Nondestructive testing device and method for pressure-bearing equipment with thermal insulation layer

Abstract

The invention provides a nondestructive testing device and method for pressure-bearing equipment with a heat preservation layer, and belongs to the technical field of nondestructive testing, the device comprises a plurality of waveguide rods, the rear end of at least one waveguide rod in the plurality of waveguide rods is connected with an ultrasonic transmitter, and the rear end of at least another waveguide rod in the plurality of waveguide rods is connected with an acoustic emission sensor; the fixing assembly is used for fixing the waveguide rod on the outer wall of the pressure-bearing equipment; the controller is in communication connection with the ultrasonic transmitter and the acoustic emission sensor and is used for detecting whether the pressure-bearing equipment has defects or not; the front end of the waveguide rod is provided with an arc-shaped coupling face matched with the curvature of the outer wall of the pressure-bearing equipment, the front end of the waveguide rod penetrates through the heat preservation layer, and the arc-shaped coupling face is covered with the heat preservation layer and attached to the outer wall of the pressure-bearing equipment. The method solves the problems of high detection cost and complicated operation caused by the need of dismounting a thermal insulation layer or difficulty in low-temperature coupling in a traditional method, and has the advantages of simplicity and convenience in operation, low cost, high detection efficiency, high accuracy, high adaptability and the like.

Description

Non-destructive testing device and method for pressure-bearing equipment with insulation layer Technical Field

[0001] This application relates to the field of nondestructive testing technology, and in particular to a nondestructive testing device and method for a pressure-bearing device with an insulation layer. Background Technology

[0002] Liquefied natural gas (LNG) and liquid hydrogen are important clean and efficient energy sources. Cryogenic pipelines, as key infrastructure in the storage and transportation systems for these media, directly impact the stability, economy, and public safety of energy supply. In recent years, with the accelerated construction of LNG receiving terminals, liquid hydrogen storage and transportation facilities, and related long-distance pipelines, the use of cryogenic pipelines has become increasingly widespread, placing higher demands on their structural health monitoring and non-destructive testing technologies. Especially in extreme low-temperature environments, such as approximately -162°C for LNG and approximately -253°C for liquid hydrogen, issues such as changes in pipeline material properties, weld integrity, low-temperature brittleness, fatigue cracking, and corrosion become more prominent. Therefore, there is an urgent need to develop online detection and real-time monitoring technologies applicable to various cryogenic media environments to achieve preventative maintenance and safe operation of pipelines.

[0003] Currently, among non-destructive testing technologies for cryogenic pipelines, acoustic emission testing (AE) has attracted considerable attention due to its ability to achieve dynamic monitoring and damage source localization. However, in practical applications of cryogenic pipelines with insulation layers, directly deploying AE sensors faces the following significant problems: sensor coupling is difficult; the performance of conventional coupling agents degrades in low-temperature environments, making it difficult for sensors to achieve stable acoustic coupling with the pipeline surface; signal distortion and attenuation are severe; ultrasonic signal propagation attenuation is aggravated in cryogenic media and is easily affected by the pipeline's insulation layer, outer protective pipe, and other multi-layered structures, leading to difficulty in signal acquisition and low signal-to-noise ratio; testing costs are high and cycles are long; traditional testing methods typically require the removal of the pipeline's insulation layer and outer protective layer, which is not only complex and time-consuming but also has high restoration costs, making it difficult to achieve routine online monitoring. Therefore, how to accurately achieve non-destructive testing of cryogenic pipelines, improve the accuracy and stability of pipeline testing, and reduce the operational complexity of pipeline testing are problems that need to be solved in this field. Summary of the Invention

[0004] This application provides a non-destructive testing device and method for pressure-bearing equipment with insulation layer, which solves the problems of high testing cost and complicated operation caused by the need to remove the insulation layer or the difficulty of low-temperature coupling in traditional methods. It has the advantages of simple operation, low cost, high testing efficiency, high accuracy and strong adaptability.

[0005] In a first aspect, this application provides a non-destructive testing device for a pressure-bearing device with an insulation layer, comprising: multiple waveguide rods, at least one of which has an ultrasonic transmitter connected to its rear end, and at least another of which has an acoustic emission sensor connected to its rear end; a fixing assembly for fixing the waveguide rods to the outer wall of the pressure-bearing device; and a controller, communicatively connected to the ultrasonic transmitter and the acoustic emission sensor, for detecting whether the pressure-bearing device has defects; wherein, the front end of each waveguide rod has an arc-shaped coupling surface adapted to the curvature of the outer wall of the pressure-bearing device, the front end of the waveguide rod passes through the insulation layer, and the arc-shaped coupling surface is covered by the insulation layer and adheres to the outer wall of the pressure-bearing device.

[0006] In some embodiments, the front end of the waveguide rod is provided with an acoustic wave converging structure, and / or the rear end of the waveguide rod is provided with an acoustic wave converging structure; wherein the acoustic wave converging structure is designed based on the acoustic black hole theory.

[0007] In some embodiments, the device includes: four waveguide rods, one waveguide rod having an ultrasonic transmitter connected to its rear end, and three waveguide rods having acoustic emission sensors connected to their rear ends, the four waveguide rods being symmetrically distributed on the outer wall of the pressure-bearing device.

[0008] In some embodiments, the waveguide rod is permanently fixed to the outer wall of the pressure-bearing equipment, and the installation of the waveguide rod requires only one insulation layer treatment.

[0009] In some embodiments, the waveguide rod is made of a low-temperature resistant material that matches the acoustic impedance of the pressure-bearing device.

[0010] In some embodiments, the fixing component includes: at least one locking clamp, which covers the body portion of the outer waveguide rod of the insulation layer and is securely connected to the pressure-bearing device, the locking clamp being made of a non-metallic material.

[0011] Secondly, embodiments of this application also provide a non-destructive testing method for pressure-bearing equipment with an insulation layer, executed by the non-destructive testing device for pressure-bearing equipment with an insulation layer as described in the first aspect. The non-destructive testing method for pressure-bearing equipment with an insulation layer includes: controlling the ultrasonic transmitter to emit a first ultrasonic signal, the first ultrasonic signal being transmitted to the outer wall of the pressure-bearing equipment via a corresponding waveguide rod; acquiring a second ultrasonic signal received by the acoustic emission sensor, the second ultrasonic signal originating from the outer wall of the pressure-bearing equipment and being transmitted to the acoustic emission sensor via a corresponding waveguide rod; and detecting whether the pressure-bearing equipment has defects based on the second ultrasonic signal.

[0012] In some embodiments, detecting whether the pressure-bearing equipment has defects based on the second ultrasonic signal includes: performing preset processing on the electrical signal corresponding to the second ultrasonic signal to extract signal features; comparing the extracted signal features with preset standard signal features; and detecting whether the pressure-bearing equipment has defects based on the comparison result; wherein the standard signal features are obtained through non-destructive samples of the same specifications as the pressure-bearing equipment to be tested.

