Detection and early warning method and system for pressure-bearing equipment

Through deep learning models, the problem that the existing technology cannot effectively monitor micro cracks, and low-cost and high-precision crack monitoring and early warning are achieved, and the risk of safety accidents is reduced.

CN118941490APending Publication Date: 2024-11-12CHINA PETROLEUM & CHEMICAL CORP +1
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202310539770.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-12
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing technology cannot effectively monitor and early warning of tiny cracks in refining equipment online, and cannot monitor the development status of cracks in real time, resulting in strong potential damage concealment and increasing the risk of safety accidents.

Method used

Deep learning model is used to identify surface cracks of the pressure-bearing equipment, collect images in real time, measure crack size, and dynamically diagnose changes according to preset thresholds, real-time monitoring and early warning of micro cracks are achieved.

Benefits of technology

It realizes low-cost and high-precision rapid inspection of surface cracks of pressure-bearing equipment, which can effectively reflect the expansion of cracks within a certain period of time, reduces the inspection costs of enterprises, and improves safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941490A_ABST
    Figure CN118941490A_ABST
Patent Text Reader

Abstract

The invention discloses a detection and early warning method for pressure-bearing equipment. The method comprises the following steps: acquiring an image containing surface characteristics of the pressure-bearing equipment to be detected in real time; identifying and marking cracks in the surface image by using a preset identification model, and then measuring crack sizes representing crack development characteristics according to the real-time crack image; according to the crack size, whether the dynamic change of the crack reaches a preset threshold value or not is diagnosed, and when the dynamic change of the crack reaches or exceeds the preset threshold value, early warning prompting is carried out. According to the invention, low-cost and high-precision rapid inspection of the surface cracks of the pressure-bearing equipment can be realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of nondestructive testing of pressure-bearing equipment, and in particular to a detection and early warning method and system for pressure-bearing equipment. Background Art

[0002] Affected by factors such as the chloride ion environment and alkaline environment in petrochemical equipment, production equipment is prone to pitting and stress corrosion cracking damage, and as the material grade of petrochemical equipment continues to improve, the damage is becoming more and more hidden. Traditional detection methods are costly or highly dependent on manual labor, which is not conducive to large-scale application by enterprises. At the same time, the knowledge reserve of enterprise professionals has declined, and leaks and safety accidents caused by subjective misjudgment have continued to increase, and there is a lack of auxiliary means for detection.

[0003] In 1983, Westinghouse Electric Company of the United States first proposed a combination of bright, dark and weak lighting in the detection system and used high-intensity linear light sources and linear array CCD cameras to detect surface defects of steel plates. The In-Sight2800 visual system developed by Cognex combines deep learning technology with traditional rule-based visual tools, adopts multi-color lighting (RGBW) integrated light source, optimizes the image formation process, and maximizes contrast and illumination. The IV3 series image recognition sensor developed by Keyence integrates white light and red light lighting systems. The intensity, light emitting method, exposure time, etc. of the lighting are automatically set by AI, and stable detection is achieved by deriving suitable conditions for various environments and workpiece states. Domestically, there are LabelRoll and PackRoll series visual inspection systems developed by Lingyun Company, as well as 41 major series, nearly 1,000 standardized light source products and more than 30,000 light source customization solutions developed by Opt.

[0004] Although the above-mentioned prior art can detect the damage of production equipment through the participation of auxiliary means, the prior art still does not have a targeted design for online monitoring of tiny cracks in refining equipment, and cannot provide early warning of the development status of cracks.

[0005] Therefore, it is necessary to provide a solution that can monitor the crack development state of production equipment in real time when micro cracks appear. Summary of the invention

[0006] The purpose of the present invention is to provide a solution for monitoring the crack development state when tiny cracks appear in production equipment.

[0007] In order to solve the above technical problems, an embodiment of the present invention provides a detection and early warning method for pressure-bearing equipment, including: step one, real-time acquisition of images containing surface features of the pressure-bearing equipment to be tested; step two, using a preset recognition model, identifying and marking cracks in the surface image, and then measuring the crack size characterizing the crack development characteristics based on the real-time crack image; step three, based on the crack size, diagnosing whether the dynamic change of the crack reaches a preset threshold, wherein an early warning prompt is issued when the preset threshold is reached or exceeded.

[0008] Preferably, the step of measuring the crack size characterizing the crack development characteristics based on the real-time crack image includes: obtaining a crack image marked with crack contour characteristics; calculating the area of ​​each crack by regarding the same crack contour as the same connected domain, and / or calculating the maximum inscribed circle in each crack region, thereby taking the maximum value of the diameters of the maximum inscribed circle in each crack region as the crack width of the current crack image.

[0009] Preferably, in the step of diagnosing whether the dynamic change of the crack reaches a preset threshold based on the crack size, it includes: calculating the crack change characteristics based on the crack widths recorded three times in a row, and judging whether the crack change characteristics reach or exceed the preset first width threshold or second width threshold, so as to issue an early warning when it is reached or exceeded.

