segment the acoustic emission signal data. Divide the original acoustic emission signal data into N segments. In order to obtain better frequency resolution in the frequency domain and make the power spectrum smoother, set a 20% overlap between each window.

[0086] Step S12: Window the acoustic emission signal data and perform window function processing on each data segment to reduce the leakage effect of spectrum estimation and reduce the frequency resolution loss introduced by the window function.

[0087] Step S13: Perform Fourier transform on the signal within each acoustic emission signal data window.

[0088] Step S14: Calculate the power spectral density estimate for each frequency band of the acoustic emission signal data.

[0089] The peak frequency of the acoustic emission signal data is extracted, a peak frequency threshold is set, and acoustic emission signal data with a peak frequency lower than the threshold is filtered out as noise to obtain the denoised acoustic emission signal data f(t) of the composite material connector.

[0090] Step S2: Use wavelet basis functions and discrete wavelet inverse transform equations to perform wavelet packet decomposition on the acoustic emission signal data of the composite material connector after denoising in step S1.

[0091] Step S21: Wavelet transform of a finite-length, attenuating wavelet basis acts like a mobile signal microscope, simultaneously acquiring the frequency and time information of the signal; the acoustic emission signal data of the composite material connector after denoising in step S1 is decomposed using wavelet basis functions as follows:

[0092]

[0093] Where f(t) is the processed acoustic emission signal data; F(ω) is the frequency domain function of the acoustic emission signal data after Fourier transform of f(t); ω is the imaginary unit; t is the angular frequency of the acoustic emission signal data; WT(a,τ) is the time-frequency function of the acoustic emission signal data after wavelet transform; τ is the time shift of the acoustic emission signal data; a is the time scaling of the acoustic emission signal data; and ψ is the mother wavelet function.

[0094] Step S22: Perform inverse discrete wavelet transform (DWT) on the acoustic emission signal data f(t) of the composite material connector using the discrete wavelet transform (DWT). This transforms the original waveform to obtain the first and second components of the acoustic emission signal data. The first component represents the approximate components of the acoustic emission signal data, and the second component represents the detail components of the acoustic emission signal data. The inverse discrete wavelet transform equation is:

[0095]

[0096] Where DWT(i,k) are the wavelet transform coefficients of the acoustic emission signal data; i is the decomposition level of the acoustic emission signal data; and k is the discrete time-domain shift.

[0097] In this embodiment of the invention, the Meyer wavelet is selected as the mother wavelet. The Meyer wavelet is a compactly supported, smooth, and asymmetric wavelet function with multi-resolution analysis characteristics. Its compact support allows it to effectively capture local details of the signal, while its smoothness and asymmetry endow it with sensitivity to the overall characteristics and transient properties of the signal. As a tool for multi-resolution analysis, the Meyer wavelet can provide frequency information at different scales, with the wavelet packet energy features extracted in this study being the most prominent. However, when processing acoustic emission signals, wavelet transform does not further decompose high-frequency detail signals. Figure 3 As shown, it only continuously decomposes the low-frequency part of the acoustic emission signal, which leads to the loss of a large amount of high-frequency detail information. For acoustic emission random signals with high sampling rates, its decomposition and representation capabilities are relatively limited. Figure 4 The wavelet decomposition process is shown. Wavelet packet transform is a further development of wavelet transform. It not only retains the time-frequency analysis characteristics of wavelet transform, but also enables simultaneous decomposition of the low-frequency and high-frequency components of the signal. By selecting different resolutions at different frequency levels, the original acoustic emission signal can be completely decomposed across the entire frequency band. Figure 5 The wavelet packet decomposition process shown demonstrates how wavelet packet transform decomposes and reconstructs signals at different scales, making it highly effective in processing acoustic emission signals with multi-scale characteristics.

[0098] The first component of the acoustic emission signal data of the composite material connector is calculated using the inverse discrete wavelet transform equation, specifically:

[0099]

[0100] Where AVC is the bandwidth of the first component after wavelet packet decomposition of the acoustic emission signal data; f s The sampling rate of the acoustic emission signal data; 2 i The number of components in the i-th layer after wavelet packet decomposition of acoustic emission signal data.

[0101] The second component of the acoustic emission signal data transmitted by the composite material connector is calculated using the inverse discrete wavelet transform equation, specifically:

[0102]

[0103] Where DQ is the bandwidth of the second component of the acoustic emission signal data after wavelet packet decomposition.

