Damage imaging method, device and medium based on density clustering of distance optimization RAPID

Abstract

The application relates to a damage imaging method, device and medium based on distance optimization density clustering RAPID. The method comprises the following steps: S1, one sensor of a sensor network is selected as an exciter to drive the sensor to generate a guided wave signal, so as to obtain a health signal SH and a damage signal SD of each sensing path. S2, a damage factor DI and a distance optimization parameter beta of the corresponding sensing path are calculated according to the health signal SH and the damage signal SD of each sensing path. S3, a to-be-measured region is determined according to a to-be-measured structure, and a damage probability value matrix P of the to-be-measured region is calculated according to the distance optimization parameter beta and the damage factor DI based on a probability reconstruction algorithm. S4, density clustering processing is performed on the damage probability value matrix P to eliminate false damage probability values, so as to form a new damage probability value matrix TP. S5, a damage image of the to-be-measured region is drawn according to the damage probability value matrix TP. The application can obtain damage imaging results more in line with the real damage shape.

Description

Technical Field

[0001] This invention relates to the field of engineering structure health monitoring technology, and in particular to a damage imaging method, device and medium based on distance-optimized density clustering RAPID. Background Technology

[0002] Structural health monitoring technology employs a novel concept of smart material structures. Driving and sensing elements are pre-attached into the structure in a network (a sensor network is deployed on the structure under test). Sensor signals are collected in real-time online, and combined with advanced signal processing algorithms, relevant signal characteristic parameters are extracted to assess the structural condition and even evaluate the location and size of structural damage. Structural health monitoring technology can monitor and control the condition of engineering structures throughout the design, testing, and service processes. Therefore, it can ensure the application of new materials, guide the reliable design of structures, thereby improving structural safety and reducing maintenance costs.

[0003] Because conventional on-site nondestructive evaluation (NDE) or nondestructive testing (NDT) techniques are overly complex, a more intuitive and easily understood method for monitoring structural damage is needed. Structural health monitoring (SHM) technology integrates sensors into the structure according to a predefined network, acquiring sensor signals online in real time, providing an excellent means of determining the structural health status. Among these, guided wave-based detection technology is one of the most widely used aircraft structural damage detection technologies. Guided waves have slow energy attenuation, long propagation distances, high sensitivity and accessibility, and sufficient detection range. When guided waves interact with defects, phenomena such as transmission, reflection, and scattering occur, causing changes in the amplitude and phase of the original excitation signal. Therefore, ultrasonic guided waves carry all damage information along the propagation path. By analyzing the changes in signal components, structural damage can be identified and assessed. Establishing a real-time health monitoring system requires many technologies, such as visualization algorithms, sensor networks, and signal processing. Using visualization technology to provide accurate damage images is crucial.

[0004] The Reconstruction Algorithm for Probabilistic Inspection (RAPID) method for arrayed waveguide damage imaging is a promising damage detection and localization technique in structural health monitoring (SHM), offering advantages such as simple principle, fast operation, and accurate localization. This algorithm typically requires acquiring a healthy signal in the early stages of the structure under test, before any damage occurs. This signal is then stored and compared with subsequently acquired data throughout the system's lifespan. At each inspection, damage factors (Damage Index, DI) are calculated based on the healthy reference signal and the currently acquired data, enabling the detection of eventual damage and determination of the most probable location within the monitored area.

[0005] However, current engineering applications demand increasingly higher levels of damage detection capabilities. Simply providing damage location information is insufficient to meet practical needs. Traditional RAPID algorithms and their extensions struggle to provide information on the degree and shape of damage. According to the logic of the RAPID algorithm, paths closer to the damage have higher DI values ​​(damage factor values). When the results are calculated using elliptic localization and probability assignment algorithms and reflected in the damage image, blocks with higher thermal values ​​appear. After fusing the images of all sensing paths to obtain the damage thermal image of the entire measured area, sensing paths with higher DI values ​​cause the damage image to extend along the path from the extreme point. Therefore, the RAPID algorithm not only locates the damage but also, to some extent, reflects the general trend of damage propagation. The spurious damage values ​​in the damage probability matrix are the cause of shape deviations in the image drawn using pixel values ​​as element values.

