Damage detection method and system for composite material based on virtual time reversal

Through the virtual time inversion method, the virtual transfer function is constructed using frequency domain transformation, which solves the problem of signal reconstruction quality and accuracy in composite material damage detection, and realizes high-precision damage positioning and imaging, avoids errors in physical reversal process , improves detection efficiency.

CN118914375APending Publication Date: 2024-11-08SOUTHERN UNIVERSITY OF SCIENCE AND TECHNOLOGY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410935755.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the prior art, in the detection of composite material damage, the signal reconstruction quality and damage imaging accuracy are limited, and the damage position and damage situation cannot be accurately positioned, and the physical time reversal process is dependent on the physical time reversal process, with large errors.

Method used

Using the detection method based on virtual time inversion, virtual transfer function is constructed through frequency domain transformation, high-precision damage detection without baseline is realized, discrete Fourier transform converts the signal to the frequency domain, calculates the transfer function of the forward transmission path, and avoids the The physical inversion process is carried out, and the virtual time inversion process is performed.

Benefits of technology

It realizes high-precision damage detection and imaging, can accurately locate the damage location and measure the damage situation, reduces errors during physical reversal, and improves processing efficiency and signal reconstruction quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118914375A_ABST
    Figure CN118914375A_ABST
Patent Text Reader

Abstract

The invention provides a composite material damage detection method and system based on virtual time reversal. The method comprises the steps that a first sensor transmits an original input signal to a second sensor; the second sensor performs virtual time reversal on the original input signal to generate a reversal signal, and simulates and sends the reversal signal to the first sensor; the first sensor converts the inversion signal into a reconstruction signal, and calculates the similarity between the reconstruction signal and the original input signal so as to calculate a damage index; and completing damage positioning and damage imaging according to the damage index. According to the virtual time reversal operation based on frequency domain transformation, input and output signals are converted into a frequency domain through discrete Fourier transformation to calculate a transfer function of a forward transfer path, a virtual time reversal process is constructed, baseline-free and high-precision damage detection is achieved, and high-efficiency and high-precision damage detection is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of damage detection of composite material structures, and in particular to a damage detection method and system for composite materials based on virtual time reversal. Background Art

[0002] Reinforced carbon fiber composites are composite materials with good structure and performance. Reinforced carbon fiber composites containing epoxy resin interlayers have a sandwich structure, which is usually composed of two or more rigid panels separated by lightweight core materials (such as honeycombs or foams) through bonding and other methods. The separation between the panels enhances the overall structural strength and stability of the sandwich structure. However, due to low-speed impact, cyclic loads and harsh working environments, the sandwich structure is prone to various damages, which affect the overall performance of the reinforced carbon fiber sandwich composites.

[0003] Minor visible indentation damage is usually difficult to detect through visual inspection, and will lead to gradual degradation of structural performance over time. Existing damage detection methods, such as conventional time reversal methods, usually require the use of baseline information, while probabilistic damage detection methods based on amplitude attenuation usually rely on reference signals and physical inversion processes, which limits their applicability in practical engineering applications. In addition, the damage detection methods of the prior art still rely on the physical time reversal process of the signal, requiring at least two signal excitations to complete the physical time reversal process. The signal reconstruction quality and damage imaging accuracy are limited, with large errors, and it is impossible to accurately locate the damage position and damage condition. How to reduce the error in the physical reversal process and quickly and accurately locate the damage is a difficulty faced by current technology.

[0004] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0005] In view of this, the present application proposes a detection method and system for composite materials based on virtual time reversal. Through an improved virtual time reversal method, baseline-free detection is adopted to avoid the physical reversal process and errors. The virtual time reversal operation based on frequency domain transformation calculates the transfer function of the forward transmission path by converting the input and output signals into the frequency domain through discrete Fourier transform, constructs a virtual time reversal process, realizes baseline-free and high-precision damage detection, realizes efficient and high-precision damage detection, and solves the technical problems in the prior art that the signal reconstruction quality and damage imaging accuracy are limited, and the damage position and damage condition cannot be accurately located. High-precision damage imaging can be formed, and the damage position and damage condition can be accurately located.

[0006] The technical solution of this application is as follows:

[0007] The first aspect of the present application provides a damage detection method for composite materials based on virtual time reversal, comprising the following steps:

[0008] The first sensor transmits a raw input signal to the second sensor;

[0009] The second sensor generates an inversion signal by performing virtual time inversion on the original input signal, and sends the inversion signal to the first sensor;

[0010] The first sensor converts the inverted signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate a damage index;

[0011] Injury localization and injury imaging are completed according to the injury index.

[0012] Specifically, the second sensor generates an inversion signal by virtually time-reversing the original input signal, and simulates sending the inversion signal to the first sensor, including:

[0013] The second sensor receives the original input signal and generates a first response signal;

[0014] Constructing a virtual transfer function through frequency domain conversion according to the original input signal and the first response signal;

[0015] Based on the virtual transfer function, a virtual inversion signal is generated by the virtual time inversion process;

[0016] The process of sending the virtual inversion signal along an inversion path to the first sensor is simulated.

[0017] Specifically, the generating of a virtual inversion signal based on the virtual transfer function through the virtual time reversal processing comprises the steps of:

[0018] Performing a first dispersion term elimination on the first response signal to generate a first optimized signal;

[0019] The first optimized signal is subjected to discrete Fourier transform and inverted in the time domain to generate the virtual inverted signal.

[0020] Specifically, after calculating the transfer function according to the original input signal and the first response signal, the method further includes the following steps:

[0021] The transfer function is conjugated to obtain a time-reversed transfer function.

[0022] Specifically, the first sensor converts the inversion signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate the damage index, including the steps of:

[0023] Simulating a process in which the inversion signal is transmitted from the second sensor to the first sensor to obtain a virtual second response signal;

[0024] obtaining a reconstructed signal based on the second response signal;

[0025] The similarity between the reconstructed signal and the original input signal is calculated to calculate a damage index.

