Method and device for locating structure impact in variable temperature environment based on multiple signal classification algorithm

Abstract

The application discloses a structure shock positioning method in a variable-temperature environment based on a multiple signal classification algorithm, which comprises the following steps: obtaining a shock positioning image of a single-frequency shock stress wave signal at a wave number initial value based on the multiple signal classification algorithm; performing iteration in a wave number iteration interval and a wave number iteration step length, and obtaining a shock positioning image corresponding to each wave number in the wave number iteration interval; marking a maximum pixel value of the shock positioning image corresponding to each wave number in the wave number iteration interval; selecting a shock positioning image with the maximum maximum pixel value in all shock positioning images as a representative shock positioning image, and determining a grid point position of the maximum pixel value of the representative shock positioning image as a shock position. The application further discloses a corresponding device. The method and device provided by the application do not need to measure wave field information of a structure in a variable-temperature environment in advance, and the practical value of the MUSIC algorithm in engineering application is improved.

Description

Technical Field

[0001] This invention belongs to the field of structural health monitoring, specifically relating to a structural impact localization method and device based on Multi-Signal Classification (MUSIC). Background Technology

[0002] With the development of technologies such as smart materials, structural optimization, and advanced manufacturing, the demand for lightweight aircraft is gradually increasing. Compared with conventional metal materials, composite materials have advantages such as high specific modulus, high specific strength, radiation resistance, fatigue resistance, and strong designability, and have been widely used and developed in the aerospace field. The amount of composite materials used has also become an important indicator for measuring advanced aircraft. However, composite materials are very sensitive to low-speed impacts. During the manufacturing, service, and maintenance of aerospace composite materials, they are inevitably subjected to external impact events such as falling maintenance tools, bird strikes, hail, and runway debris, which can lead to invisible micro-damage such as delamination and cracks within the composite materials, resulting in a decrease in structural load-bearing capacity and seriously threatening the service safety of aircraft. Therefore, impact monitoring has become a very important topic in current structural health monitoring.

[0003] Lamb-based structural health monitoring methods, with their wide monitoring range and high sensitivity, are considered one of the most promising approaches. Due to the ease of arrangement, fewer leads, and orientation estimation capabilities of dense sensor arrays, array signal processing methods are increasingly being incorporated into structural health monitoring. The multi-signal classification impact localization method is a typical array signal processing and feature search algorithm that utilizes the orthogonality between the noise and signal subspaces to construct a spectral function to estimate the direction and location of the impact source.

[0004] The MUSIC impact localization method applied to dense piezoelectric arrays and Lamb waves requires prior wavenumber calculation, direction vector construction, eigenvalue decomposition of the covariance matrix, acquisition of the signal and noise subspaces, and finally spectral peak search to achieve impact location estimation. However, the temperature environment in which aircraft operate varies greatly, and wavenumbers are affected by temperature changes. Some methods pre-calculate the group velocity of narrowband Lamb waves using full-field measurements or elastic wave theory, and then calculate the corresponding wavenumber. However, these methods lack adaptability and robustness in varying temperature environments. Therefore, to better meet the needs of engineering applications for impact localization algorithms, it is highly significant to make the MUSIC method adaptable to varying temperature environments. Summary of the Invention

[0005] To address the shortcomings of the existing technology, the present invention aims to provide a structural impact location method based on a Multi-Signal Classification (MUSIC) algorithm under varying temperature conditions, thereby achieving adaptability in varying temperature environments without the need for prior wavenumber measurement. The method includes the following steps: constructing a Cartesian coordinate system in the monitoring area of ​​the structure; arranging a sensor array in the monitoring area; marking the position coordinates of the sensor array in the Cartesian coordinate system and recording the impact stress wave signals acquired by the sensor array; extracting a single-frequency impact stress wave signal from the impact stress wave signal; setting a wavenumber iteration interval and an iteration wavenumber, the wavenumber iteration interval including an initial wavenumber value; obtaining an impact location image of the single-frequency impact stress wave signal at the initial wavenumber value based on the MUSIC algorithm; iterating using the set wavenumber iteration interval and wavenumber iteration step size to obtain an impact location image corresponding to each wavenumber within the wavenumber iteration interval; marking the maximum pixel value of the impact location image corresponding to each wavenumber within the wavenumber iteration interval; selecting the impact location image with the largest maximum pixel value among all impact location images as a representative impact location image, and determining the grid point location where the maximum pixel value of the representative impact location image is located as the impact location.

