A method and system for identifying corrosion defects based on multimodal mechanical wave characteristics

Abstract

This invention discloses a corrosion defect identification method and system based on multimodal mechanical wave characteristics, comprising: exciting multimodal mechanical waves and planning their propagation paths; arranging a sensor array to acquire and synchronously process multimodal mechanical wave signals; performing time-frequency analysis on the acquired multimodal mechanical wave signals to extract phase change characteristics; locating corrosion defects based on the phase change characteristics and propagation path information; and identifying and classifying different types of corrosion defects by combining phase change, amplitude attenuation, and propagation path offset information. This invention achieves comprehensive coverage of the material detection area through the excitation and propagation path planning of multimodal mechanical waves, reducing the possibility of signal distortion and missed detections. Through time-frequency analysis and phase difference analysis, it improves the ability to identify minute defects.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology for materials, and in particular to a method and system for identifying corrosion defects based on multimodal mechanical wave characteristics. Background Technology

[0002] In recent years, with the aging of industrial equipment and the increasingly harsh service environment, material corrosion has become one of the important factors affecting the safe operation of equipment. Corrosion not only reduces the physical properties of materials, leading to a decrease in structural strength and durability, but may also trigger catastrophic accidents, causing huge economic losses and social impacts. Therefore, the effective detection and assessment of corrosion defects has become a key technical means to ensure the long-term stable operation of industrial facilities.

[0003] Traditional methods typically use only a single mode of mechanical wave (such as ultrasound or electromagnetic waves) for detection. However, when dealing with materials of uneven thickness or complex structures, this approach is prone to signal distortion due to reflection and refraction, thus affecting the accuracy of defect identification. Furthermore, single-mode mechanical waves cannot comprehensively cover all internal defect information within the material, limiting the comprehensiveness and reliability of the detection. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides a corrosion defect identification method and system based on multimodal mechanical wave characteristics to solve the problem that single mechanical wave defect detection cannot fully cover the interior and surface of materials.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] In a first aspect, the present invention provides a corrosion defect identification method based on multimodal mechanical wave characteristics, comprising:

[0009] Excite multimodal mechanical waves and plan the propagation path of the mechanical waves;

[0010] Based on the planned mechanical wave propagation path, a sensor array is deployed to collect and synchronously process multimodal mechanical wave signals;

[0011] Time-frequency analysis was performed on the acquired multimodal mechanical wave signals to extract the phase change characteristics of the multimodal mechanical waves;

[0012] Based on the phase change characteristics of multimodal mechanical waves and combined with the propagation path information of mechanical waves, the location of corrosion defects can be determined.

[0013] Based on the location of corrosion defects, combined with information on phase changes, amplitude attenuation, and propagation path offset, different types of corrosion defects are identified and classified.

[0014] As a preferred embodiment of the corrosion defect identification method based on multimodal mechanical wave characteristics described in this invention, wherein:

[0015] The process of exciting multimodal mechanical waves and planning their propagation paths includes the following steps:

[0016] Use an ultrasonic thickness gauge to determine the thickness distribution of the material;

[0017] Using a laser scanner, a three-dimensional model of the material geometry is created to obtain the material's geometric properties;

[0018] Based on the material thickness distribution, the wavelength range matching the material thickness is calculated.

[0019] Multimode mechanical waves are formed by combining low-frequency longitudinal waves, high-frequency transverse waves, and surface waves.

[0020] The reflection coefficients of different modal mechanical waves at the material boundary were calculated by finite element simulation, and the wave mode with the highest reflection value was selected.

[0021] Multimodal mechanical waves are excited using a transducer, and the excitation angle is adjusted.

[0022] Based on the material's geometry and thickness distribution, finite element analysis software is used to simulate the propagation path of mechanical waves within the material.

[0023] The wave propagates from the excitation point, passes through reflection or refraction within the material, and reaches each sensor along its propagation path, with each path corresponding to a sensor location.

[0024] Based on the simulation results, the excitation point was adjusted to fully cover the material detection area.

[0025] As a preferred embodiment of the corrosion defect identification method based on multimodal mechanical wave characteristics described in this invention, the step of arranging the sensor array to acquire and synchronously process multimodal mechanical wave signals includes the following steps:

[0026] A highly sensitive array of ultrasonic sensors is arranged on the surface of the material, with each sensor capturing signals of different mechanical wave modes;

[0027] Based on the multimodal mechanical wave propagation path and boundary reflection characteristics obtained from finite element simulation, the sensor arrangement was adjusted.

[0028] The sensor spacing is set based on the excitation wavelength of multimodal mechanical waves;

[0029] Based on the material's geometric properties, sensors are uniformly arranged along the material surface, with increased density at geometric changes.

[0030] For materials with uneven thickness and complex structures, sensors are placed on the sides and back of the material;

[0031] Using standard samples with no defects and known defects, multimodal mechanical waves are excited, and response signals are collected through sensors. The response delay and sensitivity of each sensor are calibrated.

[0032] High-speed acquisition of multimodal mechanical wave signals;

[0033] Each sensor synchronously and in real time collects amplitude, phase and propagation time information of different modes of mechanical waves;

[0034] During the acquisition process, a low-pass filter is used to remove high-frequency environmental noise, and a band-pass filter is used for directional filtering of the excitation wave frequency range.

[0035] The amplitude, phase, and propagation time of the acquired multimodal mechanical waves are normalized.

