MUSIC damage location method based on search space constraint strategy

Abstract

The application discloses a MUSIC damage positioning method based on a search space constraint strategy, and the damage scattering signal obtained through single array element excitation is subjected to characteristic decomposition, the signal noise subspace is directly acquired, and the structural damage is preliminarily positioned. According to the prior information of the complex form of the structure, a search neighborhood constraint factor is defined; the algorithm is determined according to the damage position of the preliminary positioning and the neighborhood constraint factor to further search the constraint space; and finally, the damage scattering signal is focused and enhanced in the constraint space, and the accurate damage position is searched. The search process of the traditional MUSIC method in the structural health monitoring for damage positioning needs to perform the full domain subdivision, signal focusing, eigenvalue decomposition to acquire the noise subspace and space spectrum peak comparison on the monitored structure, thereby leading to a large number of calculation times and low efficiency. The method can greatly reduce the calculation times of the MUSIC method focusing signal, noise subspace and space spectrum peak, and improve the damage positioning efficiency of the MUSIC method.

Description

Technical Field

[0001] This invention belongs to the field of engineering structure health monitoring technology, specifically involving a MUSIC damage localization method based on a search space constraint strategy. Background Technology

[0002] Boeing has developed an aircraft health management system that improves operational efficiency through diagnostic tools such as real-time fault monitoring and data analysis, enabling troubleshooting and targeted maintenance even during flight. Airbus has introduced a flight health monitoring system that collects aircraft diagnostic information in real time, organizes, tracks, and analyzes alerts, allowing for rapid and effective decision-making and preparation of optimal solutions. Therefore, efficient algorithms are currently a key demand driver. In practical engineering applications, the enormous computational demands have become a bottleneck hindering the practical engineering application of MUSIC damage localization methods, especially in projects with high real-time requirements.

[0003] Due to the advantages of guided waves, such as long propagation distance in structures and sensitivity to small damage, guided wave-based structural health monitoring methods are considered one of the most promising online monitoring methods for engineering applications. Multiple signal classification (MUSIC) damage localization is a newly introduced array signal processing method for guided wave structural health monitoring in recent years. The basic idea of ​​this method is to perform eigenvalue decomposition on the covariance matrix of arbitrary array output data to obtain a signal subspace corresponding to the signal components and a noise subspace orthogonal to the signal components. Then, the orthogonality of these two subspaces is used to estimate the direction of the signal source.

[0004] However, in existing structural health monitoring methods, the MUSIC damage localization method applied to piezoelectric guided wave dense arrays requires a search process that involves subdividing the entire monitored structure, focusing the signal, performing eigenvalue decomposition to obtain the noise subspace, comparing spatial spectral peaks, and finally locating the damage. This process involves numerous calculations and low computational efficiency, making online real-time damage localization of the structure extremely difficult. Therefore, improving the efficiency of the MUSIC damage localization method is essential. Summary of the Invention

[0005] To address the shortcomings of the existing technology, the present invention aims to provide a MUSIC damage localization method based on a search space constraint strategy, thereby solving the problem of high computational complexity in existing MUSIC damage localization methods; the method of the present invention improves the computational efficiency of MUSIC damage localization methods.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] The present invention provides a MUSIC damage localization method based on a search space constraint strategy, comprising the following steps:

[0008] (1) Acquire structural health status sensing signals: When the structure is in a healthy state, a single element in the excitation array will excite an excitation signal into the structure, and the sensing array will receive the response signal of the structure.

[0009] (2) Acquiring structural damage state sensing signals: During the damage monitoring process, a single element in the excitation array is used to excite the structure with an excitation signal, and the sensing array receives the response signal of the structure.

[0010] (3) Obtain the damage scattering array signal under a single excitation source: Subtract the structural damage state sensing signal and the structural health state sensing signal under the same excitation sensing channel to obtain the damage scattering array signal under a single excitation source.

[0011] (4) Obtaining the noise subspace: Using the damage scattering array signal under a single excitation source, calculate the covariance matrix of the array signal, and then perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues ​​arranged from large to small. The eigenvectors corresponding to the small eigenvalues ​​form the noise subspace, and the eigenvectors corresponding to the large eigenvalues ​​form the signal subspace.

