A large shaft forging blank defect identification method based on ultrasonic flaw detection data

Abstract

The present application relates to the technical field of embryo material defect detection, and particularly relates to a large shaft forging embryo material defect identification method based on ultrasonic flaw detection data, which comprises the following steps: obtaining an ultrasonic A-scan signal of a large shaft forging embryo material; extracting a backscattering signal from the ultrasonic A-scan signal; preprocessing the backscattering signal; calculating the power spectral density of the backscattering signal; obtaining a characteristic parameter related to depth based on the power spectral density; calculating a dynamic attenuation coefficient changing with depth based on the characteristic parameter; obtaining a compensation factor through exponential operation; compensating the original echo peak amplitude of a detected defect in the A-scan signal by using the compensation factor to obtain a defect signal amplitude; comparing the defect signal amplitude with a threshold of a standard DAC curve at a corresponding depth to determine whether a defect exists. The present application improves the accuracy and reliability of defect identification.

Description

Technical Field

[0001] This invention relates to the field of billet defect detection technology, and in particular to a method for identifying defects in large shaft forging billets based on ultrasonic flaw detection data. Background Technology

[0002] Large shaft forgings, such as generator rotors and marine main shafts, are core components of the nation's heavy industry, and their internal quality directly affects the service safety and lifespan of the entire equipment. Ultrasonic testing is a crucial process for ensuring the quality of these forgings, identifying potential internal metallurgical defects by analyzing ultrasonic echo signals. In practical testing, the commonly used method is distance amplitude correction (DAC) technology. This technology pre-measures the echo signal intensity of artificial defects at different depths on a standard test block, plotting a reference curve to compensate for the natural weakening of the ultrasonic signal as it propagates through the material. This allows defects at different depths to be compared and evaluated under the same standard.

[0003] Chinese patent application CN114755304A discloses an ultrasonic flaw detection process for the hardened layer of cold-rolled forged steel work rolls. The method is as follows: a reference test block is made using a series of 10mm deep blind holes as artificial reflectors. A shear wave angle probe wedge is ground to a shape that matches the curved surface of the work roll. This series of shear wave angle probes is used for circumferential detection, aligning the main acoustic beam of the probe with the series of 10mm blind holes under the hardened layer of the reference test block, and a distance-amplitude correction curve is generated. The series of shear wave angle probes are mounted on the measuring slide of a roll grinding machine, and the cold-rolled forged steel work roll process and the ultrasonic flaw detection process are performed simultaneously. In the cold-rolled forged steel work roll process, the cold-rolled forged steel work roll rotates and moves at a speed not exceeding 150mm / s. The flaw detection time base scanning speed is adjusted at a horizontal ratio of 1:2. The ultrasonic probe performs circumferential detection on the surface of the hardened layer of the cold-rolled forged steel work roll to quickly determine the defect echo.

[0004] However, existing DAC technology has significant limitations when applied to large shaft forging blanks. The core problem lies in the use of a uniform attenuation model pre-set before detection, assuming a uniform microstructure of the material. In reality, large forgings exhibit differences in grain size and microstructure due to uneven cooling and stress, leading to microstructure inhomogeneity. Using a fixed DAC curve results in insufficient compensation in areas of strong attenuation, causing excessive attenuation of defect signals and missed detections; conversely, overcompensation in areas of weak attenuation amplifies noise or background signals, leading to misjudgments and reducing the accuracy and reliability of defect detection. Summary of the Invention

[0005] To address the failure of traditional DAC correction in large shaft forging blanks due to uneven internal structure, this invention provides a defect identification method for large shaft forging blanks based on ultrasonic flaw detection data.

