A shock positioning method based on signal arrival time accurate estimation

Abstract

The application provides an impact positioning method based on accurate estimation of signal arrival time, and belongs to the technical field of structural health monitoring, and comprises the following steps: collecting a dynamic strain signal generated by an impact event by using a fiber grating sensor network, and performing adaptive modal decomposition on the original signal by using a variational modal decomposition method; combining a short-time Fourier transform time-frequency technology, dynamically optimizing a time-frequency analysis window function, accurately extracting the arrival time of the impact signal at each sensing node, and realizing high-precision spatial positioning of the impact source based on a time difference positioning algorithm. The application is suitable for impact identification and positioning of key composite material plate and shell structures in the fields of aerospace and rail transit, and has the characteristics of strong real-time performance and outstanding anti-interference capability.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring technology, and in particular to an impact location method based on accurate estimation of signal arrival time. Background Technology

[0002] In fields such as aerospace and rail transportation, where critical composite material plate and shell structures are widely used, these structures are susceptible to low-velocity damage, such as impacts from foreign objects, during service. While these impact events may not immediately lead to structural failure, they can induce potential cracks or delamination defects, thereby affecting safety.

[0003] Current mainstream methods, such as piezoelectric sensor arrays and electromagnetic acoustic wave technology, suffer from problems such as complex deployment and severe electromagnetic interference. Fiber Bragg grating (FBG) sensors, due to their advantages such as resistance to electromagnetic interference, small size, and versatility in composite applications, have become an important direction for structural health monitoring.

[0004] Therefore, this invention proposes an impact location method based on accurate estimation of signal arrival time. Summary of the Invention

[0005] This invention provides an impact location method based on accurate signal arrival time estimation. It adaptively separates noise and impact signals using variational mode decomposition (VMD) and combines dynamic window-length STFT time-frequency analysis to accurately extract the time of arrival (TOA), significantly improving the accuracy of low-velocity impact location for composite materials. The method employs a VMD algorithm with optimized mode number K and penalty factor α to effectively suppress environmental interference and adaptively adjusts the STFT analysis window length based on the impact pulse width characteristics, achieving accurate TOA extraction with an error of less than 0.5 μs. Finally, a high-precision TDOA algorithm and a hyperbolic impact location model are used to locate low-velocity impacts on carbon fiber composite materials.

[0006] This invention provides an impact location method based on accurate estimation of signal arrival time, comprising:

[0007] S1: Obtain basic structural information, material performance data and impact monitoring area distribution information of the composite material structure under test. Combine the directional characteristics of shock wave propagation in the composite material structure under test and the time delay positioning requirements, attach multiple fiber optic grating sensors at regular intervals on the surface of the composite material plate to construct a two-dimensional distributed sensing network covering the key area.

[0008] S2: In a laboratory environment, standard impact loads are applied at multiple preset locations using a controllable impact loading device. The strain-time history signals of each sensing node in the two-dimensional distributed sensing network are collected synchronously. The collected signals are adaptively decomposed using a variational mode decomposition algorithm. By adjusting the mode number K and the penalty factor α, noise interference and effective impact response in the signal are distinguished. Time-frequency analysis is performed using a dynamic window length STFT to calibrate the transmission delay and time-frequency characteristics of each sensing node and retain the effective impact mode signals.

[0009] S3: For the retained effective impact mode signals, the multi-scale joint analysis method is used to identify the impact arrival time and obtain the TDOA matrix between sensing nodes;

[0010] S4: Construct a hyperbolic impact positioning trajectory with the reference node position and any other sensor node position as the focus based on the TDOA matrix between sensor nodes, and solve the intersection of the hyperbolic impact positioning trajectory to determine the impact positioning result of the impact source. Each set of TDOA values ​​corresponds to a hyperbolic solution set, and multiple nodes are combined to form multiple hyperbolas.

[0011] S5: Establish an error correction mechanism for impact positioning results and achieve real-time visualization display.