[0012] (5) Set the search start position and search step size: Set the search start position to (r s ,θ s That is, the distance r from the search position to the reference element in the sensor array. s The search position is θ relative to the sensor array. s The search step sizes for distance and angle are r, respectively. int and θ int ;where r s r int θ s θ int The value range is within the monitoring range;

[0013] (6) Obtain the array steering vector: Calculate the distance from the search position to each element in the sensor array, then obtain the distance difference between the search position to each element and the reference element in the sensor array, and then obtain the time difference between the time from the search position to each element and the reference element based on the waveguide propagation speed, which is defined as the array delay. Calculate the array steering vector based on the obtained array delay.

[0014] (7) Calculate the spatial spectrum based on a single excitation source, and obtain the initial location of the damage after imaging the spatial spectrum;

[0015] (8) Define the search neighborhood constraint factor based on prior information about the complex structural form;

[0016] (9) Determine the search constraint space: If the distance between the initial location of the damage and the complex structure is within R, it is determined that the initial location of the damage is on the complex structure; otherwise, it is determined that the initial location of the damage is on the non-complex structure, and the search constraint space is further adaptively set in combination with the specific structure.

[0017] (10) Focus and enhance the damage scattering array signal under multiple single excitation sources in the constrained space; calculate the time delay of each excitation array element to the search position, shift the scattering signal of each array element when it is excited forward or backward according to the time delay, and then superimpose the damage scattering signal after applying the time delay to obtain the superimposed damage scattering signal.

[0018] (11) Calculate the search constraint space spectrum and image it. The peak point of the imaging result is the damage location.

[0019] Further, step (1) includes an excitation array E and a sensing array T, each array consisting of 2K+1 array elements arranged at equal intervals. Each array element in the excitation array E is represented by E. g This means that each element in the sensing array T is represented by T. g Let g represent the array element number, and g = -K, -(K-1), ..., 0, ..., K-1, K.

[0020] Furthermore, the damage scattering array signal obtained in step (3) based on the single excitation source E0 is:

[0021]

[0022] In the formula, H(t) represents the sensor array response signal based on a single excitation source E0 under the structural health state, and D(t) represents the sensor array response signal based on a single excitation source E0 under the structural damage state. This represents the damage scattering array signal obtained by the sensing array T based on a single excitation source E0.

[0023] Furthermore, in step (4), the covariance matrix C of the damaged scattering array signal is:

[0024]

[0025] In the formula, for The Hermitian transpose, where Y is the signal sampling length;

[0026] Perform eigenvalue decomposition on the covariance matrix C:

[0027]

[0028] In the formula, Z S Zη These are the signal subspace and the noise subspace, respectively. S Σ η These are the large eigenvalues ​​corresponding to the signal subspace and the small eigenvalues ​​corresponding to the noise subspace, respectively.

[0029] Further, the elements in the array guiding vector A(r,θ) in step (6) are represented as follows: ω0 is the center frequency of signal propagation, and j is the imaginary unit. b is the spacing between adjacent array elements, v is the signal propagation speed, r and θ are the distance and angle of the search position relative to the sensing array T, respectively, and τ g This represents the signal arrival time delay of each array element in array T relative to the reference array element T0.

[0030] Furthermore, in step (7), the noise subspace Z from step (4) is utilized. η And the array steering vector A(r,θ) in step (6), calculate the spatial spectrum J corresponding to the search position (r,θ). MUSIC The spatial spectrum expression is:

[0031]

[0032] In the formula, A H (r,θ) is the Hermitian transpose of A(r,θ). For Z η Hermitian transpose;

[0033] Imaging the spatial spectrum reveals a distinct peak, which represents the initial damage localization location (r0, θ0) obtained based on a single excitation source.

[0034] Furthermore, in step (10), the damage scattering array signal G', which is focused and enhanced using a coherent superposition method within the constrained space, is:

[0035]

[0036] In the formula, G g Let ω0 represent the damage scattering array signal obtained by the sensing array T when the g-th array element in the excitation array E is excited, where ω0 is the center frequency of signal propagation and t is the relative time delay from each array element in the excitation array E to the search space position.

