Method and system for detecting defects of petrochemical equipment

Abstract

The invention discloses a method and system for detecting petrochemical equipment defects, and the method comprises the steps: dividing to-be-evaluated petrochemical equipment into a plurality of detection regions, and obtaining defect point cloud data of each detection region through three-dimensional shape measurement; according to the defect point cloud data of each inspection area, utilizing a defect identification model to identify defects of each inspection area and obtain vertex features of each defect; according to the vertex features and the defect point cloud of each defect, defect evaluation features of each inspection area are calculated, and the defect evaluation features comprise a defect opening area coefficient, a defect depth coefficient and a defect metal loss coefficient; and according to the defect evaluation features of each inspection area, obtaining a defect risk coefficient for comprehensively representing the severity of the overall defect features of the petrochemical equipment to be evaluated. According to the invention, the defect detection precision and the automatic detection efficiency can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of nondestructive testing of petrochemical equipment, and in particular to a method and system for detecting defects of petrochemical equipment. Background Art

[0002] Petrochemical equipment operates in harsh conditions, and the internal medium is highly corrosive. After long-term operation, its inner and outer walls are prone to defects such as dents, cracks, and holes. If the defects are not discovered in time or handled improperly, it is easy to cause equipment explosions, casualties and other accidents. Therefore, it is necessary to conduct defect detection and safety assessment on related equipment regularly. Minor defects are highly hidden and difficult to detect. Commonly used non-destructive testing methods are difficult to accurately assess the size and depth information of minor defects. Therefore, a detection solution that can measure the depth of defects is urgently needed.

[0003] Three-dimensional shape measurement is a technology that measures and analyzes the three-dimensional shape of an object's surface, and can understand the details of the object's surface shape. The accuracy of three-dimensional measurement can reach the micron level, and while measuring the defect area, the depth information of the defect can be obtained. Compared with other detection methods, it has the advantages of high accuracy, good repeatability, and the ability to detect tiny defects, and is suitable for defect detection on the inner and outer walls of the equipment. However, after obtaining the point cloud data of the equipment through three-dimensional measurement, the reference surface must be manually selected to obtain the defect depth; and the defect opening must be manually selected to calculate the opening area. This process is not conducive to further improving the detection efficiency.

[0004] The prior art CN115541697B uses a magnetic flux leakage detection probe, which contacts the pipeline through a magnetizing mechanism to form a magnetic field inside the pipeline. The magnetic field inside the pipeline is then observed through the detection mechanism. If magnetic flux leakage occurs, it proves that the pipe wall at the magnetic flux leakage location is abnormal. The equipment is relatively light and can be held by the inspector, which improves the detection efficiency. However, the probe proposed in this patent is not suitable for the inspection of non-magnetic material equipment such as fiberglass in petrochemical scenes.

[0005] The prior art CN115264324B designs a mobile pipeline corrosion detection device for detecting corrosion defects in pipelines. However, the track used in this solution is not easy to move, which reduces the efficiency of pipeline detection. The track space on which the detection equipment relies occupies a large space, resulting in poor portability of the detection equipment.

[0006] In summary, the prior art needs to provide a technical solution that can quantitatively and accurately evaluate the three-dimensional size of tiny defects and automatically and quickly realize the defect detection function. Summary of the invention

[0007] The purpose of the present invention is to provide a solution to the technical problem that commonly used non-destructive testing methods are difficult to accurately and quantitatively assess the three-dimensional size information of tiny defects, especially the depth information of defects. It can also solve the technical problems that after the existing three-dimensional measurement obtains the point cloud data of the defective part, each defect needs to be manually distinguished, the efficiency of defect identification and defect three-dimensional size detection is low, and the automated defect detection function needs to be improved.

[0008] In order to solve the above technical problems, an embodiment of the present invention provides a method for detecting defects in petrochemical equipment, comprising: dividing the petrochemical equipment to be evaluated into multiple inspection areas, and using three-dimensional morphology measurement to obtain defect point cloud data of each inspection area; based on the defect point cloud data of each inspection area, using a defect recognition model, identifying the defects of each inspection area and obtaining the vertex features of each defect; based on the vertex features and defect point cloud of each defect, calculating the defect evaluation features of each inspection area, the defect evaluation features including a defect opening area coefficient, a defect depth coefficient and a defect metal loss coefficient; based on the defect evaluation features of each inspection area, obtaining a defect risk coefficient for comprehensively characterizing the severity of the overall defect features of the petrochemical equipment to be evaluated.

[0009] Preferably, the process of dividing the petrochemical equipment to be evaluated into multiple inspection areas includes: dividing the petrochemical equipment to be evaluated into multiple inspection areas according to the type of the petrochemical equipment to be evaluated and according to at least one factor of the function, structural composition and size of the equipment.

[0010] Preferably, the defect recognition model is constructed by the following method: using a three-dimensional scanning instrument to scan different inspection areas in various types of petrochemical equipment to obtain an original point cloud set; marking the minimum circumscribed cuboid of each defect in the original point cloud set; establishing a basic neural network model, and using the original point cloud set and the minimum circumscribed cuboid of each marked defect to train and test the basic neural network model to obtain the defect recognition model.

[0011] Preferably, point cloud data belonging to the same equipment do not appear in the model training set and the model test set at the same time; the petrochemical equipment to be evaluated is not among the petrochemical equipment in the original point cloud set.