[0012] Preferably, the basic structural information of the composite material plate includes dimensional parameters, thickness distribution, layup method, boundary constraints, and typical working conditions. The material performance data includes the material's elastic modulus, shear modulus, Poisson's ratio, density, and anisotropic wave velocity characteristics. The impact monitoring area distribution information is determined based on the operating conditions, stress environment, and historical impact records to identify key areas where impacts may occur.

[0013] Preferably, a multi-scale joint analysis method is used to identify the impact arrival time, obtaining the TDOA matrix between sensing nodes, including:

[0014] In the time domain, the arrival time of the first local maximum point of the strain signal change rate is solved by differential calculation, and in the frequency domain, short-time Fourier transform with dynamic window length is used for verification, and the time difference between the time domain and frequency domain detection results is determined.

[0015] The time difference between the time domain and frequency domain detection results is less than When the time domain and frequency domain detection results are taken as the final TOA, the sensor node that receives the impact signal earliest is taken as the reference node. The arrival time difference between the corresponding arrival time and the signal reception time of each other sensor node is calculated, and the TDOA matrix between sensor nodes is calculated.

[0016] Preferably, an error correction mechanism for impact positioning results is established and real-time visualization is achieved, including:

[0017] The obtained impact location results are compared and analyzed with historical impact test data and modeling parameters of composite material structures to construct an error distribution database;

[0018] Based on the error distribution database and using the least squares method to establish an error function, the impact positioning results are fitted and corrected to obtain the corrected coordinates.

[0019] The corrected impact source coordinates are synchronized with real-time sensor network data, and the error-corrected impact source coordinates are displayed in real time on the monitoring platform.

[0020] Preferably, the error function in Let N be the coordinates of the i-th sensor node, and N be the total number of sensor nodes. The coordinates of the impact source are: Let be the position coordinates of the reference node, and It is the distance difference, and speed of transmission; This is the difference between the arrival time of the reference node and the time of the i-th sensor node.

[0021] Preferably, time-frequency analysis is performed using a dynamically time-windowed STFT to calibrate the propagation delay and time-frequency characteristics of each sensing node, retaining the effective impulse mode signal, including:

[0022] In composite plate test specimens Impact tests were conducted at preset impact locations, and standard impact loads were applied using a controllable impact loading device, while the dynamic strain response signals of all sensing nodes were collected simultaneously.

[0023] Record No. The raw strain signals collected by each sensing node Where N is the total number of sensor nodes;

[0024] Signals for each node Variational mode decomposition is performed to obtain eigenmode functions in, For the signal of the i-th sensing node The kth eigenmode function obtained after variational mode decomposition;

[0025] The required impact modes are selected based on the kurtosis criterion, and the initial mode set is denoted as . And reconstruct the denoised signal ;

[0026] right Verification will be conducted to determine if it is a valid impact signal. If the verification passes, Determined as a valid impact mode signal And retain.

[0027] Preferably, determining the time difference between the time-domain and frequency-domain detection results includes:

[0028] Effective impact mode signals obtained through variational mode decomposition and screening ,and As the object of analysis;

[0029] The differential acquisition method is used to solve for the arrival time in the time domain of the signal. ;

[0030] For each selected valid impact mode signal The frequency domain arrival time is obtained by performing short-time Fourier time-frequency analysis. The short-time Fourier transform (STFT) is defined as follows:

[0031] ;

[0032] in, For window functions, the window length is... According to the The impact signal characteristics of each sensing node are adaptively set. The standard deviation of background noise;

[0033] Sure and The difference.

[0034] Preferably, the method further includes: using the fminsearc function to correct and optimize the error function and using the Nelder-Mead simplex method to perform gradient-free optimization of the error function.

[0035] Compared with the prior art, the beneficial effects of this application are as follows:

[0036] To achieve high-precision monitoring of invisible damage caused by low-energy impacts during the service life of composite materials, a fiber Bragg grating sensor network is strategically attached to the surface of carbon fiber composite materials or embedded within them. This method effectively addresses the problem of inaccurate shock wave arrival time by acquiring the time-domain signal response values ​​of the fiber Bragg grating sensors and combining them with a hyperbolic positioning algorithm, thereby enabling accurate prediction of the impact location in composite laminates. Experimental results show that this scheme not only improves the accuracy of impact positioning but also achieves parameterized identification of the health status of composite materials, providing reliable technical support for the intelligent monitoring of composite materials.