[0006] This invention provides a method for identifying defects in large shaft forging blanks based on ultrasonic flaw detection data, employing the following technical solution:

[0007] A method for defect identification of large shaft forging blanks based on ultrasonic flaw detection data includes the following steps: ultrasonic testing of the large shaft forging blank using a probe to obtain the ultrasonic A-scan signal; extracting a backscattered signal from the ultrasonic A-scan signal within a stable time window after the interface echo and before the expected defect area; preprocessing the backscattered signal; calculating the power spectral density of the backscattered signal; obtaining depth-related characteristic parameters based on the power spectral density, wherein the characteristic parameters are positively correlated with the difference between the probe center frequency and the centroid frequency of the power spectral density; calculating a dynamic attenuation coefficient varying with depth based on the cumulative deviation of the characteristic parameters along the ultrasonic propagation path and a set reference attenuation coefficient; accumulating the difference between the dynamic attenuation coefficient and the reference attenuation coefficient along the round-trip path and obtaining a compensation factor through exponential calculation; compensating the original echo peak amplitude of the defect detected in the A-scan signal using the compensation factor to obtain the defect signal amplitude; comparing the defect signal amplitude with the threshold of the standard DAC curve at the corresponding depth to determine whether a defect exists.

[0008] By analyzing the changes in the acoustic properties of materials along depth using backscattered signals, a dynamic attenuation coefficient and compensation factor that vary with depth were constructed to dynamically and accurately compensate and correct the original echo of detected defects. This ensures that the amplitude of the compensated defect signal accurately reflects the actual size of the defect, unaffected by local differences in material microstructure. This avoids missed detections due to insufficient compensation in areas of strong attenuation and misjudgments due to overcompensation in areas of weak attenuation, thus improving the accuracy and reliability of ultrasonic testing of large shaft forgings.

[0009] The preferred method for calculating the characteristic parameters is as follows:

[0010] ;

[0011] In the formula, Indicates the depth inside the blank. Characteristic parameters at the location; It is a frequency variable; In internal depth The power spectral density of the backscattered signal at that location; It is the center frequency of the ultrasonic probe; It is the fourth central moment of the power spectral density; It is the second central moment of the power spectral density; It is a reference value for the kurtosis of the power spectrum distribution.

[0012] By combining the centroid frequency shift term, which reflects high-frequency attenuation, with the power spectrum kurtosis enhancement term, which reflects the degree of microstructural disorder, the characteristic parameters can not only characterize the overall attenuation properties of the material, but also more sensitively capture local anomalies such as abnormal aggregation of inclusions or structural distortion.

[0013] Preferably, the fourth central moment of the power spectral density is calculated using the fourth moment in higher-order statistics.

[0014] Using higher-order statistics for calculations ensures greater accuracy and robustness in quantifying the non-Gaussianity of signals (i.e., kurtosis calculation).

[0015] The preferred method for calculating the dynamic attenuation coefficient is as follows:

[0016] ;

[0017] In the formula, It is at a depth inside the blank. The dynamic attenuation coefficient at the location; It is the reference attenuation coefficient under ideal uniform conditions; Indicates the depth of the blank Characteristic parameters at the location; These are characteristic parameters The baseline value under uniform conditions; It is the characteristic length of exponential decay; It is the base of the natural logarithm; It is the cumulative integral process from surface 0 to the current depth d.

[0018] By employing an exponentially weighted integral method, the contribution of tissue regions closer to the current depth to the attenuation coefficient is increased. This better reflects the physical reality of sound wave propagation, where the tissue state along the nearest path has a greater impact. Consequently, the calculated dynamic attenuation coefficient more accurately reflects the true cumulative attenuation effect experienced by ultrasound waves at any depth.

[0019] Preferred, compensation factor The calculation method is as follows:

[0020] ;

[0021] In the formula, In internal depth Compensation factor at the location; This is the current depth; In internal depth The dynamic attenuation coefficient at the location; It is the baseline attenuation coefficient of the material under ideal homogeneous conditions; It is an exponential function with the natural constant as its base.

[0022] By accumulating the difference between the dynamic attenuation coefficient and the baseline attenuation coefficient on the round trip path and performing an exponential operation, a compensation factor that is exponentially related to the additional attenuation is constructed.

