A method and system for online monitoring of high-pressure gas pipeline leaks

Abstract

This application discloses an online monitoring method and system for high-pressure gas pipeline leaks, relating to the field of gas leak detection. The method includes acquiring a monitoring image; determining a displacement vector map based on the monitoring image and a reference image; the reference image being an image of the high-pressure gas pipeline under test when no leak exists; obtaining a test field region based on the monitoring image; the test field region being a square region containing a jet cone; the jet cone being a conical region generated by the gas leak when a gas leak exists in the high-pressure gas pipeline under test; determining the refractive index at the boundary of the test field region; determining the density image of the test field region based on the displacement vector map and the refractive index at the boundary of the test field; and performing online monitoring of the density image of the test field based on a convolutional neural network to obtain monitoring results; the monitoring results include whether a leak exists in the high-pressure gas pipeline under test. This application achieves low-cost, high-sensitivity online leak monitoring.

Description

Technical Field

[0001] This application relates to the field of gas leak detection technology, and in particular to an online monitoring method and system for high-pressure gas pipeline leaks. Background Technology

[0002] The operation of large-scale special electrical equipment involves the pipeline transportation and storage of various flammable gases and high-temperature, high-pressure gases. Leaks can cause not only chemical pollution but, in severe cases, explosions, endangering the safety of on-site personnel and causing irreversible equipment damage. Therefore, online monitoring of high-pressure gas pipeline leaks is essential.

[0003] Currently, relevant gas leak monitoring methods in industry include the soap film method, fixed-point catalytic combustion method, infrared thermal imaging method, and schlieren method. However, these technologies struggle to achieve both low cost and high sensitivity for online monitoring of high-pressure gas pipeline leaks. Gas leaks remain a significant safety hazard for workers operating large-scale special electrical equipment. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for online monitoring of high-pressure gas pipeline leaks, which can realize low-cost and high-sensitivity online leak monitoring of high-pressure gas pipelines in large special electrical equipment.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] Firstly, this application provides an online monitoring method for high-pressure gas pipeline leaks, including:

[0007] Acquire monitoring images; the monitoring images are images obtained in real time from the high-pressure gas pipeline under test;

[0008] A displacement vector diagram is determined based on the monitoring image and the reference image; the reference image is an image of the high-pressure gas pipeline under test when there is no leakage.

[0009] The test area is obtained based on the monitoring image; the test area is a square area containing a jet cone; the jet cone is a cone-shaped area generated by the gas leak when there is a gas leak in the high-pressure gas pipeline under test;

[0010] Determine the refractive index at the boundary of the test field region;

[0011] The density image of the test field region is determined based on the displacement vector diagram and the refractive index at the boundary of the test field.

[0012] The density image of the test area is monitored online using a convolutional neural network to obtain monitoring results; the monitoring results include whether there is a leak in the high-pressure gas pipeline under test.

[0013] Optionally, determining a displacement vector map based on the monitored image and the reference image specifically includes:

[0014] Image enhancement is performed on the monitoring image based on the principle of overlay.

[0015] The enhanced monitoring image and the reference image are cross-correlated to obtain a displacement vector map; the displacement vector map is used to represent the position vector relationship of the same ray on the monitoring image and the reference image.

[0016] Optionally, determining the refractive index at the boundary of the test field region specifically includes:

[0017] Substituting the gas pressure and gas temperature inside the high-pressure gas pipeline into the ideal gas law, the gas density at the boundary of the test field is calculated.

[0018] The gas density at the boundary of the test field is substituted into the Glaston-Dale equation to calculate the refractive index at the boundary of the test field.

[0019] Optionally, the enhanced monitoring image and the reference image are cross-correlated to obtain a displacement vector map, specifically including:

[0020] The enhanced monitoring image is calibrated to obtain a calibrated monitoring image;

[0021] A query window is used to perform cross-correlation calculations between the calibrated monitoring image and the reference image to obtain a displacement vector diagram; the size of the query window is selected according to the different gas pressures in the high-pressure gas pipeline to be measured.

[0022] Optionally, determining the density image of the region to be measured based on the displacement vector diagram and the refractive index at the boundary of the field to be measured specifically includes:

[0023] Displacement vector data is determined based on the displacement vector diagram; the displacement vector data includes: the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image, in the X and Y directions; the background is the camera focusing position;

[0024] Substitute the refractive index at the boundary of the field to be measured and the displacement vector data into the Poisson equation and iteratively solve to obtain the refractive index of the field to be measured.

[0025] Substituting the refractive index of the field to be measured into the Glaston-Dale equation, the density image of the field to be measured is calculated.

[0026] Optionally, the formula for calculating the gas density at the boundary of the field to be measured is:

[0027]

[0028] Where ρ is the gas density at the boundary of the test field, P is the gas pressure inside the high-pressure gas pipeline, M is the molar mass obtained based on the type of gas inside the high-pressure gas pipeline, and R... g Let T be the ideal gas constant, and T be the gas temperature inside the high-pressure gas pipeline to be measured.