[0037] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0038] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0039] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0040] Figure 1 This is a flowchart of the hyperbolic positioning method based on the arrival time of the impact signal according to the present invention;

[0041] Figure 2 The strain measurement data obtained from experiments using different sensors in the examples are shown.

[0042] Figure 3 The strain signal curves for each mode in the embodiment are shown below;

[0043] Figure 4 The energy diagrams of the strain signals for each mode in the embodiment are shown.

[0044] Figure 5 This refers to the STFT of each sensor's modal impact signal in the embodiment;

[0045] Figure 6 The arrival time of the impact signals from each sensor in the embodiment;

[0046] Figure 7 Image showing the precise location results of impact signals used for health monitoring of composite material structures. Detailed Implementation

[0047] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0048] like Figures 1 to 7 As shown, the present invention provides an impact location method based on accurate estimation of signal arrival time, comprising:

[0049] S1: Obtain basic structural information, material performance data and impact monitoring area distribution information of the composite material structure under test. Combine the directional characteristics of shock wave propagation in the composite material structure under test and the time delay positioning requirements, attach multiple fiber optic grating sensors at regular intervals on the surface of the composite material plate to construct a two-dimensional distributed sensing network covering the key area.

[0050] In this embodiment, during the deployment of the fiber Bragg grating sensor, the resolution capability of the shock wave arrival time difference is improved through a deployment spacing optimization strategy. The deployment spacing optimization includes the following steps:

[0051] S11: Based on the composite material plate structure dimensions obtained in S1, establish a finite element model, extract the response of the FBG sensor, and set strain acquisition aggregation points at typical sensor deployment locations to record the strain-time variation curves of corresponding node elements to simulate the signal reception process of the actual FBG sensor. Determine the shock wave velocity range and the minimum detectable time difference. ;

[0052] S12: According to And the shock wave propagation velocity v, calculate the theoretical minimum deployment spacing ;

[0053] S13: Set the actual deployment spacing d to... Where m is a safety redundancy coefficient greater than 1, ensuring that the system can still accurately extract TDOA even in the presence of synchronization errors and structural attenuation, thereby improving local positioning accuracy.

[0054] S2: In a laboratory environment, standard impact loads are applied at multiple preset locations using a controllable impact loading device. The strain-time history signals of each sensing node in the two-dimensional distributed sensing network are collected synchronously. The collected signals are adaptively decomposed using a variational mode decomposition algorithm. By adjusting the mode number K and the penalty factor α, noise interference and effective impact response in the signal are distinguished. Time-frequency analysis is performed using a dynamic window length STFT to calibrate the transmission delay and time-frequency characteristics of each sensing node and retain the effective impact mode signals.

[0055] In this embodiment, the shock wave signal received by each fiber Bragg grating node is identified, and the original signal is decomposed by VMD in the following manner:

[0056] S21: Extract the impact strain signal of each fiber Bragg grating sensor and perform baseline correction to eliminate low-frequency interference such as environmental temperature drift;

[0057] S22: Perform VMD decomposition on the strain signal of each sensor, setting the mode number K = 2~5 and the penalty factor α = 2000~5000, to obtain... Each intrinsic mode function (IMF)

[0058] Calculate the energy of each mode and perform FFT analysis on its spectral characteristics to screen out the modes that contain the main impact information (usually the second or third mode).

[0059] S23: Perform a short-time Fourier transform (STFT) on the selected effective modes, using an adaptive Hanning window (window length...). Based on the dynamic adjustment of the impact pulse width, the signal energy change is detected in the time-frequency joint domain to analyze each modal signal and extract the effective impact signal.

