Shear wave ultrasonic backscattering micro crack detection method based on time-frequency variation threshold

Abstract

The invention discloses a shear wave ultrasonic backscattering micro-crack detection method based on a time-frequency variation threshold, comprising the following steps: S1, expressing the backscattering coefficient in a polycrystalline material according to the elastic constant covariance and the two-point correlation function; S2, deducing a shear wave time-frequency backscattering model within a set time interval according to the backscattering coefficient, scanning the processed sample and collecting reference signals and oblique incident backscattering signals, inputting various parameters required in the shear wave time-frequency backscattering model according to the material of the tested block, and obtaining a theoretical space variance surface and an experimental space variance surface; S3, calculating a backscattering maximum value surface and a confidence interval according to the obtained space variance surface, and deriving an upper bound as a time-frequency variation threshold; S4, calculating a signal space correlation coefficient, scanning a tested defective test block of the same material, processing the time-frequency signals of each point with the time-frequency variation threshold, and imaging the micro-cracks.

Description

Technical Field

[0001] The invention relates to the technical field of nondestructive testing, in particular to a shear wave ultrasonic backscattering micro crack detection method based on a time-frequency variation threshold. Background Art

[0002] When alloys are made into various parts through rolling, forging and other processes, various defects such as cracks, inclusions, and delamination may occur due to the influence of the microscopic characteristics of the material, the processing process and the service environment. If these defects are not discovered as early as possible, they will not only seriously affect the product quality, but may also cause safety hazards and economic losses. For pressure-bearing equipment and load-bearing structures in service, more than half of the cracks are caused by fatigue or aging. When there are tiny defects in the workpiece with a size smaller than a certain threshold, the actual fatigue limit of the alloy is reduced, which can easily cause catastrophic accidents under extremely harsh working conditions. Therefore, it is necessary to strictly control the tiny defects inside the workpiece, and effectively detecting its tiny internal defects is the premise and basis for its quality control.

[0003] However, my country's metal material detection capabilities still need to be improved, mainly due to the inability to effectively detect This also seriously affects the damage tolerance design capability of metal materials, violates the goal of lightweight design, and restricts its engineering process. Especially for polycrystalline materials, tiny defect signals are easily submerged by backscattered noise. When the defect signal is small and the grain noise is large, the signal-to-noise ratio is low and the defect missed detection rate will increase. Backscattered noise weakens the incident signal and defect signal, which will have an adverse effect on defect detection. The false detection rate of tiny defects in polycrystalline materials is high and the defect detection rate is low, so it needs to be solved urgently. Summary of the invention

[0004] In order to avoid and overcome the technical problems existing in the prior art, the present invention provides a shear wave ultrasonic backscattering micro crack detection method based on time-frequency variation threshold. The present invention greatly improves the detection rate of polycrystalline material defects and has a low false detection rate.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The method for detecting micro cracks by shear wave ultrasonic backscattering based on time-frequency variation threshold comprises the following steps:

[0007] S1, expressing the backscattering coefficient in polycrystalline materials based on the elastic constant covariance and the two-point correlation function;

[0008] S2. Introduce a shear wave time-frequency backscattering model within a set time interval based on the backscattering coefficient, scan the processed sample and collect reference signals and oblique incident backscattering signals, input various parameters required in the shear wave time-frequency backscattering model according to the material of the tested block, and obtain a theoretical spatial variance surface and an experimental spatial variance surface;

[0009] S3. Based on the obtained spatial variance surface, calculate the backscatter maximum value surface and confidence interval, and derive the upper bound as the time-frequency change threshold;

[0010] S4. Calculate the signal spatial correlation coefficient, scan the defective test blocks of the same material, process the time-frequency signals of each point with the time-frequency change threshold, and image the tiny cracks.

[0011] As a further solution of the present invention: the backscatter coefficient in step S1 can be expressed as:

[0012]

[0013] Among them, k T =2πf / c T is the wave number at frequency f;

[0014] <(δC 1313 ) 2 > is the elastic constant covariance;

[0015] ρ is the density of the polycrystalline material;

[0016] c T is the shear wave speed of sound in polycrystalline materials;

[0017] r(s) is the spatial correlation function, which represents the probability that two points with a distance |s| fall in the same scatterer, and integrates the displacement field Ω of the scatterer;

[0018] e is a unit vector, reflecting the beam direction.

[0019] As a further solution of the present invention: in step S2, a shear wave time-frequency backscattering model within a set time interval is derived based on the backscattering coefficient obtained in step S1 in combination with the independent scattering hypothesis; a reference signal is collected from a macro-defect-free sample that has undergone surface roughness treatment using an immersion ultrasonic scanning device, and various parameters required in the shear wave time-frequency backscattering model are input according to the material of the tested block to obtain a theoretical spatial variance surface;

[0020] The oblique incidence backscattering signal of the macro-defect-free sample with surface roughness treatment was collected by using an immersion ultrasonic scanning device. According to the material of the tested block, the various parameters required in the shear wave time-frequency backscattering model were input to obtain the experimental space variance surface.

