Elliptic probability fusion-based damage location method for elastic body layer carbon fiber composite material

Abstract

This invention discloses a damage localization method for elastomeric layered carbon fiber composite materials based on elliptic probability fusion, comprising: acquiring Lamb wave signals from multiple sensing paths using different sensors as excitation sources; performing continuous wavelet transform using complex Morlet wavelets to extract the measured flight time; introducing a layered wave velocity correction model to optimize the elliptic trajectory method and calculating the theoretical flight time and damage probability; introducing a dynamic minor axis optimization model to optimize the probability weighting method and calculating the damage probability; generating a prior distribution using the results calculated by the elliptic trajectory method and the probability weighting method; constructing a likelihood function based on the difference between the measured flight time and the theoretical flight time; obtaining the posterior distribution of the damage location by combining the prior distribution and the likelihood function using Bayes' theorem; sampling the posterior distribution using an adaptive MCMC algorithm and generating a probability cloud map of the damage location; and outputting the final imaging result of the damage location for damage localization.

Description

Technical Field

[0001] This invention belongs to the field of composite material damage self-sensing technology, and specifically relates to a damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion. Background Technology

[0002] Carbon fiber composites with elastomeric layers have been widely used in a wide range of fields, including aerospace, new energy vehicles, and high-end equipment maintenance, due to their lightweight and high-strength properties. For example, in the new energy vehicle sector, they are used for battery pack casings and lightweight body components (such as doors and chassis reinforcements). In the aerospace sector, they are used in structural components such as aircraft skin and spacecraft shells. In the high-end equipment maintenance sector, wind turbine blades (carbon fiber-elastomeric sandwich structures) and industrial robot arms also utilize this material extensively. However, these materials are prone to damage during use due to factors such as impact and fatigue. If damage is not detected and located in a timely manner, it can lead to serious consequences. For instance, in new energy vehicles, damage to the battery pack casing can cause battery safety issues and shorten battery pack life. In the aerospace sector, the accumulation of latent damage to aircraft skin and spacecraft shells can lead to structural failure and endanger flight safety. Therefore, high-precision damage localization of carbon fiber composites with elastomeric layers is crucial.

[0003] Because Lamb waves are sensitive to minute damage, Lamb wave damage localization technology based on the elliptical trajectory method is currently the mainstream technology for high-precision damage localization of carbon fiber composites containing elastomer layers. Existing technologies, such as patent "CN110376282A A Lamb Wave Damage Localization Method Based on Elliptic Probability and Bayesian Estimation," integrate the elliptical trajectory method and the probabilistic damage reconstruction method. By combining the non-linear path detection capability of the elliptical trajectory method with the linear path detection capability of the probabilistic damage reconstruction method through Bayesian estimation, high-precision damage localization is achieved. However, existing technologies have several shortcomings: the material properties of the elastomer layer and the carbon fiber layer cause the Lamb wave propagation velocity to exhibit layered variations, but existing technologies use a single group velocity assumption and do not consider the wave velocity layering changes caused by the elastomer layer. Therefore, they cannot accurately model the wave velocity distribution of multi-material composite structures, ultimately leading to large damage localization errors. Simultaneously, the abrupt change in wave velocity at the interface between the elastomer layer and the carbon fiber layer is not modeled, resulting in inaccurate elliptical trajectory generation. Furthermore, the probability weighting method relies on empirical parameters in application and is difficult to adapt to the complex signal propagation characteristics of CFRP materials containing elastomer layers. Moreover, the elastomer layer may introduce nonlinear changes in wave velocity or path interference, leading to enhanced noise interference. However, traditional methods have not optimized the Bayesian fusion framework for such complex structures, making the localization results susceptible to noise and unstable. The multi-layered heterogeneous characteristics of carbon fiber composites containing elastomer layers complicate the Lamb wave propagation path, but existing methods are mainly for homogeneous materials or simple laminates and do not consider the wave velocity correction requirements at the interface between the elastomer layer and the carbon fiber layer. Therefore, they cannot be directly applied to composite materials containing elastomer layers. Summary of the Invention

[0004] The purpose of this invention is to provide a damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion, which can significantly improve the accuracy and robustness of damage localization in multilayer heterogeneous composite materials and adapt to complex wave velocity distributions and noise interference.

