Damage location method based on Lamb wave temperature compensation

Abstract

The present invention relates to the field of structural health monitoring technology, and in particular, provides a damage location method based on Lamb wave temperature compensation. The method includes establishing a Lamb wave-based structural health monitoring system and grouping sensing paths within a monitoring area; collecting and grouping healthy Lamb wave signals at a reference temperature to obtain a reference signal and construct a reference mapping matrix; collecting and grouping damaged Lamb wave signals at a target temperature to obtain a target signal and construct a target mapping matrix; determining the damage scattering signal at the target temperature; extracting damage location features, calculating the path damage probability, and determining the damage location. During the damage location process, the method reduces the Lamb wave's dependence on temperature, achieves rapid and synchronous location of the damage location, and improves its wide applicability.

Description

Technical Field

[0001] The present invention relates to the technical field of structural health monitoring, and in particular to a damage location method based on Lamb wave temperature compensation. Background Art

[0002] Existing structural health monitoring systems widely utilize lamb wave technology. Its operating principle is to collect lamb wave signals through a piezoelectric sensor array and analyze the signal characteristics to detect and locate damage. However, the propagation characteristics of lamb waves in structures are susceptible to changes in ambient temperature, manifesting as fluctuations in wave velocity and time-domain signal deformation. This results in significant differences in the signal response to the same damage at different temperatures. Traditional temperature compensation and location methods primarily include neural network-based compensation methods and piecewise linear compensation methods.

[0003] Neural network-based compensation methods primarily construct training datasets by collecting Lamb wave signals (such as wave velocity, flight time, amplitude, and dispersion characteristics) at different temperatures and their corresponding temperature values. Neural networks (such as BP, LMBP, and CNN) take temperature or temperature-related signal features as input and output compensated signal parameters (such as corrected wave velocity and time-domain signals). Neural network methods can automatically capture nonlinear relationships and adapt to complex scenarios, but they suffer from the following limitations: Strong data dependence and insufficient generalization: Neural network methods require a large number of training samples of temperature and signal responses. However, obtaining high-quality data across the entire temperature range is costly in practical engineering, and the damage mechanisms of the structures under test are complex, limiting the generalization ability of the models. Compensation effectiveness can be significantly reduced, especially in extreme temperatures or untrained temperature ranges. High computational complexity and poor real-time performance: Neural network training involves numerous iterative computations (such as optimizing weights in the LM algorithm), significantly increasing training time, especially when the number of hidden layer nodes is large. For scenarios requiring real-time monitoring (such as aviation structures), rapid response times are difficult to achieve. Traditional BP networks are prone to local minima and overfitting risks: They are prone to falling into local minima, resulting in unstable compensation results. Furthermore, network performance is sensitive to initial parameter values, and network structure design (such as the number of hidden layer nodes) requires trial and error, potentially introducing overfitting.

[0004] The piecewise linear compensation method works by pre-dividing the ambient temperature range into multiple adjacent sub-intervals (e.g., 10°C intervals). Based on the signals collected at the boundary temperatures of these intervals, a linear mapping model is established between the signal parameters and temperature. In practical applications, linear compensation is performed by estimating the range of the current detected temperature. This method simplifies the model based on physical assumptions and is computationally efficient, but it has the following limitations: Inaccuracy due to model simplification: Piecewise linear compensation typically divides the temperature range into several segments and establishes a linear relationship. However, Lamb wave propagation characteristics (such as wave velocity and dispersion) exhibit a nonlinear relationship with temperature. Especially in variable temperature environments (e.g., -20°C to 50°C), simple linear models cannot accurately describe complex temperature effects, resulting in large compensation residuals; Reliance on a priori temperature segmentation and calibration: Segmentation requires pre-determined temperature interval boundaries and calibration points. In practice, temperature variations may span multiple intervals or exhibit a non-uniform distribution, requiring frequent adjustments to the segmentation strategy and increasing system maintenance complexity; Poor robustness and adaptability: The segmented model is sensitive to long-term factors such as sensor drift and material aging, requiring regular recalibration. In addition, in anisotropic structures such as composite materials, the effect of temperature on wave propagation is direction-dependent, and piecewise linearization is difficult to adapt to multi-dimensional nonlinear changes. Summary of the Invention

[0005] In view of this, the present invention provides a damage localization method based on Lamb wave temperature compensation, which is used to reduce the dependence of Lamb wave on temperature during the damage localization process, achieve rapid and synchronous localization of the damage position, and improve wide applicability.

[0006] In a first aspect, the present invention provides a damage localization method based on Lamb wave temperature compensation, the method comprising:

[0007] Step 1: Build a structural health monitoring system based on Lamb waves and group the sensing paths within the monitoring area;

[0008] Step 2: Using the system and sensing path grouping in step 1, collect and group the healthy Lamb wave signals at the reference temperature to obtain the reference signal and construct a reference mapping matrix;

[0009] Step 3: Through the system and sensing path grouping in step 1 and step 2, the damage Lamb wave signals at the target temperature are collected and grouped to obtain the target signal and construct the target mapping matrix;

[0010] Step 4: Using step 3, determine the damage scattering signal at the target temperature;

[0011] Step 5: According to step 4, extract the damage location features, calculate the path damage probability, and determine the damage location.

