Intelligent Visual Evaluation System for Pipeline Integrity

Abstract

The present invention discloses a pipeline integrity intelligent evaluation and visualization system, comprising a three-dimensional visualization module and a pipeline intelligent evaluation module. The three-dimensional visualization module includes a three-dimensional model visualization display submodule, a supporting environment submodule, an evaluation module data interface submodule, a pipeline information directory tree submodule, and a pipeline engineering data management submodule. The pipeline intelligent evaluation module includes a basic data management submodule, a detection management submodule, a defect management submodule, and an integrity intelligent evaluation and maintenance decision submodule. The integrity intelligent evaluation and maintenance decision submodule includes intelligent maintenance decisions for pipeline anomaly signals whose defect size and type cannot be quantified, as well as automatic evaluation and maintenance decisions for quantified defects. The three-dimensional model visualization display submodule performs three-dimensional dynamic interactive visualization. By interacting with a pipeline three-dimensional visualization platform, the present invention can realize intelligent evaluation of pipeline integrity, especially pipelines within refinery stations and gathering and transportation networks.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline integrity intelligent evaluation visualization, and in particular to a pipeline integrity intelligent evaluation visualization system. Background Art

[0002] Integrity assessment is a core component of pipeline integrity management. It can be used to determine the safety status of pipelines and provide guidance for pipeline repair and maintenance. Currently, pipelines in refineries, stations, and oil and gas field gathering and transportation networks, where internal inspection and excavation testing are not possible, face the following key challenges: 1) Refineries and stations are home to numerous feedstock lines, interconnecting lines, and power lines. Most power lines are located underground, and the variety and complexity of underground pipelines make unified management challenging. 2) Oil and gas field gathering and transportation networks, due to the large number of pipelines and the lack of data on older pipelines, often experience discrepancies between pipeline routes and actual site conditions, as well as the management team's understanding. This presents significant challenges for construction, inspection and maintenance, as well as for facility renovation, expansion, and equipment maintenance and upgrades. 3) Complex underground conditions lead to widespread corrosion and thinning of pipelines. Some older pipelines are overloaded and operating beyond their service life. Corrosion, pressure, and other factors can lead to pipe bursts, leaks, and cross-contamination, resulting in frequent leaks and significant safety hazards. 4) Internal inspections are impossible, hindering timely monitoring of pipeline safety status. 5) The existing method of evaluating test data through manual experience cannot guarantee the accuracy of the evaluation results. Summary of the Invention

[0003] In response to the problems existing in the prior art, an embodiment of the present invention proposes a pipeline integrity intelligent evaluation visualization system, specifically a pipeline integrity intelligent evaluation visualization system that realizes real-time dynamic interaction with a pipeline three-dimensional model.

[0004] Specifically, the embodiments of the present invention provide the following technical solutions:

[0005] The embodiment of the present invention provides a pipeline integrity intelligent evaluation visualization system, comprising: a three-dimensional visualization module and a pipeline intelligent evaluation module;

[0006] The three-dimensional visualization module includes a three-dimensional model visualization display submodule, a supporting environment submodule, an evaluation module data interface submodule, a pipeline information directory tree submodule, and a pipeline engineering data management submodule;

[0007] The pipeline intelligent evaluation module includes a basic data management submodule, a detection management submodule, a defect management submodule, and an integrity intelligent evaluation and maintenance decision submodule;

[0008] The evaluation module data interface submodule is used to provide a data interface for the basic data management submodule, the detection management submodule, the defect management submodule and the integrity intelligent evaluation and maintenance decision submodule;

[0009] The intelligent integrity evaluation and maintenance decision submodule includes intelligent maintenance decisions for pipeline abnormality signals whose defect size and type cannot be quantified, as well as automatic evaluation and maintenance decisions for quantified defects. For pipeline abnormality signals whose defect size and type cannot be quantified, the detection signal graph and detection signal strength value of the pipe section are first obtained from the detection management submodule, and an intelligent maintenance decision model based on pre-trained signal graphs or signal data based on deep learning is used to make intelligent maintenance decisions.

[0010] For the automatic evaluation and repair decision-making of quantitative defects, batch and single defects with size and type information are detected. Different evaluation methods are used for different defect types. After selecting an evaluation method for a certain type of defect, the built-in evaluation algorithm is called to perform batch evaluation, single evaluation and repair decision-making on defect points of the same type or a certain defect using the same evaluation method; the three-dimensional model visualization display submodule performs three-dimensional dynamic interactive visualization of the intelligent evaluation and repair decision results.

