A method and apparatus for online monitoring of acoustic emission in laser shock enhancement quality

Abstract

This invention discloses an online monitoring method and apparatus for acoustic emission of laser shock stabilization (LSS) enhancement quality. The method involves converting each sub-signal containing one modal information into its spectral information; constructing a first complex network model for each mode based on the spectral information of each sub-signal, and extracting the degree distribution features of the first complex network model; calculating the degree distribution mutual information index matrix between different sub-signals based on the degree distribution features; converting the degree distribution mutual information index matrix into a sparse adjacency matrix; constructing a second complex network model based on the sparse adjacency matrix, and extracting the average weight degree feature of the second complex network model; and fitting the changing trend of the average weight degree feature to obtain a fitting index for characterizing the quality of laser shock stabilization. This invention can fully utilize the multimodal information of acoustic emission signals, combine it with complex network theory to construct quality-sensitive features, stably characterize processing quality, is unaffected by the environment, and has high monitoring accuracy.

Description

Technical Field

[0001] This invention belongs to the field of laser shock peening, and specifically relates to an online monitoring method and device for acoustic emission of laser shock peening quality. Background Technology

[0002] Laser Shock Peening (LSP) is a novel surface strengthening technology that can significantly improve the mechanical properties, corrosion resistance, and fatigue performance of metallic materials, extending their service life. Compared to traditional surface strengthening techniques, it can introduce a deeper residual compressive stress layer onto the metal surface. Therefore, it has broad application prospects in aerospace, nuclear energy, and petrochemical fields. The essence of LSP is to induce high-temperature, high-pressure shock waves on the material surface, causing microstructural changes such as dislocation movement and grain refinement in the subsurface material, thereby forming plastic deformation to improve the macroscopic properties of the material.

[0003] To address existing online monitoring methods for laser shock annealing, Chinese Patent No. CN 113340494 A invented an online monitoring method for laser shock annealing quality based on harmonic wavelet band energy. This method characterizes the processing quality by extracting wavelet packet band energy features, thereby achieving online monitoring of laser shock annealing. However, this method does not fully utilize the multimodal characteristics of acoustic emission signals, only using the energy proportion characteristics of a single mode band. Furthermore, because the correlation between band energy and processing quality is low and easily affected by the processing environment, this method clearly suffers from low monitoring accuracy. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides an online monitoring method and device for acoustic emission of laser shock enhancement quality. It can fully utilize the multimodal information of acoustic emission signals, combine complex network theory to construct quality-sensitive features, stably characterize processing quality, is unaffected by the environment, and has high monitoring accuracy.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0006] A method for online monitoring of acoustic emission of laser-strengthened mass includes:

[0007] Real-time acquisition of laser-shock-enhanced acoustic emission downsampling signals and acoustic emission multimodal spectrum information;

[0008] The acoustic emission downsampled signal is divided into multiple modes based on the acoustic emission multimodal spectrum information to obtain the frequency domain range of each mode;

[0009] The acoustic emission downsampled signal is decomposed according to the frequency domain range of each mode, and several decomposed sub-signals containing one mode information are selected from the decomposition results.

[0010] Each of the sub-signals containing one modal information is converted into the spectral information of the sub-signal;

[0011] Construct a first complex network model for each mode based on the spectral information of each decomposed sub-signal, and extract the degree distribution features of the first complex network model;

[0012] Calculate the degree distribution mutual information index matrix between different decomposed sub-signals based on the degree distribution characteristics;

[0013] The degree distribution mutual information index matrix is ​​transformed into a sparse adjacency matrix;

[0014] A second complex network model is constructed based on the sparse adjacency matrix, and the average weight feature of the second complex network model is extracted.

[0015] By fitting the changing trend of the average weighting characteristic, a fitting index for characterizing the quality of laser shock strengthening is obtained.

[0016] Further, the construction of the first complex network model for each mode based on the spectral information of each of the decomposed sub-signals includes:

[0017] Based on the spectral information of each of the decomposed sub-signals, a first complex network model for each modality is constructed using the visibility graph algorithm, specifically as follows:

[0018] The amplitude of the frequency points contained in each of the decomposed sub-signals is determined based on the spectral information of each of the decomposed sub-signals;

[0019] Based on the magnitude of the frequency points contained in the decomposed sub-signals, a visibility algorithm is used to determine the visibility between pairs of frequency points.