[0026] Specifically, the process of simulating the inversion signal being transmitted from the second sensor to the first sensor to obtain a virtual second response signal comprises the steps of:

[0027] The inversion signal is converted into the frequency domain by Fourier transformation, and then multiplied by the time reversal transfer function, and then the result of the multiplication is converted into the time domain by inverse Fourier transformation, so as to obtain the second response signal.

[0028] Specifically, obtaining the reconstruction signal based on the second response signal comprises the steps of:

[0029] performing a second dispersion term elimination on the second response signal to generate a second optimized signal;

[0030] Performing inverse discrete Fourier transform on the second optimized signal to perform time domain inversion to generate the reconstructed signal.

[0031] Specifically, the reinforced carbon fiber composite material is a reinforced carbon fiber epoxy resin sandwich composite material with a sandwich structure.

[0032] A second aspect of the present application is to provide a damage detection system for reinforced carbon fiber composite materials based on virtual time reversal, including a sensor network device and a data acquisition and processing device;

[0033] Specifically, the sensor network device includes a piezoelectric sensor array, the piezoelectric sensor array includes a plurality of piezoelectric sensors, each of the piezoelectric sensors is embedded with a microcontroller, and the piezoelectric sensors serve as Lamb wave excitation signalers and Lamb wave signal second sensors;

[0034] Specifically, the data acquisition and processing device includes:

[0035] A signal excitation module, used to excite and generate an original input signal, wherein the original input signal is a Lamb wave;

[0036] The signal receiving and analyzing module comprises a first optimization unit, a second optimization unit and a time reversal unit; the first optimization unit is used to generate a first response signal according to the received original input signal, and to eliminate the dispersion term of the first response signal to generate a first optimized signal; the time reversal unit is used to virtually time-reverse the first optimized signal, and calculate the transfer function and the time reversal transfer function;

[0037] The second optimization unit is used for simulating the process of inversion transmission, simulating the inversion signal to be transmitted from the second sensor to the first sensor to generate a second response signal, and performing dispersion elimination on the second response signal to generate a second optimization signal;

[0038] The damage information acquisition module is used to compare the similarity between the reconstructed signal and the original signal, analyze and calculate the damage index, and provide damage data;

[0039] The imaging analysis module reconstructs the area where crack damage may exist in the test structure; it uses the probabilistic imaging principle to reconstruct the crack damage image and evaluate the length of the crack.

[0040] Specifically, the piezoelectric sensor array is a square array or a circular array including a plurality of piezoelectric sensors; each of the piezoelectric sensors is embedded with a microcontroller, and the piezoelectric sensors serve as excitation signal generators and receivers of the original input signal of the Lamb wave.

[0041] The third aspect of the present application is to provide a non-transitory computer-readable storage medium, on which is stored a program for a damage detection method for a reinforced carbon fiber composite material based on virtual time reversal, and when the program is executed by a processor, the following steps are implemented:

[0042] The first sensor transmits a raw input signal to the second sensor;

[0043] The second sensor generates a virtual inversion signal by performing virtual time inversion on the original input signal, and simulates sending the inversion signal to the first sensor;

[0044] The first sensor converts the inverted signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate a damage index;

[0045] Injury localization and injury imaging are completed according to the injury index.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] (1) The present application provides high damage detection accuracy for slightly visible indentation damage of reinforced carbon fiber honeycomb sandwich structures of different sizes.

[0048] (2) The present application obtains the transfer function through frequency domain transformation and constructs a virtual time reversal process to achieve baseline-free damage detection, which greatly reduces the impact of scattering / boundary reflection waves on the physical time reversal process and improves the time reversibility and spatial reciprocity of ultrasonic guided waves (Lamb).

[0049] (3) Based on the improved virtual time reversal method, the present application obtains the highest detection accuracy near the sweet spot frequency (200kHz) where Lamb waves have the highest detectability and excitation.

[0050] (4) The present application significantly reduces the physical inversion and excitation signal measurement time / process required in the traditional physical time reversal method, saves time, and does not require physical conversion of the excitation sequence of the actuator and the receiver, thereby improving processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 It is a flow chart of a damage detection method for honeycomb composite materials based on virtual time reversal according to an embodiment of the present application.

[0053] Figure 2 It is a schematic diagram of the signal processing process of a damage detection method for honeycomb composite materials based on virtual time reversal in an embodiment of the present application.

[0054] Figure 3 Schematic diagram of a damage detection system 10 for a reinforced carbon fiber composite material based on virtual time reversal according to an embodiment of the present application.

[0055] Figure 4 Schematic diagram of the experimental device of the embodiment of the present application.

[0056] Figure 5 Schematic diagram of the PZT sensor array and indentation position of an embodiment of the present application.

[0057] Figure 6 It is a similarity comparison diagram between the original excitation signal and the reconstructed signal in the embodiment of the present application.

[0058] Figure 7 It is a comparison diagram of the similarity between the original excitation signal and the reconstructed signal.

[0059] Figure 8It is a 5 mm indentation damage image reconstructed by IVTRM in the embodiment of the present application and a 5 mm indentation damage image reconstructed by TRM in the comparative example.

[0060] Fig. 9 It is a 3 mm indentation damage image reconstructed by IVTRM in the embodiment of the present application and a 3 mm indentation damage image reconstructed by TRM in the comparative example. DETAILED DESCRIPTION

[0061] The following will be combined with the implementation methods of this application to clearly and completely describe the technical solutions in the implementation methods of this application. Obviously, the described implementation methods are only part of the implementation methods of this application, not all of the implementation methods. Based on the implementation methods in this application, all other implementation methods obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0062] It should be understood that the terms used in this specification of the embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the embodiments of the present application. As used in the specification of the embodiments of the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms of "a", "an" and "the" are intended to include plural forms.