[0029] Optionally, the formula for calculating the refractive index at the boundary of the field to be measured is:

[0030] n0 = kρ + 1;

[0031] Where n0 is the refractive index at the boundary of the field to be measured, k is the Glaston-Dale coefficient, and ρ is the gas density at the boundary of the field to be measured.

[0032] Optionally, the formula for calculating the refractive index of the field to be measured is:

[0033]

[0034] Where n is the refractive index of the field to be measured, n0 is the refractive index at the boundary of the gas to be measured at room temperature and pressure, W is the thickness of the field to be measured, and Z is the refractive index of the gas to be measured at room temperature and pressure. d Let x be the distance between the area to be tested and the background, x be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the X direction, y be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, x be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, y ... x Δ is the displacement in the X direction of the difference between the position of a ray emitted from the same location in the background after passing through the test area in the monitoring image and the initial position of the same ray emitted from the same location in the background in the reference image. y The displacement in the Y direction is the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image.

[0035] Optionally, after obtaining the monitoring results by online monitoring of the density image of the area to be tested based on a convolutional neural network, the online monitoring method for high-pressure gas pipeline leakage further includes:

[0036] A database is constructed; the database includes background schlieren measurement results data obtained in calibration experiments for different leak diameters and pressures of the experimental high-pressure gas pipeline; the background schlieren measurement results data includes the density distribution of the test field area corresponding to the gas type leakage of the experimental high-pressure gas pipeline; the experimental high-pressure gas pipeline and the test high-pressure gas pipeline are the same pipeline;

[0037] If the monitoring result indicates a leak, the density distribution of the test area of ​​the high-pressure gas pipeline to be tested is compared with the corresponding background schlieren measurement result data in the database to obtain the similarity; the density distribution of the test area of ​​the high-pressure gas pipeline to be tested is determined based on the density image of the test area.

[0038] An alarm signal is issued when the similarity exceeds a set threshold.

[0039] Secondly, this application provides an online monitoring system for high-pressure gas pipeline leaks, including: an imaging device and a computer;

[0040] The imaging device is positioned above the high-pressure gas pipeline under test; the imaging device is connected to the computer, and is used to capture real-time images of the high-pressure gas pipeline under test to obtain monitoring images and send the monitoring images to the computer; the computer includes: an image acquisition module, a displacement vector diagram determination module, a test field area determination module, a refractive index determination module at the boundary of the test field, a density image determination module, and an online monitoring module;

[0041] Image acquisition module, used to acquire the monitoring image;

[0042] The displacement vector diagram determination module is used to determine the displacement vector diagram based on the monitoring image and a reference image; the reference image is an image of the high-pressure gas pipeline under test when there is no leakage.

[0043] The test area determination module is used to obtain the test area based on the monitoring image; the test area is a square area containing a jet cone; the jet cone is a cone-shaped area generated by the gas leak when there is a gas leak in the high-pressure gas pipeline under test;

[0044] The module for determining the refractive index at the boundary of the test field is used to determine the refractive index at the boundary of the test field region.

[0045] A density image determination module is used to determine the density image of the region of the field to be measured based on the displacement vector diagram and the refractive index at the boundary of the field to be measured.

[0046] The online monitoring module is used to perform online monitoring of the density image of the field under test based on a convolutional neural network to obtain monitoring results; the monitoring results include whether there is a leak in the high-pressure gas pipeline under test.

[0047] According to the specific embodiments provided in this application, this application has the following technical effects:

[0048] This application provides a method and system for online monitoring of high-pressure gas pipeline leaks. The method involves determining a displacement vector map based on a monitoring image and a reference image; obtaining the target area based on the monitoring image and determining the refractive index at the target area boundary; determining the density image of the target area based on the displacement vector map and the refractive index at the target area boundary; and performing online monitoring of the density image of the target area using a convolutional neural network to obtain the monitoring results. This application eliminates the need for lens groups and complex optical paths, obtaining the density image of the target area through the above scheme. This achieves fewer devices, lower environmental requirements, and a larger field of view. By determining the density image of the target area and using the real-time and efficient recognition of the convolutional neural network, online monitoring of high-pressure gas leaks can be achieved, improving monitoring sensitivity. This enables low-cost, high-sensitivity online leak monitoring of high-pressure gas pipelines in large special electrical equipment. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a schematic flowchart of an online monitoring method for high-pressure gas pipeline leakage according to an embodiment of this application;

[0051] Figure 2 This is a schematic diagram of the optical path for a natural background texture provided in an embodiment of this application;

[0052] Figure 3 This is a simulation diagram of the overall layout of an online monitoring system for high-pressure gas pipeline leakage provided in an embodiment of this application;

[0053] Figure 4 This is a schematic diagram of an online monitoring system for high-pressure gas pipeline leakage provided in one embodiment of this application;

[0054] Figure 5 This is a schematic diagram of an image formed by an imaging device provided in one embodiment of this application;

[0055] Figure 6This is a schematic diagram of the imaging device arrangement in one embodiment of this application;

[0056] Figure 7 This is a schematic diagram of the functional modules of an online monitoring system for high-pressure gas pipeline leakage in one embodiment of this application;

[0057] Figure 8 A schematic diagram illustrating the installation and monitoring steps of a high-pressure gas pipeline monitoring system for a steam turbine provided in an embodiment of this application;

[0058] Figure 9 This is a partial leakage diagram of the leakage displacement vector diagram provided in one embodiment of this application;

[0059] Figure 10 for Figure 9 The corresponding diagram shows the structural division of the leakage field image.

