Ultrasonic phased array based high precision non-destructive testing system for welds

Abstract

The application relates to the technical field of nondestructive testing, in particular to a high-precision nondestructive testing system for weld seams based on an ultrasonic phased array, which comprises an environment sensing and parameter acquisition module, a sensor integrated in a probe wedge is used to collect temperature data of the probe wedge and a workpiece surface, and temperature monitoring parameters containing wedge temperature values and workpiece surface temperature values are generated; an acoustic model dynamic solving module is used to calculate the sound velocity values of the wedge material and the weld seam material to be measured at the current temperature and the wedge delay value based on the temperature monitoring parameters and in combination with a preset material temperature change characteristic model, and dynamic sound velocity and zero point offset compensation parameters are generated. In the application, a real-time temperature acquisition and sound velocity thermal sensitivity coefficient correction mechanism is introduced in the acoustic model dynamic solving module, the scheme can dynamically update the wedge sound velocity and the weld seam sound velocity under high-temperature working conditions, systematic deviation of propagation time caused by sound velocity drift with temperature is avoided, and the physical accuracy of the acoustic model is stabilized from the source.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and in particular to a high-precision nondestructive testing system for welds based on ultrasonic phased array. Background Technology

[0002] In ultrasonic phased array testing of critical structures such as welds, the propagation velocity of ultrasonic waves in the wedge and workpiece is a fundamental parameter for focusing law calculation and imaging positioning. Its accuracy directly determines the focusing position of the sound beam and the precision of defect location. In actual industrial testing scenarios, especially under high-temperature post-weld testing or online testing conditions, the wedge and workpiece are often in a non-constant temperature environment, and the sound velocity exhibits significant thermosensitive characteristics with temperature changes. Therefore, it is necessary to effectively correct the acoustic parameters during the testing process to ensure the reliability of the test results.

[0003] Existing technologies typically use static sound velocity parameters based on a fixed sound velocity or a single reference temperature to calculate focusing rules. This fails to reflect the thermal drift effect caused by the change in sound velocity with temperature under high-temperature conditions in real time. As a result, there is a systematic deviation between the actual propagation time of the sound wave and the theoretical calculation value, which causes the focus point position to shift and the imaging distortion, thus reducing the accuracy and stability of defect detection. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a high-precision non-destructive testing system for welds based on ultrasonic phased arrays. It aims to improve the problem that existing technologies typically use static sound velocity parameters based on fixed sound velocity or a single reference temperature for focusing law calculations, which fail to reflect the thermal drift effect caused by sound velocity changes with temperature under high-temperature conditions in real time.

[0005] This invention provides the following technical solution: a high-precision non-destructive testing system for welds based on an ultrasonic phased array, comprising:

[0006] The environmental perception and parameter acquisition module collects temperature data of the probe wedge and the workpiece surface through sensors integrated in the probe wedge, and generates temperature monitoring parameters including the wedge temperature value and the workpiece surface temperature value.

[0007] The acoustic model dynamic solution module, based on the temperature monitoring parameters and combined with the preset material temperature change characteristic model, calculates the sound velocity value and wedge delay value of the wedge material and the weld material to be tested at the current temperature, and generates dynamic sound velocity and zero-point offset compensation parameters.

[0008] The phase law calculation module calculates the excitation delay time of each array chip based on the dynamic sound speed and zero-point offset compensation parameters and according to Snell's law, and generates phase correction excitation timing instructions.

[0009] The array excitation and echo acquisition module, based on the phase correction excitation timing command, drives the multi-channel piezoelectric crystal of the phased array probe to emit ultrasonic waves and receive reflected waves, converting physical sound waves into electrical signals and generating multi-channel raw echo time-domain signals.

[0010] The image reconstruction and geometric correction module, based on the multi-channel original echo time-domain signal and the dynamic sound velocity and zero-point offset compensation parameters, uses the dynamic sound velocity and zero-point offset compensation parameters to calculate the spatial coordinates of the echo signal, maps the time-domain signal into spatial location distribution data, and generates geometrically corrected imaging data.

[0011] The defect feature extraction and evaluation module extracts the geometric features of areas where the amplitude intensity exceeds a preset threshold based on the geometrically corrected imaging data, and generates a weld quality inspection and evaluation result that includes defect depth location data and geometric size data.

[0012] Preferably, in the environmental sensing and parameter acquisition module, the step of generating temperature monitoring parameters including wedge temperature values ​​and workpiece surface temperature values ​​includes:

[0013] The system acquires a first analog voltage signal output by a first temperature sensor embedded inside the probe wedge substrate, and acquires a second analog voltage signal output by a second temperature sensor located at the bottom of the wedge.

[0014] The first analog voltage signal and the second analog voltage signal are subjected to analog-to-digital conversion and low-pass filtering to generate the original digital temperature sampling sequence.

[0015] The pre-stored sensor voltage-temperature characteristic calibration curve is invoked to map the original digital temperature sampling sequence into physical temperature scale values, and the wedge temperature value and workpiece surface temperature value are calculated respectively.

[0016] The wedge temperature value and the workpiece surface temperature value are encapsulated in a preset data format to generate temperature monitoring parameters.

[0017] Preferably, in the acoustic model dynamic solution module, the step of calculating the sound velocity and wedge delay values ​​of the wedge material and the weld material to be tested at the current temperature by combining a preset material temperature change characteristic model includes:

[0018] Read the first reference sound velocity and first sound velocity temperature coefficient of the wedge material, and the second reference sound velocity and second sound velocity temperature coefficient of the weld material to be tested from the system's preset storage unit.

[0019] Using the wedge temperature value in the temperature monitoring parameters, a linear correction calculation is performed based on the first reference sound velocity and the first sound velocity temperature coefficient to generate the current wedge sound velocity value.

[0020] Using the workpiece surface temperature value in the temperature monitoring parameters, a linear correction calculation is performed based on the second reference sound velocity and the second sound velocity temperature coefficient to generate the current weld sound velocity value.

[0021] The pre-stored wedge geometric sound path data is called, the wedge geometric sound path data is divided by the current wedge sound velocity value, the propagation time of the ultrasonic wave inside the wedge is calculated, and the current wedge delay value is generated.

[0022] The current wedge block sound velocity value, the current weld sound velocity value, and the current wedge block delay value are correlated and encoded to generate dynamic sound velocity and zero-point offset compensation parameters.

[0023] Preferably, the step of reading the first reference sound velocity and first sound velocity temperature coefficient of the wedge material, and the second reference sound velocity and second sound velocity temperature coefficient of the weld material to be tested from the system preset storage unit includes:

[0024] The system receives the wedge block model identification code and workpiece material type identification code input by the user through the human-computer interaction interface;

[0025] Access the pre-built acoustic material property database within the system, and use the wedge block model identification code and the workpiece material type identification code as index keys for matching and retrieval;

[0026] Extract the corresponding room temperature sound velocity calibration value and the sound velocity-temperature change slope value from the matching records in the database;

[0027] The extracted room temperature sound velocity calibration values ​​were assigned as the first reference sound velocity and the second reference sound velocity, respectively. The extracted sound velocity-temperature change slope values ​​were assigned as the first sound velocity temperature coefficient and the second sound velocity temperature coefficient, respectively.