[0010] Preferably, an early warning is issued when one of the following conditions is met: n ≥δw n-1 And w n+1 ≥δw n , where w n-1 、w n and w n+1 represents the crack widths of the same crack measured for the n-1th, nth and n+1th times respectively, δ represents the first width threshold; or, w n ≥ηw n-1 And w n+1 ≥ηw n-1 , where η represents the second width threshold.

[0011] Preferably, in the step of diagnosing whether the dynamic change of the crack reaches a preset threshold based on the crack size, it includes: calculating the ratio between the total crack area of ​​the same crack image and the area of ​​the corresponding surface image; judging whether the current crack area ratio data reaches or exceeds the preset area threshold, so as to issue an early warning when it reaches or exceeds.

[0012] Preferably, the detection and early warning method further comprises: repeating the step 2 multiple times, so as to perform a diagnostic operation using the average value of the crack size calculated in each round.

[0013] Preferably, in the process of repeatedly performing step 2, the interval between each round is 1 minute, and the number of repetitions is not less than 3 times.

[0014] Preferably, the step one further comprises: the difference between the section angles of the camera and the surface of the pressure-bearing equipment to be tested does not exceed 30°, wherein the collection time interval of the real-time crack image is 1 hour.

[0015] Preferably, the recognition model is constructed using the following method: collecting multiple test surface images containing different positions, different crack shapes and different environmental conditions of the pressure-bearing equipment, and marking the outline of each crack in each of the test surface images, wherein the number of the test surface images is no less than 1,000, and the resolution of each test surface image is no less than 200 dpi; building an image segmentation neural network, and performing multiple iterative training on the built image segmentation neural network based on multiple test surface images with unmarked and marked crack outlines, so as to obtain a crack recognition model.

[0016] Preferably, the detection and early warning method also includes: issuing early warning prompts for different risk levels, wherein the first width threshold is constructed to determine the first width sub-thresholds of different levels according to the different materials of the pressure-bearing equipment to be tested, wherein, when the material of the pressure-bearing equipment to be tested is 16MnDR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.03, 1.06 and 1.09 respectively; when the material of the pressure-bearing equipment to be tested is Q345R, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.05, 1.1 and 1.15 respectively; when the material of the pressure-bearing equipment to be tested is 15CrMoR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.04, 1.08 and 1.12 respectively.

[0017] Preferably, the detection and early warning method also includes: issuing early warning prompts for different risk levels, wherein the second width threshold is constructed to determine different levels of second width sub-thresholds according to the different materials of the pressure-bearing equipment to be tested, wherein, when the material of the pressure-bearing equipment to be tested is 16MnDR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.02, 1.04 and 1.06 respectively; when the material of the pressure-bearing equipment to be tested is Q345R, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.04, 1.08 and 1.12 respectively; when the material of the pressure-bearing equipment to be tested is 15CrMoR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.03, 1.06 and 1.09 respectively.

[0018] Preferably, the detection and early warning method also includes: issuing early warning prompts for different risk levels, wherein the area threshold is constructed to determine different levels of area sub-thresholds according to the different materials of the pressure-bearing equipment to be tested, wherein, when the material of the pressure-bearing equipment to be tested is 16MnDR, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.06, 1.12 and 1.18 respectively; when the material of the pressure-bearing equipment to be tested is Q345R, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.09, 1.18 and 1.27 respectively; when the material of the pressure-bearing equipment to be tested is 15CrMoR, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.07, 1.14 and 1.21 respectively.

[0019] On the other hand, the present invention also provides a detection and early warning system for pressure-bearing equipment, which is used to implement the above-mentioned detection and early warning method, including: a camera, which is used to collect images containing surface features of the pressure-bearing equipment to be tested in real time; a processor, which is connected to the camera, and is used to use a preset recognition model to identify and mark cracks in the surface image, and then measure the crack size that characterizes the crack development characteristics based on the real-time crack image, and diagnose whether the dynamic change of the crack reaches a preset threshold based on the crack size, wherein an early warning prompt is given when the preset threshold is reached or exceeded.

[0020] Compared with the prior art, one or more embodiments of the above scheme may have the following advantages or beneficial effects:

[0021] The present invention proposes a detection and early warning method and system for pressure-bearing equipment. The system and method use a deep learning model to identify cracks, and have better resistance to interference from environmental factors; the crack width is determined by obtaining the diameter of the largest inscribed circle in the crack area; in addition to early warning of the crack area through a preset threshold, the crack width is also compared in continuous detection results to better reflect the expansion of cracks on the surface of the equipment within a certain period of time. The present invention can be applied to the energy and chemical industries to achieve low-cost, high-precision rapid inspection of surface cracks on pressure-bearing equipment, which is beneficial to cost reduction and efficiency improvement for various enterprises, so as to solve the negative impact of defects such as cracks that are not considered when they expand too quickly.