[0012] Preferably, in the process of calculating the defect risk coefficient, the process includes: calculating the unit risk coefficient of each inspection area according to the defect evaluation characteristics of each inspection area, wherein the unit risk coefficient is calculated using the following expression:

[0013] f k =D×[ωS+(1-ω)M]

[0014] Among them, f k represents the unit risk coefficient, D represents the defect depth coefficient, S represents the defect opening area coefficient, M represents the defect metal loss coefficient, and ω represents the weight coefficient; the defect risk coefficient is obtained according to the unit risk coefficient of each inspection area.

[0015] Preferably, the method also includes: determining the defect severity level of the current petrochemical equipment to be evaluated based on the defect risk coefficient, and the defect severity level is divided into high, medium and low in order from high to low severity, wherein the high severity and the medium severity use the first coefficient threshold as the distinguishing boundary, and the medium severity and the low severity use the second coefficient threshold as the distinguishing boundary, and the first coefficient threshold is less than the second coefficient threshold.

[0016] Preferably, the defect opening area coefficient of each inspection area is calculated according to the rectangular vertex features and defect point cloud of each defect, which includes: projecting the defect point cloud contour of each identified defect onto three mutually perpendicular faces of the corresponding rectangular area to obtain the projection area of ​​each face, so as to take the maximum value of the three projection areas as the opening area of ​​the current defect; according to the opening area of ​​each defect, combined with the area of ​​the rectangular area where the corresponding maximum projection area is located, the opening area coefficient of the current defect is calculated, wherein the opening area coefficient is calculated using the following expression:

[0017] S=(ss loss ) / s

[0018] Among them, S represents the defect opening area coefficient, s represents the area of ​​the rectangular area where the maximum projection area corresponding to each defect is located, and s loss Indicates the opening area of ​​each defect; based on the opening area coefficient of each defect, determine the defect opening area coefficient of each inspection area.

[0019] Preferably, the defect depth coefficient of each inspection area is calculated according to the rectangular vertex features and defect point cloud of each defect, which includes: projecting the defect point cloud contour of each identified defect onto three mutually perpendicular faces of the corresponding rectangular area to obtain the projection area of ​​each face, so as to take the maximum value of the three projection areas as the available rectangular face of the current defect; the defect depth of each defect is obtained by calculating the distance between the available rectangular face of each defect and its corresponding parallel face, based on which, combined with the design wall thickness of the petrochemical equipment to be evaluated, the defect depth coefficient of each defect is obtained, wherein the defect depth coefficient is calculated using the following expression:

[0020] D=(dd loss ) / d

[0021] Where D represents the defect opening area coefficient, d represents the designed wall thickness, and d loss Indicates the defect depth of each defect; according to the defect depth coefficient of each defect, the defect depth coefficient of each inspection area is determined.

[0022] Preferably, the defect metal loss coefficient of each inspection area is calculated according to the rectangular vertex features and defect point cloud of each defect, which includes: extracting the corresponding local point cloud from the standard point cloud data for each identified rectangle vertex coordinate to record the number of points of the local point cloud of each defect, wherein the standard point cloud data is the point cloud data of the petrochemical equipment to be evaluated in a defect-free state; and obtaining the defect metal loss coefficient of each defect by calculating the absolute value of the difference between the number of points of the local point cloud of each defect and the number of points of the defect point cloud of the corresponding defect, wherein the defect metal loss coefficient is calculated using the following expression:

[0023] M=(mm max ) / m

[0024] Where M represents the defect metal loss coefficient, m represents the number of points in the local point cloud of each defect, and m max The number of points in the defect point cloud representing each defect; the defect metal loss coefficient of each inspection area is determined according to the defect metal loss coefficient of each defect.

[0025] On the other hand, an embodiment of the present invention provides a system for detecting defects in petrochemical equipment, the system being used to implement the method described above, wherein the system includes: a point cloud acquisition module, configured to divide the petrochemical equipment to be evaluated into multiple inspection areas, and use three-dimensional morphology measurement to obtain defect point cloud data of each inspection area; a defect recognition module, configured to identify the defects of each inspection area and obtain vertex features of each defect based on the defect point cloud data of each inspection area using a defect recognition model; a statistical analysis module, configured to calculate the defect evaluation features of each inspection area based on the vertex features and defect point cloud of each defect, the defect evaluation features including a defect opening area coefficient, a defect depth coefficient and a defect metal loss coefficient; a risk assessment module, configured to obtain a defect risk coefficient for comprehensively characterizing the severity of the overall defect features of the petrochemical equipment to be evaluated based on the defect evaluation features of each inspection area.

[0026] Preferably, the point cloud acquisition module is further configured to divide the petrochemical equipment to be evaluated into multiple inspection areas according to the type of the petrochemical equipment to be evaluated and at least one factor of the function, structural composition and size of the equipment.

[0027] Preferably, the defect recognition model is constructed according to the following process: use a three-dimensional scanning instrument to scan different inspection areas in various types of petrochemical equipment to obtain an original point cloud set; mark the minimum circumscribed cuboid of each defect in the original point cloud set; establish a basic neural network model, and use the original point cloud set and the minimum circumscribed cuboid of each marked defect to train and test the basic neural network model to obtain the defect recognition model.

