Assembly type pipeline part intelligent detection system and method

Abstract

The invention discloses an assembly type pipeline part intelligent detection system and method.The system comprises a detection device and a control cabinet, the detection device comprises a bottom support, a detection mechanism is installed on one side of the top end of the bottom support, and two fixing mechanisms are fixedly connected to the other side of the top end of the bottom support; the two fixing mechanisms are distributed in a mirror image mode, each fixing mechanism comprises a supporting frame, each supporting frame is fixedly connected with the bottom support, a sliding way is arranged at the top end of each supporting frame, a sliding block is connected to the inner surface of each sliding way in a sliding mode, and the inner surface of each sliding block is sleeved with a first lead screw in a threaded mode; the control cabinet comprises a cabinet body, an industrial personal computer, a display and a control panel. The device can adapt to various pipelines with different sizes, and is high in universality and good in detection effect.

Description

Technical Field

[0001] The present invention belongs to the technical field of assembly-type pipeline component detection, and in particular relates to an intelligent detection system and method for assembly-type pipeline components. Background Art

[0002] Prefabricated piping is a new type of piping project in which pipe components are designed, cut, grooved, welded, and flanged in a factory to form pipe components whose shapes and sizes meet the requirements of on-site construction. These components are then assembled on-site by the client.

[0003] Currently, there are many types of equipment used to produce and manufacture prefabricated pipes, but there are no tools to verify that the manufactured pipes conform to the drawings. Due to the large size of the pipe components, up to 5 meters in length, with diameters ranging from 100mm to 600mm, and a wide variety of shapes, the number of branch pipes varies, and the lengths of the branches vary. The angles between the branch pipes are often non-planar. Current measurement methods, using tape measures, result in large length deviations and make it impossible to measure position and non-planar angles. This results in unsuccessful on-site assembly, high repair rates, and low installation efficiency. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an intelligent detection system for assembled pipeline components in response to the above-mentioned deficiencies in the existing technology, which can improve the factory qualification rate of assembled pipelines and enhance the installation efficiency of pipelines.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: an intelligent detection system for assembled pipeline components, including a detection device and a control cabinet, the detection device includes a bottom bracket, a detection mechanism is installed on one side of the top of the bottom bracket, and two fixing mechanisms are fixedly connected to the other side of the top of the bottom bracket, and the two fixing mechanisms are distributed in a mirror image, and the fixing mechanism includes a support frame, the support frame is fixedly connected to the bottom bracket, and the top of the support frame is provided with a slide, the inner surface of the slide is slidably connected to a slider, the inner surface of the slider is threaded with a screw rod 1, one end of the screw rod 1 passes through the slide and extends to the outside of the slide, and the top of the slide is provided with two A fixed frame, one of which is fixedly connected to the slider, and the other fixed frame is fixedly connected to the slide, the inner surfaces of the two fixed frames are rotatably connected to transmission rollers, one end of one of the transmission rollers passes through the fixed frame and extends outside the fixed frame, the outer surface of the transmission roller outside the fixed frame is fixedly sleeved with a worm gear, one side of the slide is fixedly connected to a power motor, the output end of the power motor is fixedly connected to a worm, and the worm and the worm gear are meshed with each other; the detection mechanism includes a detection platform, the top of the detection platform is fixedly connected to an electric slide rail, the top of the electric slide rail is slidably connected to a slide, and the top of the slide is equipped with a laser and a detection camera.

[0006] Preferably, a fixed rod is fixedly connected to one side of the top of the slide, and a movable rod is slidably connected to the inner surface of the fixed rod. The fixed rod is hollow, and a through-type adjustment slot is provided at both ends of the fixed rod. A limiting slot is provided on the inner surface of the movable rod, and two limiting plates are slidably connected to the inner surface of the limiting slot. The corresponding ends of the two limiting plates are jointly fixedly connected to a return spring, and the opposite ends of the two limiting plates are fixedly connected to an adjusting rod used in conjunction with the adjusting slot, and the outer surfaces of the two adjusting rods are slidably connected to the inner surface of the movable rod.

[0007] Preferably, one end of the fixed rod is rotatably connected to rotating block 1, the top end of the adjusting rod is rotatably connected to a pressure rod, one end of the pressure rod is rotatably connected to rotating block 2, the inner surface of the rotating block 2 is threadedly sleeved with screw rod 2, and the other end of the screw rod 2 is rotatably connected to rotating block 1.

[0008] Preferably, one end of the screw rod 1 passes through the slideway and is fixedly connected to the rotating handle 1, and one end of the screw rod 2 passes through the rotating block 2 and is fixedly connected to the rotating handle 2.

[0009] Preferably, one end of the pressure rod away from the second rotating block is rotatably connected to two rollers.

[0010] Preferably, both ends of the electric slide rail are fixedly connected to limit blocks, and the two limit blocks are fixedly connected to the detection platform.

[0011] The present invention also discloses an intelligent detection method for assembled pipeline components, which includes the following steps: Step 1: Turn on the power, start the industrial computer, and set the detection parameters; Step 2: Perform the inspection operation on the operation interface of the display. The industrial computer analyzes the CAD drawing of the pipeline component to be inspected and displays the graphics of the pipeline component on the display. Step 3: Use a crane to place the pipe component on the transmission rollers of the two fixing mechanisms and adjust it to a suitable position to accommodate the width of the pipe component; Step 4: Operate the components on the two fixing mechanisms in sequence to press the pipeline components tightly; Step 5: Install the laser inclinometer on the flange of the pipeline. The laser inclinometer and the flange are attracted together by magnets. Step 6. Press the "Pipeline Attitude Adjustment" button on the display, adjust the pipeline position through the dialog box, and the power motor drives the pipeline to rotate. When the laser inclinometer shows 90 degrees or the laser is horizontal, the pipeline attitude adjustment is completed. Press the OK button to return to the main interface; Step 7. Press the "Auto Measure" button on the monitor. The synchronous belt drives the laser and the detection camera to scan the pipeline. After the industrial computer collects the image, it uses visual recognition and visual measurement technology to complete the pipeline inspection. Step 8. After the image analysis is completed, press the "Measurement Result" button to display and print the measurement results.

[0012] Compared with the prior art, the present invention has the following advantages: 1. The present invention provides a screw rod. When the screw rod is rotated by turning the handle, the slider and one of the fixing brackets are moved by rotating the screw rod, thereby quickly adjusting the distance between the two fixing brackets. After that, the pipe can be fixed by turning the second handle. This allows the device to adapt to pipes of various sizes, effectively improving the versatility of the device. 2. The present invention provides a pressure rod. Under the action of the pressure rod and the roller, after the pressure rod cooperates with the roller to fix the pipe, the motor drives the transmission roller to rotate, so that the transmission roller can drive the upper pipe to rotate. Comprehensive inspection can be carried out without adjusting the clamping device, effectively improving the inspection efficiency. 3. The present invention can adapt to pipelines of various sizes, has strong versatility and good detection effect.

[0013] 4. The present invention acquires three-dimensional pipeline data through high-density point cloud acquisition, extracts pipeline geometric features using point cloud segmentation and registration algorithms, and converts the local coordinate system into a global coordinate system. A high-precision pipeline model is then generated through three-dimensional reconstruction, key dimensional data is extracted, and deviations are evaluated. For flange connections, the present invention uses normal vector analysis and least squares optimization to calculate torsional deviations, and simulates stress distribution through finite element analysis. Finally, an adaptive adjustment algorithm is used to optimize installation parameters, and virtual assembly verification is performed. The present invention achieves high-precision three-dimensional reconstruction and precise calibration of complex pipeline systems, which can effectively improve pipeline installation accuracy and connection stability, and provides strong support for the design optimization and installation and commissioning of industrial pipeline systems.