[0060] Reference numerals: 1-Imaging equipment, 2-Light source, 3-High-pressure gas pipeline to be measured, 4-Computer, 5-Alarm device. Detailed Implementation

[0061] The technical solutions of the embodiments 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, and 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.

[0062] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0063] Among relevant gas leak detection methods, the soap film method involves applying soapy water to the surface of the pipe to be tested and observing whether bubbles appear to determine if a leak has occurred. Although it has the lowest cost, it is labor-intensive and cannot be applied to the detection of high-pressure gas leaks. The fixed-point catalytic combustion method involves placing a detection device at a fixed point to continuously catalytically combust the gas in the air to determine whether there is a leak of flammable gas. It has a low cost but cannot quickly locate the leak point and requires the leak of the flammable gas to reach a certain scale before it can be detected. The infrared thermal imaging method uses the infrared absorption characteristics of the gas for detection. It is more sensitive but has a high cost and is difficult to apply to situations where high-pressure air or water vapor leaks into relatively open air domains. The schlieren method uses the density gradient between the leaking gas and the surrounding air environment for detection. It can provide high-resolution real-time imaging of the leak field, but the imaging window is limited by the size of the optical devices and the optical path is complex and precise, making it difficult to promote on a large scale in industrial production practice. Therefore, in order to eliminate the safety hazards caused by high-pressure gas leakage during the operation of large special electrical equipment, this application provides a method and system for online monitoring of high-pressure gas pipeline leakage, which can realize low-cost and high-sensitivity online leakage monitoring of high-pressure gas pipelines in large special electrical equipment.

[0064] In one exemplary embodiment, such as Figure 1 As shown, an online monitoring method for high-pressure gas pipeline leakage is provided, including the following S1 to S6. Wherein:

[0065] S1: Acquire monitoring images; the monitoring images are images obtained by real-time imaging of the high-pressure gas pipeline under test; the pressure inside the high-pressure gas pipeline under test is above 1 MPa. As an optional implementation, the monitoring images are obtained by real-time imaging of the high-pressure gas pipeline under test using the natural background schlieren method, with a background of a non-test high-pressure gas pipeline within the high-pressure gas pipeline group located at the camera's focus position; the high-pressure gas pipeline group includes multiple high-pressure gas pipelines. During monitoring, the entire high-pressure gas pipeline group is divided into two parts according to spatial distribution. Each part has an imaging device on one side. When the first part is the high-pressure gas pipeline under test, the first imaging device is used for imaging, and the second part serves as the background; when the second part is the high-pressure gas pipeline under test, the second imaging device is used for imaging, and the first part serves as the background; thus, the entire pipeline group is monitored through the complementary use of the two imaging devices.

[0066] S2: Determine a displacement vector diagram based on the monitored image and the reference image; the reference image is an image of the high-pressure gas pipeline under test when there is no leakage. As an optional implementation, the reference image is an image of the high-pressure gas pipeline under test when there is no leakage, obtained using the natural background schlieren method.

[0067] S3: The test area is obtained based on the monitoring image; the test area is a square area containing a jet cone; the jet cone is a cone-shaped area generated by the high-pressure gas leak when there is a gas leak in the high-pressure gas pipeline under test.

[0068] S4: Determine the refractive index at the boundary of the test field region.

[0069] S5: Determine the density image of the region to be measured based on the displacement vector diagram and the refractive index at the boundary of the field to be measured.

[0070] S6: Online monitoring of the density image of the test area is performed based on a convolutional neural network to obtain monitoring results; the monitoring results include whether there is a leak in the test high-pressure gas pipeline.

[0071] In another exemplary embodiment of this application, step S2 specifically includes:

[0072] S21: Image enhancement is performed on the monitoring images based on the superposition principle. Specifically, S211, the monitoring images are grouped chronologically. Depending on the gas pressure inside the high-pressure gas pipeline under test, each group can contain 10-100 images. The higher the gas pressure in the pipeline, the smaller the group size should be. This part involves superimposing the images within each group to achieve noise reduction and resolution enhancement. S212, after superimposing all the images contained in each group, the average value is taken to obtain a super-resolution image.

[0073] S22: Perform cross-correlation calculations on the enhanced monitoring image (super-resolution image) and the reference image to obtain a displacement vector map. 。

[0074] In another exemplary embodiment of this application, the refractive index at the boundary of the field to be measured is obtained based on the atmospheric refractive index and the gas pressure in the pipeline. S4 specifically includes:

[0075] S41: Substitute the gas pressure and gas temperature in the high-pressure gas pipeline to be tested into the ideal gas law to calculate the gas density at the boundary of the test field.