[0022] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0024] Figure 1 A schematic diagram of the steps of a detection and early warning method for pressure-bearing equipment according to an embodiment of the present application.

[0025] Figure 2 This is an example diagram of a crack image in a detection and early warning method for pressure-bearing equipment according to an embodiment of the present application.

[0026] Figure 3 This is an example diagram of the crack size calculation principle in the detection and early warning method for pressure-bearing equipment in an embodiment of the present application.

[0027] Figure 4 This is a schematic diagram of the structure of a detection and early warning system for pressure-bearing equipment according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] The following will describe the implementation methods of the present invention in detail with reference to the accompanying drawings and embodiments, so that the implementation process of how the present invention applies technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly. It should be noted that as long as there is no conflict, the various embodiments of the present invention and the various features in the embodiments can be combined with each other, and the technical solutions formed are all within the protection scope of the present invention.

[0029] In addition, the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions. Also, although a logical sequence is shown in the flowchart, in some cases, the steps shown or described can be performed in a sequence different from that here.

[0030] The terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments. Unless the context clearly indicates otherwise, the singular forms "one", "one" and "item" used herein are also intended to include plural numbers. It should also be understood that the terms "include" and / or "comprise" used herein specify the existence of stated features, integers, steps, operations, units and / or components, without excluding the existence or addition of one or more other features, integers, steps, operations, units, components and / or combinations thereof.

[0031] Affected by factors such as the chloride ion environment and alkaline environment in petrochemical equipment, production equipment is prone to pitting and stress corrosion cracking damage, and as the material grade of petrochemical equipment continues to improve, the damage is becoming more and more hidden. Traditional detection methods are costly or highly dependent on manual labor, which is not conducive to large-scale application by enterprises. At the same time, the knowledge reserve of enterprise professionals has declined, and leaks and safety accidents caused by subjective misjudgment have continued to increase, and there is a lack of auxiliary means for detection.

[0032] In 1983, Westinghouse Electric Company of the United States first proposed a combination of bright, dark and weak lighting in the detection system and used high-intensity linear light sources and linear array CCD cameras to detect surface defects of steel plates. The In-Sight2800 visual system developed by Cognex combines deep learning technology with traditional rule-based visual tools, adopts multi-color lighting (RGBW) integrated light source, optimizes the image formation process, and maximizes contrast and illumination. The IV3 series image recognition sensor developed by Keyence integrates white light and red light lighting systems. The intensity, light emitting method, exposure time, etc. of the lighting are automatically set by AI, and stable detection is achieved by deriving suitable conditions for various environments and workpiece states. Domestically, there are LabelRoll and PackRoll series visual inspection systems developed by Lingyun Company, as well as 41 major series, nearly 1,000 standardized light source products and more than 30,000 light source customization solutions developed by Opt.

[0033] Although the above-mentioned prior art can detect the damage of production equipment through the participation of auxiliary means, the prior art still does not have a targeted design for online monitoring of tiny cracks in refining equipment, and cannot provide early warning of the development status of cracks.

[0034] Therefore, it is necessary to provide a solution that can monitor the crack development state of production equipment in real time when micro cracks appear.

[0035] In order to solve the technical problems in the above-mentioned background technology, the embodiment of the present application proposes a detection and early warning method and system for pressure-bearing equipment. The system and method identify cracks through a deep learning model, and have better resistance to interference from environmental factors; the crack width is determined by obtaining the diameter of the largest inscribed circle in the crack area; in addition to early warning of the crack area through a preset threshold, it also performs online monitoring and early warning of small cracks by comparing the degree of change in the crack width in the continuous detection results, which can better reflect the expansion of cracks on the surface of the equipment within a certain period of time, thereby realizing monitoring and operation under the damaged state of production equipment.

[0036] Embodiment 1

[0037] Figure 1This is a schematic diagram of the steps of the detection and early warning method for pressure-bearing equipment in the embodiment of the present application. Figure 1 , the detection and early warning method described in the embodiment of the present invention is explained.

[0038] Step S110 collects images containing surface features of the pressure-bearing equipment to be tested in real time, wherein the difference between the angle of the collecting camera and the section angle of the surface of the pressure-bearing equipment to be tested does not exceed 30°.

[0039] In step S110, the real-time crack image is collected at an interval of 1 hour. That is, step S110 collects an image containing the surface features of the pressure-bearing equipment to be tested at a preset collection time interval (eg, 1 hour) to obtain a corresponding surface image.

[0040] refer to Figure 1 In step S120, a preset recognition model is used to identify and mark cracks in the surface image collected in real time in step S110, and then the crack size representing the crack development characteristics is measured according to the real-time crack image.

[0041] In step S120 , first, a preset recognition model is used to recognize and mark cracks in the surface image acquired in real time in step S110 .