[0011] Furthermore, the three-dimensional visualization module is based on the principle of human-computer interaction. By conducting on-site surveys of refinery equipment and devices, ground markers, and ground and underground pipeline objects connected to the equipment and devices, the module determines the geographic information or relative position of the modeling object, establishes a model of the object in the three-dimensional model, and adds, checks, deletes, and modifies the established pipeline, device, and equipment-related engineering information through the three-dimensional model or directory tree. The pipeline information serves as the data basis for the integrity evaluation of the buried pipeline.

[0012] Furthermore, the basic data management submodule is used to obtain basic information of the pipeline where the pipeline is located from the three-dimensional visualization module, and is used for adding, checking, deleting and modifying the pipe diameter, wall thickness, pipe material, conveying medium and connection equipment information.

[0013] Furthermore, the detection management submodule includes three functions: detection pipeline management, detection project management and detection data management, which is aimed at the logical management of detection operations, detection signals and detection results;

[0014] Among them, the inspected pipeline management is used to manage the inspection pipeline and its related inspection projects. The inspected pipeline management has an inspection scope for the inspected pipeline. Based on the built-in pipeline number naming rules, the pipeline number of the newly added inspection pipeline is automatically generated based on the pipeline where the inspected pipeline is located. The information of the newly added inspected pipeline includes: inspection length, inspection location, and remarks. The information is maintained, modified, and viewed. At the same time, after a new inspection pipeline is created, the inspected pipeline is highlighted on the 3D model by the pipeline number. In addition, the inspection project and related information of the inspected pipeline can be added and modified under the added inspected pipeline. Check, delete, and modify. After a pipeline inspection, manually locate the inspection marker information on the pipeline 3D model and click on the pipeline to be inspected to determine the inspection range of the inspection pipeline. The system automatically generates the inspection pipeline number based on the pipeline where the inspection pipeline is located. At the same time, based on the inspection pipeline number, the inspection pipeline is highlighted and located in the 3D model. The information of the inspected pipeline is maintained and modified, including the entry and modification of the inspection length, inspection positioning point, design coefficient information, and the deletion of the inspected pipeline. Among them, after the inspected pipeline is deleted, the inspection project, defect point location, and defect data associated with it are also deleted.

[0015] Inspection project management: This module is used to establish an inspection project after conducting an inspection on a pipeline section. This includes creating a new inspection project under the newly inspected pipeline and managing the addition, query, deletion, modification, and maintenance of the inspection project and its related information, including the inspection project name, inspection method, inspection equipment, inspection time, and inspection unit. Deleting an inspection project also deletes the associated defect location and defect data. This module also allows users to view a list of all inspection projects under the inspected pipeline by inspected pipeline number.

[0016] Detection data management: used for the management of detection data for a certain detection project, with functions including detection signal diagram management, detection signal data management, and detection result data management; among them, detection signal diagram management is the management of the detection signal diagram of the pipeline; detection signal data management is in the form of a form, and the strength values of the detection signal along the detection mileage are imported into the system in the form of a form according to the detector sampling results; the detection result data is also managed in the form of importing the system into the form; it also includes detection signal, result template download, data import, and data addition, query, deletion, and modification.

[0017] Furthermore, the defect management submodule is used to manage defect data obtained after a pipeline inspection through expert signal analysis, excavation verification, or intelligent maintenance decision-making. Defect data can interact with the three-dimensional model to achieve the addition, query, deletion, and modification of defect information of the inspected pipeline. At the same time, defect information is managed separately according to different defect types. Among them, the corrosion rate of corrosion defects is directly calculated based on the defect depth and commissioning time through the built-in background algorithm.

[0018] New defects are handled differently:

[0019] For individual defects identified through excavation verification and expert analysis, the defect location is directly added to the 3D pipeline model by measuring with a ruler using markers on the inspected pipeline, and then the defect data is entered.