[0013] In some embodiments, after detecting a defect in the pressure-bearing equipment based on the second ultrasonic signal, the method further includes: inputting the signal features into a pre-trained defect classification model; and obtaining the type and / or extent of the defect based on the output of the defect classification model.

[0014] In some embodiments, before inputting the signal features into a pre-trained defect classification model, the method further includes: constructing a standardized damage sample database; wherein the damage sample database includes second ultrasonic signal feature data corresponding to defect samples of various known types, sizes, and locations; performing multi-level feature extraction on the second ultrasonic signal data; wherein the features include at least one of time-domain features, frequency-domain features, and time-frequency domain multi-dimensional features; establishing and training a hybrid classification model; wherein the hybrid classification model includes a first-level classifier and a second-level classifier, the first-level classifier using a decision tree-based ensemble learning algorithm to process the high-dimensional feature vector obtained after feature extraction, and the second-level classifier using a one-dimensional convolutional neural network to process the filtered original time-series waveform; obtaining the type and / or degree of the defect based on the output of the defect classification model, including: fusing the outputs of the first-level classifier and the second-level classifier through a weighted voting method to obtain the type and / or degree of the defect.

[0015] This application solves the problems of high detection costs and complex operation caused by traditional methods that require removing the insulation layer or making low-temperature coupling difficult by setting the waveguide rod's front end to pass through the insulation layer and directly adhere to the equipment wall. This eliminates the need for the sensing process to be exposed to a low-temperature environment and the use of a low-temperature coupling agent. Simultaneously, the arc-shaped coupling surface design at the front end maximizes the contact area between the waveguide rod and the equipment wall, reducing energy loss and signal distortion of ultrasonic waves at the contact interface, providing a high-quality initial signal for subsequent reliable analysis. Furthermore, since the waveguide rod can be fixedly installed without removing the insulation layer, it provides a basis for long-term, online, real-time non-destructive testing of pressure-bearing equipment with insulation layers in service, facilitating timely defect detection. It offers advantages such as simple operation, low cost, high detection efficiency, high accuracy, and strong adaptability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced one by one below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 is a three-dimensional structural schematic diagram of a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application; Figure 2 is a front view structural schematic diagram of a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application; Figure 3 is a cross-sectional structural schematic diagram of a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application at the waveguide rod; Figure 4 is a cross-sectional structural schematic diagram of the waveguide rod in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application; Figure 5 is a three-dimensional structural schematic diagram of the waveguide rod in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application; Figure 6 is a three-dimensional structural schematic diagram of the fixing components in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application; Figure 7 is a three-dimensional structural schematic diagram of some fixing components in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application; Figure 8 is a flowchart of a non-destructive testing method for a pressure-bearing device with an insulation layer provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Figure 1 is a three-dimensional structural schematic diagram of a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application. Figure 2 is a front view structural schematic diagram of a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application. Figure 3 is a cross-sectional structural schematic diagram of the non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application at the waveguide rod. Figure 4 is a cross-sectional structural schematic diagram of the waveguide rod in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application. Figure 5 is a three-dimensional structural schematic diagram of the waveguide rod in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application. Figure 6 is a three-dimensional structural schematic diagram of the fixing components in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application. Figure 7 is a three-dimensional structural schematic diagram of some fixing components in a non-destructive testing device for a pressure-bearing device with an insulation layer provided in an embodiment of this application.

[0020] Referring to Figures 1 to 7, the non-destructive testing device for the pressure-bearing equipment 10 with insulation layer includes multiple waveguide rods 1, a fixing component 9, and a controller. At least one of the waveguide rods 1 has an ultrasonic transmitter 5 connected to its rear end, and at least another waveguide rod 1 has an acoustic emission sensor 6 connected to its rear end. The fixing component 9 is used to fix the waveguide rods 1 to the outer wall of the pressure-bearing equipment 10. The controller is communicatively connected to the ultrasonic transmitter 5 and the acoustic emission sensor 6, respectively, and is used to detect whether there are defects in the pressure-bearing equipment 10. The front end of the waveguide rod 1 is provided with an arc-shaped coupling surface 2 that matches the curvature of the outer wall of the pressure-bearing equipment 10. The front end of the waveguide rod 1 passes through the insulation layer 11, and the arc-shaped coupling surface 2 is covered by the insulation layer 11 and fits against the outer wall of the pressure-bearing equipment 10.

[0021] Specifically, the pressure-bearing equipment 10 with an insulation layer refers to equipment such as cryogenic pipelines, high-temperature pipelines, and storage tanks that are externally covered with a layer of insulating material to withstand internal pressure. This is commonly seen in scenarios involving the transportation of cryogenic media such as liquefied natural gas and liquid hydrogen. The accompanying drawings in this application use a cryogenic pipeline as an example for detailed illustration. The waveguide rod 1 is a rod-shaped component used to guide and transmit ultrasonic signals. The ultrasonic transmitter 5 is a device capable of generating and emitting ultrasonic waves of a specific frequency. The acoustic emission sensor 6 is a sensitive element used to receive and convert acoustic signals into electrical signals. The arc-shaped coupling surface 2 refers to the arc-shaped surface at the front end of the waveguide rod 1, which is processed to match the curvature of the outer wall of the pressure-bearing equipment 10, used to achieve efficient transmission of acoustic signals from the waveguide rod 1 to the wall of the pressure-bearing equipment 10.

[0022] At least one waveguide rod 1 has an ultrasonic transmitter 5 at its rear end that generates an ultrasonic signal, which is transmitted along the waveguide rod 1 to its front end. The front end of the waveguide rod 1 is designed with an arc-shaped coupling surface 2 that matches the curvature of the outer wall of the pressure-bearing equipment 10 under test. This end passes through the insulation layer 11, and its arc-shaped coupling surface 2 is covered by the insulation layer 11 and tightly fitted to the outer wall of the pressure-bearing equipment 10. In this way, the ultrasonic signal can be effectively transmitted into the outer wall of the pressure-bearing equipment 10 through the arc-shaped coupling surface 2. When defects exist in the outer wall of the pressure-bearing equipment 10, such as cracks or corrosion, the propagation characteristics of the ultrasonic waves change. These ultrasonic signals carrying defect information—i.e., waves after reflection, scattering, or mode conversion—propagate through the outer wall and are captured by at least one other waveguide rod 1, whose front end is also attached to the outer wall of the pressure-bearing equipment 10. These waves are then received by the acoustic emission sensor 6 at its rear end and converted into electrical signals. The controller is responsible for coordinating the signal transmission of the ultrasonic transmitter 5 and receiving and processing the electrical signals from the acoustic emission sensor 6, thereby completing the detection process.