[0012] Optionally, step 1 includes:

[0013] A structural health monitoring system is constructed using a piezoelectric sensor array, a Lamb wave detector, a host computer, and a structure to be tested. The piezoelectric sensor array consists of L piezoelectric sensors, which are evenly attached to the surface of the structure to be tested in a circular or square shape with equal spacing. They are used to transmit and receive Lamb wave signals from the structure to be tested. The number of sampling points used by the Lamb wave detector to obtain the signal is represented by m. One end of the Lamb wave detector is connected to the sensor and the other end is connected to the host computer, which is responsible for transmitting the excitation and collected Lamb wave signals. The host computer is responsible for setting the excitation signal parameters and displaying and storing the collected signals. The structure to be tested is a flat plate or circular tubular structure made of aluminum alloy or carbon fiber.

[0014] Group the sensing paths within the monitoring area: a. The straight line connecting the two different sensors is called a sensing path. For a piezoelectric sensing array consisting of L piezoelectric sensors, the number of sensing paths is b. According to the direction of the sensing path, all the sensing paths are divided into N groups: path1 to pathN; the directions of the sensing paths in the same group are consistent, and the directions of the sensing paths in different groups are different; for the kth path group, if Represents the number of sensing paths contained within it, then the total number of sensing paths is , 1≤k≤N.

[0015] Optionally, step 2 includes:

[0016] Under the reference temperature, the healthy Lamb wave signals are collected and grouped, with room temperature as the reference temperature T r , using the structural health monitoring system to collect T r The Lamb wave signal at the temperature and the healthy state of the structure to be measured is recorded as the reference signal SR; according to the grouping of the sensing path, the reference signal SR is divided into N groups, which are recorded as to ; For the kth group of reference signals , which contains A column vector of dimension m×1, using Indicates that , The dimension is m×1; ;1≤k≤N;

[0017] At the reference temperature, the reference mapping matrix within the sensing path group is constructed. For the reference signal SR, the dynamic time warping DTW algorithm is used to calculate the mapping matrix between the reference signals within each path group, which is recorded as the reference mapping matrix. The reference mapping matrix is expressed as Indicates that represents the mapping matrix between reference signals in the kth path group.

[0018] Optionally, take the reference signal of the kth path group , specifically introduces the mapping matrix between reference signals in the kth path group based on the DTW algorithm The build process:

[0019] ①For signal and , first, construct an m×m dimensional matrix D to represent and The Euclidean distance of any element in the matrix , 1≤i≤m, 1≤j≤m; then, the cumulative distance matrix C is calculated according to formula (1) m×m , where parameter β ≥ 1. Parameter β is the bending coefficient introduced to avoid over-compensation of the signal and to impose nonlinear constraints on the signal:

[0020] (1);

[0021] ②After obtaining the cumulative distance matrix C, from the last element c of the matrix mm Start backtracking along the minimum path to c 11 , the selection principle of each element on the minimum path is: if c ij is an element on the minimum path, then the next element on the path is The smallest one in the cumulative distance matrix C is obtained by analogy. mm to c 11 The minimum path P;

[0022] ③According to the row and column number set of each element on the minimum path P, the signal is obtained To signal The mapping matrix , m ≤ q ≤ 2m+1; mapping matrix It is a two-dimensional matrix with 2 rows, where the data in the oth column [i o ,j o ] T express The i-th o Elements and The jth o Elements Match, that is, if The i-th o Elements Assigned as The jth o elements, and the signal Matched estimated signal ;

[0023] ④ Repeat steps ①-③ to obtain the reference signal on any path f To the reference signal on path u Mapping matrix , and then get the reference signal The mapping matrix , whose expression is:

[0024] .

[0025] Optionally, step 3 includes:

[0026] At the target temperature, the damaged Lamb wave signals are collected and grouped, and the ambient temperature of the structure to be tested is recorded as the target temperature T o ,and ; Use the structural health monitoring system to collect T o The Lamb wave signal at the temperature and the damage state of the structure to be measured is recorded as the target signal SO; according to the grouping of the sensing path, the target signal SO is divided into N groups, which are recorded as to ; For the kth group of reference signals , which contains A column vector of dimension m×1, using Indicates that , The dimension is m×1; ;1≤k≤N;

[0027] At the target temperature, the target mapping matrix in the sensing path group is constructed, and the target temperature T o Next, the mapping matrix between Lamb wave signals in each path group when the structure to be tested is in a healthy state is recorded as the target mapping matrix ,in The target mapping matrix of the kth path group is expressed as:

[0028] ;

[0029] in, Represents the mapping matrix from the Lamb wave signal of path i to the Lamb wave signal of path j at the target temperature, 1≤i≤n k , 1≤j≤n k ;

[0030] When detecting ambient temperature changes, the reference mapping matrix and the target mapping matrix are not equal, that is, ; The deviation between the target mapping matrix WO and the reference mapping matrix WR is recorded as the correction matrix , whose expression is: ;

[0031] in, , represents the kth path group, target mapping matrix With the reference mapping matrix The deviation between ; Represents the target mapping matrix With the reference mapping matrix Deviation between

[0032] Based on the correction matrix and the reference mapping matrix , calculate the target mapping matrix ,Right now .