[0020] For defects obtained through intelligent maintenance decisions, the intelligent maintenance decision model outputs the mileage location and maintenance level of the pipeline abnormal area, automatically marks the abnormal interval on the corresponding pipe section on the 3D platform, and automatically synchronizes the maintenance level to the abnormal interval of the corresponding pipeline. At the same time, the defects and maintenance decision results are saved to the defect management submodule so that the maintenance decision results can be displayed when entering the intelligent evaluation interface again. Among them, the defects obtained through intelligent decision-making can be saved together with the defects obtained through expert analysis and excavation verification through the defect addition method field; the defect addition methods include: expert analysis, excavation verification or intelligent decision-making;

[0021] For batch defects, the batch defects obtained through intelligent decision-making are intelligently added and located in the same way. For other defects, batch defect data is imported through data import. The specific operations are as follows: download the defect information template, edit the defect information, and then import it into the system. After the batch defect data is successfully imported, the defect location is manually entered for each defect data, and the newly added defects are associated and displayed in the 3D model, or the defects to be added and located are selected, and the defect is added to the 3D pipeline through the Add button of the defect location, and the defect information is automatically updated to the 3D defect point.

[0022] Furthermore, there are three cases for defect deletion:

[0023] A. Delete the defect point location to retain the defect data. When you add the pipeline abnormal area mileage location with the same defect number again, the defect data will be automatically updated.

[0024] B. Delete the defect data but keep the 3D defect point location. After adding the defect point information at that location again, it will be automatically associated with the 3D defect point.

[0025] C. Delete the defect data and the mileage position of the abnormal area of the pipeline on the 3D model pipeline at the same time;

[0026] The defect management submodule realizes the highlight positioning of defect points in three dimensions through defect point numbering; on the defect management page, by selecting the year, device, pipeline number, and inspection project through the page drop-down list, all defect points and defect information under the inspection project can be viewed.

[0027] Furthermore, the integrity intelligent evaluation and maintenance decision submodule includes intelligent maintenance decisions for pipeline abnormality signals whose defect size and type cannot be quantified, as well as automatic evaluation and maintenance decisions for quantified defects;

[0028] For pipeline abnormal signals whose defect size and type cannot be quantified, the detection signal graph and detection signal strength value of the pipe section are first obtained from the detection management submodule. For the detection signal graph and detection signal strength value, the pre-trained intelligent maintenance decision model based on the detection signal graph and the intelligent maintenance decision model based on the detection signal strength value are respectively used to determine a signal graph or signal data value of a standard length based on the pipeline signal graph and signal strength data. If the signal graph or signal value length The signal graph and value will be scaled to the standard size; if the signal length Starting from the left end point of the signal array, create a Sliding window, each The signal is trimmed once until the right side of the sliding window reaches the boundary of the maximum value of the signal, and the signal diagrams or signal arrays; these signal diagrams or signal arrays are input into the intelligent maintenance decision model for prediction, and the mileage interval where the abnormal signal location is located and the maintenance level of the abnormal signal are output according to the prediction results;

[0029] If the signal includes ultrasonic guided waves and magnetic stress signals, the results of both are output simultaneously. The mileage interval of the abnormal signal is the union of the two, and the maintenance decision level is the result of the higher level. The abnormal signal mileage interval of the pipeline and the corresponding maintenance decision result are output; or the results of the two are fused according to the weight and then output;

[0030] Based on the output abnormal mileage interval, the background automatically marks the abnormal interval in red on the pipe section in the three-dimensional platform, and automatically synchronizes the maintenance decision result to the abnormal interval of the pipeline section. At the same time, the evaluation result is saved and can be displayed when entering the intelligent evaluation interface next time.

[0031] Furthermore, for the automatic evaluation and repair decision-making of quantitative defects, batch and single defects with size and type information are detected. Different defect types have different defect evaluation methods. After selecting the evaluation method for a certain type of defect, the system can call the built-in evaluation algorithm to perform batch evaluation, single evaluation and repair decision-making on the same type of defect points or a single defect using the same evaluation method;

[0032] After the system evaluation is successful, the evaluation results are saved, and the evaluation results will be displayed the next time you enter the evaluation page; the automatic evaluation of single defects is for defect points with known defect types and defect sizes. In the single evaluation module, select the defect points to be evaluated, select the evaluation method, and automatically evaluate the defect points. The defect evaluation results are saved for the evaluated defect points, and the evaluation results are displayed when you enter again. The evaluation method can be modified and re-evaluated; the automatic evaluation of batch defects is on the batch evaluation page. After selecting the device, pipeline number, inspection project, and defect type, for the defect points to be evaluated, select the evaluation method, enter the information required for evaluation, and then use the system's built-in evaluation algorithm to batch evaluate a certain type of defects. The evaluation method of the evaluated defect points can be modified and re-evaluated.