[0006] Therefore, in damage imaging based on probability detection reconstruction algorithms, it is necessary to introduce distance optimization parameters and density clustering algorithms to optimize the calculation process of damage probability, and to eliminate false damage through density clustering algorithms in order to obtain damage imaging results that better match the shape of the real damage. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a damage imaging method, device and medium based on distance-optimized density clustering RAPID.

[0008] The technical solution adopted by this invention to solve its technical problem is: a damage imaging method based on distance-optimized density clustering RAPID, wherein a sensor network is arranged on the structure under test, the sensor network including multiple sensing paths, and the method includes the following steps:

[0009] S1. In the healthy state and the damaged state of the structure under test, respectively, one of the sensors in the sensor network is selected as the exciter to drive it to generate guided wave signals, so as to obtain the health signal SH and the damage signal SD of each sensing path.

[0010] S2. Calculate the damage factor DI and distance optimization parameter β for each sensing path based on the health signal SH and damage signal SD of each sensing path.

[0011] S3. Determine the test area based on the structure to be tested, and calculate the damage probability matrix P of the test area based on the distance optimization parameter β and the damage factor DI, using the probability reconstruction algorithm.

[0012] S4. Perform density clustering on the damage probability value matrix P to eliminate false damage probability values, thereby forming a new damage probability value matrix TP.

[0013] S5. Draw the damage image of the area to be tested based on the damage probability matrix TP.

[0014] Furthermore, in the damage imaging method of the present invention, step S2 includes the following steps:

[0015] S2-1. Calculate the various signal feature values ​​of each sensing path based on the health signal SH and damage signal SD of each sensing path, and fuse all the various signal feature values ​​of each sensing path to obtain the damage factor DI corresponding to each sensing path.

[0016] S2-2. Calculate the corresponding distance optimization parameter β based on the damage factor DI for each sensing path.

[0017] Furthermore, in the damage imaging method described in this invention, step S2-1 includes the following steps:

[0018] S2-1-1. The eight signal characteristic values ​​for each sensing path are obtained sequentially using the following calculation method:

[0019]

[0020] DI2=rms(SH)-rms(SD) Formula (2)

[0021] DI3=var(SH)-var(SD) Formula (3)

[0022] DI4=means(|SH|)-means(|SD|) Formula (4)

[0023]

[0024] DI6 = max(SH) - max(SD) Formula (6)

[0025]

[0026] DI8=kurtosis(SH)-kurtosis(SD) Formula (8)

[0027] S2-1-2. The damage factor DI for each sensing path is obtained using the following full-sample fusion calculation method:

[0028]

[0029] Where sqrt is for calculating the square root, rms is for calculating the root mean square error, var is for calculating the variance, means is for calculating the mean, max is for calculating the maximum value, min is for calculating the minimum value, kurtosis is for calculating the kurtosis, i is the type number of the signal feature value, and n = 1, 2, ..., 8.

[0030] Furthermore, in the damage imaging method of the present invention, step S2-2 includes:

[0031] The distance optimization parameter β for the corresponding sensing path is obtained through the following calculation method. k :

[0032]

[0033] Where k is the sequence number of the k-th sensing path, k = 1, 2, 3, ..., β 最优 Let β be the optimal value of the distance optimization parameter obtained through experimental research, and α be the constraint parameter. k Let be the damage factor of the k-th sensing path, min(DI) be the smallest damage factor among all sensing paths, and max(DI) be the largest damage factor among all sensing paths.

[0034] Furthermore, in the damage imaging method described in this invention, step S3 includes the following steps:

[0035] S3-1. The area to be measured, determined based on the position of the structure to be measured, is gridded, and a Cartesian coordinate system is established for the area to be measured. Then, the distance coefficient R from each grid position (x, y) to each sensing path is calculated sequentially. k (x, y);

[0036] S3-2, the distance coefficient R k (x, y) and the corresponding distance optimization parameter β k By comparison, the first probability L of the corresponding grid position is obtained. k (x, y);

[0037] S3-3, Based on the damage factor DI of the k-th sensing path k For the first probability L k The probability reconstruction of (x, y) is performed to obtain the damage probability value P(x, y) at the corresponding grid position, and a damage probability value matrix P of the area to be tested is formed.