[0063] Lamb waves are a special form of stress waves formed by the coupling of longitudinal waves and transverse waves in an acoustic waveguide (such as a thin plate) whose thickness is the same order of magnitude as the wavelength of the excitation sound wave. Since it is a guided wave in a plate, it is also commonly called a plate wave. Piezoelectric ceramic materials (PZT) can convert the voltage excited by the signal generator into mechanical strain, and form Lamb waves on the reinforced carbon fiber honeycomb sandwich composite plate through the piezoelectric effect. However, unlike body waves, due to the limitations of boundary conditions, the propagation of Lamb waves in thin plates exhibits dispersion and multi-mode characteristics.

[0064] Due to the unique dispersion characteristics of Lamb waves, the pulse waves excited by them will be deformed during propagation due to the change of group velocity, and the arrival time and amplitude of the waveform are difficult to measure, which limits the application of Lamb waves in structural damage identification.

[0065] Time reversal means that after the sensor receives the time domain signal emitted by the sound source, it reverses the time of the signal and then transmits it out, that is, the first to arrive is the last to be transmitted, and the last to arrive is the first to be transmitted. Time reversal is one of the applications of the principle of acoustic reciprocity, which can focus energy in space and time. Through this focusing, the sound source (input) signal can be reconstructed. When damage occurs in the structure, the shape of the reconstructed signal deviates from the shape of the original input signal. The damage can be identified by examining the deviation between the reconstructed signal and the known initial input signal, without the need to compare with the reference signal.

[0066] The improved virtual time-reversal method (IVTRM) of the present application uses PZT sensors as brakes and receivers of original input signals, configures PZT sensor arrays as sensor network devices, and configures actuator units and receiving units in each PZT sensor for transmitting and receiving Lamb waves. The Lamb wave signals are then processed and analyzed in combination with a data acquisition and processing device. After two optimizations of eliminating reflection-scattering wave dispersion, a virtual time reversal process is constructed in combination with time reversal theory. Finally, high-precision and high-efficiency Lamb wave signal processing is achieved without physical inversion. The reconstructed signal is obtained and compared with the original input signal to obtain a damage function, and then the damage imaging is reconstructed. The present application adopts baseline-free detection and a virtual time reversal operation based on frequency domain transformation. The transfer function of the forward transmission path is calculated by converting the input and output signals into the frequency domain through discrete Fourier transform, and a virtual time reversal process is constructed to achieve baseline-free and high-precision damage detection. Efficient and high-precision damage detection is achieved, avoiding the physical reversal process and errors, and solving the technical problems of the time reversal method (TRM) in the prior art, such as limited signal reconstruction quality and damage imaging accuracy, and inability to accurately locate the damage position and damage condition. High-precision damage imaging can be formed, and the damage position and damage condition can be accurately located.

[0067] Specifically, the first sensor and the second sensor in the present application are both PZT sensors. Specifically, in a PZT sensor array composed of multiple PZT sensors, each PZT sensor can be used as an actuator to send an original input signal to other PZT sensors, and can receive an original input signal from another PZT sensor. The virtual time reversal operation processing process based on frequency domain transformation of the present application is described by two PZT sensors and the corresponding data processing process, and the two PZT sensors are respectively the first sensor PZT1 and the second sensor PZT2.

[0068] Specifically, the first aspect of the present application is to provide a damage detection method for a reinforced carbon fiber composite material based on virtual time reversal, such as Figure 1 FIG. 1 is a flowchart of a damage detection method for a reinforced carbon fiber composite material based on virtual time reversal of the present application, which specifically includes the following steps:

[0069] S1, the first sensor transmits the original input signal to the second sensor;

[0070] S2, the second sensor generates an inversion signal by performing virtual time inversion on the original input signal, and sends the inversion signal to the first sensor;

[0071] S3, the first sensor converts the inversion signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate a damage index;

[0072] S4. Obtain damage imaging according to the damage index.

[0073] Specifically, step S1 is a physical process, a real signal transmission and reception process. The time reversal after PZT2 receives the signal is to construct a virtual reversal process.

[0074] Specifically, due to the material properties of the PZT sensor, the excited voltage can be converted into mechanical strain acting on the reinforced carbon fiber honeycomb sandwich composite material. The first sensor and the second sensor in the present application both use PZT sensors. Moreover, the PZT sensor is dual-functional, having both the functions of an actuator and a receiver. Each PZT sensor can transmit and receive Lamb waves and inversion signals. The following describes the processing process of the improved virtual time reversal of the present application through the transmission and processing of signals between the two PZT sensors PZT1 and PZT2.

[0075] First, the wave energy U of the original input signal is generated in the first sensor PZT1 A (ω):

[0076] U A (ω) = k aa (ω)V A (ω), (1)

[0077] Among them, U A (ω) represents the wave energy excited at A, V A (ω) is the excitation signal, k aa (ω) is the electromechanical effect conversion coefficient of PZT1, and ω represents a single-frequency Lamb wave.

[0078] Specifically, the power of the excitation signal is set, and the original input signal can be calculated based on the excitation signal and the electro-mechanical effect conversion coefficient of the PZT sensor. The electro-mechanical effect conversion coefficient is known and is related to the performance of the PZT sensor. In terms of signal generation and data acquisition, a sinusoidal sonic boom (peak value of 60V) enclosed in the Hanning window is used as the input signal of the first sensor. The activation and acquisition of the waveform signal are performed by a PZT piezoelectric sensor. Preferably, the power of the excitation signal is set to 200-250kHz, where 200kHz is the sweet spot frequency.

[0079] In step S1, the first sensor PZT1 generates the wave energy U of the original input signal A (ω) is sent to the second sensor PZT2. This process is a physical transmission process.

[0080] Specifically, step S2 includes the following steps:

[0081] S21, the second sensor receives the original input signal and generates a first response signal;

[0082] S22, constructing a virtual transfer function through frequency domain conversion according to the original input signal and the first response signal;

[0083] S23, generating a virtual inversion signal through the virtual time inversion processing based on the virtual transfer function;

[0084] S24, simulating a process of sending the virtual inversion signal along the inversion path to the first sensor.

