A method and system for ultrasonic testing of weld penetration

Abstract

The present application belongs to the technical field of ultrasonic detection, and particularly relates to a welding seam penetration ultrasonic detection method and system, which comprises the following steps: arranging ultrasonic sensors on both sides of the welding seam, exciting the welding seam area by double-frequency sequential excitation, collecting time domain signals under different frequencies, and extracting corresponding envelope sequences; calculating the energy value of the envelope sequence in a pure noise time window, calculating the scattering index of the current welding seam coarse grain area based on the energy value, separating the envelope sequence in the envelope domain by using the scattering index to obtain the residual envelope after denoising; calculating the energy weighted centroid time of the residual envelope by taking the envelope amplitude square as the weight in the theoretical diffraction wave arrival time window, and combining the extracted upper surface direct wave arrival time to calculate the time difference for penetration inversion. The present application can accurately separate the diffraction signal and eliminate the systematic deviation of the arrival time extraction system.

Description

Technical Field

[0001] This invention belongs to the field of ultrasonic testing technology, specifically relating to an ultrasonic testing method and system for weld penetration depth. Background Technology

[0002] In the manufacturing and in-service testing of load-bearing welded structures such as pressure vessels, offshore platforms, and heavy steel structures, weld penetration depth is a core process quality indicator that directly determines the load-bearing capacity of the joint. Insufficient penetration depth implies the potential presence of incomplete fusion defects, which can easily induce fatigue crack propagation under alternating loads. Therefore, accurate and non-destructive numerical measurement of penetration depth has irreplaceable engineering value.

[0003] In current engineering practice, ultrasonic testing is the mainstream method for numerically measuring weld penetration depth. Among them, the time-of-arrival (TOA) diffraction method is currently the most accurate engineering application solution: An angled probe sensor is placed on each side of the weld, serving as the transmitter and receiver respectively. Utilizing the time difference of arrival between the geometrically diffracted wave generated at the fusion line tip and the direct wave on the upper surface, combined with the known sound velocity and probe spacing, the penetration depth is calculated through geometric inversion. This method has been incorporated into several international nondestructive testing standards.

[0004] However, under conditions such as high-heat-input multi-layer welding and narrow-gap welding, the weld metal will form a coarse-grained austenite or columnar crystal structure, with grain sizes reaching hundreds of micrometers. When ultrasound propagates in such coarse-grained regions, each grain interface will generate random elastic wave scattering, and the superposition of a large number of scattered waves will form continuous structural noise. The time-domain broadening of this structural noise precisely covers the arrival time window of the fusion line diffraction wave, and its local peak amplitude is even higher than that of the diffraction wave itself. The signal processing flow of the traditional diffraction time-of-flight method system relies on amplitude threshold determination of the received signal envelope to lock the arrival time of the diffraction wave. In the context of coarse-grained structural noise, the time corresponding to the maximum amplitude is no longer the true diffraction wave front, but a false peak point formed by the random superposition of scattered noise and diffraction wave. This causes a systematic shift in the extraction of the arrival time, ultimately resulting in uncontrollable errors in the obtained weld depth value. Summary of the Invention

[0005] This invention provides a method and system for ultrasonic testing of weld penetration depth to solve the technical problem of systematic offset in arrival time extraction under the background of noise in coarse-grained weld structures.

[0006] In a first aspect, the present invention provides an ultrasonic testing method for weld penetration depth, comprising the following steps: S1, ultrasonic sensors are arranged on both sides of the weld. The weld area is pulse-excited by dual-frequency sequential excitation, and time-domain signals at different frequencies are collected to extract the corresponding envelope sequence. S2, calculate the energy value of the envelope sequence within the pure noise time window, deduce the scattering index of the current coarse grain region of the weld based on the energy value, and use the scattering index to separate the envelope sequence in the envelope domain to obtain the denoised residual envelope; S3. Within the theoretical diffracted wave arrival time window, the energy-weighted centroid time of the residual envelope is calculated with the square of the envelope amplitude as the weight, and the time difference used for melt depth inversion is calculated in combination with the arrival time of the extracted upper surface through wave. S4. Extract the arrival time of the bottom reflected wave from the received signal to calculate the current real-time sound velocity. Substitute the current real-time sound velocity and time difference into the geometric inversion formula to calculate and output the current weld penetration value.