[0021] Assuming that the polycrystalline material is a cubic crystal, the spatial correlation function in step S1 is r(s)=exp(-|s| / a), and the backscattering coefficient can be simplified in numerical calculation as follows:

[0022]

[0023] Where a is the grain size;

[0024] Then the theoretical spatial variance of shear waves within the set time interval can be expressed as:

[0025]

[0026] where Γ 2 (τ, f; a) is the theoretical spatial variance surface of shear waves when the time is τ and the frequency is f and the grain size is a;

[0027] D is the diffraction correction;

[0028] D R is the Rayleigh distance;

[0029] ρ f and ρ represent the density of water and the density of the sample, respectively;

[0030] c f and c T They represent the sound velocity of water and the shear wave speed of the sample respectively;

[0031] z f and z ref denote the hydroacoustic distances for the backscattering experiment and the reference test, respectively;

[0032] α(f) is the frequency domain attenuation coefficient in water, α T (f) is the frequency domain shear wave attenuation coefficient of the sample, which can be obtained by Weaver attenuation theory;

[0033] T f is the transmission coefficient of water to the sample;

[0034] R ff is the reflection coefficient;

[0035] is the frequency domain reference signal;

[0036] p(x,y,z,f) is the normalized sound field at position (x,y,z) and frequency f;

[0037] The range of the integration limit z is related to the Fourier transform window function w(t).

[0038] As a further solution of the present invention: collecting the input parameters required for the shear wave time-frequency backscattering model, the input parameters include the focal length F of the focusing probe, the radius b, the frequency range f, the density ρ of the material being measured L and the speed of sound c L , attenuation coefficient α f ; Liquid density ρ f and the longitudinal wave speed c f , water sound distance z f , the single crystal elastic constants c11, c12 and c33 of the object being measured, the pulse width σ of the incident wave;

[0039] Perform immersion ultrasonic testing on the test block without macro defects after surface roughness treatment, set the pulse generator parameters: voltage, gain, energy, pulse impedance, damping, high / low pass; focus the probe vertically on the upper surface of the sample to obtain the surface echo reference signal; experimental reference signal in the time domain It can be obtained by cutting off the zero points at both ends of the surface echo; after time-frequency processing of the surface echo reference signal, substitute it into the formula Γ 2 The theoretical space variance surface is obtained in (τ,f;a).

[0040] As a further solution of the present invention, shear waves are generated by oblique incident pulse reflection method, and a focusing probe is used to perform a plane scan on a polycrystalline test block without macroscopic defects to obtain backscattered waves, and a short-time Fourier transform is performed on the backscattered signals:

[0041]

[0042] Where w(t) is a window function with a width of Δt:

[0043]

[0044] For the obtained N frequency domain signals S i The variance surface of the experimental space can be obtained by calculating the variance. The variance surface of the experimental space is as follows:

[0045]

[0046] As a further solution of the present invention: in step S3, according to the distribution characteristics of the ultrasonic signals in the time domain and frequency domain, combined with the obtained spatial variance surface, the extreme value analysis theory is used to calculate the backscatter maximum value surface and the confidence interval, and the upper limit is derived as the time-frequency change threshold;

[0047] When N independent and identically distributed scattered signals V i When (t) obeys a normal distribution with zero mean, their corresponding frequency domain signals S i Obeying Rayleigh distribution, experimental maximum surface The frequency domain signal obtained by short-time Fourier transform is represented as:

[0048]

[0049] Maximum surface of backscattered signal in time-frequency domain for:

[0050]

[0051] Wherein the Euler–Mascheroni constant γ = 0.5772, Γ is the theoretical space standard deviation surface, which can be obtained through the theoretical space variance surface; The upper bound of can be expressed as:

[0052]

[0053] Where α is the confidence level.

[0054] As a further solution of the present invention: in step S4, before the defect imaging process, the defect-free part of the sample is analyzed to verify the maximum value surface of the backscattered signal in the time-frequency domain and upper bound U the The effectiveness of (τ,f).

[0055] As a further solution of the present invention: in step S4, the signal spatial correlation coefficient is calculated, and the tested defective test block of the same material is subjected to water immersion ultrasonic scanning, and the time-frequency signal of each point is processed with the time-frequency change threshold value to image the tiny cracks;

[0056] For the signal set collected in actual defect detection, each backscattered wave is expressed by the following formula:

[0057]

[0058] Perform time-frequency transformation and take the absolute value;

[0059] If the amplitude of the signal within the considered time-frequency range exceeds the time-frequency variation threshold, there is a defect at the point where the backscattered wave is located, and the time-frequency amplitude of the time-frequency point exceeding the time-variable threshold is recorded to image the defect;

[0060] If the amplitude of any echo within the considered time-frequency range does not exceed the time-frequency change threshold, there is no defect at that point and it is recorded as 0V;

[0061] The above steps complete the processing of all signal sets and complete the ultrasonic detection imaging.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] 1. The present invention improves the frequency domain scattering model under the multivariate Gaussian framework to obtain the time-frequency scattering model of shear waves, which can evaluate the grain size of the material, and then combine the extreme value analysis theory to obtain the time-frequency change threshold, providing a multidimensional analysis tool for material micro-crack detection. After establishing the time-frequency change threshold for a material, the defect analysis can be directly performed on the tested samples of the same material; due to its multi-dimensional analysis advantage, the time-frequency correlation threshold can effectively reduce the signal-to-noise ratio in actual defect detection, has a more stable defect imaging effect, greatly improves the detection rate of polycrystalline material defects, and has a low false detection rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a distribution law diagram of the time domain backscattering signal and the frequency domain backscattering signal in the present invention.

