Method for detecting welding seam by using ultrasonic flaw detector

Abstract

The invention provides a method for detecting a welding seam by adopting an ultrasonic flaw detector, and belongs to the technical field of welding seam detection. The welding seam is divided into small welding sections, and each small welding section is subjected to multi-direction ultrasonic scanning; a scattering attenuation coefficient and a signal-to-noise ratio parameter are calculated to determine an optimal detection position direction vector group and an optimal time frequency group, and then a high-frequency narrow pulse probe is matched with a phased array focusing technology to carry out second detection according to the optimized parameter; and the identified defects are classified and evaluated according to the near surface blind area characteristic parameters, secondary verification is performed by adopting a surface wave supplementary detection algorithm, and finally a weld defect position distribution diagram is generated, so that the technical problem of low defect detection accuracy caused by serious ultrasonic scattering attenuation of a coarse grain structure weld material is solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of weld detection, and in particular relates to a method for detecting a weld using an ultrasonic flaw detector. BACKGROUND

[0002] In the field of weld quality detection, ultrasonic flaw detection technology is widely used for non-destructive detection of internal defects of welded joints. Traditional ultrasonic detection methods usually use single-crystal straight probes or angle probes to scan the weld at a fixed angle, and identify and locate internal defects by analyzing the time-domain characteristics of the echo signals. However, in practical applications, many welding materials, especially stainless steel welds and nickel-based alloy welds, are coarse-grained materials, and the grain size of these materials often exceeds 1 millimeter. The coarse grains of such materials can cause strong scattering of ultrasonic waves, resulting in a significant attenuation of ultrasonic energy during propagation, which severely weakens the defect echo signals and even causes them to be submerged in scattering noise. Scattering also causes spectral broadening and a sharp decline in signal-to-noise ratio. Traditional detection methods often use fixed detection angles and fixed frequency parameters when facing coarse-grained welds, and cannot adaptively adjust to the scattering characteristics of different regions, resulting in insufficient detection sensitivity, high defect miss rate, poor positioning accuracy, and other problems. That is, the prior art has the technical problem of low defect detection accuracy due to severe ultrasonic scattering attenuation in coarse-grained weld materials. SUMMARY

[0003] Therefore, the present application provides a method for detecting a weld using an ultrasonic flaw detector, which can solve the technical problem of low defect detection accuracy due to severe ultrasonic scattering attenuation in coarse-grained weld materials in the prior art.

[0004] The present application is implemented in the following manner: the present application provides a method for detecting welds using an ultrasonic flaw detector, comprising the following steps: dividing the weld to be detected into a plurality of small weld sections along the welding direction and coating a coupling agent layer on the surface; using a double crystal probe to perform multi-directional scanning on the small weld sections to record first echo data; performing spectral analysis on the first echo data to extract signal-to-noise ratio parameters and echo amplitude parameters, calculating scattering attenuation coefficients and screening a set of candidate direction angles; selecting the detection direction angle with the largest signal-to-noise ratio parameter in the set of candidate direction angles as the optimal detection direction angle, recording excitation frequency parameters and pulse width parameters to form an optimal detection position direction vector group and an optimal time frequency group; repeating the above operations on all small weld sections in turn to obtain their respective optimal detection position direction vector groups and optimal time frequency groups; using a high-frequency narrow pulse probe in combination with phased array focusing technology to perform a second scan according to the optimal detection position direction vector group and the optimal time frequency group to obtain second echo data; performing defect recognition processing on the second echo data; classifying and evaluating the recognized defects; generating a weld defect position distribution map; and determining the optimal detection parameters for each small weld section to reduce the influence of scattering attenuation of coarse-grained materials.

[0005] In the step of dividing the weld to be detected into a plurality of small weld sections along the welding direction, surface roughness measurement is performed on each small weld section, and when the surface roughness exceeds a preset value, surface polishing treatment is performed to reduce the surface roughness to below a specified value before coating the coupling agent layer.

[0006] The calculation process of the scattering attenuation coefficient is to divide the actually measured echo amplitude parameter by a standard echo amplitude parameter in an ideal state without scattering to obtain an attenuation ratio value, and the standard echo amplitude parameter is obtained by measuring a fine-grained standard test block. The attenuation ratio value reflects the degree of scattering loss of ultrasonic waves by coarse-grained materials.

[0007] The optimal detection position direction vector group includes three angle parameters: an optimal detection direction angle, an optimal probe deflection angle, and an optimal probe tilt angle. The optimal time frequency group includes three time domain and frequency domain parameters: an excitation frequency parameter, a pulse width parameter, and a repetition frequency parameter.

[0008] The phased array focusing technology forms a sound beam convergence at a set depth position by controlling the excitation delay time of each array element in the probe array, which can concentrate acoustic energy inside the near-surface blind area, thereby improving the detection resolution and defect recognition capability of the near-surface blind area.

[0009] The calculation method of the near-surface blind area characteristic parameter is to divide the difference between the echo arrival time of the defect and the start wave end time by the time length corresponding to the pulse width parameter. When the difference is less than a preset multiple of the time length corresponding to the pulse width parameter, it is determined that the defect is located in the near-surface blind area.

[0010] In the step of classifying and evaluating the identified defects, when the value of the near-surface blind zone feature parameter is greater than the first threshold (the empirical value is 0.5, that is, the difference between the arrival time of the defect echo and the end time of the initial wave is exactly equal to half of the pulse width), the surface wave supplementary detection algorithm is activated to perform secondary verification of the defect area. The surface wave supplementary detection algorithm uses a surface wave probe to perform transverse wave detection along the weld surface.

[0011] In the step of classifying and evaluating the identified defects, when the value of the near-surface blind zone characteristic parameter is less than the second threshold (empirical value is 0.7, that is, the difference between the arrival time of the defect echo and the end time of the initial wave is less than 0.3 times the pulse width), the conventional ultrasonic testing method is used to complete the quantitative analysis of the defects. The conventional ultrasonic testing method is to use a single crystal straight probe to perform longitudinal wave detection inside the workpiece in a vertical incident manner.

[0012] The spectrum analysis process includes performing a fast Fourier transform on the first echo data to obtain the frequency domain distribution characteristics, and extracting the dominant frequency component and bandwidth parameter from the frequency domain distribution characteristics. The dominant frequency component reflects the influence characteristics of the material grain structure on the ultrasonic wave.

[0013] The signal-to-noise ratio parameter is calculated by dividing the peak amplitude of the defect echo signal by the root mean square amplitude of the background noise signal, which is extracted from the echo data of the defect-free region.

[0014] Optionally, the dual-crystal probe is a transceiver separate structure probe. The transceiver separate structure probe can be a dual-crystal straight probe or a dual-crystal angled probe available on the market. The dual-crystal straight probe is suitable for vertical incidence detection, and the dual-crystal angled probe is suitable for oblique incidence detection.

[0015] Optionally, the transceiver-separated structure probe includes a transmitting chip and a receiving chip, with a sound insulation layer between the transmitting chip and the receiving chip. The transceiver-separated structure probe emits ultrasonic waves through the transmitting chip and receives the reflected echoes through the receiving chip, thereby eliminating the initial wave interference.

[0016] In the second scan using a high-frequency narrow pulse probe in conjunction with phased array focusing technology, when the excitation frequency parameter in the optimal time frequency group of a small weld segment is lower than the preset frequency value, the small weld segment is determined to be a strong scattering area. At this time, the center frequency of the high-frequency narrow pulse probe is adjusted to a higher frequency to reduce the scattering effect.

[0017] In the defect identification process of the second echo data, when multiple adjacent small weld segments are detected with defects, the extended scanning mode is activated, merging the adjacent small weld segments into an extended detection area and using a multi-angle composite scanning method for comprehensive detection.