[0014] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the overall structure of the present invention; Figure 2 This is a structural diagram of the fixing mechanism of the present invention; Figure 3 For the present invention Figure 2 A schematic diagram of the structure at center A; Figure 4 Schematic cross-sectional view of the fixed rod and the movable rod of the present invention; Figure 5Schematic diagram of the detection mechanism structure of the present invention; Figure 6 This is the image processing flow chart of the present invention.

[0016] In the figure: 1. Bottom bracket; 2. Detection mechanism; 3. Fixing mechanism; 4. Support frame; Electric slide rail; 22. Detection platform; 23. Slide; 24. Limit block; 25. Laser; 26. Detection camera; 31. Screw rod 1; 32. Turning handle 1; 33. Slide; 34. Slider; 35. Fixed frame; 36. Transmission roller; 37. Roller; 38. Movable rod; 39. Pressure rod; 40. Turning handle 2; 41. Turning block 2; 42. Screw rod 2; 43. Turning block 1; 44. Fixed rod; 45. Worm; 46. Power motor; 47. Worm gear; 441. Adjusting slot; 442. Adjusting rod; 443. Return spring; 444. Limiting slot; 445. Limiting plate. DETAILED DESCRIPTION

[0017] Example 1

[0018] like Figure 1 As shown, the intelligent detection system for assembled pipeline components of the present invention includes a detection device and a control cabinet. The detection device includes a bottom bracket 1, a detection mechanism 2 is installed on one side of the top of the bottom bracket 1, and two fixing mechanisms 3 are fixedly connected to the other side of the top of the bottom bracket 1. The two fixing mechanisms 3 are distributed in a mirror image. The fixing mechanism 3 includes a support frame 4, which is fixedly connected to the bottom bracket 1. A slide 33 is provided at the top of the support frame 4. The inner surface of the slide 33 is slidably connected to a slider 34. The inner surface of the slider 34 is threadedly sleeved with a screw rod 31. One end of the screw rod 31 passes through the slide 33 and extends to the outside of the slide 33. Two fixing frames 35 are provided at the top of the slide 33, one of which is fixedly connected to the slider 34, and the other is fixedly connected to the slide 33. The inner surfaces of the two fixing frames 35 are rotatably connected to drive rollers 36, one end of which passes through the fixing frame 35 and extends to the outside of the fixing frame 35. The outer surface of the drive roller 36 outside the fixing frame 35 is fixedly sleeved with a worm gear 47. One side of the slide 33 is fixedly connected to a power motor 46. The power motor 4 The output end of 6 is fixedly connected to a worm 45, which is meshed with a worm gear 47, thereby driving the worm gear 47 and the transmission roller 36 to rotate, thereby realizing automatic driving of the pipeline component to rotate, facilitating its all-round detection and improving the detection efficiency; the detection mechanism 2 includes a detection platform 22, the top of the detection platform 22 is fixedly connected to an electric slide rail 21, and the top of the electric slide rail 21 is slidably connected to a slide 23. The slide 23 serves as a carrying platform for the laser 25 and the detection camera 26, which can be flexibly moved to fully scan the pipeline components. The top of the slide 23 is equipped with a laser 25 and a detection camera 26. The laser 25 is used to emit laser light and the detection camera 26 captures images. The two can cooperate to perform high-precision detection on the size, shape, weld quality and other aspects of the pipeline components to ensure the quality of the products leaving the factory; The control cabinet includes a cabinet body, a display arranged on the top of the cabinet body, and an industrial computer and a control board arranged in the cabinet body. The detection camera 26 and the display are both connected to the industrial computer. A control circuit module is integrated on the control board, and the control circuit module is connected to and communicates with the industrial computer.

[0019] In specific implementations, the industrial computer is equipped with a high-speed image acquisition card responsible for real-time acquisition of images captured by the detection camera 26. The control circuit module includes a four-axis motion controller and a power supply module. The input of the four-axis motion controller is connected to a photoelectric limit switch, and the output of the four-axis motion controller is connected to a motor driver. The power motor 46 is connected to the output of the motor driver. The four-axis motion controller is connected to the industrial computer via a network cable, receiving instructions from the industrial computer after image acquisition and analysis. It is also connected to the motor driver via a signal cable, driving the power motor 46 to rotate at a given speed and direction to a given position.

[0020] Specifically, the outer surface of the transmission roller 36 is coated with an anti-skid coating, which can increase the friction between the transmission roller 36 and the pipeline components, prevent the pipeline components from slipping during the rotation detection process, and ensure that the detection process is carried out stably.

[0021] like Figures 1 to 5 As shown, a fixed rod 44 is fixedly connected to one side of the top of the slide 33, and the inner surface of the fixed rod 44 is slidably connected to the movable rod 38. The fixed rod 44 is hollow, and both ends of the fixed rod 44 are provided with a through-type adjustment slot 441. A limiting slot 444 is provided on the inner surface of the movable rod 38. Two limiting plates 445 are slidably connected to the inner surface of the limiting slot 444. The corresponding ends of the two limiting plates 445 are fixedly connected to the return spring 443. The opposite ends of the two limiting plates 445 are fixedly connected with an adjusting rod 442 used to cooperate with the adjusting slot 441. The outer surfaces of the two adjusting rods 442 are slidably connected to the inner surface of the movable rod 38. The adjusting rod 442 cooperates with the adjusting slot 441 to accurately control the lifting and lowering of the movable rod 38 to meet the fixing requirements of pipeline components of different heights.

[0022] Furthermore, one end of the fixed rod 44 is rotatably connected to the rotating block 1 43. The top end of the adjusting rod 442 is rotatably connected to the pressure rod 39. One end of the pressure rod 39 is rotatably connected to the rotating block 2 41. The inner surface of the rotating block 2 41 is threadedly connected to the screw rod 2 42. The other end of the screw rod 2 42 is rotatably connected to the rotating block 1 43. By rotating the screw rod 2 42, the degree of compression of the pressure rod 39 can be precisely controlled to accommodate the fixation of pipe components of different diameters. The other end of the screw rod 2 42 is rotatably connected to the rotating block 1 43.

[0023] like Figures 1 to 5 As shown, one end of the screw rod 1 31 passes through the slide 33 and is fixedly connected to the rotating handle 1 32, and one end of the screw rod 2 42 passes through the rotating block 2 41 and is fixedly connected to the rotating handle 2 40, which makes it easy for the operator to accurately control the pressure rod 39 and the slider 34, thereby improving the operating convenience.

[0024] It is worth noting that the end of the pressure rod 39 away from the rotating block 2 41 is rotatably connected to two rollers 37. When the pipe components are compressed, the rollers 37 can roll along the shape of the pipe components to avoid scratching the pipe surface, while reducing friction resistance and making the compression process smoother.

[0025] It is worth emphasizing that both ends of the electric slide rail 21 are fixedly connected to the limit blocks 24, and the two limit blocks 24 are fixedly connected to the detection platform 22. The limit blocks 24 can prevent the slide 23 from sliding out of the electric slide rail 21, thereby ensuring the safety of the movement of the slide 23 and avoiding damage to the equipment.

[0026] The power motor 46, laser 25, and detection camera 26 are prior art and will not be described in detail. The wiring diagram of the motor in the present invention is common knowledge in the art, and its operating principle is already known technology. The model is selected according to actual use, so the control method and wiring layout of the motor will not be explained in detail.

[0027] In specific implementation, the control cabinet is the core of the intelligent detection device.