[0036] As a preferred embodiment of the corrosion defect identification method based on multimodal mechanical wave characteristics described in this invention, the step of performing time-frequency analysis on the acquired multimodal mechanical wave signals to extract the phase change characteristics of the multimodal mechanical waves includes the following steps:

[0037] The wavelet transform is used for time-domain and frequency-domain analysis, as shown in the following expression:

[0038]

[0039] Where W(a,b) are the coefficients after wavelet transform, a is the scaling parameter, b is the translation parameter, x(t) is the original multimodal mechanical wave propagation signal at time t, and ψ(t) is the wavelet basis function. * (t) is the complex conjugate of the wavelet basis functions. Let b be the local feature of the signal at scale a, j be the imaginary unit, π be the value of pi, ω0 be the center frequency of the wavelet, and dt represent the integration over time t.

[0040] Extract the instantaneous phase of the multimodal mechanical wave signal after wavelet transform processing;

[0041] Calculate the phase difference between adjacent time points and find the phase abrupt change point;

[0042] Based on the wave propagation characteristics and noise level of the material, a phase change threshold Δφ is set using experimental data. th When the phase difference Δφ(t) between adjacent times is greater than Δφ th When this occurs, it is determined to be a phase abrupt change point.

[0043] As a preferred embodiment of the corrosion defect identification method based on multimodal mechanical wave characteristics described in this invention, the method for locating the corrosion defect location based on the phase change characteristics of multimodal mechanical waves and combined with mechanical wave propagation path information includes the following steps:

[0044] Based on the simulation results of the propagation path of mechanical waves using finite element analysis software, and according to the mechanical properties of the material and the modes of the waves, the propagation speed of mechanical waves of different modes in the material is determined through experimental measurement.

[0045] Each sensor calculates the time from excitation to detection of the mechanical wave by simulating the propagation path length and wave speed using mechanical waves.

[0046] A reference sensor is placed at the mechanical wave excitation point, and the time when the reference sensor receives the mechanical wave is the reference time.

[0047] The time when the mechanical wave is received by the other sensors is compared with the reference time to obtain the time difference between the reference sensor and the other sensors;

[0048] Based on the 3D modeling structure of the material, the positioning coordinates (x, y, y) of each sensor in the sensor array are set. i ,y i ,z i );

[0049] Draw circles with the coordinates of each sensor as centers and the product of the time difference and wave speed as radii. The defect location is at the intersection of these circles. The expression is:

[0050]

[0051] Among them, (x f ,y f ,z f ) represents the coordinates of the defect location, v represents the propagation speed of the mechanical wave in the material, and ΔT represents the velocity. i Let i be the time difference between the i-th sensor and the reference sensor, where i is the index of the number of sensors.

[0052] A nonlinear equation solving algorithm is used to minimize the error of multiple circles and obtain the location of corrosion defects.

[0053] As a preferred embodiment of the corrosion defect identification method based on multimodal mechanical wave characteristics described in this invention, the step of using a nonlinear equation solving algorithm to minimize the error of multiple circles and obtain the location of the corrosion defect includes the following steps:

[0054] A target function is constructed based on the least squares method to calculate the sum of squared errors S(x) of the sensor relative to the reference sensor. f ,y f ,z f The expression is as follows:

[0055]

[0056] Where N is the number of sensors;

[0057] The gradient descent method is used to iteratively update the position coordinates along the gradient direction by calculating the partial derivative of the objective function with respect to the defect location.

[0058] As a preferred embodiment of the corrosion defect identification method based on multimodal mechanical wave characteristics described in this invention, the step of identifying the type of corrosion defect by combining phase change, amplitude attenuation, and propagation path offset information includes the following steps:

[0059] Extracting the characteristics of mechanical wave propagation in materials;

[0060] The amplitude attenuation value of the mechanical wave is obtained by calculating the difference between the amplitude when the mechanical wave is excited and the amplitude when the sensor receives it.

[0061] By calculating the product of the time difference and the wave speed when the mechanical wave arrives at the same position with and without defects, the propagation path offset and phase can be obtained.

[0062] The amplitude attenuation value, propagation path offset value, and phase difference obtained from each sensor are used as a data sample;

[0063] By controlling the experimental conditions, samples with different types of corrosion defects were manufactured, and the types of corrosion defects and their corresponding features were detected. The features were then labeled with corrosion defect categories.

[0064] The input features are mapped to a high-dimensional space using radial basis functions. The decision function f(B) of the support vector machine model is calculated by comparing the similarity between the new sample B1 and the training sample B2, and is expressed as follows:

[0065]

[0066] Where sign is the sign function, H is the number of samples, j is the sample index, α is the Lagrange multiplier, and C... jLet K(B1,B2) be the class label of the j-th training sample, K(B1,B2) be the radial basis function, and d be the bias term.

[0067] K(B1,B2)=exp(-γ||B1-B2|| 2 );

[0068] Where γ is the width of the radial basis function, and ||B1-B2|| is the Euclidean distance between the new sample B1 and the training sample B2;

[0069] The support vector machine model is trained using labeled feature data, and the regularization parameter D and the width γ of the radial basis function in the support vector machine model are optimized.

[0070] The regularization parameter D and the width γ of the radial basis function are optimized through cross-validation.