[0042] The recognition model is an intelligent model used for online crack recognition of surface images collected in real time. The model can mark the contour of each crack contained in each input surface image with several color blocks and output the marked crack image, such as Figure 2 shown.

[0043] In the embodiment of the present invention, firstly, a plurality of test surface images containing different crack distribution positions, different crack shapes and different environmental conditions of the pressure-bearing equipment are collected, and the outline of each crack is marked in each test surface image. The collected test surface images are not less than 1,000, and the resolution of each test surface image is not less than 200 dpi.

[0044] Then, an image segmentation neural network is built, and the built image segmentation neural network is iteratively trained multiple times based on multiple test surface images without crack contours and multiple test surface images with crack contours, so as to obtain a crack recognition model after completing the model training.

[0045] It should be noted that the output data enhancement operation needs to be performed during model training. The trained crack recognition model is an instance segmentation model based on deep learning, and can mark each crack with a different color block.

[0046] In this way, the crack recognition model obtained by the present invention can accurately identify cracks of different shapes and distribution positions under the influence of different environmental factors, and accurately mark the outline of each crack.

[0047] Therefore, step S120 will first use the above crack recognition model to identify and mark the crack contour in each surface image, so as to obtain a real-time crack image.

[0048] Next, after marking the cracks in each surface image, step S120 also measures the crack size representing the crack development characteristics according to each crack image obtained above. In the embodiment of the present invention, the crack size includes the crack area and / or the crack width.

[0049] Specifically, before measuring the crack size, each crack image marked with crack contour features is first obtained, and each crack image is preprocessed. The preprocessing includes, but is not limited to, several morphological corrosion and expansion operations. When measuring the crack size, in the same crack image, by considering the same crack contour as the same connected domain, the area of ​​each crack is calculated, and / or the maximum inscribed circle in each crack region is calculated, so that the maximum value of the diameter of the maximum inscribed circle in each crack region is used as the crack width of the current crack image.

[0050] In one embodiment, step S120 regards the marks of different colors in a single crack image as different connected domains, and records the total area of ​​all connected domains in the current crack image to be detected (the sum of the areas of each crack), which is recorded as the crack area s. ′ .

[0051] In another embodiment, step S120 sequentially calculates the maximum inscribed circle of each connected domain of the marked crack contour in the current crack image to be tested, obtains a series of inscribed circles and their corresponding inscribed circle diameter data from the crack image to be tested, and regards the maximum value of the diameter data as the crack width w of the current crack image. 1 ,like Figure 3 shown.

[0052] In addition, the detection and early warning method described in the embodiment of the present invention also includes: repeating step S120 multiple times to use the average value of the crack size (crack area and / or crack width) calculated in each round to perform the diagnostic operation of step S130.

[0053] In the process of repeatedly executing step S120, the interval between each round is 1 minute, and the number of repetitions is not less than 3 times. In the process of obtaining crack size data, step S120 is repeated at least 3 times, each time with an interval of 1 minute, and the obtained crack size is taken as the average value of the 3 operations and then used as the final crack size, thereby entering step S130.

[0054] Continue to refer Figure 1 Step S130 will also diagnose whether the dynamic change of the crack size reaches a preset threshold value based on the crack size of the image obtained in step S120. If the preset threshold value is reached or exceeded, an early warning prompt will be issued.

[0055] In the first embodiment, the crack variation characteristics are calculated based on the crack widths of three consecutively recorded images, and it is determined whether the crack variation characteristics reach or exceed a preset first width threshold or a second width threshold, so as to issue an early warning when the crack width reaches or exceeds the preset first width threshold.

[0056] Otherwise, if the crack variation characteristic does not reach the preset first width threshold or the second width threshold does not meet the aforementioned conditions, no warning is required.

[0057] Specifically, an early warning is issued when one of the following conditions is met: n ≥δw n-1 And w n+1 ≥δw n , or w n ≥ηw n-1 And w n+1 ≥ηw n-1 Among them, w n-1 、w n and w n+1 They represent the crack widths of the surface images to be measured measured for the n-1th, nth and n+1th times respectively, δ represents the first width threshold, and η represents the second width threshold.

[0058] In the second embodiment, the ratio between the total crack area of ​​the same crack image and the area of ​​the corresponding surface image is calculated to obtain crack area percentage data, and then it is determined whether the current crack area percentage data reaches or exceeds a preset area threshold, so as to issue an early warning when it is reached or exceeded.

[0059] Specifically, when τ≥φ and If the condition is met, an early warning is issued. Otherwise, if τ≥φ is not met and If the conditions are met, no warning is required. Wherein, τ represents the crack area ratio data, s represents the area of ​​the surface image, and φ represents the preset area threshold.

[0060] Furthermore, the embodiment of the present invention can also provide early warning prompts for different risk levels in step S130.

[0061] In the first embodiment, the first width threshold is constructed to determine first width sub-thresholds of different levels according to different materials of the pressure-bearing equipment to be tested.