[0023] Therefore, this embodiment of the application, by setting the front end of the waveguide rod 1 to pass through the insulation layer 11 and directly adhere to the outer wall of the pressure-bearing equipment 10, eliminates the need for the sensing process to be exposed to a low-temperature environment and the use of a low-temperature coupling agent. This solves the problems of high detection costs and complex operation caused by the need to remove the insulation layer 11 or the difficulty of low-temperature coupling in traditional methods. Simultaneously, the design of the arc-shaped coupling surface 2 at the front end maximizes the contact area between the waveguide rod 1 and the outer wall of the pressure-bearing equipment 10, reducing energy loss and signal distortion of ultrasonic waves at the contact interface, providing a high-quality initial signal for subsequent reliable analysis. Furthermore, since the waveguide rod 1 can be fixedly installed without removing the insulation layer 11, it provides a basis for long-term, online, real-time non-destructive testing of the pressure-bearing equipment 10 with the insulation layer 11 in service. This facilitates timely defect detection and offers advantages such as simple operation, low cost, high detection efficiency, high accuracy, and strong adaptability.

[0024] In some embodiments, the front end of the waveguide rod 1 is provided with an acoustic wave converging structure 4, and / or the rear end of the waveguide rod 1 is provided with an acoustic wave converging structure 4; wherein, the acoustic wave converging structure 4 is designed based on the acoustic black hole (ABH) theory. Figure 5 exemplarily shows that the front end of the waveguide rod 1 is provided with an acoustic wave converging structure 4, and the rear end of the waveguide rod 1 is provided with an acoustic wave converging structure 4.

[0025] Specifically, the acoustic wave converging structure 4 is a special geometric structure designed on the body 3 of the waveguide rod 1. Its cross-sectional dimensions, such as thickness or diameter, vary along the length according to the acoustic black hole theory, which can guide and converge the energy of the curved waves propagating within it. The acoustic black hole theory, by designing the geometric or material parameters of the structure according to specific laws, such as the power function theory gradually changing to zero, makes the wave velocity of the curved waves gradually decrease to zero and the phase tend to infinity, thereby guiding the curved waves to propagate to a specific region and converging energy, while suppressing wave reflection.

[0026] When ultrasonic waves propagate in waveguide rod 1, if the cross-sectional dimensions of the rod body 3 of waveguide rod 1 gradually decrease along the axial direction according to the acoustic black hole theory, the wave velocity of the curved wave propagating to this region will slow down, and the wavelength will shorten. Ideally, at the edge where the thickness approaches zero, the wave velocity approaches zero, and the wave will never reach the boundary to be reflected; the energy is captured and concentrated in the black hole region. For example, an acoustic wave converging structure 4 can be provided at the front or rear end of waveguide rod 1, or simultaneously at both ends. Providing an acoustic wave converging structure 4 at the front end helps to more effectively converge and couple the signal from the ultrasonic transmitter 5 into the outer wall of the pressure-bearing device 10, reducing signal reflection loss at the front interface and increasing the energy of the incident signal. Providing an acoustic wave converging structure 4 at the rear end helps to converge and enhance the weak signal carrying defect information from the outer wall before it reaches the acoustic emission sensor 6, improving the signal amplitude and signal-to-noise ratio received by the acoustic emission sensor 6.

[0027] Specifically, in the design of waveguide rod 1 in this embodiment, based on the one-dimensional acoustic black hole theory, the cross section of waveguide rod 1 with acoustic wave converging structure 4 satisfies the following bending wave equation of the one-dimensional acoustic black hole structure:

[0028] in, For bending stiffness, in thin plate theory, bending stiffness is expressed as follows:

[0029] in, It is the density of the material, and the unit is g / cm³. 3 , For the thickness of the thin plate, in the sound wave converging structure 4, the thickness varies. Changes, in mm This is the independent variable, and its physical meaning is the distance from a point on the axis of the acoustic wave converging structure 4 at the end of waveguide rod 1 to its small end face. Time, in the denominator of the aforementioned bending wave equation (i.e., a partial differential equation), represents the sum of the accelerations generated by the material's own motion and the accelerations generated by its deformation. It is a thin plate in The bending displacement at point , its complex form displacement function is: ,in and They are the amplitude and the cumulative phase, respectively, and satisfy the following formula:

[0030]

[0031] Among them, the wave number in the plate , The longitudinal wave velocity is expressed in m / s. It is the frequency of the sound wave. When In When, accumulated phase The value tends towards infinity, indicating that the curved wave will not reach the edge of the sound wave converging structure 4, and therefore there is no reflection. The curved wave can converge to the edge of the sound wave converging structure 4, thus achieving the purpose of energy concentration; among which... is the acoustic black hole constant, and m is the power exponent. Both are structural parameters of the acoustic wave converging structure 4. By exploring their influence on the acoustic emission signal through experimental and numerical simulation methods, suitable structural parameters can be determined.

[0032] Based on the above theory, it is shown that the end acoustic wave converging structure 4 has a good enhancement effect. On this basis, this application optimizes the front and rear structure of the waveguide rod 1 by combining the acoustic black hole theory, and proposes a new type of waveguide rod 1. The waveguide rod 1 has acoustic wave converging structures 4 set at the front and rear ends, and the arc-shaped coupling surface 2 is designed in combination with the outer diameter of the pressure-bearing device 10 under test to ensure that the waveguide rod 1 is in close contact with the pressure-bearing device 10 under test.

[0033] Therefore, in this embodiment, the cross-section of the waveguide rod 1's body 3 is designed based on the one-dimensional acoustic black hole theory. By concentrating sound wave energy through the acoustic black hole effect, it effectively compensates for the energy attenuation of ultrasonic waves during long-distance transmission through the waveguide rod 1 and at the interface of multiple media, thus improving the signal-to-noise ratio and clarity of the signal finally reaching the acoustic emission sensor 6. A stronger received signal means that weak signal changes caused by minute defects can be identified more clearly, thereby improving the sensitivity of detecting minute defects such as microcracks and early corrosion, and the reliability of the overall detection results.

[0034] In some embodiments, the non-destructive testing device for the pressure-bearing equipment 10 with insulation layer includes four waveguide rods 1, one waveguide rod 1 has an ultrasonic transmitter 5 connected to its rear end, three waveguide rods 1 have acoustic emission sensors 6 connected to their rear ends, and the four waveguide rods 1 are symmetrically distributed on the outer wall of the pressure-bearing equipment 10.

[0035] Specifically, the symmetrical distribution of waveguide rods 1 on the outer wall of the pressure-bearing equipment 10 refers to multiple waveguide rods 1 arranged around the circumference or axis of the pressure-bearing equipment 10 in a specific geometrical symmetry relationship, such as axial symmetry or centrosymmetry, to ensure comprehensive detection coverage and balanced signal reception. Four waveguide rods 1 form a symmetrical layout of one transmitter and three receivers. That is, four waveguide rods 1 are used, one of which acts as a transmitter connected to the ultrasonic transmitter 5, and the other three act as receivers connected to an acoustic emission sensor 6. These four waveguide rods 1 are arranged symmetrically around the outer wall of the pressure-bearing equipment 10 under test. For example, the transmitter rod and one receiver rod are located on one side of the pipe, and the other two receiver rods are symmetrically located on the other side, and are evenly distributed around the circumference of the pressure-bearing equipment 10. This layout allows the ultrasonic signal to radiate from the transmitter point in all directions and be captured by multiple receivers located at different spatial positions. The signal received by each receiver reflects the state of the ultrasonic wave after propagation along different paths and its impact by defects.