[0071] After training, the phase difference, amplitude attenuation, and propagation path offset of the detected multimodal mechanical waves are used as comprehensive features to input into the support vector machine model for prediction, and the category label of corrosion defects is output.

[0072] Secondly, the present invention provides a corrosion defect identification system based on multimodal mechanical wave characteristics, comprising:

[0073] The mechanical wave excitation module is used to excite multimodal mechanical waves and plan the propagation path of the mechanical waves;

[0074] The signal acquisition module is used to arrange a sensor array according to the planned mechanical wave propagation path, acquire and synchronously process multimodal mechanical wave signals;

[0075] The feature extraction module is used to perform time-frequency analysis on the acquired multimodal mechanical wave signals and extract the phase change features of the multimodal mechanical waves;

[0076] The positioning module is used to locate the position of corrosion defects based on the phase change characteristics of multimodal mechanical waves and the information on the propagation path of mechanical waves.

[0077] The identification module is used to identify and classify different types of corrosion defects based on the location of the corrosion defect, combined with information on phase change, amplitude attenuation, and propagation path offset.

[0078] Thirdly, the present invention provides a computing device, comprising:

[0079] Memory, used to store programs;

[0080] A processor is configured to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the corrosion defect identification method based on multimodal mechanical wave characteristics.

[0081] Fourthly, the present invention provides a computer-readable storage medium comprising: when the program is executed by a processor, implementing the corrosion defect identification method based on multimodal mechanical wave characteristics.

[0082] The beneficial effects of this invention are as follows: By using multimodal mechanical wave excitation and propagation path planning, this invention achieves comprehensive coverage of the material detection area, reducing the possibility of signal distortion and missed detections; through a high-sensitivity sensor array and precise signal processing, it improves signal quality and signal-to-noise ratio; through time-frequency analysis and phase difference analysis, it enhances the ability to identify minute defects; through a precise positioning algorithm, it improves the location accuracy of corrosion defects; and through an intelligent identification algorithm, it improves the classification accuracy of corrosion defect types. These beneficial effects collectively ensure the high efficiency, accuracy, and reliability of corrosion defect detection, providing strong technical support for the safe operation of industrial equipment and infrastructure. Attached Figure Description

[0083] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0084] Figure 1 This is a schematic diagram of the basic process of a corrosion defect identification method based on multimodal mechanical wave characteristics, provided as an embodiment of the present invention. Detailed Implementation

[0085] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0086] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0087] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0088] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0089] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0090] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0091] Example 1

[0092] Reference Figure 1 As an embodiment of the present invention, a corrosion defect identification method based on multimodal mechanical wave characteristics is provided, comprising:

[0093] S1: Excite multimodal mechanical waves and plan the propagation path of the mechanical waves;

[0094] In this embodiment, an ultrasonic thickness gauge is used to determine the thickness distribution of the material, especially for complex structures with non-uniform thickness, to ensure that complete thickness data is obtained.

[0095] In this embodiment of the application, the material surface, especially the detection area, is cleaned with a lint-free cloth and cleaning solution to remove dust, oil, and impurities, thereby reducing the interference of material surface contaminants on the propagation of mechanical waves.

[0096] In this embodiment of the application, a laser scanner is used to perform three-dimensional modeling of the material geometry and obtain the material's geometric properties;

[0097] In this embodiment, based on the material thickness distribution, the wavelength range matching the material thickness is calculated, and preferably a wavelength range of about 0.5 to 3 times the material thickness is used to ensure effective propagation of the wave in the material;

[0098] In this embodiment, a lower frequency longitudinal wave is used, which has strong penetration and is suitable for internal defect detection, while a higher frequency transverse wave and surface wave have high surface sensitivity and are suitable for detecting surface and near-surface corrosion defects, and are combined into a multimodal mechanical wave.

[0099] In this embodiment of the application, the reflection coefficients of different modes of mechanical waves at the material boundary are calculated by finite element simulation, and the wave mode with the highest reflection value is selected.

[0100] In the embodiments of this application, the optimal parameters for each wave mode can be selected in the entire "multimodal" system. For example, in a certain wave mode (such as the S0 mode or A0 mode in surface waves), the mode with the highest reflection coefficient is selected to ensure that the wave mode has the best rebound effect at the material boundary, thereby improving the detection sensitivity.

[0101] It should be noted that optimizing a certain mode (such as enhancing the boundary bounce value) is not the same as abandoning the multimodal detection method, but rather a refinement step aimed at improving the detection capability of each mode within its applicable range.

[0102] In the embodiments of this application, a transducer is used to excite multimodal mechanical waves, and the excitation angle is adjusted to meet the incident conditions of the waves in the material. For longitudinal waves, the waves are injected into the material at 0° to ensure that the waves enter the material perpendicularly and maximize their penetration depth. For transverse waves and surface waves, the waves are injected at 30° to 45° so that the waves can propagate better on the material surface and in thin-walled structures.

[0103] In this embodiment of the application, based on the geometry and thickness distribution of the material, finite element analysis software is used to simulate the propagation path of mechanical waves inside the material, focusing on the material boundaries, corners and curved areas to optimize the wave propagation path and avoid signal loss in these areas;

[0104] In the embodiments of this application, the propagation path of the wave originating from the excitation point, passing through reflection or refraction within the material, and reaching each sensor, with each path corresponding to a sensor position.