[0028] During specific implementation, the on-site installation is as follows: (1) The site floor must be cement and flat; (2) Mark the footing position of the pipeline component rotating platform on the ground, install expansion screws, and fix the platform; (3) Draw the center line of the 3D camera moving platform at a distance of 2030 mm from the center of the pipeline component rotating platform. The two center lines must be parallel with an error of less than 2 mm. Install the expansion screws to fix the 3D camera moving platform. (4) Install the control cabinet, the location can be determined according to user needs; (5) Connect the wires according to the instructions; (6) Adjust the horizontal adjustment device of the rotating platform of the pipeline component to make the platform level with an error within 1mm; (7) Adjust the horizontal adjustment device of the 3D camera moving platform to make the platform level with an error within 1mm; On-site calibration is: (1) Prepare pipes with diameters of 100mm-600mm and lengths of 5000mm respectively; (2) Hoist the above-mentioned pipelines onto the pipeline component rotating platform respectively, install a spirit level on the upper part of the pipeline, turn the handwheel on the pipeline level adjustment device to make it level, and record the pipe diameter and the corresponding handwheel scale.

[0029] Example 2

[0030] The intelligent detection method for assembled pipeline components of the present invention comprises the following steps: Step 1: Turn on the power, start the industrial computer, and set the detection parameters; In specific implementation, the detection parameters include the starting speed, acceleration, and pulse unit of the pipeline component rotation platform, the pulse unit, starting speed, and acceleration of the 3D camera mobile platform, the sampling rate, sampling area, triggering mode, laser intensity, x-axis sampling spacing, y-axis sampling spacing, and z-axis sampling spacing of the 3D camera; Step 2: Perform the inspection operation on the operation interface of the display. The industrial computer analyzes the CAD drawing of the pipeline component to be inspected and displays the graphics of the pipeline component on the display. During the specific implementation, press the "Load Drawing" button on the operation interface of the monitor to pop up the file selection dialog box, select the CAD drawing of the pipeline component to be tested, press OK, and the industrial computer will automatically parse the CAD file and display the graphics of the pipeline component on the monitor. At the same time, the drawing number, main pipe diameter, and branch pipe angle are displayed in the table on the right; Step 3: Use a crane to place the pipe component on the transmission rollers 36 of the two fixing mechanisms 3 and adjust them to a suitable position to accommodate the width of the pipe component; Step 4: Operate the components on the two fixing mechanisms 3 in sequence to press the pipeline components; Step 5: Install the laser inclinometer on the flange of the pipeline. The laser inclinometer and the flange are attracted together by magnets. Step 6. Press the "Pipeline Attitude Adjustment" button on the display and adjust the pipeline position through the dialog box. The power motor 46 drives the pipeline to rotate. When the laser inclinometer shows 90 degrees or the laser is horizontal, the pipeline attitude adjustment is completed. Press the OK button to return to the main interface. Step 7: Press the "Auto Measure" button on the display, and the synchronous belt drives the laser (25) and the detection camera 26 to scan the pipeline. After the industrial computer collects the image, it uses visual recognition and visual measurement technology to complete the pipeline detection; Step 8. After the image analysis is completed, press the "Measurement Result" button to display and print the measurement results.

[0031] In specific implementation, the pipeline detection includes position judgment, flange position measurement, flange three-dimensional torsion angle measurement, main pipeline length measurement and branch pipeline size measurement; the image processing process is as follows Figure 6 shown.

[0032] In addition, during the specific implementation, the pipeline quality can also be inspected in detail. The specific inspection methods include: Step S101, 1. Perform high-precision point cloud data collection on the pipeline system in a complex three-dimensional environment using laser scanning technology, collecting data at a density of at least 10,000 points per square meter to obtain an initial point cloud data set.

[0033] Laser scanning technology is used to comprehensively collect data on pipeline systems in complex environments, construct an initial point cloud dataset, and obtain high-density point cloud information covering three-dimensional space. Based on the initial point cloud dataset, point cloud registration technology is used to spatially align the collected data to obtain a unified point cloud model. If there are noise points or redundant data in the point cloud model, the point cloud model is optimized through filtering to determine the optimized point cloud dataset. For the optimized point cloud dataset, segmentation technology is used to extract features of the pipeline system structure and obtain geometric feature information of the pipeline system. Based on this geometric feature information, the pipeline system is spatially modeled using three-dimensional reconstruction technology to obtain a three-dimensional model of the pipeline system. If there is data missing in local areas of the three-dimensional model, interpolation technology is used to complete the data in the missing areas and determine the completeness of the completed model. Based on the completed model, accuracy analysis methods are used to evaluate the deviation between the model and the actual pipeline system to determine the final three-dimensional modeling result.

[0034] Step S102, 2: Based on the initial point cloud data set, a point cloud segmentation algorithm is used to extract the central axis and end face plane of the pipeline, determine the spatial position of the pipeline and the relative position relationship between the flange, and obtain the pipeline geometric feature data.

[0035] Through point cloud data processing, segmentation techniques are used to analyze the initial input, separating relevant point sets for the pipe center and end face planes to obtain preliminary segmentation results. Based on these preliminary segmentation results, axis extraction is performed on the pipe center point set, and a preset geometric fitting method is used to determine the axis direction and position data of the pipe center. This axis direction and position data, combined with the end face plane point set, are used to perform spatial geometry calculations to obtain the spatial position information of the pipe and determine its distribution characteristics in three-dimensional space. If the spatial position information does not meet the preset threshold range, the point cloud data is re-segmented in a local area to obtain corrected spatial position data. Based on this corrected spatial position data, combined with the point cloud distribution of the relative position of the flange, the positional relationship between the two is analyzed to determine the relative geometric constraints between the pipe and flange. Using the relative geometric constraints, the feature data for the pipe center and end face planes are integrated to derive the complete pipe geometric features and obtain the final feature description data. If the final feature description data contains missing or abnormal data, the initial input point cloud data is supplemented with sampling, and the segmentation techniques and axis extraction are re-applied to obtain the complete geometric feature data.

[0036] Step S103, 3 If the spatial deviation threshold between the central axis of the pipeline and the theoretical position exceeds 2 mm, the initial point cloud data set is aligned using the iterative closest point algorithm, and at least 10 convergence calculations are performed to control the proportion of point cloud noise outliers within 5% to obtain the corrected point cloud data.

[0037] By comparing the center of the pipeline with the theoretical position, the specific value of the axis deviation is obtained to determine whether the deviation exceeds the preset spatial threshold. If the axis deviation exceeds the preset spatial threshold, the iterative nearest point algorithm is used to perform data alignment for the initial point cloud data set to obtain the point cloud data after preliminary alignment. Based on the point cloud data after preliminary alignment, multiple convergence calculations are performed to adjust the point cloud position to obtain the optimized point cloud data set. For the optimized point cloud data set, the noise outliers are detected to determine whether the ratio of noise outliers is within the specified range. If the ratio of noise outliers exceeds the specified range, the redundant outliers are removed through filtering to obtain the denoised point cloud data set. Based on the denoised point cloud data set, the axis position of the pipeline center is recalculated to determine whether the corrected point cloud data meets the requirements of the theoretical position. If the corrected point cloud data still deviates from the theoretical position, the alignment and denoising processes are repeated to obtain the final corrected data that meets the requirements.

[0038] For example, when comparing the deviation of the pipeline center axis, we can start with the definition of the theoretical position. The theoretical position usually refers to the ideal position of the pipeline axis in the design drawing, while the actual point cloud data may deviate due to the accuracy of the acquisition equipment or environmental interference. Suppose in a certain industrial pipeline inspection scenario, the theoretical axis position is a straight line along the X-axis, and the axis position calculated by the actual point cloud data deviates by 0.5 meters. After obtaining the deviation value through comparison, if the preset spatial threshold is 0.2 meters, it is obviously out of range and requires further processing. This comparison method intuitively reflects potential problems in pipeline installation or data acquisition, which is helpful for subsequent correction.

[0039] For example, using the iterative closest point algorithm for data registration to address deviations exceeding a threshold can be considered a point cloud alignment technique. Suppose the initial point cloud dataset contains 100,000 points, some of which have shifted due to device vibration. The registration process, through multiple iterations, gradually aligns these points to the theoretical model. After initial registration, the overall point cloud offset may be reduced from 0.5 meters to 0.1 meters, but further optimization is still required. The advantage of this method is that it gradually approaches the ideal position, improving data accuracy.