[0020] The adjacency matrix of each of the decomposed sub-signals is determined based on the visibility between the pairwise frequency points.

[0021] A first complex network model for each modality is constructed based on the adjacency matrix of each of the decomposed sub-signals;

[0022] The degree distribution features of the first complex network model are extracted using the following formula:

[0023] P(k)=N k / N

[0024] In the formula: P(k) represents the degree distribution characteristic; N is the number of nodes in the first complex network model; N k Let k be the number of nodes of degree k in the first complex network model.

[0025] Furthermore, the degree distribution mutual information index matrix I between different decomposed sub-signals is calculated based on the degree distribution characteristics, and the calculation formula is as follows:

[0026]

[0027]

[0028]

[0029] In the formula, α and β are any two different decomposed sub-signals, and their corresponding degree distribution characteristics are P(k) and P(k) respectively. [α] ) and P(k [β] );k [α] and k [β] These are the degree values ​​in α and β, respectively; I α,β P(k) is the degree distribution mutual information index matrix between α and β; [α] ,k [β] To find a value of degree k in α [α] The node and the search for the degree k in β [β] The joint probability of the nodes; For a degree of k in α [α] And in β, the degree is k [β] The number of nodes.

[0030] Furthermore, the elements of the sparse adjacency matrix A are:

[0031]

[0032] In the formula, A ij Let I be the value in the i-th row and j-th column of the sparse adjacency matrix A; ij is the value of the i-th row and j-th column in the degree distribution mutual information index matrix I; v is the preset threshold.

[0033] Furthermore, the extraction formula for the average weight feature of the second complex network model is as follows:

[0034]

[0035] In the formula, <k>ω represents the average weight feature; M is the number of nodes in the second complex network model; ij Let be the element in the i-th row and j-th column of the sparse adjacency matrix.

[0036] Further, the fitting of the changing trend of the average weighting feature to obtain a fitting index for characterizing the quality of laser shock strengthening includes:

[0037] By fitting the changing trend of the average weighting characteristic using a linear function, a fitting index for characterizing the quality of laser shock strengthening is obtained.

[0038] Further, the process of decomposing the acoustic emission downsampled signal according to the frequency domain range of each mode, and selecting several decomposed sub-signals containing one mode information from the decomposition results, includes:

[0039] Based on the frequency domain range of each mode, the acoustic emission downsampled signal is decomposed using the harmonic wavelet packet decomposition method, specifically as follows:

[0040] The bandwidth of harmonic wavelet packet decomposition is determined based on the frequency domain range of each mode;

[0041] The number of sub-signals containing at most one modal information is determined based on the bandwidth.

[0042] The acoustic emission downsampled signal is subjected to harmonic wavelet packet decomposition based on the number of decomposed sub-signals containing at most one modal information, and several decomposed sub-signals containing one modal information are selected from the decomposition results.

[0043] Furthermore, the real-time acquisition of the laser-strengthened acoustic emission downsampling signal and acoustic emission multimodal spectrum information includes:

[0044] Real-time acquisition of acoustic emission signals enhanced by laser shock;

[0045] Perform a Fourier transform on the acoustic emission signal to obtain the acoustic emission multimodal spectrum information;

[0046] The acoustic emission signal is downsampled according to Shannon's sampling theorem to obtain the downsampled acoustic emission signal.

[0047] Further, the step of converting each of the decomposed sub-signals containing one modal information into the spectral information of the decomposed sub-signals includes:

[0048] Perform a Fourier transform on each of the decomposed sub-signals containing one mode information to obtain the spectral information of the decomposed sub-signals.

[0049] An online monitoring device for acoustic emission of laser-strengthened mass includes:

[0050] The acquisition module is used to acquire the acoustic emission downsampling signal and acoustic emission multimodal spectrum information of laser shock enhancement in real time;

[0051] The segmentation module is used to perform multimodal segmentation on the acoustic emission downsampled signal based on the acoustic emission multimodal spectrum information to obtain the frequency domain range of each mode;

[0052] The decomposition module is used to decompose the acoustic emission downsampled signal according to the frequency domain range of each mode, and select several decomposed sub-signals containing one mode information from the decomposition results;

[0053] The first conversion module is used to convert each of the decomposed sub-signals containing one modal information into the spectral information of the decomposed sub-signals;

[0054] The first construction module is used to construct a first complex network model for each mode based on the spectral information of each decomposed sub-signal, and to extract the degree distribution features of the first complex network model.