[0005] The technical solution provided by this invention is as follows:

[0006] A damage localization method for elastomeric layer carbon fiber composites based on elliptic probability fusion includes:

[0007] Step 1: Install sensors on the damaged carbon fiber composite laminate containing elastomeric layers, use different sensors as excitation sources, and use the ultrasonic guided wave damage diagnosis platform system to collect Lamb wave signals from multiple sensing paths.

[0008] Step 2: Perform continuous wavelet transform on the Lamb wave signal using complex Morlet wavelets to extract the measured flight time;

[0009] Step 3: Introduce a layered wave velocity correction model to optimize the elliptical trajectory method, and use the optimized elliptical trajectory method to calculate the theoretical flight time and damage probability;

[0010] Step 4: Introduce a dynamic short-axis optimization model to optimize the probability weighting method, and use the optimized probability weighting method to calculate the damage probability;

[0011] Step 5: Using the damage probability calculated by the elliptical trajectory method and the probability weighting method, generate a prior distribution; based on the difference between the measured flight time and the theoretical flight time, construct a likelihood function; and by combining the prior distribution and the likelihood function with Bayes' theorem, obtain the posterior distribution of the damage location.

[0012] Step 6: Use the adaptive MCMC algorithm to sample the posterior distribution and generate a probability cloud map of the damage location; output the final imaging result of the damage location to locate the damage location.

[0013] Preferably, the process of introducing a layered wave velocity correction model to optimize the elliptical trajectory method and calculating the theoretical flight time and damage probability is as follows: Group velocity samples in different directions are obtained using a DC calculator; a continuous group velocity function is constructed using linear interpolation; a layered wave velocity correction model is introduced to obtain a corrected continuous group velocity function; the theoretical flight time is calculated using the corrected continuous group velocity function; and the damage probability is calculated using the measured flight time and the theoretical flight time.

[0014] Preferably, the modified continuous group velocity function is:

[0015]

[0016] In the formula, v g (θ) is the modified continuous group velocity function, reflecting the change of wave velocity with direction in CFRP containing an elastic layer; i is the sensing path number, the i-th sensing path; θ is the target point, a given angle parameter; θ i θ i+1 Let θ be the two nearest neighbors of the target point; To find the two nearest neighbors θ i θ i+1 The corresponding velocity values; θ1 is the initial angle; θ2 is the angle greater than θ1 and closest to θ1; This is the velocity value corresponding to θ1; This is the velocity value corresponding to θ2; θ n The last angle; θ n-1 For less than θ n And with θ n The closest angle; For θ n The corresponding speed value; For θ n-1 The corresponding speed value.

[0017] Preferably, the formula for calculating the theoretical flight time using the modified continuous group velocity function is as follows:

[0018]

[0019] In the formula, (x,y) represents each grid intersection point, i.e., the location of the damage; T i TH (x,y) represents the theoretical flight time of the i-th sensing path when the damage location is (x,y); (x r ,y r ) is the receiving point; θ r,i The angle of each grid intersection point relative to the receiving point; v g (θ r,i The continuous group velocity function of the angle between each grid intersection point and the receiving point, by θ r,i Substituting the corrected continuous group velocity function v g (θ) is obtained; (x) e ,y e ) is the excitation point; θ e,i The angle of each grid intersection point relative to the excitation point; v g (θ e,i The continuous group velocity function of the angle between each grid intersection point and the excitation point, expressed as θ. e,i Substituting the corrected continuous group velocity function v g It is obtained from (θ).

[0020] Preferably, the formula for calculating the damage probability using the optimized probability-weighted method is as follows:

[0021]

[0022] In the formula, p elliptical (x,y) represents the damage probability calculated at the damage location (x,y) using the optimized elliptical trajectory method, i.e., the probability distribution of the elliptical trajectory; N P The total number of sensing paths; τ0 is the attenuation factor; T i Let be the measured flight time of the i-th sensor path.