[0040] For example, when performing multiple convergence calculations to adjust point cloud positions, the dataset can be optimized by gradually narrowing the error range. Suppose that after each convergence calculation, the point cloud position error decreases from 0.1 meter to 0.05 meter. After three iterations, the error stabilizes within 0.02 meters, resulting in an optimized point cloud dataset. This approach effectively improves the geometric consistency of the point cloud and lays the foundation for subsequent analysis.

[0041] For example, when determining the proportion of noise outliers, suppose that 5% of the points in the optimized point cloud dataset deviate from the main point cloud by more than 0.3 meters, exceeding the 3% limit. These outliers can be removed through filtering, such as statistical filtering methods, to remove points that deviate too far, resulting in a denoised dataset. This process significantly improves data quality and prevents noise from interfering with subsequent calculations.

[0042] For example, when recalculating the center axis position of a pipeline, based on the denoised point cloud dataset, the axis deviation is assumed to be reduced to 0.01 meters after correction, meeting the theoretical position requirement. This correction process embodies the closed-loop logic of data processing and ensures the reliability of the final data.

[0043] For example, if there are still deviations after correction, additional point cloud sampling density adjustments can be introduced when repeating the registration and denoising processes. Assuming that the initial point cloud density is insufficient and the deviations recur, the sampling points can be increased in the local area, from 1,000 points per cubic meter to 2,000 points, and the registration and denoising can be performed again to finally obtain the corrected data that meets the requirements. This method can effectively deal with the problem of insufficient data and improve processing accuracy. Through the above multi-angle implementation methods, it can be seen that each technical theme plays an important role in pipeline point cloud data processing. Whether it is deviation comparison, data registration, convergence calculation, noise removal or axis correction, they all revolve around improving data accuracy and geometric consistency, ensuring the reliability of pipeline spatial position analysis, and providing a solid foundation for subsequent industrial applications.

[0044] Step S104, 4: By point cloud coordinate system conversion, the corrected point cloud data is converted from the local coordinate system to the global coordinate system, the reference frame is unified, and the globally corrected point cloud data is obtained.

[0045] Initial processing of the point cloud data involves retrieving the raw point cloud data from the storage medium. A preliminary analysis of the data distribution in local coordinates is performed to obtain a local coordinate point cloud set. Based on this local coordinate point cloud set, a pre-established coordinate transformation matrix is ​​used to map the local coordinates to global coordinates, determining the transformed point cloud coordinate positions. If the transformed point cloud coordinate positions deviate, the point cloud coordinate positions are adjusted by comparing them to the standard coordinate distribution in the reference frame, resulting in a corrected intermediate point cloud dataset. The intermediate point cloud dataset is then tested for data consistency in global coordinates. If the test results indicate uneven data distribution, point cloud processing techniques are used to smooth the data, resulting in a more consistent point cloud dataset. This more consistent point cloud dataset is then integrated using coordinate mapping rules within a unified framework to determine whether it meets the global correction standard requirements. If the global correction standard requirements are not met, the integrated point cloud dataset is iteratively optimized, and outliers are identified and removed using the random forest algorithm to obtain the optimized global point cloud data. A final data verification is performed on the optimized global point cloud data, and the verification results are recorded and stored to determine the final globally corrected point cloud data.

[0046] For example, in the initial stages of processing point cloud data, when acquiring raw data from storage media, point cloud information in a specific format can be extracted by reading the data file. For example, in a pipeline inspection project, the raw point cloud data is stored on a local server and contains the 3D coordinate information of the pipeline surface, with a data volume of approximately 5 million points. During preliminary analysis, the density and coverage of the point cloud can be focused on the data distribution in local coordinates. For example, if the point cloud density in certain areas is low, which may affect subsequent processing, these areas need to be recorded for subsequent data supplementation.

[0047] For example, when mapping local coordinates to global coordinates, the local point cloud data can be mapped to a unified global coordinate system using a pre-established coordinate transformation matrix. For example, in a pipeline inspection scenario, the local coordinate system uses the starting point of a certain section of the pipeline as its origin, while the global coordinate system uses the reference point of the entire project as its origin. After adjustment through the transformation matrix, all point cloud data is ensured to be in the same reference frame, facilitating overall analysis. If a coordinate position deviation is detected after the transformation, such as a point cloud of a certain section of the pipeline deviating by 5 mm from the standard position, it is necessary to compare the standard coordinate distribution under the reference frame, adjust the point cloud position, and form an intermediate data set.

[0048] For example, when performing a global coordinate consistency check on an intermediate point cloud dataset, one can examine whether the data distribution is uniform. If the inspection reveals a sparse distribution of point clouds in a certain area, potentially compromising the integrity of the pipeline model, point cloud smoothing techniques can be employed to fill in the sparse areas through interpolation, resulting in a more consistent dataset. This process helps improve data continuity and lays the foundation for subsequent integration.

[0049] For example, in data integration, combining coordinate mapping rules within a unified framework allows point cloud data from different sources to be fused into a single, integrated dataset. For example, in pipeline inspection, data collected by multiple devices needs to be integrated. If the integrated data still does not meet global calibration standards, such as insufficient overlap in certain areas, iterative optimization is required.

[0050] Preferably, a random forest algorithm is used to identify outliers. Assuming that about 2% of the points are identified as outliers, these points are removed to optimize the global point cloud data and ensure data quality.

[0051] For example, during the final data verification phase, the optimized global point cloud data can be verified across multiple dimensions, such as to check whether coordinate accuracy meets expected standards. If the verification results indicate that the error is within 1 mm, the relevant information is recorded and stored to form the final globally corrected point cloud data. This meticulous verification and recording facilitates subsequent tracing and analysis, providing a reliable basis for pipeline inspection.

[0052] Step S105, 5: Based on the global correction point cloud data, a 3D reconstruction technology is used to generate a surface model of the pipeline system, ensuring that the surface reconstruction accuracy deviation is less than 0.5 mm, and an STL format model with at least 5000 triangles per meter of pipeline is generated to obtain a pipeline geometric model.

[0053] The collected point cloud data is preprocessed using filtering techniques to remove noise points and outliers, resulting in a preliminary cleansed point cloud dataset. Based on this cleansed point cloud dataset, a global correction method is applied to spatially register the data, adjust the overall consistency of the point cloud, and determine the corrected point cloud structure. Based on the corrected point cloud structure, an initial surface model is constructed using 3D reconstruction techniques, and basic triangular facets are generated using meshing methods to obtain a preliminary surface morphology. This preliminary surface morphology is optimized, and curvature smoothing techniques are applied to adjust the connectivity between facets, resulting in a smoother surface model. If the surface model data's accuracy deviation exceeds a preset threshold, detailed adjustments are made to the local area, and facets are encrypted based on the point cloud density distribution to determine a surface result that meets the accuracy requirements. Based on the surface result that meets the accuracy requirements, the triangular facet distribution density is adjusted to ensure that the number of facets per unit length meets the predetermined standard, generating a high-density geometric form. Using standard conversion tools, the high-density geometric form is exported as an STL file to complete the final pipeline geometry model.

[0054] For example, filtering technology is a key step in preprocessing collected point cloud data. The goal of filtering is to remove noise points and outliers to ensure data quality. A common implementation method is to use voxel grid filtering, which divides the point cloud data into fixed-size voxel units, such as 0.05-meter per voxel. The points within each voxel are then averaged as a representative point, reducing data volume and smoothing noise. This method is particularly suitable for high-density point cloud data and can effectively reduce the computational burden of subsequent processing.