[0023] Preferably, the defect identification method further includes: setting a maximum compensation factor; when the calculated compensation factor is greater than the maximum compensation factor, using the maximum compensation factor as the final compensation factor to compensate and correct the original echo peak amplitude of the defect.

[0024] By adding a maximum compensation factor limit, significant robustness is provided for practical applications. Compared to unlimited compensation, it effectively prevents false defect signals from being generated due to over-amplification in areas with extremely high noise or extremely weak signals, thus avoiding misjudgments. While ensuring effective compensation for real defects, it also improves the signal-to-noise ratio and stability of the entire detection system.

[0025] Preferably, the maximum compensation multiple The value range is 80~120.

[0026] Preferably, the amplitude of the defective signal is compared with the threshold of the standard DAC curve at the corresponding depth to determine whether a defect exists, including: if the amplitude of the corrected defective signal is greater than the preset amplitude threshold, a defect is determined to exist; if the amplitude of the corrected defective signal is less than or equal to the preset amplitude threshold, a defect is determined not to exist.

[0027] This ensures that the final determination of defects is based on a signal value that has eliminated the influence of material inhomogeneity and is closer to its true reflectivity, thus making the determination result more fair and accurate.

[0028] Preferably, the preprocessing of the backscattered signal includes: performing detrending processing on the backscattered signal to eliminate baseline drift; and performing bandpass filtering on the detrending backscattered signal to filter out noise unrelated to the detection frequency.

[0029] By using detrending processing and bandpass filtering, baseline drift and frequency band noise that are irrelevant to detection that may occur during signal acquisition can be effectively eliminated.

[0030] Preferably, the backscattered signal corresponds to a depth The calculation method is as follows: ,in, The speed at which ultrasound propagates in the blank. It is the center moment of the stable time window.

[0031] By using the center moment of a stable time window for calculation, a unified and clear standard is provided for the mapping from time-domain signals to spatial depth. This ensures that the calculated depth characteristic parameters, attenuation coefficients, etc., can accurately correspond to the physical positions inside the forging, thus guaranteeing the accuracy of the entire spatial variation model.

[0032] The present invention has the following technical effects:

[0033] 1. This invention proposes a dynamic attenuation compensation method based on backscattered signals. This method analyzes the power spectrum characteristics of the backscattered signal to construct a dynamic attenuation model that reflects the changes in the material's microstructure along depth in real time. Based on this model, a compensation factor is generated to accurately correct defect echoes. This fundamentally eliminates the interference of material inhomogeneity on the detection results, transforming a fixed compensation mode into dynamic adaptive compensation, thereby improving the accuracy and reliability of defect identification.

[0034] 2. By analyzing the backscattered signal in the ultrasonic A-scan signal, frequency domain parameters reflecting local microstructure characteristics such as grain size and inclusion distribution are extracted, and the power spectral density of the backscattered signal is calculated. A dynamic attenuation coefficient that varies with depth is constructed, and a depth-adaptive compensation factor is generated accordingly. The measured defect signal is corrected point by point, which breaks through the limitation of the traditional DAC technology that relies on a set uniform attenuation model. It realizes accurate restoration and reliable determination of the defect echo amplitude in the background of non-uniform materials. Attached Figure Description

[0035] Figure 1 This is a flowchart of a method for identifying defects in large shaft forging blanks based on ultrasonic flaw detection data, according to the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] This invention discloses a method for identifying defects in large shaft forging blanks based on ultrasonic flaw detection data, referring to... Figure 1 This includes the following steps:

[0038] S1: Acquire ultrasound A-scan signal and backscatter signal, and perform preprocessing.