[0060] S3: For the retained effective impact mode signals, the multi-scale joint analysis method is used to identify the impact arrival time and obtain the TDOA matrix between sensing nodes;

[0061] S4: Construct a hyperbolic impact positioning trajectory with the reference node position and any other sensor node position as the focus based on the TDOA matrix between sensor nodes, and solve the intersection of the hyperbolic impact positioning trajectory to determine the impact positioning result of the impact source. Each set of TDOA values ​​corresponds to a hyperbolic solution set, and multiple nodes are combined to form multiple hyperbolas.

[0062] The hyperbolic model for impact source localization, based on the TDOA principle, is established and solved as follows:

[0063] S41: Based on the arrival times of multiple fiber optic grating nodes obtained in S3, and taking the sensor that receives the signal first as the reference, construct the arrival time difference between each sensor, and construct a hyperbolic impact positioning trajectory model with the reference node position and the position of any other sensor node as the focus.

[0064] Let the planar position of the sensor be... The location of the impact source is For unknown parameters, wave speed The arrival time of each sensor is The sensor that received the earliest signal was selected as the reference sensor, and its number was [number missing]. Its arrival time is .

[0065] Because the wave velocity is uniform and constant, the signal propagation time is directly proportional to the distance:

[0066]

[0067] in This represents the Euclidean distance.

[0068] The time difference is defined with reference to a benchmark sensor:

[0069]

[0070] Corresponding distance difference:

[0071]

[0072] Right now

[0073]

[0074] S42: Each set of TDOA values ​​corresponds to a hyperbola solution set. Multiple nodes can be combined to form multiple hyperbolas. The location coordinates of the impact source are determined by solving the intersection of these hyperbolas.

[0075] The distance difference is the difference in distance between two points to an unknown point, and it defines a set of points on a hyperbola.

[0076]

[0077] Expand:

[0078]

[0079] This equation defines the... A hyperbola with focus centered at foci.

[0080] for There are multiple such hyperbolas. The unknown point P must satisfy the equations of all hyperbolas.

[0081] S43: By combining the boundary conditions and wave velocity characteristics of the composite material plate, the wave propagation path constraint within the structure is established, thereby improving the accuracy and reliability of positioning.

[0082] S5: Establish an error correction mechanism for impact positioning results and achieve real-time visualization display.

[0083] Preferably, the basic structural information of the composite material plate includes dimensional parameters, thickness distribution, layup method, boundary constraints, and typical working conditions. The material performance data includes the material's elastic modulus, shear modulus, Poisson's ratio, density, and anisotropic wave velocity characteristics. The impact monitoring area distribution information is determined based on the operating conditions, stress environment, and historical impact records to identify key areas where impacts may occur.

[0084] Preferably, a multi-scale joint analysis method is used to identify the impact arrival time, obtaining the TDOA matrix between sensing nodes, including:

[0085] In the time domain, the arrival time of the first local maximum point of the strain signal change rate is solved by differential calculation, and in the frequency domain, short-time Fourier transform with dynamic window length is used for verification, and the time difference between the time domain and frequency domain detection results is determined.

[0086] The time difference between the time domain and frequency domain detection results is less than When the time domain and frequency domain detection results are taken as the final TOA, the sensor node that receives the impact signal earliest is taken as the reference node. The arrival time difference between the corresponding arrival time and the signal reception time of each other sensor node is calculated, and the TDOA matrix between sensor nodes is calculated.

[0087] In this embodiment, the sensor r that detects the impact signal earliest is selected, and its arrival time is... This serves as a benchmark. The arrival time differences of other sensors relative to the benchmark sensor are calculated. in, The arrival time of the impact signal detected by the i-th sensing node.

[0088] In this embodiment, the construction of the TDOA matrix provides a guarantee for the subsequent input data based on the hyperbolic positioning algorithm.

[0089] Preferably, an error correction mechanism for impact positioning results is established and real-time visualization is achieved, including:

[0090] The obtained impact location results are compared and analyzed with historical impact test data and modeling parameters of composite material structures to construct an error distribution database;

[0091] Based on the error distribution database and using the least squares method to establish an error function, the impact positioning results are fitted and corrected to obtain the corrected coordinates.