[0062] In a second embodiment, the second width threshold is constructed to determine second width sub-thresholds of different levels according to different materials of the pressure-bearing equipment to be tested.

[0063] In a third embodiment, the area threshold is constructed to determine area sub-thresholds of different levels according to different materials of the pressure-bearing equipment to be tested.

[0064] Specifically, δ, η and Depending on the material, a series of different values ​​are set, corresponding to different levels of risk warning. Taking δ as an example, you can set δ 1 ,δ 2 and δ 3 , and correspond to medium risk, high risk and high risk warning respectively. Further, according to different materials, δ 1 ,δ 2 and δ 3 Can take a series of different values.

[0065] Embodiment 2

[0066] Based on the above-mentioned first embodiment, the process involved in the detection and early warning method described in the embodiment of the present invention is described below by taking the above-mentioned first width threshold as the basis for early warning diagnosis.

[0067] S1: Obtain surface image data of the device to be inspected, such as Figure 2 In step S1, during the process of acquiring the surface image of the device to be inspected, the angle between the camera and the device surface or the section of the device surface does not exceed 30°.

[0068] In S1, first, a plurality of test surface images containing different crack distribution positions, different crack shapes and different environmental conditions of the pressure-bearing equipment are collected, and the outline of each crack is marked in each test surface image. The collected test surface images are no less than 1,000, and the resolution of each test surface image is no less than 200 dpi.

[0069] S2: Use the crack recognition model to identify the image data obtained in S1. If the image to be inspected is found to contain cracks (see Figure 2 1 part in the image), different cracks are marked with different colors. The area of ​​the image to be inspected is s. Figure 3 As shown, the independent crack color blocks (see Figure 3 The two parts in the image are the color blocks of one of the cracks) are regarded as the connected domains of different cracks, and their areas are recorded as s'. The crack recognition model in step S2 is an instance segmentation model based on deep learning.

[0070] Specifically, an image segmentation neural network is constructed, and the constructed image segmentation neural network is iteratively trained multiple times based on a plurality of test surface images without crack contours and a plurality of test surface images with crack contours, so as to obtain a crack recognition model after completing the model training.

[0071] It should be noted that the output data enhancement operation needs to be performed during model training. The trained crack recognition model is an instance segmentation model based on deep learning, and can mark each crack with a different color block.

[0072] S3: Preprocess the image obtained in S2. The preprocessing in step S3 is a morphological erosion and dilation operation of the image.

[0073] S4: Find the maximum inscribed circle for each connected domain in turn (see Figure 3 The maximum value of a series of inscribed circle diameters is regarded as the crack width of the crack image, denoted as w 1 .

[0074] Execute steps S2-S4 and repeat at least 3 times to obtain the crack width respectively. The interval between each operation is 1 minute. The crack width data of the crack image is the average value of the values ​​obtained by the 3 operations.

[0075] S5: Every 1 hour, collect surface images at the same location in the same way and repeat steps S1-S4 to obtain the crack width w 2 , w 3 ...when w n-1 ≠0.

[0076] If n ≥δw n-1 And w n+1 ≥δw n , then an early warning is issued.

[0077] Table 1 shows the first width threshold δ corresponding to different materials. The first width threshold is constructed as the first width sub-threshold representing different risk levels, which are set to δ 1 ,δ 2 ,δ 3 .

[0078] Table 1 Values ​​of δ corresponding to different materials

[0079] Medium risk Higher risk High risk Material <![CDATA[δ 1 ]]> <![CDATA[δ 2 ]]> <![CDATA[δ 3 ]]> 16MnDR 1.03 1.06 1.09 Q345R 1.05 1.10 1.15 15CrMoR 1.04 1.08 1.12

[0080] As shown in Table 1, when the material of the pressure-bearing equipment to be tested is 16MnDR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.03, 1.06 and 1.09 respectively.

[0081] As shown in Table 1, when the material of the pressure-bearing equipment to be tested is Q345R, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.05, 1.1 and 1.15 respectively.

[0082] As shown in Table 1, when the material of the pressure-bearing equipment to be tested is 15CrMoR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.04, 1.08 and 1.12 respectively.

[0083] Embodiment 3

[0084] Based on the above-mentioned first embodiment, the process involved in the detection and early warning method described in the embodiment of the present invention is described below by taking the above-mentioned second width threshold as the basis for early warning diagnosis.

[0085] S1: Obtain surface image data of the device to be inspected, such as Figure 2 In step S1, during the process of acquiring the surface image of the device to be inspected, the angle between the camera and the device surface or the section of the device surface does not exceed 30°.

[0086] In S1, first, a plurality of test surface images containing different crack distribution positions, different crack shapes and different environmental conditions of the pressure-bearing equipment are collected, and the outline of each crack is marked in each test surface image. The collected test surface images are no less than 1,000, and the resolution of each test surface image is no less than 200 dpi.