[0037] Furthermore, in step (11), the search constraint space spectrum J is calculated by combining the focused and enhanced damage scattering array signal obtained in step (10) within the constraint space. MUSIC2 (r,θ):

[0038]

[0039] In the formula, Z' η This represents the noise subspace formed by eigenvectors corresponding to small eigenvalues ​​obtained through eigenvalue decomposition of the covariance matrix of the focused and enhanced damage scattering array signal.

[0040] The beneficial effects of this invention are:

[0041] This invention first performs feature decomposition on the damage scattering array signal obtained by excitation of a single array element to directly obtain the signal noise subspace and initially locate the structural damage location; then, it defines a search neighborhood constraint factor based on prior information of the complex structure; next, it further searches the constraint space based on the initially located damage location and the neighborhood constraint factor determination method; finally, it focuses and enhances the damage scattering signal within the constraint space to search for the precise damage location.

[0042] This invention significantly reduces the number of calculations required for the focused signal, noise subspace, and spatial spectrum peaks in the MUSIC damage localization method, thereby improving the efficiency of the MUSIC damage localization method. Attached Figure Description

[0043] Figure 1 This is a flowchart of the method of the present invention;

[0044] Figure 2 This is a schematic diagram of the composite material structure and array arrangement in the embodiment;

[0045] Figure 3 This is a schematic diagram of the damage location in the embodiment;

[0046] Figure 4 This is a signal diagram of the damage scattering array in the embodiment;

[0047] Figure 5 The image shown is a preliminary damage localization imaging result of MUSIC based on a single excitation source in the embodiment.

[0048] Figure 6 The image shows the MUSIC damage localization imaging results based on the search space constraint strategy in this embodiment. Detailed Implementation

[0049] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to embodiments and accompanying drawings. The content mentioned in the embodiments is not intended to limit the present invention.

[0050] Reference Figure 1 As shown, the MUSIC damage localization method based on a search space constraint strategy of the present invention comprises the following steps:

[0051] (1) Acquire structural health status sensing signals: When the structure is in a healthy state, a single element in the excitation array will excite an excitation signal into the structure, and the sensing array will receive the response signal of the structure.

[0052] (2) Acquiring structural damage state sensing signals: During the damage monitoring process, a single element in the excitation array is used to excite the structure with an excitation signal, and the sensing array receives the response signal of the structure.

[0053] (3) Obtain the damage scattering array signal under a single excitation source: Subtract the structural damage state sensing signal and the structural health state sensing signal under the same excitation sensing channel to obtain the damage scattering array signal under a single excitation source.

[0054] (4) Obtaining the noise subspace: Using the damage scattering array signal under a single excitation source, calculate the covariance matrix of the array signal, and then perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues ​​arranged from large to small. The eigenvectors corresponding to the small eigenvalues ​​form the noise subspace, and the eigenvectors corresponding to the large eigenvalues ​​form the signal subspace.

[0055] (5) Set the search start position and search step size: Set the search start position to (r s ,θ s That is, the distance r from the search position to the reference element in the sensor array. s The search position is θ relative to the sensor array. s The search step sizes for distance and angle are r, respectively. int and θ int ;where r s r int θ s θ int The value range is within the monitoring range;

[0056] (6) Obtain the array steering vector: Calculate the distance from the search position to each element in the sensor array, then obtain the distance difference between the search position to each element and the reference element in the sensor array, and then obtain the time difference between the time from the search position to each element and the reference element based on the waveguide propagation speed, which is defined as the array delay. Calculate the array steering vector based on the obtained array delay.

[0057] (7) Calculate the spatial spectrum based on a single excitation source, and obtain the initial location of the damage after imaging the spatial spectrum;

[0058] (8) Define the search neighborhood constraint factor based on the prior information of the complex structural form; for example, the search neighborhood constraint factor for the stiffened structural form is defined as I1, the search neighborhood constraint factor for the wing rib structural form is defined as I2, and the search neighborhood constraint factor for the non-complex structural form is defined as I3.