[0105] In this embodiment, based on the simulation of the propagation path and three-dimensional model of mechanical waves inside the material using finite element analysis software, areas with thinner thickness and complex geometry are identified. Based on the optimization of the propagation path by the finite element analysis software, the coverage and signal reflection intensity are maximized by adjusting the excitation angle and the position of the excitation point.

[0106] S2: Based on the planned mechanical wave propagation path, deploy a sensor array to collect and synchronously process multimodal mechanical wave signals;

[0107] In this embodiment, a highly sensitive array of ultrasonic sensors is arranged on the material surface, and each sensor is capable of capturing signals of different mechanical wave modes (such as longitudinal waves, transverse waves, surface waves, etc.).

[0108] It should be noted that these sensors should have a high frequency response capability and a bandwidth that covers the frequency range of the excitation wave (e.g., 1MHz to 10MHz) to ensure that the complete signal of the multimode wave can be captured.

[0109] In this embodiment of the application, the sensor arrangement is determined based on the multimodal mechanical wave propagation path and boundary reflection characteristics obtained from finite element simulation.

[0110] In this embodiment, based on the excitation wavelength of multimodal mechanical waves, the sensor spacing is set to 1 / 3 to 1 / 2 of the excitation wavelength to ensure that the propagation path of the wave in the material can be captured by each sensor without signal aliasing.

[0111] In this embodiment of the application, the propagation path and three-dimensional model of mechanical waves inside the material are simulated based on finite element analysis software. In areas where the material thickness varies greatly or the geometry is complex, the density of sensor arrangement is increased.

[0112] In this embodiment of the application, for materials with uneven thickness and complex structure, sensors are arranged on the side and back of the material to ensure that multiple reflections of the wave inside the material are fully captured.

[0113] It should be noted that areas with significant signal interference, such as signal overlap or interference caused by multiple reflections of mechanical wave signals, should be avoided to improve the quality of signal acquisition and ensure sufficient signal acquisition and accurate identification of defects.

[0114] In this embodiment, multimodal mechanical waves are excited using standard samples with no defects and known defects. Response signals are collected by sensors, and the response delay and sensitivity of each sensor are calibrated to ensure that the data acquisition from all sensors has consistent timing and response characteristics.

[0115] In this embodiment of the application, high-speed acquisition of multimodal mechanical wave signals is performed, and the sampling frequency should be more than 10 times the excitation wave frequency to ensure complete signal acquisition;

[0116] In the embodiments of this application, each sensor synchronously and in real time collects the amplitude information, phase information, and propagation time of different modes of mechanical waves;

[0117] In this embodiment of the application, during the acquisition process, a low-pass filter is used to remove high-frequency environmental noise. The cutoff frequency of the filter should be set to 1.5 to 2 times the excitation wave frequency to ensure that the effective signal passes through and remove higher frequency environmental noise. A band-pass filter is used to perform directional filtering for the excitation wave frequency range to avoid low-frequency and high-frequency interference and ensure that the received signal is concentrated in a specific frequency band.

[0118] In this embodiment of the application, the amplitude information, phase information and propagation time of the acquired multimodal mechanical waves are normalized.

[0119] In this embodiment of the application, the signal amplitude of all sensors is adjusted to the same standard range, and a normalization range of 0 to 1 or -1 to 1 is selected.

[0120] In this embodiment of the application, a reference signal is set, and phase normalization is performed by calculating the phase difference;

[0121] In this embodiment of the application, the time of all signals is aligned by calculating the time difference of arrival of the signals.

[0122] S3: Perform time-frequency analysis on the acquired multimodal mechanical wave signals to extract the phase change characteristics of the multimodal mechanical waves;

[0123] In this embodiment, wavelet transform is used for time-domain and frequency-domain analysis, as shown in the following expression:

[0124]

[0125] Where W(a,b) are the coefficients after wavelet transform, representing the local features of the signal at scale a and displacement b, a is the scale parameter controlling the compression and stretching of the wavelet, b is the translation parameter representing the translation of the wavelet function on the time axis, x(t) is the original multimodal mechanical wave propagation signal at time t, and ψ(t) is the wavelet basis function, representing the mother wavelet function used to decompose the signal. * (t) is the complex conjugate of the wavelet basis functions. Let b be the local feature of the signal at scale a, j be the imaginary unit, π be the mathematical constant π, ω0 be the center frequency of the wavelet, which determines the frequency characteristics of the wavelet, and dt represent the integration over time t.

[0126] In this embodiment, the instantaneous phase of the multimodal mechanical wave signal after wavelet transform processing is extracted, which can reflect the wave propagation characteristics in the material. The expression is as follows:

[0127] φ(t) = arg(W(a,b));

[0128] Where φ(t) is the instantaneous phase at time t, and arg(W(a,b)) is the phase angle of the wavelet transform coefficients;

[0129] It should be noted that during the propagation of multimodal mechanical waves, when the wave passes through defects inside the material (such as corrosion areas), a significant abrupt change in phase usually occurs.

[0130] In this embodiment, a phase change detection algorithm is used to accurately calculate the phase difference between adjacent time points and find the phase change point. The expression is as follows:

[0131] Δφ(t)=∣φ(t)-φ(t-Δt)∣;

[0132] In this embodiment, the phase change threshold Δφ is set based on the wave propagation characteristics and noise level of the material, using experimental data. th When the phase difference Δφ(t) between adjacent times is greater than Δφ th When this occurs, it is determined to be a phase abrupt change point.