[0055] The calculation module is used to calculate the degree distribution mutual information index matrix between different decomposed sub-signals based on the degree distribution characteristics.

[0056] The second transformation module is used to transform the degree distribution mutual information index matrix into a sparse adjacency matrix.

[0057] The second construction module is used to construct a second complex network model based on the sparse adjacency matrix and extract the average weight feature of the second complex network model.

[0058] The fitting module is used to fit the changing trend of the average weighting characteristics to obtain a fitting index for characterizing the quality of laser shock strengthening.

[0059] An apparatus includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the laser shock-enhanced mass acoustic emission online monitoring method.

[0060] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the laser shock-enhanced mass acoustic emission online monitoring method.

[0061] Compared with the prior art, the present invention has at least the following beneficial effects:

[0062] This invention provides a highly efficient and accurate method for monitoring acoustic emission in laser shock enhancement, effectively solving the problems of low monitoring accuracy caused by the weak correlation of existing feature definitions and susceptibility to environmental influences. On one hand, this invention introduces a complex network method to extract degree distribution features and average network weight features to measure the degree of distortion in the signal spectrum caused by processing quality, thus effectively addressing the problem of low correlation between existing features and quality. On the other hand, it decomposes the signal into multiple sub-signals through harmonic wavelet packet decomposition, and based on these sub-signals, constructs two-layer complex network models from both single-mode and multi-mode perspectives, effectively solving the problems of low information utilization, significant limitations, and low monitoring accuracy in existing methods. This invention comprehensively and completely explores the correlation between signal and processing quality. The method is simple, the physical meaning of the features is clear, it has good real-time performance, and possesses high reliability and robustness, providing an effective approach for online monitoring of laser shock enhancement.

[0063] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the specific embodiments of the present invention, the drawings used in the description of the specific embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0065] Figure 1 This is a technical flowchart of an online monitoring method for acoustic emission of laser-strengthened materials according to the present invention;

[0066] Figure 2 This is a schematic diagram of online monitoring of acoustic emission of laser-strengthened quality in an embodiment of the present invention;

[0067] Figure 3 The diagram shows the laser shock path, plate dimensions, and sensor installation in this embodiment of the invention, where (a) is a front view and (b) is a side view.

[0068] Figure 4 The following are the spectrum diagrams of the acoustic emission signal before and after downsampling in an embodiment of the present invention, where (a) is the spectrum diagram before downsampling and (b) is the spectrum diagram after downsampling;

[0069] Figure 5 The above are the time-domain and frequency-domain waveforms of the first 6 decomposed sub-signals after harmonic wavelet packet decomposition in an embodiment of the present invention.

[0070] Figure 6 The first complex network model with 7 modalities in this embodiment of the invention: (a) to (g) are modalities 1 to 7 respectively;

[0071] Figure 7 The average network weight under different energies in the embodiments of the present invention. <k>feature;

[0072] Figure 8 'a' represents the fitting parameter 'a' for different energies in this embodiment of the invention.

[0073] In the diagram: 1-Industrial control computer, 2-Data acquisition board, 3-Preamplifier, 4-Acoustic emission sensor, 5-Laser main control console, 6-Robotic arm, 7-Laser generator, 8-Workpiece to be impacted, 9-Absorption protective layer, 10-Constraint layer. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0075] Current methods for monitoring acoustic emission enhanced by laser shock mostly rely on the overall characteristics of the signal or the energy characteristics of a single mode to achieve quality monitoring. The features defined and extracted in existing methods are too simple, and the feature definitions have low correlation with the processing process, which has certain limitations. They are also easily affected by the environment in practical applications. Therefore, existing methods have the disadvantage of low monitoring accuracy.

[0076] As a specific embodiment of the present invention, combined with Figure 1 As shown, the present invention provides an online monitoring method for acoustic emission of laser shock-enhanced quality, which specifically includes the following steps:

[0077] S1. Real-time acquisition of laser-strengthened acoustic emission downsampling signals and acoustic emission multimodal spectrum information.

[0078] In this embodiment, the downsampled acoustic emission signal enhanced by laser shock is acquired, reducing the data length and improving the data processing speed. Acquiring the multimodal spectrum information of acoustic emission can intuitively reflect the modal components contained in the laser shock enhanced acoustic emission signal, and can also determine the maximum frequency component of the laser shock enhanced acoustic emission signal.