[0076] S42: Substitute the gas density at the boundary of the field to be measured into the Glaston-Dale equation to calculate the refractive index at the boundary of the field to be measured.

[0077] In another exemplary embodiment of this application, S22 specifically includes:

[0078] S221: The enhanced monitoring image is calibrated to obtain a calibrated monitoring image. Calibration refers to adding coordinate scales to the image using objects of specific known dimensions in the image (such as the diameter of the pipe to be measured) to facilitate subsequent calculations.

[0079] S222: After calibrating the obtained super-resolution image, cross-correlation calculation is performed between it and a previously captured reference image. Specifically, a query window is used to perform cross-correlation calculation between the calibrated monitoring image and the reference image to obtain a displacement vector map. The size of the query window is selected based on the different gas pressures within the high-pressure gas pipeline being measured. The gas pressure within the power plant's transmission pipeline is generally known and constant. If a high-pressure pipeline is being monitored, a smaller query window is selected; the higher the gas pressure in the pipeline, the larger the selected query window should be. Finally, the displacement vector map is calculated. The displacement vector map represents the position vector relationship of the same ray on the monitoring image and the reference image. When light passes through a non-uniform density field, it will be deflected. The position of the deflected light reaching the camera's imaging plane (i.e., the monitoring image) will deviate from its position when the non-uniform density field does not exist (i.e., the reference image). The displacement vector refers to the position vector relationship of the same ray on these two images.

[0080] In another exemplary embodiment of this application, the refractive index at the location of the pipeline leak is difficult to measure or estimate. Therefore, its density is calculated based on the type, pressure, and temperature of the gas flowing inside the pipeline, and then substituted into the Glastondale formula to calculate the refractive index at the corresponding pipeline leak location. The refractive index of the remaining boundary regions can be considered as the refractive index of the ambient gas when using a larger query window. After inputting the refractive index at the boundary of the test field based on the pressure of the high-pressure pipeline, the density image of the test field region is obtained through inversion calculation. S5 specifically includes:

[0081] S51: Determine displacement vector data based on the displacement vector diagram; the displacement vector data includes: the displacement in the X and Y directions of the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image; the background is the camera focusing position. Specifically, the camera focusing position is the non-tested high-pressure gas pipeline in the high-pressure gas pipeline group.

[0082] S52: Substitute the refractive index at the boundary of the field to be measured and the displacement vector data into the Poisson equation and iteratively solve to obtain the refractive index of the field to be measured.

[0083] S53: Substitute the refractive index of the field to be measured into the Glaston-Dale equation to calculate the density image of the region of the field to be measured.

[0084] In another exemplary embodiment of this application, the formula for calculating the gas density at the boundary of the field to be measured is:

[0085]

[0086] Where ρ is the gas density at the boundary of the test field, P is the gas pressure inside the high-pressure gas pipeline, M is the molar mass obtained based on the type of gas inside the high-pressure gas pipeline, and R... g Let T be the ideal gas constant, and T be the gas temperature inside the high-pressure gas pipeline to be measured.

[0087] In another exemplary embodiment of this application, the formula for calculating the refractive index at the boundary of the field to be measured is:

[0088] n0 = kρ + 1.

[0089] Where n0 is the refractive index at the boundary of the field to be measured, k is the Glaston-Dale coefficient, and ρ is the gas density at the boundary of the field to be measured.

[0090] In another exemplary embodiment of this application, the formula for calculating the refractive index of the field to be measured is:

[0091]

[0092] Where n is the refractive index of the test field, n0 is the refractive index at the boundary of the gas under normal temperature and pressure, W is the thickness of the test field, which is usually neglected in actual calculations, and Z... d Let x be the distance between the area to be tested and the background, x be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the X direction, y be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, x be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, y ... x Δ is the displacement in the X direction of the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image. y The displacement in the Y direction is the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image.

[0093] In another exemplary embodiment of this application, after S6, the online monitoring method for high-pressure gas pipeline leakage further includes:

[0094] S61: Construct a database; the database includes background schlieren measurement results data obtained in calibration experiments for different leakage diameters and pressures of the experimental high-pressure gas pipeline; the background schlieren measurement results data includes the density distribution of the test field area corresponding to the gas type leakage of the experimental high-pressure gas pipeline; the experimental high-pressure gas pipeline and the test high-pressure gas pipeline are the same pipeline; specifically, the experimental high-pressure gas pipeline and the test high-pressure gas pipeline have the same gas type, pipeline diameter, and gas pressure.

[0095] S62: If the monitoring result indicates a leak, the density distribution of the test area of ​​the high-pressure gas pipeline to be tested is compared with the background schlieren measurement result data in the database to obtain the similarity; the density distribution of the test area of ​​the high-pressure gas pipeline to be tested is determined based on the density image of the test area.