[0036] Therefore, this embodiment of the application, through multiple symmetrically arranged receiving points, can capture defect scattering or reflection signals from different directions, avoiding the detection blind spots that may exist in single-point reception and improving the detection rate of defects in the entire tested area. By comparing the differences in signal characteristics from different receiving rods, such as arrival time difference and amplitude difference, spatial information can be provided for subsequent analysis, which helps to achieve preliminary judgment or even precise positioning of defect location, as well as more accurate identification of defect type, such as directional cracks. In addition, by comprehensively utilizing the signals from multiple receiving points for analysis, random errors caused by local coupling differences, environmental noise, etc., can be averaged or canceled out, making the detection results more reliable.

[0037] In some embodiments, the waveguide rod 1 is permanently fixed to the outer wall of the pressure-bearing device 10, and the installation of the waveguide rod 1 only requires one insulation layer 11 treatment.

[0038] Specifically, "permanent fixation" means that the waveguide rod 1, after being installed by reliable mechanical means, such as by forcefully locking it with the fixing component 9, remains fixed throughout the entire service life of the equipment or a considerable testing cycle, without the need for frequent disassembly. "One-time insulation layer 11 treatment" means that during the installation of the waveguide rod 1, only a one-time local opening or treatment is needed on the insulation layer 11 corresponding to each installation position of the waveguide rod 1, so that the front end of the waveguide rod 1 can pass through and contact the equipment wall. After installation, the insulation layer 11 can be restored or kept intact, and subsequent testing does not require further damage or treatment of the insulation layer 11.

[0039] The waveguide rod 1 is designed to be permanently fixed to the outer wall of the pressure equipment 10. During installation, the operator only needs to perform a localized, small-scale treatment of the insulation layer 11 at each planned installation location of the waveguide rod 1, such as drilling or grooving, so that the front end of the waveguide rod 1 can pass through the insulation layer 11 to reach the equipment wall. Once the front end of the waveguide rod 1 is in contact with the outer wall of the pressure equipment 10 through its arc-shaped coupling surface 2 and reliably locked by the fixing component 9, the installation is complete. Afterwards, whether for initial testing, subsequent periodic re-inspections, or continuous monitoring, it is not necessary to disassemble the insulation layer 11 or reinstall the waveguide rod 1. For example, for existing pressure equipment 10, installing the above device only requires a small-scale removal of the insulation layer 11, after which the waveguide rod 1 is permanently fixed to the pressure equipment 10, and subsequent testing does not require repeated removal of the insulation layer 11. For pressure equipment 10 planned for use, the device can be laid in conjunction with the pressure equipment 10, and then uniformly covered with the insulation layer 11 before use. Both installation methods allow for the testing of pressure equipment 10 without removing the insulation layer 11, significantly reducing testing costs and time.

[0040] Therefore, this embodiment avoids the tedious and costly process of removing, inspecting, and restoring the insulation layer 11 for each test, which is particularly suitable for occasions requiring frequent or long-term monitoring, resulting in significant overall economic benefits. The one-time installation causes minimal damage to the insulation layer 11, and the insulation performance is effectively restored and maintained after installation, reducing interference with normal equipment operation, especially for low-temperature equipment with strict insulation requirements. Simultaneously, the permanently fixed waveguide rod 1 ensures that the coupling state and relative position of the acoustic emission sensor 6 are completely identical for each test, eliminating errors caused by changes in coupling state due to repeated installation and improving the accuracy of long-term monitoring data.

[0041] In some embodiments, the waveguide rod 1 is made of a low-temperature resistant material that matches the acoustic impedance of the pressure-bearing device 10.

[0042] Specifically, acoustic impedance matching means that the acoustic impedance value of the selected waveguide rod 1 material is as close as possible to the acoustic impedance value of the main material of the pressure-bearing equipment 10 under test. Acoustic impedance is the product of the material density and the speed of sound propagation in the material, and it is a key physical quantity that determines the reflection and transmission behavior of sound waves at the interface between two media. Low-temperature resistant materials refer to materials that can maintain good mechanical and physical properties in low-temperature environments.

[0043] Specifically, when an ultrasonic wave is incident perpendicularly from the first medium to the planar interface formed by the first and second media, reflection and transmission phenomena will occur. That is, part of the sound energy is reflected and propagates in the opposite direction to the incident wave; the other part of the sound energy passes through the interface and propagates in the same direction as the incident wave. i Indicates the incident wave acoustic intensity, I r Indicates the intensity of the reflected wave, I t Let Z1 represent the acoustic impedance of the first medium and Z2 represent the acoustic impedance of the second medium. Let the sound pressure of the incident wave be p. i The sound pressure of the reflected wave is p r The sound pressure of the transmitted wave is p t The ratio of the sound pressure of the reflected wave to the sound pressure of the incident wave at the interface is called the sound pressure reflectivity, denoted by R. p It represents and satisfies the following formula:

[0044] The ratio of the transmitted wave sound pressure to the incident wave sound pressure at the interface is called the sound pressure transmittance, denoted by T. p It represents and satisfies the following formula:

[0045] According to the law of conservation of energy, the total sound pressure on both sides of the interface is always equal, and the amplitude of the particle vibration velocity is also equal. Therefore, the following boundary conditions apply:

[0046] By combining the above formulas, we can obtain the sound pressure reflectivity R. p and sound pressure transmittance T p They respectively satisfy the following formulas:

[0047]

[0048] As can be seen from the derivation of the above formulas, sound pressure reflectivity and transmittance are determined by acoustic impedance. For a pipe wall thickness measurement system based on a waveguide structure, when the acoustic impedance of the waveguide rod 1 and the pressure-bearing device 10 being measured is close (i.e., Z1≈Z2), the sound pressure reflectivity at the interface is very low, and the sound pressure transmittance is close to 1, resulting in almost full transmission. Therefore, the material of the waveguide rod 1 should be selected to be close to the acoustic impedance of the pressure-bearing device 10 being measured, to ensure that most of the sound energy is transmitted into the pressure-bearing device 10. For example, if the pressure-bearing device 10 is a stainless steel pipe, the waveguide rod 1 can be made of 316L stainless steel. For example, as shown in Figure 4, the diameter of the rod body 3 of the waveguide rod 1 can be set to 6mm. To avoid the difficulty in fixing a long waveguide rod 1 during the experiment, the overall length of the waveguide rod 1 can be set to 218mm, and the axial length of the arc-shaped coupling surface 2 can be set to 40mm to ensure sufficient contact area and reduce signal distortion. Taking the pressure-bearing device 10 as a cryogenic pipe as an example, the outer diameter of the cryogenic pipe can be 152mm. In addition, the radius of the frustum 7 at the end of the waveguide rod 1 can be set to 60mm and the thickness to 2mm, so as to stably support the acoustic emission sensor 6 and the ultrasonic transmitter 5 and ensure less energy loss.