[0033] Furthermore, when excavation data is not available in the early stage, the intelligent maintenance decision model is trained and tested using expert analysis data in the defect management submodule;

[0034] After the degree of excavation data accumulation meets the preset conditions, the excavation defect data is obtained from the defect management submodule, and the intelligent maintenance decision model is optimized through data migration.

[0035] Furthermore, the training process of the intelligent maintenance decision model based on the detection signal graph includes:

[0036] Obtaining sample magnetic stress images from expert analysis data in the defect management submodule; wherein the sample magnetic stress images need to cover various defect pixel locations and various repair levels;

[0037] Obtaining the mileage location and maintenance level of the pipeline abnormal area in the sample magnetic stress image from the expert analysis data in the defect management submodule;

[0038] The model uses sample magnetic stress images with predetermined defect pixel locations and repair levels as input, and uses the mileage locations and repair levels of pipeline abnormality areas in the sample magnetic stress images as output. The neural network model is trained and tested based on an intelligent learning algorithm to obtain an intelligent maintenance decision model.

[0039] The training process of the intelligent maintenance decision model based on the detection signal strength value includes:

[0040] Obtaining sample magnetic stress detection signal strength values from expert analysis data in the defect management submodule; wherein the sample magnetic stress detection signal strength values need to cover various defective pixel locations and various repair levels;

[0041] Obtaining the pipeline abnormal area location and maintenance level corresponding to the sample magnetic stress detection signal strength value from the expert analysis data in the defect management submodule; wherein the data is divided into a training set and a test set;

[0042] The sample magnetic stress detection signal strength values with predetermined defective pixel locations and maintenance levels are used as the model input, and the mileage locations and maintenance levels of pipeline abnormal areas corresponding to the sample magnetic stress detection signal strength values are used as the model output. The neural network model is trained and tested based on an intelligent learning algorithm to obtain an intelligent maintenance decision-making model.

[0043] As can be seen from the above technical solution, the pipeline integrity intelligent evaluation visualization system provided by the embodiment of the present invention can realize intelligent evaluation of pipeline integrity, especially pipelines within refineries, stations and gathering and transportation networks, by interacting with the pipeline three-dimensional visualization platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 It is a structural diagram of a pipeline integrity intelligent evaluation system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following embodiments of the present invention are further described in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0047] The pipeline integrity intelligent evaluation and visualization system provided by the embodiment of the present invention interacts with the buried pipeline three-dimensional visualization platform to realize intelligent analysis and evaluation of pipe body defects and maintenance decisions for buried pipelines, especially underground pipelines within refineries, gathering and transportation networks, and stations.

[0048] Figure 1 The schematic diagram of the structure of the pipeline integrity intelligent evaluation visualization system provided by the embodiment of the present invention is shown. Figure 1 As shown, the pipeline integrity intelligent evaluation visualization system provided by the embodiment of the present invention includes: a three-dimensional visualization module and a pipeline intelligent evaluation module;

[0049] The three-dimensional visualization module includes a three-dimensional model visualization display submodule, a supporting environment submodule, an evaluation module data interface submodule, a pipeline information directory tree submodule, and a pipeline engineering data management submodule;

[0050] The pipeline intelligent evaluation module includes a basic data management submodule, a detection management submodule, a defect management submodule, and an integrity intelligent evaluation and maintenance decision submodule;

[0051] The evaluation module data interface submodule is used to provide a data interface for the basic data management submodule, the detection management submodule, the defect management submodule and the integrity intelligent evaluation and maintenance decision submodule;

[0052] The intelligent integrity evaluation and maintenance decision submodule includes intelligent maintenance decisions for pipeline abnormality signals whose defect size and type cannot be quantified, as well as automatic evaluation and maintenance decisions for quantified defects. For pipeline abnormality signals whose defect size and type cannot be quantified, the detection signal graph and detection signal strength value of the pipe section are first obtained from the detection management submodule, and an intelligent maintenance decision model based on pre-trained signal graphs or signal data based on deep learning is used to make intelligent maintenance decisions.