[0133] S4: Based on the phase change characteristics of multimodal mechanical waves and combined with the mechanical wave propagation path information, locate the position of corrosion defects;

[0134] In this embodiment, based on the simulation results of the propagation path of mechanical waves using finite element analysis software, and according to the mechanical properties of the material and the wave modes, the elastic modulus E, Poisson's ratio ν, and density ρ of the material are obtained by conducting tensile and compression experiments on standard components of the material. The propagation speed of mechanical waves of different modes in the material is then determined, as expressed below:

[0135]

[0136] v r =0.87v p ;

[0137] Among them, v p Let v be the longitudinal wave velocity. s v is the transverse wave velocity. r Surface wave velocity;

[0138] In this embodiment of the application, each sensor calculates the time from excitation to detection of the mechanical wave by using the mechanical wave simulation propagation path length and wave speed;

[0139] In this embodiment of the application, a reference sensor is arranged at the mechanical wave excitation point, and the time when the reference sensor receives the mechanical wave is the reference time;

[0140] In this embodiment, the time when the mechanical wave is received by the other sensors is compared with the reference time to obtain the time difference between the reference sensor and the other sensors;

[0141] In the embodiments of this application, the phase abrupt change point usually corresponds to the location where the wave encounters a defect in the material. The larger the time difference, the more significant the influence on the wave during propagation.

[0142] In the embodiments of this application, when a significant time delay occurs in the wave signal received by a sensor, it usually indicates that the wave has encountered defects or discontinuities (such as corrosion areas) inside the material during propagation, resulting in a slowdown in the wave propagation speed or a longer path.

[0143] In this embodiment, based on the three-dimensional modeling structure of the material, the positioning coordinates (x, y, y) of each sensor in the sensor array are set. i ,y i ,z i );

[0144] In this embodiment, circles are drawn with the coordinates of each sensor as the center and the product of the time difference and the wave speed as the radius. The defect location is at the intersection of these circles, as shown in the following expression:

[0145]

[0146] Among them, (x f ,y f ,z f ) represents the coordinates of the defect location, v represents the propagation speed of the mechanical wave in the material, and ΔT represents the velocity. i Let i be the time difference between the i-th sensor and the reference sensor, where i is the index of the number of sensors.

[0147] In this embodiment, a nonlinear equation solving algorithm is used to minimize the error of multiple circles and obtain the location of corrosion defects.

[0148] In this embodiment, since the time difference and wave velocity measured by each sensor have certain errors, an objective function is constructed based on the least squares method to calculate the sum of squared errors S(x) of the sensors relative to the reference sensor. f ,y f ,z f The expression is as follows:

[0149]

[0150] Where N is the number of sensors;

[0151] In this embodiment, the gradient descent method is used. The position coordinates are iteratively updated along the gradient direction by calculating the partial derivative of the objective function with respect to the defect location until the sum of squared errors reaches its minimum.

[0152] S5: Based on the location of corrosion defects, combined with phase change, amplitude attenuation and propagation path offset information, different types of corrosion defects are identified and classified.

[0153] In this embodiment of the application, the characteristics of mechanical wave propagation in the material are extracted;

[0154] In the embodiments of this application, amplitude attenuation reflects the energy loss of the wave during propagation, especially when encountering defects, the amplitude of the wave will be significantly reduced;

[0155] In this embodiment of the application, the amplitude attenuation value ΔA of the mechanical wave is obtained by calculating the difference between the amplitude when the mechanical wave is excited and the amplitude when the sensor receives it.

[0156] In the embodiments of this application, defects can cause changes in the propagation path of waves, especially transverse waves and surface waves. The propagation path offset of waves is calculated by using the time difference collected by sensors.

[0157] In this embodiment of the application, the propagation path offset value ΔL and the phase difference Δφ(t) are obtained by calculating the product of the time difference and the wave speed when the mechanical wave arrives at the same position with and without defects.

[0158] In this embodiment of the application, the amplitude attenuation value, propagation path offset value and phase difference obtained by each sensor are used as a data sample;

[0159] In this embodiment of the application, by controlling the experimental conditions, samples with different types of corrosion defects are manufactured, and the types of corrosion defects and their corresponding features (ΔA, ΔL, Δφ(t)) are detected. The features are then labeled with corrosion defect categories (pitting corrosion, uniform corrosion, crack propagation, and no defects).

[0160] For example: No defects: ΔA<0.05, ΔL<1.5, Δφ(t)<0.1π;

[0161] Pitting corrosion: (0.05<ΔA<0.1,1.5<ΔL<4.5,0.1π<Δφ(t)<0.5π);

[0162] Uniform corrosion: (0.1≤ΔA<0.3, 1.5≤ΔL<15, 0.5π≤Δφ(t)<π);

[0163] Crack propagation: (0.3≤ΔA, 15≤ΔL, π≤Δφ(t));

[0164] In this embodiment, due to the nonlinearity of the defect features, radial basis functions are used to map the input features to a high-dimensional space to handle complex nonlinear classification problems. The decision function f(B) of the support vector machine model is calculated by comparing the similarity between the new sample B1 and the training sample B2, and is expressed as follows:

[0165]

[0166] Where `sign` is the sign function, which converts the output of the decision function into a class label; a positive result indicates a positive class, and a negative result indicates a negative class. `H` is the number of samples, `j` is the sample index, `α` is the Lagrange multiplier representing the weights of the support vectors, and `C`... j Let K(B1,B2) be the class label of the j-th training sample, K(B1,B2) be the radial basis function, and d be the bias term, which is the decision boundary offset of the support vector machine model.