[0085] Specifically, the second sensor in step S21 refers to the second sensor PZT2. The second sensor PZT2 generates a single-frequency Lamb wave of a first response signal according to the received original input signal, which can be expressed as:

[0086] U B (ω,r)=k rb A(ω,r)e -i(k·r-ωt) , (2)

[0087] Among them, A is the amplitude of the single-frequency Lamb wave, A(ω,r) is the attenuation of the Lamb wave amplitude, r is the propagation path of the Lamb wave, k is the wavelength of the single-frequency Lamb wave, and k rb is the mechanical-electrical energy conversion coefficient of PZT2, e -i(k·r-ωt) Represents the dispersion term of Lamb wave (which decays with propagation distance), i represents an imaginary number, and t represents the propagation time of the wave.

[0088] When a single-frequency Lamb wave propagates in a composite material plate, the amplitude of the Lamb wave is attenuated to a certain extent due to the material damping and the geometric diffusion of the composite material plate itself. This phenomenon can be expressed as:

[0089]

[0090] Where T(ω,r) is the linear characteristic of the Lamb wave mode, U A (ω) represents the frequency domain representation of the original input signal excited by PZT1, and r is the modulus of r. It represents the geometric attenuation of the Lamb wave amplitude, which is caused by the geometric diffusion phenomenon of the Lamb wave, that is, the farther the Lamb wave propagates, the greater the amplitude attenuation.-η(θ,ω)r It is the wave packet energy dissipation caused by the nonlinear characteristics of the CFEP honeycomb sandwich composite material. In addition, the attenuation coefficient η depends on the propagation direction θ of the Lamb wave and the excitation frequency ω.

[0091] In general, the PZT type of the actuating-receiving pair is the same, the actuating corresponds to the actuator, and the receiving corresponds to the receiver, that is, the parameters and performance of the first sensor PZT1 and the second sensor PZT2 in this application are the same, that is:

[0092]

[0093] Among them, k aa Indicates the electrical-to-mechanical energy conversion coefficient of PZTa, k ab Represents the electrical-to-mechanical energy conversion coefficient of PZTb, represents the mechanical-electrical energy conversion coefficient of PZTa, It represents the mechanical-electrical energy conversion coefficient of PZTa. In this application, PZTa corresponds to PZT1, and PZTb corresponds to PZT2.

[0094] Specifically, in step S22, the system transfer function H(ω,r) of the Lamb wave propagating in the CFEP honeycomb sandwich composite plate is derived by combining formulas (1) to (4):

[0095]

[0096] Among them, U B (ω,r) is u B The discrete Fourier frequency domain representation of (t,r), e -i(k·r-ωt) Represents the dispersion term of Lamb wave (which decays with propagation distance). Due to the existence of the dispersion term, the final reconstruction process will be affected. Therefore, it is necessary to eliminate the dispersion term to reduce its influence.

[0097] Specifically, step S23 includes the following steps:

[0098] S231, performing a first dispersion term elimination on the first response signal to generate a first optimized signal;

[0099] S232. Perform discrete Fourier transform on the first optimized signal, and invert it in the time domain to generate the virtual inversion signal.

[0100] Specifically, in step S231, the calculation method for eliminating the dispersion term, that is, eliminating the boundary reflection wave / scattered wave signal includes formulas (6) to (7):

[0101]

[0102] Among them, U(A, t) and U(B, t) are the original input signal and the time domain received signal of PZT2 respectively, V peak-Direct and V peak-A is the energy peak of the direct path and the original input Lamb wave signal, such as Figure 2 As shown, the direct path refers to a unidirectional path from PZT1 to PZT2;

[0103] D t =[n peak -n left-Direct ,n peak -n right-Direct ], (7)

[0104] Among them, n peak Represents the location of the direct wave energy peak point, n left-Direct and n right-Direct It is the left and right boundaries of the truncation window, which is determined by the ratio of the peak point to the left and right boundaries.

[0105] The above process of eliminating the dispersion term is achieved by considering the peak point ratio of the Lamb wave itself and its linear similarity with the original input signal.

[0106] Specifically, step S22 further includes taking a conjugate of the transfer function to obtain a time reversal transfer function:

[0107] H TR (ω,r)=H * (ω,r)H(ω,r), (8)

[0108] Among them, H TR (ω,r) is the time reversal transfer function, H * (ω,r) is the conjugate of the transfer function H(ω,r). The introduction of the time reversal transfer function replaces the operation of reversing the forward transmission path and performing dispersion compensation (reversing the signal) in the physical reversal process in the prior art. Compared with the prior art, the construction of the virtual transfer function and the time reversal transfer function avoids the error in the physical reversal process, makes the data processing process more accurate, and greatly improves the processing efficiency, which helps to reconstruct the accuracy and efficiency of the signal.

[0109] Specifically, step S3 includes the following steps:

[0110] S31, simulating a process in which the inversion signal is transmitted from the second sensor to the first sensor to obtain a virtual second response signal;

[0111] S32. Obtain a reconstruction signal based on the second response signal;

[0112] S33. Calculate the similarity between the reconstructed signal and the original input signal to calculate a damage index.

[0113] Specifically, the process of step S3 is a process in which PZT2 simulates transmitting an inverted signal (virtual reconstructed Lamb wave signal) to PZT1, and obtains a reconstructed signal after optimization processing and time domain inversion.

[0114] Specifically, step S31 simulates the process of transmitting the inversion signal from the second sensor to the first sensor to obtain a virtual second response signal, including converting the inversion signal into the frequency domain by Fourier transform, multiplying it with the time reversal transfer function, and then converting the multiplication result into the time domain by inverse Fourier transform, thereby obtaining the second response signal.

[0115] Specifically, the process of emitting the simulated inversion signal from PZT2 to PZT1 is that the virtual inversion signal is Fourier transformed into the frequency domain, then multiplied by the time reversal transfer function, and then Fourier inverse transformed into the time domain to obtain the second response signal U R (ω), the second response signal is related to the transfer function and the inverse signal, so the expression is as follows:

[0116] U R (ω)=U B * (ω,r)H(ω,r), (9)

[0117] in, Indicates U B The complex conjugate of (ω,r) in the frequency domain. Substituting equation (5) into equation (9), we can get equation (10):

[0118]

[0119] Among them, e -η(θ,ω)r It is the energy dissipation of the wave packet caused by the nonlinear characteristics of the CFEP honeycomb sandwich composite material, representing the dispersion term of the Lamb wave (which decays with the propagation distance). Due to the existence of the dispersion term, it is also necessary to eliminate the dispersion term to reduce the influence of the dispersion term on the reconstructed signal.