[0006] In some embodiments, obtaining the impact positioning image corresponding to each wavenumber within the wavenumber iteration interval includes: constructing a covariance matrix for the extracted single-frequency signal, performing eigenvalue decomposition to obtain the signal subspace and noise subspace of each single-frequency signal; dividing the monitoring area into a grid, calculating the distance difference between the distance from each grid point to each sensor and the distance to a reference sensor, and establishing an array direction vector with the wavenumber and the distance difference as variables; using the initial wavenumber value in the wavenumber iteration interval, the array direction vector, and the noise subspace to calculate the spatial spectrum of all grid points, and using the spatial spectrum of all grid points as pixel values ​​to image and obtain the impact positioning image of the entire monitoring area under the initial wavenumber value.

[0007] In some embodiments, the single-frequency impact stress wave signal in the impact stress wave signal is extracted using short-time Fourier transform.

[0008] In some embodiments, the iteration with the set wavenumber iteration interval and wavenumber iteration step size includes iterating from the initial wavenumber value to the final wavenumber value of the wavenumber iteration interval.

[0009] In some embodiments, based on the single-frequency shock stress wave signal, the data covariance matrix R is calculated using maximum likelihood estimation of a finite number of data points: eigenvalue decomposition is performed on the covariance matrix R to obtain M eigenvalues. and the corresponding feature vector The eigenvalues ​​are arranged in order of magnitude. The largest eigenvalues, equal to the number D of signal sources, and their corresponding eigenvectors are divided into the signal part. The remaining smaller eigenvalues, MD, and their corresponding eigenvectors are divided into the noise part. The signal subspace is constructed using the eigenvectors of the signal part. Let Vs be the feature vectors of the noise component; construct the noise subspace. , denoted as V N The eigenvalue decomposition of the covariance matrix R can be expressed as: In the formula, and These represent the larger eigenvalues ​​corresponding to the signal subspace and the smaller eigenvalues ​​corresponding to the noise subspace, respectively. for Hermitian transpose, for Hermitian transpose.

[0010] In some embodiments, the set wavenumber interval [k] min, k max The initial value k min As the wave number k, based on the distance difference r i (x, y) and wavenumber k min Construct array direction vector ; Utilizing the noise subspace V N and the array direction vector Calculate the spatial spectrum corresponding to the search grid points. Using the grid point division intervals Δx and Δy as search step sizes, search for all grid points (x, y) in the monitoring area; calculate the spatial spectrum over the entire monitoring area, and use the spatial spectrum as pixel values ​​to obtain the wavenumber k. min Impact positioning image I(k) min (x, y), where the location of the spatial spectral peak is the impact position at wavenumber k. .

[0011] In some embodiments, the iteration wave number k is determined according to the set wave number iteration step size, and the equation is... Substitution Impact localization imaging is performed by iteratively calculating the pixel values ​​of all grid points within the monitoring area at different wavenumbers k until... When the iteration terminates, the impact positioning images I(k, x, y) corresponding to different wavenumbers within the wavenumber interval are obtained; and the impact positioning image with the largest maximum pixel value in the impact positioning images I(k, x, y) is taken as the representative impact positioning image.

[0012] In some embodiments, the wavenumber iteration interval is associated with the single frequency.