[0028] Compared with the prior art, one or more embodiments of the above scheme may have the following advantages or beneficial effects:

[0029] The present invention proposes a method and system for detecting defects in petrochemical equipment. The method and system can be widely used in non-contact non-destructive defect detection and measurement of equipment in the petroleum and chemical fields, without any requirements on the material of the equipment being inspected, and can effectively improve the accuracy of defect detection and the efficiency of automated detection. Among them, the three-dimensional measurement technology is introduced into the detection of defects on the inner and outer walls of petrochemical equipment, and the beneficial effects produced are as follows:

[0030] 1. Improve the measurement accuracy of non-contact and non-destructive testing in the field of petrochemical equipment testing to sub-millimeter level; sub-millimeter dimensional measurement accuracy provides high-precision initial measurement values ​​for subsequent equipment remaining life prediction methods and calculation models based on measured defects such as crack depth and length, which is conducive to accurately predicting the remaining life of the equipment under test according to its actual situation;

[0031] 2. The deep learning-based 3D target detection method is used to realize the automatic identification of complex petrochemical equipment parts and their surface defects and the automatic positioning of defect positions, improve the efficiency of equipment defect detection, and realize the automatic identification of equipment defects on the point cloud data set: automatically identify and locate defects within the tested area, and automatically find the position and corresponding size information of the deepest defect or the defect with the largest opening area according to the defect depth or defect opening area;

[0032] 3. In order to accurately measure the three-dimensional morphology and size of defects, it is proposed to use the defect opening area, defect depth and metal loss measured automatically to quantitatively judge the severity of equipment defects without human intervention. By proposing a quantitative evaluation scheme for the severity of petrochemical equipment defects, it provides data support for the comprehensive management and preventive maintenance of petrochemical equipment.

[0033] 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

[0034] 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:

[0035] Figure 1 Schematic diagram of the steps of a method for detecting defects in petrochemical equipment according to an embodiment of the present application.

[0036] Figure 2 This is a schematic diagram of a specific process in a method for detecting defects in petrochemical equipment according to an embodiment of the present application.

[0037] Figure 3 A schematic diagram of defect marking in a method for detecting defects in petrochemical equipment according to an embodiment of the present application.

[0038] Figure 4 Schematic diagram of projection area calculation in a method for detecting defects in petrochemical equipment according to an embodiment of the present application.

[0039] Figure 5 This is a schematic diagram of the structure of a system for detecting defects in petrochemical equipment according to an embodiment of the present application. DETAILED DESCRIPTION

[0040] 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.

[0041] 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.

[0042] 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.

[0043] Petrochemical equipment operates in harsh conditions, and the internal medium is highly corrosive. After long-term operation, its inner and outer walls are prone to defects such as dents, cracks, and holes. If the defects are not discovered in time or handled improperly, it is easy to cause equipment explosions, casualties and other accidents. Therefore, it is necessary to conduct defect detection and safety assessment on related equipment regularly. Minor defects are highly hidden and difficult to detect. Commonly used non-destructive testing methods are difficult to accurately assess the size and depth information of minor defects. Therefore, a detection solution that can measure the depth of defects is urgently needed.

[0044] Three-dimensional shape measurement is a technology that measures and analyzes the three-dimensional shape of an object's surface, and can understand the details of the object's surface shape. The accuracy of three-dimensional measurement can reach the micron level, and while measuring the defect area, the depth information of the defect can be obtained. Compared with other detection methods, it has the advantages of high accuracy, good repeatability, and the ability to detect tiny defects, and is suitable for defect detection on the inner and outer walls of the equipment. However, after obtaining the point cloud data of the equipment through three-dimensional measurement, the reference surface must be manually selected to obtain the defect depth; and the defect opening must be manually selected to calculate the opening area. This process is not conducive to further improving the detection efficiency.

[0045] The prior art CN115541697B uses a magnetic flux leakage detection probe, which contacts the pipeline through a magnetizing mechanism to form a magnetic field inside the pipeline. The magnetic field inside the pipeline is then observed through the detection mechanism. If magnetic flux leakage occurs, it proves that the pipe wall at the magnetic flux leakage location is abnormal. The equipment is relatively light and can be held by the inspector, which improves the detection efficiency. However, the probe proposed in this patent is not suitable for the inspection of non-magnetic material equipment such as fiberglass in petrochemical scenes.

[0046] The prior art CN115264324B designs a mobile pipeline corrosion detection device for detecting corrosion defects in pipelines. However, the track used in this solution is not easy to move, which reduces the efficiency of pipeline detection. The track space on which the detection equipment relies occupies a large space, resulting in poor portability of the detection equipment.

[0047] In summary, the prior art needs to provide a technical solution that can quantitatively and accurately evaluate the three-dimensional size of tiny defects and automatically and quickly realize the defect detection function.

[0048] In order to solve the technical problems in the above-mentioned background technology, the embodiment of the present application proposes a method and system for detecting defects in petrochemical equipment. The system and method are widely used in the field of defect detection. Specifically, by processing the three-dimensional point cloud, the position of the defect is directly identified in the three-dimensional scene, and the defect depth and the amount of metal loss caused by the defect are calculated. The present invention can effectively improve the defect detection efficiency of petrochemical equipment and provide data support for the standardized management and preventive maintenance of equipment.

[0049] Embodiment 1

[0050] Figure 1 Schematic diagram of the steps of a method for detecting defects in petrochemical equipment according to an embodiment of the present application. Figure 2 This is a schematic diagram of a specific process in the method for detecting defects in petrochemical equipment in an embodiment of the present application. Figure 1 and Figure 2 , the specific steps and processes of the method for detecting defects in petrochemical equipment (hereinafter referred to as "defect detection method") described in an embodiment of the present invention are described.

[0051] In step S110, the petrochemical equipment to be evaluated is divided into a plurality of inspection areas, and three-dimensional shape measurement technology is used to obtain defect point cloud data (ie, original three-dimensional point cloud data containing defect features) of each inspection area.

[0052] In step S110, the petrochemical equipment to be evaluated is first divided into a plurality of inspection areas. Specifically, the petrochemical equipment to be evaluated is divided into a plurality of inspection areas according to the type of the petrochemical equipment to be evaluated and at least one of the function, structural composition and size of the current equipment.