[0065] Figure 2 It is a schematic diagram of the maximum value surface and theoretical upper limit (confidence level 99.99%) of the crack defect sample in the present invention.

[0066] Figure 3 Schematic diagram of the defect surface and theoretical upper limit of a crack defect at a certain location in the present invention

[0067] Figure 4 Schematic diagram of amplitude imaging and time imaging of a crack defect sample at frequencies of 2.8 MHz to 7.2 MHz in the present invention.

[0068] Figure 5 It is a schematic diagram of amplitude imaging and time imaging of a crack defect sample at a frequency of 1.4 MHz to 4.0 MHz in the present invention.

[0069] Figure 6 It is a schematic diagram of amplitude imaging and time imaging of a crack defect sample at frequencies of 1.0 MHz to 9.0 MHz in the present invention.

[0070] Figure 7 This is a schematic diagram of the amplitude imaging and crack length measurement of the crack defect sample using a fixed threshold (0.3V) in the present invention.

[0071] Figure 8 The figure is a schematic diagram of the amplitude imaging and crack length determination of the crack defect sample using a time-varying threshold (confidence level 99.99%) in the present invention.

[0072] Fig. 9 The figure is a schematic diagram of the amplitude imaging and the crack length measured by using the time-frequency correlation threshold (confidence level 99.99%) for the crack defect sample in the present invention. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] See also Figures 1 to 9 In the embodiment of the present invention, the shear wave ultrasonic backscattering micro-crack detection method based on the time-frequency variation threshold is used. The defective sample and the reference sample used in the present invention are both 304 stainless steel plates that have been processed and surface roughened. The defective sample is a compact tensile sample designed according to GB / T 6398-2017 and fatigue cracks are prefabricated and extended. After the sample is polished, it is photographed with a microscope, and the crack length is measured to be about 9.72 mm.

[0075] The present invention requires that a test block be fixed in a water tank filled with sufficient water, an ultrasonic pulse generator / receiver be used to excite an ultrasonic focusing probe, the ultrasonic focusing probe be clamped on a motion platform through a probe frame, a motion control card installed on a computer is connected to a control circuit to control the movement of the motion platform, the position of the ultrasonic focusing probe in the water tank is adjusted according to the needs of collecting reference signals and backscattered signals, and a high-speed data acquisition card on the computer is used to acquire and store the original ultrasonic A-wave signal output by the ultrasonic instrument, and finally further analysis and modeling are performed on the computer.

[0076] The detection method of the present invention is as follows:

[0077] S1. The backscattering coefficient in polycrystalline materials is represented by the elastic constant covariance and the two-point correlation function. To obtain a theoretical estimate of the backscattering coefficient, it is necessary to consider the elastic constant covariance and the two-point correlation function of the grains. The characteristics of backscattering are reflected in the backscattering coefficient η(ω), which is defined as the differential scatterer per unit volume in the backscattering direction.

[0078] The statistical ultrasonic backscattering information in polycrystalline materials can be explained by mathematical random theory. This theory shows that the parameters of the effective medium can be obtained by the Voigt average of the elastic constants of the single crystal, and the backscattering signal of a single grain still has incoherent components after averaging.

[0079] In a scatterer in a homogeneous and isotropic polycrystalline body, the far-field scattering amplitude of a shear wave can be expressed as

[0080]

[0081] Among them B i is the component of the scattering amplitude B in the i direction;

[0082] ρ is the density of the main medium (i.e. the polycrystalline material);

[0083] c T is the shear wave speed of the main medium;

[0084] k T =2πf / c T is the wave number at frequency f;

[0085] δρ and δC are the changes of density and elastic stiffness tensors with the change of the main medium scatterer;

[0086] C jklm is the fourth-order elastic stiffness tensor;

[0087] ω=2πf is the angular frequency;

[0088] δ ij is the Kronecker symbol, which represents the dot product of any two orthogonal unit vectors;

[0089] e i It can be expressed as the i-direction component of the propagation direction unit vector e,

[0090] d i It can be expressed as the i-direction component of the polarization direction unit vector d.