[0167] K(B1,B2)=exp(-γ||B1-B2|| 2 );

[0168] Where γ is the width of the radial basis function, which determines the similarity between two samples, and ||B1-B2|| is the Euclidean distance between the new sample B1 and the training sample B2;

[0169] In this embodiment, the labeled feature data is used to train the support vector machine model, and the regularization parameter D and the width γ of the radial basis function in the support vector machine model are optimized.

[0170] In this embodiment, cross-validation is used to optimize the regularization parameter D and the width γ of the radial basis function to ensure that the model has good generalization ability;

[0171] In this embodiment, K-fold cross-validation is used to divide the dataset into K subsets. Common values ​​for K are 5 or 10. Each time, K-1 subsets are selected from the K subsets as the training set, and the remaining 1 subset is used as the validation set. The support vector machine model is trained and the performance of the validation set is evaluated. This process is repeated K times, so that each subset is used as a validation set exactly once.

[0172] In this embodiment of the application, after all K training and validation cycles are completed, the average accuracy of the support vector machine model on the validation set is calculated;

[0173] In this embodiment, a set of regularization parameters D and the width γ of the radial basis function are first defined. K-fold cross-validation is performed on each pair of regularization parameters D and the width γ of the radial basis function, and the average accuracy of the support vector machine model is recorded.

[0174] In this embodiment of the application, for all parameter combinations, the parameter combination with the highest cross-validation performance accuracy is selected as the final regularization parameter D and the width γ of the radial basis function;

[0175] In this embodiment of the application, after training is completed, the phase difference, amplitude attenuation and propagation path offset of the detected multimodal mechanical waves are used as comprehensive features to input the support vector machine model for prediction, and the category label of corrosion defects is output.

[0176] This embodiment also provides a corrosion defect identification system based on multimodal mechanical wave characteristics, including:

[0177] The mechanical wave excitation module is used to excite multimodal mechanical waves and plan the propagation path of the mechanical waves;

[0178] The signal acquisition module is used to arrange a sensor array according to the planned mechanical wave propagation path, acquire and synchronously process multimodal mechanical wave signals;

[0179] The feature extraction module is used to perform time-frequency analysis on the acquired multimodal mechanical wave signals and extract the phase change features of the multimodal mechanical waves;

[0180] The positioning module is used to locate the position of corrosion defects based on the phase change characteristics of multimodal mechanical waves and the information on the propagation path of mechanical waves.

[0181] The identification module is used to identify and classify different types of corrosion defects based on the location of the corrosion defect, combined with information on phase change, amplitude attenuation, and propagation path offset.

[0182] Furthermore, this also includes:

[0183] Memory, used to store programs;

[0184] A processor is used to load the program to execute the corrosion defect identification method based on multimodal mechanical wave characteristics.

[0185] This embodiment also provides a computer-readable storage medium storing a program that, when executed by a processor, implements the corrosion defect identification method based on multimodal mechanical wave characteristics.

[0186] The storage medium proposed in this embodiment belongs to the same inventive concept as the corrosion defect identification method based on multimodal mechanical wave characteristics proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0187] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0188] Example 2

[0189] Referring to Table 1, an embodiment of the present invention provides a corrosion defect identification method based on multimodal mechanical wave characteristics. To further verify the technical solution of the present invention, experimental simulation data of the corrosion defect identification method based on multimodal mechanical wave characteristics are given.

[0190] To verify the effectiveness of a corrosion defect identification method based on multimodal mechanical wave characteristics, this experiment compares the performance of this method with existing single-mode ultrasonic testing techniques (such as commonly used ultrasonic phased array testing techniques).

[0191] The experimental material used was an aluminum alloy sample with uneven thickness and complex geometry, measuring 500mm × 500mm × 10mm. Three types of corrosion defects were pre-designed on the surface: pitting corrosion, linear corrosion, and regional corrosion. Two detection methods were used for comparative analysis: one based on ultrasonic phased array technology, and the other based on the multimodal mechanical wave feature recognition method of this invention.

[0192] A three-dimensional model of the aluminum alloy sample was created using a laser scanner to construct its geometric property model. The thickness distribution of the sample was measured using an ultrasonic thickness gauge, and the thickness fluctuated between 9.8 mm and 10.2 mm.

[0193] Based on the geometric characteristics and thickness distribution, the wavelength range suitable for the material was obtained through finite element simulation. Multimode mechanical wave excitation was carried out by a combination of longitudinal waves, transverse waves and surface waves, with excitation frequencies of 5MHz (longitudinal wave), 8MHz (transverse wave) and 10MHz (surface wave). The wave mode with the highest reflection coefficient was selected through simulation, and the excitation angle was adjusted using a transducer to ensure that the mechanical wave could propagate throughout the material.

[0194] Twenty-five array-type ultrasonic sensors (spaced 20 mm apart) were uniformly arranged on the material surface, and the sensor density was increased at geometric changes. Each sensor synchronously acquired phase information, amplitude information, and propagation time of multimodal mechanical waves. During the acquisition process, a bandpass filter was used for directional filtering of the excitation wave frequency band, and the acquired signals were normalized.