[0007] Its effect is that by using dual-frequency sequential excitation and envelope domain scattering separation, the interference of coarse-grained structure noise on the diffraction signal is suppressed, the signal-to-noise ratio of the diffraction signal in the high-frequency residual envelope is improved, and the diffraction wave characteristics that were originally submerged by noise are clearly presented.

[0008] Furthermore, pulse excitation of the weld region is performed using dual-frequency sequential excitation, and time-domain signals at different frequencies are acquired to extract the corresponding envelope sequences, including: Switch the transmission frequencies sequentially without moving the probe; The same weld area was sequentially pulsed with low frequency and high frequency, and the corresponding low-frequency time domain signal and high-frequency time domain signal were collected. The low-frequency time-domain signal and the high-frequency time-domain signal are respectively subjected to analytical signal extraction, and finally their respective envelope sequences are extracted.

[0009] Its effect is as follows: by limiting the probe to a fixed position, low-frequency and high-frequency signals are sequentially excited to the same weld area and the envelope is extracted by obtaining analytical signals respectively, ensuring that the probe coupling, sound path, and detection area of ​​the two acquisitions are completely consistent, with only the excitation frequency being different. This eliminates the additional sound path interference introduced by probe displacement and repeated coupling, and provides two sets of original time-domain envelope data that can be directly compared without additional variable interference for subsequent differentiation of noise and diffraction signals based on frequency scattering characteristics, thus ensuring the consistency of the basic data for scattering index calculation.

[0010] Furthermore, the energy value of the envelope sequence is calculated within a pure noise time window, and the scattering index of the current coarse-grained region of the weld is inferred based on the energy value, including: Within a pure noise time window earlier than the theoretical earliest arrival time of the diffracted wave, the root mean square energy values ​​of the two envelope sequences are calculated respectively. Calculate the measured scattering ratio of the two root-mean-square energy values; Based on the relationship between the measured scattering ratio and the theoretical frequency, the scattering index of the coarse-grained region of the current weld was finally calculated.

[0011] Its effect is that by calculating the scattering index within a pure noise window, without the need for any pre-setting of material parameters, this method can be adaptively applied to weld structures with different coarse grain levels, and has good versatility in working conditions.

[0012] Furthermore, the envelope sequence is separated in the envelope domain using the scattering index to obtain the denoised residual envelope, including: Multiplying the low-frequency envelope sequence by a frequency scaling factor constructed based on the scattering index yields a reference estimate of the high-frequency structural noise envelope. The reference estimate is subtracted from the high-frequency envelope sequence to obtain the denoised residual envelope.

[0013] Furthermore, within the theoretical diffracted wave arrival time window, the energy-weighted centroid time of the residual envelope is calculated using the square of the envelope amplitude as the weight, including: The boundaries of the theoretical diffraction wave arrival time window are determined based on the probe spacing, nominal sound velocity, and weld plate thickness. Within the theoretical diffracted wave arrival time window, the square of the residual envelope amplitude is used as a weighting factor. By performing weighted statistics on the sampling point times, the energy-weighted centroid time of the residual envelope is finally calculated.

[0014] Its effect is that by replacing the single-point peak value with the energy-weighted centroid as the positioning basis for arrival time, the sensitivity of the single-point amplitude extreme value to residual random noise is eliminated, and the time extraction result has stronger stability to residual scattering disturbance.

[0015] Furthermore, by combining the arrival times of the extracted upper surface through-wave, the time difference used for melt depth inversion is calculated, including: The arrival time of the direct wave on the upper surface is extracted at the front end of the receiving signal and used as the acoustic path reference point. The difference between the energy-weighted centroid time and the arrival time of the direct wave on the upper surface is calculated. By simultaneously eliminating trigger delay and cable delay errors, the time difference used for melt depth inversion is finally calculated.

[0016] Furthermore, the arrival time of the bottom-reflected wave in the received signal is extracted to calculate the current real-time speed of sound, including: After each excitation acquisition, the arrival time of the bottom surface reflected wave generated by the bottom surface of the weldment is synchronously extracted from the received signal; Obtain the nominal thickness of the weldment and the trigger time of the emission pulse; Based on the nominal thickness of the weldment, the arrival time of the bottom reflected wave, and the trigger time of the emission pulse, the current real-time speed of sound is finally calculated.