[0092] The corrected impact source coordinates are synchronized with real-time sensor network data, and the error-corrected impact source coordinates are displayed in real time on the monitoring platform.

[0093] In this embodiment, the historical impact test data includes at least 20 sets, with each set covering impact energy from 5J to 20J, and the impact location covering uniform grid points (grid spacing 50mm) of the monitoring area; outliers are removed when the positioning error exceeds 3 times the standard deviation of the mean.

[0094] Preferably, the error function in, Let N be the coordinates of the i-th sensor node, and N be the total number of sensor nodes. The coordinates of the impact source are: Let be the position coordinates of the reference node, and It is the distance difference, and For the speed of transmission; This is the difference between the arrival time of the reference node and the time of the i-th sensor node.

[0095] Preferably, time-frequency analysis is performed using a dynamically time-windowed STFT to calibrate the propagation delay and time-frequency characteristics of each sensing node, retaining the effective impulse mode signal, including:

[0096] In composite plate test specimens Impact tests were conducted at preset impact locations, and standard impact loads were applied using a controllable impact loading device, while the dynamic strain response signals of all sensing nodes were collected simultaneously.

[0097] Record No. The raw strain signals collected by each sensing node Where N is the total number of sensor nodes;

[0098] Signals for each node Variational mode decomposition is performed to obtain eigenmode functions in, For the first The signal of each sensor node The kth eigenmode function obtained after variational mode decomposition;

[0099] The required impact modes are selected based on the kurtosis criterion, and the initial mode set is denoted as . And reconstruct the denoised signal ;

[0100] right Verification will be conducted to determine if it is a valid impact signal. If the verification passes, Determined as a valid impact mode signal And retain.

[0101] In this embodiment, the verification criteria are: the kurtosis value of the reconstructed signal is greater than 3, the kurtosis value of the background noise is less than 1.5, and the signal amplitude is greater than 3 times the standard deviation of the background noise.

[0102] Preferably, determining the time difference between the time-domain and frequency-domain detection results includes:

[0103] Effective impact mode signals obtained through variational mode decomposition and screening and As the object of analysis;

[0104] The differential acquisition method is used to solve for the arrival time in the time domain of the signal. ;

[0105] For each selected valid impact mode signal The frequency domain arrival time is obtained by performing short-time Fourier time-frequency analysis. The short-time Fourier transform (STFT) is defined as follows:

[0106]

[0107] in, For window functions, the window length is... According to the The impact signal characteristics of each sensing node are adaptively set. The standard deviation of background noise;

[0108] Sure The difference.

[0109] In this embodiment, For Hanning window, and Based on the pulse width of the impact signal and background noise standard deviation Confirmed, specifically: in, Extracted from the modal signals after VMD decomposition. Calculations are performed by collecting strain signals under no-impact conditions.

[0110] Preferably, the method further includes: using the fminsearc function to correct and optimize the error function and using the Nelder-Mead simplex method to perform gradient-free optimization of the error function.

[0111] In this embodiment, initialization is performed using an initial solution. Starting from the triangle, construct the initial simplex (triangle search domain):

[0112] ;

[0113] Iterative search, for the k-th iteration:

[0114] Reflection operation: Calculate the highest point Regarding the center of gravity Reflection point ;

[0115] like Then it contracts inward:

[0116] Convergence condition: Termination occurs when the volume of the simplex is less than the threshold ϵ, i.e.: .

[0117] In this embodiment, the initial search point of the fminsearc function is set to the initial localization result. The initial point set is: {(1.1x0,y0),(0.9x0,y0),(x0,1.1y0),(x0,0.9y0)}; the Nelder-Mead algorithm uses an initial simplex with 3 vertices, the vertex spacing is 50% of the initial positioning error, and the convergence threshold is when the function value change is less than 1e-8 and the simplex volume is less than 1e-6.