[0053] Specifically, according to the type of the inspected equipment (i.e., the petrochemical equipment to be evaluated), the inspected equipment is divided into K inspection areas according to at least one factor such as function, structural composition, and size. In one embodiment, the division of the inspection area can be determined according to the organizational structure of the equipment. In another embodiment, each divided inspection area (e.g., the area divided according to the structural composition) can be divided again according to factors such as size.

[0054] Then, step S110 will also use three-dimensional morphology technology to scan each divided inspection area in each evaluated petrochemical equipment in turn, so as to obtain defect point cloud data of each inspection area.

[0055] In addition, in step S110, the equipment structure data in the equipment ledger is referenced to establish standard point cloud data of each inspection area of ​​the petrochemical equipment to be evaluated in a defect-free state.

[0056] Step S120 identifies the defects in each inspection area and obtains the vertex features of each defect using a pre-built defect recognition model based on the defect point cloud data of each inspection area obtained in step S110.

[0057] In an embodiment of the present invention, the defect recognition model is a computer vision recognition model based on deep learning. The defect recognition model is constructed by the following method flow:

[0058] S101, using a 3D scanner to scan different inspection areas in various petrochemical equipment to obtain an original point cloud set. According to the method for dividing the inspection area of ​​the equipment described in step S110, the inspection areas of various petrochemical equipment used for model construction are divided, and then, the defect point cloud data of each inspection area in each equipment is obtained using a 3D scanner.

[0059] It should be noted that the petrochemical equipment to be evaluated described in the embodiment of the present invention is not in the petrochemical equipment (used for model construction) in the original point cloud set.

[0060] S102, marking the minimum circumscribed cuboid of each defect in the original point cloud set. Specifically, for each defect in each inspection area in each device, the minimum circumscribed cuboid corresponding to the defect size is used to mark the defect, and the coordinates of the eight vertices of the minimum circumscribed cuboid of each defect are marked, see Figure 3 .

[0061] S103, establish a basic neural network model, use the original point cloud set and the smallest circumscribed cuboid of each marked defect to train and test the basic neural network model to obtain a defect recognition model. Preferably, the basic neural network model is a commonly used three-dimensional object detection model.

[0062] Specifically, the point cloud data set annotated with S2 is divided into a data training set and a data test set. The data training set is used to train the 3D object detection model. Then, the data test set is used to adjust the data, model structure and training strategy according to the model accuracy until the tested model meets the required accuracy.

[0063] It should be noted that, in the embodiment of the present invention, point cloud data belonging to the same device do not appear in the data training set and the data test set at the same time.

[0064] In this way, step S120 can use the defect recognition model constructed based on three-dimensional target detection to identify the defect point cloud data of each inspection area. The defect recognition model will use a rectangular block to identify the defective parts in the original three-dimensional point cloud data, and output the 8 vertex coordinates corresponding to the minimum circumscribed rectangular block of each defect, thereby entering step S130.

[0065] Step S130 calculates the defect evaluation features of each inspection area according to the vertex features (8 vertex coordinates of the cuboid) and the defect point cloud of each defect identified in step S120. The defect evaluation features include: defect opening area coefficient, defect depth coefficient and defect metal loss coefficient. It should be noted that in the embodiment of the present invention, the defect point cloud of each defect is the corresponding point cloud area extracted from the original point cloud data containing the defect features by extracting the rectangular vertex coordinates of each defect.

[0066] Specifically, step S130 calculates the defect opening area coefficient, defect depth coefficient and defect metal loss coefficient of each inspection area according to the rectangular vertex features and defect point cloud of each defect in the petrochemical equipment to be evaluated.

[0067] In the process of calculating the defect opening area coefficient of each inspection area, first, the defect point cloud contour of each identified defect is projected onto the three mutually perpendicular faces of the corresponding minimum circumscribed rectangular parallelepiped area, see Figure 4 , calculate the projection area of ​​each surface, and then take the maximum value of the three projection areas as the opening area of ​​the current defect, recorded as s loss .

[0068] Specifically, for each rectangular area output by the defect recognition model, the defect point cloud opening contour is projected onto three mutually perpendicular faces in the rectangular area, and the projection area is calculated, so that the maximum value of the projection area is taken as the opening area s of the current defect. loss , and the area of ​​the face of the cuboid where the maximum projection lies is recorded as s.

[0069] Then, according to the opening area s of each defect loss , combined with the area s of the rectangular area where the corresponding maximum projection area is located, calculate the opening area coefficient S of the current defect. Among them, the opening area coefficient is calculated using the following expression:

[0070] S=(ss loss ) / s

[0071] Among them, S represents the defect opening area coefficient, s represents the area of ​​the rectangular area where the maximum projection area corresponding to each defect is located, and s loss Indicates the opening area of ​​each defect.

[0072] Finally, the defect opening area coefficient of each inspection area is determined according to the opening area coefficient of each defect. In one embodiment, the maximum value of the opening area coefficients of the defects in the current inspection area is taken as the defect opening area coefficient of the current inspection area.

[0073] In the process of calculating the defect depth coefficient of each inspection area, the defect point cloud contour of each identified defect is projected onto the three mutually perpendicular faces of the corresponding rectangular area, and the projection area of ​​each face is calculated, see Figure 4 , so that the surface corresponding to the maximum value of the three projection areas is taken as the available rectangular surface of the current defect.

[0074] Specifically, for each rectangular area output by the defect recognition model, the opening contour of the defect point cloud is projected onto three mutually perpendicular faces in the rectangular area, the projection area is calculated, and then the face where the maximum projection face is located is taken as the available rectangular face.