[0017] Furthermore, by substituting the current real-time sound velocity and time difference into the geometric inversion formula, the current weld penetration value is calculated and output, including: A geometric inversion formula is constructed by performing coordinate transformation based on the mid-plane of the plate thickness. Replace the fixed sound speed with the current real-time sound speed, and substitute the time difference into the geometric inversion formula; Perform successive data acquisition and real-time compensation, and finally calculate and output the current weld penetration value.

[0018] Its effect is that by calculating the sound velocity in real time through the bottom surface reflected wave, the influence of sound velocity temperature drift is incorporated into the point-by-point compensation process of each acquisition, so that the melting depth inversion result does not depend on the preset fixed sound velocity parameter, thus improving the reliability of the inversion result under the on-site detection conditions.

[0019] Furthermore, after calculating and outputting the current weld penetration value, the process also includes: synchronously archiving the penetration value, time difference, current real-time sound velocity, and scattering index; ultimately forming a dataset for process quality assessment and traceability.

[0020] Secondly, the present invention provides an ultrasonic testing system for weld penetration depth, including a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned ultrasonic testing method for weld penetration depth is implemented.

[0021] The beneficial effects are: by acquiring time-domain signals at different frequencies through dual-frequency sequential excitation and calculating the scattering index in the envelope domain, effective separation of noise from coarse-grained structures is achieved. This process does not rely on preset material parameters, can adapt to weld microstructures with different degrees of coarseness, and improves the universality and anti-interference ability of the detection method.

[0022] Meanwhile, this invention uses energy-weighted centroid time instead of traditional single-point peak positioning, and combines real-time sound velocity back calculation for point-by-point compensation, eliminating the sensitivity of single-point extreme values ​​to random noise and the error caused by temperature drift, thus improving the accuracy and reliability of engineering testing. Attached Figure Description

[0023] Figure 1 This is a flowchart of the ultrasonic testing method for weld penetration depth in this invention.

[0024] Figure 2 This is a line graph comparing the root mean square error of the two methods for melting depth inversion under different coarse grain noise intensities in this invention. Detailed Implementation

[0025] 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.

[0026] An embodiment of the ultrasonic testing method for weld penetration depth provided by this invention: like Figure 1 As shown, the ultrasonic testing method for weld penetration depth includes the following steps: S1, extract the envelope sequence using dual-frequency sequential excitation.

[0027] In an optional embodiment, broadband phased array ultrasonic sensors are arranged on both sides of the weld, each acting as both the transmitter and receiver. Without moving or re-contacting the probes, the system controller sequentially switches the transmission frequencies, using... and Pulse excitation is applied to the same weld area. For example, The value can be 2.25MHz. The value can be 5MHz. The frequency switching time interval between two excitations does not exceed the device's minimum switching cycle. This switching time interval is typically on the order of milliseconds to ensure that the geometric path conditions, probe contact state, and beam coverage area are exactly the same in both acquisitions, with only the excitation frequency differing. The receiving sensor acquires the corresponding A-scan time-domain signal, denoted as the low-frequency time-domain signal. and high-frequency time domain signals This serves as the foundational data for subsequent processing.

[0028] Furthermore, according to Rayleigh scattering theory, when the ultrasonic wave length is much larger than the grain size, the scattering intensity of the ultrasonic wave by a single grain is... With excitation frequency It is proportional to the fourth power, that is, it satisfies Therefore, the ratio of structural noise intensity under low-frequency to high-frequency conditions theoretically follows... The deterministic relationship. Taking this parameter value as an example, the ratio is approximately This indicates that the structural noise energy carried by the high-frequency signal is approximately 24 times that of the low-frequency signal. In contrast, the geometric diffraction signal generated at the weld fusion line tip belongs to Huygens wavelet radiation, and its arrival time is determined by both the sound path and the sound velocity, showing complete consistency at both frequencies; the amplitude of the geometric diffraction signal varies with frequency in a similar manner to... The laws of frequency dependence differ significantly, with the slope of change being considerably smaller. This fundamental difference in frequency dependence between the two signals forms the physical basis for subsequent separation of scattering noise and diffraction signals in the envelope domain.