[0120] Specifically, in step S32, obtaining a reconstructed signal based on the second response signal includes the following steps:

[0121] S321, performing a second dispersion term elimination on the second response signal to generate a second optimized signal;

[0122] S322: Perform inverse discrete Fourier transform on the second optimized signal to perform time domain inversion to generate the reconstructed signal.

[0123] Specifically, the calculation method for eliminating the second dispersion term in step S321 is consistent with the principle of eliminating the first dispersion term, and is also implemented by considering the peak point ratio of the Lamb wave itself and its linear similarity with the original input signal. The specific calculation method is:

[0124]

[0125] Among them, U(A, t) and U(B, t) are the original input signal and the time domain received signal of PZT2 respectively, V peak-Direct and V peak-A is the energy peak of the direct path and the original input Lamb wave signal, such as Figure 2 As shown in the figure, the direct path refers to the unidirectional path from PZT1 to PZT2; the calculation formula of the direct path is as follows:

[0126] D t =[n peak -n left-Direct ,n peak -n right-Direct ],(7)

[0127] Among them, n peak Represents the location of the direct wave energy peak point, n left-Direct and n right-Direct It is the left and right boundaries of the truncation window, which is determined by the ratio of the peak point to the left and right boundaries.

[0128] Since Lamb waves can be divided into symmetric and asymmetric modes (i.e., S mode and A mode), at low frequencies, they mainly exist in the form of S0 and A0 modes. Therefore, the linear characteristic T(ω,r) of Lamb waves can be expressed as:

[0129]

[0130] Among them, S0 and A0 are the amplitudes of the S0 mode and A0 mode of the Lamb wave, respectively. and are the corresponding wave numbers of Lamb waves.

[0131] Substituting formula (11) into formula (10), the second response signal U R (ω) can be expressed as:

[0132]

[0133] in, The energy of the main part of the Lamb wave signal representing the second response signal. Formulas (12) and (13) show that the reconstructed Lamb wave is time-reversed. At PZT1, r = 0, so the dispersion term e -i(k·r-ωt)The influence on the main wave packet is completely eliminated and the signal is refocused at PZT1.

[0134] Traditional physical inversion only considers the complete separation of Lamb waves in S0 and A0 modes, and completely reconstructs the Lamb wave signal. The superposition of reflected waves or waves in other modes with the main wave packet will seriously interfere with the physical inversion process and destroy the time reversibility of Lamb waves. This application proposes an improved virtual time reversal method by constructing a virtual transfer function and a time reversal transfer function, considering the impact of the damage path on the reversal process, and implementing the main wave packet interception based on the S0 mode. The virtual reversal signal generated at PZT2 can be expressed as:

[0135] U B (ω,r)=U 3-Direct (ω,r1)+U B-Boudary (ω,r2), (14)

[0136] And the virtual second response signal obtained at PZT1 according to the virtual inversion signal can be expressed as:

[0137]

[0138] Where r1 and r2 represent the direct path and boundary reflection path of the Lamb wave, respectively.

[0139] Since the Lamb wave of the S mode is faster than the A mode, the wave of the S0 mode in the low-frequency mode arrives first, and the remaining wave dispersion terms are eliminated through the process of two dispersion terms elimination. The final frequency domain reconstructed signal The result of reconstructing the Lamb wave signal based on eliminating redundant wave packets can be expressed as:

[0140]

[0141] Finally, based on the inverse discrete Fourier transform, the reconstructed signal in the time domain is obtained. It is expressed as:

[0142]

[0143] The reconstructed Lamb wave (ie, the reconstructed signal) contains only the S0 mode, thereby improving the signal quality and eliminating the dispersion term of the Lamb wave.

[0144] Specifically, in step S33, the similarity between the reconstructed signal and the original input signal is calculated to calculate the damage index.

[0145] Specifically, the damage index (DamageIndex, DI) is used to calculate the similarity between the reconstructed signal and the original input signal to evaluate the damage degree of the reinforced carbon fiber honeycomb sandwich composite material, which can be expressed as follows:

[0146]

[0147] Among them, DI is used to evaluate the similarity between the original input signal and the reconstructed signal, and n is the number of data recording points. n and Y n are the original input signal and the reconstructed signal, respectively, ρ represents the Pearson similarity (the difference between the two signals, the smaller the difference, the smaller the value), and μ is the average value of the signal amplitude. On the boundary path and the non-boundary path, the smaller the DI, the higher the similarity, the better the signal refocusing effect, and the lower the impact of the damage on the current path. On the damaged path, the larger the DI, the lower the similarity, and the better the detection effect of the damage position.

[0148] Preferably, in order to more clearly distinguish the DI difference between the damaged path and the healthy sensing path, in another embodiment, the new DI calculation method is as follows:

[0149] DI accelerate =1-ρ 2 , (19).

[0150] Specifically, step S4, obtaining damage imaging according to the damage index, and calculating the damage accuracy and size.

[0151] Specifically, a probabilistic damage reconstruction algorithm including weighted DIs realizes damage imaging. The probability distribution is largest in the middle area of ​​the intuition path and decreases as the distance between the damage and the intuition path increases. In this case, DI represents the possibility of damage in the sensing path. By linearly superimposing the spatial probability distribution of all sensing paths in sequence, the damage perception of the area surrounded by the first sensor-second sensor network can be enhanced, and probabilistic damage imaging can be realized. In this application, the first sensor-second sensor network realizes the sensing path network by arranging and combining the same PZT sensors in a certain way.

[0152] Specifically, the sensing path network may be a square PZT sensor array composed of a number of PZT sensors, or may be rectangular, circular, honeycomb, or the like.