[0075] Then, the defect depth of each defect is obtained by calculating the distance between the available rectangular surface of each defect and its corresponding parallel surface. loss , combined with the design wall thickness d of the petrochemical equipment to be evaluated, the defect depth coefficient of each defect is obtained. The defect depth coefficient of each defect is calculated using the following expression:

[0076] D=(dd loss ) / d

[0077] Where D represents the defect opening area coefficient, d represents the designed wall thickness, and d loss Indicates the defect depth of each defect.

[0078] Finally, the defect depth coefficient of each inspection area is determined according to the defect depth coefficient of each defect. In one embodiment, the maximum value of the defect depth coefficients of the defects in the current inspection area is taken as the defect depth coefficient of the current inspection area.

[0079] In the process of calculating the defect metal loss coefficient of each inspection area, first, the corresponding local point cloud is extracted from the standard point cloud data for each identified rectangle vertex coordinate, thereby recording the actual total number of points of the local point cloud of each defect. The standard point cloud data is a point cloud data set of the petrochemical equipment to be evaluated in a defect-free state.

[0080] Specifically, for each rectangular region output by the defect recognition model, a corresponding local point cloud is selected from the standard point cloud data obtained in step S110 according to its vertex coordinates, and the number of points of the local point cloud of each defect is recorded as m.

[0081] Then, the absolute value m of the difference between the number of local point cloud points of each defect and the number of defect point cloud points of the corresponding defect is calculated. loss , and obtain the defect metal loss coefficient of each defect. The defect metal loss coefficient of each defect is calculated using the following expression:

[0082] M=(mm max ) / m

[0083] Where M represents the defect metal loss coefficient, m represents the number of points in the local point cloud of each defect, and m max The number of points in the defect point cloud representing each defect.

[0084] Finally, the defective metal loss coefficient of each inspection area is determined according to the defective metal loss coefficient of each defect. In one embodiment, the maximum value of the defective metal loss coefficients of the defects in the current inspection area is taken as the defective metal loss coefficient of the current inspection area.

[0085] After calculating various defect evaluation features of each defect, the process proceeds to step S140.

[0086] Step S140 obtains a defect risk coefficient for comprehensively characterizing the severity of the overall defect characteristics of the petrochemical equipment to be evaluated based on the defect evaluation characteristics of each inspection area obtained in step S130.

[0087] In step S140, first, the unit risk coefficient of each inspection area is calculated according to the defect evaluation characteristics of each inspection area. The unit risk coefficient of each inspection area in the current device is calculated using the following expression:

[0088] f k =D×[ωS+(1-ω)M]

[0089] Among them, f k represents the unit risk coefficient of each inspection area, and ω represents the weight coefficient. In the embodiment of the present invention, 0≤ω≤1. In one embodiment, ω is preferably 0.5.

[0090] Then, according to the unit risk coefficient of each inspection area, a defect risk coefficient for comprehensively representing the overall defect status of the entire device to be evaluated is obtained.

[0091] In one embodiment, the minimum value of the unit risk coefficient of each inspection area in the current device is taken as the defect risk coefficient of the current device to be evaluated. The defect risk coefficient is calculated using the following expression:

[0092] f=min(f k )

[0093] Where f represents the defect risk factor.

[0094] In addition, in step S150, the embodiment of the present invention also determines the defect severity level of the petrochemical equipment to be evaluated according to the calculated defect risk coefficient. The defect severity level is divided into three levels: high, medium and low according to the severity from high to low.

[0095] In one embodiment, the high severity and the medium severity use the first coefficient threshold μ1 as the distinguishing boundary, and the medium severity and the low severity use the second coefficient threshold μ2 as the distinguishing boundary, and the first coefficient threshold is less than the second coefficient threshold. The first coefficient threshold μ1 is preferably in the range of 0.01 to 0.30. The second coefficient threshold μ2 is preferably in the range of 0.30 to 0.50.

[0096] In the embodiment of the present invention, the severity of equipment defects can be divided according to Table 1:

[0097] Table 1 Classification of equipment defect severity

[0098] Risk Level f high <![CDATA[0≤f<μ1]]> middle <![CDATA[μ1≤f<μ2]]>

[0099]

[0100] Embodiment 2

[0101] Based on the above-mentioned first embodiment, the storage tank is taken as the petrochemical equipment to be evaluated to illustrate the process involved in the defect detection method described in the embodiment of the present invention.

[0102] S1: The storage tank is divided into four inspection areas: cylinder, upper head, lower head and short circuit, and secondary division is performed according to the size of each inspection area.

[0103] S2: Use the 3D shape detection method to scan each inspection area of ​​each device to obtain the defect point cloud data of each inspection area. According to the design parameters, the standard point cloud data of each inspection area is established.

[0104] S3: Use a defect recognition model based on 3D target detection to identify defect point cloud data of each inspection area. The model uses a cuboid to identify defect locations in the point cloud data and outputs the coordinates of the eight vertices corresponding to the cuboid.

[0105] The steps of obtaining the defect recognition model include:

[0106] S31: Divide the inspection area of ​​the storage tank according to the method of S1, use a 3D scanner to obtain defect point cloud data, and mark each defect with the corresponding minimum circumscribed cuboid, such as Figure 2 As shown;

[0107] Wherein, the inspected storage tank described in S1 is not among the storage tanks described in S31;

[0108] S32: Use the point cloud data annotated in S31 to train a 3D object detection model, and adjust the data, model structure, and training strategy according to the model accuracy until the required accuracy is met;

[0109] Among them, the point cloud data belonging to the same tank do not appear in the training set and the test set at the same time.