[0023] Preferably, the process of introducing a dynamic short-axis optimization model to optimize the probability weighted method is as follows: for the Lamb wave signals of carbon fiber composite laminates with elastomer layers in healthy and damaged states, the first wave packet data of the Lamb wave signal collected by the receiver of each sensing path is selected, the main damage path is screened by dynamic shape parameter calculation, and an adaptive imaging model is constructed based on the screening results.

[0024] Preferably, the formula for calculating the damage probability using the optimized probability weighting method is as follows:

[0025]

[0026] In the formula, p rapid (x,y) represents the damage probability calculated at the damage location (x,y) using the optimized probability weighting method, i.e., the probability distribution of the probability weighting; k is the ellipse number, i.e., the ellipse determined by the k-th sensing path; M is the total number of ellipses; ρ is the correlation coefficient; W′ k (x,y) is the weighting function; c k Let α be the coordinates of the center of the ellipse; k For shape parameters.

[0027] Preferably, the prior distribution is:

[0028]

[0029] In the formula, p prior (x,y) is the prior distribution.

[0030] Preferably, the formula for constructing the likelihood function is:

[0031]

[0032] In the formula, p(D|U) is the likelihood function; D is the set of representations of the measured flight times. U represents the uncertain parameter, U = (x, y, σ); σ is an adjustable parameter that reflects the uncertainty of measurement noise.

[0033] The beneficial effects of this invention are:

[0034] The damage localization method for elastomeric layered carbon fiber composites provided by this invention accurately locates damage and reduces errors through a layered wave velocity correction model and a dynamic minor axis optimization algorithm. The adaptive MCMC algorithm and optimized Bayesian fusion framework enhance the anti-interference capability. It is also applicable to a variety of multilayered heterogeneous composite materials and has been widely used in many fields. It also has higher computational efficiency and more flexible parameter optimization. It effectively improves the accuracy, robustness and material adaptability of damage localization in elastomeric layered carbon fiber composites. Attached Figure Description

[0035] Figure 1 This is a flowchart of the damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion as described in this invention.

[0036] Figure 2This is a schematic diagram showing the sensor location and damage location on the carbon fiber composite laminate containing an elastomer layer as described in this invention.

[0037] Figure 3 This is a schematic diagram of the scattering signal described in this invention.

[0038] Figure 4 This is the time-frequency spectrum of the continuous wavelet transform of the scattering signal described in this invention.

[0039] Figure 5 This is a graph illustrating the ToF calculation based on wavelet transform as described in this invention.

[0040] Figure 6 This is a diagram showing the location result of the damage location 1 using existing technology as described in this invention.

[0041] Figure 7 This is a diagram showing the location result of damage location 1 using the elliptic probability fusion-based elastomeric carbon fiber composite material damage location method described in this invention.

[0042] Figure 8 This is a diagram showing the location result of damage location 2 using existing technology as described in this invention.

[0043] Figure 9 This is a diagram showing the location result of damage location 2 using the elliptic probability fusion-based elastomeric layer carbon fiber composite material damage location method described in this invention.

[0044] Figure 10 This is a diagram showing the location result of the damage location 3 using existing technology as described in this invention.

[0045] Figure 11 This is a diagram showing the location result of damage location 3 using the elliptic probability fusion-based elastomeric layer carbon fiber composite material damage location method described in this invention. Detailed Implementation

[0046] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0047] like Figure 1 As shown, this invention provides a damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion. The specific implementation process is as follows:

[0048] Step 1: Sensors are installed on the damaged carbon fiber reinforced polymer (CFRP) laminate containing an elastomeric layer, and the relative arrangement of the sensors is determined. Multiple sensing paths are formed using different sensors as excitation sources. The ultrasonic guided wave damage diagnosis platform system is used to collect Lamb wave signals from these multiple sensing paths. The ultrasonic guided wave damage diagnosis platform system includes: an excitation source sensor, a receiving sensor, and a wave signal processing algorithm module. The excitation source sensor is used to transmit excitation Lamb wave signals, and the receiving sensor is used to collect Lamb wave signals in both the damaged and healthy states of the CFRP laminate containing the elastomeric layer. Under the same sensing path, the wave signal processing algorithm module calculates the difference between the Lamb wave signals in the damaged and healthy states collected by the receiving sensor to obtain the scattered signal.