[0153] The second aspect of the present application is to provide a damage detection system for a reinforced carbon fiber composite material based on virtual time reversal, such as Figure 3 , which is a schematic diagram of a damage detection system 10 for reinforced carbon fiber composite materials based on virtual time reversal of the present application, including a sensor network device 20 and a data acquisition and processing device 30;

[0154] The sensor network device 20 includes a piezoelectric sensor array, which includes a plurality of piezoelectric sensors, each of which has a microcontroller embedded therein, and the piezoelectric sensors serve as Lamb wave excitation signal generators and Lamb wave signal second sensors. Each of the piezoelectric sensors serves as a first sensor to generate the original input signal of the Lamb wave, and as a second sensor to receive the inversion signal of the aforementioned Lamb wave sent by other piezoelectric sensors. Figure 4 , which is a schematic diagram of a PZT sensor array according to an embodiment of the present application, including a square composed of 16 PZT sensors, and black dots represent the positions of the PZT sensors.

[0155] Preferably, the piezoelectric sensor array may be in a square, rectangular, circular or other shape.

[0156] The data acquisition and processing device 30 comprises:

[0157] A signal excitation module 31 is used to excite and generate an original input signal, wherein the original input signal is a Lamb wave;

[0158] The signal receiving and analyzing module 32 includes a first optimization unit 321, a second optimization unit 323 and a time reversal unit 322; the first optimization unit 321 is used to generate a first response signal according to the received original input signal, and to eliminate the dispersion term of the first response signal to generate a first optimized signal; the time reversal unit 322 is used to time-reverse the first optimized signal, calculate the transfer function and the time reversal transfer function; the second optimization unit 323 is used to generate a second response signal according to the received inversion signal, and to eliminate the dispersion term of the second response signal to generate a second optimized signal;

[0159] A damage information acquisition module 33 is used to compare the similarity between the reconstructed signal and the original signal, analyze and evaluate the damage index, and provide damage data;

[0160] The imaging analysis module 34 reconstructs the area where crack damage may exist in the detection structure; reconstructs the crack damage image using the probability imaging principle and evaluates the length of the crack.

[0161] The third aspect of the present application is to provide a non-transitory computer-readable storage medium, on which is stored a program for a damage detection method for a reinforced carbon fiber composite material based on virtual time reversal, and when the program is executed by a processor, the following steps are implemented:

[0162] The first sensor transmits a raw input signal to the second sensor;

[0163] The second sensor generates an inversion signal by performing virtual time inversion on the original input signal, and sends the inversion signal to the first sensor;

[0164] The first sensor converts the inverted signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate a damage index;

[0165] Injury imaging is obtained according to the damage index.

[0166] The following is a further description of the damage detection method for reinforced carbon fiber composite materials based on virtual time reversal of the present application through specific embodiments, showing how to achieve high-quality signal reconstruction through an improved time reversal method.

[0167] like Figure 4 , which is a schematic diagram of an experimental device of an embodiment of the present application. Figure 4 (a) is a reinforced carbon fiber honeycomb sandwich panel subjected to indentation. Figure 4 (b) is the surface contour modeling of the sandwich panel after indentation.

[0168] The material properties of the reinforced carbon fiber honeycomb sandwich panel made of plain fabric are shown in Table 1:

[0169] Table 1. Material properties of plain fabric reinforced carbon fiber honeycomb sandwich panels

[0170]

[0171] A sensor array consisting of 16 circular PZTs was created on a reinforced carbon fiber honeycomb sandwich panel, with the PZT elements spaced 100 mm apart to form a square area. Figure 5 As shown, Figure 5 The diagram of the PZT sensor array and indentation position of the embodiment of the present application, each element (P1-P16) is 10mm in diameter and 1mm thick, and is surface mounted on a reinforced carbon fiber honeycomb sandwich plate using Loctite super glue. One of the PZT elements (P1) acts as the first sensor for the start wave signal, while the remaining elements (P2-P16) act as sensors for capturing these signals. The role of the first sensor is always alternating until all PZT elements act as the first sensor.

[0172] In order to ensure accurate placement of the PZT elements and minimize possible errors between different PZT elements due to positional offset or electromechanical coefficient variations, a template was made and used when mounting the PZT elements. In addition, a consistent amount of adhesive was used when bonding each PZT element, and the curing time and temperature were strictly controlled.

[0173] In general, different frequency components decay at different rates, and the physical time reversal process cannot fully compensate. Therefore, by exciting with narrowband pulses, the amplitude dispersion can be minimized, and the original excitation signal can be reconstructed to a greater extent through the TR process. In the embodiment of the present application, in terms of signal generation and data acquisition, a sinusoidal sonic boom (with a peak value of 60V) enclosed in a Hanning window is used as the input signal of the first sensor. The activation and acquisition of the waveform signal is performed by an actuator.

[0174] The specific process is: the reinforced carbon fiber honeycomb composite sandwich panel is further indented until the indentation reaches 3mm (Example 1) and 5mm (Example 2) to produce slightly visible indentations, and the designed PZT sensor network array is used for damage detection. Then, the application of Lamb waves to detect slightly visible indentations in reinforced carbon fiber honeycomb sandwich panels is studied in depth.

[0175] like Figure 4 The red area in the figure is the area where the indentation in Example 1 and Example 2 is located. The detection area includes an area of ​​400×400mm 2 The 16 PZTs are firmly glued to the edges of this square area with uniform spacing. Probabilistic damage imaging technology involves dividing the damage detection area into 400×400 blocks, forming 160,000 individual pixels. Each pixel corresponds to 1×1mm 2 square area.

[0176] In order to evaluate the sensitivity of the IVTRM of the present application and the TRM of the prior art and to verify the reliability of the present application, three specific paths were randomly selected in the above-mentioned reinforced carbon fiber honeycomb sandwich panel: paths 2-6 representing boundary paths, paths 3-8 representing non-boundary paths, and damage paths 9-15 passing through the damage location.