[0028] Preferably, in the phase law calculation module, the step of calculating the excitation delay time of each array wafer based on Snell's law and generating phase correction excitation timing instructions includes:

[0029] Based on the preset scanning strategy, a series of virtual focusing target points are discretized and generated within the weld area to be tested, and their spatial coordinates are generated.

[0030] For each virtual focusing target point, using Fermat's principle and the current wedge sound velocity value and the current weld sound velocity value in the dynamic sound velocity and zero-point offset compensation parameters, the acoustic propagation path conforming to Fermat's principle from the center point of each wafer in the array to the virtual focusing target point is calculated. The acoustic propagation path conforming to Fermat's principle satisfies Snell's law of refraction at the interface between the wedge and the workpiece.

[0031] Calculate the absolute transit time required for the ultrasonic wave to propagate along each of the acoustic propagation paths that conform to Fermat's principle. The absolute transit time is the sum of the path length of the sound wave in the wedge divided by the current sound velocity value of the wedge and the path length of the sound wave in the workpiece divided by the current sound velocity value of the weld.

[0032] The peak value of the absolute transit time of all wafers corresponding to the same virtual focus target point is selected as the reference time. The difference between the absolute transit time of each wafer and the reference time is calculated to obtain the relative excitation delay time of each wafer.

[0033] The relative excitation delay times of all wafers are encoded in the array arrangement order to generate phase-corrected excitation timing instructions.

[0034] Preferably, in the array excitation and echo acquisition module, the step of generating multi-channel raw echo time-domain signals includes:

[0035] The phase correction excitation timing command is analyzed, and the multi-channel high-voltage pulse transmitting circuit is controlled to apply high-frequency high-voltage excitation pulses to each piezoelectric crystal of the phased array probe according to the delay time set in the command;

[0036] The piezoelectric crystal receives the ultrasonic mechanical wave reflected from the weld seam under test, and uses the piezoelectric effect to convert the ultrasonic mechanical wave into a continuously changing analog echo voltage signal.

[0037] The analog echo voltage signal of each channel is pre-amplified and bandpass filtered;

[0038] The processed analog echo voltage signal is synchronously sampled and quantized using an analog-to-digital converter to generate a discrete digital echo sequence;

[0039] The discretized digital echo sequences of all channels are encapsulated in parallel according to the channel index to generate multi-channel original echo time-domain signals.

[0040] Preferably, in the image reconstruction and geometric correction module, the step of generating geometrically corrected imaging data includes:

[0041] A two-dimensional imaging grid is constructed to cover the area of ​​the weld to be tested. The two-dimensional imaging grid consists of several pixels with independent spatial coordinates.

[0042] For each pixel, the current wedge sound velocity value, current weld sound velocity value, and current wedge delay value in the dynamic sound velocity and zero-point offset compensation parameters are used to calculate the two-way propagation time of the ultrasonic wave from each chip in the array to the pixel and back, and generate a dynamic focusing delay rule table.

[0043] Based on the dynamic focusing delay rule table, the instantaneous amplitude data corresponding to each channel is extracted from the multi-channel original echo time-domain signal, and the extracted data is coherently superimposed and summed to calculate the synthetic sound intensity value of the pixel.

[0044] The synthetic sound intensity value is calculated by traversing all pixels, and the synthetic sound intensity values ​​of all pixels are arranged according to the spatial coordinates of the pixels to generate geometrically corrected imaging data.

[0045] Preferably, in the defect feature extraction and evaluation module, the step of generating weld quality inspection and evaluation results that include defect depth location data and geometric dimension data includes:

[0046] The geometrically corrected imaging data is subjected to spatial smoothing filtering to generate an enhanced imaging matrix;

[0047] A threshold for amplitude determination is set, and pixels in the enhanced imaging matrix whose sound intensity values ​​are lower than the threshold are set to zero, while pixels whose sound intensity values ​​are not less than the threshold are retained, thus generating a binarized feature map.

[0048] A connected component labeling algorithm is used to perform cluster analysis on the non-zero pixels in the binarized feature map, and adjacent connected non-zero pixels are identified as independent defect candidate targets.

[0049] For each of the defect candidate targets, the depth coordinates of the preset feature points of the defect candidate target in the imaging grid coordinate system are extracted as the defect depth position data, and the span of the defect candidate target in the horizontal and depth directions is calculated as the geometric dimension data.

[0050] The defect depth location data and the geometric dimension data are structured and encapsulated to output the weld quality inspection and evaluation results.

[0051] The present invention has the following beneficial effects:

[0052] 1. In this invention, by introducing a real-time temperature acquisition and sound velocity thermal sensitivity coefficient correction mechanism into the dynamic solution module of the acoustic model, this solution can dynamically update the sound velocity of the wedge and the sound velocity of the weld under high temperature conditions, avoiding systematic deviations in propagation time caused by sound velocity drift with temperature, and stabilizing the physical accuracy of the acoustic model from the source.

[0053] 2. In this invention, the phase law calculation module directly uses the temperature-corrected current wedge sound velocity value and the current weld sound velocity value during the path solving process, and recalculates the refraction point position based on Fermat's principle and Snell's law, so that the refraction angle of the ultrasonic wave at the interface between the wedge and the workpiece can be adaptively adjusted with temperature changes, avoiding the incident path deflection caused by sound velocity mismatch.

[0054] 3. In this invention, by uniformly incorporating the dynamic sound velocity parameter and the wedge zero-point delay compensation parameter into the calculation process of the absolute transit time and the relative excitation delay time, the focusing law is ensured to be consistent with the actual sound path, so that the focusing position of the ultrasonic wave inside the weld is highly consistent with the calculation model, thereby significantly reducing the systematic positioning error in the defect depth calculation. Attached Figure Description

[0055] Figure 1 This is an architecture diagram of the high-precision non-destructive testing system for welds based on ultrasonic phased array proposed in this invention. Detailed Implementation

[0056] The technical solutions in 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 embodiments of the present invention, and not all embodiments. 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.

[0057] This invention provides a high-precision non-destructive testing system for welds based on ultrasonic phased arrays, such as... Figure 1 As shown, it includes:

[0058] The environmental perception and parameter acquisition module collects temperature data of the probe wedge and the workpiece surface through sensors integrated in the probe wedge, and generates temperature monitoring parameters including the wedge temperature value and the workpiece surface temperature value.

[0059] Furthermore, in the environmental perception and parameter acquisition module, the steps for generating temperature monitoring parameters, including the wedge temperature value and the workpiece surface temperature value, include:

[0060] The system acquires a first analog voltage signal output by a first temperature sensor embedded inside the probe wedge substrate, and acquires a second analog voltage signal output by a second temperature sensor located at the bottom of the wedge.