[0079] The preferred method for obtaining the laser-shock-enhanced acoustic emission downsampled signal and acoustic emission multimodal spectrum information is as follows:

[0080] Real-time acquisition of acoustic emission signals enhanced by laser shock;

[0081] Perform a Fourier transform on the acoustic emission signal to obtain the acoustic emission multimodal spectrum information;

[0082] The acoustic emission signal is downsampled according to Shannon's sampling theorem to obtain the downsampled acoustic emission signal.

[0083] This invention selects the relatively mature acoustic emission detection technology to achieve online monitoring of laser shock strengthening quality. Acoustic emission sensing technology can capture the elastic waves excited when the material undergoes plastic deformation in real time, and has a high sampling rate, making it particularly suitable for the high transient characteristics of laser shock strengthening. Therefore, the key parameters of the acquisition system are specified and briefly explained here. Figure 2 As shown, this embodiment uses an acoustic emission acquisition system to acquire the acoustic emission signal enhanced by laser shock in real time. In this embodiment, the acoustic emission acquisition system consists of an acoustic emission sensor 4, a preamplifier 3, a data acquisition board 2, and an industrial control computer 1 connected sequentially. The acoustic emission sensor 4 in the acoustic emission acquisition system is an RS-54A type wideband piezoelectric sensor. The DS2-8B series preamplifier 3 and data acquisition board 2 are used in the acoustic emission acquisition system to amplify, convert analog to digital, and acquire data of the acoustic emission signal, respectively. The preamplifier 3 has a preamplification gain of 40dB, and the data acquisition board 2 has a sampling rate of 3MHz to ensure no signal distortion. Compared to traditional resonant acoustic emission sensors, wideband piezoelectric sensors do not have a center frequency and have a larger frequency bandwidth, enabling the acquisition of more information. The sensitivity of the RS-54A type wideband piezoelectric sensor is 70dB±5dB.

[0084] The acoustic emission sensor 4 is mounted on the surface of the workpiece 8 to be impacted. Industrial coupling agent and fixtures are used to ensure that the acoustic emission sensor 4 and the workpiece 8 are in close contact. During the processing, the acoustic emission acquisition system collects the acoustic emission signal x(t) enhanced by laser shock in real time.

[0085] S2. Divide the acoustic emission downsampled signal into multiple modes according to the acoustic emission multimodal spectrum information to obtain the frequency domain range of each mode.

[0086] S3. Decompose the acoustic emission downsampled signal according to the frequency domain range of each mode, and select several decomposed sub-signals containing one mode information from the decomposition results.

[0087] Preferably, based on the frequency domain range of each mode, the acoustic emission downsampled signal is decomposed using the harmonic wavelet packet decomposition method, as follows:

[0088] The bandwidth of harmonic wavelet packet decomposition is determined based on the frequency domain range of each mode;

[0089] The number of sub-signals containing at most one modal information is determined based on the bandwidth, and the calculation formula is as follows:

[0090] j = f s / 2B

[0091] In the formula, j is the number of sub-signals; f s B is the sampling frequency of the acoustic emission downsampled signal; B is the bandwidth of the harmonic wavelet packet decomposition.

[0092] The acoustic emission downsampled signal is subjected to harmonic wavelet packet decomposition based on the number of decomposed sub-signals containing at most one modal information, and several decomposed sub-signals containing one modal information are selected from the decomposition results.

[0093] In this embodiment, the acoustic emission downsampled signal is decomposed using the harmonic wavelet packet decomposition method, which can ensure that the decomposed sub-signals have a strict box-shaped spectrum, making it particularly suitable for decomposing multimodal signals.

[0094] In this embodiment, the harmonic wavelet packet decomposition method selects harmonic wavelets as the basis functions, and its frequency domain expression is:

[0095]

[0096] Its time-domain expression is:

[0097]

[0098] In the formula, W m,n (ω) is the frequency domain expression; ω m,n (x) is the time-domain expression; m and n are scale parameters; ω is the frequency; and x is the time.

[0099] S4. Convert the decomposed sub-signal containing one modal information into the spectral information of the decomposed sub-signal.

[0100] Preferably, a Fourier transform is performed on each of the decomposed sub-signals to obtain the spectral information of the decomposed sub-signals.

[0101] S5. Construct a first complex network model for each mode based on the spectral information of each decomposed sub-signal, and extract the degree distribution features of the first complex network model.