[0049] Therefore, this embodiment significantly reduces the energy reflection loss of ultrasonic waves at the interface between the front end of the waveguide rod 1 and the outer wall of the pressure-bearing device 10 through acoustic impedance matching. This allows the transmitted signal to be transmitted to the outer wall of the pressure-bearing device 10 more efficiently, and also allows weak defect signals from the outer wall of the pressure-bearing device 10 to be transmitted to the receiving waveguide rod 1 more efficiently, thereby ensuring that the detection system has a sufficiently strong effective signal. The use of low-temperature resistant materials ensures that the waveguide rod 1 itself will not experience performance degradation or damage in extreme low-temperature operating environments, guaranteeing the long-term stability and durability of the entire detection device in low-temperature environments.

[0050] In some embodiments, referring to Figures 6 and 7, the fixing component 9 includes at least one locking clamp, which covers the rod body 3 portion of the outer waveguide rod 1 of the insulation layer 11 and is securely connected to the pressure-bearing device 10. The locking clamp is made of non-metallic material.

[0051] Specifically, the locking clamp, for example, is composed of two or more parts and is a ring or semi-ring clamp that can be locked with fasteners such as bolts, used to hold and fix the waveguide rod 1. The locking clamp is designed to cover the rod body 3 of the waveguide rod 1 located outside the insulation layer 11. By tightening the bolts connecting the various parts of the locking clamp and between the locking clamp and the pressure-bearing device 10, a strong clamping force and downward pressure are generated. This force, on the one hand, firmly fixes the waveguide rod 1 in the preset position, preventing it from loosening or shifting; on the other hand, it is transmitted to the front end through the rod body 3 of the waveguide rod 1, ensuring that the arc-shaped coupling surface 2 at the front end of the waveguide rod 1 maintains a tight and stable contact pressure with the outer wall of the pressure-bearing device 10, achieving good acoustic coupling. Using non-metallic materials such as ABS (Acrylonitrile Butadiene Styrene) to make the locking clamp can avoid unnecessary interference, scattering, or absorption of ultrasonic signals propagating in its vicinity by metallic materials.

[0052] For example, the radius of curvature of the curved portion of the locking clamp can be set to be equal to the outer diameter of the insulation layer 11, and the inner diameter of the portion of the waveguide rod 1 body 3 covered by the insulation layer 11 can be the same as the outer diameter of the waveguide rod 1 body 3, to ensure that the waveguide rod 1 body 3 is locked on the outer side of the insulation layer 11. Referring to Figures 1, 6 and 7, a locking clamp can be set including a first locking clamp portion 91, a second locking clamp portion 92, a third locking clamp portion 93 and a fourth locking clamp portion 94. The first locking clamp portion 91 and the second locking clamp portion 92 cover the waveguide rod 1 body 3 in the upper left corner of Figure 2, and the third locking clamp portion 93 and the fourth locking clamp portion 94 cover the waveguide rod 1 body 3 in the lower left corner of Figure 2. Another locking clamp includes a fifth locking clamp section 95, a sixth locking clamp section 96, a seventh locking clamp section 97, and an eighth locking clamp section 98. The fifth locking clamp section 95 and the sixth locking clamp section 96 cover the body 3 of the waveguide rod 1 in the upper right corner of Figure 2, while the seventh locking clamp section 97 and the eighth locking clamp section 98 cover the body 3 of the waveguide rod 1 in the lower right corner of Figure 2. The first locking clamp section 91 is connected to the second locking clamp section 92, the third locking clamp section 93 is connected to the fourth locking clamp section 94, the fifth locking clamp section 95 is connected to the sixth locking clamp section 96, and the seventh locking clamp section 97 is connected to the eighth locking clamp section 98, each connected by four bolts. The first locking clamp section 91 and the fourth locking clamp section 94, the second locking clamp section 92 and the third locking clamp section 93, the fifth locking clamp section 95 and the eighth locking clamp section 98, and the sixth locking clamp section 96 and the seventh locking clamp section 97 are each connected by a bolt to ensure that the waveguide rod 1 and the pressure-bearing equipment 10 under test are in close contact and do not move relative to each other during the testing process.

[0053] Therefore, the locking clamp structure of this embodiment is simple and reliable, providing sufficient fastening force to ensure that the waveguide rod 1 maintains stable coupling with the equipment wall under conditions such as vibration and temperature changes. The non-metallic locking clamp avoids introducing metallic heterostructures near the ultrasonic signal path, reducing additional scattering and energy loss that the signal may encounter during propagation, helping to maintain signal purity and simplifying subsequent signal analysis. The locking clamp structure is easy to install and adjust on-site and is suitable for use on equipment with different pipe diameters.

[0054] In summary, the embodiments of this application transmit ultrasonic signals through waveguide rod 1, which can effectively suppress energy attenuation and extend signal duration, thereby improving the accuracy and reliability of acoustic emission detection and providing technical support for the safe operation of cryogenic pipelines. An ultrasonic wave is emitted by ultrasonic transmitter 5, the waveguide rod 1 conducts and amplifies the signal, and the acoustic emission sensor 6 receives the signal. Based on the comparison of acoustic signals under different micro-defects, a quantitative relationship between signal characteristics and micro-defects is established, enabling non-destructive testing and location of defects such as micro-cracks and micro-corrosion on the inner wall of cryogenic pipelines. This efficiently identifies internal defects of pressure equipment 10 under service conditions, not only simplifying the non-destructive testing steps of pressure equipment 10 but also achieving higher detection efficiency and more accurate results. The use of an ultrasonic source during detection results in a more stable and reliable signal compared to other sound sources.

[0055] This application embodiment also provides a non-destructive testing method for pressure-bearing equipment with an insulation layer, which is executed by the non-destructive testing device for pressure-bearing equipment with an insulation layer as described in the above embodiments. Figure 8 is a schematic flowchart of the non-destructive testing method for pressure-bearing equipment with an insulation layer provided in this application embodiment. As shown in Figure 8, the non-destructive testing method for pressure-bearing equipment with an insulation layer includes the following steps: S101, controlling an ultrasonic transmitter to emit a first ultrasonic signal, the first ultrasonic signal being transmitted to the outer wall of the pressure-bearing equipment via a corresponding waveguide rod.

[0056] S102. Acquire the second ultrasonic signal received by the acoustic emission sensor. The second ultrasonic signal comes from the outer wall of the pressure-bearing equipment and is transmitted to the acoustic emission sensor through the corresponding waveguide rod.

[0057] S103. Detect whether there are defects in the pressure-bearing equipment based on the second ultrasonic signal.