[0039] When performing ultrasonic testing on large shaft forging blanks, the system records the complete ultrasonic A-scan signal as the probe moves to each testing point. Ultrasound A-scan is an amplitude-time scan, a current technology, which simultaneously records the three-dimensional position of the probe. Scan signal It includes various echoes reflected from the surface, internal defects, and internal grains of large shaft forging blanks. Among them, the signal formed by random scattering by internal grains or inclusions is called the backscattered signal. In other words, the backscattered signal refers to the continuous signal background formed by the random scattering of ultrasonic waves by the microscopic non-uniform structure such as grains or inclusions inside the large shaft forging blank.

[0040] In this embodiment, a stable time window without macroscopic defects is selected from the ultrasonic A-scan signal, located after the interface echo and before the expected defect region. The signal extracted during this time period is recorded as the backscattered signal. ,in , The depth of the large shaft forging blank corresponding to the center of this stable time window is determined by ultrasonic speed. The result is obtained after zero-point calibration, and the calculation method is as follows: ,in, It is the center moment of the stable time window.

[0041] To improve the signal-to-noise ratio in subsequent analysis, the backscattered signal... Detrending processing is performed to eliminate baseline drift in the signal; and bandpass filtering is used for preprocessing to remove random noise unrelated to the detection frequency. After preprocessing, the backscattered signal is obtained. .

[0042] S2: Obtain the original echo peak amplitude of the suspected defect and correct the original echo peak amplitude.

[0043] By analyzing the backscattered signal, the acoustic characteristics of shaft forging blanks varying along the depth direction are quantified, and the defect signal is compensated accordingly.

[0044] a. Construct characteristic parameters to describe local acoustic features.

[0045] The microstructure of the raw material affects the scattering characteristics of ultrasound waves. The difference between the spectral centroid and the probe center frequency reflects the attenuation of high-frequency components. The more severe the microstructure inhomogeneity of the raw material, such as grain coarsening and differences in inclusion distribution, the more significant the difference between the spectral centroid and the probe center frequency. Therefore, by calculating the degree of deviation of the spectral centroid from the probe center frequency, the energy absorption variation of large shaft forgings can be characterized. The degree of deviation of the power spectrum distribution kurtosis from the reference value reflects the degree of local structural disorder of the scattering source. The more severe the microstructure inhomogeneity of the material, the more the power spectrum distribution kurtosis deviates from the reference value. Therefore, by calculating the difference between the power spectrum distribution kurtosis and the reference value, abnormal aggregation or distortion of inclusions in the microstructure can be captured. Finally, these two calculations are combined to describe the characteristic parameters of local acoustic features, reflecting the influence of microstructure inhomogeneity at various depths.

[0046] Specifically, for the preprocessed backscattered signal Perform a Fourier transform to obtain its frequency domain representation. ,in Represents frequency variables. The depth of the large shaft-type forging blank corresponding to the center of this stable time window is determined, and the power spectral density of the backscattered signal is calculated. .

[0047] Power spectral density of backscattered signal The centroid frequency relative to the probe center frequency The normalized offset term and the kurtosis deviation of the power spectral distribution from the reference value Multiplying the absolute enhancement terms, a characteristic parameter is constructed that comprehensively reflects the local acoustic characteristics of large shaft forging blanks. The calculation expression is as follows:

[0048] ;

[0049] In the formula, Indicates the depth inside the blank. Characteristic parameters at the location; It is a frequency variable; In internal depth The power spectral density of the backscattered signal at that location; It is the center frequency of the ultrasonic probe (before actual testing, the operator can pre-select a probe with a suitable center frequency according to the material, thickness, detection sensitivity and other requirements of the tested blank). It is the fourth central moment of the power spectral density, calculated using the fourth moment in higher-order statistics (HOS), reflecting the sharpness of the power spectral distribution and used to measure whether there is abnormally strong scattering in the signal; It is the second central moment of the power spectral density, that is, the variance of the power spectrum, which reflects the degree of energy dispersion of the power spectrum at the frequency. The larger the value, the wider and less concentrated the frequency components are. It is a reference value for the kurtosis of the power spectrum distribution. Under the standard condition of uniform material structure and no abnormalities, it is the average value of the kurtosis of the power spectrum distribution obtained through experimental calibration or statistical analysis.