[0061] The first and second analog voltage signals are subjected to analog-to-digital conversion and low-pass filtering to generate the original digital temperature sampling sequence.

[0062] The pre-stored sensor voltage-temperature characteristic calibration curve is invoked to map the original digital temperature sampling sequence into physical temperature scale values, and the wedge temperature value and workpiece surface temperature value are calculated respectively.

[0063] The wedge temperature value and the workpiece surface temperature value are encapsulated according to a preset data format to generate temperature monitoring parameters.

[0064] Specifically, the physical implementation of the environmental perception and parameter acquisition module relies on a high-precision hardware sensor network and digital signal processing logic. The system first uses a first temperature sensor embedded inside the phased array probe wedge substrate and avoiding the area of ​​the ultrasonic main beam propagation path to sense the overall average temperature of the wedge material in real time. At the same time, a second temperature sensor installed at the bottom of the wedge and in close contact with the surface of the workpiece under test senses the surface temperature of the weld area. Both sensors use industrial-grade thermistors with linear output characteristics, which can convert the detected thermodynamic state into a continuously changing analog voltage signal. The first analog voltage signal represents the thermal state inside the wedge, and the second analog voltage signal represents the thermal state of the workpiece surface. In order to eliminate the influence of high-frequency electromagnetic interference in the industrial environment on the weak sensor signal, the module's built-in analog-to-digital conversion interface discretizes the above analog voltage signal at a set sampling frequency to generate the original digital temperature sampling sequence.

[0065] Subsequently, the system employs a moving average filtering algorithm to perform digital low-pass filtering on the sequence to remove random noise pulses. For a given voltage sample, the filtered voltage value is calculated using the following formula:

[0066] ;

[0067] The symbols in the formula are explained as follows:

[0068] Represents the current number The voltage value output after being processed by the filtering algorithm at each sampling time is used as the effective voltage reading at the current time.

[0069] The value represents the length of the sliding filter window. This value is a preset positive integer, for example, between 5 and 20, and is used to determine the amount of historical data to be included in the average calculation.

[0070] This represents the loop index variable in the summation operation, with values ​​incrementing from 0 to... Subtract 1;

[0071] Represents the current moment Move forward to the first The algorithm uses the original digital voltage sample values ​​collected at each sampling time to smooth out the instantaneous jitter of the voltage signal by taking the arithmetic average of the sampled data within the time window.

[0072] After obtaining a high signal-to-noise ratio digital voltage signal, the module calls the sensor voltage and temperature characteristic calibration curve pre-stored in the system's non-volatile memory. This calibration curve is a data set obtained by calibrating the sensor and its conditioning circuit under constant temperature conditions in a laboratory. It exhibits linear characteristics within the operating temperature range. Based on the current filtered voltage value, the system performs numerical mapping within the linear interval of the calibration curve to calculate the corresponding physical temperature scale value. The specific temperature conversion calculation is performed according to the following linear regression formula:

[0073] ;

[0074] The symbols in the formula are explained as follows:

[0075] This represents the final calculated physical temperature value, in degrees Celsius, corresponding to the wedge temperature or the workpiece surface temperature.

[0076] This represents the current voltage value output by the aforementioned filtering steps, in volts.

[0077] The voltage-temperature response coefficient of the sensor, i.e., the slope of the calibration curve, represents the amount of temperature change corresponding to a unit change in voltage, and its unit is degrees Celsius per volt.

[0078] The zero-point drift correction constant of the system, i.e. the intercept of the calibration curve, is used to correct the basic deviation. Its unit is degrees Celsius. The system performs the above calculation on the signals of the two sensors respectively, and independently calculates the current wedge temperature value and the workpiece surface temperature value. Finally, the module packages these two independent temperature values ​​according to the predefined data protocol to generate temperature monitoring parameters.

[0079] By combining hardware filtering with digital algorithms, the temperature fluctuations of the detection environment can be captured in real time, providing an accurate physical environmental benchmark for the subsequent acoustic model dynamic calculation module to correct the sound velocity and delay.

[0080] The acoustic model dynamic solution module, based on temperature monitoring parameters and combined with a preset material temperature change characteristic model, calculates the sound velocity and wedge delay values ​​of the wedge material and the weld material to be tested at the current temperature, and generates dynamic sound velocity and zero-point offset compensation parameters.

[0081] Furthermore, in the acoustic model dynamic solution module, the steps for calculating the sound velocity and wedge delay values ​​of the wedge material and the weld material under test at the current temperature, based on the preset material temperature change characteristic model, include:

[0082] Read the first reference sound velocity and first sound velocity temperature coefficient of the wedge material, and the second reference sound velocity and second sound velocity temperature coefficient of the weld material to be tested from the system's preset storage unit.

[0083] Using the wedge temperature value from the temperature monitoring parameters, a linear correction calculation is performed based on the first reference sound velocity and the first sound velocity temperature coefficient to generate the current wedge sound velocity value.

[0084] Using the workpiece surface temperature value from the temperature monitoring parameters, a linear correction calculation is performed based on the second reference sound velocity and the second sound velocity temperature coefficient to generate the current weld sound velocity value.

[0085] Call the pre-stored wedge geometric sound path data, divide the wedge geometric sound path data by the current wedge sound velocity value, calculate the propagation time of the ultrasonic wave inside the wedge, and generate the current wedge delay value;

[0086] The current wedge block sound velocity value, the current weld sound velocity value, and the current wedge block delay value are correlated and encoded to generate dynamic sound velocity and zero-point offset compensation parameters.

[0087] Furthermore, the steps of reading the first reference sound velocity and first sound velocity temperature coefficient of the wedge material, and the second reference sound velocity and second sound velocity temperature coefficient of the weld material to be tested from the system's preset storage unit include:

[0088] The system receives the wedge block model identification code and workpiece material type identification code input by the user through the human-computer interaction interface;

[0089] Access the pre-built acoustic material property database within the system and use the wedge block model identification code and the workpiece material type identification code as index keys for matching and retrieval;

[0090] Extract the corresponding room temperature sound velocity calibration value and the sound velocity-temperature change slope value from the matching records in the database;

[0091] The extracted room temperature sound velocity calibration values ​​were assigned as the first reference sound velocity and the second reference sound velocity, respectively. The extracted sound velocity-temperature change slope values ​​were assigned as the first sound velocity temperature coefficient and the second sound velocity temperature coefficient, respectively.

[0092] Specifically, the acoustic model dynamic solution module, as the parameter calibration center of the system, is mainly responsible for providing accurate acoustic parameters with temperature correction for subsequent imaging and positioning calculations. This module first runs based on a pre-built acoustic material property database. The system maintains this database in non-volatile memory, where each record corresponds to a specific wedge model or workpiece material type and stores a unique index identifier, the first or second reference sound velocity at the standard reference temperature, and the sound velocity temperature coefficient characterizing the acoustic thermosensitive properties of the material.