[0033] Optionally, take path 1 and path 2 in the kth path group and specifically introduce the correction matrix The calculation process:

[0034] For the signal and , based on the DTW algorithm, we can get arrive The mapping matrix is expressed as: , m ≤ q ≤ 2m+1; According to formula (2), the correction matrix is obtained , that is, the correction matrix The first row of is numerically equal to the matrix The difference between the second row and the first row, The second row of is numerically equal to Compared with the path The product of is expressed as:

[0035] (2);

[0036] Among them, the path ratio is the ratio of the length of path 2 to the length of path 1;

[0037] Repeat the above process to obtain the correction matrix of the kth path group , and then obtain the correction matrix .

[0038] Optionally, step 4 includes:

[0039] Calculate the target estimation signal based on the target mapping matrix and target signal , calculate the target estimation signal ,in, represents the target estimation signal of the kth path group, and its expression is:

[0040] ;

[0041] in, Indicates the target signal based on the fth path and target mapping matrix The target estimated signal on the u-th path is obtained, 1≤f≤n k , 1≤u≤n k ;1≤k≤N;

[0042] Determine the damage scattering signal at the target temperature, based on the target signal and target estimation signal , calculate the damage scattering signal ,in, represents the damage scattering signal of the kth path group, and its expression is:

[0043] ;

[0044] Among them, 1≤k≤N.

[0045] Optionally, take the path f and path u in the kth path group and describe the target estimation signal calculation process in detail:

[0046] For the target mapping matrix , where the oth column data [i o ,j o ] T represents the i-th target signal on path f o element and the jth element of the target signal on path u o elements in the time domain; the target signal on path f The i-th o Elements are assigned as signals jth o elements, that is, the target signal on path u is obtained Target estimation signal matched in time domain .

[0047] Optionally, step 5 includes:

[0048] h. Extract the damage location features based on the orthogonal matching pursuit (OMP) algorithm, and then calculate the damage scattering signal. , calculate the damage characteristic matrix ,in, represents the damage characteristic of the kth path group, and its expression is:

[0049] ;

[0050] Where, 1 ≤ k ≤ N;

[0051] For the kth path group, Indicates the scattered signal based on the OMP algorithm The damage characteristics obtained by sparse decomposition are calculated as follows:

[0052] I. Construct an overcomplete dictionary D, whose atoms correspond to wave packet basis functions with different time delays, expressed as:

[0053] (3)

[0054] Where T is the duration of Lamb wave signal; m 1 is the number of atoms, which is numerically equal to the number of sampling points of the Lamb wave detector to obtain the signal; is the Lamb wave basis function;

[0055] II. Parameter initialization: the current number of iterations u = 0, the initial residual , support set ;

[0056] III. Calculate the current residual The inner product with all atoms in dictionary D, the most relevant atom index according to formula (4) , and update the support set: ; The expression is:

[0057] (4);

[0058] in, is the vth column of dictionary D, i.e. the vth atom;

[0059] IV. Update the sparse coefficient vector: , For the dictionary The submatrix composed of the columns of the index updates the residual: ;

[0060] V. Order , if u=4, the process ends, otherwise go to step III;

[0061] According to the above process, we get the support set Contains four elements, assigned to , a sparse coefficient vector There are four non-zero elements in , which are assigned to ;

[0062] i. Construct the search point location feature matrix within the monitoring area According to the damage identification accuracy requirements, the monitoring area of the structure to be tested is discretized into r×g search points with equal spacing. The position coordinate matrix PS of the search points is expressed as:

[0063] ;

[0064] in, Indicates the The position coordinates of the search points, 1 ≤ i ≤ r, 1 ≤ j≤ g;

[0065] Calculate the feature matrix of the search point position on the sensing path ,in, Represents the position feature matrix of the search point on the path within the kth path group, that is, , represents the number of sensing paths in the kth path group; represents the position characteristics of the search point on the jth path in the kth path group, It is a two-dimensional matrix with dimensions r×g, and its expression is:

[0066] (5)

[0067] Where, 1 ≤ k ≤ N; 1 ≤ j ≤ , represents the coordinates of the excitation sensor on the jth path; represents the coordinates of the receiving sensor on the jth path; Represents the search point coordinates PS and The Euclidean distance of , dimension is r×g; Represents the search point coordinates PS and The Euclidean distance of , dimension is r×g; It represents the propagation speed of Lamb wave in the structure to be measured;

[0068] j. Calculate the path damage probability and determine the damage location. The total damage probability within the monitoring area is: , where N is the number of groups in the sensing path; PD k is the damage probability of the kth path group, and PD is calculated according to formula (6): k :

[0069] (6)

[0070] in, is the number of sensing paths in the kth path group; According to the characteristics of the damage location and search point location features The damage probability of sensing path j in the kth path group is calculated according to formula (7): :

[0071] (7)

[0072] in, is the time threshold, which is set to 1.5 times the Lamb wave excitation period;

[0073] Total damage probability is a two-dimensional matrix of dimension r×g, Indicates the Search point locations The probability of damage occurring at the monitoring location is the maximum damage probability concentration area, where r and g are the number of rows and columns of discrete points in the monitoring area, respectively.