[0059] (9) Determine the search constraint space: If the distance between the initial location of the damage and the complex structure is within R, it is determined that the initial location of the damage is on the complex structure; otherwise, it is determined that the initial location of the damage is on the non-complex structure. Further, the search constraint space is adaptively set in combination with the specific structure. For example, the search constraint space for the stiffened structure is adaptively set to the monitoring area × I1, the search constraint space for the wing rib structure is adaptively set to the monitoring area × I2, and the search constraint space for the non-complex structure is adaptively set to the monitoring area × I3.

[0060] (10) Focus and enhance the damage scattering array signal under multiple single excitation sources in the constrained space; calculate the time delay of each excitation array element to the search position, shift the scattering signal of each array element when it is excited forward or backward according to the time delay, and then superimpose the damage scattering signal after applying the time delay to obtain the superimposed damage scattering signal.

[0061] (11) Calculate the search constraint space spectrum and image it. The peak point of the imaging result is the damage location.

[0062] The object in this embodiment is a reinforced composite material structure with dimensions of 90cm × 90cm × 0.3cm, such as... Figure 2 As shown, an excitation array E and a sensing array T are arranged on the surface of the structure. There are two reinforcing ribs in the middle of the composite material plate, each 4 cm wide. The distance between the arrays is set to 19.8 cm. The excitation array E and the sensing array T are connected by the reinforcing ribs. Each array has 7 elements, and the spacing between adjacent elements is 1.2 cm.

[0063] The specific implementation method for damage localization in composite material structures is as follows:

[0064] 1. Obtain the damage scattering array signal under a single excitation source;

[0065] In a healthy structural state, the array element E0 in the excitation array excites the structure with an excitation signal, and the sensing array receives the structure's response signal; during damage monitoring, such as... Figure 3 As shown, the array element E0 in the excitation array excites the structure with an excitation signal, and the sensing array receives the response signal of the structure; the difference between the structural damage state sensing signal and the structural health state sensing signal under the same excitation sensing channel is used to obtain the damage scattering array signal under a single excitation source:

[0066]

[0067] In the formula, H(t) represents the sensor array response signal based on a single excitation source E0 under the structural health state, and D(t) represents the sensor array response signal based on a single excitation source E0 under the structural damage state. This represents the damage scattering array signal obtained by the sensing array T based on a single excitation source E0, such as... Figure 4 As shown.

[0068] 2. Calculate the spatial spectrum based on a single excitation source;

[0069] (1) Using the damage scattering array signal under a single excitation source, the covariance matrix of the array signal is calculated as follows:

[0070]

[0071] In the formula, for The Hermitian transpose, where Y is the signal sampling length;

[0072] (2) Perform eigenvalue decomposition on the covariance matrix C:

[0073]

[0074] In the formula, Z S Z η These are the signal subspace and the noise subspace, respectively. S Σ η These are the large eigenvalues ​​corresponding to the signal subspace and the small eigenvalues ​​corresponding to the noise subspace, respectively.

[0075] (3) Set the search start position and search step size: Set the search start position to (1mm, 1°), that is, the distance from the search position to the reference element in the sensor array is 1mm, the direction of the search position relative to the sensor array is 1°, and the search step size for distance and angle is 1mm and 1° respectively.

[0076] (4) Calculate the distance from the search position to each element in the array, then obtain the distance difference between the search position and each element relative to the distance to the reference element in the sensing array. Based on the waveguide propagation speed of 1026.8 m / s at 40 kHz, obtain the time difference between the search position and each element relative to the reference element, which is defined as the array delay. Calculate the array steering vector based on the obtained array delay:

[0077] The elements in the array guide vector A(r,θ) are represented as j is the imaginary unit. r and θ are the distance and angle of the search position relative to the sensor array T, respectively, and τ g The signal arrival time delay of each element in array T relative to the reference element T0 is represented by g, where g = -3, -2, ..., 0, ..., 2, 3.

[0078] (5) Utilizing the noise subspace Z ηAnd the array steering vector A(r,θ), calculate the spatial spectrum J corresponding to the search position (r,θ). MUSIC The spatial spectrum expression is:

[0079]

[0080] In the formula, A H (r,θ) is the Hermitian transpose of A(r,θ). For Z η Hermitian transpose;

[0081] like Figure 5 As shown, when imaging the spatial spectrum, there is a distinct peak in the spatial spectrum, which is the initial damage location (r0, θ0) obtained based on a single excitation source.