[0018] In the step of generating the weld defect position distribution map, when the weld defect position distribution map shows that the proportion of the number of defects detected to the total number of small weld segments exceeds a preset proportion, it is determined that the weld to be detected has a systematic quality problem, and at this time, the welding process parameters need to be reviewed.

[0019] The reduction of the pulse width parameter of the high-frequency narrow pulse probe can shorten the width range of the near-surface blind area, so that the defect echo signal of the near-surface area is separated from the initial wave in the time domain.

[0020] In the step of multi-directionally scanning the small weld segments by using the double-crystal probe, the first echo data is repeatedly collected multiple times at each detection direction angle, and the average value is calculated as the final first echo data of the detection direction angle.

[0021] The present application divides the weld into multiple small weld segments and performs multi-directional angle ultrasonic scanning on each small weld segment, calculates the scattering attenuation coefficient and signal-to-noise ratio parameter under different detection directions, thereby determining the optimal detection position direction vector group and the optimal time frequency group for each small weld segment, and then using a high-frequency narrow pulse probe in combination with a phased array focusing technology to perform a second accurate detection according to the optimized parameters. The method can screen out detection directions with less scattering influence through the quantitative evaluation of the scattering attenuation coefficient, ensure the effective identification of defect signals through the optimization of the signal-to-noise ratio parameter, concentrate sound energy on the key detection area through the phased array focusing technology to compensate for the energy loss caused by scattering, and use a surface wave supplementary detection algorithm to achieve full-depth coverage for the near-surface blind area. In summary, the present application solves the technical problem of low defect detection accuracy caused by severe ultrasonic scattering attenuation in the coarse-grained structure weld material mentioned in the background art. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The flowchart of the method of the present application.

[0023] Figure 2 The detection direction angle and signal-to-noise ratio parameter relationship curve in the embodiment.

[0024] Figure 3 The scattering attenuation coefficient distribution graph in the embodiment.

[0025] Figure 4 The near-surface blind area characteristic parameter distribution graph in the embodiment.

[0026] Figure 5 The weld defect position distribution graph in the embodiment. DETAILED DESCRIPTION

[0027] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0028] As Figure 1 shown is a flow chart of a method for detecting a weld joint by using an ultrasonic flaw detector, and the method comprises the following steps:

[0029] S01, the weld joint to be detected is evenly divided into a plurality of small weld sections with a length in the range of 15mm to 30mm along the welding direction, and a coupling agent layer is coated on the surface of each small weld section, and the thickness of the coupling agent layer is 0.5mm to 1.2mm;

[0030] S02, a first small weld section is selected as an initial detection object, a double-crystal probe is used to perform multi-directional scanning on the surface of the initial detection object, and first echo data of ultrasonic waves received under different detection direction angles are recorded, the change range of the detection direction angle is 0 degrees to 90 degrees, and the change step of the detection direction angle is 5 degrees;

[0031] S03, the first echo data are processed by spectrum analysis, a signal-to-noise ratio parameter and an echo amplitude parameter corresponding to each detection direction angle are extracted, a scattering attenuation coefficient of each detection direction angle is calculated, the scattering attenuation coefficient is determined by the ratio of the echo amplitude parameter to a standard echo amplitude parameter, and a detection direction angle set with a scattering attenuation coefficient less than 0.6 is screened out to form a candidate direction angle set;

[0032] S04, a detection direction angle with the largest signal-to-noise ratio parameter in the candidate direction angle set is selected as an optimal detection direction angle of the initial detection object, and an excitation frequency parameter and a pulse width parameter under the optimal detection direction angle are recorded, and the optimal detection direction angle, the excitation frequency parameter and the pulse width parameter form an optimal detection position direction vector set and an optimal time frequency set of the initial detection object;

[0033] S05, the operation processes of steps S02 to S04 are repeatedly performed on each remaining small weld section in turn, and optimal detection position direction vector sets and optimal time frequency sets corresponding to all small weld sections are obtained;

[0034] S06, a high-frequency narrow pulse probe is used in combination with a phased array focusing technology to perform second scanning detection according to the parameter settings of the optimal detection position direction vector set and the optimal time frequency set of each small weld section, and ultrasonic second echo data are obtained, the center frequency of the high-frequency narrow pulse probe is 10MHz to 20MHz, and the focal point depth of the phased array focusing technology is set to a range of 1mm to 4mm from the surface;

[0035] S07, performing defect identification processing on the second echo data, determining that a defect exists in a certain small welding segment when a ratio of a peak amplitude of the second echo data of the small welding segment to a bottom surface echo amplitude is greater than a determination threshold value, recording position coordinates and echo characteristic parameters of the defect, and an empirical value of the determination threshold value is 0.25;

[0036] S08, performing classification and evaluation on the identified defects, calculating echo spectrum characteristic parameters and near-surface blind area characteristic parameters of the defects, starting a surface wave supplementary detection algorithm to perform secondary verification on the defect area when a value of the near-surface blind area characteristic parameter is ∈ (0.7, 1], and completing defect quantitative analysis by using a conventional ultrasonic detection method when the value of the near-surface blind area characteristic parameter is ∈ [0, 0.7];

[0037] S09, generating a welding defect position distribution map according to results of defect identification processing and defect classification and evaluation of all small welding segments, the welding defect position distribution map containing position coordinates of the defects, type identifiers of the defects, and size parameters of the defects.

[0038] The dual-crystal probe is a transmitting-receiving separation type probe, the transmitting-receiving separation type probe comprises a transmitting wafer and a receiving wafer, a sound insulation layer is arranged between the transmitting wafer and the receiving wafer, the thickness of the sound insulation layer is 0.3 mm to 0.8 mm, the transmitting-receiving separation type probe transmits ultrasonic waves through the transmitting wafer and receives reflected echoes through the receiving wafer, so as to eliminate initial wave interference, the initial wave interference is a shielding effect of a direct wave signal generated by a probe excitation pulse on a defect echo signal on a workpiece surface.

[0039] The calculation process of the scattering attenuation coefficient is to divide the actually measured echo amplitude parameter by a standard echo amplitude parameter in an ideal state without scattering to obtain an attenuation ratio value, the standard echo amplitude parameter is obtained by measurement on a fine-grained standard block, the attenuation ratio value reflects the scattering loss degree of a coarse-grained material on ultrasonic waves, and the coarse-grained material is a metal material with a grain size greater than 1 mm.

[0040] The optimal detection position direction vector group includes three angle parameters of an optimal detection direction angle, an optimal probe deflection angle and an optimal probe tilt angle, the optimal time frequency group includes three time domain and frequency domain parameters of the excitation frequency parameter, the pulse width parameter and the repetition frequency parameter, and the optimal detection position direction vector group and the optimal time frequency group jointly determine a detection configuration scheme for obtaining the best signal-to-noise ratio parameter in the coarse-grained material.

[0041] The phased array focusing technology forms a sound beam convergence at a set depth position by controlling the excitation delay time of each array element in the probe array, which can concentrate sound energy inside the near-surface blind area to improve the detection resolution and defect recognition ability of the near-surface blind area, the excitation delay time is determined according to the focal depth and the array element position through sound path calculation, and the near-surface blind area is an area within 2-5mm below the surface of the workpiece which cannot be identified due to the overlap of defect echo and initial wave caused by probe excitation pulse width and protective film delay.

[0042] The near-surface blind area characteristic parameter is calculated by dividing the difference between the echo arrival time of the defect and the end time of the initial wave by the time length corresponding to the pulse width parameter, and when the difference is less than 1.5 times the time length corresponding to the pulse width parameter, it is determined that the defect is located in the near-surface blind area, and at this time the near-surface blind area characteristic parameter value is larger.