[0013] Other embodiments of this application provide a structural impact positioning device under varying temperature conditions based on a multi-signal classification algorithm. The device includes a memory and a processor. The memory stores computer program code, and the processor is configured to execute the computer code to perform the following: recording impact stress wave signals from a sensor array arranged in a monitoring area; constructing a Cartesian coordinate system in the monitoring area and marking the position coordinates of the sensor array in the Cartesian coordinate system; extracting a single-frequency signal from the impact stress wave signal; setting a wavenumber iteration interval and an iteration wavenumber, the wavenumber iteration interval including an initial wavenumber value; obtaining an impact positioning image of the single-frequency impact stress wave signal at the initial wavenumber value based on the multi-signal classification algorithm; iterating with the set wavenumber iteration interval and wavenumber iteration step size to obtain an impact positioning image corresponding to each wavenumber within the wavenumber iteration interval; marking the maximum pixel value of the impact positioning image corresponding to each wavenumber within the wavenumber iteration interval; selecting the impact positioning image with the largest maximum pixel value among all impact positioning images as a representative impact positioning image, and determining the grid point location where the maximum pixel value of the representative impact positioning image is located as the impact location.

[0014] The beneficial technical effects of this invention are as follows: The method and device provided in this application do not require prior measurement of the wave field information of the structure under varying temperature conditions, which makes the impact positioning method temperature adaptive and improves the practical value of the MUSIC algorithm in engineering applications. Attached Figure Description

[0015] Figure 1A This is a flowchart of the first part of the adaptive wavenumber MUSIC method for structural impact localization under varying temperature environments according to this application.

[0016] Figure 1B This is a flowchart of the second part of the adaptive wavenumber MUSIC method for structural impact localization under varying temperature environments according to this application.

[0017] Figure 2 This is a schematic diagram of the adaptive wavenumber MUSIC method for structural impact localization under varying temperature environments according to this application.

[0018] Figure 3 This is a schematic diagram of the composite material structure and sensor array arrangement according to this application;

[0019] Figure 4A The time-domain waveform of the impulse response signal of array element 1 according to the method of this application is shown.

[0020] Figure 4B The spectrum of the impulse response signal of array element 1 according to the method of this application;

[0021] Figure 5 The method of this application is a single-frequency shock stress wave signal with a center frequency of 3kHz for array element 1.

[0022] Figure 6A This is a schematic diagram of the final impact positioning result according to an embodiment of this application, using angle and distance as coordinates;

[0023] Figure 6B This is a schematic diagram of the final impact positioning result according to an embodiment of this application, using length and width as coordinates;

[0024] Figure 7A This is a schematic diagram showing the results of an adaptive impact positioning method in a variable temperature environment according to an embodiment of this application, with angle and distance as coordinates.

[0025] Figure 7B This is a schematic diagram showing the results of the adaptive impact positioning method under varying temperature conditions according to an embodiment of this application, with length and width as coordinates. Detailed Implementation

[0026] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0027] The specific structural and functional details disclosed herein are merely representative and are intended to describe exemplary embodiments of this application. However, this application may be implemented in many alternative forms and should not be construed as being limited solely to the embodiments set forth herein.

[0028] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments. Unless the context clearly indicates otherwise, the singular forms “a” and “an” as used herein are also intended to include the plural. It should also be understood that the terms “comprising” and / or “including” as used herein specify the presence of the stated features, integers, steps, operations, units, and / or components, without excluding the presence or addition of one or more other features, integers, steps, operations, units, components, and / or combinations thereof.

[0030] It should also be mentioned that in some alternative implementations, the functions / actions / steps mentioned may occur in a different order than those shown in the figures. For example, depending on the functions / actions / steps involved, the two figures shown successively may actually be executed substantially simultaneously or sometimes in reverse order.

[0031] In one embodiment of the present invention, a storage medium is also provided, the storage medium storing computer program instructions, the computer program instructions being executed according to the sleep aid control method described in Embodiment 1 or Embodiment 2.

[0032] like Figure 1A , Figure 1B As shown, the structural impact localization method based on a multi-signal classification algorithm under varying temperature conditions in this application can specifically include the following steps:

[0033] Step 1: Construct a rectangular coordinate system, mark the position coordinates of the sensor array, and record the impact stress wave signal.