[0082] 3. Determine the search constraint space;

[0083] First, the neighborhood constraint factor for reinforced structures is defined as 1%, and the neighborhood constraint factor for unreinforced structures is defined as 0.5%. Then, based on the initial damage location and the neighborhood constraint factor, the algorithm further searches the constraint space. If the distance between the initial damage location and the reinforced structure is within 5cm, the initial damage location is determined to be on the reinforced structure; otherwise, the initial damage location is determined to be on the unreinforced structure. Furthermore, the search constraint space is adaptively set based on the initial damage location result. The search constraint space for reinforced structures is adaptively set to 1% of the monitoring area, and the search constraint space for unreinforced structures is adaptively set to 0.5% of the monitoring area.

[0084] 4. Focusing and enhancing damage scattering signals from multiple single excitation sources within a constrained space;

[0085] Calculate the time delay of each excitation array element reaching the search position, shift the scattered signal of each array element at excitation forward or backward according to the time delay, and then superimpose the damage scattered signals after applying the time delay to obtain the focused and enhanced damage scattered array signal G':

[0086]

[0087] In the formula, G g Let ω0 represent the damage scattering array signal obtained by the sensing array T when the g-th array element in the excitation array E is excited, where ω0 is the center frequency of signal propagation and t is the relative time delay from each array element in the excitation array E to the search space position.

[0088] 5. Calculate the search constraint space spectrum and image it; the peak points in the imaging results are the damage locations.

[0089] The search constraint space spectrum J is calculated by combining the focused and enhanced damage scattering array signal obtained within the constraint space. MUSIC2 (r,θ):

[0090]

[0091] In the formula, Z' η This represents the noise subspace formed by eigenvectors corresponding to small eigenvalues ​​obtained through eigenvalue decomposition of the covariance matrix of the focused and enhanced damage scattering array signal.

[0092] Imaging the constrained spatial spectrum, such as Figure 6 As shown, a distinct peak exists in the spatial spectrum, indicating the precise location of the damage. The horizontal axis represents the direction of arrival (ROA) of the signal source, providing a precise estimate of the damage direction. The vertical axis represents the distance to the signal source, which is a precise estimate of the damage distance.

[0093] This invention has many specific applications. The above description is only a preferred embodiment of this invention. It should be noted that for those skilled in the art, several improvements can be made without departing from the principle of this invention, and these improvements should also be considered within the scope of protection of this invention.

Claims

1. A MUSIC damage localization method based on a search space constraint strategy, characterized in that, The steps are as follows: (1) Acquiring structural health status sensing signals: When the structure is in a healthy state, a single element in the excitation array is used to excite the structure with an excitation signal, and the sensing array receives the response signal of the structure. (2) Acquiring structural damage state sensing signals: During the damage monitoring process, a single element in the excitation array is used to excite the structure with an excitation signal, and the sensing array receives the response signal of the structure. (3) Obtain the damage scattering array signal under a single excitation source: Subtract the structural damage state sensing signal and the structural health state sensing signal under the same excitation sensing channel to obtain the damage scattering array signal under a single excitation source. (4) Obtaining the noise subspace: Using the damage scattering array signal under a single excitation source, calculate the covariance matrix of the array signal, and then perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues ​​arranged from large to small. The eigenvectors corresponding to the small eigenvalues ​​form the noise subspace, and the eigenvectors corresponding to the large eigenvalues ​​form the signal subspace. (5) Set the search start position and search step size: Set the search start position to That is, the distance from the search position to the reference element in the sensor array is The search position is relative to the direction of the sensor array. The search step sizes for distance and angle are respectively and ; (6) Obtain the array steering vector: Calculate the distance from the search position to each element in the sensor array, then obtain the distance difference between the search position to each element and the reference element in the sensor array, and then obtain the time difference between the time from the search position to each element and the reference element based on the waveguide propagation speed, which is defined as the array delay. Calculate the array steering vector based on the obtained array delay. (7) Calculate the spatial spectrum based on a single excitation source, and obtain the initial location of the damage after imaging the spatial spectrum; (8) Define the search neighborhood constraint factor based on the prior information of the complex structural form; the search neighborhood constraint factor for the stiffened structural form is defined as I1, the search neighborhood constraint factor for the wing rib structural form is defined as I2, and the search neighborhood constraint factor for the non-complex structural form is defined as I3. (9) Based on the initial location of the damage and the neighborhood constraint factor, the algorithm further searches the constraint space: when the distance between the initial location of the damage and the complex structure is within R, it is determined that the initial location of the damage is on the complex structure; otherwise, it is determined that the initial location of the damage is on the non-complex structure, and the search constraint space is further adaptively set in combination with the specific structure; the search constraint space at the stiffened structure is adaptively set to the monitoring area × I1, the search constraint space at the wing rib structure is adaptively set to the monitoring area × I2, and the search constraint space at the non-complex structure is adaptively set to the monitoring area × I3. (10) Focusing and enhancing the damage scattering array signal under multiple single excitation sources within the constrained space; Calculate the time delay of each excitation array element to the search position, shift the scattered signal of each array element at the time of excitation forward or backward according to the time delay, and then superimpose the damage scattered signal after applying the time delay to obtain the superimposed damage scattered signal. (11) Calculate the search constraint space spectrum and image it. The peak point of the imaging result is the damage location.