[0055] For example, when applying global correction methods for spatial registration, point cloud alignment can be achieved through an iterative closest point algorithm. Consider two sets of point cloud data that require registration. First, one set is selected as a reference and the other as the object to be aligned. By finding the closest point pair, the transformation matrix between the two point clouds is calculated. The position and orientation of the point cloud to be aligned are then adjusted to ensure that it coincides as closely as possible with the reference point cloud. This method is particularly important when processing point cloud data collected from multiple perspectives, ensuring overall consistency.

[0056] For example, when performing 3D reconstruction on a corrected point cloud structure, the initial surface model can be constructed using the Poisson reconstruction method. This method constructs a continuous surface based on the point cloud's normal vector information. For example, in the case of pipeline point cloud data, a preliminary cylindrical surface structure can be generated. This method is suitable for processing complex geometries and can effectively restore the overall contour of the pipeline.

[0057] For example, when optimizing the initial surface form, curvature smoothing can be used to adjust the connections between facets. For example, if the triangles in certain areas of the initial surface model are not naturally connected, the curvature differences between adjacent facets can be calculated and the vertex positions adjusted to achieve a smoother transition. This process is particularly critical in curved sections of pipe models, improving the visual continuity of the model.

[0058] For example, if the surface model's accuracy deviation exceeds a preset threshold, such as 0.01 meters, detailed adjustments to the local area are necessary. Based on the point cloud density distribution, the number of triangles in denser areas is increased, for example, from 100 to 200 per square meter, to enhance detail. This approach can effectively improve model accuracy in critical areas.

[0059] For example, when adjusting the density of triangles, we can prioritize increasing the density of triangles at key locations in the pipeline, such as connection points or elbows, to ensure that the number of triangles per unit length meets a predetermined standard, such as no fewer than 500 triangles per meter. This high-density geometry better captures the subtle features of the pipeline.

[0060] For example, when exporting to STL format, standard conversion tools ensure model data compatibility. If the generated pipeline geometry needs to be integrated with other modeling software, STL, as a universal triangular facet format, can be seamlessly imported into other platforms for subsequent editing or analysis. This greatly facilitates data sharing and application.

[0061] Step S106, 6: Extract the pipeline outer diameter and wall thickness data based on the pipeline geometric model, determine whether the pipeline diameter deviation is within the ±0.3 mm threshold, and obtain pipeline size assessment data.

[0062] Raw data is obtained from the pipeline geometry model, and data extraction is performed on the pipeline outer diameter and wall thickness to obtain a preliminary dimensional data set. Based on this preliminary dimensional data set, the pipeline outer diameter and wall thickness are calibrated using a preset standardization method to determine the calibrated dimensional data. Based on the calibrated dimensional data, the diameter deviation of the pipeline outer diameter is calculated. If the diameter deviation exceeds the preset threshold, it is marked as abnormal data, and an abnormality marking result is obtained. Based on the abnormality marking result, the data marked as abnormal is further analyzed using the support vector machine algorithm to determine whether there are systematic errors and obtain the deviation analysis conclusion. Based on the deviation analysis conclusion, the data with systematic errors is corrected to obtain corrected dimensional data. Using the corrected dimensional data, a comprehensive assessment of the pipeline dimensions is performed. If the assessment results show that the deviation is still outside the preset threshold, it is recorded as unqualified data, and the final assessment result is determined. Based on the final assessment result, classified pipeline dimension data is generated, and the unqualified data is archived to obtain a classified archive record.

[0063] For example, when acquiring raw data from a pipe geometry model, key points related to the pipe's outer diameter and wall thickness can be extracted through analysis of 3D scan data. For outer diameter data extraction, the circumferential point set of the pipe cross section can be focused on, and a circular contour can be fitted to initially determine the diameter range. For wall thickness data, preliminary values ​​can be obtained through distance analysis between the inner and outer surface point clouds. This approach ensures comprehensive data and lays the foundation for subsequent processing.

[0064] For example, when standardizing a preliminary dimensional data set, a pre-set calibration benchmark can be used. For example, the extracted OD data can be compared and adjusted against the industry-standard OD tolerance. For example, if the original OD data for a pipe section is 100.2 mm and the standard value is 100.0 mm, a calibration algorithm can be used to adjust the deviation to an acceptable range. This approach helps standardize the data format and facilitates subsequent analysis.

[0065] For example, a preset threshold of 0.3 mm can be set for calculating and flagging diameter deviations. If the outer diameter deviation of a pipe section reaches 0.4 mm, the system will automatically flag it as abnormal data. This flagging mechanism quickly identifies potential problem areas and provides clear guidance for subsequent processing.

[0066] For example, when using support vector machine algorithms for deviation analysis, the model can be trained using historical data to identify whether the deviations exhibit systematic characteristics. For example, consistently high outer diameter deviations across multiple pipe sections may indicate an equipment calibration issue rather than random errors. This analysis method can help distinguish different sources of deviation and improve the accuracy of problem location.

[0067] For example, systematic errors can be corrected by adjusting scanning device parameters or recalibrating the measurement benchmark. If the deviation is found to be due to device offset, the outer diameter deviation can be reduced from 0.4 mm to within 0.1 mm by correcting the device angle or updating the calibration data. This correction effectively improves data reliability.

[0068] For example, when comprehensively evaluating pipeline dimensions, multiple indicators, such as outer diameter and wall thickness, can be combined for a comprehensive assessment. If a section of pipeline has an outer diameter deviation of 0.2 mm and a wall thickness deviation of 0.1 mm, while each individual indicator may pass, the overall assessment may still fall short and be recorded as unqualified. This multi-dimensional assessment approach provides a more comprehensive picture of pipeline quality.

[0069] For example, when generating and archiving classified data, unqualified data can be categorized and stored by deviation type, such as abnormal outer diameter and abnormal wall thickness, filed separately, along with detailed deviation values ​​and detection times. This categorized archiving facilitates subsequent traceability and improvement, providing data support for pipeline quality management.

[0070] Step S107, 7: Obtain the point cloud data of the flange part in the pipeline geometric model, calculate the flange surface angle distribution through the normal vector analysis algorithm, and determine the preliminary value of the flange torsion angle deviation.

[0071] By extracting geometric data from the pipeline model, a point cloud dataset of the flange is obtained, initially forming the basic information for subsequent analysis. Using point cloud analysis technology, the normal vector value of each point in the obtained flange point cloud dataset is calculated to obtain a set of local directional features of the flange surface. Based on the calculated set of normal vector values, the angular distribution characteristics of the flange surface are analyzed to determine the angular variation patterns of each surface area. If there are significant areas of non-uniform distribution in the angular distribution characteristics, the difference in normal vector values ​​of adjacent areas is compared to determine whether there are preliminary signs of torsion angle deviation. After obtaining the judgment results, the specific torsion angle deviation values ​​are calculated within a preset threshold range for areas with signs of deviation to obtain preliminary deviation assessment results. By comparing the preliminary deviation assessment results with the overall geometric data of the pipeline model, it is determined whether the deviation affects the overall structural characteristics and form the final deviation analysis conclusion.

[0072] For example, when extracting geometric data from a pipe model to obtain a point cloud dataset of a flange section.

[0073] Understandably, this process primarily aims to build a high-precision 3D data foundation for subsequent analysis. For example, for a flange connection in a piping system, high-precision scanning equipment is used to acquire point cloud data, initially forming a collection of millions of points representing the 3D coordinates of the flange surface. This data acquisition method provides reliable raw information for subsequent feature analysis, which is particularly important when the tightness and stability of the flange connection are of particular concern.

[0074] For example, when using point cloud analysis to calculate the normal vector value for each point, the local surface orientation of each point can be determined by analyzing the spatial relationship between each point in the point cloud and its neighboring points. For example, in the point cloud data of a flange surface, points in a certain area show a significant deviation in the normal vector direction. This may indicate localized surface deformation or machining defects. Such analysis helps quickly locate potential problem areas and provides a basis for subsequent deviation assessment.