[0038] Furthermore, in the damage imaging method of the present invention, step S3-2 includes:

[0039] The first probability L of each grid location is obtained through the following calculation method. k (x, y):

[0040]

[0041] Step S3-3 includes:

[0042] The damage factor DI of the k-th sensing path k As a weight, for the first probability L k The probability reconstruction calculation is performed on (x, y) to obtain the damage probability value P(x, y) at the corresponding grid position, and a damage probability value matrix P of the area to be tested is formed.

[0043] Furthermore, in the damage imaging method described in this invention, step S4 includes the following steps:

[0044] S4-1. Calculate the standard deviation of the damage probability matrix P to obtain the degree of damage in the area to be tested.

[0045] S4-2. Using the degree of damage as a threshold for filtering the damage probability value of each grid position, density clustering is performed on the damage probability value matrix P to eliminate false damage probability values, thereby forming a new damage probability value matrix TP.

[0046] Furthermore, in the damage imaging method of the present invention, step S4-1 includes:

[0047] The degree of damage (STD) of the area to be tested is obtained using the following calculation method. p :

[0048]

[0049] Where num(P) is the capacity of the damage probability matrix P; P(n) is the nth element in the damage probability matrix; and P is the average value of all elements in the damage probability matrix.

[0050] Furthermore, in the damage imaging method described in this invention, after step S5, the following steps are also included:

[0051] S6. Determine the damage shape of the structure under test based on the damage image.

[0052] Furthermore, in the damage imaging method of the present invention, step S5 includes:

[0053] The damage image is drawn using each element of the damage probability matrix TP as a pixel value.

[0054] In addition, the present invention provides a computer-readable storage medium storing a computer program adapted for loading by a processor to perform the steps of the distance-optimized density clustering RAPID damage imaging method described above.

[0055] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the distance-optimized density clustering RAPID damage imaging method as described above by calling the computer program stored in the memory.

[0056] The damage imaging method, device, and medium based on distance-optimized density clustering RAPID of the present invention have the following beneficial effects: Addressing the problem that traditional RAPID algorithms and their extensions only provide damage location information and cannot provide damage degree and shape information, thus failing to meet practical needs, the present invention proposes a distance optimization algorithm for damage factors. This algorithm adaptively adjusts the probability reconstruction range during the probability reconstruction process of the RAPID algorithm, combining the results of density clustering with the distance-optimized RAPID algorithm to optimize damage imaging results. This improves the accuracy of damage imaging and has broad application prospects in the diagnosis of damage in practical engineering structures.

[0057] A multi-type guided wave signal eigenvalue fusion method is proposed to improve the reliability of guided wave signal feature extraction. This further enhances the accuracy of damage degree assessment, damage imaging, and damage shape extraction. Attached Figure Description

[0058] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0059] Figure 1 This is a schematic flowchart of a damage imaging method provided in an embodiment of the present invention;

[0060] Figure 2 This is a schematic flowchart of a damage imaging method provided in some embodiments of the present invention;

[0061] Figure 3 This is a schematic flowchart of a damage imaging method provided in some embodiments of the present invention;

[0062] Figure 4 This is a schematic flowchart of a damage imaging method provided in some embodiments of the present invention;

[0063] Figure 5 This is a schematic flowchart of a damage imaging method provided in some embodiments of the present invention;

[0064] Figure 6 This is a diagram showing the arrangement of three groups of damage and a schematic diagram of all sensing paths provided in an embodiment of the present invention.

[0065] Figure 7 yes Figure 6 The scattering signal diagrams of the first group of damages are shown in the figure. (a) is the signal diagram of the sensing path PZT1-PZT7, and (b) is the signal diagram of the sensing path PZT5-PZT9.