2. The MUSIC damage localization method based on a search space constraint strategy according to claim 1, characterized in that, Step (1) includes an excitation array E and a sensing array T. Each array consists of 2K+1 array elements, which are arranged at equal intervals. Each element in the excitation array E is represented by E. g This means that each element in the sensing array T is represented by T. g Let g represent the array element number, and g = -K, -(K-1), ..., 0, ..., K-1, K.

3. The MUSIC damage localization method based on a search space constraint strategy according to claim 2, characterized in that, The damage scattering array signal obtained in step (3) based on the single excitation source E0 is: ； In the formula, H(t) represents the sensor array response signal based on a single excitation source E0 under the structural health state, and D(t) represents the sensor array response signal based on a single excitation source E0 under the structural damage state. This represents the damage scattering array signal obtained by the sensing array T based on a single excitation source E0.

4. The MUSIC damage localization method based on a search space constraint strategy according to claim 3, characterized in that, In step (4), the covariance matrix of the damaged scattering array signal : ； In the formula, for The Hermitian transpose, where Y is the signal sampling length; For covariance matrix Perform eigenvalue decomposition: ； In the formula, , These are the signal subspace and the noise subspace, respectively. , These are the large eigenvalues ​​corresponding to the signal subspace and the small eigenvalues ​​corresponding to the noise subspace, respectively.

5. The MUSIC damage localization method based on a search space constraint strategy according to claim 4, characterized in that, In step (6), the array guiding vector The elements in are represented as , The center frequency of signal propagation is j, where j is the imaginary unit. b is the spacing between adjacent array elements, v is the signal propagation speed, and r and These represent the distance and angle of the search position relative to the sensor array T, respectively. This represents the signal arrival time delay of each array element in array T relative to the reference array element T0.

6. The MUSIC damage localization method based on a search space constraint strategy according to claim 5, characterized in that, In step (7), the noise subspace from step (4) is utilized. and the array guide vector in step (6) Calculate the search position Corresponding spatial spectrum The spatial spectrum expression is: ； In the formula, for Hermitian transpose, for Hermitian transpose; Imaging the spatial spectrum reveals a distinct peak, which represents the initial damage localization location obtained based on a single excitation source. .

7. The MUSIC damage localization method based on a search space constraint strategy according to claim 6, characterized in that, In step (10), the enhanced damage scattering array signal is focused using a coherent superposition method within the constrained space. for: ； In the formula, This represents the damage scattering array signal obtained by the sensing array T when the g-th element in the excitation array E is excited. Let t be the center frequency of signal propagation, and t be the relative time delay from each element in the excitation array E to its position in the search space.

8. The MUSIC damage localization method based on a search space constraint strategy according to claim 7, characterized in that, In step (11), the search constrained space spectrum is calculated by combining the focused and enhanced damage scattering array signal obtained in step (10) within the constrained space. : ； In the formula, This represents the noise subspace formed by eigenvectors corresponding to small eigenvalues ​​obtained through eigenvalue decomposition of the covariance matrix of the focused and enhanced damage scattering array signal.

Images