[0053] For the automatic evaluation and repair decision-making of quantitative defects, batch and single defects with size and type information are detected. Different evaluation methods are used for different defect types. After selecting an evaluation method for a certain type of defect, the built-in evaluation algorithm is called to perform batch evaluation, single evaluation and repair decision-making on defect points of the same type or a certain defect using the same evaluation method; the three-dimensional model visualization display submodule performs three-dimensional dynamic interactive visualization of the intelligent evaluation and repair decision results.

[0054] As can be seen from the above technical solution, the pipeline integrity intelligent evaluation visualization system provided in this embodiment can realize intelligent evaluation of pipeline integrity, especially pipelines within refineries, stations, and gathering and transportation networks, by interacting with the pipeline three-dimensional visualization platform.

[0055] In this embodiment, the intelligent pipeline integrity assessment and visualization system includes two components: 3D visualization and intelligent pipeline integrity assessment. The 3D visualization function primarily includes simplified visualization of pipeline connection equipment, devices, ground markers, and 3D models of all connected pipelines, enabling pipeline information management. The intelligent pipeline assessment function primarily includes basic pipeline data management, inspection data management, defect management, and intelligent pipeline integrity assessment and maintenance decision-making through dynamic data interaction with the 3D model.

[0056] The 3D visualization module is introduced as follows:

[0057] The 3D visualization module includes 3D model visualization, a secure and reliable support environment, and an application interface for the intelligent integrity assessment module (including pipeline basic data management, inspection management, defect management, and intelligent assessment). Following the principles of human-computer interaction, the module conducts on-site surveys of refinery equipment and devices, above-ground landmarks (such as roads and manhole covers), and all above-ground and underground pipelines connected to the equipment and devices. The geographic information or relative positions of the modeled objects are determined and then models of these objects are created within the 3D model. Furthermore, engineering information related to the established pipelines, devices, and equipment can be added, checked, deleted, and modified through the 3D model or directory tree. Pipeline information serves as the data foundation for buried pipeline integrity assessment.

[0058] The pipeline intelligent evaluation module is introduced as follows:

[0059] The pipeline intelligent evaluation module mainly includes: basic data management, detection management, defect management, integrity intelligent evaluation and maintenance decision-making sub-modules that realize dynamic data interaction with three-dimensional models.

[0060] The basic data management submodule includes the basic information of the pipeline where the pipeline is located obtained from the 3D visualization platform, including the addition, query, deletion and modification of information such as pipe diameter, wall thickness, pipe material, conveying medium, and connecting equipment.

[0061] The inspection management submodule is mainly divided into three functions: inspection pipeline management, inspection project management, and inspection data management. It mainly focuses on the logical management of inspection data such as inspection operations, inspection signals, and inspection results. ① Inspection pipeline management: It mainly manages the inspection pipeline and its related inspection projects. Inspection pipeline management mainly focuses on the inspection range of the inspection pipeline. Based on the built-in pipeline number naming rules, the pipeline number of the newly added inspection pipeline is automatically generated based on the pipeline where the inspected pipeline is located, so as to maintain, modify and view the information of the newly added inspection pipeline, such as: inspection length, inspection location (or inspection starting point and inspection end point markers), and remarks. At the same time, after the new inspection pipeline is created, the inspection pipeline is highlighted on the three-dimensional model by the pipeline number; in addition, the inspection project and related information of the inspected pipeline can be added, checked, deleted, and modified under the added inspection pipeline. After a pipeline inspection, manual positioning is performed on the pipeline model using information such as inspection markers. The inspection range of the pipeline to be inspected is determined by clicking on the pipeline to be inspected. The system automatically generates an inspection pipeline number based on the pipeline where the inspection pipeline is located. The inspection pipeline number is then used to highlight the inspected pipeline in the 3D model. Information about the inspected pipeline can be maintained and modified, including the entry and modification of information such as inspection length, inspection location, and design coefficients, as well as deletion of the inspected pipeline. Deleting an inspected pipeline deletes its associated inspection project, defect location, and defect data. ② Inspection Project Management: After an inspection is conducted on a pipeline section, an inspection project is created, which involves creating a new inspection project under the newly inspected pipeline. Inspection projects and their related information (project name, inspection method, inspection equipment, inspection time, inspection unit, etc.) can be added, checked, deleted, and modified. Deleting an inspection project deletes its associated defect location and defect data. This module also allows users to view a list of all inspection projects associated with the inspected pipeline by inspected pipeline number. ③ Inspection Data Management: This module manages inspection data for a specific inspection project. Functions include test signal graph management, test signal data management, and test result data management. Test signal graph management primarily manages pipeline test signal graphs. Test signal data management involves importing test signal strength values along the test route into the system in a spreadsheet format based on detector sampling results. Test result data is also managed through this form import. Furthermore, functions include downloading test signal and result templates, importing data, and adding, querying, deleting, and modifying data.