[0087] S2: Use the crack recognition model to identify the image data obtained in S1. If the image to be inspected is found to contain cracks (see Figure 2 1 part in the image), different cracks are marked with different colors. The area of ​​the image to be inspected is s. Figure 3 As shown, the independent crack color blocks (see Figure 3 The two parts in the image are the color blocks of one of the cracks) are regarded as the connected domains of different cracks, and their areas are recorded as s'. The crack recognition model in step S2 is an instance segmentation model based on deep learning.

[0088] Specifically, an image segmentation neural network is constructed, and the constructed image segmentation neural network is iteratively trained multiple times based on a plurality of test surface images without crack contours and a plurality of test surface images with crack contours, so as to obtain a crack recognition model after completing the model training.

[0089] It should be noted that the output data enhancement operation needs to be performed during model training. The trained crack recognition model is an instance segmentation model based on deep learning, and can mark each crack with a different color block.

[0090] S3: Preprocess the image obtained in S2. The preprocessing in step S3 is a morphological erosion and dilation operation of the image.

[0091] S4: Find the maximum inscribed circle for each connected domain in turn (see Figure 3 The maximum value of a series of inscribed circle diameters is regarded as the crack width of the crack image, denoted as w 1 .

[0092] Execute steps S2-S4 and repeat at least 3 times to obtain the crack width respectively. The interval between each operation is 1 minute. The crack width data of the crack image is the average value of the values ​​obtained by the 3 operations.

[0093] S5: Every 1 hour, collect surface images at the same location in the same way and repeat steps S1-S4 to obtain the width w 2 , w 3 ...when w n-1 ≠0.

[0094] If n ≥ηw n-1 、w n+1 ≥ηw n-1 And w n+2 ≥ηw n-1 , then an early warning is issued.

[0095] Table 2 shows the second width threshold η corresponding to different materials. The second width threshold is constructed as a second width sub-threshold representing different risk levels, which is set to η 1 , η 2 , η 3 .

[0096] Table 2 Values ​​of η corresponding to different materials

[0097] Medium risk Higher risk High risk Material <![CDATA[η 1 ]]> <![CDATA[η 2 ]]> <![CDATA[η 3 ]]> 16MnDR 1.02 1.04 1.06 Q345R 1.04 1.08 1.12 15CrMoR 1.03 1.06 1.09

[0098] As shown in Table 2, when the material of the pressure-bearing equipment to be tested is 16MnDR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.02, 1.04 and 1.06 respectively.

[0099] As shown in Table 2, when the material of the pressure-bearing equipment to be tested is Q345R, the first width sub-thresholds corresponding to medium risk, relatively high risk, and high risk are 1.04, 1.08, and 1.12, respectively;

[0100] As shown in Table 2, when the material of the pressure-bearing equipment to be tested is 15CrMoR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.03, 1.06 and 1.09, respectively.

[0101] Embodiment 4

[0102] Based on the above-mentioned embodiment 1, the process involved in the detection and early warning method described in the embodiment of the present invention is described below using the above-mentioned area threshold as a basis for early warning diagnosis.

[0103] S1: Obtain surface image data of the device to be inspected, such as Figure 2 In step S1, during the process of acquiring the surface image of the device to be inspected, the angle between the camera and the device surface or the section of the device surface does not exceed 30°.

[0104] In S1, first, a plurality of test surface images containing different crack distribution positions, different crack shapes and different environmental conditions of the pressure-bearing equipment are collected, and the outline of each crack is marked in each test surface image. The collected test surface images are no less than 1,000, and the resolution of each test surface image is no less than 200 dpi.

[0105] S2: Use the crack recognition model to identify the image data obtained in S1. If the image to be inspected is found to contain cracks (see Figure 2 1 part in the image), different cracks are marked with different colors. The area of ​​the image to be inspected is s. Figure 3 As shown, the independent crack color blocks (see Figure 3 The two parts in the figure are the color blocks of one of the cracks) are regarded as the connected domains of different cracks, and their areas are recorded as s ′ The crack recognition model in step S2 is an instance segmentation model based on deep learning.

[0106] Specifically, an image segmentation neural network is constructed, and the constructed image segmentation neural network is iteratively trained multiple times based on a plurality of test surface images without crack contours and a plurality of test surface images with crack contours, so as to obtain a crack recognition model after completing the model training.

[0107] It should be noted that the output data enhancement operation needs to be performed during model training. The trained crack recognition model is an instance segmentation model based on deep learning, and can mark each crack with a different color block.

[0108] S3: Preprocess the image obtained in S2. The preprocessing in step S3 is a morphological erosion and dilation operation of the image.

[0109] S4: Calculate the crack area ratio data τ.

[0110] Among them, when obtaining the crack area in steps S2-S4, the relevant steps are performed 3 times with an interval of 1 minute, and the crack area percentage data is obtained by taking the average value of the data obtained by the 3 operations.

[0111] S5: If Give early warning.