[0177] like Figure 6 This is a similarity comparison diagram of the original excitation signal and the reconstructed signal of the boundary path 2-6, non-boundary path 3-8, and damaged path 9-15 using IVTRM in this application. The indentation is 5mm and the power of the excitation signal is 200kHz. 2-6 =0.1187,DI 3-8 =0.0877,DI 9-15 =0.3058.

[0178] Conventional TRM is used to detect boundary paths 2-6, non-boundary paths 3-8, and damaged paths 9-15, such as Figure 7 It is a similarity comparison diagram between the original excitation signals and the reconstructed signals of the boundary paths 2-6, the non-boundary paths 3-8, and the damaged paths 9-15 measured by conventional physical TRM in the comparative example.

[0179] like Figure 7 It is a reinforced carbon fiber honeycomb composite sandwich panel with the same material as above. In the same boundary path 2-6, non-boundary path 3-8, and damage path 9-15, the indentation is 5mm and the power of the excitation signal is 200kHz. The signal reconstruction results based on the main wave packet S0 mode Lamb wave using conventional TRM, the similarity comparison diagram of the original excitation signal and the reconstructed signal of different paths, DI 2-6 =0.2525,DI 3-8 =0.1323,DI 9-15 =0.1239.

[0180] Figure 6 The damage index DI values ​​are 0.1187 (corresponding to path 2-6) and 0.0877 (corresponding to path 3-8), which are significantly smaller than those under the same path. Figure 7 The values ​​of the damage index DI in the comparative example are 0.2525 (corresponding to path 2-6) and 0.1323 (corresponding to path 3-8). The reason why the DI of path 2-6 is greater than that of path 3-8 is because path 2-6 is closer to the boundary than path 3-8, and therefore path 2-6 is more affected by boundary reflection waves and additional modes. Since on boundary paths and non-boundary paths, the smaller the D1 value, the higher the similarity between the reconstructed signal and the original input signal, the signal reconstruction and focusing effect of the embodiment is better than that of the comparative example. On the damaged path, the larger the DI, the lower the similarity, and the better the detection effect of the damaged position. Figure 6 (c) DI 9-15 =0.3058, much higher than Figure 7 (c) DI 9-15 =0.1239, indicating that the detection result of the damage position of the present application is better than that of the control example. Specifically, in the processing of the present application, a virtual transfer function is constructed, and the dispersion term is eliminated twice, which reduces the influence of the scattered wave and ensures that the time reversibility of the Lamb wave is not destroyed. The larger DI changes in the boundary path and the non-boundary path in the control example are attributed to the obvious impact of the damage on the transfer function extraction process. This impact destroys the time reversibility of the Lamb wave, and further stems from the mixing of scattered waves from the damage in the main wave packet, thereby affecting the virtual time reversal process. In contrast, the reconstruction quality of the IVTRM of the present application is higher, and it can correctly distinguish between damaged paths and healthy paths. This also means that the IVTRM of the present application has better consistency and reduces the risk of false alarms of damage.

[0181] In addition, the IVTRM of the present application can effectively reconstruct the signal with higher fidelity to the original input signal, whether the sensing path is close to the boundary or far from the boundary. This highlights that the IVTRM of the present application has strong anti-interference ability. The IVTRM of the present application performs relatively well in time reversal processing in the presence of damage, and the difference between the reconstructed signal of path 2-6 and the original signal is much smaller. Compared with Figure 7 The TRM results, Figure 6 It can be seen from the reconstruction results of the IVTRM of the present application for paths 2-6 and 3-8 that the IVTRM of the present application corrects the reconstruction anomaly caused by the boundary scattering wave, and the obtained reconstructed signal is closer to the original excitation signal (the DI value of IVTRM is lower than that of TRM, and the influence of its scattered waves on the main wave packet is significantly reduced, and the main wave packet is more completely reconstructed).

[0182] like Figure 8 The 5 mm indentation damage image obtained by IVTRM based on the probabilistic damage reconstruction method in the embodiment of the present application and the 5 mm indentation damage image reconstructed by TRM in the comparative example are obtained by superposition of all sensing paths. Figure 8 Among them, (a) the setting condition is: conventional TRM, 200kHz; (b) the setting condition is: conventional TRM, 250kHz; (c) the setting condition is: IVTRM of this application, 200kHz; (d) the setting condition is: IVTRM of this application, 250kHz.

[0183] like Fig. 9 The 3mm indentation damage image obtained by the IVTRM method based on the probabilistic damage reconstruction method in the embodiment of the present application and the 3mm indentation damage image reconstructed by the TRM method in the comparative example are obtained by superposition of all sensing paths. Fig. 9 Among them, (a) the setting conditions are: conventional TRM method, 200kHz; (b) the setting conditions are: conventional TRM, 250kHz; (c) the setting conditions are: IVTRM of the present application, 200kHz; (d) the setting conditions are: IVTRM of the present application, 250kHz.

[0184] As can be seen from the figure, compared with the indentation damage imaging image obtained by the conventional TRM, the IVTRM of the present application can effectively improve the quality of time reversal reconstruction, enhance the time reversibility and spatial linearity of Lamb waves, significantly reduce the influence of boundary reflection waves, scattered waves and multimodality on time reversal reconstruction and damage imaging, and greatly improve the damage detection accuracy and imaging quality. In addition, since the IVTRM does not need to physically perform the time reversal process, but obtains the transfer function through frequency domain transformation and constructs a virtual time reversal process to achieve it. Therefore, compared with the conventional TRM, the detection time and operation process of the IVTRM of the present application can be greatly reduced. Therefore, IVTRM is a reliable, fast, accurate, baseline-free Lamb wave damage detection method. Compared with the prior art, the present application reduces the dependence on baseline signals, the dispersion characteristics of Lamb waves, low-quality physical inversion / reconstruction signals, and limited damage detection accuracy, and introduces an improved virtual time reversal method for baseline-free damage detection of reinforced carbon fiber composite sandwich structures.