[0104] The inverse discrete wavelet transform equation is determined based on the mathematical model of the discrete wavelet transform, specifically:

[0105]

[0106] Where c represents the discrete wavelet transform coefficients.

[0107] by Figure 2 Taking the tensile mechanical test of the composite material joint specimen as an example, the sampling frequency was set to 2MHz. According to the sampling theorem, the frequency range at the wavelet packet decomposition node (0,0) should be 0-1MHz. In the preliminary univariate statistical characteristic analysis, based on the acoustic emission signal data and the distribution density of the centroid frequency in the acoustic emission impact group, the energy of the acoustic emission signal is mainly concentrated in the 80-500kHz frequency band. For example... Figure 6 The diagram shows the process logic of wavelet packet decomposition of the signal according to the present invention. The acoustic emission signal undergoes four-level wavelet packet decomposition, and the wavelet packet coefficients of the first eight nodes of the fourth level are collected: (4,0), (4,1), (4,2), (4,3), (4,4), (4,5), (4,6), (4,7) for wavelet packet signal reconstruction. The eight frequency bands corresponding to these eight nodes are: 0-62.5kHz, 62.5-125kHz, 125-187.5kHz, 187.5-250kHz, 250-312.5kHz, 312.5-375kHz, 375-437.5kHz, and 437.5-500kHz. Figure 7 The image shown is a time-domain signal diagram of the acoustic emission signal of this invention. This is the original signal diagram; the wavelet packet decomposition process involves... Figure 7 It is broken down into Figure 8 and Figure 9, as shown below.

[0108] After wavelet packet decomposition of the acoustic emission signal, starting from 0 kHz, the wavelet packet coefficients are rearranged in frequency band order, with increments of 62.5 kHz. This ensures that the eight nodes correspond to the following eight frequency bands: 0-62.5 kHz, 62.5-125 kHz, 125-187.5 kHz, 187.5-250 kHz, 250-312.5 kHz, 312.5-375 kHz, 375-437.5 kHz, and 437.5-500 kHz. After rearrangement, the acoustic emission impact signal is reconstructed in the time domain using the decomposed wavelet packet coefficients, resulting in eight components, as shown in Figure 8. Figure 8a The time-domain signal is reconstructed from wavelet packets in the frequency band of 0-62.5kHz; Figure 8b The time-domain signal is reconstructed from wavelet packets in the frequency band 62.5-125kHz; Figure 8c The time-domain signal is reconstructed from wavelet packets in the frequency band of 125-187.5kHz; Figure 8d The time-domain signal is reconstructed from wavelet packets in the frequency band of 187.5-250kHz; Figure 8e The time-domain signal is reconstructed from wavelet packets in the frequency band of 250-312.5kHz; Figure 8f The time-domain signal is reconstructed from wavelet packets in the frequency band of 312.5-375kHz; Figure 8g The time-domain signal is reconstructed from wavelet packets in the frequency band of 375-437.5kHz; Figure 8h The time-domain signal is reconstructed from wavelet packets in the 437.5-500kHz frequency band. The frequency domain signal is shown in Figure 9. Figure 9a The spectrum of the wavelet packet reconstructed signal in the frequency band 0-62.5kHz;

[0109] Figure 9b The spectrum of the wavelet packet reconstructed signal in the frequency band 62.5-125kHz; Figure 9c The spectrum of the wavelet packet reconstructed signal in the frequency band of 125-187.5kHz; Figure 9d The spectrum of the wavelet packet reconstructed signal in the frequency band of 187.5-250kHz; Figure 9e The spectrum of the wavelet packet reconstructed signal in the frequency band of 250-312.5kHz; Figure 9f The spectrum of the wavelet packet reconstructed signal in the frequency band of 312.5-375kHz; Figure 9g The spectrum of the wavelet packet reconstructed signal in the frequency band of 375-437.5kHz; Figure 9h The spectrum of the wavelet packet reconstructed signal in the frequency band 437.5-500kHz.

[0110] Step S3: Obtain the acoustic emission signal data of the composite material connector after processing in step S2, and extract the characteristic energy of each acoustic emission signal data.