[0112] Table 3 shows the area thresholds corresponding to different materials Among them, the area threshold is constructed to represent the area sub-thresholds of different risk levels, which are set as

[0113] Table 3 Correspondence of different materials The value of

[0114]

[0115] As shown in Table 3, when the material of the pressure-bearing equipment to be tested is 16MnDR, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.06, 1.12 and 1.18, respectively.

[0116] As shown in Table 3, when the material of the pressure-bearing equipment to be tested is Q345R, the area sub-thresholds corresponding to medium risk, relatively high risk, and high risk are 1.09, 1.18, and 1.27, respectively.

[0117] As shown in Table 3, when the material of the pressure-bearing equipment to be tested is 15CrMoR, the area sub-thresholds corresponding to medium risk, relatively high risk, and high risk are 1.07, 1.14, and 1.21, respectively.

[0118] Embodiment 5

[0119] Based on the detection and early warning methods described in the first to fourth embodiments above, the present invention further provides a detection and early warning system for pressure-bearing equipment. The detection and early warning system is used to implement the detection and early warning methods described in the first to fourth embodiments above.

[0120] Figure 4 Schematic diagram of the structure of the detection and early warning system for pressure-bearing equipment in the embodiment of the present application. Figure 4 As shown, the detection and warning system described in the embodiment of the present invention includes: a camera 41 and a processor 42.

[0121] The camera 41 is implemented according to the method described in the above step S110 to collect images containing surface features of the pressure-bearing equipment to be tested in real time.

[0122] The processor 42 is connected to the camera 41 and implemented according to the method described in step S110, for identifying and marking cracks in the surface image using a preset recognition model, and then measuring the crack size representing the crack development characteristics according to the real-time crack image, and diagnosing whether the dynamic change of the crack reaches a preset threshold according to the crack size. In which, an early warning prompt is issued when the preset threshold is reached or exceeded.

[0123] The present invention discloses a detection and early warning method and system for pressure-bearing equipment. The system and method identify cracks through a deep learning model, and have better resistance to interference from environmental factors; the crack width is determined by obtaining the diameter of the largest inscribed circle in the crack area; in addition to early warning of the crack area through a preset threshold, the crack width is also compared in continuous detection results to better reflect the expansion of cracks on the surface of the equipment within a certain period of time. The present invention can be applied to the energy and chemical industries to achieve low-cost, high-precision rapid inspection of cracks on the surface of pressure-bearing equipment, which is beneficial to cost reduction and efficiency improvement of various enterprises, so as to solve the negative impact of defects such as cracks that are not considered when they expand too quickly.

[0124] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any person familiar with the technology within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0125] In the description of the present invention, unless otherwise specified, "plurality" means two or more than two; the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail", etc. are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0126] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0127] It should be understood that the embodiments disclosed in the present invention are not limited to the specific structures, processing steps or materials disclosed herein, but should be extended to equivalent substitutions of these features understood by ordinary technicians in the relevant field. It should also be understood that the terms used herein are only used for the purpose of describing specific embodiments and are not meant to be limiting.

[0128] The "one embodiment" or "embodiment" mentioned in the specification means that the specific features, structures or characteristics described in conjunction with the embodiment are included in at least one embodiment of the present invention. Therefore, the phrases "one embodiment" or "embodiment" appearing in various places throughout the specification do not necessarily refer to the same embodiment.

[0129] The embodiments of the present invention are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

[0130] Although the embodiments disclosed in the present invention are as above, the contents described are only embodiments adopted for facilitating the understanding of the present invention and are not intended to limit the present invention. Any technician in the technical field to which the present invention belongs can make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in the present invention, but the patent protection scope of the present invention shall still be subject to the scope defined in the attached claims.

Claims

1. A detection and early warning method for pressure-bearing equipment, characterized in that: include: Step 1: real-time acquisition of images containing surface features of the pressure-bearing equipment to be tested; Step 2: using a preset recognition model to identify and mark cracks in the surface image, and then measuring the crack size that characterizes the crack development characteristics based on the real-time crack image; Step three: based on the crack size, diagnose whether the dynamic change of the crack reaches a preset threshold, wherein an early warning prompt is issued when the preset threshold is reached or exceeded.

2. The detection and early warning method according to claim 1, characterized in that: The step of measuring the crack size representing crack development characteristics according to the real-time crack image includes: Obtaining a crack image marked with crack contour features; By considering the same crack contour as the same connected domain, the area of ​​each crack is calculated, and / or the maximum inscribed circle in each crack region is calculated, so that the maximum value of the diameters of the maximum inscribed circle in each crack region is used as the crack width of the current crack image.

3. The detection and early warning method according to claim 2, characterized in that: The step of diagnosing whether the dynamic change of the crack reaches a preset threshold according to the crack size includes: The crack variation characteristics are calculated based on the crack widths recorded three times in succession, and it is determined whether the crack variation characteristics reach or exceed a preset first width threshold or a second width threshold, so as to issue an early warning when the threshold is reached or exceeded.