[0034] like Figure 2 As shown, a rectangular coordinate system is established on the monitoring area of ​​the monitored structure, with the length direction of the structure as the coordinate system. The positive axis direction is taken as the width direction of the structure. Positive axis direction. Arrange a sensor array with M elements, and label sensor S. i The position coordinates are (Sx i Sy i (i=1,2,…,M). The system monitors impact events on the structure and acquires the impact stress wave signal from the sensor array, denoted as X. i (t).

[0035] Step 2: Extract the narrowband wave signal of a single frequency as a single-frequency shock stress wave signal.

[0036] Short-Time Fourier Transform (STFT) possesses excellent local time-frequency domain analysis capabilities and is often used in impact signal processing. Therefore, this embodiment utilizes STFT to extract a single-frequency impact stress wave signal with a specific center frequency from a broadband impact stress wave signal. Signal X i The expression for the short-time Fourier transform of (t) is:

[0037] (1)

[0038] In the formula, Indicates frequency, Represents a time window function. This indicates the center position of the window function when the window is moved to the beginning of the signal. The width of the window function is selected based on the characteristics of the shock stress wave signal, and the center frequency of interest is chosen from the time-spectrum graph. Its corresponding amplitude in the time-frequency spectrum The center frequency is Impact stress wave signal C i (t).

[0039] Step 3: Set the initial parameters for the adaptive wavenumber MUSIC impact location algorithm.

[0040] Setting the iterative parameters for the adaptive wavenumber MUSIC impact localization algorithm includes setting the initial wavenumber k of the impact stress wave iteration. min Termination value k max The iterative interval formed And the iteration step size Δk.

[0041] Step 4: Obtain the impact positioning image based on the adaptive wavenumber MUSIC impact positioning algorithm.

[0042] Based on the single-frequency shock stress wave signal obtained from the actual acquisition and signal processing described above, the data covariance matrix R is calculated using maximum likelihood estimation of a finite number of data points:

[0043] (2)

[0044] In the formula, for The Hermitian transpose, where Y is the sampling length of the signal.

[0045] Eigenvalue decomposition of the covariance matrix R yields M eigenvalues. and the corresponding feature vector The eigenvalues ​​are arranged in order of magnitude. The largest eigenvalues, equal to the number D of signal sources, and their corresponding eigenvectors are designated as the signal component. The remaining smaller eigenvalues, MD, and their corresponding eigenvectors are designated as the noise component. A signal subspace is constructed using the eigenvectors of the signal component. Let Vs be the feature vectors of the noise component. The noise subspace is constructed from these feature vectors. , denoted as V N Therefore, the eigenvalue decomposition of the covariance matrix R can be expressed as:

[0046] (3)

[0047] In the formula, and These are the larger eigenvalues ​​corresponding to the signal subspace and the smaller eigenvalues ​​corresponding to the noise subspace, respectively.

[0048] The monitored area of ​​the structure is divided into grid points, where each grid point represents a potential impact location. The starting coordinates of the grid point search are set to (x0, y0), and this point is used as the assumed impact source (x, y).

[0049] (4)

[0050] Calculate the distance S from this point to each sensor. i The distance, denoted as L i (x0, y0):

[0051] (5)

[0052] Calculate the distance from the assumed impact source (x, y) to the sensor S. i The distance difference between the reference sensor S1 and the reference sensor S1 is denoted as r. i (x, y), where the reference sensor is an arbitrarily selected sensor in the sensor array, for example, sensor S1 can be used as the reference sensor:

[0053] (6)

[0054] The initial value k of the wavenumber interval set in step 3 min As wave number k:

[0055] (7)

[0056] Based on the distance difference r i (x, y) and wavenumber k min Construct array direction vector :

[0057] (8)

[0058] In the formula, j is the imaginary unit.