[0043] The surface wave supplementary detection algorithm uses a surface wave probe to detect transverse waves along the weld surface, and the surface wave propagates on the material surface and produces a reflected echo for defects within 2-5mm below the surface, so the surface wave supplementary detection algorithm can cover the near-surface blind area of the conventional ultrasonic detection method to achieve comprehensive detection of the defects.

[0044] The conventional ultrasonic detection method is a method of using a single crystal straight probe to detect longitudinal waves vertically on the inside of the workpiece, which is suitable for detecting internal defects at a depth of more than 5mm from the surface, and the conventional ultrasonic detection method completes defect positioning and quantification by analyzing the time domain characteristics of the bottom surface echo and the defect echo.

[0045] The frequency spectrum analysis processing includes performing fast Fourier transform on the first echo data to obtain frequency domain distribution characteristics, extracting the main frequency component and the frequency band width parameter in the frequency domain distribution characteristics, the main frequency component reflects the influence characteristics of the material grain structure on the ultrasonic wave, and the frequency band width parameter reflects the degree of spectral broadening caused by scattering.

[0046] The signal-to-noise ratio parameter is calculated by dividing the peak amplitude of the defect echo signal by the root mean square amplitude of the background noise signal, and the background noise signal is extracted from the echo data of the defect-free area, and when the signal-to-noise ratio parameter is greater than 6, it is considered that the defect echo signal can be effectively identified.

[0047] The echo characteristic parameters include three time domain parameters of the peak amplitude, arrival time and waveform duration, and the echo characteristic parameters are used to distinguish different types of internal defects in the weld and calculate the size parameters of the defects.

[0048] The echo spectrum characteristic parameters include a center frequency offset and a bandwidth expansion coefficient, the center frequency offset is a difference between a center frequency of the echo of the defect and a center frequency of an excitation signal, and the bandwidth expansion coefficient is a ratio between a frequency band width of the echo of the defect and a frequency band width of the excitation signal, and the echo spectrum characteristic parameters are used for classification evaluation of the defect.

[0049] The standard echo amplitude parameter in step S03 is obtained by performing ultrasonic detection on a standard defect test block with a known depth and size, and recording the echo amplitude parameter, the material type and grain size of the standard defect test block match the weld to be detected, and the standard echo amplitude parameter is used as a reference value for evaluating the scattering attenuation degree.

[0050] The determination threshold in step S07 is determined by an experimental method, the experimental process is to perform ultrasonic detection on a standard weld sample with a known defect size, collect 100 groups of second echo data containing defects and 100 groups of second echo data without defects, calculate the ratio of the peak amplitude to the bottom echo amplitude of the second echo data respectively, statistically analyze 200 groups of ratio data, draw a ratio distribution curve, and select the ratio that makes the defect detection rate reach 95% and the false positive rate is lower than 5% as the determination threshold, and the experimental data analysis result shows that the experience value of the determination threshold is 0.25.

[0051] In step S08, when the value of the near-surface blind area characteristic parameter is ∈ [0.5, 0.7], the phased array focusing technology is used to enhance detection and confirm the defect, and when the value of the near-surface blind area characteristic parameter is ∈ [0, 0.5), it is determined that the defect is located in the conventional detection range and the conventional ultrasonic detection method is used to complete the quantitative analysis of the defect.

[0052] Further, in step S01, the surface roughness of each small weld section is measured, and when the surface roughness exceeds 25 μm, the surface is polished to reduce the surface roughness to less than 10 μm before the coupling agent layer is applied. The large surface roughness will cause poor coupling and affect the propagation effect of ultrasonic waves.

[0053] Further, in step S06, when the excitation frequency parameter in the optimal time-frequency group of a certain small weld section is lower than 5 MHz, it is determined that the small weld section is a strong scattering area, and at this time, the center frequency of the high-frequency narrow pulse probe is adjusted to more than 15 MHz to reduce the scattering effect and improve the penetration ability.

[0054] Further, when the three adjacent small weld sections are detected with the defect in step S07, the extended scanning mode is started, the three adjacent small weld sections are combined into one extended detection area, and the multi-angle composite scanning mode is used for comprehensive detection to accurately determine the continuous distribution range of the defect.

[0055] Further, when the proportion of the number of defects detected in the weld defect position distribution map to the total number of small weld sections exceeds 30% in step S09, it is determined that the weld seam to be detected has a systematic quality problem, and at this time, the welding process parameters need to be reviewed and the weld seam to be detected needs to be re-detected and verified.

[0056] Further, the pulse width parameter of the high-frequency narrow pulse probe is 50nS0 to 150nS0, and the reduction of the pulse width parameter can shorten the width range of the near-surface blind area, separate the near-surface area defect echo signal from the initial wave in time domain, and thus improve the detection capability of the near-surface defect.

[0057] Further, in step S02, the first echo data is repeatedly collected 5 times at each detection direction angle, and the average value is calculated as the final first echo data of the detection direction angle. Repeated collection can reduce the influence of random noise and improve the reliability of the data.

[0058] Further, the type identification in the weld defect position distribution map includes pore defect identification, crack defect identification, incomplete fusion defect identification, and slag inclusion defect identification. Different types of defects are automatically classified and identified according to the differences in the echo characteristic parameters and the echo spectrum characteristic parameters.

[0059] The specific implementation of the above steps is described in detail below.

[0060] The specific implementation of step S01 is to first use a laser range finder to continuously measure along the center line of the weld to obtain total length data of the weld, and to calculate the number of small weld sections based on the total length data. When dividing, ensure that the length of each small weld section is within the range of 15 mm to 30 mm. This length range is determined based on the balance between the ultrasonic beam divergence angle and the detection resolution. Too short a weld section will result in reduced detection efficiency, and too long a weld section will affect the consistency of the detection parameters due to differences in material grain orientation. After completing the weld section division, use a surface roughness meter to measure the surface of each small weld section. When the surface roughness exceeds 25 μm, use sandpaper or a grinder to treat the surface to reduce the roughness to below 10 μm. Control of the surface roughness is a prerequisite for effective coupling of the ultrasonic wave. A rough surface will form an air gap between the probe and the workpiece, resulting in a loss of acoustic energy. Then evenly coat a coupling agent layer on the treated surface. The thickness of the coupling agent layer is controlled within the range of 0.5 mm to 1.2 mm. This thickness range can fill the surface micro concave-convex and form a stable sound propagation path, while avoiding the problem of increased sound path length and increased attenuation due to an excessively thick coupling agent layer.

[0061] The specific implementation of step S02 is to select the first small weld section as the initial detection object and use a transmit-receive separated double crystal probe for detection. The double crystal probe is composed of a transmitting wafer and a receiving wafer, with a sound insulation layer of 0.3 mm to 0.8 mm in thickness between them. The role of the sound insulation layer is to prevent the transmission of the initial wave interference from the transmitting wafer to the receiving wafer, which can shield the echo signal of near-surface defects. During detection, the probe is placed on the surface of the small weld section, and the detection direction angle is adjusted from 0 degrees to 90 degrees in steps of 5 degrees. At each detection direction angle, the probe is excited to transmit ultrasonic waves and receive reflected echoes. To improve data reliability, the first echo data is repeatedly collected 5 times at each angle position, and the average value is calculated as the final data for that angle. Repeated collection can reduce the influence of random noise and transient interference through statistical averaging. The change in the detection direction angle is to address the anisotropic problem of sound beam scattering caused by uneven grain orientation in coarse-grained materials. The interaction between ultrasonic waves and grain boundaries at different angles produces differential scattering and attenuation effects. By multi-angle scanning, the propagation path with the least scattering can be found.