[0049] Step 2: Perform continuous wavelet transform on the scattered signal using complex Morlet wavelets; selecting the complex Morlet wavelet as the basis function, the mapping relationship between each frequency and the scale parameter when analyzing the scattered Lamb wave using the complex Morlet wavelet is expressed as follows:

[0050]

[0051] In the formula, f is the excitation frequency of the Lamb wave, f mz It is the center frequency of the complex Morlet wavelet, f sc is the sampling frequency of the scattered signal; 'a' is the scale parameter, which determines the time-frequency resolution of the basis function and corresponds to the frequency domain distribution characteristics of the scattered signal. It is used to control the up and down scaling (frequency change) of the wavelet function.

[0052] The continuous wavelet transform of the scattered signal can be defined as the inner product operation of the basis function and the scattered signal at different scales and locations, expressed as:

[0053]

[0054] In the formula, CWT(a,b) are the continuous wavelet transform (CWT) coefficients, representing the local characteristics of the scattered signal at the corresponding time and frequency; b is the displacement parameter, characterizing the time localization of the basis function, reflecting the time evolution of the scattered signal characteristics, and used to control the left and right shifts (time variations) of the wavelet function; S(t) is the scattered signal; ψ is the basis function, which is a function with localization characteristics in both time and frequency, used to extract the time-frequency characteristics of the scattered signal; ψ * t represents the complex conjugate of the basis functions; t is time.

[0055] Using the frequency change corresponding to the scale parameter a (obtained through the frequency-scale mapping relationship) as the ordinate and the time change corresponding to the displacement parameter b as the abscissa, the time-frequency spectrum of the continuous wavelet transform of the scattered signal is generated.

[0056] The scattered signal after continuous wavelet transform and the excitation Lamb wave signal emitted by the excitation source sensor are combined in the ToF calculation diagram based on wavelet transform. The time corresponding to the first peak of the scattered signal is subtracted from the time corresponding to the peak of the excitation signal to calculate the time of flight characteristic, i.e., the measured time of flight (ToF).

[0057] Step 3: Discretize the detection area into multiple grid structures, with each grid intersection having an independent theoretical flight time; assuming that the intersection of each grid is the damage location, calculate the damage probability of each grid intersection using the measured flight time and the theoretical flight time.

[0058] When calculating theoretically predicted flight time, a layered wave velocity correction model is introduced to perform wave velocity anisotropy correction and adjustment, avoiding the errors caused by assuming isotropic wave velocities in traditional methods.

[0059] Group velocity samples in different directions were obtained using a DC (Dispersion Calculator). A continuous group velocity function was constructed using linear interpolation, and a hierarchical wave velocity correction model was introduced to obtain the corrected continuous group velocity function, expressed as:

[0060]

[0061] In the formula, v g (θ) is the modified continuous group velocity function, reflecting the change of wave velocity with direction in CFRP containing an elastic layer; i is the sensing path number, the i-th sensing path; θ is the target point, a given angle parameter; θ i θ i+1 Let θ be the two nearest neighbors of the target point; To find the two nearest neighbors θ i θ i+1 The corresponding velocity values; θ1 is the initial angle; θ2 is the angle greater than θ1 and closest to θ1; This is the velocity value corresponding to θ1; This is the velocity value corresponding to θ2; θ n The last angle; θ n-1 For less than θ n And with θ n The closest angle; For θ n The corresponding speed value; For θ n-1 The corresponding speed value.