[0074] In the technical solution provided by the present invention, the method includes building a structural health monitoring system based on Lamb waves and grouping the sensing paths within the monitoring area; collecting and grouping healthy Lamb wave signals at a reference temperature to obtain a reference signal, and constructing a reference mapping matrix; collecting and grouping damaged Lamb wave signals at a target temperature to obtain a target signal, and constructing a target mapping matrix; determining the damage scattering signal at the target temperature; extracting damage location characteristics, and calculating the path damage probability to determine the damage location; in the damage location process, the method reduces the dependence of the Lamb wave on temperature, realizes rapid and synchronous location of the damage location, and improves wide applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0076] Figure 1 A flowchart of a damage location method based on Lamb wave temperature compensation provided in an embodiment of the present invention;

[0077] Figure 2 A flowchart of another damage localization method based on Lamb wave temperature compensation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0078] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0079] It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0080] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the embodiments of the present invention are also intended to include plural forms, unless the context clearly indicates other meanings.

[0081] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. Furthermore, the character " / " in this document generally indicates an "or" relationship between the associated objects.

[0082] The word "if," as used herein, may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to the determination" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)," depending on the context.

[0083] The present invention provides a damage location method based on Lamb wave temperature compensation, such as Figure 1 and Figure 2 As shown, the method includes:

[0084] Step 1: Build a structural health monitoring system based on Lamb waves and group the sensing paths within the monitoring area.

[0085] In the embodiment of the present invention, step 1 includes:

[0086] A structural health monitoring system is built using a piezoelectric sensor array, a Lamb wave detector, a host computer, and a structure to be tested. The piezoelectric sensor array consists of L piezoelectric sensors, which are evenly attached to the surface of the structure to be tested in a circular or square shape with equal spacing. They are used to transmit and receive Lamb wave signals from the structure to be tested. The number of sampling points used by the Lamb wave detector to obtain the signal is represented by m. The Lamb wave detector is connected to the sensor at one end and to the host computer at the other end. It is responsible for transmitting the excitation and collected Lamb wave signals and has functions such as filtering and amplification. The host computer is responsible for setting the excitation signal parameters (such as excitation frequency, amplitude, filter upper and lower limits, etc.) and displaying and storing the collected signals. The structure to be tested is a flat plate or circular tubular structure made of aluminum alloy or carbon fiber.

[0087] Group the sensing paths within the monitoring area: a. The straight line connecting the two different sensors is called a sensing path. For a piezoelectric sensing array consisting of L piezoelectric sensors, the number of sensing paths is b. According to the direction of the sensing path, all the sensing paths are divided into N groups: path1 to pathN; the directions of the sensing paths in the same group are consistent, and the directions of the sensing paths in different groups are different; for the kth path group, if Represents the number of sensing paths contained within it, then the total number of sensing paths is , 1≤k≤N.

[0088] Step 2: Using the system and sensing path grouping in step 1, collect and group the healthy Lamb wave signals at the reference temperature, obtain the reference signal, and construct a reference mapping matrix.

[0089] In the embodiment of the present invention, the room temperature is 23° C., and step 2 includes:

[0090] Under the reference temperature, the healthy Lamb wave signals are collected and grouped, with room temperature as the reference temperature T r , using the structural health monitoring system to collect T r The Lamb wave signal at the temperature and in the healthy state of the structure to be tested (no damage on the surface and inside of the structure) is recorded as the reference signal (SR). According to the grouping of the sensing path, the reference signal SR is divided into N groups, which are recorded as to ; For the kth group of reference signals , which contains A column vector of dimension m×1, using Indicates that , The dimension is m×1; ;1≤k≤N;

[0091] At the reference temperature, the reference mapping matrix within the sensing path group is constructed. For the reference signal SR, the dynamic time warping (DTW) algorithm is used to calculate the mapping matrix between the reference signals within each path group, which is recorded as the reference mapping matrix. The reference mapping matrix is expressed as Indicates that represents the mapping matrix between reference signals within the kth path group. The reference mapping matrix of each path group is independently calculated based on the reference signals contained in the group using the same method.

[0092] In the embodiment of the present invention, the reference signal of the kth path group is taken , specifically introduces the mapping matrix between reference signals in the kth path group based on the DTW algorithm The build process:

[0093] ①For signal and , first, construct an m×m dimensional matrix D to represent and The Euclidean distance of any element in the matrix , 1≤i≤m, 1≤j≤m; then, the cumulative distance matrix C is calculated according to formula (1) m×m , where parameter β ≥ 1. Parameter β is the bending coefficient introduced to avoid over-compensation of the signal and to impose nonlinear constraints on the signal:

[0094] (1);

[0095] ②After obtaining the cumulative distance matrix C, from the last element c of the matrix mm Start backtracking along the minimum path to c 11 , the selection principle of each element on the minimum path is: if c ij is an element on the minimum path, then the next element on the path is The smallest one in the cumulative distance matrix C is obtained by analogy. mm to c 11 The minimum path P;

[0096] ③According to the row and column number set of each element on the minimum path P, the signal is obtained To signal The mapping matrix , m ≤ q ≤ 2m+1; mapping matrix It is a two-dimensional matrix with 2 rows, where the data in the oth column [i o ,j o ] T express The i-tho Elements and The jth o Elements Match, that is, if The i-th o Elements Assigned as The jth o elements, and the signal Matched estimated signal ; Matching signal at this time Although with the signal There are certain differences in the values, but they are consistent in the time domain.