[0062] The specific implementation of step S03 is to perform spectral analysis processing on the collected first echo data, which converts the time domain echo signal into frequency domain distribution characteristics using a fast Fourier transform algorithm. The fast Fourier transform algorithm can efficiently calculate the frequency component distribution of a discrete signal. The main frequency component reflects the influence of material grain size on the frequency selective scattering of ultrasonic waves, and the frequency bandwidth parameter represents the degree of spectral energy diffusion caused by scattering. The signal-to-noise ratio parameter corresponding to each detection direction angle is calculated based on the frequency domain characteristics. The signal-to-noise ratio parameter is obtained by dividing the peak amplitude of the defect echo signal by the root mean square amplitude of the background noise signal extracted from the defect-free area. When the signal-to-noise ratio parameter is greater than 6, it is considered that the echo signal has sufficient recognizability. At the same time, the echo amplitude parameter at each angle is extracted and compared with the standard echo amplitude parameter. The standard echo amplitude parameter is obtained by performing ultrasonic testing on a standard defect test block with a material type and grain size matching the weld to be detected and recording the echo amplitude. The scattering attenuation coefficient is calculated by dividing the actual measured echo amplitude parameter by the standard echo amplitude parameter to obtain an attenuation ratio value. This value directly reflects the degree of scattering loss of ultrasonic waves by coarse-grained materials. Coarse-grained materials refer to metal materials with grain sizes greater than 1 mm, whose grain boundary size is close to the wavelength of ultrasonic waves, resulting in strong scattering effects. The detection direction angles with a scattering attenuation coefficient less than 0.6 are selected to form a set of candidate direction angles. A scattering attenuation coefficient less than 0.6 means that the sound energy loss in this direction is relatively small, making it suitable for effective detection.

[0063] The specific implementation of step S04 is to iterate through all candidate angles in the set of candidate direction angles, compare the signal-to-noise ratio parameter values corresponding to each angle, and select the detection direction angle with the largest signal-to-noise ratio parameter as the optimal detection direction angle for the small weld segment. The reason for selecting the angle with the largest signal-to-noise ratio is that the defect echo signal at this angle has the strongest distinguishability relative to noise, which can improve the accuracy and sensitivity of defect detection. At the same time, the excitation frequency parameter and pulse width parameter of the probe at the optimal detection direction angle are recorded. These parameters together constitute the configuration scheme for obtaining the best detection effect of the small weld segment in coarse-grained materials. The optimal detection direction angle, together with the probe deflection angle and probe tilt angle, forms the optimal detection position direction vector group. The excitation frequency parameter, together with the pulse width parameter and the repetition frequency parameter, forms the optimal time frequency group. The combination of these two groups of parameters determines the spatial pose and time domain excitation characteristics of the probe during subsequent precise detection, providing an optimized parameter basis for the second high-precision scan.

[0064] The specific implementation of step S05 is to select each remaining small welding segment in sequence along the weld line, and repeat the operation process described in steps S02 to S04 for each small welding segment to collect first echo data by multi-angle scanning, perform spectral analysis and scattering attenuation coefficient calculation, screen the candidate direction angle set and determine the optimal detection position direction vector group and the optimal time-frequency group. The purpose of this repeated process is to establish adaptive detection parameter configuration for the local differences in the microstructure of the material at different positions along the weld. Due to the non-uniformity of heat input and cooling speed during welding, the grain size and orientation vary along the weld direction. The step-by-step optimization strategy ensures that each local region is detected under optimal detection conditions, and finally the optimal detection parameter set covering all small welding segments of the entire weld is obtained.

[0065] The specific implementation of step S06 is to perform a second scanning detection using a high-frequency narrow pulse probe with a center frequency of 10-20 MHz in combination with phased array focusing technology. The pulse width parameter of the high-frequency narrow pulse probe is set to The narrow pulse width can shorten the width range of the near-surface blind zone, allowing the near-surface region defect echo signal to be separated from the primary wave in the time domain. The phased array focusing technology achieves the convergence of the ultrasonic beam at the set depth position by precisely controlling the excitation delay time of each array element in the probe array. The excitation delay time is calculated based on the target focal depth and the spatial position of each array element through the sound path geometry, so that the sound waves emitted by different array elements are in phase at the focal point, forming a high sound pressure region. The focal depth is set to a range of 1-4 mm from the workpiece surface. This depth range exactly covers the near-surface blind zone of conventional ultrasonic detection. The near-surface blind zone refers to the region within 2-5 mm below the workpiece surface where the defect echo and primary wave overlap due to the probe excitation pulse width and the protective film delay, making it impossible to identify. The sound beam convergence concentrates sound energy inside the near-surface blind zone, significantly improving the detection resolution and defect recognition ability of this region. During scanning, the spatial pose parameters of the probe are set according to the pre-determined optimal detection position direction vector group for each small welding segment, and the frequency and pulse parameters of the excitation signal are set according to the optimal time-frequency group. When the excitation frequency parameter in the optimal time-frequency group for a small welding segment is below 5 MHz, it is determined that this region is a strong scattering region. At this time, the center frequency of the high-frequency narrow pulse probe is adjusted to above 15 MHz to reduce scattering and improve penetration. Through this adaptive parameter configuration, a second accurate scanning of all small welding segments is completed and second echo data is obtained.

[0066] ​The specific implementation of step S07 is to perform a defect identification process on the second echo data. The process first extracts the peak amplitude of each small weld segment second echo data and the bottom surface echo amplitude, the bottom surface echo being the echo signal reflected back after the ultrasonic wave propagates to the workpiece bottom surface, calculates the ratio of the peak amplitude to the bottom surface echo amplitude, compares the ratio with a judgment threshold, and determines the threshold to be 0.25 by an experimental method. The experimental process is to collect 100 groups of data containing defects and 100 groups of data without defects on a standard weld sample with known defect size, statistically analyze the ratio of the 200 groups of data, and select the ratio that makes the defect detection rate reach 95% and the misjudgment rate be lower than 5% as the judgment threshold. When the ratio of a small weld segment is greater than 0.25, it is determined that there is a defect in the small weld segment, and the position coordinates and echo characteristic parameters of the defect are recorded. The position coordinates are calculated according to the arrival time of the defect echo and the sound speed, and the echo characteristic parameters include three time domain parameters of the peak amplitude, the arrival time and the waveform duration. These parameters are used for subsequent defect type identification and size quantitative analysis. When defects are detected in three adjacent small weld segments in succession, an extended scanning mode is started, the three adjacent small weld segments are merged into an extended detection area, and a multi-angle composite scanning mode is used for comprehensive detection to accurately determine the continuous distribution range of the defects.

[0067] The specific implementation of step S08 is to classify and evaluate the identified defects. The evaluation process calculates the echo spectrum characteristic parameters and near-surface blind area characteristic parameters of the defects. The echo spectrum characteristic parameters include the center frequency offset and the bandwidth expansion coefficient. The center frequency offset is the difference between the center frequency of the defect echo and the center frequency of the excitation signal. The bandwidth expansion coefficient is the ratio of the defect echo bandwidth to the excitation signal frequency bandwidth. These spectral characteristics can reflect the geometric shape and surface roughness characteristics of the defects, thereby achieving automatic classification of defect types. The near-surface blind area characteristic parameter is calculated by dividing the difference between the defect echo arrival time and the beginning wave end time by the time length corresponding to the pulse width parameter. When the difference is less than 1.5 times the time length corresponding to the pulse width, it is determined that the defect is located in the near-surface blind area. At this time, the near-surface blind area characteristic parameter value is larger. According to the value range of the near-surface blind area characteristic parameter, different processing strategies are selected. When the value is between 0.7 and 1, the surface wave supplementary detection algorithm is started to perform secondary verification on the defect area. The surface wave supplementary detection algorithm uses a surface wave probe to perform transverse wave detection along the weld surface. The surface wave propagates on the material surface and produces a reflected echo for defects within a depth range of 2mm to 5mm below the surface. This algorithm can cover the near-surface blind area of the conventional ultrasonic detection method, thereby achieving comprehensive detection of defects. When the near-surface blind area characteristic parameter value is between 0.5 and 0.7, the phased array focusing technology is used to enhance detection and confirm the defects. When the value is between 0 and 0.5, it is determined that the defect is located in the conventional detection range and the conventional ultrasonic detection method is used to complete the defect quantitative analysis. The conventional ultrasonic detection method is a method of using a single crystal straight probe to perform longitudinal wave detection on the workpiece interior in a vertical incidence manner. This method is suitable for detecting internal defects with a depth of more than 5mm from the surface. Defect positioning and quantification are completed by analyzing the time domain characteristics of the bottom surface echo and the defect echo.