[0096] S63: When the similarity exceeds a set threshold, an alarm signal is issued. The threshold is set according to the leakage hazard level, and the leakage diameter and leakage pressure are output based on the leakage image. The reference image, monitoring image, and leakage image are all images of the same pipeline under different conditions obtained by the same imaging equipment: the reference image is the image obtained during the equipment commissioning phase when no leakage is confirmed; the monitoring image is the image obtained during equipment operation. Under normal circumstances, since there is no leakage, the light is not deflected due to the cone-shaped area caused by the high-pressure gas leakage, that is, its displacement vector field is zero displacement throughout; the leakage image is obtained during the monitoring process when a high-pressure gas leakage occurs, and a cone-shaped area appears in the displacement vector field, after which inversion calculation and neural network judgment are performed.

[0097] In another exemplary embodiment of this application, S6 specifically includes:

[0098] Density images (field cloud maps) of experimental high-pressure gas pipeline leaks with known leakage pressures and corresponding gas types at different diameters are input into a convolutional neural network for learning. The convolutional neural network continuously performs image recognition and comparison on the density images of the test area obtained in S5 to obtain monitoring results.

[0099] like Figure 2 As shown, this application uses a high-pressure gas pipeline as the background pattern and employs the natural background schlieren method to obtain monitoring and reference images. Based on schlieren technology, a background schlieren technique has been developed, specifically using a planar light source and random scattered points as the background to detect the density field of the leaking gas and the surrounding air environment. The natural background schlieren method in this application is a novel optical measurement technique based on the background schlieren method. While sharing the same principle as the schlieren method—both imaging and detection are based on the density gradient between the measured field and the surrounding environment—it differs from the background schlieren method in that it utilizes a natural background with a certain contrast instead of the traditional background composed of a planar light source and random scattered points. This application uses the high-pressure gas pipeline as the natural background, eliminating the need for the fabrication and arrangement of a luminous background plate, thus overcoming the limitations of the luminous background plate's fabrication and arrangement on the field of view and further reducing monitoring costs.

[0100] Based on the near-optical axis assumption: the deflection angle of the light ray relative to the principal optical axis is extremely small, that is, the magnitude of the light ray deflection angle is approximately equal to the magnitude of its own tangent. It is also assumed that the thickness W of the test field is negligible relative to its distance from the background. Finally, the relationship between the light ray deflection angle ε and the displacement Δ of the light ray after passing through the non-uniform density test field is obtained:

[0101]

[0102] In the formula, ε x ε is the deflection angle of the light ray in the X direction after passing through the area to be measured. y The angle of deflection in the Y direction is the angle of light rays after passing through the area to be measured.

[0103] Z D Z represents the distance from the area to be measured to the background. A Z is the distance from the lens group in the camera (imaging device) to the high-pressure gas pipeline group being measured. B Z is the distance from the lens group in the camera (imaging device) to the background. B The value is Z D Value and Z A The sum of values, Z i Δ is the distance from the lens group to the photosensitive element in a camera (imaging device). x Δ is the displacement in the X direction (not shown in the figure) of the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image. y Δ is the displacement in the Y direction of the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image; y ′ represents the displacement of the light ray in the Y direction relative to its initial position on the background after passing through the test area. The relationship between the displacement Δ of the light ray after passing through the non-uniform density test area can be obtained directly by calibrating the (two-dimensional) displacement vector diagram.

[0104] This application uses the natural background schlieren method to obtain displacement vector data, eliminating the need for lens groups and complex optical paths, thereby achieving less equipment, lower environmental requirements, and a larger field of view. By performing continuous cross-correlation calculations between the monitored image and a set reference image, and using convolutional neural networks for real-time and efficient identification, online monitoring of high-pressure gas leaks can be realized.

[0105] Based on the same inventive concept, this application also provides a high-pressure gas pipeline leakage online monitoring system for implementing the above-mentioned high-pressure gas pipeline leakage online monitoring method. The solution provided by this system is similar to the implementation scheme described in the above method; therefore, the specific limitations of one or more high-pressure gas pipeline leakage online monitoring system embodiments provided below can be found in the limitations of the high-pressure gas pipeline leakage online monitoring method described above, and will not be repeated here.

[0106] In one exemplary embodiment, a high-pressure gas pipeline leak online monitoring system is provided, including hardware devices: an imaging device and a computer; and software components of the computer: an image enhancement algorithm, a cross-correlation algorithm, and a convolutional neural network algorithm. The specific solution is as follows:

[0107] like Figure 3 and Figure 4 As shown, the online monitoring system for high-pressure gas pipeline leaks includes: an imaging device 1 and a computer 4; the imaging device 1 is positioned above the high-pressure gas pipeline 3 to be monitored, ensuring that the pipeline in the image is as free from overlap or obstruction as possible; the image captured by the imaging device 1 should be as shown... Figure 5 As shown, the imaging device 1 is connected to the computer 4. The imaging device 1 is used to capture real-time images of the high-pressure gas pipeline 3 under test and send the images to the computer 4. The imaging device 1 is preferably an industrial CCD camera or an industrial camera with equivalent spatial and temporal resolution.