[0111] The energy of the wavelet packet coefficients and the energy proportion of each frequency band in the acoustic emission signal data of the composite material connector are calculated as follows:

[0112] Energy criteria are used to distinguish the energy distribution of each composite material connector, and the energy level of each acoustic emission signal data is set as follows:

[0113]

[0114] Where j = 1,...,2 i Number the j-th component of the wavelet packet decomposed to the i-th level; To decompose the components of the i-th layer of the signal; t0 represents the frequency band energy of the acoustic emission signal data; t0 represents the start time of the acoustic emission signal data.

[0115] The energy of the wavelet packet coefficients of the acoustic emission signal data is set as follows:

[0116]

[0117] Among them, E Total (t) represents the energy of the wavelet packet coefficients of the acoustic emission signal data.

[0118] Based on the energy level of each acoustic emission signal data and the total energy of the acoustic emission signal data, the relative energy distribution of each energy level of the composite material connector is determined. for:

[0119]

[0120] in, Let represent the relative energy distribution of the i-th layer energy level in the composite material connector; E represents the frequency band energy of the acoustic emission signal data. Total (t) represents the energy of the wavelet packet coefficients of the acoustic emission signal data.

[0121] Each acoustic emission signal data is decomposed to the i-th level using wavelet packet decomposition. The acoustic emission signal data is decomposed to the i-th level using 2... i The energy ratio characteristics of each wavelet packet form a 2 i The characteristic energies of each energy level in the composite material connector are:

[0122]

[0123] Among them, X i Let be the characteristic energy of the i-th layer energy level of the composite material connector.

[0124] Since the acoustic emission signals generated by different damages in composite material connectors are concentrated in different frequency bands—for example, matrix-shaped damage is generally concentrated in low frequencies, while fiber-shaped damage is generally concentrated in high frequencies—the wavelet packet energy features actually contain a great deal of acoustic emission damage information. Unsupervised machine learning algorithms can be used to mine this information and identify different acoustic emission damage signals. In this embodiment of the invention, for the acoustic emission signals generated by damage in the composite material connector structure, if the sampling rate is 1MHz, wavelet packet decomposition is used to the i=3rd layer to form a 22 3 =8-dimensional eigenvectors.

[0125] The embodiments of the present invention are based on Figure 2 Taking the tensile mechanical test of composite material joint specimens as an example, the square root of the sum of squares of each element is calculated to determine the energy of the wavelet packet signal at each node. After summing the total energy, the energy percentage of the wavelet packet at each node is calculated, and a bar chart is plotted as follows. Figure 10 As shown, nodes 1 to 8 correspond to eight frequency bands: F1 (0-62.5kHz), F2 (62.5-125kHz), F3 (125-187.5kHz), F4 (187.5-250kHz), F5 (250-312.5kHz), F6 (312.5-375kHz), F7 (375-437.5kHz), and F8 (437.5-500kHz). These eight energy proportions form an eight-dimensional feature vector.

[0126] Step S4: Use the k-means clustering method to perform cluster analysis on the characteristic energies of each energy level of the composite material connectors in Step S3. The k-means clustering algorithm is a simple, fast, widely applicable, and highly interpretable unsupervised learning method, belonging to the field of machine learning. Since the algorithm only involves calculating the distance between samples and cluster centers, it can achieve efficient cluster analysis with appropriate data structures and algorithm optimization, especially suitable for large-scale datasets. It allows for flexible selection of the number of clusters k, making it suitable for various types of data. The clustering results it produces are intuitive and easy to understand, with a rigorous mathematical theoretical foundation. The goal of the k-means clustering algorithm is to divide the samples in the dataset into k clusters, minimizing the distance between each sample and the center of its cluster. The algorithm flow is as follows: Figure 11 As shown, data points are iteratively assigned to k clusters to minimize the distance between each data point and the center of its cluster. The cluster centers are continuously updated until a convergence condition is met. The goal is to minimize the variance of data points within a cluster and maximize the difference between clusters, thereby grouping data points with similar characteristics together. Specifically, the characteristic energies X = {X1, X2…X...} of each energy level in n sets of composite material connectors are assigned. i …Xn The input is k-means clustering, which divides the data into k clusters such that the distance between each feature energy data sample and the cluster center is minimized. The k cluster centers are initialized as follows: The Euclidean distance from each acoustic emission signal data point to each cluster center is calculated as follows:

[0127]

[0128] Among them, Dis(X) i C p X is the Euclidean distance from the i-th acoustic emission signal data point to the cluster center of the p-th cluster; p is the cluster center number; X i C represents the characteristic energy of the i-th energy level of the composite material connector; p Let {C1, ..., C2 be the cluster center of the p-th class.} p …,C k}, 1 < p ≤ n; X it C represents the coordinate value of the t-th feature energy of the i-th acoustic emission feature energy data sample. pt Let t be the t-th feature energy coordinate value of the cluster center of the p-th cluster; t is the feature energy coordinate value number; m is the feature energy dimension. Taking a sampling rate of 1MHz and wavelet packet decomposition to the 3rd layer as an example, then m = 8, t = 1, ..., 8.