[0091] Integrate the displacement field of the scatterer, where x is the vector at a certain point on the scatterer. Assume that the polarization direction of the incident wave is +1(x), the polarization direction of the incident wave is +2(y), and the propagation direction of the incident wave is +3(z). The subscripts 1, 2, and 3 in the following text represent the corresponding directions. Then, according to the relevant operation rules of tensors, we can get

[0092] e1=e2=d2=d3=0

[0093] e3=d1=1

[0094] Due to the assumption that the polycrystalline material is homogeneous and isotropic, the density variation δρ is zero, so

[0095]

[0096] B3=0

[0097] The differential shear backscattering cross section is given by |B| 2 Given, its ensemble average is equal to the backscattering coefficient, that is,

[0098]

[0099] Where Ω is the volume of the sample. By converting the variables r = (x + x') / 2, we can simplify the above formula to obtain

[0100]

[0101] Assuming that the elastic and spatial components of the covariance are independent, we can obtain

[0102] δC 1313 δC 1313 >(s)=<(δC 1313 ) 2 >r(s),

[0103] δC 2313 δC 2313 >(s)=<(δC 2313 ) 2 >r(s)

[0104] The symbol <> on the right side of the equal sign represents the arithmetic mean of the entire space, and r(s) is the two-point correlation function, which represents the probability that two points with a distance |s| apart fall in the same crystallite. Finally, we can get

[0105]

[0106] Among them, k T =2πf / c T is the wave number at frequency f;

[0107] <(δC 1313 ) 2 > is the elastic constant covariance;

[0108] ρ is the density of the polycrystalline material;

[0109] c T is the shear wave speed of sound in polycrystalline materials;

[0110] r(s) is the spatial correlation function, which represents the probability that two points with a distance |s| fall in the same scatterer, and integrates the displacement field Ω of the scatterer;

[0111] e is a unit vector, reflecting the beam direction.

[0112] The backscattering contribution of shear waves comes from two terms: the first term is due to the backscattered shear waves polarized in the x-direction, which is called the vertical shear shear wave; the second term is due to the backscattered shear waves polarized in the y-direction, which is called the horizontal shear shear wave.

[0113] In the water immersion pulse echo setup of the present invention, the backscattered shear wave polarized along the y direction cannot propagate through the solid-liquid interface and return to the ultrasonic probe. Therefore, the measured shear wave backscatter coefficient is only the first term in the above formula.

[0114] S2. Based on the backscattering coefficient obtained in step S1 and the independent scattering hypothesis, a shear wave time-frequency backscattering model within a set time interval is derived; a reference signal is collected from a macro-defect-free sample that has undergone surface roughness treatment using an immersion ultrasonic scanning device, and various parameters required in the shear wave time-frequency backscattering model are input according to the material of the tested block to obtain a theoretical spatial variance surface;

[0115] The oblique incidence backscattering signal of the macro-defect-free sample with surface roughness treatment was collected by using an immersion ultrasonic scanning device. According to the material of the tested block, the various parameters required in the shear wave time-frequency backscattering model were input to obtain the experimental space variance surface.

[0116] For equiaxed crystals, the spatial correlation function is r(s) = exp(-|s| / a), which can be simplified to

[0117]

[0118] The theoretical spatial variance surface of shear waves within a set time interval can be expressed as

[0119]

[0120] where Γ 2 (τ, f; a) is the theoretical spatial variance surface of shear waves when the time is τ and the frequency is f and the grain size is a;

[0121] D is the diffraction correction;

[0122] D R is the Rayleigh distance;

[0123] ρ f and ρ represent the density of water and the density of the sample, respectively;

[0124] c f and c T They represent the sound velocity of water and the shear wave speed of the sample respectively;

[0125] z f and z ref denote the hydroacoustic distances for the backscattering experiment and the reference test, respectively;

[0126] α(f) is the frequency domain attenuation coefficient in water, α T (f) is the frequency domain shear wave attenuation coefficient of the sample, which can be obtained by Weaver attenuation theory;

[0127] T f is the transmission coefficient of water to the sample;

[0128] R ff is the reflection coefficient;

[0129] is the frequency domain reference signal;

[0130] p(x,y,z,f) is the normalized sound field at position (x,y,z) and frequency f;

[0131] The range of the integration limit z is related to the Fourier transform window function w(t).

[0132] In order to calculate the shear wave time-frequency backscattering model, it is necessary to collect the various input parameters required by the model, including: the focal length F of the focusing probe, the radius b, the frequency range f, the density ρ of the material being measured L and the speed of sound c L , attenuation coefficient α f ; Liquid density ρ f and the longitudinal wave speed c f , water sound distance z f , the single crystal elastic constants c11, c12 and c33 of the object being measured, and the pulse width σ of the incident wave.

[0133] Perform water immersion ultrasonic testing on the test block without macro defects after surface roughness treatment, and set the pulse generator parameters including voltage, gain, energy, pulse impedance, damping, high / low pass, etc. Focus the probe vertically on the upper surface of the sample to obtain the surface echo reference signal. Experimental reference signal in the time domain It is obtained by cutting off the zero points at both ends of the surface echo. After time-frequency processing of the surface echo reference signal, substitute it into the formula Γ 2 (τ,f; a), we obtain the theoretical space variance surface.