[0118] The beneficial effects of the above technical solution are as follows: To achieve high-precision monitoring of invisible damage caused by low-energy impacts during the service of composite materials, a fiber optic grating sensor network is rationally attached to or embedded within the carbon fiber composite material. This method effectively solves the problem of inaccurate shock wave arrival time by acquiring the time-domain signal response values ​​of the fiber optic grating sensors and combining them with a hyperbolic positioning algorithm, thereby achieving accurate prediction of the impact location of the composite laminate. Experimental results show that this solution not only improves the accuracy of impact positioning but also enables parameterized identification of the health status of composite materials, providing reliable technical support for the intelligent monitoring of composite materials.

[0119] This invention provides an impact location method based on accurate signal arrival time estimation. Combining the composite material plate structural dimensions obtained in S1, a finite element model is established, and the response of the FBG sensor is extracted. Strain acquisition points need to be set at typical sensor deployment locations to record the strain-time variation curves of corresponding node elements, simulating the signal reception process of the actual FBG sensor, determining the shock wave velocity range, and identifying the minimum detectable time difference. ,include:

[0120] The dimensional parameters, layup method, and boundary constraints of the composite material plate under test are obtained, and a three-dimensional solid finite element model is established. The element type of the three-dimensional solid finite element model adopts eight-node reduced integral solid elements. During mesh generation, an FBG equivalent sensing region element mapping rule is introduced, that is, the actual sensing length Ls of the FBG sensor is mapped to the equivalent element size le of the finite element mesh, satisfying le = Ls / ne, where ne is the number of elements in the sensing region, and ne ≥ 3 to ensure acquisition accuracy. Furthermore, the mesh size of the non-sensing region is no greater than twice le to ensure that the strain gradient on the shock wave propagation path can be effectively captured. The dimensional parameters include: length L1, width W1, and thickness H1. The layup method includes: layup angle sequence. With single-layer thickness h1;

[0121] Based on the identified key areas for impact monitoring, at least five typical sensor deployment locations are selected in the finite element model. A strain acquisition aggregation point Pj is set at each typical location. The strain acquisition aggregation point Pj consists of the center nodes of all elements within a circular area with a radius not exceeding Ls / 2 surrounding the deployment location, and the number of nodes at each aggregation point is not less than eight, in order to simulate the average strain sensing characteristics of the FBG sensor during actual operation. The five typical sensor deployment locations are related to the edge, center, and impact-prone parts of the coverage area.

[0122] Three standard impact loads of different energy levels were selected within the key area of ​​impact monitoring and applied to the model's preset impact point Qk, where k=1,2,3. The loading time history of each impact load was represented by a half-sine wave function, where the loading peak time was... Total duration Synchronously output the strain-time history curves of all nodes in each strain acquisition set point Pj. And m1 is the node number of the set point Pj, and calculate the average strain-time curve for each set point Pj. ,in, The total number of nodes at the aggregation point Pj is used to simulate the signal reception process of the actual FBG sensor at the corresponding deployment location using this average curve.

[0123] The propagation velocity range of the shock wave within the composite material plate [Vmin, Vmax] is calculated as follows:

[0124] Modal analysis was performed on the three-dimensional solid finite element model to extract the natural frequencies f1 to f5 of the first five in-plane vibration modes. Combined with the elastic modulus E1, shear modulus G1, and density ρ1 of the composite material, the reference wave velocity was calculated. ,in, Poisson's ratio of the material;

[0125] Select any two strain acquisition points Pa and Pb (a≠b), and measure their straight-line distance in the finite element model. ,extract Calculate the wave velocity between the two points based on the peak strain arrival times ta and tb. ;

[0126] The calculation was performed for all pairwise combinations of points (a total of C(5,2)=10 groups). After statistical analysis and removal of outliers exceeding three standard deviations, the minimum remaining value is taken as Vmin, and the maximum value as Vmax. A ply angle correction factor is also introduced. The corrected wave velocity range is obtained by averaging the ply angle sequence. ;

[0127] Based on the determined corrected wave velocity range, combined with the sampling rate fs of the FBG sensing system and the strain measurement error caused by system noise, and preset positioning resolution requirements Calculate the minimum detectable time difference Where SNR is the signal-to-noise ratio of the FBG sensing system, and it must satisfy the following conditions:

[0128] If the calculation yields Adjust Until .