[0097] ④ Repeat steps ①-③ to obtain the reference signal on any path f To the reference signal on path u Mapping matrix , and then get the reference signal The mapping matrix , whose expression is:

[0098] .

[0099] Step 3: Through the system and sensing path grouping in step 1 and step 2, the damage Lamb wave signals at the target temperature are collected and grouped to obtain the target signal and construct the target mapping matrix.

[0100] In this embodiment of the present invention, step 3 includes:

[0101] At the target temperature, the damaged Lamb wave signals are collected and grouped, and the ambient temperature of the structure to be tested is recorded as the target temperature T o ,and ; Use the structural health monitoring system to collect T o The Lamb wave signal at the temperature and the damage state of the structure to be measured (damage on the surface and inside of the structure) is recorded as the target signal (objective signal), represented by SO; according to the grouping of the sensing path, the target signal SO is divided into N groups, which are respectively recorded as to ; For the kth group of reference signals , which contains A column vector of dimension m×1, using Indicates that , The dimension is m×1; ;1≤k≤N;

[0102] At the target temperature, the target mapping matrix in the sensing path group is constructed, and the target temperature T o Next, the mapping matrix between Lamb wave signals in each path group when the structure to be tested is in a healthy state is recorded as the target mapping matrix ,in The target mapping matrix of the kth path group is expressed as:

[0103] ;

[0104] in, Represents the mapping matrix from the Lamb wave signal of path i to the Lamb wave signal of path j at the target temperature, 1≤i≤n k , 1≤j≤n k ;

[0105] When the detection environment temperature changes, the propagation speed of Lamb waves in the structure will change, resulting in a time delay between signals at different temperatures and on the same path. At this time, the signal mapping matrix constructed based on the reference temperature can no longer represent the mapping relationship between signals at the target temperature. The reference mapping matrix and the target mapping matrix are not equal, that is, ; The deviation between the target mapping matrix WO and the reference mapping matrix WR is recorded as the modified matrix (modifiedmatrix), and It means that its expression is: ;

[0106] in, , represents the kth path group, target mapping matrix With the reference mapping matrix The deviation between ; Represents the target mapping matrix With the reference mapping matrix The correction matrix of each path group is calculated independently using the same method.

[0107] Based on the correction matrix and the reference mapping matrix , calculate the target mapping matrix ,Right now .

[0108] In the embodiment of the present invention, take path 1 and path 2 in the kth path group, and specifically introduce the correction matrix The calculation process:

[0109] For the signal and , based on the DTW algorithm, we can get arrive The mapping matrix is expressed as: , m ≤ q ≤ 2m+1; According to formula (2), the correction matrix is obtained , that is, the correction matrix The first row of is numerically equal to the matrix The difference between the second row and the first row, The second row of is numerically equal to Compared with the path The product of is expressed as:

[0110] (2);

[0111] Among them, the path ratio is the ratio of the length of path 2 to the length of path 1;

[0112] Repeat the above process to obtain the correction matrix of the kth path group , and then obtain the correction matrix .

[0113] Step 4: Using step 3, determine the damage scattering signal at the target temperature.

[0114] In this embodiment of the present invention, step 4 includes:

[0115] Calculate the target estimation signal based on the target mapping matrix and target signal , calculate the target estimation signal ,in, represents the target estimation signal of the kth path group, and its expression is:

[0116] ;

[0117] in, Indicates the target signal based on the fth path and target mapping matrix The target estimated signal on the u-th path is obtained, 1≤f≤n k , 1≤u≤n k ; 1≤k≤N; The target estimation signal of each path group and each path within the same path group is independently calculated using the same method.

[0118] Determine the damage scattering signal at the target temperature, based on the target signal and target estimation signal , calculate the damage scattering signal ,in, represents the damage scattering signal of the kth path group, and its expression is:

[0119] ;

[0120] Among them, 1≤k≤N.

[0121] In the embodiment of the present invention, the path f and path u in the kth path group are taken to specifically describe the target estimation signal calculation process:

[0122] For the target mapping matrix , where the oth column data [i o ,j o ] T represents the i-th target signal on path f o element and the jth element of the target signal on path u o elements in the time domain; the target signal on path f The i-th o Elements are assigned as signals jth o elements, that is, the target signal on path u is obtained Target estimation signal matched in time domain .

[0123] Step 5: According to step 4, extract the damage location features, calculate the path damage probability, and determine the damage location.