[0059] Using the noise subspace V N and array direction vector Calculate and search the spatial spectrum corresponding to a grid point, denoted as P. music (k min The spatial spectrum expression is as follows: (x, y)

[0060] (9)

[0061] Using the grid point division intervals Δx and Δy as the search step size, search for all grid points (x, y) in the monitoring area:

[0062] (10)

[0063] Calculate the spatial spectrum over the entire monitoring area using the steps described above, and use the spatial spectrum as pixel values ​​to obtain the wavenumber k. min MUSIC impact localization image I(k) min The location of the spatial spectral peak (x, y) is the impact location at wavenumber k. r (k),y r (k)):

[0064] (11)

[0065] Step 5: Iterate the wavenumber to achieve adaptive wavenumber impact positioning.

[0066] Impact imaging is performed on each point in the monitoring area within a reasonable wavenumber range. The correct wavenumber under the monitoring environment can be determined based on the maximum pixel value of the image under different wavenumbers. A spatial spectral peak will be formed under the correct wavenumber, meaning that the wavenumber under the monitoring environment can be retrieved without prior measurement. This process is called adaptive wavenumber.

[0067] Since this method does not require prior measurement of the full-field wavenumber information under varying temperature conditions, the iteration wavenumber k is determined according to the wavenumber iteration step size set in step 3:

[0068] (12)

[0069] Substituting equation (11) into equation (8), the pixel values ​​of all grid points within the monitoring area at different wavenumbers k are iteratively calculated for impact localization imaging. Until... When the iteration terminates, the MUSIC impact location image I(k, x, y) corresponding to different wavenumbers within the wavenumber interval is obtained. Only with the actual wavenumber and impact location of a single-frequency Lamb wave can the optimal impact location image result be obtained. Therefore, the predicted impact location (x, y) is determined by the grid point location containing the maximum pixel value among all MUSIC impact location images. r ,y r ), and its corresponding wavenumber k r This is the correct wavenumber at that temperature:

[0070] (13)

[0071] Adaptive wavenumber MUSIC impact localization image .

[0072] Example

[0073] like Figure 3The diagram shows the arrangement of the carbon fiber composite plate 100, the sensor array 200, and the passive impact monitoring system 300. The carbon fiber composite plate 100 has dimensions of 400mm × 400mm × 2mm and a layup of [0 / 90]s5. The sensor array 200 can be formed by an array of piezoelectric sensors, which can be P-51 type piezoelectric ceramic discs with a diameter of 8 mm and a thickness of 0.25 mm. Following step 1 above, a rectangular coordinate system is established with the lower left corner of the carbon fiber composite plate 100 as the origin. The coordinates of the piezoelectric sensor array 200 and the impact source are recorded. An impact event is triggered at a position on the carbon fiber composite plate 100 with coordinates (200mm, 300mm). The passive dynamic signal monitoring system 300 collects and records the impact stress wave signal received by the piezoelectric sensors. The system sampling rate is set to 200kHz. Figure 4A The image shows a time-domain plot of a typical impact stress wave signal received by the No. 1 piezoelectric sensor. Figure 4B The image shows the normalized spectrum of a typical impact stress wave signal received by the No. 1 piezoelectric sensor at room temperature (25 degrees Celsius). The spectrum indicates that the main frequency components of the impact stress wave signal are in the range of 0-10 kHz. Following step 2 above, the short-time Fourier transform is used to extract... Figure 5 The shock stress wave signal shown has a center frequency of 3kHz. Following step 3 above, the wave number iteration range for the shock stress wave propagating in the tested structure is set to [0.01, 0.06]. The iteration step size is 0.01 rad / m. Following the above steps, the result is as follows: Figure 6A , Figure 6B The impact localization imaging results shown are as follows, in which Figure 6A Using angle and distance as coordinates, Figure 6B Using length and width as coordinates, where dark "×" represents the actual impact location coordinates and light "×" represents the impact location coordinates predicted by the algorithm. The structure's temperature was increased to 35°C by irradiating it with two halogen lamps on both sides of plate 100, and the above steps were repeated to obtain the following... Figure 7A , Figure 7B The positioning results are shown. Figure 7A Using angle and distance as coordinates, Figure 7B The coordinates are length and width. "○" represents the actual impact location coordinates, and the light-colored "Δ" represents the impact location coordinates predicted by the algorithm.