[0066] Figure 8 yes Figure 6 The measured data of damage factor DI and distance optimization parameter β for all 132 sensing paths of the first group of damage;

[0067] Figure 9 yes Figure 6 A comparison chart of the actual damage levels and the damage levels obtained by the algorithm for the three groups of damages;

[0068] Figure 10 yes Figure 6 Damage images of the first group of injuries;

[0069] Figure 11 yes Figure 6 Damage images of the second group of injuries;

[0070] Figure 12 yes Figure 6 Damage images of the third group of injuries. Detailed Implementation

[0071] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0072] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0073] In a preferred embodiment, the distance-optimized density clustering RAPID damage imaging method of this embodiment requires deploying a sensor network (i.e., a sensor array) on the structure under test. The sensor network includes multiple sensing paths. Specifically, the sensor network consists of multiple sensors, with sensing paths formed between each pair of sensors. N sensing paths are formed between n sensors, and the paths are independent of the shape of the sensor network. The formula for calculating the sensing paths is as follows:

[0074] N = n*(n-1)

[0075] For example, 12 PZT-5A piezoelectric ceramic sensors can be selected to form a sensor network and attached to the structure under test using a coupling agent. The shape of the sensor network includes, but is not limited to, circular arrays, square arrays, etc.

[0076] refer to Figure 1 The method includes the following steps:

[0077] S1. In the healthy state and the damaged state of the structure under test, respectively, one of the sensors in the sensor network is selected as the exciter to drive it to generate guided wave signals, so as to obtain the health signal SH and the damage signal SD of each sensing path.

[0078] S2. Calculate the damage factor DI and distance optimization parameter β for each sensing path based on the health signal SH and damage signal SD for each sensing path.

[0079] S3. Determine the test area based on the structure to be tested, and calculate the damage probability matrix P of the test area based on the distance optimization parameter β and the damage factor DI using the probability reconstruction algorithm.

[0080] S4. Perform density clustering on the damage probability matrix P to eliminate false damage probability values, thereby forming a new damage probability matrix TP.

[0081] S5. Draw the damage image of the area to be tested based on the damage probability matrix TP. Specifically, draw the damage image using the pixel value of each element in the damage probability matrix TP.

[0082] This embodiment addresses the problem that traditional RAPID algorithms and their extensions only provide damage location information and cannot provide damage degree and shape information, thus failing to meet practical needs. It proposes a distance optimization algorithm for damage factors, which adaptively adjusts the probability reconstruction range during the RAPID algorithm's probability reconstruction process. By combining the results of density clustering with the distance-optimized RAPID algorithm, the damage imaging results are optimized, improving the accuracy of damage imaging and showing broad application prospects in the diagnosis of damage in practical engineering structures.

[0083] Optionally, refer to Figure 2 Following step S5, the following steps are also included:

[0084] S6. Determine the damage shape of the structure under test based on the damage image. Specifically, the damage shape of the structure under test is identified by searching for the highlighted areas of the entire damage image.

[0085] This embodiment combines the results of density clustering with the distance optimization RAPID algorithm to optimize damage imaging results and extract damage shape. This overcomes the problem that the RAPID algorithm cannot obtain the degree and shape of damage, and improves the accuracy of damage localization and damage imaging.

[0086] In some embodiments, reference Figure 3 Step S2 includes the following steps:

[0087] S2-1. Calculate the various signal feature values ​​of each sensing path based on the health signal SH and damage signal SD of each sensing path, and fuse all the various signal feature values ​​of each sensing path to obtain the damage factor DI corresponding to each sensing path.

[0088] S2-2. Calculate the corresponding distance optimization parameter β based on the damage factor DI for each sensing path.

[0089] In this embodiment, a multi-type guided wave signal feature value fusion method is proposed to improve the reliability of guided wave signal feature extraction. This further improves the accuracy of damage degree and damage imaging, as well as the accuracy of damage shape extraction.

[0090] In some embodiments, reference Figure 4 Step S3 includes the following steps:

[0091] S3-1. Grid the area to be measured, determined by the location of the structure to be measured, and establish a Cartesian coordinate system for the area. Then, calculate the distance coefficient R from each grid position (x, y) to each sensing path. k (x, y).

[0092] S3-2, Distance coefficient R k (x, y) and the corresponding distance optimization parameter β k By comparison, the first probability L of the corresponding grid position is obtained. k (x, y).