[0124] In this embodiment of the present invention, step 5 includes:

[0125] h. Extract the damage location features based on the Orthogonal Matching Pursuit (OMP) algorithm, and , calculate the damage feature matrix ,in, represents the damage characteristic of the kth path group, and its expression is:

[0126] ;

[0127] Where, 1 ≤ k ≤ N;

[0128] For the kth path group, Indicates the scattered signal based on the OMP algorithm The damage characteristics obtained by sparse decomposition are calculated as follows:

[0129] I. Construct an overcomplete dictionary D, whose atoms correspond to wave packet basis functions with different time delays, expressed as:

[0130] (3)

[0131] Where T is the duration of Lamb wave signal; m1 is the number of atoms, which is numerically equal to the number of sampling points of the Lamb wave detector to obtain the signal; is the Lamb wave basis function;

[0132] II. Parameter initialization: the current number of iterations u = 0, the initial residual , support set ;

[0133] III. Calculate the current residual The inner product with all atoms in dictionary D, the most relevant atom index according to formula (4) , and update the support set: ; The expression is:

[0134] (4);

[0135] in, is the vth column of dictionary D, i.e. the vth atom;

[0136] IV. Update the sparse coefficient vector: , For the dictionary The submatrix composed of the columns of the index updates the residual: ;

[0137] V. Order , if u=4, the process ends, otherwise go to step III;

[0138] According to the above process, we get the support set Contains four elements, assigned to , a sparse coefficient vector There are four non-zero elements in , which are assigned to ;

[0139] i. Construct the search point location feature matrix within the monitoring area According to the damage identification accuracy requirements, the monitoring area of the structure to be tested is discretized into r×g search points with equal spacing. The position coordinate matrix PS of the search points is expressed as:

[0140] ;

[0141] in, Indicates the The position coordinates of the search points, 1 ≤ i ≤ r, 1 ≤ j≤ g;

[0142] Calculate the feature matrix of the search point position on the sensing path ,in, Represents the position feature matrix of the search point on the path within the kth path group, that is, , represents the number of sensing paths in the kth path group; represents the position characteristics of the search point on the jth path in the kth path group, It is a two-dimensional matrix with dimensions r×g, and its expression is:

[0143] (5)

[0144] Where, 1 ≤ k ≤ N; 1 ≤ j ≤ , represents the coordinates of the excitation sensor on the jth path; represents the coordinates of the receiving sensor on the jth path; Represents the search point coordinates PS and The Euclidean distance of , dimension is r×g; Represents the search point coordinates PS and The Euclidean distance of , dimension is r×g; It represents the propagation speed of Lamb wave in the structure to be measured;

[0145] j. Calculate the path damage probability and determine the damage location. The total damage probability (probability of damage) in the monitoring area is: , where N is the number of groups in the sensing path; PD k is the damage probability of the kth path group, and PD is calculated according to formula (6): k :

[0146] (6)

[0147] in, is the number of sensing paths in the kth path group; According to the characteristics of the damage location and search point location features The damage probability of sensing path j in the kth path group is calculated according to formula (7): :

[0148] (7)

[0149] in, is the time threshold, which is set to 1.5 times the Lamb wave excitation period;

[0150] Total damage probability is a two-dimensional matrix of dimension r×g, Indicates the Search point locations The probability of damage occurring at the monitoring location is the maximum damage probability concentration area, where r and g are the number of rows and columns of discrete points in the monitoring area, respectively.

[0151] Compared with the prior art, the present invention has the following beneficial effects:

[0152] (1) In the damage location process, this method reduces the dependence on healthy Lamb wave signals at unknown temperatures: by using the dynamic time warping (DTW) algorithm and the healthy signal at the reference temperature, the mapping matrix between the signals at the reference temperature, i.e., the reference mapping matrix, can be pre-calculated; then, based on the damaged signal at the target temperature and the healthy signal at the reference temperature, the DTW algorithm is used to obtain the deviation matrix of the mapping matrix, i.e., the correction matrix; based on the reference mapping matrix and the correction matrix, the signal mapping matrix between different paths in the same direction at any unknown temperature, i.e., the target mapping matrix, can be dynamically constructed, and then, combined with the Lamb wave damage signal at the unknown temperature, the healthy signal at the current temperature can be estimated and reconstructed. Based on the difference signal between the target estimated signal and the target signal, i.e., the damage scattering signal, the damage location identification process can be continued.

[0153] (2) This method can achieve rapid and simultaneous positioning of multiple damage locations: First, by grouping sensor paths and calculating independent mapping matrices within each group, the amount of redundant data processing is reduced. Then, the orthogonal matching pursuit (OMP) algorithm is introduced to perform sparse decomposition of the damage scattering signal, which can simultaneously extract the location characteristics of multiple damages in a single path. Furthermore, by fusing the multi-path group information based on the damage probability matrix, combining the search point location characteristics with the time domain threshold judgment, the simultaneous positioning of multiple damage locations can be achieved.

[0154] (3) The implementation process of this method is simple and has wide applicability: During the implementation process, there is no need to collect a large number of standard Lamb wave signals at different temperatures and in the healthy state of the structure in advance, nor is there any need to estimate the detection environment temperature. It supports various material structures such as aluminum alloy, steel, and carbon fiber, and is suitable for damage detection of different geometric shapes such as flat plates, L-shaped, and round tubes. The system has strong scalability and provides high-reliability health monitoring solutions for aerospace, rail transportation and other fields. In the technical solution provided by the present invention, the method includes building a structural health monitoring system based on Lamb waves and grouping the sensing paths within the monitoring area; collecting and grouping healthy Lamb wave signals at a reference temperature to obtain reference signals and construct a reference mapping matrix; collecting and grouping damaged Lamb wave signals at a target temperature to obtain target signals and construct a target mapping matrix; determining the damage scattering signal at the target temperature; extracting damage location features, and calculating the path damage probability to determine the damage location; in the damage location process, the method reduces the dependence of Lamb waves on temperature, realizes rapid and synchronous location of the damage location, and improves wide applicability.