[0075] For example, when analyzing the angular distribution characteristics of a flange surface for a set of normal vector values.

[0076] Specifically, statistically analyzing the angle variations between normal vectors can be used to determine whether a surface is smooth or contains abnormal areas. For example, if the angle variation between normal vectors in one area of ​​a flange exceeds a preset 10-degree range, while remaining within 2 degrees in other areas, this could indicate uneven machining or distorted installation. This analysis can initially identify patterns in angle variation, laying the foundation for further deviation assessment.

[0077] For example, when identifying initial signs of torsional deviation, one can compare the differences in normal vector values ​​between adjacent areas. For example, in the aforementioned abnormal area, the normal vector directions of adjacent points differ by 15 degrees, while in the normal area, the difference is only 1 degree. This significant difference may indicate the presence of torsional deviation. This method can help technicians quickly focus on the problem area, avoiding the time-consuming process of a comprehensive inspection.

[0078] For example, when calculating torsion angle deviation using a preset threshold range, a deviation threshold of 5 degrees might be set. Suppose the calculation results show a torsion angle deviation of 7 degrees in a certain area, exceeding the threshold range. This indicates that this area requires further attention. This quantitative analysis can help identify issues and facilitate subsequent processing.

[0079] For example, when comparing the preliminary deviation assessment results with the overall geometric data of the pipeline model, it is assumed that a 7-degree torsion angle deviation affects the overall alignment of the flange connection, potentially affecting the sealing performance of the pipeline system. This comparative analysis can help determine whether the deviation will have a substantial impact on the overall structure, thus providing direction for subsequent improvements. Through the above multi-faceted analysis and examples, the close connection and logical progression of each link are demonstrated, ensuring the integrity from data extraction to final conclusions, while also providing strong support for pipeline system quality control.

[0080] Step S108, 8 If the preliminary value of the flange torsion angle deviation exceeds the preset threshold value of 5 degrees, the flange surface angle is optimized and fitted by the least square method to determine the accurate torsion angle deviation data.

[0081] By collecting the initial detection data of the flange torsion angle, the initial deviation value is obtained and compared with the preset threshold to determine whether it exceeds the range. If the initial deviation value exceeds the preset threshold, the collected data is preliminarily cleaned to remove abnormal points and obtain the processed initial data set. Based on the processed initial data set, the least squares method is used to optimize the fitting of the flange surface angle and calculate the adjusted angle analysis results. Through the angle analysis results, combined with the deviation calculation logic, the precise deviation value of the flange torsion angle is determined to form usable deviation data. After obtaining the precise deviation value, the torsion angle data is verified twice to determine whether it meets the preset reasonable range and obtain the verified torsion angle data. If there is still a deviation in the verified torsion angle data, the data set is iteratively adjusted through the data optimization process to determine the final optimization result. Based on the final optimization result, the precise data record of the flange torsion angle is generated, completing the entire deviation correction process.

[0082] For example, during the initial flange torsion angle test data collection process, a high-precision sensor can scan the flange at the pipe joint to obtain the initial deviation value. Suppose, during a particular test, the initial deviation value shows 3.5 degrees, while the preset threshold is 2.0 degrees, clearly exceeding the range. In this case, preliminary data cleaning is required to remove outliers caused by environmental interference or equipment jitter. For example, points that significantly deviate from the average value should be removed to obtain a more stable initial data set.

[0083] For example, when using the least squares method to optimize the flange surface angle fitting based on a cleaned data set, the surface angle data can be segmented and the angle values ​​of each small area can be smoothly adjusted. Assuming the angle values ​​in a certain area range from 2.8 to 3.2 degrees, the fitting results in a stable value close to 3.0 degrees, which serves as the adjusted angle analysis result. This method can effectively reduce the impact of local noise on the overall analysis.

[0084] For example, when determining the precise deviation of the flange torsion angle, the deviation calculation logic can be combined to compare the fitted angle result with the standard design value. If the design value is 0 degrees and the actual fitted value is 3.0 degrees, the torsion angle deviation can be preliminarily determined to be 3.0 degrees. This process ensures data representativeness to avoid distorted deviation values ​​due to insufficient sampling points. Such precise deviation values ​​provide a reliable basis for subsequent corrections.

[0085] For example, during secondary verification of torsion angle data, multiple test results can be compared to determine if they fall within a predefined acceptable range, such as whether the deviation remains stable between 2.5 and 3.5 degrees. If a verification result shows a sudden jump to 5.0 degrees, re-examination of the data collection process is necessary. This verification method helps improve data credibility and ensures the stability of analysis results.

[0086] For example, if deviations still exist in the verified torsion angle data, iterative adjustments can be made through the data optimization process. For example, if, during a particular iteration, the deviation values ​​in a certain area are consistently high, the sampling density in that area can be increased from 10 points per square centimeter to 20 points to obtain a more comprehensive data distribution, and the final results can be adjusted accordingly. This iterative approach can gradually approach the true deviation value and improve correction accuracy.

[0087] For example, when generating precise data records for flange torsion angles, the final optimization results can be stored in a table format, documenting the deviation value for each area, comparison data before and after adjustment, and verification status. For example, if the final deviation value for a flange stabilizes at 2.9 degrees and passes reasonable range verification, this can be used as the final record. This recording method facilitates subsequent traceability and analysis, providing data support for the long-term maintenance of the pipeline system.

[0088] For example, during the entire deviation correction process, flange torsion angle processing can be combined with historical data for trend analysis. For example, if the deviation of a flange gradually increases from 2.0 degrees to 2.9 degrees over multiple inspections, it can be inferred that there is a potential structural fatigue issue, requiring proactive maintenance measures. This analysis method can help identify hidden risks and improve overall system reliability.

[0089] Step S109, 9: Based on the accurate torsion angle deviation data, a finite element analysis algorithm is used to simulate the stress distribution at the flange connection, determine the stress concentration area and the maximum stress value, and obtain the flange connection stress data.

[0090] Deviation data related to flange connections is obtained from a pre-established torsional deviation database. Data cleaning methods are used to remove outliers, resulting in a processed deviation dataset. Based on this processed deviation dataset, finite element analysis (FEM) is used to simulate the stress distribution of the flange connection and generate an initial stress distribution map. Based on this initial stress distribution map, meshing techniques are used to refine the connection area, identify specific areas of stress concentration, and determine the distribution characteristics of these areas. Based on the distribution characteristics of these stress concentration areas, stress values ​​at key nodes are extracted. If the stress values ​​at these key nodes exceed a preset threshold, they are marked as high-risk areas, resulting in a high-risk area identification result. Based on the identified high-risk areas and combined with the FEM simulation data, the maximum stress value is accurately calculated to determine the specific location and range of the maximum stress value. Based on the specific location and range of the maximum stress value, comprehensive stress distribution data for the flange connection is generated. This distribution data is then graphically processed using visualization tools to generate a final stress distribution visualization map. Based on the final stress distribution visualization map, outliers are identified from the distribution data. If an outlier coincides with a high-risk area, it is prioritized and its risk level determined.

[0091] For example, when processing deviation data related to flange connections, a pre-established torsional angle deviation database can be used to retrieve historical deviation information. Imagine that during a particular inspection, the initial deviation data for a flange connection contained multiple angle values, some of which exhibited abnormally high values ​​due to measurement error. Data cleaning methods can be used to remove these outliers, such as marking data points exceeding two standard deviations from the mean as outliers, resulting in a more stable deviation dataset. This cleaning process enhances the accuracy of subsequent analysis.