[0108] like Figure 6 As shown, in practical applications, the natural background schlieren method uses an imaging device to focus on a background pattern of high-pressure gas pipelines. The area to be measured is located between the imaging device and the background pattern. Multiple high-pressure gas pipelines form a high-pressure gas pipeline group. Imaging devices are placed on both sides of the high-pressure gas pipeline group. The entire high-pressure gas pipeline group is spatially divided into two parts, left and right. Taking the first imaging device as an example, the left half of the pipeline is the first high-pressure gas pipeline group to be measured, and the right half is the first background high-pressure gas pipeline group, used as the background pattern for the first imaging device. The first imaging device focuses on the right half of the first background high-pressure gas pipeline group and monitors the left half of the first high-pressure gas pipeline group to be measured. Conversely, the second imaging device operates in the same way. Figure 6 The right half of the pipeline serves as the second high-pressure gas pipeline group to be tested for the second imaging device, while the left half of the pipeline serves as the second background high-pressure gas pipeline group, used as the background pattern for the second imaging device. The two imaging devices work simultaneously to achieve real-time monitoring of the entire high-pressure gas pipeline group to be tested.

[0109] like Figure 7As shown, the computer 4 includes: an image acquisition module, a displacement vector diagram determination module, a test field region determination module, a refractive index determination module at the boundary of the test field, a density image determination module, and an online monitoring module.

[0110] An image acquisition module is used to acquire the monitoring image.

[0111] The displacement vector diagram determination module is used to determine the displacement vector diagram based on the monitoring image and the reference image; the reference image is an image of the high-pressure gas pipeline under test when there is no leakage.

[0112] The test area determination module is used to obtain the test area based on the monitoring image; the test area is a square area containing a jet cone; the jet cone is a cone-shaped area generated by the high-pressure gas leak when there is a gas leak in the high-pressure gas pipeline under test.

[0113] The module for determining the refractive index at the boundary of the test field is used to determine the refractive index at the boundary of the test field region.

[0114] The density image determination module is used to determine the density image of the region to be measured based on the displacement vector diagram and the refractive index at the boundary of the field to be measured.

[0115] The online monitoring module is used to perform online monitoring of the density image of the field under test based on a convolutional neural network to obtain monitoring results; the monitoring results include whether there is a leak in the high-pressure gas pipeline under test.

[0116] In another exemplary embodiment of this application, the online monitoring system for high-pressure gas pipeline leakage further includes an alarm device and a light source; the alarm device is connected to the computer and is used to receive alarm signals sent by the computer and to issue an alarm; the light source is positioned above the high-pressure gas pipeline under test, so that the high-pressure gas pipeline under test in the field of view can be uniformly illuminated by the light source, and the light source can be an industrial LED light source and the industrial LED can be kept on.

[0117] During system operation, the computer receives monitoring images from the imaging device in real time and performs image enhancement based on the superposition principle. The enhanced super-resolution image is then cross-correlated with a reference image to obtain a displacement vector map. By inputting the refractive index at the boundary of the test field into the displacement vector map, the density distribution of the test field can be obtained through inversion calculation. Finally, the density image of the obtained test field region is monitored online based on a convolutional neural network. When the similarity exceeds 90%, a signal is sent to the alarm device to trigger an alarm. This application enables low-cost and high-sensitivity online monitoring of high-pressure pipeline leaks.

[0118] Natural gas power plants involve the transportation of various high-pressure gases, including blast furnace gas, high-pressure oxygen, high-pressure air, high-pressure steam, high-pressure nitrogen, and high-pressure argon, during actual production. Therefore, the implementation method of this application will be explained below using a power plant as an example. The system installation and monitoring steps are as follows: Figure 8 As shown.

[0119] The specific layout of a single device will be illustrated using the online monitoring of the high-temperature and high-pressure gas pipeline of the steam turbine, a typical high-temperature and high-pressure gas pipeline transportation link in the power plant production process, as an example. First, the monitoring system will be assembled according to the system simulation structure, and the specific setup requirements are as follows:

[0120] The high-pressure gas pipeline 3 to be tested is the high-pressure gas pipeline of the steam turbine. The imaging device 1 is a high-speed camera, which is installed at a designated position above the steam turbine pipeline. The imaging device should ensure that the pipeline in the high-pressure gas pipeline to be tested has as little overlap or obstruction as possible in the captured image.

[0121] Light source 2 is an industrial LED, installed near the installed imaging equipment, and its position is continuously adjusted until the high-pressure gas pipeline in the field of view is evenly illuminated by this light source. Afterwards, the industrial LED is kept on, and once it is confirmed that there are no leaks in any of the high-pressure gas pipelines in the steam turbine, images are captured using imaging equipment 1 and transmitted to computer 4 as reference images. At this point, the system preparation is complete. The system can be deployed above multiple high-pressure gas pipelines, and the results are aggregated at the control console to achieve online monitoring of high-pressure gas pipeline leaks throughout the plant's production process.