[0093] To ensure the accuracy of the sound velocity temperature coefficient in the database, the system performs a variable-temperature sound velocity measurement experiment during the factory calibration phase. For the acoustic medium to be calibrated, several sets of discrete temperature values ​​and corresponding measured sound velocity values ​​are collected within a preset temperature range. The collected data are then subjected to linear regression analysis using the least squares method to calculate the linear slope of sound velocity as a function of temperature. This calibration process is performed according to the following linear regression formula:

[0094] ;

[0095] The symbols in the formula are explained as follows:

[0096] The calculated temperature coefficient of sound speed is the slope of the fitted straight line, and its unit is meters per second per degree Celsius.

[0097] This represents the total number of discrete sample points collected in the calibration experiment;

[0098] Representing the The ambient temperature measurement value corresponding to each sample point is in degrees Celsius.

[0099] The arithmetic mean of the temperature values ​​of all sample points;

[0100] Represents temperature The measured sound velocity of the material is given in meters per second.

[0101] This represents the arithmetic mean of the sound speed values ​​at all sample points.

[0102] During the actual testing process, the module first obtains the wedge model and workpiece material identification input by the operator through the human-machine interface, and then retrieves the corresponding reference sound velocity and sound velocity temperature coefficient from the database. Simultaneously, the module reads temperature data in real time from the first temperature sensor attached to the probe wedge surface and the second temperature sensor attached to the workpiece surface through the temperature sensing interface. Using the retrieved static parameters and the collected dynamic temperature data, the module calculates the wedge sound velocity and weld sound velocity values ​​at the current temperature based on a linear acoustic model. The calculation process follows the sound velocity correction formula:

[0103] ;

[0104] The symbols in the formula are explained as follows:

[0105] This represents the calculated current real-time sound velocity, specifically corresponding to the current wedge sound velocity value or the current weld sound velocity value, with the unit being meters per second.

[0106] The reference speed of sound at the standard reference temperature recorded in the database is expressed in meters per second.

[0107] This represents the corresponding sound speed temperature coefficient, with units of meters per second per degree Celsius.

[0108] This represents the first or second temperature data collected in real time, in degrees Celsius.

[0109] This represents the standard reference temperature used to define the baseline speed of sound, typically set to 20 degrees Celsius.

[0110] Based on the calculated current sound velocity value of the wedge, the module further calculates the propagation time offset inside the wedge caused by the change in sound velocity, i.e., the current wedge delay value. This parameter is used to compensate for the zero-point error in the propagation time of the ultrasonic wave in the wedge medium before entering the workpiece. Its calculation is based on the following propagation time formula:

[0111] ;

[0112] The symbols in the formula are explained as follows:

[0113] This represents the calculated current wedge delay value, in microseconds.

[0114] This represents the inherent geometric sound path length inside the wedge, a value determined by the wedge's geometry and the probe's mounting position, expressed in millimeters.

[0115] This represents the current sound velocity value of the wedge block calculated in the preceding steps, in meters per second;

[0116] The constant 1000 in the formula is used to convert the speed unit from meters per second to millimeters per millisecond, thus adapting it to the microsecond-level time unit.

[0117] By packaging the current wedge sound velocity value, the current weld sound velocity value, and the current wedge delay value obtained in real time into dynamic sound velocity and zero-point offset compensation parameters, and sending them to the subsequent signal processing unit, this step eliminates the nonlinear influence of temperature fluctuations on sound velocity by introducing real-time physical field correction, providing a reliable physical basis for achieving high-precision focusing law calculation.

[0118] The phase law calculation module calculates the excitation delay time of each array chip based on dynamic sound speed and zero-point offset compensation parameters, according to Snell's law, and generates phase correction excitation timing instructions.

[0119] Furthermore, in the phase law calculation module, the steps of calculating the excitation delay time of each array wafer according to Snell's law and generating phase correction excitation timing instructions include:

[0120] Based on the preset scanning strategy, a series of virtual focusing target points are discretized and generated within the weld area to be tested, and their spatial coordinates are generated.

[0121] For each virtual focusing target point, using Fermat's principle and the current wedge sound velocity value and current weld sound velocity value in the dynamic sound velocity and zero-point offset compensation parameters, the acoustic propagation path conforming to Fermat's principle from the center point of each wafer in the array to the virtual focusing target point is calculated. The acoustic propagation path conforming to Fermat's principle satisfies Snell's law of refraction at the interface between the wedge and the workpiece.

[0122] Calculate the absolute transit time required for the ultrasonic wave to propagate along each acoustic propagation path that conforms to Fermat's principle. The absolute transit time is the sum of the path length of the sound wave in the wedge divided by the current sound velocity value of the wedge and the path length of the sound wave in the workpiece divided by the current sound velocity value of the weld.

[0123] The peak value of the absolute transit time of all chips corresponding to the same virtual focus target point is selected as the reference time. The difference between the absolute transit time of each chip and the reference time is calculated to obtain the relative excitation delay time of each chip.

[0124] The relative excitation delay times of all wafers are encoded according to the array arrangement order to generate phase-corrected excitation timing instructions.

[0125] Specifically, the phase law calculation module acts as a bridge connecting the acoustic physical model and the hardware execution unit, and is responsible for converting abstract acoustic parameters into specific circuit control timing. The module first reads the dynamic sound velocity and zero-point offset compensation parameters output by the aforementioned acoustic model dynamic solution module. This parameter set includes the current wedge sound velocity value, the current weld sound velocity value, and the current wedge delay value after temperature correction. According to the preset scanning strategy, such as sector scanning or linear scanning, the module discretizes and generates a series of virtual focusing target points in the cross-section of the weld area to be measured. These target points have definite geometric positions in the workpiece coordinate system and together constitute the focusing grid required for imaging.

[0126] For each virtual focusing target point, the system needs to calculate the acoustic path required for each chip unit in the array to emit ultrasonic waves and converge at that point. Since ultrasonic waves will be refracted at the interface when they enter the workpiece from the wedge, the module uses Fermat's principle, i.e. the principle of the shortest optical path, to solve the propagation path. The system uses the horizontal coordinates on the interface between the wedge and the workpiece as the iterative variable to establish a path time function. The system uses a numerical optimization algorithm to search for the point where the total propagation time reaches a minimum value. This point is the physical refraction point that satisfies Snell's law of refraction. After the refraction point is determined, the sound wave propagation path is precisely divided into the first path segment located in the wedge and the second path segment located in the workpiece.