[0029] Next, the low-frequency time-domain signal and high-frequency time domain signals Perform analytic signal extraction separately, i.e., perform Hilbert transform to obtain the analytic signal, and extract their respective envelope sequences. and It should be noted that there are many other types of sensors in this field capable of achieving broadband ultrasonic excitation. Those skilled in the art can select different probe models and frequency combinations based on the actual weld conditions, plate thickness range, and equipment conditions, and are not limited to the typical parameter values ​​mentioned above.

[0030] By completing dual-frequency sequential excitation and signal acquisition under the condition that the probe position is completely fixed, two original A-scan signal pairs with strictly consistent geometric sound path conditions and different frequencies were obtained. This provides reliable basic data for subsequent calculation of the current coarse grain region scattering characteristics of the weld based on frequency dependence, and fundamentally avoids the sound path consistency error introduced by two independent coupling operations.

[0031] S2, extracting the residual envelope based on scattering index separation.

[0032] In an optional embodiment, the root mean square energy values ​​of the two envelope sequences are calculated within a pure noise time window that does not contain fusion line geometric diffraction waves, i.e., a time period earlier than the theoretically earliest arrival time of the diffraction wave:

[0033] in This represents the total number of sampling points within the pure noise window. For the first The time of each sampling point. The ratio of the root mean square energy values ​​of the two paths is defined as the measured scattering ratio. .

[0034] Furthermore, based on the measured scattering ratio Theoretical frequency dependence of Rayleigh scattering Calculate the actual grain scattering index of the coarse-grained region of the current weld:

[0035] The closer the value is to 4, the smaller the grain size is to the wavelength, and the closer the scattering behavior is to ideal Rayleigh scattering; the more significantly the value deviates from 4, the greater the true deviation in reflecting the dependence of the current coarse-grained structure on the scattering frequency. The value is directly determined by the root mean square energy ratio within the measured pure noise window, without the need for any pre-setting of material parameters or empirical coefficients.

[0036] Next, using the obtained scattering index The structure noise component of the high-frequency signal is separated in the envelope domain. The low-frequency envelope sequence... Multiply by a frequency scaling factor constructed based on the scattering index A reference estimate of the high-frequency structure noise envelope is obtained. :

[0037] Then from the high-frequency envelope sequence Subtract the reference estimate from the middle Finally, the denoised residual envelope is obtained:

[0038] Because the structure noise envelope follows the same principle in both signals. The law of proportional correction and subtraction can effectively cancel each other out; however, the envelope amplitude of geometric diffraction signals does not follow this law. The law is preserved after subtraction, and the signal-to-noise ratio is improved compared to the original. Significant improvement was achieved in residual envelope. It can be directly used as the input for subsequent centroid localization, without the need to perform Hilbert transform again.

[0039] By calculating the scattering index from measured data and performing proportional subtraction in the envelope domain, the diffraction signal that was originally submerged in the noise of the coarse-grained structure is effectively separated, and the diffraction wave characteristics in the residual envelope are clearly presented, providing high signal-to-noise ratio input data for subsequent accurate time positioning.

[0040] S3, calculate the time difference at the centroid moment.

[0041] In an optional embodiment, within the theoretical diffracted wave arrival time window Inner residual envelope Process it. and From the known probe spacing Nominal speed of sound and weld plate thickness The upper and lower limits of the process are calculated together, and the window boundary is directly determined by the process parameters, without relying on manual settings.

[0042] Furthermore, within the theoretical arrival time window of the diffracted wave, instead of directly using the point of maximum amplitude as the arrival time, the square of the amplitude of the residual envelope is used as a weighting factor to calculate the energy-weighted centroid time of the residual envelope. :

[0043] in For the first The time corresponding to each sampling point The sampling interval is... and These are the sampling point numbers corresponding to the two endpoints of the time window.

[0044] Next, the arrival time of the direct-through wave on the upper surface is extracted at the receiving signal front end. This serves as the reference point for the sound path. The moment of the energy-weighted center of mass is... Arrival time of the direct wave on the upper surface The difference calculation is performed to synchronously eliminate systematic errors such as trigger delay and cable delay errors. The final time difference used for melt depth inversion is:

[0045] By replacing single-point peak values ​​with energy-weighted centroids as the location basis for the arrival time of diffracted waves, the sensitivity of single-point amplitude extrema to residual random noise is eliminated, and the time extraction results can remain stable under residual scattering disturbances, providing a time difference input with stronger anti-interference capability for subsequent melting depth inversion.