[0050] It should be noted that, It is the centroid frequency of the backscattered signal power spectral density relative to the probe center frequency. The normalized offset term, when there is grain coarsening or high overall inclusion density, exacerbates the attenuation of high-frequency components. The larger the value of this term, the worse the final characteristic parameters. The larger the value, the better; conversely, if the material has a fine and uniform structure, weak high-frequency attenuation, and the center-of-gravity frequency is close to the probe's center frequency, the smaller this value will be, resulting in a lower final characteristic parameter. The smaller the value.

[0051] The power spectrum distribution kurtosis deviates from the reference value. The absolute enhancement term is used to detect local anomalies in the microstructure, such as clusters of inclusions or micro-regional distortions. When these local structural anomalies are present, the backscattered signal exhibits non-Gaussian, strong impulse characteristics, with kurtosis significantly deviating from the reference value. This value increases rapidly; conversely, even if the overall grain size of the material is relatively coarse, as long as there is no obvious abnormal aggregation, the kurtosis will still be close to [the expected value]. This value remains at a low level.

[0052] For example, suppose the blank is in There is slight non-uniformity at this point. (Probe center frequency) Power spectrum kurtosis reference value Assume that at this depth, the measured centroid frequency of the backscattered signal is... Assuming the measured power spectrum kurtosis Then the characteristic parameters .

[0053] b. Calculate the dynamic attenuation coefficient.

[0054] Since the non-uniformity of the microstructure of large shaft forging blanks along the depth direction has a cumulative effect, an exponential decay function is used as the weight to perform weighted integration on the deviation of each layer along the path, thereby constructing a dynamic decay coefficient that varies with depth, and realizing point-by-point correction of energy decay caused by the non-uniformity of the microstructure of the material.

[0055] Specifically, the characteristic parameters Relative to its benchmark value The deviation term along the ultrasound propagation path Perform an exponentially weighted integral, with weights varying according to depth difference. The coefficient increases while decreasing to reflect the cumulative effect of each layer of the microstructure along the path on the attenuation characteristics at the current depth, thus constructing a dynamic attenuation coefficient that comprehensively reflects the propagation effect of material inhomogeneity. The calculation expression is as follows:

[0056] ;

[0057] In the formula, It is at a depth inside the blank. The dynamic attenuation coefficient at the location; It is the reference attenuation coefficient under ideal uniform conditions; Indicates the depth of the blank Characteristic parameters at the location; These are characteristic parameters The baseline value under uniform conditions; It is the characteristic length of exponential decay, representing the effective distance of the influence of the non-uniformity of the tissue at a certain point on the subsequent path. This value can be determined based on the material tissue evolution characteristics and ultrasonic testing calibration. It is the base of the natural logarithm; It is a cumulative integral process from the surface (0) to the current depth (d).

[0058] These are characteristic parameters Relative to its benchmark value The larger the deviation term, the more uneven the organization at that location, resulting in more significant energy loss. Therefore, the dynamic attenuation coefficient... The larger the value, the lower the value; conversely, when the deviation term is close to zero or negative, it indicates that the material structure at that depth is close to a uniform state, and the dynamic attenuation coefficient is lower. The smaller the value.

[0059] This ensures that when calculating the attenuation coefficient at the current depth d, the distance to the current position is closer, i.e. Smaller characteristic parameters Those assigned higher weights contribute more to α(d); while those farther away, i.e. Larger shallow characteristic parameters Its weight decays exponentially with distance.

[0060] For example, in The calculation is performed at the reference attenuation coefficient. ; Uniform state reference value ; Exponential decay characteristic length Simplifying assumptions about the path Above, characteristic parameters Constant as The dynamic attenuation coefficient is:

[0061] .