[0134] Shear waves are generated by the oblique incident pulse reflection method, and a focused probe is used to perform a plane scan on a polycrystalline test block without macroscopic defects to obtain a series of time domain A-type waves. These waveforms are the part between a surface echo and a bottom echo, namely, backscattered waves.

[0135] The backscattering model of the present invention requires that the backscattering signal obeys a standard Gaussian distribution, so the experiment needs to be set up as follows:

[0136] (1) The microstructure inside the sample is uniform, including random grain orientation, independent grains and no coupling structure;

[0137] (2) The sample has no defects in the macroscopic view, and the upper surface of the sample is sufficiently smooth and parallel to the scanning plane;

[0138] (3) The signal collection interval is large enough to ensure that adjacent backscattered signals are uncorrelated;

[0139] (4) The influence of the ultrasonic system (such as electromagnetic interference, etc.) can be ignored. The backscattered signals that meet these settings are independent and identically distributed, and the short-time Fourier transform of these backscattered signals is performed.

[0140]

[0141] In the formula, the i in normal font is an imaginary unit, and the i in italics represents the i-th signal.

[0142] Where w(t) is the window function with a width of Δt

[0143]

[0144] Although any window function can truncate the signal, using only a rectangular window will result in severe signal leakage.

[0145] In order to minimize the signal leakage caused by window function truncation, the present invention uses a Hanning window and adopts an energy correction factor Ensure that the energy is not affected by the window function. After Fourier transform, the energy correction factor should make the energy of the window function equal to that of a rectangular window of the same length. It can be calculated that the energy correction factors of the rectangular window and Hanning window are 1 and 1.63 respectively.

[0146] For the obtained N frequency domain signals S i The variance surface of the experimental space can be obtained by calculating the variance. The variance surface of the experimental space is as follows:

[0147]

[0148] The time window slides between the surface echo and the bottom echo, and the frequency domain variance curve of each moment τ can be obtained. In this way, the experimental space variance surface can be obtained.

[0149] When performing immersion ultrasonic testing on a test block with surface roughness treatment and no macro defects, it should be noted that the experimental configuration and related parameters should be kept the same as those in the reference signal acquisition. The test block is placed horizontally, the probe is incident obliquely, the inclination angle is between the first critical angle and the second critical angle, and the focus is on the center of the sample. The water sound distance is z f The pulse generator parameters should remain unchanged, and a larger gain should be set to ensure that the collected backscattered signal is more obvious.

[0150] In order to calculate the time-frequency change threshold, it is necessary to collect the various input parameters required by the model, including: the focal length F of the focusing probe, the radius b, the frequency range ω, the density ρ of the material being tested L and the speed of sound c L , attenuation coefficient α f ; Liquid density ρ f and the longitudinal wave speed c f , water sound distance z f , the single crystal elastic constants c11, c12 and c33 of the object being measured, and the pulse width σ of the incident wave.

[0151] For a 304 stainless steel sample without macroscopic defects, its single crystal elastic constant is c 11 =204.6GPa,c 12 =137.7GPa,c 44 =126.2GPa, the measured density is ρ=7930kg / m 3 Using a 10MHz Olympus circular focusing probe, the longitudinal wave velocity was measured by pulse reflection method as c L =5940m / s, which is slightly different from the calculated Voigt average longitudinal wave speed of 5959m / s. A spherical focusing probe with a center frequency of 5MHz was used in the experiment (nominal focal length 50.8mm, nominal chip diameter 12.7mm). The probe was focused vertically on the upper surface of the sample to obtain the surface echo reference signal. Then the backscattered signal was collected, and a larger gain was required to ensure that the collected backscattered signal was more obvious. The probe was tilted and focused on the depth center of the sample, and the probe was moved horizontally to collect backscattered signals at different test block positions. The vertical distance between the probe and the test block must remain unchanged. In order to obtain incoherent signals, a sufficiently large collection step (generally greater than 2 times the wavelength) is required to obtain a certain number of backscattered signal sets V i (t) The backscatter signal set is processed in step S22 to obtain the experimental space variance surface.

[0152] When collecting backscattered signals from the 304 stainless steel test block, the probe focusing depth is m p =10mm. In order to obtain incoherent signals, ultrasonic scanning was performed in a rectangular area of ​​about 25mm×25mm with a step size of 0.5mm. About 2500 backscattered A-wave signals were collected. The specific settings for collecting backscattered signals are shown in the table below.

[0153]

[0154] Theoretical space variance surface and experimental space variance surface The least squares fitting method was used to obtain the theoretical grain size of the tested sample, a=22.6±0.77 μm.