[0046] S4 calculates the current real-time sound velocity to output the melting depth value.

[0047] In an optional embodiment, a residual temperature gradient may exist in the weldment temperature at the welding site during the detection process, leading to an increase in the longitudinal wave velocity in the steel. Deviation from nominal room temperature For example, the nominal sound velocity of longitudinal waves in steel can be taken as 5900 m / s. To eliminate the influence of sound velocity temperature drift on the accuracy of melt depth inversion, after each excitation acquisition, the arrival time of the bottom surface reflected wave generated by the bottom surface of the weldment is synchronously extracted from the received signal. The nominal plate thickness of the weldment is obtained from the mechanical process drawings. and the trigger time of the transmission pulse Based on the nominal thickness of the weldment, the arrival time of the reflected wave from the bottom surface, and the trigger time of the emission pulse, the current real-time speed of sound is finally calculated:

[0048] in The current real-time sound velocity is calculated from the measured bottom surface echo, reflecting the actual acoustic parameters of the weld area under the current temperature conditions.

[0049] Furthermore, a geometric inversion formula is constructed by performing coordinate transformation based on the mid-plane of the plate thickness. This scheme employs a symmetrical dual-probe arrangement, with the two probes located on either side of the weld centerline, each at a horizontal distance of [missing information]. The diffraction point is located at the root of the fusion line near the lower surface of the weldment. Taking the upper surface of the weldment as the zero point, the original form of the standard diffraction time-of-flight method geometric formula is:

[0050] in The depth of the diffraction point from the upper surface. Let be the time difference between the diffracted wave and the direct wave. Since the detection target is the remaining unmelted depth from the root of the fusion line to the lower surface of the weldment, a coordinate transformation is performed using the mid-plane of the plate thickness as a reference, and the above formula is rewritten as a geometric inversion formula:

[0051] Current real-time speed of sound Replace fixed speed of sound And combined with time difference Substituting the values ​​into the geometric inversion formula, performing successive acquisitions and real-time compensation, and finally calculating and outputting the current weld penetration value. .because After each acquisition, the results are updated by back-calculation from the bottom reflected wave. The influence of sound velocity temperature drift on the melt depth inversion results is incorporated into the point-by-point compensation process, and the melt depth inversion results no longer depend on the preset fixed sound velocity parameters.

[0052] Next, after calculating and outputting the current weld penetration value, the system will also output the penetration value. Time difference Current real-time speed of sound and scattering index Synchronous archiving is performed to ultimately form a dataset for process quality assessment and traceability. It should be noted that there are many signal processing methods in this field capable of extracting bottom-surface reflected waves and calculating sound velocity; those skilled in the art can choose different implementation methods based on the actual plate thickness range, sampling rate, and on-site temperature conditions.

[0053] In this way, by calculating the sound velocity in real time through the bottom reflected wave and substituting it into the melt depth inversion formula, the influence of sound velocity temperature drift is incorporated into the point-by-point compensation process of each acquisition, so that the melt depth inversion results can remain reliable under the temperature fluctuation conditions detected on site.

[0054] like Figure 2 As shown, it can be observed that with the increase of noise intensity, the error curve of the existing method shows an obvious nonlinear growth trend, while the error curve of the method of the present invention remains at a low level and the increase is gradual, indicating that the present invention has more stable inversion performance under strong coarse-grained noise interference.

[0055] Embodiments of the ultrasonic weld penetration testing system provided by this invention: The ultrasonic weld penetration testing system includes a processor and a memory. The memory stores computer program instructions, which are executed by the processor to implement the ultrasonic weld penetration testing method described above.

[0056] The ultrasonic testing system for weld penetration depth also includes other components well known to those skilled in the art, such as communication interfaces. Their setup and functions are known in the art and will not be described in detail here.