[0155] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0156] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A damage location method based on Lamb wave temperature compensation, characterized in that: The method comprises: Step 1: Build a structural health monitoring system based on Lamb waves and group the sensing paths within the monitoring area; Step 2: Using the system and sensing path grouping in step 1, collect and group the healthy Lamb wave signals at the reference temperature to obtain the reference signal and construct a reference mapping matrix; Step 3: Through the system and sensing path grouping in step 1 and step 2, the damage Lamb wave signals at the target temperature are collected and grouped to obtain the target signal and construct the target mapping matrix; Step 4: Using step 3, determine the damage scattering signal at the target temperature; Step 5: According to step 4, extract the damage location features, calculate the path damage probability, and determine the damage location; The step 4 comprises: Calculate the target estimation signal based on the target mapping matrix and target signal , calculate the target estimation signal ,in, represents the target estimation signal of the kth path group, and its expression is: ； in, Indicates the target signal based on the fth path and target mapping matrix The target estimated signal on the u-th path is obtained, 1≤f≤n k , 1≤u≤n k ;1≤k≤N; Determine the damage scattering signal at the target temperature, based on the target signal and target estimation signal , calculate the damage scattering signal ,in, represents the damage scattering signal of the kth path group, and its expression is: ； Among them, 1≤k≤N; Take the path f and path u in the kth path group and describe the target estimation signal calculation process in detail: For the target mapping matrix , where the oth column data [i o ,j o ] T represents the i-th target signal on path f o element and the jth element of the target signal on path u o elements in the time domain; the target signal on path f The i-th o Elements are assigned as signals The jth o elements, that is, the target signal on path u is obtained Target estimation signal matched in time domain .

2. The method according to claim 1, characterized in that The step 1 comprises: A structural health monitoring system is constructed using a piezoelectric sensor array, a Lamb wave detector, a host computer, and a structure to be tested. The piezoelectric sensor array consists of L piezoelectric sensors, which are evenly attached to the surface of the structure to be tested in a circular or square shape with equal spacing. They are used to transmit and receive Lamb wave signals from the structure to be tested. The number of sampling points used by the Lamb wave detector to obtain the signal is represented by m. One end of the Lamb wave detector is connected to the sensor and the other end is connected to the host computer, which is responsible for transmitting the excitation and collected Lamb wave signals. The host computer is responsible for setting the excitation signal parameters and displaying and storing the collected signals. The structure to be tested is a flat plate or circular tubular structure made of aluminum alloy or carbon fiber. Group the sensing paths within the monitoring area: a. The straight line connecting the two different sensors is called a sensing path. For a piezoelectric sensing array consisting of L piezoelectric sensors, the number of sensing paths is b. According to the direction of the sensing path, all the sensing paths are divided into N groups: path1 to pathN; the directions of the sensing paths in the same group are consistent, and the directions of the sensing paths in different groups are different; for the kth path group, if Represents the number of sensing paths contained within it, and the total number of sensing paths is , 1≤k≤N.

3. The method according to claim 1, characterized in that The step 2 includes: Under the reference temperature, the healthy Lamb wave signals are collected and grouped, with room temperature as the reference temperature T r , using the structural health monitoring system to collect T r The Lamb wave signal at the temperature and the healthy state of the structure to be measured is recorded as the reference signal SR; according to the grouping of the sensing path, the reference signal SR is divided into N groups, which are recorded as to ; For the kth group of reference signals , which contains A column vector of dimension m×1, using Indicates that , The dimension is m×1; ;1≤k≤N; At the reference temperature, the reference mapping matrix within the sensing path group is constructed. For the reference signal SR, the dynamic time warping DTW algorithm is used to calculate the mapping matrix between the reference signals within each path group, which is recorded as the reference mapping matrix. The reference mapping matrix is expressed as Indicates that represents the mapping matrix between reference signals in the kth path group.