[0062] Calculating the theoretical flight time at each grid intersection point requires determining the group velocity for each path segment using angles:

[0063] Each grid intersection point (x, y) is relative to the excitation point (x). e ,y e ) angle θ e,i It can be represented as:

[0064] θ e,i =arctan 2(y e -y,x e -x)

[0065] In the formula, (x, y) represents each grid intersection point, i.e., the location of the damage; (x e ,y e ) is the excitation point; θ e,i Given the angle of each grid intersection point relative to the excitation point; arctan 2 is the arctangent function, which can correctly handle the quadrant of the angle and returns a value in the range (-π, π].

[0066] Each grid intersection point (x, y) is relative to the receiving point (x). r ,y r ) angle θ r,i It can be represented as:

[0067] θ r,i =arctan 2(y r -y,x r -x)

[0068] In the formula, (x r ,y r ) is the receiving point; θ r,i The angle of each grid intersection point relative to the receiving point;

[0069] Based on the above perspective, the theoretical flight time for each grid intersection point is calculated using the optimized elliptical trajectory method as follows:

[0070]

[0071] In the formula, T i TH (x,y) represents the theoretical flight time of the i-th sensing path when the damage location is (x,y); v g (θ r,i The continuous group velocity function of the angle between each grid intersection point and the receiving point, by θ r,i Substituting the corrected continuous group velocity function v g obtained from (θ); v g (θe,i The continuous group velocity function of the angle between each grid intersection point and the excitation point, expressed as θ. e,i Substituting the corrected continuous group velocity function v g It is obtained from (θ).

[0072] The error between the theoretically predicted damage location and the actual damage location is denoted by ε. The relationship between the measured flight time and the theoretical flight time of the i-th sensor path is as follows:

[0073] T i =T i TH (x,y)+ε

[0074] In the formula, T i Let be the measured flight time of the i-th sensor path, in seconds; ε is the error variable.

[0075] The damage probability at each grid intersection point is calculated using the measured flight time and the theoretical flight time:

[0076]

[0077] In the formula, p elliptical (x,y) represents the damage probability calculated at the damage location (x,y) using the optimized elliptical trajectory method, i.e., the probability distribution of the elliptical trajectory; N P τ represents the total number of sensing paths; τ0 is the attenuation factor, which is set to 0.01ms in this study.

[0078] By incorporating the anisotropic characteristics of wave velocity into the theoretical time-of-flight (ToF) calculation of the elliptical trajectory method, the damage localization accuracy is optimized; the optimized elliptical trajectory method significantly improves the damage localization accuracy.

[0079] Step 4: For the Lamb wave signals of the CFRP laminate with elastomer layer in both healthy and damaged states, select the first wave packet data of the Lamb wave signal collected by the receiving sensor at the receiving end of each sensing path. Introduce a dynamic short axis optimization model and use dynamic shape parameter calculation to replace the dependence on fixed parameters in traditional methods. By screening the main damage paths and eliminating redundant or interfering information, the data quality is optimized. Based on the screening results, an adaptive imaging model is constructed to enhance the damage localization capability. The damage probability of each grid intersection point is calculated using the optimized probability weighting method:

[0080]

[0081] In the formula, p rapid(x,y) represents the damage probability calculated at the damage location (x,y) using the optimized probability weighting method, i.e., the probability distribution of the probability weighting; k is the ellipse number, i.e., the ellipse determined by the k-th sensing path; M is the total number of ellipses; ρ is the correlation coefficient; W′ k (x,y) is the weighting function; c k Let α be the coordinates of the center of the ellipse; k For shape parameters.

[0082] The optimized probability weighting method significantly improves the accuracy of damage localization.

[0083] Step 5: Fuse the probability distribution of the elliptical trajectory method and the probability distribution of the probability weighting method to generate the prior distribution, expressed as:

[0084]

[0085] In the formula, p prior (x,y) is the prior distribution;

[0086] The prior distribution is calculated using the contributions of the geometric mean balanced elliptical trajectory method and the probability weighting method, and then normalized.