[0068] The specific implementation of step S09 is to generate a weld defect position distribution map according to the defect identification processing results and defect classification evaluation results of all small weld segments. The distribution map uses a two-dimensional coordinate system to represent the position coordinates of the weld along the line direction on the horizontal axis and the position coordinates of the weld depth direction on the vertical axis. The position coordinates, type identification, and size parameters of each defect are marked in the coordinate system. The type identification includes pore defect identification, crack defect identification, incomplete fusion defect identification, and slag inclusion defect identification. Different types of defects are automatically classified and identified according to the differences in echo characteristic parameters and echo spectrum characteristic parameters. The size parameters include geometric size information such as the length, width, and depth of the defect. The weld quality is evaluated by analyzing the number and distribution of defects in the distribution map. When the proportion of defect detection number to the total number of small weld segments exceeds 30%, it is determined that the weld to be detected has a systematic quality problem. At this time, the welding process parameters need to be reviewed and the weld needs to be re-detected and verified. The distribution map provides an intuitive visual basis for welding quality control and defect repair.

[0069] It should be noted that the key technical idea of the present application is to realize high precision and full coverage of weld detection through multi-aspect innovation synergy. The first key technical idea is an adaptive detection direction optimization method based on scattering attenuation coefficient. The method obtains the scattering attenuation characteristics of different propagation paths in coarse-grained materials through multi-angle scanning, and uses the scattering attenuation coefficient to quantitatively evaluate the sound energy loss degree of each direction, so as to select the detection direction with the smallest scattering. This technology can significantly reduce the signal attenuation and noise interference caused by grain boundary scattering in coarse-grained materials compared with traditional fixed-angle detection, and improve the signal-to-noise ratio and detection sensitivity of defect echoes. The second key technical idea is the synergistic application of phased array focusing technology and high-frequency narrow pulse probe. Phased array focusing realizes the accurate convergence of sound beams in the near-surface blind area by controlling the excitation delay of array elements, and high-frequency narrow pulse realizes the time-domain separation of defect echoes and initial waves by shortening the pulse width. The synergistic effect of the two breaks through the technical bottleneck that the near-surface blind area cannot be effectively detected in traditional ultrasonic detection, and expands the detection depth range to the area 1mm to 4mm below the workpiece surface. The third key technical idea is a layered detection strategy based on near-surface blind area characteristic parameters. The strategy automatically selects phased array focusing enhanced detection, surface wave supplementary detection or conventional ultrasonic detection mode according to the depth position of the defect, and realizes full-depth seamless coverage detection from near-surface to deep defects. The synergy of the three technical ideas lies in building a complete technical chain from detection direction optimization to adaptive selection of detection mode. First, the scattering attenuation coefficient optimization ensures the best propagation condition in coarse-grained materials. Then, the phased array focusing and narrow pulse technology is used to overcome the near-surface blind area limitation. Finally, the layered detection strategy is used to realize targeted detection of defects at different depths. Compared with traditional single-mode detection methods, this synergistic technology system can adapt to the complexity of the microstructure of the weld material and the diversity of the defect distribution, significantly improve the detection rate and positioning accuracy of the weld defects, and provide reliable technical support for the comprehensive evaluation of the welding quality.

[0070] It should be noted that the present application also solves the technical problems that the present application solves the technical problem that defects in the near-surface blind area are difficult to effectively detect. In ultrasonic detection, the range of 2-5 mm below the surface of the workpiece causes the near-surface defect echo signal to overlap with the initial wave in the time domain due to the wide excitation pulse width of the probe and the delay effect of the protective film, and cannot be separated and identified, forming a detection blind area. The present application determines the depth position of the defect by calculating the characteristic parameters of the near-surface blind area, and when the characteristic parameter value is between 0.7 and 1, the surface wave supplementary detection algorithm is started, and the characteristics of the surface wave propagating on the material surface are used to detect the near-surface area by transverse wave, thereby covering the blind area range of the conventional longitudinal wave detection. At the same time, the high-frequency narrow pulse probe is used to control the pulse width in the range of 50-150 ns, shorten the pulse duration so that the near-surface defect echo can be separated from the initial wave in the time domain, and the phased array focusing technology is used to set the acoustic beam focus point at a depth position of 1-4 mm from the surface, thereby improving the detection resolution and defect identification capability of the near-surface area through acoustic energy convergence, thereby realizing effective detection and accurate quantitative analysis of defects in the near-surface blind area.

[0071] Specifically, the principle of the present application is that the present application can solve the fundamental reason for solving the scattering attenuation problem of the coarse-grained weld, which is to establish an adaptive detection parameter optimization mechanism based on scattering characteristics. The scattering degree of coarse-grained materials to ultrasonic waves is closely related to the detection direction and the incident angle. The grain boundary distribution and grain orientation in different directions will cause significant differences in scattering intensity. The present application can identify the detection direction with relatively small scattering influence by collecting a large amount of echo data through multi-directional scanning and calculating the scattering attenuation coefficient. These directions can make the ultrasonic wave penetrate the material with less energy loss and obtain a strong reflection echo from the defect. At the same time, the signal-to-noise ratio parameter is used as the final optimization basis to ensure that the selected direction not only has small scattering but also has low background noise, so that the defect signal can be clearly identified. After completing the parameter optimization, the high-frequency narrow pulse probe can provide better time domain resolution to distinguish the defect echo from the scattering noise, and the phased array focusing technology further improves the detection sensitivity and spatial resolution through acoustic beam convergence. This combination of scattering characteristic analysis, parameter adaptive optimization, high-frequency narrow pulse detection and phased array focusing technology can overcome the influence of coarse-grained material scattering attenuation on detection accuracy from the mechanism.

[0072] A specific embodiment 1 of the present application is provided below, and the specific implementation of each step in embodiment 1 is described in detail as follows.

[0073] In this embodiment, the specific implementation of step S01 is the same as described above, and will not be described in detail here.

[0074] The specific implementation of step S02 is to select the first small welding section as the initial detection object and use a transmitting-receiving separated dual crystal probe for detection. The transmitting-receiving separated dual crystal probe can be a mature product in the market, for example, the DC1, DC2, DC3, DC4 and DC5 series dual crystal probes of the Olympus enterprise, wherein the DC1 and DC2 are 7.5 MHz frequency suitable for near-surface detection, the DC3, DC4 and DC5 are 5 MHz frequency suitable for thin-walled material detection, or the DU series dual crystal probes, the SEB series dual crystal probes and the VSY series angle dual crystal probes of the American Veegoo Teklammer enterprise, or the Olympus D794 type dual crystal probe. These probes are all transmitting-receiving separated structures. In the detection process, the detection direction angle is adjusted from 0 degrees to 90 degrees at a step of 5 degrees. The first echo data is repeatedly collected 5 times at each detection direction angle, and the average value is calculated as the final data of the angle. The average value calculation formula of repeated collection is as follows: In the formula, is the final first echo amplitude value at the detection direction angle , and the unit is ; is the detection direction angle, and the unit is degree, the change range is 0 degree to 90 degree, and the change step is 5 degree; is the repeated collection number, and the default is 5; is the echo amplitude value collected for the th time at the detection direction angle , and the unit is ; is the collection serial number, and the value range is 1 to . The parameter acquisition method is as follows: After the ultrasonic flaw detector excites the probe and receives the reflected echo for the th time at the detection direction angle , the analog-to-digital converter collects it.