4. The method according to claim 3, characterized in that Take the reference signal of the kth path group , specifically introduces the mapping matrix between reference signals in the kth path group based on the DTW algorithm The build process: For the signal and , first, construct an m×m dimensional matrix D to represent and The Euclidean distance of any element in the matrix , 1≤i≤m, 1≤j≤m; Then, the cumulative distance matrix C is calculated according to formula (1): m×m , where parameter β ≥ 1. Parameter β is the bending coefficient introduced to avoid over-compensation of the signal and to impose nonlinear constraints on the signal: （1）； After obtaining the cumulative distance matrix C, the last element c of the matrix mm Start backtracking along the minimum path to c 11 , the selection principle of each element on the minimum path is: if c ij is an element on the minimum path, then the next element on the path is The smallest one in the cumulative distance matrix C is obtained by analogy. mm to c 11 The minimum path P; According to the row and column number set of each element on the minimum path P, the signal To signal The mapping matrix , m ≤ q ≤ 2m+1; mapping matrix It is a two-dimensional matrix with 2 rows, where the data in the oth column [i o ,j o ] T express The i-th o Elements and The jth o Elements Match, that is, if The i-th o Elements Assigned as The jth o elements, and the signal Matched estimated signal ; Repeat steps - Get the reference signal on any path f To the reference signal on path u Mapping matrix , and then get the reference signal The mapping matrix , whose expression is: 。 5. The method according to claim 1, wherein The step 3 includes: At the target temperature, the damaged Lamb wave signals are collected and grouped, and the ambient temperature of the structure to be tested is recorded as the target temperature T o ,and ; Use the structural health monitoring system to collect T o The Lamb wave signal at the temperature and the damage state of the structure to be measured is recorded as the target signal SO; according to the grouping of the sensing path, the target signal SO is divided into N groups, which are recorded as to ; For the kth group of reference signals , which contains A column vector of dimension m×1, using Indicates that , The dimension is m×1; ;1≤k≤N; At the target temperature, the target mapping matrix in the sensing path group is constructed, and the target temperature T o Next, the mapping matrix between Lamb wave signals in each path group when the structure to be tested is in a healthy state is recorded as the target mapping matrix ,in The target mapping matrix of the kth path group is expressed as: ； in, Represents the mapping matrix from the Lamb wave signal of path i to the Lamb wave signal of path j at the target temperature, 1≤i≤n k , 1≤j≤n k ; When detecting ambient temperature changes, the reference mapping matrix and the target mapping matrix are not equal, that is, ; The deviation between the target mapping matrix WO and the reference mapping matrix WR is recorded as the correction matrix , whose expression is: ； in, , represents the kth path group, target mapping matrix With the reference mapping matrix The deviation between ; Represents the target mapping matrix With the reference mapping matrix Deviation between Based on the correction matrix and the reference mapping matrix , calculate the target mapping matrix ,Right now .

6. The method according to claim 5, characterized in that Take path 1 and path 2 in the kth path group and introduce the correction matrix in detail The calculation process: For the signal and , based on the DTW algorithm, we can get arrive The mapping matrix is expressed as: , m ≤ q ≤ 2m+1; According to formula (2), the correction matrix is obtained , that is, the correction matrix The first row of is numerically equal to the matrix The difference between the second row and the first row, The second row of is numerically equal to Compared with the path The product of is expressed as: （2）； Among them, the path ratio is the ratio of the length of path 2 to the length of path 1; Repeat the above process to obtain the correction matrix of the kth path group , and then get the correction matrix .

7. The method according to claim 1, characterized in that The step 5 comprises: h. Extract the damage location features based on the orthogonal matching pursuit (OMP) algorithm, and then calculate the damage scattering signal. , calculate the damage characteristic matrix ,in, represents the damage characteristic of the kth path group, and its expression is: ； Where, 1 ≤ k ≤ N; For the kth path group, Indicates the scattered signal based on the OMP algorithm The damage characteristics obtained by sparse decomposition are calculated as follows: I. Construct an overcomplete dictionary D, whose atoms correspond to wave packet basis functions with different time delays, expressed as: （3） Where T is the duration of Lamb wave signal; m 1 is the number of atoms, which is numerically equal to the number of sampling points of the Lamb wave detector to obtain the signal; is the Lamb wave basis function; II. Parameter initialization: the current number of iterations u = 0, the initial residual , support set ; III. Calculate the current residual The inner product with all atoms in dictionary D, the most relevant atom index according to formula (4) , and update the support set: ; The expression is: （4）； in, is the vth column of dictionary D, i.e. the vth atom; IV. Update the sparse coefficient vector: , For the dictionary The submatrix composed of the columns of the index updates the residual: ; V. Order , if u=4, the process ends, otherwise go to step III; According to the above process, we get the support set Contains four elements, assigned to , a sparse coefficient vector There are four non-zero elements in , which are assigned to ; i. Construct the search point location feature matrix within the monitoring area According to the damage identification accuracy requirements, the monitoring area of the structure to be tested is discretized into r×g search points with equal spacing. The position coordinate matrix PS of the search points is expressed as: ； in, Indicates the The position coordinates of the search points, 1 ≤ i ≤ r, 1 ≤ j ≤ g; Calculate the feature matrix of the search point position on the sensing path ,in, Represents the position feature matrix of the search point on the path within the kth path group, that is, , represents the number of sensing paths in the kth path group; represents the position characteristics of the search point on the jth path in the kth path group, It is a two-dimensional matrix with dimensions r×g, and its expression is: （5） Where, 1 ≤ k ≤ N; 1 ≤ j ≤ , represents the coordinates of the excitation sensor on the jth path; represents the coordinates of the receiving sensor on the jth path; Represents the search point coordinates PS and The Euclidean distance of , dimension is r×g; Represents the search point coordinates PS and The Euclidean distance of , dimension is r×g; It represents the propagation speed of Lamb wave in the structure to be measured; j. Calculate the path damage probability and determine the damage location. The total damage probability within the monitoring area is: , where N is the number of groups in the sensing path; PD k is the damage probability of the kth path group, and PD is calculated according to formula (6): k : （6） in, is the number of sensing paths in the kth path group; According to the characteristics of the damage location and search point location features The damage probability of sensing path j in the kth path group is calculated according to formula (7): : （7） in, is the time threshold, which is set to 1.5 times the Lamb wave excitation period; Total damage probability is a two-dimensional matrix of dimension r×g, Indicates the Search point locations The probability of damage occurring at the monitoring area is the maximum damage probability concentration area, where r and g are the number of rows and columns of discrete points in the monitoring area, respectively.

Images