[0087] The uncertain parameters are represented as a vector U = (x, y, σ); the measured flight time is represented as... Based on the difference between the measured flight time and the corrected theoretically predicted flight time, the formula for constructing the likelihood function is as follows:

[0088]

[0089] In the formula, p(D|U) is the likelihood function; σ is an adjustable parameter that reflects the uncertainty of measurement noise;

[0090] Assume the error variable ε follows a normal distribution with a mean of zero and a standard deviation of σ; the likelihood function can be written as:

[0091]

[0092] Using Bayesian estimation, the prior distribution is multiplied by the likelihood function and normalized to obtain the posterior distribution of the damage location:

[0093] p(U|D)∝p(D|U)p prior (x,y)

[0094] In the formula, p(U|D) is the posterior probability density function for predicting the damage coordinates, i.e., the posterior distribution.

[0095] Step 6: Sample the posterior distribution using the adaptive MCMC algorithm. The specific process is as follows:

[0096] Step 1: Set the initial state (x0, y0, σ0) = (125, 125, 5), located near the center of the grid. This choice is based on the geometric symmetry of the grid range (0 to 300 mm) to ensure that the initial point is representative.

[0097] Step 2: Generate candidate states using a normal distribution:

[0098]

[0099] In the formula, x prop The proposed x-coordinate value; y prop The proposed y-coordinate value; σ prop The proposed standard deviation; sd x This is the standard deviation control parameter for the x-coordinate, used to adjust the exploration range of the x-coordinate; sd y This is the standard deviation control parameter for the y-coordinate, used to adjust the exploration range of the y-coordinate; sd σ This is the standard deviation control parameter, used to adjust the exploration range of the standard deviation; To represent a standard normal distribution with a mean of 0 and a variance of 1, it is a random variable used to introduce randomness;

[0100] in,

[0101] Standard deviation of normal distribution sd x sd y sd σ The jump range of candidate states is controlled to balance exploration efficiency and acceptance rate.

[0102] Step 3: Set the sampling boundary conditions; ensure that the target value meets the constraints of the actual physical environment during sampling. The sampling boundary conditions are as follows:

[0103]

[0104] Step 3 avoids sampling outside the grid range or generating non-physical noise parameters.

[0105] Step 4: Calculate the receiver rate for (x, y) and σ for each sampled value:

[0106] α=min{1,exp[log p(x prop ,y prop ,σ prop |D)-log p(U|D)]}

[0107] In the formula, α is an acceptance probability used to determine whether to accept the new proposal state.

[0108] If the random number rand < α, then accept the new state (x). prop ,y prop ,σ prop Otherwise, the current state is retained; this mechanism ensures that the chain converges to the posterior distribution.

[0109] Step 5: Perform sampling, with a total sampling count N = 100,000. Through the above steps, the adaptive MCMC algorithm successfully generates samples of the posterior distribution.

[0110] The sampling results of the adaptive MCMC algorithm are combined with the posterior probability density function for visualization to generate imaging results of the damage location; the region with a confidence level greater than 85% is selected, the final imaging result of the damage location is output, the damage location is predicted, and the damage location is located.

[0111] Step 7: By comparing the error between the actual damage location and the predicted damage location, evaluate the positioning accuracy of the improved method.

[0112] In the elliptic probability fusion-based damage localization method for carbon fiber composite materials, the machine learning model can be used to replace the layered wave velocity correction model to predict the wave velocity distribution, but it requires a large amount of training data and is costly; a statistical model can be used to replace the Bayesian framework, but the computational complexity increases significantly.

[0113] like Figure 2 As shown, in this embodiment, the CFRP laminate containing the elastomer layer is a square with a side length of 300mm; a sensor is set at each of the four corners of the CFRP laminate containing the elastomer layer, and the sensor is 50mm from the edge; three impact positions are set in the CFRP laminate containing the elastomer layer as the damage positions to be detected, namely: impact position 1 (125, 125), impact position 2 (200, 125) and impact position 3 (125, 150).

[0114] To address damage at different locations in CFRP laminates containing elastomer layers, Lamb wave technology was used. Sensors 1, 2, and 3 were used as excitation source sensors, and sensors 1, 2, 3, and 4 were used as receiving sensors. Any two sensors formed a sensing path, and Lamb wave detection signals from a total of 6 sensing paths were collected. A sampling frequency of 12MHz was used to record 4000 data points to ensure high signal resolution and integrity of time-domain information.