[0102] Preferably, in this embodiment, based on the spectral information of each of the decomposed sub-signals, a first complex network model for each modality is constructed using a visual graph algorithm, as follows:

[0103] The amplitude of the frequency points contained in each of the decomposed sub-signals is determined based on the spectral information of each of the decomposed sub-signals;

[0104] Based on the magnitude of the frequency points contained in the decomposed sub-signals, a visibility algorithm is used to determine the visibility between pairs of frequency points.

[0105] In this embodiment, a visibility algorithm is used to determine the visibility between frequency points. The spectrum of the decomposed sub-signal is treated as a one-dimensional discrete sequence, and transformed into a complex network according to the following principle:

[0106] For any three points a, b, c in a one-dimensional discrete sequence, if two of the points (x, b, c) are (x, b, c)... a y a ) and (x b y b If they are mutually visible, then for any point (x) c y c ), where x a <x c <x b The following equation is satisfied:

[0107]

[0108] The adjacency matrix of each of the decomposed sub-signals is determined based on the visibility between the pairwise frequency points.

[0109] The first complex network model for each modality is constructed based on the adjacency matrix of each of the decomposed sub-signals.

[0110] Preferably, in this embodiment, the extraction formula for the degree distribution features of the first complex network model is as follows:

[0111] P(k)=N k / N

[0112] In the formula: P(k) represents the degree distribution characteristic; N is the number of nodes in the first complex network model; N k Let k be the number of nodes of degree k in the first complex network model.

[0113] S6. Calculate the degree distribution mutual information index matrix between different decomposed sub-signals based on the degree distribution characteristics.

[0114] Preferably, in this embodiment, the degree distribution mutual information index matrix I between different decomposed sub-signals is calculated using the following formula:

[0115]

[0116]

[0117]

[0118] In the formula, α and β are any two different decomposed sub-signals, and their corresponding degree distribution characteristics are P(k) and P(k) respectively. [α] ) and P(k [β] );k [α] and k [β] These are the degree values ​​in α and β, respectively; I α,β P(k) is the degree distribution mutual information index matrix between α and β; [α] ,k [β] To find a value of degree k in α [α] The node and the search for the degree k in β [β] The joint probability of the nodes; For a degree of k in α [α] And in β, the degree is k [β] The number of nodes.

[0119] S7. Transform the degree distribution mutual information index matrix into a sparse adjacency matrix.

[0120] Specifically, the elements of the sparse adjacency matrix A are:

[0121]

[0122] In the formula, A ij Let I be the value in the i-th row and j-th column of the sparse adjacency matrix A; ij is the value of the i-th row and j-th column in the degree distribution mutual information index matrix I; v is the preset threshold.

[0123] Preferably, in this embodiment, the degree distribution mutual information index matrix is ​​transformed into a sparse adjacency matrix by setting a threshold v. The specific method is as follows:

[0124] Based on the degree distribution mutual information index matrix, a limiting threshold v is introduced and determined. j and critical threshold v i ;

[0125] The limit threshold v j The critical threshold v is the minimum threshold before a complex network reaches a fully accessible graph. i The maximum threshold before a complex network reaches a disconnected graph; based on the threshold v j and critical threshold v i Determine the threshold v;

[0126] Based on the threshold v, the degree distribution mutual information index matrix is ​​transformed into a sparse adjacency matrix.

[0127] S8. Construct a second complex network model based on the sparse adjacency matrix, and extract the average weight feature of the second complex network model.

[0128] Preferably, in this embodiment, the average weight feature of the second complex network model is extracted using the following formula:

[0129]

[0130] In the formula, <k>ω represents the average weight feature; M is the number of nodes in the second complex network model; ij Let be the element in the i-th row and j-th column of the sparse adjacency matrix.

[0131] S9. Fit the changing trend of the average weighting characteristic to obtain the fitting index used to characterize the quality of laser shock strengthening.

[0132] Preferably, because of the average weighting characteristic <k>The average weighting characteristic increases monotonically with the change of the impact point. Therefore, the variation trend of the average weighting characteristic is fitted by the linear function y = ax + b to obtain the fitting index a for characterizing the quality of laser shock strengthening, thereby realizing online monitoring of the quality of laser shock strengthening.

[0133] This embodiment also provides an online monitoring device for acoustic emission of laser shock-enhanced quality, used to implement the above monitoring method, including:

[0134] The acquisition module is used to acquire the laser-strengthened acoustic emission downsampling signal and acoustic emission multimodal spectrum information in real time.