[0195] Wavelet transform was used to perform time-frequency analysis on the acquired multimodal mechanical wave signals, the instantaneous phase was extracted, and the phase change point was identified by calculating the phase difference between adjacent moments. Based on the mechanical wave propagation path and phase change characteristics, combined with amplitude attenuation and propagation path offset information, the location and type of corrosion defects were accurately located.

[0196] The same material and defects were tested using ultrasonic phased array technology, and the test results were recorded. Phased array technology mainly identifies defects through single-mode ultrasonic signals and fails to effectively capture the multi-mode wave propagation characteristics in complex geometries, which affects the detection accuracy.

[0197] The details are shown in Table 1 below:

[0198] Table 1 Comparison of Experimental Data

[0199] parameter Multimodal mechanical wave detection method Ultrasonic phased array method Detection success rate (%) 98 85 Positioning error (mm) 0.5 2.2 Detection time (seconds) 120 180 Noise interference impact (dB) 5 15 Accurate defect identification rate (%) 96 82

[0200] The multimodal mechanical wave method achieves a detection success rate of up to 98%, while the ultrasonic phased array technology only achieves 85%. This indicates that the multimodal mechanical wave method can more comprehensively cover the material detection area, especially in complex geometries, and can more effectively identify different types of corrosion defects.

[0201] The positioning error of the multimodal mechanical wave method is less than 1 mm for all defect types, with the lowest being 0.4 mm. In contrast, the ultrasonic phased array technology has a larger error, especially in the detection of pitting defects, where the error is as high as 2.2 mm.

[0202] The detection time of multimodal mechanical wave methods is generally shorter than that of phased array technology. For example, when detecting linear corrosion defects, the multimodal method only takes 110 seconds, while the phased array technology takes 160 seconds.

[0203] Multimodal mechanical wave technology shows a significant advantage in terms of noise interference, with a noise impact of only 5dB, while phased array technology is subject to higher noise interference, at 15dB.

[0204] The multimodal mechanical wave method has a precise defect identification rate of 96%, which is much higher than the 82% of the ultrasonic phased array technology.

[0205] In summary, the corrosion defect identification method based on multimodal mechanical wave characteristics, by introducing innovative technologies such as multimodal signal detection, time-frequency analysis, and path simulation, significantly outperforms existing single-mode ultrasonic detection technologies in terms of detection success rate, positioning accuracy, detection efficiency, noise resistance, and defect identification rate.

[0206] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A corrosion defect identification method based on multimodal mechanical wave characteristics, characterized in that, include: Excite multimodal mechanical waves and plan the propagation path of the mechanical waves; Based on the planned mechanical wave propagation path, a sensor array is deployed to collect and synchronously process multimodal mechanical wave signals; Time-frequency analysis was performed on the acquired multimodal mechanical wave signals to extract the phase change characteristics of the multimodal mechanical waves; Based on the phase change characteristics of multimodal mechanical waves and combined with the propagation path information of mechanical waves, the location of corrosion defects can be determined. Based on the location of corrosion defects, combined with phase change, amplitude attenuation and propagation path offset information, different types of corrosion defects are identified and classified. The arrangement of the sensor array to acquire and synchronously process multimodal mechanical wave signals includes the following steps: A highly sensitive array of ultrasonic sensors is arranged on the surface of the material, with each sensor capturing signals of different mechanical wave modes; Based on the multimodal mechanical wave propagation path and boundary reflection characteristics obtained from finite element simulation, the sensor arrangement was adjusted. The sensor spacing is set based on the excitation wavelength of multimodal mechanical waves; Based on the material's geometric properties, sensors are uniformly arranged along the material surface, with increased density at geometric changes. For materials with uneven thickness and complex structures, sensors are placed on the sides and back of the material; Using standard samples with no defects and known defects, multimodal mechanical waves are excited, and response signals are collected through sensors. The response delay and sensitivity of each sensor are calibrated. High-speed acquisition of multimodal mechanical wave signals; Each sensor synchronously and in real time collects amplitude, phase and propagation time information of different modes of mechanical waves; During the acquisition process, a low-pass filter is used to remove high-frequency environmental noise, and a band-pass filter is used for directional filtering of the excitation wave frequency range. The amplitude, phase, and propagation time of the acquired multimodal mechanical waves are normalized. The step of performing time-frequency analysis on the acquired multimodal mechanical wave signals and extracting the phase change characteristics of the multimodal mechanical waves includes the following steps: The wavelet transform is used for time-domain and frequency-domain analysis, as shown in the following expression: ； ； in, The coefficients are those after wavelet transform. For scale parameters, For translation parameters, In time The original multimodal mechanical wave propagation signal, For wavelet basis functions, Let be the complex conjugate of the wavelet basis functions. For the signal at scale The local features below, the translation amount is , The imaginary unit, Pi The center frequency of the wavelet. Indicates time Integrate points; Extract the instantaneous phase of the multimodal mechanical wave signal after wavelet transform processing; Calculate the phase difference between adjacent time points and find the phase abrupt change point; Based on the wave propagation characteristics and noise level of the material, a threshold for phase change is set using experimental data. When the phase difference between adjacent times Greater than When this occurs, it is determined to be a phase abrupt change point; The method of locating corrosion defects based on the phase change characteristics of multimodal mechanical waves and combined with mechanical wave propagation path information includes the following steps: Based on the simulation results of the propagation path of mechanical waves using finite element analysis software, and according to the mechanical properties of the material and the modes of the waves, the propagation speed of mechanical waves of different modes in the material is determined through experimental measurement. Each sensor calculates the time from excitation to detection of the mechanical wave by simulating the propagation path length and wave speed using mechanical waves. A reference sensor is placed at the mechanical wave excitation point, and the time when the reference sensor receives the mechanical wave is the reference time. The time when the mechanical wave is received by the other sensors is compared with the reference time to obtain the time difference between the reference sensor and the other sensors; Based on the 3D modeling structure of the material, the positioning coordinates of each sensor in the sensor array are set. ; Draw circles with the coordinates of each sensor as the center and the product of the time difference and the wave speed as the radius. The defect location is at the intersection of these circles, as shown in the following expression: ； in, The coordinates of the defect location, The speed at which mechanical waves propagate in a material. For the first The time difference between the individual sensor and the reference sensor Index for the number of sensors; A nonlinear equation solving algorithm is used to minimize the error of multiple circles and obtain the location of corrosion defects; The method of identifying the type of corrosion defect by combining phase change, amplitude attenuation, and propagation path offset information includes the following steps: Extracting the characteristics of mechanical wave propagation in materials; The amplitude attenuation value of the mechanical wave is obtained by calculating the difference between the amplitude when the mechanical wave is excited and the amplitude when the sensor receives it. By calculating the product of the time difference and the wave speed when the mechanical wave arrives at the same position with and without defects, the propagation path offset and phase can be obtained. The amplitude attenuation value, propagation path offset value, and phase difference obtained from each sensor are used as a data sample; By controlling the experimental conditions, samples with different types of corrosion defects were manufactured, and the types of corrosion defects and their corresponding features were detected. The features were then labeled with corrosion defect categories. The radial basis function maps the input features to a high-dimensional space; the decision function of the support vector machine model... By calculating new samples With training samples The similarity between them is expressed as follows: ； in, For symbolic functions, For the sample size, Index for sample size For Lagrange multipliers, For the first The class labels of each training sample. For radial basis functions, For bias terms; ； in, The width of the radial basis functions. For new samples With training samples The Euclidean distance between them; The support vector machine model is trained using labeled feature data, and the regularization parameters in the support vector machine model are optimized. and the width of the radial basis functions ; Optimize regularization parameters using cross-validation. and the width of the radial basis functions ; After training, the phase difference, amplitude attenuation, and propagation path offset of the detected multimodal mechanical waves are used as comprehensive features to input into the support vector machine model for prediction, and the category label of corrosion defects is output.