[0110] S4: Calculate the defect opening area coefficient, defect depth coefficient and defect metal loss coefficient in the inspection area in sequence.

[0111] Wherein, obtaining the defect opening area coefficient of the inspection area includes the following steps:

[0112] S411: For each rectangular area output by a single defect recognition model, such as Figure 3 As shown in the figure, the opening contour of the defect point cloud is projected onto the three mutually perpendicular faces of the rectangular area, and the projection area is calculated.

[0113] S412: The maximum value of the projected area is taken as the opening area s of the defect loss , and the area of ​​the corresponding rectangular parallelepiped face is recorded as s. The defect opening area coefficient of the rectangular parallelepiped region S = (ss loss ) / s;

[0114] S413: Calculate the opening area coefficient of each rectangular region output by the defect recognition model in sequence, and take the maximum value as the defect opening area coefficient of the inspection area.

[0115] The step of obtaining the defect depth coefficient of the inspection area includes:

[0116] S421: For each rectangular area output by a single defect recognition model, after calculating the projection area according to S411-S413, take the projection surface used and the corresponding parallel surface, and the distance between the two surfaces is the defect depth d loss The design wall thickness of this area is recorded as d, and the defect depth coefficient of the cuboid is D = (dd loss ) / d.

[0117] S422: After calculating the defect depth coefficient of each rectangular area output by the defect recognition model in sequence, take the maximum value as the defect depth coefficient of the inspection area.

[0118] The step of obtaining the defect metal loss coefficient of the inspection area includes:

[0119] S431: For each rectangular area output by a single defect recognition model, select the corresponding local point cloud from the standard point cloud according to its vertex coordinates, and record the number of points as m. Subtract the selected local point cloud from the defect point cloud, and the absolute value is the metal loss amount m caused by the defect. loss The metal loss coefficient of the cuboid is M = (mm max ) / m.

[0120] S432: After calculating the metal loss coefficient of each rectangular area output by the defect recognition model in turn, take the maximum value as the metal loss coefficient of the inspection area.

[0121] S5: Based on the calculation results of S4, weight coefficients are assigned to each item and the equipment defect risk coefficient is calculated to comprehensively reflect the severity of the tank defects.

[0122] The calculation method of the regional defect risk coefficient is:

[0123] f k =D×[ωS+(1-ω)M]

[0124] The value of ω is 0.5, and k is the number of divided inspection areas.

[0125] Furthermore, the overall defect risk coefficient of the equipment is calculated as follows:

[0126] f=min(f k )

[0127] The severity of the equipment defects is classified as follows:

[0128] Risk Level f high <![CDATA[0≤f<μ1]]> middle <![CDATA[μ1≤f<μ2]]> Low <![CDATA[μ2≤f≤1]]>

[0129] Among them, μ1=0.3, μ2=0.5.

[0130] Embodiment 3

[0131] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is used to detect pipeline defects, except that in embodiment 3:

[0132] Divide the pipeline into several inspection areas at intervals of 1 meter;

[0133] The rest of the process is similar to that of the second embodiment.

[0134] Embodiment 4

[0135] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is used to detect defects in the heat exchanger, except that in embodiment 4:

[0136] Divide the heat exchanger into inspection areas according to the tube bundle and shell, and perform secondary division of the inspection areas according to the structure and size of each inspection area;

[0137] The severity of the equipment defects is classified as follows:

[0138] Risk Level f high <![CDATA[0≤f<μ1]]> middle <![CDATA[μ1≤f<μ2]]> Low <![CDATA[μ2≤f≤1]]>

[0139] Among them, μ1=0.2, μ2=0.4.

[0140] The rest of the process is similar to that of the second embodiment.

[0141] Embodiment 5

[0142] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is adopted to perform tower equipment defect detection, except that in embodiment 5:

[0143] The tower equipment is divided into inspection areas according to size; the inspection area is divided into several layers according to the location of the manholes in each inspection area; and each layer is divided three times according to its size.

[0144] The rest of the process is similar to that of the second embodiment.

[0145] Embodiment 6

[0146] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is adopted to detect furnace tube defects, except that in embodiment 6:

[0147] The furnace tube is divided into inspection areas according to its diameter, and each inspection area is further divided into five sub-areas from bottom to top; each sub-area of ​​the furnace tube is inspected according to the defect detection method described in the second embodiment of the present invention.

[0148] The rest of the process is similar to that of the second embodiment.

[0149] Embodiment 7

[0150] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is used to detect defects on the outer wall and the connecting pipe of the gas buffer tank, except that in embodiment 7:

[0151] The gas buffer tank is divided into inspection areas; the gas buffer tank and the connecting pipe part are divided into an area separately; and each area is inspected according to the defect detection method described in the second embodiment of the present invention.

[0152] The rest of the process is similar to that of the second embodiment.

[0153] Embodiment 8

[0154] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is used to detect short-circuit defects of air coolers, except that in embodiment 8:

[0155] A single short circuit on the air cooler is regarded as a single inspection area, and inspection is performed according to the defect detection method described in the second embodiment of the present invention.

[0156] The rest of the process is similar to that of the second embodiment.

[0157] Embodiment 9

[0158] Based on the above-mentioned embodiment 2, a method similar to that of embodiment 2 is used to detect defects in the instrument air tank, except that in embodiment 9:

[0159] The outer wall of the instrument air tank is divided into five inspection areas according to the upper head, cylinder, lower head, manhole and short circuit, and the inspection area is divided again according to the size.

[0160] The rest of the process is similar to that of the second embodiment.

[0161] Embodiment 10

[0162] Based on the above-mentioned embodiment 2, flange defect detection is performed using a method similar to that of embodiment 2, except that in embodiment 10:

[0163] The flange is divided into several inspection areas according to size and different surfaces.