[0057] 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 method for ultrasonic testing of weld penetration depth, characterized in that, Includes the following steps: S1, ultrasonic sensors are arranged on both sides of the weld. The weld area is pulse-excited by dual-frequency sequential excitation, and time-domain signals at different frequencies are collected to extract the corresponding envelope sequence. S2, calculate the energy value of the envelope sequence within the pure noise time window, deduce the scattering index of the current coarse grain region of the weld based on the energy value, and use the scattering index to separate the envelope sequence in the envelope domain to obtain the denoised residual envelope; S3. Within the theoretical diffracted wave arrival time window, the energy-weighted centroid time of the residual envelope is calculated with the square of the envelope amplitude as the weight, and the time difference used for melt depth inversion is calculated in combination with the arrival time of the extracted upper surface through wave. S4. Extract the arrival time of the bottom reflected wave from the received signal to calculate the current real-time sound velocity. Substitute the current real-time sound velocity and time difference into the geometric inversion formula to calculate and output the current weld penetration value.

2. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, The weld region is pulse-excited using dual-frequency sequential excitation. Time-domain signals at different frequencies are acquired, and the corresponding envelope sequences are extracted, including: Switch the transmission frequencies sequentially without moving the probe; The same weld area was sequentially pulsed with low frequency and high frequency, and the corresponding low-frequency time domain signal and high-frequency time domain signal were collected. The low-frequency time-domain signal and the high-frequency time-domain signal are respectively subjected to analytical signal extraction, and finally their respective envelope sequences are extracted.

3. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, Calculate the energy value of the envelope sequence within a pure noise time window, and deduce the scattering index of the current coarse-grained region of the weld based on the energy value, including: Within a pure noise time window earlier than the theoretical earliest arrival time of the diffracted wave, the root mean square energy values ​​of the two envelope sequences are calculated respectively. Calculate the measured scattering ratio of the two root-mean-square energy values; Based on the relationship between the measured scattering ratio and the theoretical frequency, the scattering index of the coarse-grained region of the current weld was finally calculated.

4. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, The envelope sequence is separated in the envelope domain using the scattering index to obtain the denoised residual envelope, including: Multiplying the low-frequency envelope sequence by a frequency scaling factor constructed based on the scattering index yields a reference estimate of the high-frequency structural noise envelope. The reference estimate is subtracted from the high-frequency envelope sequence to obtain the denoised residual envelope.

5. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, Within the theoretical diffracted wave arrival time window, the energy-weighted centroid time of the residual envelope is calculated using the square of the envelope amplitude as the weight, including: The boundaries of the theoretical diffraction wave arrival time window are determined based on the probe spacing, nominal sound velocity, and weld plate thickness. Within the theoretical diffracted wave arrival time window, the square of the residual envelope amplitude is used as a weighting factor. By performing weighted statistics on the sampling point times, the energy-weighted centroid time of the residual envelope is finally calculated.

6. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, Based on the arrival time of the extracted upper surface through wave, the time difference used for melt depth inversion is calculated, including: The arrival time of the direct wave on the upper surface is extracted at the front end of the receiving signal and used as the acoustic path reference point. The difference between the energy-weighted centroid time and the arrival time of the direct wave on the upper surface is calculated. By simultaneously eliminating trigger delay and cable delay errors, the time difference used for melt depth inversion is finally calculated.

7. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, Extracting the arrival time of the bottom-reflected wave from the received signal to inversely calculate the current real-time speed of sound includes: After each excitation acquisition, the arrival time of the bottom surface reflected wave generated by the bottom surface of the weldment is synchronously extracted from the received signal; Obtain the nominal thickness of the weldment and the trigger time of the emission pulse; Based on the nominal thickness of the weldment, the arrival time of the bottom reflected wave, and the trigger time of the emission pulse, the current real-time speed of sound is finally calculated.

8. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, Substituting the current real-time sound velocity and time difference into the geometric inversion formula, the current weld penetration value is calculated and output, including: A geometric inversion formula is constructed by performing coordinate transformation based on the mid-plane of the plate thickness. Replace the fixed sound speed with the current real-time sound speed, and substitute the time difference into the geometric inversion formula; Perform successive data acquisition and real-time compensation, and finally calculate and output the current weld penetration value.

9. The ultrasonic testing method for weld penetration depth according to claim 1, characterized in that, After calculating and outputting the current weld penetration value, the process also includes: synchronously archiving the penetration value, time difference, current real-time sound velocity, and scattering index; ultimately forming a dataset for process quality assessment and traceability.

10. An ultrasonic testing system for weld penetration depth, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the ultrasonic testing method for weld penetration depth as described in any one of claims 1-9 is implemented.

Images