[0062] c. Calculate the compensation factor and correct the original echo peak amplitude.

[0063] Dynamic attenuation coefficient Compared with the reference attenuation coefficient The difference along the round-trip propagation path Accumulate and construct compensation factors through exponential operations. The calculation expression is as follows:

[0064] ;

[0065] In the formula, In internal depth Compensation factor at the location; This is the current depth; In internal depth The dynamic attenuation coefficient at the location; It is the baseline attenuation coefficient of the material under ideal homogeneous conditions; It is an exponential function with the natural constant as its base.

[0066] This represents the additional attenuation coefficient caused by the non-uniformity of the raw material. This represents the total attenuation over the round-trip propagation path of the ultrasound wave. When Greater than That is, when the actual attenuation is more severe, this cumulative value is positive, and the compensation factor is obtained through exponential calculation. It will be greater than 1. The more severe the additional attenuation, the greater the value. The larger the value, the stronger the amplitude compensation required.

[0067] For example, .

[0068] When an object located at depth is found in the A scan signal When a suspected defect is detected, record its original echo peak amplitude. Before comparing with the standard DAC curve, the compensation factor corresponding to this depth is first calculated. The correction is performed to obtain the corrected defect signal amplitude. The calculation expression is as follows:

[0069] ;

[0070] In the formula, It is the corrected amplitude of the defect signal; It is detected in the A scan signal at a depth The original echo peak amplitude of the suspected defect at the location; At the depth of the defect The corresponding compensation factor.

[0071] When compensation factor A value greater than 1 indicates that the additional attenuation caused by the material at that depth is relatively severe, and the compensation force is correspondingly enhanced, thereby reducing the underestimated defect signal amplitude. By adjusting upwards, we obtain a result that more closely approximates its true reflective ability. Conversely, when When the value is close to 1, it indicates that the material at that location is close to an ideal state, and the compensation effect is weak. Compared with the original amplitude Basically the same.

[0072] To prevent over-amplification of the signal in noisy regions, this invention sets a maximum compensation factor. (Can be set to 100 times). If the calculated Exceed Then use As a compensation factor.

[0073] For example, the measured raw echo peak amplitude The corrected defect signal amplitude This correction process effectively restores the intensity of defect signals that were weakened by additional material attenuation.

[0074] S3: Compare the defect signal amplitude with the preset amplitude threshold to identify defects in large shaft forging blanks.

[0075] Since the corrected defect signal amplitude has eliminated the energy attenuation deviation caused by the non-uniformity of the material's microstructure, it can accurately reflect the reflectivity of the defect. Therefore, it is compared with the standard DAC threshold for the corresponding depth. If the corrected amplitude exceeds the threshold, it is determined to be a valid defect; otherwise, it is considered as a defect-free defect.

[0076] Specifically, this involves adjusting the amplitude of the corrected defect signal. Compared with the standard DAC curve in depth The corresponding amplitude threshold Comparison:

[0077] if If so, then the location is determined to have a valid defect;

[0078] if If the result is negative, then it is determined that there is no defect at that location.

[0079] A standard DAC curve refers to a family of reference curves obtained by measuring the echo amplitude of artificial defects of the same size at different depths on a standard test block. It typically includes an evaluation line, a quantitative line, and a rejection line, used to establish a baseline relationship between amplitude and distance. In this invention, the corrected defect signal amplitude is... The defect is compared with the preset rejection line threshold at the corresponding depth in this family of curves to determine whether the defect exceeds the standard. This setting can be flexibly adjusted according to the actual inspection standards.

[0080] For example, suppose in the standard DAC curve, the rejection threshold is... ,because If the result is negative, then it is determined that there is no defect at that location.