2. The corrosion defect identification method based on multimodal mechanical wave characteristics as described in claim 1, characterized in that: The process of exciting multimodal mechanical waves and planning their propagation paths includes the following steps: Use an ultrasonic thickness gauge to determine the thickness distribution of the material; Using a laser scanner, a three-dimensional model of the material geometry is created to obtain the material's geometric properties; Based on the material thickness distribution, the wavelength range matching the material thickness is calculated. Multimode mechanical waves are formed by combining low-frequency longitudinal waves, high-frequency transverse waves, and surface waves. The reflection coefficients of different modal mechanical waves at the material boundary were calculated by finite element simulation, and the wave mode with the highest reflection value was selected. Multimodal mechanical waves are excited using a transducer, and the excitation angle is adjusted. Based on the material's geometry and thickness distribution, finite element analysis software is used to simulate the propagation path of mechanical waves within the material. The wave propagates from the excitation point, passes through reflection or refraction within the material, and reaches each sensor along its propagation path, with each path corresponding to a sensor location. Based on the simulation results, the excitation point was adjusted to fully cover the material detection area.

3. The corrosion defect identification method based on multimodal mechanical wave characteristics as described in claim 2, characterized in that: The method employs a nonlinear equation solving algorithm to minimize the error of multiple circles and obtain the location of corrosion defects, including the following steps: A target function is constructed based on the least squares method to calculate the sum of squared errors of the sensor relative to the reference sensor. The expression is as follows: ； in, Number of sensors; The gradient descent method is used to iteratively update the position coordinates along the gradient direction by calculating the partial derivative of the objective function with respect to the defect location.

4. A corrosion defect identification system based on multimodal mechanical wave characteristics, characterized in that, The method of claim 1, comprising: The mechanical wave excitation module is used to excite multimodal mechanical waves and plan the propagation path of the mechanical waves; The signal acquisition module is used to arrange a sensor array according to the planned mechanical wave propagation path, acquire and synchronously process multimodal mechanical wave signals; The feature extraction module is used to perform time-frequency analysis on the acquired multimodal mechanical wave signals and extract the phase change features of the multimodal mechanical waves; The positioning module is used to locate the position of corrosion defects based on the phase change characteristics of multimodal mechanical waves and the information on the propagation path of mechanical waves. The identification module is used to identify and classify different types of corrosion defects based on the location of the corrosion defect, combined with information on phase change, amplitude attenuation, and propagation path offset.

5. A computing device, characterized in that, include: Memory, used to store programs; A processor for loading the program to perform the steps of the corrosion defect identification method based on multimodal mechanical wave characteristics as described in any one of claims 1-3.

6. A computer-readable storage medium storing a program, characterized in that, When the program is executed by the processor, it implements the steps of the corrosion defect identification method based on multimodal mechanical wave characteristics as described in any one of claims 1-3.

Images