[0093] S3-3, Based on the damage factor DI of the k-th sensing path k For the first probability L k The probability reconstruction of (x, y) is performed to obtain the damage probability value P(x, y) at the corresponding grid position, and a damage probability value matrix P of the area to be tested is formed.

[0094] In some embodiments, reference Figure 5 Step S4 includes the following steps:

[0095] S4-1. Calculate the standard deviation of the damage probability matrix P to obtain the degree of damage in the area to be tested.

[0096] S4-2. Using the degree of damage as the threshold for filtering the damage probability value of each grid position, density clustering is performed on the damage probability value matrix P to eliminate false damage probability values, thereby forming a new damage probability value matrix TP.

[0097] In one specific embodiment, taking a circular array of 12 PZT-5A piezoelectric ceramic sensors as an example, the 12 PZT-5A piezoelectric ceramic sensors are evenly arranged on the structure under test in a circle with a radius of 150 mm using a coupling agent.

[0098] A circular array of piezoelectric ceramic sensor elements is driven to generate guided waves. Simultaneously, guided wave signals received by all other sensors acting as receivers are acquired. Guided wave signals in both damaged and healthy states are extracted separately. Various signal characteristic values ​​are calculated to obtain the damage factor DI, and the distance optimization parameter β is obtained based on the magnitude of the damage factor DI. This process continues until all sensing paths have had their healthy and damaged signals extracted, and the damage factor DI and distance optimization parameter β are calculated for each corresponding path. In this embodiment, the following eight types of signal characteristic values ​​are preferably used for calculation:

[0099] S2-1-1. The eight signal characteristic values ​​for each sensing path are obtained sequentially using the following calculation method:

[0100]

[0101] DI2=rms(SH)-rms(SD) Formula (2)

[0102] DI3=var(SH)-var(SD) Formula (3)

[0103] DI4=means(|SH|)-means(|SD|) Formula (4)

[0104]

[0105] DI6 = max(SH) - max(SD) Formula (6)

[0106]

[0107] DI8=kurtosis(SH)-kurtosis(SD) Formula (8)

[0108] S2-1-2. The damage factor DI for each sensing path is obtained using the following full-sample fusion calculation method:

[0109]

[0110] Where sqrt is for calculating the square root, rms is for calculating the root mean square error, var is for calculating the variance, means is for calculating the mean, max is for calculating the maximum value, min is for calculating the minimum value, kurtosis is for calculating the kurtosis, i is the type number of the signal feature value, and n = 1, 2, ..., 8.

[0111] Alternatively, the distance optimization parameter β for the corresponding sensing path can be obtained through the following calculation method. k :

[0112]

[0113] Where k is the sequence number of the k-th sensing path, k = 1, 2, 3, ..., β 最优 To determine the optimal value of the distance optimization parameter β obtained through experimental research, this embodiment can select β as the optimal value. 最优 =1.05. α is a constraint parameter, and α is positively correlated with β. That is, the larger α is, the more sensitive β is to changes in the damage factor; conversely, the smaller α is, the more stable the change in β. In this embodiment, α = 0.04 can be selected. DI k Let be the damage factor of the k-th sensing path, min(DI) be the smallest damage factor among all sensing paths, and max(DI) be the largest damage factor among all sensing paths.

[0114] Next, the area to be tested is gridded, and the damage at each grid location is checked sequentially. The distance coefficient R(x, y) between each grid location and the entire sensing path is calculated:

[0115]

[0116] In the above formula, (x, y) are the coordinates of the detected grid; (x ki y ki ) and (x kj y kj ) represent the coordinates of the exciter (array element) and receiver in the k-th sensing path, that is, x ki Let y be the x-coordinate of the excitation sensor in the k-th sensing path. ki x is the ordinate of the excitation sensor in the k-th sensing path; kj Let y be the x-coordinate of the receiving sensor in the k-th sensing path. kj Let be the ordinate of the receiving sensor in the k-th sensing path.

[0117] After obtaining the distance coefficients R(x, y) for all grid locations in the area to be tested, the distance coefficients R(x, y) are compared with the distance optimization parameter β corresponding to the sensing path. The location information that meets the size condition is further input into the probabilistic reconstruction algorithm, and the damage probability value of the location is calculated by combining the distance optimization parameter. The above processing is performed on each network location to obtain the damage probability value of each grid location in the area to be tested, thereby forming the damage probability value matrix P.