[0115] like Figure 3As shown, the ultrasonic guided wave injury diagnosis platform system obtains the scattered signal by subtracting the Lamb wave signals in the damaged state and the healthy state collected by the receiving sensor; the horizontal axis of the spectrum of the scattered signal is time in seconds (s), and the vertical axis is amplitude in millimeters (mm).

[0116] like Figure 4 As shown, the time-frequency spectrum of the scattered signal after continuous wavelet transform is shown, with the horizontal axis representing time in seconds (s) and the vertical axis representing the normalized amplitude.

[0117] like Figure 5 As shown, the Time-of-Flight (ToF) calculation graph based on wavelet transform has the horizontal axis representing time in seconds (s) and the vertical axis representing the normalized amplitude. The measured Time of Flight (ToF) is obtained by subtracting the time corresponding to the first peak of the scattered signal from the time corresponding to the peak of the excitation signal.

[0118] like Figure 6-11 As shown, the black circles represent the actual damage locations, and the red crosses represent the predicted damage locations. Under the same data acquisition conditions, compared with existing technologies, the elliptic probability fusion-based damage localization method for elastomeric layer carbon fiber composite materials provided by this invention significantly improves the localization accuracy while effectively reducing the misjudgment rate of non-damaged locations.

[0119] This invention provides a damage localization method for carbon fiber composites with elastomeric layers based on elliptic probability fusion. Specifically, for carbon fiber composites containing elastomeric layers, a layered wave velocity correction model is proposed, which can accurately describe the wave velocity distribution in the elastomeric and carbon fiber layers, particularly the wave velocity transition characteristics at the interface between the elastomeric and carbon fiber layers. The layered wave velocity correction model constructs a continuous group velocity function through linear interpolation, solving the localization error problem caused by the single wave velocity assumption in traditional methods. The elliptic trajectory method is improved by introducing the layered wave velocity correction model to dynamically adjust the elliptic equation to adapt to the composite structure containing elastomeric layers. The improved elliptic trajectory method can significantly improve the accuracy of damage localization, especially in complex multilayer heterogeneous materials. A dynamic minor axis optimization algorithm is proposed, replacing the fixed shape optimization in traditional methods. The parameter dependency was optimized by filtering the main damage paths and removing redundant information, thus improving data quality and further enhancing the accuracy of damage localization. This also improved the adaptability of the probability weighting method to complex signal propagation characteristics, making it suitable for multilayer heterogeneous materials. In the Bayesian estimation framework, the damage localization results were optimized by combining the wave velocity-corrected time-of-flight data (likelihood function) with the prior distribution of path correlation. The introduction of the adaptive MCMC algorithm can dynamically adjust the sampling step size according to the posterior distribution, ensuring efficient convergence of the posterior distribution, reducing computational complexity, and improving the real-time performance of damage localization, making it suitable for real-time monitoring scenarios. To address the nonlinear changes in wave velocity and path interference that may be introduced by the elastic body layer, the Bayesian fusion framework was optimized, enhancing the robustness of the localization results and ensuring stable convergence and high-precision localization even in complex noise environments.

[0120] The present invention provides a damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion, which can monitor damage caused by collision or fatigue in real time in the field of new energy vehicles and extend battery pack life; accurately locate minute damage in the aerospace field to avoid structural failure caused by the accumulation of hidden damage; and reduce downtime maintenance time and improve equipment operation reliability in the field of high-end equipment maintenance.