[0092] For example, when using finite element analysis to simulate the stress distribution of a flange connection, the flange connection area can be divided into multiple small units, assuming each unit is 2 mm x 2 mm in size. By simulating loading conditions, such as applying a tensile force of 500 Newtons, an initial stress distribution map is generated. This method can intuitively reflect the stress conditions in different areas of the flange connection, providing a basis for subsequent optimization.

[0093] For example, to refine the initial stress distribution map, meshing techniques can be used to further divide the connection edge areas, where stress may be concentrated, into smaller mesh elements, such as 1 mm x 1 mm, to identify specific areas of stress concentration. For example, if, in an analysis, the stress values ​​near a bolt hole are significantly higher than those in other areas, this area can be identified as a stress concentration area. This refinement helps pinpoint the problem area.

[0094] For example, when extracting stress data for key nodes, we can focus on core points within the stress concentration area. For example, if the stress value of a node is 300 MPa and the preset threshold is 250 MPa, the node will be marked as a high-risk area. This method can quickly screen out areas that require special attention and improve analysis efficiency.

[0095] For example, when accurately calculating maximum stress, finite element analysis simulation data can be combined to identify the highest stress point within a high-risk area, assuming it is 320 MPa and located 3 mm from the edge of the bolt hole. This precise calculation can provide a specific reference for subsequent design improvements.

[0096] For example, when generating and visualizing comprehensive stress distribution data, graphical tools can be used to represent stress values ​​using color depth, assuming red represents high-stress areas and green represents low-stress areas, to generate a final visualization of the stress distribution. This intuitive presentation allows technicians to quickly understand the distribution characteristics.

[0097] For example, when extracting abnormal points and determining their risk levels, if an abnormal point is found to coincide with a high-risk area, such as a stress abnormality point at the edge of a bolt hole with a value of 310 MPa, it will be ranked highest. This prioritization helps focus resources on addressing the most critical problem areas and ensure the stability of the flange connection.

[0098] Step S1010, 10: Based on the flange connection stress data, an adaptive adjustment algorithm is used to optimize the installation angle between the pipeline center axis and the flange, and the point cloud registration accuracy is controlled at the millimeter level to obtain the optimized installation parameters of the pipeline and the flange.

[0099] The stress data of the flange connection and the initial position data of the pipe center are obtained. A sensor acquisition system is used to scan the actual connection between the pipe and flange, obtaining preliminary stress distribution and geometric position information. Based on this preliminary stress distribution information, point cloud registration technology is used to perform a 3D reconstruction of the relative position between the pipe center and the flange installation, determining the initial registration result of the point cloud data. Based on this initial registration result, an adaptive adjustment algorithm is used to optimize the axis of the pipe center. If the axis deviation exceeds a preset threshold, the axis position is fine-tuned to obtain the adjusted axis parameters. The adjusted axis parameters are combined with the flange installation angle data to calculate the relative angular deviation between the two. If the angular deviation exceeds a preset range, the flange installation angle is iteratively adjusted to obtain the optimized angle value. Based on the optimized angle value and axis parameters, the point cloud registration result is recalibrated, and an iterative closest point algorithm is used to improve registration accuracy, ensuring that the final registration accuracy reaches the millimeter level. Using the final registration accuracy data, the installation parameters of the pipe and flange are calculated, and an optimized installation plan is generated, resulting in an adapted installation parameter set. For the generated installation parameter set, the stress data is simulated and verified. If the simulation results show that the stress distribution is uneven, return to the axis optimization step for readjustment to obtain the final optimization result.

[0100] For example, when acquiring stress data for flange connections and initial position data for the center of the pipe, a high-precision sensor acquisition system can be used to perform real-time scanning of the pipe-flange connection. Sensors can be placed at multiple key points on the flange, such as near the connecting bolts and along the flange edge. The collected stress data may reveal elevated stress values ​​in some areas, perhaps reaching 200 MPa, while others are only 50 MPa. This discrepancy suggests possible installation deviation, requiring further analysis.

[0101] Specifically, the application of point cloud registration technology can help construct a 3D model of the pipe and flange based on preliminary stress distribution information. Assuming the scanned point cloud data contains geometric information about the pipe's central axis and the flange's mounting surface, the registration algorithm can reveal that the angle between the pipe axis and the flange surface deviates by 2 degrees, exceeding the preset threshold of 1 degree. This deviation can lead to stress concentration, necessitating subsequent adjustments.

[0102] For example, when using an adaptive adjustment algorithm to optimize the pipe centerline, if the detected axis deviation is 3 mm, exceeding the preset 1 mm threshold, it can be corrected by fine-tuning the pipe installation position or adjusting the flange angle. After adjustment, the axis deviation may be reduced to 0.5 mm, meeting the requirements. This optimization helps ensure stability during subsequent installations.

[0103] Specifically, when calculating the relative deviation between the flange installation angle and the pipe axis, if the angle deviation is found to be 1.5 degrees, exceeding the preset 1-degree range, the flange installation angle can be iteratively adjusted by fine-tuning it by 0.2 degrees each time, ultimately controlling the deviation within 0.8 degrees. This approach can effectively reduce stress anomalies caused by angle issues.

[0104] For example, the iterative closest point algorithm can further improve accuracy by performing secondary calibration on point cloud registration results. Assuming an initial registration error of 2 mm, after multiple iterations, the error can be reduced to 0.3 mm, achieving millimeter-level accuracy. This provides a reliable basis for determining subsequent installation parameters.

[0105] Specifically, when generating the optimized installation plan, the final alignment accuracy data is used to calculate the optimal installation parameters for the pipe and flange, such as a bolt tightening torque of 100 Nm and a flange angle deviation within 0.5 degrees. These parameters ensure the stability of the connection.

[0106] For example, when performing stress simulations to verify a set of installation parameters, if an uneven stress distribution is detected in a certain area—for example, a stress value as high as 250 MPa on one side and only 60 MPa on the other—then the axis optimization step must be repeated. Through multiple iterations, the stress distribution difference can ultimately be controlled to within 30 MPa. This repeated verification approach can significantly improve the reliability of the connection structure and reduce potential risks.

[0107] Step S1011, 11 performs virtual assembly simulation on the pipeline system by optimizing the installation parameters, uses a geometric constraint algorithm to verify the connection stability and multi-angle measurement accuracy, determines whether the pipeline diameter deviation and the flange torsion angle deviation meet the preset threshold, and obtains the final calibration result.

[0108] By obtaining optimized installation parameters from the pipeline system database, the virtual assembly process is initialized, a three-dimensional simulation environment for the pipeline system is constructed, and preliminary assembly model data is obtained. Based on this preliminary assembly model data, a geometric constraint algorithm is used to restrictively calculate the relative positions of the pipeline system and the connecting components, determining the geometric matching of the connection. Based on this geometric matching, multi-angle measurement data is obtained to simulate and verify the stability of the pipeline system connection. The force distribution at each key point is determined to ensure balance, and a force distribution analysis result is obtained. If the force distribution analysis results indicate anomalies in certain areas, the diameter deviation data in the installation parameters is processed and compared with a preset threshold to determine the adjustment direction for the diameter deviation. Based on the adjustment direction for the diameter deviation, relevant data on the torsion angle deviation is obtained. Through information processing, the torsion angle deviation is compared with a preset threshold to determine whether the torsion angle parameters need to be corrected, resulting in a torsion angle correction solution. Using the torsion angle correction solution and the adjusted diameter deviation data, a secondary simulation verification of the virtual assembly model of the pipeline system is performed to obtain the final calibration result. Based on the final calibration results, the data of the connection stability of the pipeline system is archived, and a set of calibrated assembly parameters is generated to determine the reference benchmark for subsequent simulation verification.

[0109] For example, when performing a preliminary simulation of a piping system using virtual assembly technology, a digital model of the piping system can be created using 3D modeling software. This model includes basic data such as the pipe length, diameter, and connection point geometry. Assuming the pipe system is designed to be 5.0 meters long and 0.2 meters in diameter, the connection points in the initial model are distributed at standard spacing. This approach provides a visual representation of the overall piping system layout, facilitating subsequent analysis of potential installation issues.