[0135] The partitioning module is used to partition the acoustic emission downsampled signal into multiple modes based on the acoustic emission multimodal spectrum information, so as to obtain the frequency domain range of each mode.

[0136] The decomposition module is used to decompose the acoustic emission downsampled signal according to the frequency domain range of each mode, and select several decomposed sub-signals containing one mode information from the decomposition results.

[0137] The first conversion module is used to convert each of the decomposed sub-signals containing one modal information into the spectral information of the decomposed sub-signals.

[0138] The first construction module is used to construct a first complex network model for each mode based on the spectral information of each decomposed sub-signal, and to extract the degree distribution features of the first complex network model.

[0139] The calculation module is used to calculate the degree distribution mutual information index matrix between different decomposed sub-signals based on the degree distribution characteristics.

[0140] The second transformation module is used to transform the degree distribution mutual information index matrix into a sparse adjacency matrix.

[0141] The second construction module is used to construct a second complex network model based on the sparse adjacency matrix and extract the average weight feature of the second complex network model.

[0142] The fitting module is used to fit the changing trend of the average weighting characteristics to obtain a fitting index for characterizing the quality of laser shock strengthening.

[0143] The following is a specific embodiment to more clearly explain the online monitoring method for acoustic emission of laser shock enhancement quality according to the present invention, as follows:

[0144] like Figure 2 The diagram shown illustrates the online monitoring of acoustic emission during laser shock peening in this embodiment. It mainly includes an industrial computer 1, a data acquisition board 2, a preamplifier 3, an acoustic emission sensor 4, a laser control console 5, a robotic arm 6, a laser generator 7, a workpiece to be impacted 8, an absorption protective layer 9, and a constraint layer 10. The industrial computer 1, data acquisition board 2, preamplifier 3, and acoustic emission sensor 4 are sequentially connected to form an acoustic emission signal acquisition system. The laser control console 5, robotic arm 6, and laser generator 7 form the laser shock peening system. During laser shock, the acoustic emission acquisition system captures the acoustic emission signals inside the material in real time.

[0145] The parameters set in this embodiment include: a sampling frequency of 3MHz, a preamplifier gain of 40dB to ensure more acoustic emission signals are collected, a laser energy of 1J, 2J, and 3J, a single impact, a spot size of 3mm, and the use of transparent water as a constraint layer and black tape as an absorption and protection layer. Figure 3 As shown, the laser impact path is a straight line, impacting 25 points consecutively with an overlap rate of 50%. By analyzing the changing trend of the average network weight of the 25 impact points under different energies, linear fitting is performed to obtain fitting parameters, which serve as the final quality monitoring index. The dimensions of the plate to be impacted are: length 250mm, width 60mm, and height 4mm. The distance between the acoustic emission sensor and the first impact point is 50mm, as detailed below. Figure 3 As shown.

[0146] In this embodiment, firstly, based on the experimental equipment and parameters, the acoustic emission signal enhanced by laser shock is acquired in real time. Then, a Fourier transform is performed on the acoustic emission signal to obtain the multimodal spectral information of the acoustic emission, such as... Figure 4 As shown in (a), the acoustic emission multimodal spectrum information reveals that the maximum frequency component of the acoustic emission signal is 500kHz, exhibiting significant multimodal characteristics. It contains multiple modes, with the first seven modes being: 31kHz, 49kHz, 70kHz, 91kHz, 115kHz, 139kHz, and 164kHz. Therefore, according to Shannon's sampling theorem, the acoustic emission signal is downsampled to obtain the downsampled acoustic emission signal. The spectrum of the downsampled acoustic emission signal is shown below. Figure 4 As shown in (b). At this time, the sampling frequency f of the acoustic emission downsampling signal s It is 1MHz.

[0147] In this embodiment, the acoustic emission downsampling signal is divided into multiple modes according to the acoustic emission multimodal spectrum information, and the number of modes is determined to be 7. The frequency domain range of each mode is shown in Table 1.

[0148] Table 1 Frequency domain range for each mode

[0149]

[0150]

[0151] In this embodiment, the bandwidth B = 10kHz for harmonic wavelet packet decomposition is determined based on the frequency domain range of each mode. Based on the bandwidth B and the formula j = f... s After determining the harmonic wavelet packet decomposition in / 2B, the number of sub-signals j containing at most one modal information is 50, and the acoustic emission downsampled signal is decomposed as follows: Figure 5 The table shows the time-domain and frequency-domain waveforms of the first six decomposed sub-signals after harmonic wavelet packet decomposition, each containing at most one modal information. Table 2 shows seven decomposed sub-signals containing one modal information selected from these sub-signals.