[0129] The feature vectors are input into a clustering algorithm for cluster analysis, and the k-means clustering results are visualized as follows: Figure 12 As shown, the results of three types of three-dimensional clustering analysis are presented. Figure 13a This is a first projection diagram visualizing the k-means clustering analysis process of acoustic emission signals according to the present invention, representing the clustering analysis results of three types of peak amplitude (dB); as shown... Figure 13b This is a second projection diagram visualizing the k-means clustering analysis process of acoustic emission signals according to the present invention, showing the clustering analysis results of three types of peak frequencies (kHz).

[0130] Based on the clustering results, acoustic emission events are divided into different clusters. Taking tensile test data of composite material connectors as an example, this method identifies three damage modes that occur in composite material connectors from intact to fracture processes, such as... Figure 14 This is a damage mode evolution diagram based on k-means clustering analysis using wavelet packet energy ratio characteristics, as presented in this invention. Type 1 is matrix fracture (MC), type 2 is adhesive debonding (AD), and type 3 is fiber fracture (FB). For the composite adhesive joint specimens in this experiment, the failure process, based on the damage signals identified by this method, is as follows:

[0131] 1. Due to the presence of the bonded portion, the fibers bear a significant amount of pressure, causing the fibers at the bonded area to fracture within a certain range first, leading to a gradual decrease in the mechanical properties of the specimen; for example... Figure 15a A histogram of wavelet packet energy ratios for the cluster center of damage type (1).

[0132] 2. As the tensile test progresses, the mechanical properties of the specimen become insufficient, and the adhesive begins to debond. The continuous debonding causes the fibers adhered to the surface to be gradually pulled out and break. For example... Figure 15b A histogram of wavelet packet energy ratios for the cluster center of damage type (2).

[0133] 3. In the final moments, the adhesive completely breaks down and detaches, causing a large number of fibers to be pulled out and break, resulting in large-scale cracking of the resin matrix, which is connected to the fibers and has weaker mechanical properties. This method allows for, to some extent, life-cycle structural health monitoring of composite material connectors. For example... Figure 15c A histogram of wavelet packet energy ratios for the cluster center of damage type (3).

[0134] K-means clustering analysis based on wavelet packet energy ratio characteristics of acoustic emission signals was used to identify three damage modes in composite bonded specimens: matrix cracking (MC), adhesive debonding (AD), and fiber breakage (FB). Figures 16, 17, and 18 show the damage type of each acoustic emission event observed using electron microscopy for these three damage modes. Figure 16a This is a graph showing the cracking and damage detection results of the CFRP matrix composite material output by an embodiment of the present invention. Figure 16b This is an image showing the matrix cracking and damage observed using an electron microscope according to the present invention.

[0135] like Figure 17a This is a graph showing the detection results of CFRP adhesive debonding damage in composite materials, as output by an embodiment of the present invention. Figure 17b This image shows the adhesive debonding damage observed using an electron microscope according to the present invention. Figure 18a This is a graph showing the results of CFRP fiber fracture damage detection in an embodiment of the present invention. Figure 18b This is an image showing the fiber fracture damage observed using an electron microscope according to the present invention.

[0136] This invention focuses on composite material bonding structures and uses real-time monitored acoustic emission signals as the basis for analysis. The Welch method is employed to calculate the signal power spectrum, enabling the extraction of traditional acoustic emission characteristics such as rise time, peak amplitude, and peak frequency. Univariate statistical analysis of the acoustic emission signals reveals the signal distribution. Mayer wavelets are used as the wavelet basis to perform wavelet packet decomposition on the acoustic emission signals, enabling the extraction of wavelet packet energy features. This provides an 8-dimensional feature vector for multivariate statistical analysis of the acoustic emission signals. A k-means clustering algorithm model is established. The extracted feature vectors are input into the k-means clustering algorithm model, and finally, through acoustic emission feature analysis, three damage types in the composite material bonding specimens are identified and observed under an electron microscope: matrix cracking, adhesive debonding, and fiber breakage.