[0075] The specific implementation of step S03 is to perform spectral analysis processing on the collected first echo data. The processing converts the time domain echo signal into frequency domain distribution characteristics by using the fast Fourier transform algorithm. The fast Fourier transform is an algorithm for efficiently calculating the discrete Fourier transform. It can convert a discrete time domain signal with a length of into a frequency domain representation, wherein is the sampling point number of the fast Fourier transform, and the value is usually a power of 2, such as 512 to 4096. The main frequency component and the frequency band width parameter are extracted from the frequency domain distribution characteristics. The signal-to-noise ratio parameter corresponding to each detection direction angle is calculated based on the frequency domain characteristics. The signal-to-noise ratio parameter calculation formula is as follows:

[0076] ;

[0077] In the formula, To detect direction angle The signal-to-noise ratio parameter is dimensionless. To detect direction angle Peak amplitude of the defect echo signal, in units of ; The normalized reference amplitude is typically set to a value of [value to be filled in]. ; The root mean square amplitude of the background noise signal, in units of The parameter acquisition method is as follows: By angle The maximum value is obtained by searching the echo data; The method employed was experimental, including step 1: selecting a defect-free area for ultrasonic scanning; step 2: acquiring echo data from this area; and step 3: calculating the root mean square value of the echo data as the background noise amplitude. Simultaneously, echo amplitude parameters at each angle were extracted and compared with standard echo amplitude parameters to calculate the scattering attenuation coefficient. The formula for calculating the scattering attenuation coefficient is as follows:

[0078] ;

[0079] In the formula, To detect direction angle The scattering attenuation coefficient under the given condition is dimensionless. To detect direction angle The actual measured echo amplitude parameters are shown in units of... This parameter is the same as the final first echo amplitude value calculated in step S02; Standard echo amplitude parameters, in units of The parameter acquisition method is as follows: The method employed is experimental, including the following steps: Step 1: Prepare a standard defect test block matching the material type and grain size of the weld to be inspected; Step 2: Perform ultrasonic testing on the standard defect test block using the same probe and excitation parameters; Step 3: Record the received echo amplitude as the standard echo amplitude parameter. Detection direction angles with a scattering attenuation coefficient less than 0.6 are selected to form a candidate direction angle set. This selection condition indicates that the acoustic energy loss in this direction is relatively small and suitable for effective detection.

[0080] The specific implementation of step S04 is to select the detection direction angle with the largest signal-to-noise ratio parameter from the candidate direction angle set as the optimal detection direction angle for the small weld segment, and at the same time record the excitation frequency parameter and pulse width parameter under the optimal detection direction angle. The optimal detection direction angle, probe deflection angle and probe forward tilt angle are combined to form the optimal detection position direction vector group, and the excitation frequency parameter, pulse width parameter and repetition frequency parameter are combined to form the optimal time frequency group.

[0081] The implementation of step S05 is the same as the foregoing, and will not be described in detail here.

[0082] The implementation of step S06 is to perform a second scan detection using a high-frequency narrow pulse probe with a center frequency of 10-20 MHz in combination with phased array focusing technology. The phased array focusing technology realizes the convergence of an ultrasonic beam at a set depth position by precisely controlling the excitation delay time of each array element in the probe array. The calculation formula of the excitation delay time is as follows:

[0083] ;

[0084] In the formula, is the excitation delay time of the i-th array element, and the unit is ; is a time normalization reference value, and the value is usually is the propagation speed of ultrasonic waves in the material, and the unit is , and the value in steel is usually is a speed normalization reference value, and the value is usually is the transverse position coordinate of the i-th array element, and the unit is is the transverse position coordinate of the focal point, and the unit is is the focal depth, and the unit is , and the setting range is 1-4 mm; is a length normalization reference value, and the value is usually is the array element number, and the value range is 1- the total number of array elements of the phased array probe. The phased array probe usually contains 16-128 array elements. The parameter acquisition method is as follows: The position of the i-th array element is determined according to the physical arrangement position of the array element in the probe, and the formula is as follows:

[0085] ;

[0086] In the formula, is the array element spacing, and the unit is , and the empirical value is 0.5-1 mm. The geometric center position of the weld to be detected is set; The near-surface blind zone depth range is set to 1-4 mm. By controlling the excitation delay time of each array element, the sound waves emitted by different array elements are in phase at the focal point position to form a high sound pressure area. The sound beam convergence concentrates sound energy inside the near-surface blind zone, thereby significantly improving the detection resolution and defect recognition ability of the region.

[0087] ​​​The specific implementation of step S07 is a defect identification process on the second echo data, which first extracts the peak amplitude and the bottom echo amplitude of each small welding segment second echo data, calculates the ratio of the peak amplitude to the bottom echo amplitude, compares the ratio with a judgment threshold, and the defect judgment ratio calculation formula is expressed as follows:

[0088] ;

[0089] In the formula, is the defect judgment ratio, dimensionless; is the peak amplitude of the second echo data, with the unit of is the bottom echo amplitude, with the unit of . Wherein, the parameter acquisition method is: obtained by searching for the maximum amplitude in the second echo data; obtained by identifying the amplitude of the echo signal reflected back after the ultrasonic wave propagates to the workpiece bottom, and the signal arrival time is calculated according to the workpiece thickness and the sound speed. When is greater than the judgment threshold 0.25, it is judged that there is a defect in the small welding segment, and the judgment threshold 0.25 is determined by experimental method. In the experimental process, 100 groups of data containing defects and 100 groups of data without defects are collected on the standard weld sample with known defect size, and the ratio of 200 groups of data is statistically analyzed to select the ratio that makes the defect detection rate reach 95% and the false judgment rate is lower than 5% as the judgment threshold.

[0090] The specific implementation of step S08 is to classify and evaluate the identified defects, which calculates the echo spectrum characteristic parameter and the near-surface blind area characteristic parameter of the defect. The echo spectrum characteristic parameter includes the center frequency offset and the bandwidth expansion coefficient, and the center frequency offset calculation formula is expressed as follows:

[0091] ;

[0092] In the formula, is the center frequency offset, dimensionless; is the center frequency of the defect echo, with the unit of is the center frequency of the excitation signal, with the unit of is the frequency normalization reference value, usually taking the value of 1 . Wherein, the parameter acquisition method is: obtained by finding the frequency corresponding to the energy peak in the frequency domain after fast Fourier transform of the defect echo signal; determined by the excitation signal setting parameter of the flaw detector. The bandwidth expansion coefficient calculation formula is expressed as follows:

[0093] ;

[0094] wherein, is a bandwidth extension coefficient, dimensionless; is a frequency band width of the defect echo, unit is is a frequency band width of the excitation signal, unit is is a frequency band width normalized reference value, usually taking 1 . Wherein, the parameter acquisition method is: obtained by calculating the frequency range corresponding to the energy falling to half of the peak energy in the spectrum of the defect echo signal, which is called the half-power bandwidth method; obtained by the same method in the spectrum of the excitation signal. The formula for calculating the near-surface blind zone characteristic parameter is as follows:

[0095] ;

[0096] wherein, is a near-surface blind zone characteristic parameter, dimensionless; is a defect echo arrival time, unit is is a beginning wave end time, unit is is a normalized reference time, usually taking is a time length corresponding to the pulse width parameter, unit is , the setting range is 0.05 to 0.15 . Wherein, the parameter acquisition method is: obtained by identifying the first arrival time of the defect echo signal in the echo data; obtained by identifying the end time of the beginning wave signal generated by the probe excitation pulse; is a width parameter of the probe transmission pulse, determined by the flaw detector setting. When is less than 1.5 times, it is determined that the defect is located in the near-surface blind zone, at this time

[0097] The larger the value is, according to the value range of the near-surface blind zone characteristic parameter, different processing strategies are selected, when the value is between 0.7 and 1, the surface wave supplementary detection algorithm is started to verify the defect area again, when the value is between 0.5 and 0.7, the phased array focusing technology is used to enhance the detection and confirm the defect, when the value is between 0 and 0.5, it is determined that the defect is located in the conventional detection range and the conventional ultrasonic detection method is used to complete the defect quantitative analysis.