[0127] Based on this, the module combines the real-time updated sound velocity parameters to calculate the absolute transit time of the ultrasonic wave along the optimal path. To ensure the accuracy of the time reference, the calculation process incorporates the current wedge delay value as an inherent zero-point compensation of the system and performs strict unit conversion to match microsecond-level timing accuracy. The array chip reaches the first The absolute transit time of each virtual focal target point is calculated using the following path integral formula:

[0128] ;

[0129] The symbols in the formula are explained as follows:

[0130] Represents the calculated first The chip to the first The absolute transit time of a sound wave at a focal point is measured in microseconds.

[0131] Representative from the first The geometric path length of a wafer center propagating through the calculated refraction point in the wedge medium, in millimeters;

[0132] This represents the current sound velocity of the wedge, in meters per second. Dividing by 1,000 in the formula converts it to millimeters per microsecond.

[0133] Represents the distance from the refraction point to the th The geometric path length of a focal point propagating in the workpiece medium, in millimeters;

[0134] This represents the current sound velocity value of the weld, in meters per second.

[0135] This represents the current wedge delay value input, which is the value calculated by the aforementioned module. It is used to compensate for the zero-point deviation of the system, and its unit is microsecond.

[0136] To achieve in-phase superposition of beams at the virtual focusing target point, the system needs to control the transmission timing to compensate for the time difference caused by the acoustic path difference between the individual chips. The module iterates through the absolute transit times of all chips for the same virtual focusing target point, selects the maximum value as the common reference time, and then calculates the relative excitation delay time of each chip based on the following delay logic formula:

[0137] ;

[0138] The symbols in the formula are explained as follows: Representing the The chip is for the first The relative excitation delay time of each focal point, which is a non-negative number and is in microseconds;

[0139] Represents the total number of chips in the probe array;

[0140] This represents the function to find the maximum value, used in all cases. Find the maximum value from the set of absolute transit times of each chip, and ensure that the chip with the longest path distance from the focal point is used as the zero point of the emission reference, and the other chips are emitted accordingly.

[0141] After the calculation is completed, the module encodes the relative excitation delay time corresponding to all wafers according to the physical arrangement order of the array and the scanning order of the focal point, and generates a phase correction excitation timing instruction containing specific nanosecond-level delay data. This instruction is directly transmitted to the subsequent array excitation module. By directly substituting the temperature-corrected dynamic sound velocity and zero-point parameters into the calculation source of the delay law, it ensures that the generated focusing law can adapt to changes in ambient temperature in real time, effectively preventing focal point defocusing and phase distortion caused by sound velocity drift, and ensuring the precise focusing of the ultrasonic beam inside the weld.

[0142] The array excitation and echo acquisition module, based on the phase correction excitation timing command, drives the multi-channel piezoelectric crystal of the phased array probe to emit ultrasonic waves and receive reflected waves, converting physical sound waves into electrical signals and generating multi-channel raw echo time-domain signals.

[0143] Furthermore, in the array excitation and echo acquisition module, the steps for generating multi-channel raw echo time-domain signals include:

[0144] The phase correction excitation timing command is analyzed, and the multi-channel high-voltage pulse transmitting circuit is controlled to apply high-frequency high-voltage excitation pulses to each piezoelectric crystal of the phased array probe according to the delay time set in the command.

[0145] The ultrasonic mechanical wave reflected from the weld under test is received by a piezoelectric crystal, and the ultrasonic mechanical wave is converted into a continuously changing analog echo voltage signal by utilizing the piezoelectric effect.

[0146] The analog echo voltage signal of each channel is pre-amplified and bandpass filtered;

[0147] The processed analog echo voltage signal is synchronously sampled and quantized using an analog-to-digital converter to generate a discrete digital echo sequence.

[0148] The discretized digital echo sequences of all channels are encapsulated in parallel according to the channel index to generate multi-channel original echo time-domain signals.

[0149] Specifically, the array excitation and echo acquisition module, as the physical front end of the system, mainly consists of a multi-channel ultrasonic transmitting and receiving circuit and a high-speed digital signal processing unit. This module first receives the phase correction excitation timing command generated by the aforementioned phase law calculation module. The field-programmable gate array main control unit inside the system performs parallel parsing of the command and extracts the specific relative excitation delay time for each physical channel of the phased array probe. Based on these delay time data, the main control unit configures the delay counter inside the multi-channel high-voltage pulse transmitting circuit to ensure that each channel is triggered in a predetermined order with nanosecond-level time accuracy. When the trigger signal arrives, the high-voltage pulse circuit applies an instantaneous high-frequency high-voltage excitation pulse to the corresponding piezoelectric crystal of the phased array probe. The voltage amplitude of this pulse is usually set between 50 volts and 200 volts, driving the crystal to emit ultrasonic waves through the inverse piezoelectric effect. These ultrasonic waves with specific phase differences interfere and superimpose inside the workpiece to form a synthetic sound beam with a specific focusing law.

[0150] When the ultrasonic wave encounters a heterogeneous interface or defect inside the weld, it is reflected. The echo signal returns along the original path and is received by the piezoelectric crystal array of the probe. At this time, each piezoelectric crystal uses the positive piezoelectric effect to convert the received weak mechanical vibration into a continuously changing analog echo voltage signal. Since the original echo signal amplitude is extremely low and accompanied by environmental noise, the analog front-end circuit in the module first performs low-noise pre-amplification processing on the signal of each channel to improve the signal-to-noise ratio. Subsequently, the signal passes through a bandpass filter. The passband frequency range of this filter is set to cover the center frequency and main bandwidth of the probe to filter out low-frequency mechanical vibration interference and high-frequency electromagnetic noise.

[0151] After analog conditioning, the signal enters a high-speed analog-to-digital converter for discretization. The system uses a unified high-precision clock source to drive the analog-to-digital converters of all channels to perform synchronous sampling, ensuring strict alignment of the data of each channel on the time axis. For any channel's analog voltage signal at a specific sampling time, the process of converting it into a digitally quantized discrete value is performed according to the following quantization formula:

[0152] ;

[0153] The symbols in the formula are explained as follows:

[0154] Representing the The discrete digital echo sequence value output by each sampling point, which is a dimensionless integer;

[0155] Representative at the The instantaneous amplitude of the analog echo voltage signal input at each sampling period, in volts;

[0156] This represents the sampling period of the system, which is the reciprocal of the sampling frequency, and its unit is seconds;

[0157] This represents the DC bias voltage at the input of the analog-to-digital converter, used to shift bipolar AC signals to the unipolar input range, and its unit is volts.

[0158] This represents the full-scale voltage range of the analog-to-digital converter, and its unit is volts.

[0159] This represents the number of quantization bits in the analog-to-digital converter, typically chosen as 12 or 14 bits to ensure dynamic range. That is, the maximum quantization order;

[0160] This represents the rounding operation function.

[0161] After quantization, the module encapsulates the discrete digital echo sequences generated by all channels in parallel according to the channel index order. The system adds the corresponding timestamp and gain information to the data packet header, and finally generates multi-channel raw echo time-domain signals. Through high-precision timing control and synchronous acquisition, the phase and amplitude information of the echo signals are completely preserved, providing a high-quality raw data foundation with rich physical features for subsequent image reconstruction and geometric correction.