[0185] The present application provides a damage detection method for a reinforced carbon fiber composite material based on virtual time reversal, comprising the following steps: a first sensor transmits an original input signal to a second sensor; the second sensor generates an inversion signal by virtual time reversing the original input signal, and simulates sending the inversion signal to the first sensor; the first sensor converts the inversion signal into a reconstructed signal, calculates the similarity between the reconstructed signal and the original input signal, so as to calculate a damage index; and completes damage location and damage imaging according to the damage index. The present application proposes a virtual time reversal operation based on frequency domain transformation. The transfer function of the forward transmission path is calculated by converting the input and output signals into the frequency domain through discrete Fourier transform, and a virtual time reversal process is constructed to realize baseline-free damage detection, thereby realizing efficient and high-precision damage detection. The influence of scattering / boundary reflection waves on the physical time reversal process is greatly reduced, and the time reversibility and spatial reciprocity of ultrasonic guided waves (Lamb) are improved. The highest detection accuracy result is obtained near the sweet spot frequency (200kHz) where the Lamb wave has the highest detectability and excitation. The measurement time / process of the physical inversion and excitation signal required in the traditional physical reversal process saves time, and there is no need to physically convert the excitation sequence of the actuator and the receiver, which improves the processing efficiency.

[0186] The above description is only a preferred implementation mode of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A damage detection method for composite materials based on virtual time reversal, characterized in that: The following steps are involved: The first sensor transmits a raw input signal to the second sensor; The second sensor generates a virtual inversion signal by performing virtual time inversion on the original input signal, and simulates sending the inversion signal to the first sensor; The first sensor converts the inverted signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate a damage index; Injury localization and injury imaging are completed according to the injury index.

2. A composite material damage detection method based on virtual time reversal according to claim 1, characterized in that: The second sensor generates an inversion signal by performing virtual time inversion on the original input signal, and simulates sending the inversion signal to the first sensor, including: The second sensor receives the original input signal and generates a first response signal; Constructing a virtual transfer function through frequency domain conversion according to the original input signal and the first response signal; Based on the virtual transfer function, a virtual inversion signal is generated by the virtual time inversion process; The process of sending the virtual inversion signal along an inversion path to the first sensor is simulated.

3. A composite material damage detection method based on virtual time reversal according to claim 2, characterized in that: The method of generating a virtual inversion signal based on the virtual transfer function through the virtual time reversal processing comprises the steps of: Performing a first dispersion term elimination on the first response signal to generate a first optimized signal; The first optimized signal is subjected to discrete Fourier transform and inverted in the time domain to generate the virtual inverted signal.

4. The damage detection method for composite materials based on virtual time reversal according to claim 2, characterized in that: After calculating the transfer function according to the original input signal and the first response signal, the step further includes: The transfer function is conjugated to obtain a time-reversed transfer function.

5. The damage detection method for composite materials based on virtual time reversal according to claim 4, characterized in that: The first sensor converts the inverted signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate the damage index, including the steps of: Simulating a process in which the inversion signal is transmitted from the second sensor to the first sensor to obtain a virtual second response signal; obtaining a reconstructed signal based on the second response signal; The similarity between the reconstructed signal and the original input signal is calculated to calculate a damage index.

6. A composite material damage detection method based on virtual time reversal according to claim 5, characterized in that: The process of simulating the inversion signal being transmitted from the second sensor to the first sensor to obtain a virtual second response signal comprises the steps of: The inversion signal is converted into the frequency domain by Fourier transformation, then multiplied by the time reversal transfer function, and then the result of the multiplication is converted into the time domain by inverse Fourier transformation, so as to obtain the second response signal.

7. The damage detection method for composite materials based on virtual time reversal according to claim 5, characterized in that: The step of obtaining a reconstruction signal based on the second response signal comprises the steps of: performing a second dispersion term elimination on the second response signal to generate a second optimized signal; Performing inverse discrete Fourier transform on the second optimized signal to perform time domain inversion to generate the reconstructed signal.

8. A composite material damage detection system based on virtual time reversal, characterized in that: It includes a sensor network device and a data acquisition and processing device; The sensor network device comprises a piezoelectric sensor array, the piezoelectric sensor array comprises a plurality of piezoelectric sensors, each of the piezoelectric sensors is embedded with a microcontroller, and the piezoelectric sensors serve as Lamb wave excitation signalers and Lamb wave signal second sensors; The data acquisition and processing device comprises: A signal excitation module, used to excite and generate an original input signal, wherein the original input signal is a Lamb wave; A signal receiving and analyzing module, comprising a first optimization unit, a second optimization unit and a time reversal unit; The first optimization unit is used to generate a first response signal according to the received original input signal, and to eliminate the dispersion term of the first response signal to generate a first optimized signal; the time reversal unit is used to virtually time-reverse the first optimized signal, and calculate the transfer function and the time reversal transfer function; The second optimization unit is used for simulating the process of inversion transmission, simulating the inversion signal to be transmitted from the second sensor to the first sensor to generate a second response signal, and performing dispersion elimination on the second response signal to generate a second optimization signal; The damage information acquisition module is used to compare the similarity between the reconstructed signal and the original signal, analyze and calculate the damage index, and provide damage data; The imaging analysis module reconstructs the area where crack damage may exist in the test structure; it uses the probabilistic imaging principle to reconstruct the crack damage image and evaluate the length of the crack.

9. A composite material damage detection system based on virtual time reversal according to claim 8, characterized in that: The piezoelectric sensor array is a square array or a circular array including a plurality of piezoelectric sensors; each of the piezoelectric sensors is embedded with a microcontroller, and the piezoelectric sensors serve as excitation signal devices and receivers of the original input signal of the Lamb wave.

10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores a program of a composite material damage detection method based on virtual time reversal, and when the program is executed by a processor, the following steps are implemented: The first sensor transmits a raw input signal to the second sensor; The second sensor generates a virtual inversion signal by performing virtual time inversion on the original input signal, and simulates sending the inversion signal to the first sensor; The first sensor converts the inverted signal into a reconstructed signal, and calculates the similarity between the reconstructed signal and the original input signal to calculate a damage index; Injury localization and injury imaging are completed according to the injury index.