[0164] The rest of the process is similar to that of the second embodiment.

[0165] Embodiment 11

[0166] Based on the defect detection methods described in the first to tenth embodiments, the present invention further provides a system for detecting defects in petrochemical equipment (also referred to as a "defect detection system"). The defect detection system is used to implement the defect detection methods described in the first to tenth embodiments.

[0167] Figure 5 Schematic diagram of the structure of the system for detecting defects in petrochemical equipment according to an embodiment of the present application. Figure 5 As shown, the defect detection system described in the embodiment of the present invention includes: a point cloud acquisition module 51, a defect identification module 52, a statistical analysis module 53 and a risk assessment module 54.

[0168] The point cloud acquisition module 51 is implemented according to the method described in the above step S110, and is configured to divide the petrochemical equipment to be evaluated into multiple inspection areas, and use three-dimensional morphology measurement to obtain defect point cloud data of each inspection area; the defect recognition module 52 is implemented according to the method described in the above step S120, and is configured to identify the defects of each inspection area and obtain the vertex features of each defect by using a defect recognition model based on the defect point cloud data of each inspection area; the statistical analysis module 53 is implemented according to the method described in the above step S130, and is configured to calculate the defect evaluation features of each inspection area based on the vertex features and defect point clouds of each defect, and the defect evaluation features include defect opening area coefficient, defect depth coefficient and defect metal loss coefficient; the risk assessment module 54 is implemented according to the method described in the above step S140, and is configured to obtain a defect risk coefficient for comprehensively characterizing the severity of the overall defect features of the petrochemical equipment to be evaluated based on the defect evaluation features of each inspection area.

[0169] In one embodiment, the point cloud acquisition module 51 is further configured to divide the petrochemical equipment to be evaluated into multiple inspection areas according to the type of the petrochemical equipment to be evaluated and at least one of the function, structural composition and size of the equipment.

[0170] In one embodiment, the defect recognition model is constructed according to the following process:

[0171] S101. Use a 3D scanning device to scan different inspection areas in various petrochemical equipment to obtain an original point cloud set;

[0172] S102, marking the smallest circumscribed cuboid of each defect in the original point cloud set;

[0173] S103, establishing a basic neural network model, using the original point cloud set and the smallest circumscribed cuboid of each marked defect, training and testing the basic neural network model to obtain a defect recognition model.

[0174] The present invention discloses a method and system for detecting defects in petrochemical equipment. The method and system can be widely used in non-contact non-destructive defect detection and measurement of equipment in the petroleum and chemical fields, without any requirements on the material of the inspected equipment, and can effectively improve the accuracy of defect detection and the efficiency of automated detection. Among them, three-dimensional measurement technology is introduced into the detection of defects on the inner and outer walls of petrochemical equipment, and the beneficial effects produced are as follows:

[0175] 1. Improve the measurement accuracy of non-contact and non-destructive testing in the field of petrochemical equipment testing to sub-millimeter level; sub-millimeter dimensional measurement accuracy provides high-precision initial measurement values ​​for subsequent equipment remaining life prediction methods and calculation models based on measured defects such as crack depth and length, which is conducive to accurately predicting the remaining life of the equipment under test according to its actual situation;

[0176] 2. The deep learning-based 3D target detection method is used to realize the automatic identification of complex petrochemical equipment parts and their surface defects and the automatic positioning of defect positions, improve the efficiency of equipment defect detection, and realize the automatic identification of equipment defects on the point cloud data set: automatically identify and locate defects within the tested area, and automatically find the position and corresponding size information of the deepest defect or the defect with the largest opening area according to the defect depth or defect opening area;

[0177] 3. In order to accurately measure the three-dimensional morphology and size of defects, it is proposed to use the defect opening area, defect depth and metal loss measured automatically to quantitatively judge the severity of equipment defects without human intervention. By proposing a quantitative evaluation scheme for the severity of petrochemical equipment defects, it provides data support for the comprehensive management and preventive maintenance of petrochemical equipment.

[0178] 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.

[0179] 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.

[0180] 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.

[0181] 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.

[0182] 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.

[0183] 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.

[0184] 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 method for detecting defects in petrochemical equipment, characterized in that: include: The petrochemical equipment to be evaluated is divided into multiple inspection areas, and three-dimensional topography measurement is used to obtain defect point cloud data of each inspection area; According to the defect point cloud data of each inspection area, using a defect recognition model, the defects of each inspection area are identified and the vertex features of each defect are obtained; Calculating defect evaluation features of each inspection area according to the vertex features of each defect and the defect point cloud, wherein the defect evaluation features include a defect opening area coefficient, a defect depth coefficient, and a defect metal loss coefficient; According to the defect evaluation characteristics of each inspection area, a defect risk coefficient for comprehensively characterizing the severity of the overall defect characteristics of the petrochemical equipment to be evaluated is obtained.

2. The method according to claim 1, characterized in that In the process of dividing the petrochemical equipment to be evaluated into multiple inspection areas, including: According to the type of the petrochemical equipment to be evaluated, and in accordance with at least one factor of the function, structural composition and size of the equipment, the petrochemical equipment to be evaluated is divided into a plurality of inspection areas.

3. The method according to claim 1 or 2, characterized in that: The defect recognition model is constructed by the following method: Use 3D scanning equipment to scan different inspection areas in various petrochemical equipment to obtain a collection of original point clouds; Marking the smallest circumscribed cuboid of each defect in the original point cloud set; A basic neural network model is established, and the basic neural network model is trained and tested using the original point cloud set and the smallest circumscribed cuboid of each marked defect to obtain the defect recognition model.