[0162] The image reconstruction and geometric correction module, based on the multi-channel original echo time-domain signal and dynamic sound velocity and zero-point offset compensation parameters, calculates the spatial coordinates of the echo signal using the dynamic sound velocity and zero-point offset compensation parameters, maps the time-domain signal into spatial location distribution data, and generates geometrically corrected imaging data.

[0163] Furthermore, in the image reconstruction and geometric correction module, the steps for generating geometrically corrected imaging data include:

[0164] A two-dimensional imaging grid covering the area of ​​the weld to be tested is constructed. The two-dimensional imaging grid consists of several pixels with independent spatial coordinates.

[0165] For each pixel, the current wedge sound velocity value, current weld sound velocity value, and current wedge delay value in the dynamic sound velocity and zero-point offset compensation parameters are used to calculate the two-way propagation time of the ultrasonic wave from each chip in the array to the pixel and back, and generate a dynamic focusing delay rule table.

[0166] Based on the dynamic focusing delay rule table, the instantaneous amplitude data corresponding to each channel is extracted from the original echo time domain signal of the multi-channel, and the extracted data is coherently superimposed and summed to calculate the synthetic sound intensity value of the pixel.

[0167] The synthetic sound intensity value is calculated by traversing all pixels, and the synthetic sound intensity values ​​of all pixels are arranged according to the spatial coordinates of the pixels to generate geometrically corrected imaging data.

[0168] Specifically, the image reconstruction and geometric correction module, as the imaging core of the ultrasonic testing system, is responsible for converting the original time-domain signal carrying physical information into an intuitive spatial image. In this process, it dynamically compensates for positional distortion caused by changes in ambient temperature. The system first performs the initialization operation of the imaging space and constructs a two-dimensional imaging grid covering the weld area to be tested. This grid defines a virtual coordinate system corresponding to the actual workpiece cross-section. Usually, the center point of the phased array probe or the position of the first crystal is taken as the origin of the coordinate system. The horizontal axis is parallel to the workpiece surface, and the vertical axis points perpendicularly into the workpiece. The grid consists of several pixels with independent spatial coordinates. The density of the pixels determines the spatial resolution of the final image. This value is usually set between 0.1 mm and 0.5 mm.

[0169] After determining the imaging grid, the module enters the calculation process of the dynamic focusing delay rule table. For each pixel in the grid, the system uses the temperature monitoring parameters generated by the aforementioned module to perform precise acoustic path calculation. Since sound waves refract when passing through the interface between the wedge and the workpiece, the system, based on Fermat's principle (the principle of the shortest sound wave propagation time), searches for the optimal refraction point at the interface between the wedge and the workpiece using a numerical iterative algorithm. After determining the refraction point, the sound wave propagation path is divided into a first path segment located inside the wedge and a second path segment located inside the workpiece. Based on the geometric lengths of these two path segments, combined with the current real-time sound velocity, and introducing the current wedge delay value as the system's inherent zero-point offset correction, the system calculates the one-way propagation time of the ultrasonic wave during reception. Each receiving channel and target pixel The focusing delay time between them is calculated using the following formula:

[0170] ;

[0171] The symbols in the formula are explained as follows:

[0172] Represents the calculated array of the first individual chips and target pixels The one-way sound propagation time between the receivers is measured in microseconds.

[0173] Representative from the first The center of each chip reaches the pixel through the point of refraction. The physical length of the path in the wedge medium, in millimeters;

[0174] This represents the current sound velocity of the wedge, which needs to be converted to millimeters per microsecond to match the unit.

[0175] Represents the distance from the refraction point to the pixel. The physical length of the path in the medium of the weld to be tested, in millimeters;

[0176] This represents the current sound velocity value of the weld, expressed in millimeters per microsecond.

[0177] The current wedge delay value represents the input and is used here as the system's time base zero offset in the calculation. Its unit is microseconds. The system traverses all combinations of chips and all pixels to generate a dynamic focus delay rule table containing all focus time indices.

[0178] Subsequently, the system performs coherent superposition and summation processing to generate the composite sound intensity value of the pixel. According to the dynamic focusing delay rule table, the module extracts the instantaneous amplitude data of each channel at the corresponding delay time from the multi-channel original echo time domain signal. To ensure accuracy, the system also uses the aforementioned current wedge sound velocity value and current weld sound velocity value to calculate the transmission propagation time of the sound wave from the transmission center to the pixel. The transmission propagation time is superimposed with the reception propagation time to obtain the total sound path time. Since the sampling point corresponding to this time may be between two integer sampling indices, the system uses an interpolation algorithm to obtain the accurate signal amplitude. The calculation of the composite sound intensity value is performed according to the following beamforming formula:

[0179] ;

[0180] The symbols in the formula are explained as follows:

[0181] Represents the target pixel The synthesized sound intensity value at a location represents the probability or intensity of the presence of a reflector at that location;

[0182] Represents the total number of array chip channels involved in imaging;

[0183] Representing the The weighting coefficients for each channel are used to control the sidelobe level, and are usually implemented using Hanning window or rectangular window functions.

[0184] Representing the Discretized original echo sequences for each channel;

[0185] This represents the distance from the emission phase center to the pixel, calculated using the current sound speed parameters. The launch propagation time is measured in microseconds.

[0186] The pixel points calculated in the preceding steps represent the values ​​obtained from the previous steps. Return to page The reception propagation time of each receiving chip is measured in microseconds.

[0187] This represents the signal sampling interval of the system, and its unit is microseconds.

[0188] After completing the calculation of all pixels, the module arranges the synthetic sound intensity values ​​of all pixels according to their spatial coordinates in the grid, generating geometrically corrected imaging data in the form of a two-dimensional matrix. By directly fusing real-time temperature and sound velocity parameters at the bottom layer of the imaging algorithm, the module effectively eliminates the sound beam deflection and positioning errors caused by the thermal expansion and contraction of the wedge and the sound velocity drift of the workpiece, and realizes high-precision and high-fidelity imaging of weld defects in the variable temperature environment of industrial sites.

[0189] The defect feature extraction and evaluation module extracts geometric features of areas where the amplitude intensity exceeds a preset threshold based on geometrically corrected imaging data, and generates weld quality inspection and evaluation results that include defect depth location data and geometric size data.

[0190] Furthermore, in the defect feature extraction and evaluation module, the steps for generating weld quality inspection and evaluation results that include defect depth and location data as well as geometric dimension data include:

[0191] Spatial smoothing filtering is applied to the geometrically corrected imaging data to generate an enhanced imaging matrix;

[0192] Set an amplitude judgment threshold, set the pixels in the enhanced imaging matrix whose sound intensity value is lower than the amplitude judgment threshold to zero, retain the pixels whose sound intensity value is not less than the amplitude judgment threshold, and generate a binarized feature map.