[0058] Specifically, the first ultrasonic signal refers to the initial ultrasonic signal generated by the ultrasonic transmitter 5 and transmitted to the outer wall of the pressure-bearing device 10 through the transmitting waveguide 1. The second ultrasonic signal refers to the ultrasonic signal that, after propagating in the outer wall of the pressure-bearing device 10 and interacting with existing defects, is captured by the receiving waveguide 1 and transmitted to the acoustic emission sensor 6, and contains characteristic information of the defects in the pressure-bearing device 10.

[0059] First, the controller instructs the ultrasonic transmitter 5 to operate, generating a first ultrasonic signal. This first ultrasonic signal is propagated along the transmitting waveguide 1 connected to it and efficiently injected into the outer wall of the pressure-bearing equipment 10 through the coupling surface at its front end. As the first ultrasonic signal propagates within the outer wall, if it encounters defects such as cracks or corrosion pits, it undergoes changes such as reflection, scattering, mode conversion, or energy attenuation. These altered ultrasonic waves, i.e., the second ultrasonic signal, continue to propagate within the outer wall and are eventually captured by the front end of the receiving waveguide 1 at another location. The signal is then propagated along the receiving waveguide 1 to the acoustic emission sensor 6 at its end. The acoustic emission sensor 6 converts the received sound waves into electrical signals. After acquiring these electrical signals, the controller analyzes and processes them using a built-in algorithm to determine whether the pressure-bearing equipment 10 has any defects.

[0060] Therefore, in this embodiment, a non-destructive testing device is installed on the pressure-bearing equipment 10 to be tested. The distribution of waveguide rods 1 on the pressure-bearing equipment 10 is positioned according to a one-transmitter, three-receiver pattern. The arc-shaped coupling surface 2 at the front end of each waveguide rod 1 is brought into contact with the outer wall of the pressure-bearing equipment 10, and the insulation layer 11 is restored and locked in place by the fixing component 9. The ultrasonic transmitter 5 and the acoustic emission sensor 6 are then tightly attached to the corresponding ends of the waveguide rods 1. Then, the ultrasonic transmitter 5 transmits a first ultrasonic signal, and the acoustic emission sensor 6 receives a second ultrasonic signal from the outer wall of the pressure-bearing equipment 10, thus achieving non-destructive testing of the pressure-bearing equipment 10.

[0061] In some embodiments, detecting whether the pressure-bearing device 10 has a defect based on the second ultrasonic signal includes: performing preset processing on the electrical signal corresponding to the second ultrasonic signal to extract signal features; comparing the extracted signal features with preset standard signal features; and detecting whether the pressure-bearing device 10 has a defect based on the comparison result; wherein the standard signal features are obtained by using a non-destructive sample with the same specifications as the pressure-bearing device 10 to be tested.

[0062] Specifically, the preset processing refers to a series of pre-defined digital electrical signal processing steps, the purpose of which is to extract characteristic information that can effectively characterize defects from the electrical signal corresponding to the original second ultrasonic signal. Signal characteristics refer to quantitative indicators calculated from the processed electrical signal, such as the average amplitude, energy, and peak frequency of a specific frequency band. The preset standard signal characteristics refer to the signal characteristic values ​​obtained by collecting and processing from a known intact and defect-free sample pressure device 10 with the same specifications as the pressure device 10 under test, using the same device and method, as a benchmark for the health status of the pressure device 10. For example, the process of obtaining the preset standard signal characteristics may include: selecting a non-destructive sample pressure device 10 with the same material and specifications as the pressure device 10 under test; installing a non-destructive testing device on it to collect multiple sets of acoustic emission signals, i.e., the second ultrasonic signal; performing noise reduction and fast Fourier transform processing on the electrical signal obtained by converting the second ultrasonic signal through the acoustic emission sensor 6; dividing the transformed spectrum into high-frequency and low-frequency bands, and calculating the average amplitude of the high-frequency band as the standard signal characteristics of the sample pressure device 10.

[0063] After acquiring the electrical signal corresponding to the second ultrasonic signal, the controller does not directly observe the original waveform, but performs a series of preset processing steps on it. For example, the preset processing flow may include: Step 1: Noise denoising, for example, using a moving average method to filter out some environmental and electrical noise contained in the electrical signal corresponding to the second ultrasonic signal, improving signal clarity. For example, the first and last 10 signal points of the original signal can be discarded, and the arithmetic mean of the 10 data points before and after the discarded signal points can be taken, replacing the original signal points with the calculated arithmetic mean. Step 2: Feature extraction, for example, performing a fast Fourier transform on the electrical signal corresponding to the denoised second ultrasonic signal to convert the time-domain signal into a frequency-domain signal, and then calculating the average amplitude of a specific high-frequency band as the signal feature. For example, the time-amplitude view of the signal can be converted to a frequency-amplitude view, thus clearly showing the core frequency components constituting the signal, and processing the high-frequency amplitude data to obtain experimental values, i.e., extracting signal features. Step 3: Comparison and judgment. The signal characteristics of the pressure-bearing equipment 10 calculated above are compared with the standard signal characteristics obtained in advance from intact samples. If the difference between the two exceeds a certain preset threshold, the pressure-bearing equipment 10 is considered to have a defect; otherwise, the pressure-bearing equipment 10 is considered to be intact.

[0064] Therefore, this embodiment of the application, through signal processing and feature extraction, transforms complex waveform comparisons into comparisons of specific values, reducing reliance on operator experience and making the judgment criteria unified and objective. Targeted signal processing can highlight the essential changes in signals caused by defects, suppress interference, and make the judgment based on feature value differences more accurate and reliable. In addition, by establishing a connection between the micro-defects of the pressure-bearing equipment 10 and the second ultrasonic signal, the detection and location of defects in the pressure-bearing equipment 10 under test can be achieved.

[0065] The following is a detailed description of the standard signal feature acquisition process and defect diagnosis process using a specific embodiment: The standard signal feature acquisition process includes: First, for existing pressure equipment 10, a non-in-service pressure equipment 10 with the same material and dimensional parameters as the pressure equipment 10 to be tested is selected as the non-destructive testing equipment 10; for pressure equipment 10 planned for use, pressure equipment 10 from the same batch of manufacturers with the same material and dimensional parameters is selected as the non-destructive testing equipment 10. Second, the non-destructive testing device for the pressure equipment 10 is installed on the selected non-destructive testing equipment 10. Third, the ultrasonic transmitter 5 and three acoustic emission sensors 6 are simultaneously turned on. The ultrasonic transmitter 5 is used to emit a first ultrasonic signal, the acoustic emission sensors 6 are used to receive a second ultrasonic signal, and the waveguide rod 1 is used to conduct the ultrasonic signal. Each acoustic emission sensor 6 collects, for example, 1500 to 2000 sets of second ultrasonic signal data. The electrical signal data tags corresponding to the second ultrasonic signal data are selected for amplitude, energy, ring count, and arrival time. The fourth step involves denoising the acquired amplitude signal using the moving average denoising method to ensure that the filtered signal is not distorted. The fifth step involves performing a Fast Fourier Transform (FFT) on the denoised amplitude data to convert the time-amplitude view of the signal to a frequency-amplitude view, thus clearly showing the core frequency components that constitute the signal. The sixth step involves dividing the discrete amplitude obtained after the FFT into high-frequency and low-frequency bands, calculating the average amplitude of the high-frequency band, and taking the arithmetic mean of the average amplitudes corresponding to the three acoustic emission sensors 6 as the standard value, i.e., the standard signal characteristic, to minimize random errors. It should be noted that the above standard signal characteristic acquisition process is based on an embodiment of a cryogenic liquefied natural gas pipeline. For other cryogenic pipelines or pressure-bearing equipment 10, a similar method can be adopted when selecting non-destructive pressure-bearing equipment 10.