4. The method according to claim 3, characterized in that Point cloud data belonging to the same device do not appear in the model training set and the model test set at the same time; The petrochemical equipment to be evaluated is not in the petrochemical equipment in the original point cloud set.

5. The method according to any one of claims 1 to 4, characterized in that: The process of calculating the defect risk factor includes: According to the defect evaluation characteristics of each inspection area, the unit risk coefficient of each inspection area is calculated, wherein the unit risk coefficient is calculated using the following expression: f k =D×[ωS+(1-ω)M] Among them, f k represents the unit risk coefficient, D represents the defect depth coefficient, S represents the defect opening area coefficient, M represents the defect metal loss coefficient, and ω represents the weight coefficient; The defect risk coefficient is obtained according to the unit risk coefficient of each inspection area.

6. The method according to claim 5, characterized in that The method further includes: determining the defect severity level of the petrochemical equipment to be evaluated according to the defect risk coefficient, wherein the defect severity level is divided into high, medium and low in descending order of severity, wherein: The high severity and the medium severity use the first coefficient threshold as a distinguishing boundary, and the medium severity and the low severity use the second coefficient threshold as a distinguishing boundary, and the first coefficient threshold is smaller than the second coefficient threshold.

7. The method according to any one of claims 1 to 6, characterized in that: According to the rectangular vertex features of each defect and the defect point cloud, the defect opening area coefficient of each inspection area is calculated, including: Project the defect point cloud contour of each identified defect onto three mutually perpendicular faces of the corresponding rectangular area to obtain the projection area of ​​each face, and then take the maximum value of the three projection areas as the opening area of ​​the current defect; According to the opening area of ​​each defect, combined with the area of ​​the rectangular area where the corresponding maximum projection area is located, the opening area coefficient of the current defect is calculated, wherein the opening area coefficient is calculated using the following expression: S=(s-s loss ) / s Among them, S represents the defect opening area coefficient, s represents the area of ​​the rectangular area where the maximum projection area corresponding to each defect is located, and s loss Indicates the opening area of ​​each defect; According to the opening area coefficient of each defect, the defect opening area coefficient of each inspection area is determined.

8. The method according to any one of claims 1 to 7, characterized in that: According to the rectangular vertex features of each defect and the defect point cloud, the defect depth coefficient of each inspection area is calculated, including: Project the defect point cloud contour of each identified defect onto three mutually perpendicular faces of the corresponding rectangular area to obtain the projection area of ​​each face, and then take the maximum value of the three projection areas as the available rectangular face of the current defect; The defect depth of each defect is obtained by calculating the distance between the available rectangular surface of each defect and its corresponding parallel surface. Based on this, combined with the designed wall thickness of the petrochemical equipment to be evaluated, the defect depth coefficient of each defect is obtained, wherein the defect depth coefficient is calculated using the following expression: D=(d-d loss ) / d Where D represents the defect opening area coefficient, d represents the designed wall thickness, and d loss Indicates the defect depth of each defect; According to the defect depth coefficient of each defect, the defect depth coefficient of each inspection area is determined.

9. The method according to any one of claims 1 to 8, characterized in that: According to the rectangular vertex features of each defect and the defect point cloud, the defect metal loss coefficient of each inspection area is calculated, including: Extracting the corresponding local point cloud from the standard point cloud data for each identified rectangle vertex coordinate, thereby recording the number of points of the local point cloud of each defect, wherein the standard point cloud data is the point cloud data of the petrochemical equipment to be evaluated in a defect-free state; The defect metal loss coefficient of each defect is obtained by calculating the absolute value of the difference between the number of local point clouds of each defect and the number of defect point clouds of the corresponding defect, wherein the defect metal loss coefficient is calculated using the following expression: M=(m-m max ) / m Where M represents the defect metal loss coefficient, m represents the number of points in the local point cloud of each defect, and m max The number of points in the defect point cloud representing each defect; According to the defect metal loss coefficient of each defect, the defect metal loss coefficient of each inspection area is determined.

10. A system for detecting defects in petrochemical equipment, characterized in that: The system is used to implement the method according to any one of claims 1 to 9, wherein the system comprises: A point cloud acquisition module configured to divide the petrochemical equipment to be evaluated into multiple inspection areas and to obtain defect point cloud data of each inspection area using three-dimensional topography measurement; A defect recognition module is configured to identify defects in each inspection area and obtain vertex features of each defect based on the defect point cloud data of each inspection area using a defect recognition model; A statistical analysis module, configured to calculate defect evaluation features of each inspection area according to the vertex features of each defect and the defect point cloud, wherein the defect evaluation features include a defect opening area coefficient, a defect depth coefficient, and a defect metal loss coefficient; The risk assessment module is configured to obtain a defect risk coefficient for comprehensively representing the severity of the overall defect characteristics of the petrochemical equipment to be evaluated based on the defect assessment characteristics of each inspection area.

11. The system according to claim 10, characterized in that The point cloud acquisition module is further configured to divide the petrochemical equipment to be evaluated into multiple inspection areas according to the type of the petrochemical equipment to be evaluated and at least one factor of the function, structural composition and size of the equipment.

12. The system according to claim 10 or 11, characterized in that The defect recognition model is constructed according to the following process: Use 3D scanning equipment to scan different inspection areas in various petrochemical equipment to obtain a collection of original point clouds; Marking the smallest circumscribed cuboid of each defect in the original point cloud set; A basic neural network model is established, and the basic neural network model is trained and tested using the original point cloud set and the smallest circumscribed cuboid of each marked defect to obtain the defect recognition model.

Images