[0155] S3. According to the distribution characteristics of ultrasonic signals in the time domain and frequency domain, combined with the obtained spatial variance surface, the extreme value analysis theory is used to calculate the backscatter maximum surface and confidence interval, and the upper bound is derived as the time-frequency change threshold:

[0156] When N independent and identically distributed scattered signals V i When (t) obeys a normal distribution with zero mean, their corresponding frequency domain signals S i Obeying Rayleigh distribution, experimental maximum surface The frequency domain signal obtained by short-time Fourier transform is represented as

[0157]

[0158] Next, we build the theoretical maximum surface The probability density function of the Rayleigh distribution is

[0159]

[0160] The cumulative distribution function F of the Rayleigh distribution belongs to the maximum attraction domain of the Gumbel extreme value distribution Λ(S) = exp[-exp(-S)]. Next, we need to find the normalization constant a N and b N The values ​​of are, like S, time-frequency dependent. N→∞ Pr{A max ≤a N S+b N}=Λ(S), we can assume and Then the cumulative distribution function of the Rayleigh distribution is

[0161]

[0162] Then, the theoretical maximum surface can be given according to the properties of Gumbel distribution The mathematical expectation and confidence interval of the backscatter signal in the time-frequency domain is for

[0163]

[0164] Wherein the Euler–Mascheroni constant γ = 0.5772, Γ is the theoretical space standard deviation surface, which can be obtained through the theoretical space variance surface. The upper bound of can be expressed as

[0165]

[0166] Where α is the confidence level.

[0167] Upper bound U the (τ,f) can be used as a time-frequency related threshold to screen outliers caused by defects. Since it changes with time and frequency, it can be called a time-frequency variation threshold. It can be seen that the upper bound increases as the confidence level increases, and the theoretical maximum value surface and the upper bound differ in shape by only one coefficient.

[0168] S4, calculate the signal spatial correlation coefficient, perform water immersion ultrasonic scanning on the defective test block of the same material, process the time-frequency signal of each point with the time-frequency change threshold, and image the tiny cracks;

[0169] Before defect imaging, the defect-free part of the sample is analyzed to verify the maximum surface of the backscattered signal in the time-frequency domain and upper bound U the For the backscatter signal set collected in the previous step, the corresponding spatial correlation coefficient is calculated under the considered time window and frequency to verify whether the independent and identically distributed conditions are met.

[0170] It should be noted that due to the definition of the spatial correlation coefficient, another independent test is required for these signals to ensure that they conform to a zero-mean Gaussian distribution in the time domain and to a Rayleigh distribution in the time-frequency domain after time-frequency processing.

[0171] like Figure 1 The left side shows the time domain backscattered signal set collected inside the 304 stainless steel material using ultrasound with an incident angle of 20°. The dotted line represents the signal amplitude at the corresponding moment selected randomly. The amplitudes at these moments are sorted and fitted with a normal distribution. The goodness of fit R 2 is 0.997, indicating that they follow a normal distribution.

[0172] like Figure 1 The right side shows the frequency domain signal obtained by fast Fourier transforming the corresponding time domain signal using a rectangular window with a length of the full length. The dotted line represents the signal amplitude of the corresponding frequency point randomly selected (generally, the frequency point near the center frequency is selected). The amplitudes of these frequency points are sorted and fitted with Rayleigh distribution. The goodness of fit R 2 is 0.995, indicating that they obey the Rayleigh distribution very well. Therefore, it can be considered that when the independent and identically distributed scattered signals V i When (t) obeys a normal distribution with zero mean, their corresponding frequency domain signals S i It obeys the Rayleigh distribution.

[0173] After obtaining the theoretical maximum surface and theoretical upper limit, the grain size obtained by fitting is input into the calculation of the theoretical maximum surface. The experimental maximum surface is obtained by taking the maximum value of each time-frequency point after time-frequency processing of each A wave. The result is as follows: Figure 2 shown.

[0174] It can be seen that the fitting effect of the theoretical maximum value surface and the experimental maximum value surface is better. In the time range of 35.0 to 41.4 μs and the frequency range of 2 to 8 MHz, the upper bound is higher than the experimental maximum value surface, so the upper bound is available in this range.

[0175] For the signal set collected in actual defect detection, each backscattered wave is expressed by the following formula:

[0176]

[0177] Perform time-frequency transformation and take the absolute value;

[0178] If the amplitude of this signal within the considered time-frequency range exceeds the time-frequency variation threshold, it can be considered that there is a defect at the point where the A wave is located, and the time-frequency amplitude of the time-frequency point that exceeds the time-varying threshold is recorded to image the defect; conversely, if the amplitude of any echo within the considered time-frequency range does not exceed the time-frequency variation threshold, it is considered that there is no defect at the point and it is recorded as 0V; this step is repeated until the processing of all signal sets is completed and the ultrasonic detection imaging is completed.

[0179] The same system parameter configuration as that used in the backscatter signal set acquisition is used to collect the ultrasonic signal of the compact tensile specimen. First, the A wave near the crack tip of the original C-scan image is analyzed and processed by time-frequency transformation to obtain the corresponding spatial standard deviation surface, as shown in Figure 3 As shown in the figure, it is a time domain slice and a frequency domain slice. The intersection of the slices is the peak of the defect wave packet, that is, the frequency is 2.8MHz and the time is 41.1μs. By comparing with the theoretical upper limit, it is higher than the upper limit at about 40.5μs, so the possibility that this place is a defect is 99.99%. The defect surface is located near 3MHz in frequency, which is significantly lower than the probe center frequency of 5MHz. It is not easy to detect using the traditional threshold.