[0152] Table 2 contains decomposed sub-signals containing modal information.

[0153]

[0154] In this embodiment, the decomposed sub-signals in Table 2 are converted into their spectral information. Based on the spectral information of each decomposed sub-signal, a first complex network model for each mode is constructed, such as... Figure 6 As shown, the degree distribution feature P(k) of the first complex network model is extracted. The degree distribution mutual information index matrix between different decomposed sub-signals is calculated based on the degree distribution feature.

[0155] In this embodiment, the limiting threshold v of the degree distribution mutual information index matrix is ​​calculated. j and critical threshold v i The final threshold v is determined, and the degree distribution mutual information matrix I is transformed into a sparse adjacency matrix A. The threshold values ​​for laser-enhanced acoustic emission signals of different energies in this embodiment are shown in Table 3.

[0156] Table 3 Thresholds for acoustic emission signals at different laser energies

[0157]

[0158] In this embodiment, a second complex network model is constructed based on the sparse adjacency matrix A, and the average weight feature of the second complex network model is extracted. <k>,like Figure 7 As shown. From Figure 7 As can be seen from this, the average weight characteristics of the second complex network model for different laser energies are... <k>The trends of change differ to some extent, but they all gradually increase as the point of impact increases.

[0159] In this embodiment, the variation trend of the average weighting characteristic is fitted using a linear function y = ax + b to obtain the fitting index 'a' characterizing the laser shock strengthening quality, thereby achieving online monitoring of the laser shock strengthening quality. Table 4 shows the fitting parameter 'a' for different laser energies, and its variation trend is as follows: Figure 8 As shown, it can be seen that the fitting parameter feature a gradually decreases as the energy increases.

[0160] Table 4 Fitting parameters for different laser energies

[0161]

[0162] As demonstrated by the above experiments and embodiments, this invention fully utilizes the multimodal information of laser-shock-enhanced acoustic emission signals. It employs a complex network method to construct a two-layer complex network model, extracts the fitting parameter feature 'a', and ultimately achieves the goal of online monitoring of laser-shock-enhanced quality. The proposed method is simple, with clear physical meaning of its features, exhibiting high reliability and robustness, and provides an effective approach for realizing online monitoring of laser-shock-enhanced processes.

[0163] In one embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used to implement the operation of an online monitoring method for acoustic emission of laser-strengthened quality.

[0164] In one embodiment of the present invention, a laser shock-enhanced quality acoustic emission online monitoring method, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data.

[0165] The computer storage medium can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs)), optical storage (e.g., CDs, DVDs, BDs, HVDs), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0166] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0167] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0168] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0169] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0170] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.< / k> < / k> < / k> < / k> < / k> < / k>

Claims

1. A method for online monitoring of laser shock peening quality acoustic emission, characterized in that, The method comprises the following steps: real-time acquisition of acoustic emission signals of laser shock peening; Fourier transform of the acoustic emission signals to obtain acoustic emission multi-modal spectrum information; down-sampling processing of the acoustic emission signals according to the Shannon sampling theorem to obtain acoustic emission down-sampled signals; multi-modal division of the acoustic emission down-sampled signals according to the acoustic emission multi-modal spectrum information to obtain the frequency domain range of each mode; based on the frequency domain range of each mode, the acoustic emission down-sampled signals are decomposed by using a harmonic wavelet packet decomposition method, specifically: determining the bandwidth of harmonic wavelet packet decomposition according to the frequency domain range of each mode; determining the number of decomposition sub-signals containing at most one mode information according to the bandwidth; harmonic wavelet packet decomposition of the acoustic emission down-sampled signals according to the number of decomposition sub-signals containing at most one mode information, and selecting a plurality of decomposition sub-signals containing one mode information from the decomposition results; converting each of the decomposition sub-signals containing one mode information into frequency spectrum information of the decomposition sub-signals; constructing a first complex network model of each mode according to the frequency spectrum information of each of the decomposition sub-signals, and extracting a degree distribution feature of the first complex network model; calculating a degree distribution mutual information index matrix between different decomposition sub-signals according to the degree distribution feature; converting the degree distribution mutual information index matrix into a sparse adjacency matrix; constructing a second complex network model according to the sparse adjacency matrix, and extracting an average weight degree feature of the second complex network model; fitting the change trend of the average weight degree feature to obtain a fitting index for representing the quality of laser shock peening.