[0074] The methods for extracting single-frequency signals in this application also include Fourier transform, autocorrelation function, fast Fourier transform, peak detection, and Kalman filtering.

[0075] In a typical configuration of this invention, the storage medium includes permanent and non-permanent, removable and non-removable media, which can be used to store information by any method or technology. The information can be computer-readable instructions, data structures, program devices, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, read-only optical disc (CD-ROM), digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0076] In one embodiment of the present invention, a computing device is also provided, comprising: a memory for storing computer program instructions and a processor for executing the computer program instructions, wherein when the computer program instructions are executed by the processor, the computing device is triggered to execute the method described in Embodiment 1 or Embodiment 2 of the present invention. In a typical configuration of the present invention, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0077] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0078] The computing device referred to in this invention includes, but is not limited to, any electronic product that can interact with a user (e.g., through a touchpad), such as smartphones, tablets, and other mobile electronic products. The mobile electronic product can use any operating system, such as Android or iOS.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the present invention. Those skilled in the art should understand that any changes, modifications, additions or substitutions made within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A structural impact localization method based on a multi-signal classification algorithm under varying temperature conditions, characterized in that: include A Cartesian coordinate system is constructed in the monitoring area of ​​the structure, a sensor array is arranged in the monitoring area, the position coordinates of the sensor array in the Cartesian coordinate system are marked, and the impact stress wave signal acquired by the sensor array is recorded. Extract a single-frequency impact stress wave signal from the impact stress wave signal; Define the wavenumber iteration interval and the iterative wavenumber, wherein the wavenumber iteration interval includes the initial wavenumber value; The impact positioning image of the single-frequency impact stress wave signal under the initial wave number value is obtained based on the multi-signal classification algorithm. The wavenumber iteration interval and wavenumber iteration step size are set to perform iteration to obtain the impact positioning image corresponding to each wavenumber within the wavenumber iteration interval. Mark the maximum pixel value of the impact positioning image corresponding to each wavenumber in the wavenumber iteration interval; select the impact positioning image with the largest maximum pixel value among all impact positioning images as the representative impact positioning image, and determine the grid point position where the maximum pixel value of the representative impact positioning image is located as the impact position; The step of obtaining the impact localization image corresponding to each wavenumber within the wavenumber iteration interval includes: A covariance matrix is ​​constructed for the extracted single-frequency signals, and eigenvalue decomposition is performed to obtain the signal subspace and noise subspace of each single-frequency signal; The monitoring area is divided into a grid, and the distance difference between each grid point and each sensor is calculated and the distance to the reference sensor is calculated. An array direction vector is established with wavenumber and the distance difference as variables. Using the initial wavenumber value in the wavenumber iteration interval, the array direction vector, and the noise subspace, the spatial spectrum of all grid points is calculated, and the spatial spectrum of all grid points is used as pixel value imaging to obtain the impact positioning image of the entire monitoring area under the initial wavenumber value.

2. The structural impact localization method based on multi-signal classification algorithm under varying temperature conditions according to claim 1, characterized in that: The single-frequency impact stress wave signal in the impact stress wave signal is extracted using short-time Fourier transform.

3. The structural impact localization method based on multi-signal classification algorithm under varying temperature conditions according to claim 1, characterized in that: The iteration using the set wavenumber iteration interval and wavenumber iteration step size includes iterating from the initial wavenumber value to the final wavenumber value of the wavenumber iteration interval.

4. The structural impact localization method based on multi-signal classification algorithm under varying temperature conditions according to claim 3, characterized in that: Based on the single-frequency impact stress wave signal, the data covariance matrix R is calculated using maximum likelihood estimation with a finite number of data points. Eigenvalue decomposition of the covariance matrix R yields M eigenvalues. and the corresponding feature vector The eigenvalues ​​are arranged in order of magnitude. The largest eigenvalues, equal to the number D of signal sources, and their corresponding eigenvectors are divided into the signal part. The remaining smaller eigenvalues, MD, and their corresponding eigenvectors are divided into the noise part. The signal subspace is constructed using the eigenvectors of the signal part. Let Vs be the feature vectors of the noise component; construct the noise subspace. , denoted as V N The eigenvalue decomposition of the covariance matrix R can be expressed as: In the formula, and These represent the larger eigenvalues ​​corresponding to the signal subspace and the smaller eigenvalues ​​corresponding to the noise subspace, respectively. for Hermitian transpose, for Hermitian transpose.