[0098] The specific implementation of step S09 is the same as the foregoing, and will not be described in detail here.

[0099] The principles and effects of some formulas in the embodiment are described as follows: the repeated collection average value formula in step S02 is based on the arithmetic mean principle in statistics, and the arithmetic mean of multiple measurement results under the same detection direction angle can effectively suppress the influence of random noise and transient interference, improve the stability and reliability of data, the effect of the formula is to reduce the accidental error that may exist in single measurement, so that the subsequent scattering attenuation coefficient calculation and optimal angle selection are more accurate. The signal-to-noise ratio parameter formula in step S03 is based on the definition principle of signal-to-noise ratio in signal processing, and by comparing the peak amplitude of defect echo signal with the root mean square amplitude of background noise, the distinguishability of defect echo signal relative to noise is quantitatively evaluated, the formula adopts normalization processing to eliminate the influence of dimension, so that the signal-to-noise ratios under different detection direction angles are comparable, when the signal-to-noise ratio is greater than 6, the defect echo can be effectively separated from the noise, the effect of the formula is to provide a quantitative criterion for the selection of optimal detection direction angle, and to ensure that the defect detection under the selected angle has high enough sensitivity and accuracy. The scattering attenuation coefficient formula in step S03 is based on the scattering attenuation mechanism of ultrasonic wave in coarse grain material, by normalizing the comparison between the actual measurement echo amplitude and the standard echo amplitude, the degree of sound energy loss caused by ultrasonic wave scattering at different detection direction angles is quantitatively characterized, and the scattering attenuation coefficient less than 0.6 represents that the scattering loss in this direction is relatively small and the sound energy propagation efficiency is high, the effect of the formula is to establish a screening criterion for detection direction optimization, which can adaptively identify the scattering minimum propagation path in coarse grain material, and significantly improve the penetration ability of ultrasonic wave and the defect detection rate. The excitation delay time formula in step S06 is based on the principles of geometric acoustics and Huygens principle, by calculating the sound path difference of each element to the focal point and converting it into time delay, the sound waves emitted by different elements can reach the preset focal point at the same time after different propagation paths, realizing phase coherent superposition and forming a high sound pressure convergence area, the sound path calculation in the formula uses the Pythagorean theorem to determine the straight line distance from the element position to the focal point , minus the focal point depth The additional sound path relative to the vertical incidence is obtained, and the dimensionless processing is further performed to ensure the dimension uniformity on both sides of the formula. The effect of the formula is to realize the accurate focusing of the ultrasonic beam in the near-surface blind area, and to concentrate the acoustic energy in the depth range of 1 mm to 4 mm, thereby breaking through the detection blind area limitation of the near-surface region caused by the initial wave shielding in the traditional ultrasonic detection, and significantly improving the detection capability and positioning accuracy of the near-surface defects. In step S07, the defect judgment ratio formula is based on the echo amplitude comparison principle. By normalizing and comparing the peak amplitude of the defect echo with the bottom surface echo amplitude, the influence of material attenuation and propagation distance on the amplitude is eliminated, so that the defect judgment is more stable and reliable. The bottom surface echo reflects the overall propagation characteristics of the ultrasonic wave in the material, and the ratio of the defect echo to the bottom surface echo can effectively represent the reflection ability of the defect to the sound wave. The effect of the formula is to establish a unified defect judgment standard. By comparing with the threshold value 0.25 determined by the experiment, automatic defect recognition is realized, and the defect detection rate is ensured to be more than 95% and the misjudgment rate is less than 5%.

[0100] For better understanding and implementation of the present application, the following provides an embodiment 2 of a specific application scenario of the present application:

[0101] The material of the annular weld of a certain pipeline is coarse-grained austenitic stainless steel, the grain size is about 1.8 mm, the weld length is 950 mm, and the wall thickness is 42 mm. Due to the strong scattering effect of ultrasonic waves on coarse-grained materials, it is difficult for traditional detection methods to obtain clear defect echo signals, so the technical team decides to use the ultrasonic detection method of the present application for detection.

[0102] The technical team first uniformly divides the 950 mm long weld into 40 small weld segments along the welding direction, with each small weld segment being 23.75 mm long. The divided small weld segments are uniformly distributed on the surface of the weld. After grinding the surface of each small weld segment, the surface roughness is measured to be 8 μm, and then a layer of water-soluble coupling agent with a thickness of 0.8 mm is applied. The technical team selects a dual-crystal probe with a frequency of 5 MHz. The probe adopts a transmit-receive separated structure, and the thickness of the sound insulation layer between the transmitting wafer and the receiving wafer is 0.5 mm, which can effectively eliminate the initial wave interference.

[0103] The technical team selects the first small weld segment as the initial detection object and performs multi-directional scanning on its surface. The detection direction angle starts from 0 degrees, increases by 5 degrees as a step increment, and gradually increases to 90 degrees, with a total of 19 angle directions of first echo data collected. The echo data is repeatedly collected 5 times at each detection direction angle and the average value is calculated to reduce the influence of random noise. As shown in Figure 2 the signal-to-noise ratio parameters at different detection direction angles show an obvious trend of change, and the signal-to-noise ratio parameters are relatively high in the range of 35 degrees to 55 degrees.

[0104] The technical team performs fast Fourier transform on the collected first echo data to obtain the frequency domain distribution characteristics, and extracts the main frequency component and the frequency band width parameter. The standard echo amplitude parameter measured on the fine-grained standard block is 850 mV, while the actual measured echo amplitude parameter at each detection direction angle of the first small weld segment is distributed between 340 mV and 620 mV. The technical team calculates the scattering attenuation coefficient at each detection direction angle, and divides the actual echo amplitude parameter by the standard echo amplitude parameter to obtain the attenuation ratio value. As shown in Figure 3 the scattering attenuation coefficient at different detection direction angles is distributed in the range of 0.4 to 0.73, and the scattering attenuation coefficient at 9 detection direction angles is less than 0.6, which are 0 degrees, 5 degrees, 10 degrees, 40 degrees, 45 degrees, 50 degrees, 55 degrees, 85 degrees and 90 degrees. These 9 angles constitute the set of alternative direction angles.

[0105] The technical team compared the signal-to-noise ratio (SNR) parameters of various candidate angles and found that the SNR parameter at 45 degrees reached a maximum of 8.7. Therefore, 45 degrees was determined as the optimal detection direction angle for the first small welding segment. Simultaneously, the excitation frequency parameters at this angle were recorded as 5.2MHz, pulse width as 120ns, probe deflection angle as 12 degrees, probe tilt angle as 8 degrees, and repetition frequency as 1000Hz. These parameters constituted the optimal detection position direction vector set and optimal time frequency set for the first small welding segment. The technical team repeated the above procedure for the remaining 39 small welding segments, obtaining the optimal parameter sets for all 40 small welding segments, as shown in Table 1.