[0193] A connected component labeling algorithm is used to perform cluster analysis on non-zero pixels in the binarized feature map, and adjacent connected non-zero pixels are identified as independent defect candidate targets.

[0194] For each defect candidate target, the depth coordinates of the preset feature points of the defect candidate target in the imaging grid coordinate system are extracted as defect depth location data, and the span of the defect candidate target in the horizontal and depth directions is calculated as geometric dimension data.

[0195] The defect depth and location data and geometric dimension data are structured and encapsulated to output the weld quality inspection and evaluation results.

[0196] Specifically, the defect feature extraction and evaluation module, as the decision-making terminal of the intelligent detection process, is responsible for automatically identifying potential welding defects from the geometrically corrected imaging data output by the aforementioned image reconstruction and geometric correction module. The system first reads the geometrically corrected imaging data stored in the form of a two-dimensional matrix. This data contains the synthetic acoustic intensity value of each pixel in the weld area. In order to suppress background noise speckles and smooth the random speckle interference caused by coherent imaging, the module uses a Gaussian smoothing filter algorithm to perform spatial domain enhancement processing on the original imaging matrix. For a pixel at any coordinate position in the imaging grid, its smoothed acoustic intensity value is calculated using the following discrete convolution formula:

[0197] ;

[0198] The symbols in the formula are explained as follows:

[0199] The representative image is located at the horizontal physical coordinates after filtering. and depth physical coordinates The pixel at that location increases the sound intensity value;

[0200] This represents the normalization coefficient, which is the sum of all weight coefficients within the convolution kernel, used to ensure energy conservation before and after filtering;

[0201] An integer value representing the radius of the filter kernel, typically 1 or 2, corresponding to a 3x3 or 5x5 convolution window;

[0202] The synthesized acoustic intensity value represents the corresponding coordinate in the original geometrically corrected imaging data.

[0203] and These represent the local offset variables in the horizontal and vertical directions within the convolution window, respectively.

[0204] The weighting coefficients of the predefined Gaussian kernel function are used to perform a weighted average of the center pixel and its neighboring pixels. These coefficients follow a two-dimensional normal distribution.

[0205] After image enhancement, the system enters the binarization segmentation stage for defect targets. The module first calculates the global maximum sound intensity value in the enhanced imaging matrix and, based on a preset relative sensitivity standard (e.g., 50% of the maximum value, corresponding to -6 dB), determines an amplitude threshold. The system iterates through the enhanced imaging matrix, setting all pixels with sound intensity values ​​below the threshold to zero and those with sound intensity values ​​not less than the threshold to one, thus generating a binarized feature map. Subsequently, the module processes this binarized feature map using a connected component labeling algorithm, scanning the image according to the eight-neighbor connectivity rule, merging interconnected non-zero pixels into the same set, thereby identifying several independent defect candidate targets. For the first... One identified defect candidate target area The set of pixels it contains is defined as follows:

[0206] ;

[0207] The symbols in the formula are explained as follows:

[0208] Representing the A separate set of candidate defect pixels;

[0209] This represents the pair of physical space coordinates corresponding to a pixel in the imaging grid, where For horizontal physical coordinates, These are physical coordinates for depth, all in millimeters.

[0210] This represents the logical value corresponding to that point in the binary feature map;

[0211] This represents the unique target index number assigned to this point by the connected component algorithm.

[0212] Represents the integer number of the defect target currently being processed.

[0213] Finally, the system performs quantitative feature extraction for each marked defect candidate target. To accurately locate the defect and estimate its equivalent size, the module traverses the target area. For all pixels within the area, their physical coordinate boundaries are extracted. The depth location data of the defect is usually taken as the depth of the pixel with the maximum original sound intensity value in the area, or the geometric center of the area in the depth direction. The geometric dimensions of the defect include the horizontal span and the depth span, and their calculation is performed according to the following boundary extreme value formula:

[0214] ;

[0215] The symbols in the formula are explained as follows:

[0216] This represents the location data of the defect depth in the output, in millimeters.

[0217] The depth physical coordinates of the pixel with the maximum sound intensity within the defect target area;

[0218] This represents the calculated horizontal span of the defect, i.e., the length component in the geometric dimension data. Its value is equal to the difference between the maximum and minimum horizontal physical coordinates within the target area, in millimeters.

[0219] This represents the calculated defect depth span, i.e., the height component in the geometric dimension data. Its value is equal to the difference between the maximum and minimum depth physical coordinates within the target area, in millimeters.

[0220] and They represent in the set The function that calculates the maximum and minimum values ​​of the physical coordinate variables within the brackets.

[0221] The system encapsulates the calculated parameters such as depth position, horizontal span, and depth span in a structured manner, and automatically determines the defect level in combination with preset acceptance standards. Finally, it outputs weld quality inspection and evaluation results containing digital features, thereby realizing the automated conversion from acoustic images to quantitative data. It can accurately capture the true defect morphology and location after temperature correction, providing objective and reliable data support for subsequent welding quality assessment.

[0222] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A high-precision non-destructive testing system for welds based on ultrasonic phased array, characterized in that, include: The environmental perception and parameter acquisition module collects temperature data of the probe wedge and the workpiece surface through sensors integrated in the probe wedge, and generates temperature monitoring parameters including the wedge temperature value and the workpiece surface temperature value. The acoustic model dynamic solution module, based on the temperature monitoring parameters and combined with the preset material temperature change characteristic model, calculates the sound velocity value and wedge delay value of the wedge material and the weld material to be tested at the current temperature, and generates dynamic sound velocity and zero-point offset compensation parameters. The phase law calculation module calculates the excitation delay time of each array chip based on the dynamic sound speed and zero-point offset compensation parameters, according to Snell's law, and generates phase correction excitation timing instructions. The array excitation and echo acquisition module, based on the phase correction excitation timing command, drives the multi-channel piezoelectric crystal of the phased array probe to emit ultrasonic waves and receive reflected waves, converting physical sound waves into electrical signals and generating multi-channel raw echo time-domain signals. The image reconstruction and geometric correction module, based on the multi-channel original echo time-domain signal and the dynamic sound velocity and zero-point offset compensation parameters, uses the dynamic sound velocity and zero-point offset compensation parameters to calculate the spatial coordinates of the echo signal, maps the time-domain signal into spatial location distribution data, and generates geometrically corrected imaging data. The defect feature extraction and evaluation module extracts the geometric features of areas where the amplitude intensity exceeds a preset threshold based on the geometrically corrected imaging data, and generates a weld quality inspection and evaluation result that includes defect depth location data and geometric size data.

2. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 1, characterized in that, In the environmental sensing and parameter acquisition module, the step of generating temperature monitoring parameters including wedge temperature values ​​and workpiece surface temperature values ​​includes: The system acquires a first analog voltage signal output by a first temperature sensor embedded inside the probe wedge substrate, and acquires a second analog voltage signal output by a second temperature sensor located at the bottom of the wedge. The first analog voltage signal and the second analog voltage signal are subjected to analog-to-digital conversion and low-pass filtering to generate the original digital temperature sampling sequence. The pre-stored sensor voltage-temperature characteristic calibration curve is invoked to map the original digital temperature sampling sequence into physical temperature scale values, and the wedge temperature value and workpiece surface temperature value are calculated respectively. The wedge temperature value and the workpiece surface temperature value are encapsulated in a preset data format to generate temperature monitoring parameters.

3. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 1, characterized in that, In the acoustic model dynamic solution module, the steps of calculating the sound velocity and wedge delay values ​​of the wedge material and the weld material under test at the current temperature, based on a preset material temperature change characteristic model, include: Read the first reference sound velocity and first sound velocity temperature coefficient of the wedge material, and the second reference sound velocity and second sound velocity temperature coefficient of the weld material to be tested from the system's preset storage unit. Using the wedge temperature value in the temperature monitoring parameters, a linear correction calculation is performed based on the first reference sound velocity and the first sound velocity temperature coefficient to generate the current wedge sound velocity value. Using the workpiece surface temperature value in the temperature monitoring parameters, a linear correction calculation is performed based on the second reference sound velocity and the second sound velocity temperature coefficient to generate the current weld sound velocity value. The pre-stored wedge geometric sound path data is called, the wedge geometric sound path data is divided by the current wedge sound velocity value, the propagation time of the ultrasonic wave inside the wedge is calculated, and the current wedge delay value is generated. The current wedge block sound velocity value, the current weld sound velocity value, and the current wedge block delay value are correlated and encoded to generate dynamic sound velocity and zero-point offset compensation parameters.

4. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 3, characterized in that, The steps of reading the first reference sound velocity and first sound velocity temperature coefficient of the wedge material, and the second reference sound velocity and second sound velocity temperature coefficient of the weld material to be tested from the system's preset storage unit include: The system receives the wedge block model identification code and workpiece material type identification code input by the user through the human-computer interaction interface; Access the pre-built acoustic material property database within the system, and use the wedge block model identification code and the workpiece material type identification code as index keys for matching and retrieval; Extract the corresponding room temperature sound velocity calibration value and the sound velocity-temperature change slope value from the matching records in the database; The extracted room temperature sound velocity calibration values ​​were assigned as the first reference sound velocity and the second reference sound velocity, respectively. The extracted sound velocity-temperature change slope values ​​were assigned as the first sound velocity temperature coefficient and the second sound velocity temperature coefficient, respectively.

5. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 1, characterized in that, In the phase law calculation module, the step of calculating the excitation delay time of each array wafer based on Snell's law and generating phase correction excitation timing instructions includes: Based on the preset scanning strategy, a series of virtual focusing target points are discretized and generated within the weld area to be tested, and their spatial coordinates are generated. For each virtual focusing target point, using Fermat's principle and the current wedge sound velocity value and current weld sound velocity value in the dynamic sound velocity and zero-point offset compensation parameters, the acoustic propagation path conforming to Fermat's principle from the center point of each wafer in the array to the virtual focusing target point is calculated. The acoustic propagation path conforming to Fermat's principle satisfies Snell's law of refraction at the interface between the wedge and the workpiece. Calculate the absolute transit time required for the ultrasonic wave to propagate along each of the acoustic propagation paths that conform to Fermat's principle. The absolute transit time is the sum of the path length of the sound wave in the wedge divided by the current sound velocity value of the wedge and the path length of the sound wave in the workpiece divided by the current sound velocity value of the weld. The peak value of the absolute transit time of all wafers corresponding to the same virtual focus target point is selected as the reference time. The difference between the absolute transit time of each wafer and the reference time is calculated to obtain the relative excitation delay time of each wafer. The relative excitation delay times of all wafers are encoded in the array arrangement order to generate phase-corrected excitation timing instructions.

6. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 1, characterized in that, In the array excitation and echo acquisition module, the step of generating multi-channel raw echo time-domain signals includes: The phase correction excitation timing command is analyzed, and the multi-channel high-voltage pulse transmitting circuit is controlled to apply high-frequency high-voltage excitation pulses to each piezoelectric crystal of the phased array probe according to the delay time set in the command; The piezoelectric crystal receives the ultrasonic mechanical wave reflected from the weld seam under test, and uses the piezoelectric effect to convert the ultrasonic mechanical wave into a continuously changing analog echo voltage signal. The analog echo voltage signal of each channel is pre-amplified and bandpass filtered; The processed analog echo voltage signal is synchronously sampled and quantized using an analog-to-digital converter to generate a discrete digital echo sequence; The discretized digital echo sequences of all channels are encapsulated in parallel according to the channel index to generate multi-channel original echo time-domain signals.

7. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 1, characterized in that, In the image reconstruction and geometric correction module, the step of generating geometrically corrected imaging data includes: A two-dimensional imaging grid is constructed to cover the area of ​​the weld to be tested. The two-dimensional imaging grid consists of several pixels with independent spatial coordinates. For each pixel, the current wedge sound velocity value, current weld sound velocity value, and current wedge delay value in the dynamic sound velocity and zero-point offset compensation parameters are used to calculate the two-way propagation time of the ultrasonic wave from each chip in the array to the pixel and back, and generate a dynamic focusing delay rule table. Based on the dynamic focusing delay rule table, the instantaneous amplitude data corresponding to each channel is extracted from the multi-channel original echo time-domain signal, and the extracted data is coherently superimposed and summed to calculate the synthetic sound intensity value of the pixel. The synthetic sound intensity value is calculated by traversing all pixels, and the synthetic sound intensity values ​​of all pixels are arranged according to the spatial coordinates of the pixels to generate geometrically corrected imaging data.

8. The high-precision non-destructive testing system for welds based on ultrasonic phased array according to claim 1, characterized in that, In the defect feature extraction and evaluation module, the step of generating weld quality inspection and evaluation results that include defect depth location data and geometric dimension data includes: The geometrically corrected imaging data is subjected to spatial smoothing filtering to generate an enhanced imaging matrix; A threshold for amplitude determination is set, and pixels in the enhanced imaging matrix whose sound intensity values ​​are lower than the threshold are set to zero, while pixels whose sound intensity values ​​are not less than the threshold are retained, thus generating a binarized feature map. A connected component labeling algorithm is used to perform cluster analysis on the non-zero pixels in the binarized feature map, and adjacent connected non-zero pixels are identified as independent defect candidate targets. For each of the defect candidate targets, the depth coordinates of the preset feature points of the defect candidate target in the imaging grid coordinate system are extracted as the defect depth position data, and the span of the defect candidate target in the horizontal and depth directions is calculated as the geometric dimension data. The defect depth location data and the geometric dimension data are structured and encapsulated to output the weld quality inspection and evaluation results.

Images