[0081] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A large shaft forging blank defect identification method based on ultrasonic flaw detection data, characterized in that, The method comprises the steps of: using a probe to perform ultrasonic detection on a large shaft forging blank to obtain an ultrasonic A-scan signal of the large shaft forging blank; intercepting a signal in a stable time window after an interface echo and before an expected defect region from the ultrasonic A-scan signal to obtain a backscattering signal; pre-processing the backscattering signal; calculating a power spectral density of the backscattering signal; obtaining a characteristic parameter related to depth based on the power spectral density, the characteristic parameter being positively correlated with a difference between a center frequency of the probe and a center-of-gravity frequency of the power spectral density; and the characteristic parameter satisfying: wherein f0 is the center frequency of the probe, f1 is the center-of-gravity frequency of the power spectral density, and f2 is the characteristic parameter. ； represents a characteristic parameter at an internal depth of the embryo; represents a characteristic parameter at an internal depth of the embryo; is a frequency variable; is a power spectral density of the backscattered signal at an internal depth is a power spectral density of the backscattered signal at an internal depth is a center frequency of the ultrasound probe; is a fourth order central moment of the power spectral density; is a second order central moment of the power spectral density; is a reference value for the kurtosis of the power spectral distribution; Based on a cumulative deviation of the characteristic parameter along an ultrasonic propagation path, and in combination with a set reference attenuation coefficient, a dynamic attenuation coefficient varying with depth is calculated, satisfying: wherein f0 is the center frequency of the probe, f1 is the center-of-gravity frequency of the power spectral density, f2 is the characteristic parameter, and a is the dynamic attenuation coefficient. ； is the dynamic attenuation coefficient at the embryo depth is the reference attenuation coefficient in the ideal homogeneous state; is the reference value of the characteristic parameter in the homogeneous state; is the characteristic parameter at the embryo depth is the characteristic parameter at the depth is the characteristic parameter at the depth is the reference value of the characteristic parameter in the homogeneous state; is the characteristic length of the exponential decay; is the base of the natural logarithm; is the cumulative integration process from the surface 0 to the current depth d; A difference between the dynamic attenuation coefficient and the reference attenuation coefficient is accumulated along a round-trip path, and an exponential operation is performed to obtain a compensation factor, satisfying: wherein f0 is the center frequency of the probe, f1 is the center-of-gravity frequency of the power spectral density, f2 is the characteristic parameter, a is the dynamic attenuation coefficient, and b is the compensation factor. ； is a compensation factor at an internal depth is a compensation factor at an internal depth is a current depth is an exponential function with a natural constant as base The original echo peak amplitude of a detected defect in the A-scan signal is compensated using the compensation factor to obtain a defect signal amplitude. The defect signal amplitude is compared with a threshold value of a standard DAC curve at a corresponding depth to determine whether a defect exists.

2. The method according to claim 1, wherein The fourth-order central moment of the power spectral density is calculated using the fourth-order moment in high-order statistics.

3. The method according to claim 1, wherein The defect recognition method further comprises: setting a maximum compensation multiple; and when the calculated compensation factor is greater than the maximum compensation multiple, using the maximum compensation multiple as the final compensation factor to compensate and correct the original echo peak amplitude of the defect.

4. The method according to claim 3, wherein the maximum compensation factor is in the range of 80 to 120.

5. The method according to claim 1, wherein The defect signal amplitude is compared with a threshold value of a standard DAC curve at a corresponding depth to determine whether a defect exists, comprising: if the corrected defect signal amplitude is greater than a preset amplitude threshold value, determining that a defect exists; and if the corrected defect signal amplitude is less than or equal to the preset amplitude threshold value, determining that no defect exists.

6. The method according to claim 1, wherein The pre-processing of the backscattering signal comprises: de-trending the backscattering signal to eliminate baseline drift; and performing band-pass filtering on the de-trended backscattering signal to filter out noise unrelated to the detection frequency.

7. The method according to claim 1, wherein Backscattered signal corresponds to depth The calculation method is: wherein, is the propagation speed of the ultrasound in the billet, is the center time of the stabilization time window.

Images