[0118] Specifically, the first probability L of each grid location can be obtained through the following calculation method. k (x, y):

[0119]

[0120] Step S3-3 includes:

[0121] The damage factor DI of the k-th sensing path k As a weight, for the first probability L k The probability reconstruction calculation is performed on (x, y) to obtain the damage probability value P(x, y) at the corresponding grid location, forming the damage probability matrix P of the area to be tested. The specific calculation method is as follows:

[0122]

[0123] In the above formula, n p This represents the total number of sensor paths.

[0124] Step S4-1 includes:

[0125] The degree of damage (STD) of the area to be tested is obtained using the following calculation method. p :

[0126]

[0127] Where num(P) is the capacity of the damage probability matrix P. P(n) is the nth element in the damage probability matrix. P is the average value of all elements in the damage probability matrix.

[0128] In this embodiment, the standard deviation is used to reflect the dispersion of each damage point relative to the damage probability value matrix. A larger standard deviation indicates that high damage probability points are dispersed relative to the average value, while a smaller standard deviation indicates that high damage probability points are closer to the average value, thus accurately reflecting the degree of damage in the measured area.

[0129] Furthermore, the degree of damage is used as a threshold to filter the damage probability value at each location, and a new damage probability matrix TP is obtained by eliminating false damage probability values ​​through density clustering. Specifically:

[0130]

[0131] In the formula, N(P(j))={P(i)∈Cond│distance(P(i),P(j))≤}.

[0132] Cond is the conditional expression; ∈ is the neighborhood, which determines the distance threshold between data points; MinPts defines the number of data points in the neighborhood required for a core point; N∈(P(j)) is the set of grids that satisfy the threshold condition.

[0133] Finally, the damage image is drawn using each element of the new damage probability matrix TP as a pixel value. By searching for the highlighted areas of the entire damage image, the shape of the damage can be identified.

[0134] For example, such as Figure 6 As shown, the experimental material was an aluminum plate measuring 500mm x 500mm x 2mm with a density of ρ = 2700 kg / m³. 3 Poisson's ratio v = 0.33, Young's modulus E = 700 GPa. A coordinate system is established with the lower left corner of the plate as the origin. Circular sensor arrays have the advantages of comprehensive detection and small error, so 12 piezoelectric ceramic plates are evenly placed on a circle with a radius of 150 mm. This experiment uses PZT-5A type piezoelectric elements, with a diameter of 8 mm and a thickness of 0.48 mm. The piezoelectric plate located at (250 mm, 100 mm) is designated as sensor number 1, and each sensor is named counterclockwise from PZT 1 to PZT 12.

[0135] The first group of damage consists of a 23mm diameter annular iron block with a 10mm diameter hole in the center, attached to the coordinates (250mm, 325mm). Figure 6 As shown in (a). The second group of damage consists of an iron block 80m long and 5mm wide, pasted at coordinates (325mm, 250mm), as shown. Figure 6 As shown in (b). The third group of damage consists of an iron block 120m long and 5mm wide, attached at coordinates (325mm, 250mm), as shown. Figure 6 As shown in (c). Figure 6 (d) shows all excitation sensing paths.

[0136] When the structure under test is healthy, each PZT element in the sensor array is driven sequentially to generate guided waves, while the guided wave signals received by all other PZT elements are collected as the reference signal SH. During damage monitoring, the response signal of the sensor array is collected using the same driving method, denoted as the damage signal SD. A total of N excitations are performed, where N is the number of sensors in the array; during each excitation, the guided wave signals received by the remaining N-1 elements are collected simultaneously; the final array signal dimension is N×(N-1)×M, where M is the sampling length.

[0137] Taking the first group of injuries as an example, such as Figure 7 As shown, crosstalk exists in the signal segments from 0 to 100 sampling points of the sensing paths PZT1-PZT7 and PZT5-PZT9. The signal waveforms collected after 400 sampling points are chaotic. The reason for this is that the edge reflection wave caused by the tested structure and damage directly passes through the damaged sensing path, resulting in significant amplitude attenuation of the actual signal. The signal segment from 100 to 300 sampling points was selected for amplification, and the residual between the actual signal and the reference signal was calculated. The maximum amplitude of the obtained signal was greater than 2V.