[0122] During normal system operation, computer 4 groups the monitoring images transmitted by the installed imaging equipment 1 according to time sequence. Since the pressure in the steam turbine pipeline is usually 15 MPa or above, it is divided into groups of 10 images. After superimposing all the images contained in each group, the average value is taken to obtain the super-resolution image.

[0123] After calibrating the obtained super-resolution image, cross-correlation calculations were performed between it and the previously captured reference image. Since the pressure in the steam turbine pipeline is usually 15 MPa or higher, the query window size was set to 64x64 mm. Finally, the displacement vector diagram was calculated.

[0124] By substituting the steam pressure and temperature in the high-pressure gas pipeline of the steam turbine into the ideal gas law, the gas density at the boundary of the test field can be calculated:

[0125]

[0126] Where ρ is the gas density at the boundary of the field to be measured, M is the molar mass of water vapor, and R... g Let T be the ideal gas constant and T be the vapor temperature.

[0127] After obtaining the gas density at the boundary of the test field, we substitute it into the Glaston-Dale equation to calculate the refractive index at the boundary of the test field:

[0128] n0 = kρ + 1.

[0129] Where n0 is the refractive index at the boundary of the field to be measured, and k is the Glaston-Dale coefficient corresponding to water vapor.

[0130] Substituting the calculated refractive index at the boundary of the field to be measured into the Poisson equation, the resulting Poisson equation is as follows:

[0131]

[0132] Where n0 is the refractive index at the boundary of the field to be measured, i.e., the water vapor refractive index at the boundary of the field to be measured.

[0133] The refractive index of the field to be measured is obtained by iteratively solving the Poisson equation, and then the density image of the field to be measured can be obtained by the Glaston-Dale equation.

[0134] A convolutional neural network, having learned from density images (density field cloud maps) of high-pressure gas pipelines of different diameters with known leakage pressures and gas types, continuously performs image recognition and comparison on the density images of the test area obtained in the previous step. Because the risk factor of leakage in the high-pressure gas pipeline of the steam turbine is too high, a threshold of 75% is set. Once the threshold is exceeded, computer 4 outputs a signal to alarm device 5 to trigger an alarm. Figure 9 As shown, the green area at the leak location is composed of individual green arrows, which are displacement vectors obtained through cross-correlation calculations. The left conical boundary of the high-pressure gas pipeline leak in the steam turbine is the leak boundary. Figure 9 The circular structure surrounded by the green area is the Mach ring structure, such as... Figure 10 As shown, the units of the horizontal and vertical axes are pixels, the leakage field is the jet cone angle, the leakage diameter is 8mm, and the leakage pressure is 1MPa.

[0135] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0136] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for online monitoring of high-pressure gas pipeline leaks, characterized in that, The online monitoring method for high-pressure gas pipeline leakage includes: Acquire monitoring images; the monitoring images are images obtained in real time from the high-pressure gas pipeline under test; A displacement vector map is determined based on the monitoring image and the reference image; the reference image is an image of the high-pressure gas pipeline under test when there is no leakage; the monitoring image is enhanced based on the superposition principle; the enhanced monitoring image and the reference image are cross-correlated to obtain the displacement vector map, specifically including: calibrating the enhanced monitoring image to obtain a calibrated monitoring image; using a query window to perform cross-correlation calculation between the calibrated monitoring image and the reference image to obtain the displacement vector map; the size of the query window is selected according to the different gas pressures in the high-pressure gas pipeline under test; the displacement vector map is used to represent the position vector relationship of the same ray on the monitoring image and the reference image; The test area is obtained based on the monitoring image; the test area is a square area containing a jet cone; the jet cone is a cone-shaped area generated by the gas leak when there is a gas leak in the high-pressure gas pipeline under test; Determine the refractive index at the boundary of the test field region; The density image of the test field region is determined based on the displacement vector diagram and the refractive index at the boundary of the test field; displacement vector data is determined based on the displacement vector diagram; the displacement vector data includes: the displacement in the X and Y directions of the difference between the position of the light ray emitted from the same position in the background after passing through the test field region in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image; the background is the camera focusing position; the refractive index at the boundary of the test field and the displacement vector data are substituted into the Poisson equation and iteratively solved to obtain the refractive index of the test field; the refractive index of the test field is substituted into the Glaston-Dale equation to calculate the density image of the test field; The density image of the test area is monitored online using a convolutional neural network to obtain monitoring results; the monitoring results include whether there is a leak in the high-pressure gas pipeline under test.

2. The online monitoring method for high-pressure gas pipeline leakage according to claim 1, characterized in that, Determining the refractive index at the boundary of the test field region specifically includes: Substituting the gas pressure and gas temperature inside the high-pressure gas pipeline into the ideal gas law, the gas density at the boundary of the test field is calculated. The gas density at the boundary of the test field is substituted into the Glaston-Dale equation to calculate the refractive index at the boundary of the test field.