[0129] In this embodiment, the impact energy can be 5J, 10J, or 15J.

[0130] In this embodiment, the results were obtained through a laboratory static calibration experiment. .

[0131] In this embodiment, .

[0132] In this embodiment, the results were obtained through a dynamic signal testing experiment. .

[0133] In this embodiment, the three-dimensional solid finite element model is a numerical model that discretizes the geometry and material properties (such as elastic modulus and density) of the composite material plate into a large number of three-dimensional small elements. The strain and displacement response under impact are simulated by solving the element mechanical equations. The eight-node reduced integral solid element is a finite element with eight vertex nodes. It adopts a reduced integral algorithm to balance the calculation accuracy and efficiency, avoid numerical problems such as volume self-locking, and is suitable for simulating wave propagation of composite materials.

[0134] In this embodiment, if the effective sensing length of the FBG ,set up ,but The grid size in the non-sensing area is set to ≤4mm so that the grid can accurately reflect strain changes when the shock wave passes through the sensing area.

[0135] In this embodiment, with For example, draw a circle with a radius of 4mm around the deployment location and select 10 unit center nodes as convergence point nodes; when the impact occurs, average the strain-time curves of the 10 nodes, and the resulting average curve is the signal actually received by the simulated FBG.

[0136] In this embodiment, the load function is: Time; 0, At that time, among them, The impact on the structure is simulated by calculating the impact energy and the area of ​​impact.

[0137] In this embodiment, modal analysis extracts the inherent vibration characteristics of the structure (such as the first 5 natural frequencies and mode shapes); the reference wave velocity is a theoretical wave velocity calculated based on the elastic modulus, shear modulus, density, and Poisson's ratio of the composite material, and is used for wave velocity range calibration.

[0138] In the embodiment, if the ply angle sequence is [0°, 60°, −60°, 90°], then Calculated If the original wave velocity range is [3000m / s, 3500m / s], the corrected range is [2991m / s, 3490m / s].

[0139] The beneficial effects of the above technical solution are as follows: by establishing a finite element model that fits the sensing characteristics of FBG, the receiving process of FBG to the impact signal is accurately simulated. By combining modal analysis, multi-energy impact testing and ply angle correction, the range of shock wave propagation speed is accurately quantified. Furthermore, by coupling the calculation of sampling rate, noise, positioning resolution and minimum detectable time difference, key parameters are provided for the subsequent optimized deployment of FBG sensor network, ensuring the accuracy and reliability of impact positioning based on time difference of arrival.

[0140] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An impact location method based on accurate estimation of signal arrival time, characterized in that, include: S1: Obtain basic structural information, material performance data and impact monitoring area distribution information of the composite material structure under test. Combine the directional characteristics of shock wave propagation in the composite material structure under test and the time delay positioning requirements, attach multiple fiber optic grating sensors at regular intervals on the surface of the composite material plate to construct a two-dimensional distributed sensing network covering the key area. S2: In a laboratory environment, standard impact loads are applied at multiple preset locations using a controllable impact loading device. The strain-time history signals of each sensing node in the two-dimensional distributed sensing network are collected synchronously. The collected signals are adaptively decomposed using a variational mode decomposition algorithm. By adjusting the mode number K and the penalty factor α, noise interference and effective impact response in the signal are distinguished. Time-frequency analysis is performed using a dynamic window length STFT to calibrate the transmission delay and time-frequency characteristics of each sensing node and retain the effective impact mode signals. S3: For the retained effective impact mode signals, a multi-scale joint analysis method is used to identify the impact arrival time, obtaining the TDOA matrix between sensing nodes, including: In the time domain, the arrival time of the first local maximum point of the strain signal change rate is solved by differential calculation, and in the frequency domain, short-time Fourier transform with dynamic window length is used for verification, and the time difference between the time domain and frequency domain detection results is determined. When the time difference between the time domain and frequency domain detection results is less than 10μs, the average of the time domain and frequency domain detection results is taken as the final TOA. The sensor node that receives the impact signal earliest is taken as the reference node. The arrival time difference between the corresponding arrival time and the signal reception time of each other sensor node is calculated, and the TDOA matrix between sensor nodes is calculated. S4: Construct a hyperbolic impact positioning trajectory with the reference node position and any other sensor node position as the focus based on the TDOA matrix between sensor nodes, and solve the intersection of the hyperbolic impact positioning trajectory to determine the impact positioning result of the impact source. Each set of TDOA values ​​corresponds to a hyperbolic solution set, and multiple nodes are combined to form multiple hyperbolas. S5: Establish an error correction mechanism for impact positioning results and achieve real-time visualization display.