2. A method for online monitoring of laser shock peening quality acoustic emission according to claim 1, characterized in that, The construction of the first complex network model of each mode according to the frequency spectrum information of each of the decomposition sub-signals comprises: based on the frequency spectrum information of each of the decomposition sub-signals, a visual graph algorithm is used to construct the first complex network model of each mode, specifically: determining the amplitude of the frequency points contained in each of the decomposition sub-signals according to the frequency spectrum information of each of the decomposition sub-signals; based on the amplitude of the frequency points contained in each of the decomposition sub-signals, a visual graph algorithm is used to determine the visibility between two frequency points; based on the visibility between two frequency points, the adjacency matrix of each of the decomposition sub-signals is determined; based on the adjacency matrix of each of the decomposition sub-signals, the first complex network model of each mode is constructed; the extraction formula of the degree distribution feature of the first complex network model is: In the formula: is the degree distribution characteristic; is the number of nodes in the first complex network model; is the number of nodes in the first complex network model with degree is the number of nodes in the first complex network model with degree 3. A method of online monitoring of laser shock peening quality acoustic emission according to claim 2, characterized in that, The degree distribution mutual information index matrix between different decomposition sub-signals is calculated according to the degree distribution characteristics The calculation formula is: wherein are the degree distribution characteristics of the corresponding two different sub-signals and ; and are the degree values in ; is the degree distribution mutual information index matrix between and ; is the joint probability of finding a node with degree in and a node with degree in ; is the number of nodes with degree in and degree in .

4. A method of online monitoring of laser shock peening quality acoustic emission according to claim 3, characterized in that, The sparse adjacency matrix whose elements are: wherein is a sparse adjacency matrix in the first row, the column value; is a degree distribution mutual information index matrix in the first row, the column value; is a predetermined threshold.

5. A method of online monitoring of laser shock peening quality acoustic emission according to claim 1, characterized in that, the extraction formula of the average weight degree feature of the second complex network model is: In the formula, is an average weight degree feature; is the number of nodes of the second complex network model; is the element in the sparse adjacency matrix in the i-th row and the j-th column. is the element in the sparse adjacency matrix in the i-th row and the j-th column. is the element in the sparse adjacency matrix in the i-th row and the j-th column.

6. A method of online monitoring of laser shock peening quality acoustic emission according to claim 1, characterized in that, the fitting of the change trend of the average weight degree feature to obtain a fitting index for representing the quality of laser shock peening comprises: a linear function is used to fit the change trend of the average weight degree feature to obtain a fitting index for representing the quality of laser shock peening.

7. A method of online monitoring of laser shock peening quality acoustic emission according to claim 1, characterized in that, The conversion of each of the decomposition sub-signals containing one mode information into frequency spectrum information of the decomposition sub-signals comprises: Fourier transform of each of the decomposition sub-signals containing one mode information to obtain frequency spectrum information of the decomposition sub-signals.

8. A laser shock peening quality acoustic emission online monitoring device, characterized in that, A laser shock peening quality acoustic emission online monitoring method for realizing any one of claims 1-7, comprising: an acquisition module for acquiring acoustic emission down-sampling signals and acoustic emission multi-modal spectrum information of laser shock peening in real time; a division module for multi-modal division of the acoustic emission down-sampling signals according to the acoustic emission multi-modal spectrum information, to obtain a frequency domain range of each mode; a decomposition module for decomposition of the acoustic emission down-sampling signals according to the frequency domain range of each mode, and selection of a plurality of decomposition sub-signals containing one mode information from the decomposition result; a first conversion module for conversion of each of the decomposition sub-signals containing one mode information into spectrum information of the decomposition sub-signals; a first construction module for construction of a first complex network model of each mode according to the spectrum information of each of the decomposition sub-signals, and extraction of a degree distribution feature of the first complex network model; a calculation module for calculation of a degree distribution mutual information index matrix between different decomposition sub-signals according to the degree distribution feature; a second conversion module for conversion of the degree distribution mutual information index matrix into a sparse adjacency matrix; a second construction module for construction of a second complex network model according to the sparse adjacency matrix, and extraction of an average weight degree feature of the second complex network model; a fitting module for fitting of a variation trend of the average weight degree feature, to obtain a fitting index for representing laser shock peening quality.

Images