3. The online monitoring method for high-pressure gas pipeline leakage according to claim 2, characterized in that, The formula for calculating the gas density at the boundary of the field to be measured is: Where ρ is the gas density at the boundary of the test field, P is the gas pressure inside the high-pressure gas pipeline, M is the molar mass obtained based on the type of gas inside the high-pressure gas pipeline, and R... g Let T be the ideal gas constant, and T be the gas temperature inside the high-pressure gas pipeline to be measured.

4. The online monitoring method for high-pressure gas pipeline leakage according to claim 2, characterized in that, The formula for calculating the refractive index at the boundary of the field to be measured is: n0 = kρ + 1; Where n0 is the refractive index at the boundary of the field to be measured, k is the Glaston-Dale coefficient, and ρ is the gas density at the boundary of the field to be measured.

5. The online monitoring method for high-pressure gas pipeline leakage according to claim 1, characterized in that, The formula for calculating the refractive index of the field to be measured is: Where n is the refractive index of the field to be measured, n0 is the refractive index at the boundary of the gas to be measured at room temperature and pressure, W is the thickness of the field to be measured, and Z is the refractive index of the gas to be measured at room temperature and pressure. d Let x be the distance between the area to be tested and the background, x be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the X direction, y be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, x be the coordinate of the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, y ... distance between the initial position of the ray emitted from the same position in the background in the reference image in the Y direction, and Δ be the distance between the area to be tested and the background, y be the distance between the initial x Δ is the displacement in the X direction of the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image. y The displacement in the Y direction is the difference between the position of the light ray emitted from the same position in the background after passing through the test area in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image.

6. The online monitoring method for high-pressure gas pipeline leakage according to claim 1, characterized in that, After obtaining the monitoring results by online monitoring of the density image of the area to be tested based on a convolutional neural network, the online monitoring method for high-pressure gas pipeline leakage further includes: A database is constructed; the database includes background schlieren measurement results data obtained in calibration experiments for different leak diameters and pressures of the experimental high-pressure gas pipeline; the background schlieren measurement results data includes the density distribution of the test field area corresponding to the gas type leakage of the experimental high-pressure gas pipeline; the experimental high-pressure gas pipeline and the test high-pressure gas pipeline are the same pipeline; If the monitoring result indicates a leak, the density distribution of the test area of ​​the high-pressure gas pipeline to be tested is compared with the corresponding background schlieren measurement result data in the database to obtain the similarity; the density distribution of the test area of ​​the high-pressure gas pipeline to be tested is determined based on the density image of the test area. An alarm signal is issued when the similarity exceeds a set threshold.

7. An online monitoring system for high-pressure gas pipeline leaks, characterized in that, The online monitoring system for high-pressure gas pipeline leaks includes: imaging equipment and a computer; The imaging device is positioned above the high-pressure gas pipeline under test; the imaging device is connected to the computer, and is used to capture real-time images of the high-pressure gas pipeline under test to obtain monitoring images and send the monitoring images to the computer; the computer includes: an image acquisition module, a displacement vector diagram determination module, a test field area determination module, a refractive index determination module at the boundary of the test field, a density image determination module, and an online monitoring module; Image acquisition module, used to acquire the monitoring image; The displacement vector diagram determination module is used to determine a displacement vector diagram based on the monitoring image and a reference image; the reference image is an image of the high-pressure gas pipeline under test when there is no leakage; the monitoring image is enhanced based on the superposition principle; the enhanced monitoring image and the reference image are cross-correlated to obtain the displacement vector diagram, specifically including: calibrating the enhanced monitoring image to obtain a calibrated monitoring image; using a query window to perform cross-correlation calculation between the calibrated monitoring image and the reference image to obtain the displacement vector diagram; the size of the query window is selected according to the different gas pressures in the high-pressure gas pipeline under test; the displacement vector diagram is used to represent the position vector relationship of the same ray on the monitoring image and the reference image; The test area determination module is used to obtain the test area based on the monitoring image; the test area is a square area containing a jet cone, and the jet cone is a cone-shaped area generated by the gas leak when there is a gas leak in the high-pressure gas pipeline under test; The module for determining the refractive index at the boundary of the test field is used to determine the refractive index at the boundary of the test field region. A density image determination module is used to determine the density image of the test field region based on the displacement vector diagram and the refractive index at the boundary of the test field; determine displacement vector data based on the displacement vector diagram; the displacement vector data includes: the displacement in the X and Y directions of the difference between the position of the light ray emitted from the same position in the background after passing through the test field region in the monitoring image and the initial position of the light ray emitted from the same position in the background in the reference image; the background is the camera focusing position; the refractive index at the boundary of the test field and the displacement vector data are substituted into the Poisson equation and iteratively solved to obtain the refractive index of the test field; the refractive index of the test field is substituted into the Glaston-Dale equation to calculate the density image of the test field; The online monitoring module is used to perform online monitoring of the density image of the field under test based on a convolutional neural network to obtain monitoring results; the monitoring results include whether there is a leak in the high-pressure gas pipeline under test.

Images