[0106] Table 1. Statistical table of optimal detection parameters for some small weld segments

[0107] The technical team discovered that the excitation frequency parameters of weld segments 15 and 25 were below 5MHz, identifying these two segments as areas of strong scattering. Subsequently, the team used a high-frequency narrow-pulse probe with a center frequency of 15MHz, combined with phased array focusing technology, for a second scan inspection. The phased array probe array contains 32 elements. The team calculated the excitation delay time of each element based on a focal depth of 2.5mm from the surface, ensuring that the ultrasonic waves converged within the near-surface blind zone. The pulse width parameter of the high-frequency narrow-pulse probe was set to 100ns, effectively shortening the width of the near-surface blind zone.

[0108] The technical team conducted a second scan based on the optimal parameter settings for each small weld segment, acquiring the second ultrasonic echo data. During defect identification processing of the second echo data, it was found that the ratio of the peak amplitude of the second echo data to the bottom echo amplitude in five small weld segments (8, 9, 10, 23, and 34) exceeded the judgment threshold of 0.25, indicating the presence of defects within these five small weld segments. Defects were detected consecutively in weld segments 8, 9, and 10. The technical team initiated an extended scan mode, merging these three small weld segments into one extended detection area and employing a multi-angle composite scan method for comprehensive inspection. Ultimately, a 68mm long non-fusion defect was identified at this location.

[0109] The technical team categorized and evaluated the identified defects, calculating echo spectrum characteristic parameters and near-surface blind zone characteristic parameters. For defect number 23, the echo arrival time was 3.2 μs, the start and end time was 2.8 μs, and the pulse width parameter corresponded to a time length of 100 ns. The calculated near-surface blind zone characteristic parameter was 0.8. (The text abruptly ends here.) Figure 4As shown, the parameter value is within the range of 0.7 to 1. The technical team initiated a secondary verification using a surface wave supplementary detection algorithm. A surface wave probe was used to perform transverse wave detection along the weld surface. The surface wave propagating on the material surface produced a significant reflected echo of the defect, confirming the presence of a porosity defect near the surface of weld segment 23, with a size of approximately [missing information]. The near-surface blind zone characteristic parameter of defect in weld segment 34 is 0.35. Quantitative analysis of the defect was performed using conventional ultrasonic testing, confirming the presence of a slag inclusion defect with a size of approximately [missing value]. .

[0110] Based on the defect identification and processing results and defect classification assessment results of all small weld segments, the technical team generated a weld defect location distribution map. For example... Figure 5 As shown in the figure, the location coordinates, type identification, and size parameters of five defects are included. The number of defects detected accounts for 12.5% ​​of the total number of small weld segments, which is below the 30% threshold for judging systemic quality problems, indicating that the overall quality of the weld is good. The technical team submitted the inspection report to the quality management department, providing a reliable technical basis for subsequent weld repair and quality assessment.

[0111] This invention represents a significant advancement over traditional ultrasonic testing methods. Traditional methods suffer from low signal-to-noise ratios due to severe scattering attenuation in coarse-grained materials, making accurate identification of defect locations and types difficult. This invention establishes an optimal set of detection parameters through multi-directional scanning, enabling the identification of the optimal signal propagation path in strong scattering environments and effectively reducing the impact of scattering loss on the detection results. Phased array focusing technology concentrates acoustic energy into the near-surface blind zone by controlling the beam convergence position, overcoming the limitations imposed by initial wave interference in traditional methods. The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for inspecting welds using an ultrasonic flaw detector, characterized in that, Includes the following steps: The weld to be inspected is divided into multiple small weld segments along the welding direction and a coupling agent layer is coated on the surface; a dual-crystal probe is used to scan the small weld segments in multiple directions and record the first echo data; The first echo data is subjected to spectral analysis to extract the signal-to-noise ratio (SNR) and echo amplitude parameters. The scattering attenuation coefficient is calculated, and a set of candidate direction angles is selected. The detection direction angle with the largest SNR parameter in the candidate direction angle set is selected as the optimal detection direction angle. The excitation frequency parameter and pulse width parameter are recorded to form the optimal detection position direction vector set and the optimal time frequency set. The above operation is repeated for all small weld segments to obtain their respective optimal detection position direction vector sets and optimal time frequency sets. A high-frequency narrow pulse probe combined with phased array focusing technology is used to perform a second scan according to the optimal detection position direction vector set and the optimal time frequency set to obtain the second echo data. The second echo data is processed for defect identification. The identified defects are classified and evaluated. A weld defect location distribution map is generated. The optimal detection parameters are determined for each small weld segment to reduce the influence of scattering attenuation of coarse-grained materials.

2. The method for inspecting welds using an ultrasonic flaw detector according to claim 1, characterized in that, In the step of dividing the weld to be inspected into multiple small weld segments along the welding direction, the surface roughness of each small weld segment is measured. When the surface roughness exceeds the preset value, the surface is first ground to reduce the surface roughness to below the specified value before the coupling agent layer is applied.

3. The method for inspecting welds using an ultrasonic flaw detector according to claim 2, characterized in that, The calculation process of the scattering attenuation coefficient is to divide the actual measured echo amplitude parameter by the standard echo amplitude parameter under ideal conditions without scattering to obtain the attenuation ratio value. The standard echo amplitude parameter is obtained by measuring on a fine-grained standard test block. The attenuation ratio value reflects the degree of scattering loss of ultrasonic waves by coarse-grained materials.

4. The method for inspecting welds using an ultrasonic flaw detector according to claim 3, characterized in that, The optimal detection position direction vector group includes three angle parameters: optimal detection direction angle, optimal probe deflection angle, and optimal probe forward tilt angle. The optimal time frequency group includes three time-domain and frequency-domain parameters: excitation frequency parameter, pulse width parameter, and repetition frequency parameter.

5. The method for inspecting welds using an ultrasonic flaw detector according to claim 4, characterized in that, The phased array focusing technology controls the excitation delay time of each element in the probe array to make the ultrasonic waves converge at a set depth position.

6. The method for inspecting welds using an ultrasonic flaw detector according to claim 5, characterized in that, The near-surface blind zone characteristic parameters are calculated by dividing the difference between the arrival time of the defect echo and the end time of the initial wave by the time length corresponding to the pulse width parameter. When the difference is less than a preset multiple of the time length corresponding to the pulse width parameter, the defect is determined to be located in the near-surface blind zone.

7. The method for inspecting welds using an ultrasonic flaw detector according to claim 6, characterized in that, In the step of classifying and evaluating the identified defects, when the value of the near-surface blind zone feature parameter is greater than the first threshold, the surface wave supplementary detection algorithm is activated to perform secondary verification of the defect area. The surface wave supplementary detection algorithm uses a surface wave probe to perform transverse wave detection along the weld surface.

8. The method for inspecting welds using an ultrasonic flaw detector according to claim 7, characterized in that, In the step of classifying and evaluating the identified defects, when the value of the near-surface blind zone characteristic parameter is less than the second threshold, the conventional ultrasonic testing method is used to complete the quantitative analysis of the defects. The conventional ultrasonic testing method is to use a single crystal straight probe to perform longitudinal wave detection inside the workpiece in a vertical incident manner.

9. The method for inspecting welds using an ultrasonic flaw detector according to claim 8, characterized in that, The spectrum analysis and processing includes performing a fast Fourier transform on the first echo data to obtain the frequency domain distribution characteristics, and extracting the dominant frequency component and bandwidth parameter from the frequency domain distribution characteristics. The dominant frequency component reflects the influence characteristics of the material grain structure on the ultrasonic wave.

10. The method for inspecting welds using an ultrasonic flaw detector according to claim 9, characterized in that, The signal-to-noise ratio parameter is calculated by dividing the peak amplitude of the defect echo signal by the root mean square amplitude of the background noise signal, which is extracted from the echo data of the defect-free region.

Images