[0137] The beneficial effects of this invention are as follows: This invention provides a method for identifying structural damage in composite material connectors based on acoustic emission wavelet packet energy characteristics. Using acoustic emission signals as the analysis basis, it exhibits high sensitivity, enabling dynamic and continuous monitoring, unaffected by structural complexity, providing real-time data, and is applicable to various composite materials. The Welch method is used to calculate the signal power spectrum, which reduces variance, minimizes leakage effects, improves frequency resolution, and better reflects the time-varying characteristics of non-stationary signals. Wavelet packet decomposition is used to extract acoustic emission signal energy characteristics, preserving not only the time-frequency analysis characteristics of wavelet transform but also enabling simultaneous analysis of both low-frequency and high-frequency components of the signal. The solution involves selecting different resolutions at different frequency levels to achieve a complete full-band decomposition of the original acoustic emission signal. Using the compactly supported, smooth, and asymmetric Mayer wavelet, which possesses multi-resolution analysis characteristics, as the wavelet basis for wavelet transform, its compact support allows it to effectively capture local details of the signal, while its smoothness and asymmetry endow it with sensitivity to the overall characteristics and transient properties of the signal. An unsupervised k-means machine learning algorithm is employed for cluster analysis of the acoustic emission signal. This method is simple, efficient, and highly scalable. It does not require pre-specifying the cluster prototype, has wide applicability, is easily parallelized, has rigorous mathematical theory, and its results are interpretable. Ultimately, this invention enables effective classification and identification of damage to composite materials using acoustic emission signals.

[0138] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics, characterized in that, It includes: S1: Acquire acoustic emission signal data of composite materials and preprocess the acoustic emission signal data; Acoustic emission sensors were mounted on the surface of the composite material structure, and a universal testing machine was used to perform force loading tests on the composite material to collect acoustic emission signal data. The acoustic emission signal data was preprocessed, and the power spectrum of the acoustic emission signal data was calculated using the Welch power spectrum method. The peak frequency of the acoustic emission signal data was extracted, and a peak frequency threshold was set. Acoustic emission signal data with peak frequencies below the threshold were filtered out as noise, resulting in the denoised acoustic emission signal data of the composite material. Specifically, it includes the following steps: S11: Segment the acoustic emission signal data, dividing the original acoustic emission signal data into N segments. In order to obtain better frequency resolution in the frequency domain and make the power spectrum smoother, a 20% overlap is set between each window. S12: Window the acoustic emission signal data and perform window function processing on each data segment to reduce the leakage effect of spectrum estimation and reduce the frequency resolution loss introduced by the window function; S13: Perform Fourier transform on the signal within each acoustic emission signal data window; S14: Calculate the power spectral density estimate for each frequency band of the acoustic emission signal data; S2: Wavelet packet decomposition is performed on the denoised composite acoustic emission signal data from step S1 using wavelet basis functions and discrete wavelet inverse transform equations; Meyer wavelet is selected as the mother wavelet. S21: Decompose the denoised composite acoustic emission signal data from step S1 using wavelet basis functions: ； in, The processed acoustic emission signal data; for Frequency domain function of acoustic emission signal data after Fourier transform; The imaginary unit; The angular frequency of the acoustic emission signal data; For time; This is the time-frequency function of the acoustic emission signal data after wavelet transform; The time shift of the acoustic emission signal data; The time scaling factor for acoustic emission signal data; For the mother wavelet function; S22: Acoustic emission signal data of composite materials Performing an inverse discrete wavelet transform yields the first and second components of the acoustic emission signal data from the composite material. The inverse discrete wavelet transform equation is as follows: ； in, These are the wavelet transform coefficients of the acoustic emission signal data; The decomposition hierarchy of acoustic emission signal data; It is a discrete time-domain translation; S3: Obtain the acoustic emission signal data of the composite material after processing in step S2, and extract the characteristic energy of each acoustic emission signal data; The energy of the wavelet packet coefficients and the energy proportion of each frequency band in the acoustic emission signal data of the composite material are calculated, and the relative energy distribution of each energy level of the composite material is extracted as follows: ； in, For composite materials The relative energy distribution of the layer levels; The frequency band energy of acoustic emission signal data; The energy of the wavelet packet coefficients of the acoustic emission signal data; Each acoustic emission signal data is decomposed into its nearest 1 using wavelet packets. Layer, acoustic emission signal data decomposed to the first Layer The energy ratio characteristics of each wavelet packet form a The characteristic energies of each energy level of the composite material are: ； in, For composite materials Characteristic energies of layer energy levels; S4: Use the k-means clustering method to perform cluster analysis on the characteristic energies of each energy level of the composite material in step S3, identify the types of structural damage in the composite material, and thus monitor the health of the composite material throughout its entire life cycle. Specifically, the k-means clustering method is used to perform cluster analysis on the characteristic energies of each energy level of the composite material in step S3. Will Characteristic energies of each energy level in the composite material Input the k-means clustering method and divide it into Clusters are defined such that the distance between each feature energy data sample and the cluster center of its respective cluster is minimized. Initialize the cluster centers: Calculate the Euclidean distance from each acoustic emission signal data point to each cluster center: ； in, For the first The acoustic emission signal data point to the first Euclidean distance between cluster centers; Assign cluster center numbers; For composite materials Characteristic energies of layer energy levels; For the first Cluster centers of classes ; For the first The first acoustic emission characteristic energy data sample Each characteristic energy coordinate value; For the first The first cluster center Each characteristic energy coordinate value; Number the characteristic energy coordinate values; Let be the characteristic energy dimension.