2. The impact location method based on accurate signal arrival time estimation according to claim 1, characterized in that, The basic structural information of the composite material plate includes dimensional parameters, thickness distribution, layup method, boundary constraints, and typical working conditions. The material performance data includes the material's elastic modulus, shear modulus, Poisson's ratio, density, and anisotropic wave velocity characteristics. The impact monitoring area distribution information is the key areas where impacts may occur, determined based on the operating conditions, stress environment, and historical impact records.

3. The impact location method based on accurate signal arrival time estimation according to claim 1, characterized in that, Establish an error correction mechanism for impact positioning results and achieve real-time visualization, including: The obtained impact location results are compared and analyzed with historical impact test data and modeling parameters of composite material structures to construct an error distribution database; Based on the error distribution database and using the least squares method to establish an error function, the impact positioning results are fitted and corrected to obtain the corrected coordinates. The corrected impact source coordinates are synchronized with real-time sensor network data, and the error-corrected impact source coordinates are displayed in real time on the monitoring platform.

4. The impact location method based on accurate signal arrival time estimation according to claim 3, characterized in that, The error function ,in Let N be the coordinates of the i-th sensor node, and N be the total number of sensor nodes. The coordinates of the impact source are: Let be the position coordinates of the reference node, and It is the distance difference, and For the speed of transmission; This is the difference between the arrival time of the reference node and the time of the i-th sensor node.

5. The impact location method based on accurate signal arrival time estimation according to claim 1, characterized in that, Time-frequency analysis was performed using a dynamically timed STFT to calibrate the propagation delay and time-frequency characteristics of each sensing node, retaining effective impulse mode signals, including: Impact tests were conducted at M preset impact locations on the composite plate specimen. A standard impact load was applied using a controllable impact loading device, and dynamic strain response signals of all sensing nodes were collected simultaneously. Record the original strain signal acquired by the i-th sensing node. and Where N is the total number of sensor nodes; Signals for each node Variational mode decomposition yields K eigenmode functions. in, For the signal of the i-th sensing node The kth eigenmode function obtained after variational mode decomposition; The required impact modes are selected based on the kurtosis criterion, and the initial mode set is denoted as . And reconstruct the denoised signal ; right Verification will be conducted to determine if it is a valid impact signal. If the verification passes, Determined as a valid impact mode signal And retain.

6. The impact location method based on accurate signal arrival time estimation according to claim 5, characterized in that, Determine the time difference between the time-domain and frequency-domain detection results, including: Effective impact mode signals obtained through variational mode decomposition and screening ,and As the object of analysis; The differential acquisition method is used to solve for the arrival time of the signal in the time domain. ; For each selected valid impact mode signal The frequency domain arrival time is obtained by performing short-time Fourier time-frequency analysis. The short-time Fourier transform (STFT) is defined as follows: ； in, For window functions, the window length is... The settings are adaptively configured based on the impact signal characteristics of the i-th sensing node. The standard deviation of background noise; Sure and The difference.

7. The impact location method based on accurate signal arrival time estimation according to claim 3, characterized in that, Also includes: The error function is corrected and optimized using the fminsearc function, and the Nelder-Mead simplex method is used to perform gradient-free optimization of the error function.

Images