[0066] The defect diagnosis process for the pressure-bearing equipment 10 includes: repeating steps two through five of the standard signal feature acquisition process described above, processing the data collected from the pressure-bearing equipment 10 under test to obtain the core frequency components of the signal. Then, the discrete amplitude obtained after the fast Fourier transform is divided into high-frequency and low-frequency bands, and the average amplitude of the high-frequency band is calculated. The arithmetic mean of the average amplitudes corresponding to the acoustic emission sensors 6 on the three waveguide rods 1 is taken to obtain the experimental value, i.e., the signal feature is extracted. Finally, the signal features of the pressure-bearing equipment 10 under test obtained above are compared with the standard signal features obtained in advance from the non-destructive pressure-bearing equipment 10. If the difference between the two exceeds a certain preset threshold, the pressure-bearing equipment 10 under test is considered to have a defect; otherwise, the pressure-bearing equipment 10 is considered to be intact and undamaged.

[0067] In some embodiments, after detecting a defect in the pressure vessel 10 based on the second ultrasonic signal, the method further includes: inputting signal features into a pre-trained defect classification model; and obtaining the type and / or extent of the defect based on the output of the defect classification model.

[0068] Specifically, the defect classification model is a trained machine learning or pattern recognition model that can automatically output predictions or classifications of defect categories and / or severity based on input signal features. Defect categories include cracks and corrosion, while severity includes depth and length. When a defect is determined to exist in the pressure equipment 10, the controller does not stop there but feeds the extracted signal features, or a feature set containing more dimensions, into a pre-trained defect classification model. This model has internally established a complex mapping relationship from various signal features to specific defect types and degrees by learning from a large amount of historical data. After calculating and analyzing the input features, the model directly outputs qualitative and / or quantitative assessments of the current defect. Qualitative results include, for example, surface microcracks, while quantitative results include, for example, a corrosion depth of approximately 2 mm.

[0069] Therefore, the embodiments of this application go beyond simple defect presence or absence judgment, providing detailed information about defect type and severity, greatly increasing the information content and practical value of the detection results, and directly guiding maintenance decisions. For the complex nonlinear relationship between features and defect types, a well-trained model can effectively identify and distinguish them, especially suitable for different defect types with similar characteristics, reducing the reliance on senior technicians for detection results.

[0070] In some embodiments, before inputting the signal features into a pre-trained defect classification model, the method further includes: constructing a standardized damage sample database; wherein the damage sample database includes second ultrasonic signal feature data corresponding to defect samples of various known types, sizes, and locations; performing multi-level feature extraction on the second ultrasonic signal data; wherein the features include at least one of time-domain features, frequency-domain features, and time-frequency domain multi-dimensional features; establishing and training a hybrid classification model; wherein the hybrid classification model includes a first-level classifier and a second-level classifier, the first-level classifier uses a decision tree-based ensemble learning algorithm to process the high-dimensional feature vector obtained after feature extraction, and the second-level classifier uses a one-dimensional convolutional neural network to process the filtered original time-series waveform; obtaining the type and / or degree of the defect based on the output of the defect classification model, including: fusing the outputs of the first-level classifier and the second-level classifier through a weighted voting method to obtain the type and / or degree of the defect.

[0071] Specifically, the damage sample database is a systematically stored collection of data, where each data entry corresponds to a known defect state, such as the defect type, size, and location. It also includes multiple secondary ultrasonic signal feature data collected from that sample; these secondary ultrasonic signal feature data refer to the electrical signal feature data corresponding to the secondary ultrasonic signal. Multi-level feature extraction refers to extracting features of different types and abstraction levels from the original signal data, such as time-domain statistical features like peak value or root mean square, frequency-domain spectral features like dominant frequency or spectral moments, and time-frequency domain features like wavelet packet energy. A hybrid classification model is a composite model composed of multiple classifiers with different principles or structures. The first level is an ensemble learning algorithm based on decision trees, such as random forests or gradient boosting trees, while the second level is a one-dimensional convolutional neural network. Weighted voting is a model fusion strategy where each sub-classifier makes its own judgment on the input, and the final result is determined by a weighted sum based on the pre-set weights of each sub-classifier.

[0072] First, a damage sample database is constructed. Using a pressure-bearing device 10 of the same material and specifications as the device under test, various typical defects are artificially prepared. Multiple sets of second ultrasonic signal feature data are collected from each defect sample using a detection device, and the defect type, size, and location label of each sample are accurately recorded to form a standardized training dataset. Multi-level feature extraction is performed on the original signals of each sample in the database to generate a high-dimensional feature vector containing time-domain, frequency-domain, and time-frequency-domain features. Furthermore, signals from multiple acoustic emission sensors 6 can be fused to calculate spatial correlation features such as signal arrival time difference for subsequent localization.

[0073] A two-stage hybrid classification model is established. The first-stage classifier processes the high-dimensional feature vectors generated in the preceding steps. It excels at handling structured features, automatically assesses feature importance, and exhibits good generalization ability. The second-stage classifier directly inputs the filtered raw time-series waveform. It can automatically learn local details and deep abstract patterns from the raw signal without relying on manually designed features. In model application, the same input sample is simultaneously processed by both classifiers, with each classifier outputting its predicted probability distribution for the defect type. The final diagnostic result is obtained by fusing these two outputs through a weighted voting method, combining the advantages of both models.

[0074] For defect localization, location labels recorded in the database and extracted spatial correlation features, such as time difference of arrival, can be used. During training, the model can learn the relationship between signal features and defect location, thereby achieving three-dimensional spatial localization of defects during diagnosis.

[0075] Therefore, the hybrid model in this embodiment combines the advantages of feature engineering with deep learning, statistical learning with neural networks, enabling a more comprehensive and in-depth extraction of defect information from signals. Compared to a single model, it achieves higher classification accuracy and generalization performance. Multi-level feature extraction provides rich information, and weighted voting fusion reduces the risk of overfitting or misjudgment by a single model, making the overall model more robust when facing noise or new variant defects. Furthermore, it not only achieves intelligent classification of defect types but also integrates precise defect localization into the model framework.