2. The method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics according to claim 1, characterized in that: The first component of the acoustic emission signal data of the composite material in step S22 is calculated using the inverse discrete wavelet transform equation, specifically: ； in, The bandwidth of the first component after wavelet packet decomposition of acoustic emission signal data; The sampling rate of the acoustic emission signal data; After wavelet packet decomposition of acoustic emission signal data, the first... Number of components in a layer.

3. The method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics according to claim 1, characterized in that: The second component of the acoustic emission signal data transmitted by the composite material in step S22 is calculated using the inverse discrete wavelet transform equation, specifically: ； in, The bandwidth of the second component of the acoustic emission signal data after wavelet packet decomposition.

4. The method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics according to claim 1, characterized in that: The discrete wavelet inverse transform equation in step S22 is determined based on the discrete wavelet transform mathematical model, specifically as follows: ； in, These are the discrete wavelet transform coefficients.

5. The method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics according to claim 1, characterized in that: Step S3 involves calculating the energy of the wavelet packet coefficients of each acoustic emission signal data in the composite material acoustic emission signal data and the energy proportion of each frequency band, specifically as follows: The energy distribution of each composite material is distinguished using an energy criterion, and the energy level of each acoustic emission signal data is set as follows: ； in, Decompose the wavelet packet to the th The first layer Each component number; To decompose the signal Layer components; The frequency band energy of acoustic emission signal data; This is the start time of the acoustic emission signal data; The energy of the wavelet packet coefficients of the acoustic emission signal data is set as follows: ； in, The energy of the wavelet packet coefficients of the acoustic emission signal data; Based on the energy level of each acoustic emission signal data and the total energy of the acoustic emission signal data, the relative energy distribution of each energy level of the composite material is determined. .

6. The method for identifying damage in composite material structures based on acoustic emission wavelet packet energy characteristics according to claim 1, characterized in that: The types of damage to the composite material structure in step S4 include matrix cracking (MC), adhesive debonding (AD), and fiber breakage (FB). The acoustic emission signal energy generated by the three types of damage is concentrated in different frequency bands. The type of damage to the composite material structure is identified by using the acoustic emission damage information contained in the wavelet packet energy. The matrix cracking (MC) damage refers to the presence of bonded portions in the composite material structure, where fibers need to withstand pressure. The fibers at the bonded portions first experience a certain range of fracture, leading to a gradual decline in the mechanical properties of the composite material structure. The adhesive debonding AD damage manifests as the mechanical properties of the composite structure decrease as the pressure on the composite structure increases, the adhesive begins to debond, and the continuous debonding of the adhesive causes the fibers bonded to the surface of the composite structure to be gradually pulled out and broken. The fiber breakage FB damage manifests as adhesive rupture and debonding, with a large number of fibers being pulled out and broken, leading to cracking of the resin matrix that is connected to the fibers and has weaker mechanical properties.

Images