[0180] Next, we consider the influence of the selected time-frequency range on the defect detection imaging results. Figure 3 It can be found that the frequency range of the defect surface is significantly lower than the center frequency. In the same time window, different amplitude imaging can be obtained by using different frequency windows.

[0181] Figure 4 The frequency window is 2.8MHz to 7.2MHz ( Figure 3 upper shaded area in the lower right);

[0182] Figure 5 The frequency window is 1.4MHz to 4MHz ( Figure 3 the middle shaded area in the lower right);

[0183] Figure 6 The frequency window is 1MHz to 9MHz ( Figure 3 lower shaded area on the lower right).

[0184] It can be seen that the imaging effect is best when the frequency window is located near the defect frequency, the imaging effect is second when the frequency window is located near the center frequency, and the imaging effect is worst when the frequency window is 1MHz to 9MHz. Figure 4 and Figure 6 It was found that when the frequency window width increased to a certain extent, the imaging effect would be severely reduced and a large number of noise points would appear.

[0185] Therefore, when using the time-frequency related threshold, the value range of the frequency window must be determined according to the defect wave frequency range, and better imaging effects can be obtained at this time. When the defect wave position is unknown, the value of the frequency window can also be determined by the center frequency of the probe, but it should be noted that the frequency window should not be too wide, especially not too close to the low frequency.

[0186] Compare the imaging results obtained using different types of thresholds, and calculate the crack length through imaging. Figure 7 , Figure 8 , Fig. 9 As shown in the figure, the left side is the amplitude imaging screened by using a fixed threshold, a time-varying threshold, and a time-frequency related threshold. It can be seen that the false detection areas of the three methods decrease in turn, and the imaging result is the best when the time-frequency related threshold is used, and the false detection area is greatly reduced.

[0187] Figure 7 , Figure 8 , Fig. 9 On the right are the amplitude diagrams of the corresponding C-scan images near 34.5 mm in the longitudinal direction (this is the approximate position of the crack direction), the horizontal line is the half-height width of the peak, and multiple amplitudes are taken near 34.5 mm and the average is taken, and the width is used as the crack length measured by the corresponding threshold.

[0188] Figure 7 A fixed threshold of 0.3 V was used and the crack length was measured to be 9.32 mm, with a relative error of 4.12% compared with the value measured by microphotography.

[0189] Figure 8 Using a time-varying threshold, the crack length was determined to be 9.56 mm with a relative error of 1.65%. Fig. 9 In the method, the time-frequency correlation threshold was used, and the crack length was measured to be 9.62 mm with a relative error of 1.03%. It can be seen that compared with the fixed threshold and the time-varying threshold, the crack length measured by the time-frequency correlation threshold method has the smallest difference with the microscopic photography measurement result, and the shape is closer to the actual situation of the crack.

[0190] It can be seen that the time-frequency variation threshold is generally better than the time-varying threshold and the fixed threshold for fatigue crack evaluation. For closed cracks with a tip width less than 2 μm, the method proposed in the present invention can effectively detect them without missed detection. However, when using the time-frequency correlation threshold, the value range of the frequency window has a great influence on the imaging results, so special attention should be paid during analysis.