[0138] like Figure 8 As shown, taking the first group of damage as an example, the eight signal feature values ​​of 132 sensing paths are calculated using the above formula. By comparing the obtained damage factor DI with the distance optimization parameter β, it can be found that the result is consistent with the expectation. When a larger damage factor is obtained, it indicates that the real damage point is closer to the sensing path. At this time, a smaller parameter is set to reduce the inspection range of the probability ellipse, and vice versa, a larger parameter is set to expand the inspection range of the probability ellipse.

[0139] The calculated damage severity for the three groups of injuries were 43.2435, 62.7333, and 64.3775, respectively. The area of ​​the adhesive surface for the three simulated injuries was used as a direct factor in assessing the damage severity. The area of ​​the first group of injuries was 336.9358, the second group was 400, and the third group was 600. A comparison and analysis of the damage severity and damage area was conducted... Figure 9 As shown, comparing the actual damage areas of the three sets with the corresponding results obtained by the algorithm reveals a linear relationship between the algorithm results and the actual damage; as the damage area increases, the degree of damage also increases proportionally. Therefore, under the same monitoring conditions, the optimized algorithm of this scheme can achieve accurate damage degree monitoring.

[0140] Figures 10 to 12 Damage images of three groups of damages obtained by the damage imaging method based on the distance-optimized density clustering RAPID algorithm are presented. It can be clearly seen that the optimized algorithm proposed in this invention can accurately obtain the shape and longitudinal extension direction of the actual damage, and can accurately identify circular damage, 80mm rectangular damage, and 120mm rectangular damage.

[0141] In another preferred embodiment, the computer-readable storage medium of the present invention stores a computer program adapted for loading by a processor to perform the steps of the distance-optimized density clustering RAPID damage imaging method as described above.

[0142] This invention proposes a distance optimization algorithm for damage factors. It adaptively adjusts the probability reconstruction range during the probabilistic reconstruction process of the RAPID algorithm and combines the results of density clustering with the distance optimization RAPID algorithm to optimize damage imaging results. This can improve the accuracy of damage imaging and has broad application prospects in the diagnosis of damage in practical engineering structures.

[0143] In another preferred embodiment, the computer device of the present invention includes a memory and a processor. The memory stores a computer program, and the processor executes the steps of the distance-optimized density clustering RAPID damage imaging method described above by calling the computer program stored in the memory.

[0144] This invention proposes a distance optimization algorithm for damage factors. It adaptively adjusts the probability reconstruction range during the probabilistic reconstruction process of the RAPID algorithm and combines the results of density clustering with the distance optimization RAPID algorithm to optimize damage imaging results. This can improve the accuracy of damage imaging and has broad application prospects in the diagnosis of damage in practical engineering structures.

[0145] The computer-readable storage medium of the present invention can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a magnetic disk, or an optical disk.

[0146] In the embodiments of this application, the processor can be a Central Processing Unit (CPU), but it can also be 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. The general-purpose processor can be a microprocessor or any conventional processor.

[0147] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0148] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0149] It is understood that the above embodiments only illustrate preferred embodiments of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can freely combine the above technical features without departing from the concept of the present invention, and can also make several modifications and improvements, all of which fall within the protection scope of the present invention. Therefore, all equivalent transformations and modifications made with respect to the scope of the claims of the present invention should fall within the scope of the claims of the present invention.