[0076] The machine learning process is described in detail below with a specific embodiment: First, a standardized damage sample database is constructed. Pipe sections of the same material and specifications as the pipe to be tested are selected, and various typical defects of known types and sizes are artificially prefabricated, including but not limited to surface microcracks, moderate corrosion, deep cracks, and localized pits. Using a non-destructive testing device, the raw time-series signals of amplitude, ring count, and energy corresponding to three acoustic emission sensors 6 are collected for each defect sample. The geometric parameters and spatial location information of the defect are recorded simultaneously, forming a standardized training dataset with multi-dimensional labels.

[0077] Secondly, multi-level feature engineering and extraction are performed. For the acquired raw time-series signals, time-domain, frequency-domain, and time-frequency-domain feature extraction are performed sequentially. Time-domain extraction includes statistical features such as peak value, root mean square (RMS), and waveform indices; frequency-domain extraction uses Fast Fourier Transform (FFT) results to extract dominant frequency, center frequency, and spectral moment features; and time-frequency domain extraction uses wavelet packet decomposition to obtain the energy distribution features of each sub-band. Furthermore, signals from multiple acoustic emission sensors are fused, and their arrival time differences and cross-correlation coefficients are calculated to construct spatial correlation features, ultimately generating high-dimensional feature vectors characterizing various damage modes.

[0078] Next, a hybrid classification model was built and trained. A hierarchical modeling strategy was adopted. The first-level classifier used an ensemble learning algorithm based on decision trees to process high-dimensional feature vectors; the second-level classifier used a one-dimensional convolutional neural network to directly process the filtered original time-series waveform to capture local modal features.

[0079] Finally, the outputs of the two-level classifiers are fused using a weighted voting method to obtain the final damage type determination result. During training, cross-validation and grid search can be used to optimize the model hyperparameters, and class weight adjustments are introduced to address the imbalanced sample problem.

[0080] In addition, the sample database in the above machine learning method synchronously records the geometric parameters and spatial location information of the defects. When performing machine learning analysis on the signal data of the pipeline under test, the relative spatial coordinates of the four waveguide rods 1 of the marking device are used, and the signals of the three acoustic emission sensors 6 are fused. The model calculates the arrival time difference and cross-correlation coefficient to construct spatial correlation features. After the spatial correlation features are processed into three-dimensional visualization, the defects can be accurately located.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A non-destructive testing device for pressure-bearing equipment with an insulation layer, characterized in that, include: The system includes multiple waveguide rods, at least one of which has an ultrasonic transmitter connected to its rear end, and at least another waveguide rod has an acoustic emission sensor connected to its rear end; a fixing assembly for fixing the waveguide rods to the outer wall of the pressure-bearing equipment; and a controller, communicatively connected to both the ultrasonic transmitter and the acoustic emission sensor, for detecting defects in the pressure-bearing equipment. The waveguide rods have an arc-shaped coupling surface at their front end that matches the curvature of the outer wall of the pressure-bearing equipment. The front end of the waveguide rod passes through the insulation layer, and the arc-shaped coupling surface is covered by the insulation layer and adheres to the outer wall of the pressure-bearing equipment.

2. The non-destructive testing device for pressure-bearing equipment with insulation layer according to claim 1, characterized in that, The waveguide rod has an acoustic wave converging structure at its front end and / or an acoustic wave converging structure at its rear end; wherein the acoustic wave converging structure is designed based on the acoustic black hole theory.

3. The non-destructive testing device for pressure-bearing equipment with insulation layer according to claim 1, characterized in that, include: Four waveguide rods are symmetrically distributed on the outer wall of the pressure-bearing equipment. One waveguide rod is connected to an ultrasonic transmitter at its rear end, and the other three waveguide rods are connected to acoustic emission sensors at their rear ends.

4. The non-destructive testing device for pressure-bearing equipment with insulation layer according to claim 1, characterized in that, The waveguide rod is permanently fixed to the outer wall of the pressure-bearing equipment, and the installation of the waveguide rod only requires one insulation layer treatment.

5. The non-destructive testing device for pressure-bearing equipment with insulation layer according to any one of claims 1-4, characterized in that, The waveguide rod is made of a low-temperature resistant material that matches the acoustic impedance of the pressure-bearing equipment.

6. The non-destructive testing device for pressure-bearing equipment with insulation layer according to any one of claims 1-4, characterized in that, The fixing component includes at least one locking clamp, which covers the body portion of the waveguide rod outside the insulation layer and is securely connected to the pressure-bearing device. The locking clamp is made of non-metallic material.

7. A non-destructive testing method for pressure-bearing equipment with insulation layer, characterized in that, Performed by the non-destructive testing device for pressure-bearing equipment with insulation layer as described in any one of claims 1-6, the non-destructive testing method for pressure-bearing equipment with insulation layer includes: controlling the ultrasonic transmitter to emit a first ultrasonic signal, the first ultrasonic signal being transmitted to the outer wall of the pressure-bearing equipment via a corresponding waveguide rod; acquiring a second ultrasonic signal received by the acoustic emission sensor, the second ultrasonic signal originating from the outer wall of the pressure-bearing equipment and being transmitted to the acoustic emission sensor via a corresponding waveguide rod; and detecting whether there is a defect in the pressure-bearing equipment based on the second ultrasonic signal.

8. The non-destructive testing method for pressure-bearing equipment with insulation layer according to claim 7, characterized in that, Detecting whether the pressure-bearing equipment has defects based on the second ultrasonic signal includes: performing preset processing on the electrical signal corresponding to the second ultrasonic signal to extract signal features; comparing the extracted signal features with preset standard signal features; and detecting whether the pressure-bearing equipment has defects based on the comparison result; wherein the standard signal features are obtained through non-destructive samples of the same specifications as the pressure-bearing equipment to be tested.

9. The non-destructive testing method for pressure-bearing equipment with insulation layer according to claim 8, characterized in that, After detecting a defect in the pressure-bearing equipment based on the second ultrasonic signal, the method further includes: inputting the signal features into a pre-trained defect classification model; and obtaining the type and / or degree of the defect based on the output of the defect classification model.

10. The non-destructive testing method for pressure-bearing equipment with insulation layer according to claim 9, characterized in that, Before inputting the signal features into the pre-trained defect classification model, the method further includes: constructing a standardized damage sample database; wherein the damage sample database includes second ultrasonic signal feature data corresponding to defect samples of various known types, sizes, and locations; performing multi-level feature extraction on the second ultrasonic signal data; wherein the features include at least one of time-domain features, frequency-domain features, and time-frequency domain multi-dimensional features; establishing and training a hybrid classification model; wherein the hybrid classification model includes a first-level classifier and a second-level classifier, the first-level classifier uses a decision tree-based ensemble learning algorithm to process the high-dimensional feature vector obtained after feature extraction, and the second-level classifier uses a one-dimensional convolutional neural network to process the filtered original time-series waveform; obtaining the type and / or degree of the defect based on the output of the defect classification model, including: fusing the outputs of the first-level classifier and the second-level classifier through a weighted voting method to obtain the type and / or degree of the defect.

Images