[0110] For example, when applying geometric constraint algorithms to analyze connection stability, force distribution simulation can be used to determine whether the connection points are uniform. For example, suppose a connection point experiences a force of 200 Newtons in the initial layout, while an adjacent connection point experiences only 100 Newtons, significantly exceeding the expected uniformity. In this case, the spacing between the connection points can be adjusted or support structures can be added to ensure a more balanced force distribution. This method helps identify design weaknesses early on.

[0111] For example, when performing multi-angle measurement, laser scanning equipment can be used to obtain diameter and torsion angle data from different angles. Suppose the measurement reveals a diameter deviation of 0.5 mm and a torsion angle deviation of 1.2 degrees for a particular pipe section, both exceeding the preset thresholds of 0.3 mm and 0.8 degrees. Recording these deviations provides an accurate basis for subsequent corrections. This measurement method provides a comprehensive picture of the geometric condition of the pipe system.

[0112] For example, to correct deviation data, relevant parameters can be adjusted in the virtual assembly model. For example, the angle or position of a pipe connection can be fine-tuned to reduce torsional deviation. Assume that after adjustment, the torsional deviation is reduced to 0.6 degrees, meeting the threshold requirement. This correction method can effectively improve model accuracy and provide a reliable reference for actual installation.

[0113] For example, when combining a geometric constraint algorithm with secondary verification, the stress conditions of the entire piping system can be re-simulated to check whether the adjusted connection points meet the design requirements. If the secondary verification reveals that a certain connection point still experiences slight stress imbalances, the support design can be further optimized to ensure overall stability. This verification method can improve system reliability.

[0114] For example, after final calibration of installation parameters and the generation of a dataset, all optimized data can be consolidated into a complete parameter set, including information such as pipe position, angle, and connection point forces. This integration allows for direct reference during subsequent actual operations, ensuring an efficient and smooth installation process.

[0115] For example, when outputting the final virtual assembly model, 3D visualization technology can be used to present the optimized piping system layout, visually demonstrating whether each component meets the required accuracy and stability. Assuming the final model shows that all deviations are within the preset range, this output method can provide clear guidance to the engineering team and improve installation efficiency.

[0116] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An intelligent detection system for assembled pipeline components, characterized in that: The invention comprises a detection device and a control cabinet, wherein the detection device comprises a bottom bracket (1), a detection mechanism (2) is installed on one side of the top of the bottom bracket (1), two fixing mechanisms (3) are fixedly connected to the other side of the top of the bottom bracket (1), and the two fixing mechanisms (3) are distributed in a mirror image; the fixing mechanism (3) comprises a support frame (4), the support frame (4) is fixedly connected to the bottom bracket (1), a slideway (33) is provided at the top of the support frame (4), a slider (34) is slidably connected to the inner surface of the slideway (33), a screw rod (31) is threadedly sleeved on the inner surface of the slider (34), one end of the screw rod (31) passes through the slideway (33) and extends to the outside of the slideway (33), and two fixing frames (35) are provided at the top of the slideway (33), one of the fixing frames (35) is fixedly connected to the slider (34), and the other fixing frame (35) is fixedly connected to the slider (34). The frame (35) is fixedly connected to the slideway (33), and the inner surfaces of the two fixed frames (35) are rotatably connected to the transmission rollers (36), one end of one of the transmission rollers (36) passes through the fixed frame (35) and extends to the outside of the fixed frame (35), and the outer surface of the transmission roller (36) located outside the fixed frame (35) is fixedly sleeved with a worm gear (47), and one side of the slideway (33) is fixedly connected to a power motor (46), and the output end of the power motor (46) is fixedly connected to a worm (45), and the worm (45) and the worm gear (47) are meshed with each other; the detection mechanism (2) includes a detection platform (22), the top of the detection platform (22) is fixedly connected to an electric slide rail (21), the top of the electric slide rail (21) is slidably connected to a slide (23), and the top of the slide (23) is equipped with a laser (25) and a detection camera (26); The control cabinet comprises a cabinet body, a display arranged on the top of the cabinet body, and an industrial computer and a control board arranged in the cabinet body. The detection camera (26) and the display are both connected to the industrial computer. A control circuit module is integrated on the control board, and the control circuit module is connected to and communicates with the industrial computer.

2. The intelligent detection system for assembled pipeline components according to claim 1, characterized in that: A fixed rod (44) is fixedly connected to one side of the top of the slideway (33), and a movable rod (38) is slidably connected to the inner surface of the fixed rod (44). The fixed rod (44) is hollow, and both ends of the fixed rod (44) are provided with a through-type adjustment groove (441). A limiting groove (444) is provided on the inner surface of the movable rod (38). Two limiting plates (445) are slidably connected to the inner surface of the limiting groove (444). The corresponding ends of the two limiting plates (445) are fixedly connected to the return spring (443). The opposite ends of the two limiting plates (445) are fixedly connected to an adjustment rod (442) used in conjunction with the adjustment groove (441). The outer surfaces of the two adjustment rods (442) are slidably connected to the inner surface of the movable rod (38).

3. The intelligent detection system for assembled pipeline components according to claim 1, characterized in that: One end of the fixed rod (44) is rotatably connected to the rotating block 1 (43), the top end of the adjusting rod (442) is rotatably connected to the pressure rod (39), one end of the pressure rod (39) is rotatably connected to the rotating block 2 (41), the inner surface of the rotating block 2 (41) is threadedly sleeved with the screw rod 2 (42), and the other end of the screw rod 2 (42) is rotatably connected to the rotating block 1 (43).

4. The intelligent detection system for assembled pipeline components according to claim 1, characterized in that: One end of the screw rod 1 (31) passes through the slideway (33) and is fixedly connected to the rotating handle 1 (32), and one end of the screw rod 2 (42) passes through the rotating block 2 (41) and is fixedly connected to the rotating handle 2 (40).

5. The intelligent detection system for assembled pipeline components according to claim 1, characterized in that: One end of the pressure rod (39) away from the second rotating block (41) is rotatably connected to two rollers (37).

6. The intelligent detection system for assembled pipeline components according to claim 1, characterized in that: Both ends of the electric slide rail (21) are fixedly connected to limit blocks (24), and the two limit blocks (24) are fixedly connected to the detection platform (22).

7. A detection method for the intelligent detection system for assembled pipeline components according to claim 1, characterized in that: The method comprises the following steps: Step 1: Turn on the power, start the industrial computer, and set the detection parameters; Step 2: Perform the inspection operation on the operation interface of the display. The industrial computer analyzes the CAD drawing of the pipeline component to be inspected and displays the graphics of the pipeline component on the display. Step 3: Use a crane to place the pipe component on the transmission rollers (36) of the two fixing mechanisms (3), and adjust it to a suitable position to accommodate the width of the pipe component; Step 4: Operate the components on the two fixing mechanisms (3) in sequence to compress the pipeline components; Step 5: Install the laser inclinometer on the flange of the pipeline. The laser inclinometer and the flange are attracted together by magnets. Step 6. Press the "Pipeline Attitude Adjustment" button on the display, adjust the pipeline position through the dialog box, and the power motor (46) drives the pipeline to rotate. When the laser inclinometer shows 90 degrees or the laser is horizontal, the pipeline attitude adjustment is completed. Press the OK button to return to the main interface; Step 7: Press the "Auto Measure" button on the display, and the synchronous belt drives the laser (25) and the detection camera (26) to scan the pipeline. After the industrial computer collects the image, it uses visual recognition and visual measurement technology to complete the pipeline detection; Step 8. After image analysis is completed, press the "Measurement Result" button to display and print the measurement results.

Images