5. The structural impact localization method based on a multi-signal classification algorithm under varying temperature conditions according to claim 4, characterized in that: The set wavenumber interval [k] min, k max The initial value k min As the wave number k, based on the distance difference r i (x, y) and wavenumber k min Construct array direction vector ; Utilizing the noise subspace V N and the array direction vector Calculate the spatial spectrum corresponding to the search grid points. Using the grid point division intervals Δx and Δy as the search step size, search all grid points (x, y) in the monitoring area; Calculate the spatial spectrum over the entire monitoring area, and use the spatial spectrum as pixel values ​​to obtain the wavenumber k. min Impact positioning image I(k) min (x, y), where the location of the spatial spectral peak is the impact position at wavenumber k. .

6. The structural impact localization method based on a multi-signal classification algorithm under varying temperature conditions according to claim 5, characterized in that: Determine the iteration wave number k according to the set wave number iteration step size, and then use the formula... Substitution Impact localization imaging is performed by iteratively calculating the pixel values ​​of all grid points within the monitoring area at different wavenumbers k until... When the iteration terminates, the impact positioning images I(k, x, y) corresponding to different wavenumbers within the wavenumber interval are obtained; and the impact positioning image with the largest maximum pixel value in the impact positioning images I(k, x, y) is taken as the representative impact positioning image.

7. The structural impact localization method based on multi-signal classification algorithm under varying temperature conditions according to claim 1, characterized in that: The wavenumber iteration interval is associated with the single frequency.

8. A structural impact positioning device based on a multi-signal classification algorithm under varying temperature conditions, characterized in that, Includes a memory and a processor, wherein the memory stores computer program code, and the processor is configured to execute the computer program code to perform: Record the impact stress wave signal, which comes from a sensor array arranged in the monitoring area, construct a Cartesian coordinate system in the monitoring area, and mark the position coordinates of the sensor array in the Cartesian coordinate system; Extract a single-frequency signal from the shock stress wave signal; Define the wavenumber iteration interval and the iterative wavenumber, wherein the wavenumber iteration interval includes the initial wavenumber value; The impact location image of the single-frequency impact stress wave signal under the initial wave number value is obtained based on the multi-signal classification algorithm. The wavenumber iteration interval and wavenumber iteration step size are set to perform iteration to obtain the impact positioning image corresponding to each wavenumber within the wavenumber iteration interval. Mark the maximum pixel value of the impact positioning image corresponding to each wavenumber in the wavenumber iteration interval; select the impact positioning image with the largest maximum pixel value among all impact positioning images as the representative impact positioning image, and determine the grid point position where the maximum pixel value of the representative impact positioning image is located as the impact position; The step of obtaining the impact localization image corresponding to each wavenumber within the wavenumber iteration interval includes: A covariance matrix is ​​constructed for the extracted single-frequency signals, and eigenvalue decomposition is performed to obtain the signal subspace and noise subspace of each single-frequency signal; The monitoring area is divided into a grid, and the distance difference between each grid point and each sensor is calculated and the distance to the reference sensor is calculated. An array direction vector is established with wavenumber and the distance difference as variables. Using the initial wavenumber value in the wavenumber iteration interval, the array direction vector, and the noise subspace, the spatial spectrum of all grid points is calculated, and the spatial spectrum of all grid points is used as pixel value imaging to obtain the impact positioning image of the entire monitoring area under the initial wavenumber value.

9. The structural impact positioning device based on a multi-signal classification algorithm under varying temperature conditions according to claim 8, characterized in that: The processor further executes the computer program code to perform the method of any one of claims 2 to 7.

Images