4. The detection and early warning method according to claim 3, characterized in that: An early warning is issued when one of the following conditions is met: w n ≥δw n-1 And w n+1 ≥δw n , where w n-1 、w n and w n+1 represents the crack widths of the same crack measured for the n-1th, nth and n+1th times respectively, and δ represents the first width threshold; or, w n ≥ηw n-1 And w n+1 ≥ηw n-1 , where η represents the second width threshold.

5. The detection and early warning method according to claim 2, characterized in that: The step of diagnosing whether the dynamic change of the crack reaches a preset threshold according to the crack size includes: Calculating the ratio between the total crack area of ​​the same crack image and the area of ​​the corresponding surface image; Determine whether the current crack area ratio data reaches or exceeds the preset area threshold, so as to issue an early warning when it reaches or exceeds the threshold.

6. The detection and early warning method according to any one of claims 1 to 5, characterized in that: The detection and early warning method also includes: The step 2 is repeated multiple times to perform a diagnostic operation using the average value of the crack size calculated in each round.

7. The detection and early warning method according to claim 6, characterized in that: In the process of repeatedly executing step 2, the interval between each round is 1 minute, and the number of repetitions is not less than 3 times.

8. The detection and early warning method according to any one of claims 1 to 7, characterized in that: The step one also includes: the difference between the section angles of the camera and the surface of the pressure-bearing equipment to be tested does not exceed 30°, wherein the time interval for collecting the real-time crack image is 1 hour.

9. The detection and early warning method according to any one of claims 1 to 8, characterized in that: The recognition model is constructed using the following method: Collect multiple test surface images containing different positions, different crack shapes and different environmental conditions of the pressure-bearing equipment, and mark the outline of each crack in each of the test surface images, wherein the test surface images are no less than 1,000, and the resolution of each test surface image is no less than 200 dpi; An image segmentation neural network is constructed, and multiple iterations of training are performed on the constructed image segmentation neural network according to multiple test surface images of unlabeled and labeled crack contours, so as to obtain a crack recognition model.

10. The detection and early warning method according to claim 3 or 4, characterized in that: The detection and warning method further includes: providing warning prompts for different risk levels, wherein the first width threshold is constructed to determine first width sub-thresholds of different levels according to different materials of the pressure-bearing equipment to be tested, wherein: When the material of the pressure-bearing equipment to be tested is 16MnDR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.03, 1.06 and 1.09 respectively; When the material of the pressure-bearing equipment to be tested is Q345R, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.05, 1.1 and 1.15 respectively; When the material of the pressure-bearing equipment to be tested is 15CrMoR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.04, 1.08 and 1.12 respectively.

11. The detection and early warning method according to claim 3 or 4, characterized in that: The detection and early warning method further includes: providing early warning prompts for different risk levels, wherein the second width threshold is constructed to determine second width sub-thresholds of different levels according to different materials of the pressure-bearing equipment to be tested, wherein: When the material of the pressure-bearing equipment to be tested is 16MnDR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.02, 1.04 and 1.06 respectively; When the material of the pressure-bearing equipment to be tested is Q345R, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.04, 1.08 and 1.12 respectively; When the material of the pressure-bearing equipment to be tested is 15CrMoR, the first width sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.03, 1.06 and 1.09 respectively.

12. The detection and early warning method according to claim 5, characterized in that: The detection and early warning method further includes: providing early warning prompts for different risk levels, wherein the area threshold is constructed to determine different levels of area sub-thresholds according to different materials of the pressure-bearing equipment to be tested, wherein: When the material of the pressure-bearing equipment to be tested is 16MnDR, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.06, 1.12 and 1.18 respectively; When the material of the pressure-bearing equipment to be tested is Q345R, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.09, 1.18 and 1.27 respectively; When the material of the pressure-bearing equipment to be tested is 15CrMoR, the area sub-thresholds corresponding to medium risk, relatively high risk and high risk are 1.07, 1.14 and 1.21 respectively.

13. A detection and early warning system for pressure-bearing equipment, characterized in that: The detection and early warning system is used to implement the detection and early warning method according to any one of claims 1 to 12, including: A camera, which is used to collect images containing surface features of the pressure-bearing equipment to be tested in real time; A processor is connected to the camera and is used to use a preset recognition model to identify and mark cracks in the surface image, and then measure the crack size that characterizes the crack development characteristics based on the real-time crack image, and diagnose whether the dynamic change of the crack reaches a preset threshold based on the crack size, wherein an early warning prompt is given when the preset threshold is reached or exceeded.

Citation Information

Patent Citations

  • Device and method for automatically detecting surface cracks of soft material

    CN107991307A

  • High-strength alloy structural steel plate surface fine crack identification and monitoring device

    CN111612771A

  • Non-contact crack measurement method and system

    CN113446932A