Claims

1. A damage imaging method based on distance-optimized density clustering RAPID, characterized in that, A sensor network, comprising multiple sensing paths, is deployed on the structure under test. The method includes the following steps: S1. In the healthy state and the damaged state of the structure under test, respectively, one of the sensors in the sensor network is selected as the exciter to drive it to generate guided wave signals, so as to obtain the health signal SH and the damage signal SD of each sensing path. S2. Calculate the damage factor DI and distance optimization parameter β for each sensing path based on the health signal SH and damage signal SD of each sensing path. S3. Determine the test area based on the structure to be tested, and calculate the damage probability matrix P of the test area based on the distance optimization parameter β and the damage factor DI, using the probability reconstruction algorithm. S4. Perform density clustering on the damage probability value matrix P to eliminate false damage probability values, thereby forming a new damage probability value matrix TP. S5. Draw the damage image of the area to be tested based on the damage probability matrix TP; Step S2 includes the following steps: S2-1. Calculate the various signal feature values ​​of each sensing path based on the health signal SH and damage signal SD of each sensing path, and fuse all the various signal feature values ​​of each sensing path to obtain the damage factor DI corresponding to each sensing path. S2-2. Calculate the corresponding distance optimization parameter β based on the damage factor DI for each sensing path; Step S3 includes the following steps: S3-1. The area to be tested, determined based on the position of the structure to be tested, is gridded, and a Cartesian coordinate system is established for the area to be tested. Then, the distance coefficient R from each grid position (x, y) to each sensing path is calculated sequentially. k (x, y); S3-2, the distance coefficient R k (x, y) and the corresponding distance optimization parameter β k By comparison, the first probability L of the corresponding grid position is obtained. k (x, y); S3-3, Based on the damage factor DI of the k-th sensing path k For the first probability L k (x, y) is subjected to probability reconstruction processing to obtain the damage probability value P(x, y) at the corresponding grid position, and a damage probability value matrix P of the area to be tested is formed. Step S2-2 includes: The distance optimization parameter β for the corresponding sensing path is obtained through the following calculation method. k : Where k is the sequence number of the k-th sensing path, k = 1, 2, 3, ..., β 最优 Let β be the optimal value of the distance optimization parameter obtained through experimental research, and α be the constraint parameter. k Let be the damage factor of the k-th sensing path, min(DI) be the minimum damage factor among all sensing paths, and max(DI) be the maximum damage factor among all sensing paths. Step S4 includes the following steps: S4-1. Calculate the standard deviation of the damage probability matrix P to obtain the degree of damage in the area to be tested. S4-2. Using the degree of damage as a threshold for filtering the damage probability value of each grid position, density clustering is performed on the damage probability value matrix P to eliminate false damage probability values, thereby forming a new damage probability value matrix TP.

2. The damage imaging method according to claim 1, characterized in that, Step S2-1 includes the following steps: S2-1-1. The eight signal characteristic values ​​for each sensing path are obtained sequentially using the following calculation method: S2-1-2. The damage factor DI for each sensing path is obtained using the following full-sample fusion calculation method: Where sqrt is for calculating the square root, rms is for calculating the root mean square error, var is for calculating the variance, means is for calculating the mean, max is for calculating the maximum value, min is for calculating the minimum value, kurtosis is for calculating the kurtosis, i is the type number of the signal feature value, and n=1, 2, ..., 8.

3. The damage imaging method according to claim 1, characterized in that, Step S3-2 includes: The first probability L of each grid location is obtained through the following calculation method. k (x, y): Step S3-3 includes: The damage factor DI of the k-th sensing path k As a weight, for the first probability L k The probability reconstruction calculation is performed on (x, y) to obtain the damage probability value P(x, y) at the corresponding grid position, and a damage probability value matrix P of the area to be tested is formed.

4. The damage imaging method according to claim 1, characterized in that, Step S4-1 includes: The degree of damage (STD) of the area to be tested is obtained using the following calculation method. p : Where num(P) is the capacity of the damage probability matrix P; P(n) is the nth element in the damage probability matrix. This is the average value of all elements in the damage probability matrix.

5. The damage imaging method according to claim 1, characterized in that, Following step S5, the following steps are also included: S6. Determine the damage shape of the structure under test based on the damage image.

6. The damage imaging method according to claim 1, characterized in that, Step S5 includes: The damage image is drawn using each element of the damage probability matrix TP as a pixel value.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted for loading by a processor to perform the steps of the damage imaging method based on distance-optimized density clustering RAPID as described in any one of claims 1 to 6.

8. A computer device, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the distance-optimized density clustering RAPID damage imaging method as described in any one of claims 1 to 6 by calling the computer program stored in the memory.

Images