[0121] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion, characterized in that, include: Step 1: Install sensors on the damaged carbon fiber composite laminate containing elastomeric layers, use different sensors as excitation sources, and use the ultrasonic guided wave damage diagnosis platform system to collect Lamb wave signals from multiple sensing paths. Step 2: Perform continuous wavelet transform on the Lamb wave signal using complex Morlet wavelets to extract the measured flight time; Step 3: Introduce a layered wave velocity correction model to optimize the elliptical trajectory method, and use the optimized elliptical trajectory method to calculate the theoretical flight time and damage probability; The corrected continuous group velocity function in the layered wave velocity correction model is: ； In the formula, The corrected continuous group velocity function reflects the change of wave velocity with direction in CFRP containing an elastic layer; The number of the sensing path, the first One sensing path; Given the angle parameters for the target point; , For target point The two nearest neighbors; , To the two nearest neighbors , The corresponding speed value; The starting angle; greater than And with The closest angle; To and The corresponding speed value; To and The corresponding speed value; The last angle; Less than And with The closest angle; To and The corresponding speed value; To and The corresponding speed value; Step 4: Introduce a dynamic short-axis optimization model to optimize the probability weighting method, and use the optimized probability weighting method to calculate the damage probability; Step 5: Using the damage probability calculated by the elliptical trajectory method and the probability weighting method, generate a prior distribution; based on the difference between the measured flight time and the theoretical flight time, construct a likelihood function; and by combining the prior distribution and the likelihood function with Bayes' theorem, obtain the posterior distribution of the damage location. Step 6: Use the adaptive MCMC algorithm to sample the posterior distribution and generate a probability cloud map of the damage location; output the final imaging result of the damage location to locate the damage location.

2. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 1, characterized in that, The process of introducing a layered wave velocity correction model to optimize the elliptical trajectory method and calculating the theoretical flight time and damage probability is as follows: Group velocity samples in different directions are obtained using a DC calculator; a continuous group velocity function is constructed using linear interpolation; a layered wave velocity correction model is introduced to obtain the corrected continuous group velocity function; the theoretical flight time is calculated using the corrected continuous group velocity function; and the damage probability is calculated using the measured flight time and the theoretical flight time.

3. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 2, characterized in that, The formula for calculating the theoretical flight time using the modified continuous group velocity function is as follows: ； In the formula, For each grid intersection point, i.e., the location of the damage; The location of the injury is Time The theoretical flight time of each sensing path; For receiving points; The angle of each grid intersection point relative to the receiving point; The continuous group velocity function of the angle between each grid intersection point and the receiving point, by... Substituting the modified continuous group velocity function Obtained from; As an incentive point; The angle of each grid intersection point relative to the excitation point; The continuous group velocity function of the angle of each grid intersection point relative to the excitation point is obtained by... Substituting the modified continuous group velocity function It was obtained from the middle.

4. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 3, characterized in that, The formula for calculating the damage probability using the optimized elliptical trajectory method is as follows: ； In the formula, To utilize the optimized elliptical trajectory method at the damage location The damage probability calculated at that point is the probability distribution of the elliptical trajectory. This represents the total number of sensor paths; It is the attenuation factor; For the first Measured flight time of the sensor path.

5. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 1, characterized in that, The process of introducing a dynamic short-axis optimization model to optimize the probability weighted method is as follows: For the Lamb wave signals of carbon fiber composite laminates with elastomer layers in both healthy and damaged states, the first wave packet data of the Lamb wave signal collected by the receiver of each sensing path is selected, and the main damage paths are screened by dynamic shape parameter calculation. Based on the screening results, an adaptive imaging model is constructed.

6. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 5, characterized in that, The formula for calculating the damage probability using the optimized probability-weighted method is as follows: ； In the formula, To utilize the optimized probability weighting method at the damage location The damage probability calculated at that point is the probability-weighted probability distribution. The number of the ellipse, i.e., the _____. The ellipse determined by the individual sensing paths; This represents the total number of ellipses. The correlation coefficient; For weighting functions; The coordinates of the center of the ellipse; For shape parameters.

7. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 4 or 6, characterized in that, The prior distribution is: ； In the formula, This is the prior distribution.

8. The damage localization method for elastomeric layer carbon fiber composite materials based on elliptic probability fusion according to claim 7, characterized in that, The formula for constructing the likelihood function is: ； In the formula, It is the likelihood function; This is the set of representations of the measured flight times. ; For the representation of uncertain parameters, ; This is an adjustable parameter that reflects the uncertainty of measurement noise.

Images