[0191] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0192] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0193] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0194] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0195] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A shear wave ultrasonic backscattering micro-crack detection method based on time-frequency variation threshold, characterized in that: The steps include: S1, expressing the backscattering coefficient in polycrystalline materials based on the elastic constant covariance and the two-point correlation function; S2. Introduce a shear wave time-frequency backscattering model within a set time interval based on the backscattering coefficient, scan the processed sample and collect reference signals and oblique incident backscattering signals, input various parameters required in the shear wave time-frequency backscattering model according to the material of the tested block, and obtain a theoretical spatial variance surface and an experimental spatial variance surface; S3. Based on the obtained spatial variance surface, calculate the backscatter maximum value surface and confidence interval, and derive the upper bound as the time-frequency change threshold; S4. Calculate the signal spatial correlation coefficient, scan the defective test block of the same material, process the time-frequency signal of each point with the time-frequency change threshold, and image the tiny cracks; The backscatter coefficient in step S1 can be expressed as: Among them, k T =2πf / c T is the wave number at frequency f; <(δC 1313 ) 2 > is the elastic constant covariance; ρ is the density of the polycrystalline material; c T is the shear wave speed of sound in polycrystalline materials; r(s) is the spatial correlation function, which represents the probability that two points with a distance |s| fall in the same scatterer, and integrates the displacement field Ω of the scatterer; e is a unit vector, reflecting the beam direction; In step S2, a shear wave time-frequency backscattering model within a set time interval is derived based on the backscattering coefficient obtained in step S1 and the independent scattering hypothesis; a reference signal is collected from a macro-defect-free sample that has undergone surface roughness treatment using an immersion ultrasonic scanning device, and various parameters required in the shear wave time-frequency backscattering model are input according to the material of the tested block to obtain a theoretical spatial variance surface; The oblique incidence backscattering signal of the macro-defect-free sample with surface roughness treatment was collected by using an immersion ultrasonic scanning device. According to the material of the tested block, the various parameters required in the shear wave time-frequency backscattering model were input to obtain the experimental space variance surface. Assuming that the polycrystalline material is a cubic crystal, the spatial correlation function in step S1 is r(s)=exp(-|s| / a), and the backscattering coefficient can be simplified in numerical calculation as follows: Where a is the grain size; Then the theoretical spatial variance of shear waves within the set time interval can be expressed as: where Γ 2 (τ, f; a) is the theoretical spatial variance surface of shear waves when the time is τ and the frequency is f and the grain size is a; D is the diffraction correction; D R is the Rayleigh distance; ρ f and ρ represent the density of water and the density of the sample, respectively; c f and c T They represent the sound velocity of water and the shear wave speed of the sample respectively; z f and z ref denote the hydroacoustic distances for the backscattering experiment and the reference test, respectively; α(f) is the frequency domain attenuation coefficient in water, α T (f) is the frequency domain shear wave attenuation coefficient of the sample, which can be obtained by Weaver attenuation theory; T f is the transmission coefficient of water to the sample; R ff is the reflection coefficient; is the frequency domain reference signal; p(x,y,z,f) is the normalized sound field at position (x,y,z) and frequency f; The range of the integration limit z is related to the Fourier transform window function w(t); The shear wave is generated by the oblique incident pulse reflection method, and the polycrystalline test block without macro defects is scanned in the plane using a focusing probe to obtain the backscattered wave, and the backscattered signal is short-time Fourier transformed: Where w(t) is a window function with a width of Δt: For the obtained N frequency domain signals S i The variance surface of the experimental space can be obtained by calculating the variance. The variance surface of the experimental space is as follows: In step S3, according to the distribution characteristics of the ultrasonic signals in the time domain and frequency domain, combined with the obtained spatial variance surface, the extreme value analysis theory is used to calculate the backscatter maximum value surface and the confidence interval, and the upper bound is derived as the time-frequency change threshold; When N independent and identically distributed scattered signals V i When (t) obeys a normal distribution with zero mean, their corresponding frequency domain signals S i Obeying Rayleigh distribution, experimental maximum surface The frequency domain signal obtained by short-time Fourier transform is: Maximum surface of backscattered signal in time-frequency domain for: Wherein the Euler–Mascheroni constant γ = 0.5772, Γ is the theoretical space standard deviation surface, which can be obtained through the theoretical space variance surface; The upper bound of can be expressed as: Where α is the confidence level; In step S4, the signal spatial correlation coefficient is calculated, and a water immersion ultrasonic scanning is performed on the defective test block of the same material, and the time-frequency signal of each point is processed with the time-frequency change threshold to image the tiny cracks; For the signal set collected in actual defect detection, each backscattered wave is expressed by the following formula: Perform time-frequency transformation and take the absolute value; If the amplitude of the signal within the considered time-frequency range exceeds the time-frequency variation threshold, there is a defect at the point where the backscattered wave is located, and the time-frequency amplitude of the time-frequency point exceeding the time-variable threshold is recorded to image the defect; If the amplitude of any echo within the considered time-frequency range does not exceed the time-frequency change threshold, there is no defect at that point and it is recorded as 0V; The above steps complete the processing of all signal sets and complete the ultrasonic detection imaging.

2. The shear wave ultrasonic backscattering micro-crack detection method based on time-frequency variation threshold according to claim 1 is characterized in that: The input parameters required for collecting the shear wave time-frequency backscattering model include the focal length F of the focusing probe, the radius b, the frequency range f, and the density ρ of the material being measured. L and the speed of sound c L , attenuation coefficient α f ; Liquid density ρ f and the longitudinal wave speed c f , water sound distance z f , the single crystal elastic constants c11, c12 and c33 of the object being measured, the pulse width σ of the incident wave; Perform immersion ultrasonic testing on the test block without macro defects after surface roughness treatment, set the pulse generator parameters: voltage, gain, energy, pulse impedance, damping, high / low pass; focus the probe vertically on the upper surface of the sample to obtain the surface echo reference signal; experimental reference signal in the time domain It can be obtained by cutting off the zero points at both ends of the surface echo; after time-frequency processing of the surface echo reference signal, substitute it into the formula Γ 2 The theoretical space variance surface is obtained in (τ,f;a).

3. The shear wave ultrasonic backscattering micro crack detection method based on time-frequency variation threshold according to claim 1 is characterized in that: In step S4, before defect imaging, the defect-free part of the sample is analyzed to verify the maximum surface of the backscattered signal in the time-frequency domain and upper bound U the The effectiveness of (τ,f).

Images