[0062] The defect management submodule primarily manages defect data (including defect number, defect type, defect size, defect location, defect corrosion rate, defect location, and evaluation results) obtained after a pipeline inspection through expert signal analysis, excavation verification, or intelligent maintenance decisions. Defect data can interact with the 3D model, enabling the addition, review, deletion, and modification of inspected pipeline defect information. Furthermore, defect information can be managed separately based on different defect types, such as corrosion defects, manufacturing defects, deformation defects, and weld anomalies. The corrosion rate of corrosion defects can be directly calculated based on defect depth, commissioning time, and other factors using a built-in backend algorithm. There are two ways to add new defects: ① For single defects obtained through excavation verification and expert analysis, under the premise of known defect location, the defect point location can be directly added on the 3D pipeline model by measuring the markers on the inspected pipeline (pipeline starting point, road, manhole cover, etc.), and then the defect data can be entered; ② For defects obtained by intelligent maintenance decision-making, the defect location and maintenance level are output through the intelligent maintenance decision-making model, and the abnormal interval is automatically marked on the corresponding pipe section on the 3D platform, and the maintenance level is automatically synchronized to the abnormal interval of the corresponding pipeline. At the same time, the defects and maintenance decision results are saved in the defect management submodule, so that the maintenance decision results can be displayed when entering the intelligent evaluation interface again. Among them, the defects obtained by intelligent decision-making can be Defects identified through expert analysis and excavation verification can be saved simultaneously using the "Defect Add Method" field (Expert Analysis, Excavation Verification, Intelligent Decision). For batch defects, the batch defects identified through intelligent decision-making can be added and located in the same manner as in step 2. Other defects can be imported through data import. The specific steps are: download the defect information template, edit the defect information, and then import it into the system. After the batch defect data is successfully imported, manually enter the defect location for each defect data item and associate the newly added defects with the 3D model. Alternatively, select the defect to be located and click the Add button to add the defect to the 3D pipeline. The defect information will automatically update the 3D defect point. Deleting defects can be done in three ways: ① Deleting the defect point location retains the defect data. Adding a defect location with the same defect number automatically updates the defect data. ② Deleting the defect data retains the 3D defect point location. Adding the defect point information at the same location automatically associates it with the 3D defect point. ③ Deleting the defect data and the defect location on the 3D pipeline model simultaneously. This module highlights the location of defect points in 3D by numbering them. On the Defect Management page, select Year, Device, Pipeline Number, or Inspection Project from the drop-down list to view all defects and defect information for that inspection project.

[0063] The integrity intelligent evaluation and maintenance decision module mainly includes intelligent maintenance decisions for pipeline abnormal signals whose defect size and type cannot be quantified, and automatic evaluation and maintenance decisions for quantified defects.

[0064] For pipeline abnormal signals whose defect size and type cannot be quantified, first obtain the detection signal graph and detection signal strength value of the pipe section from the "Detection Data Management". Use the pre-trained signal graph or signal data integrity intelligent maintenance decision model to determine a standard length of signal graph or signal data value based on the pipeline signal graph and signal strength data. If the signal graph or signal value length The signal graph and value will be scaled to the standard size; if the signal length Starting from the left end point of the signal array, create a Sliding window, each The signal is trimmed once until the right side of the sliding window reaches the boundary of the maximum value of the signal, and the (Round down) signal graphs or signal arrays. These signal graphs or signal arrays are input into the intelligent maintenance decision model for prediction. Based on the network prediction results, the mileage interval where the abnormal signal is located and the maintenance level of the abnormal signal (immediate repair, planned repair, and monitoring use) are output respectively. If the signal includes ultrasonic guided waves and magnetic stress signals, the results of both can be output simultaneously (the mileage interval of the abnormal signal is the union of the two, and the maintenance decision level is the higher level result, and the abnormal signal mileage interval of the pipeline and the corresponding maintenance decision result (defect type, maintenance level) are output). Alternatively, the results of the two can be fused according to the weights and output. Then, based on the output abnormal mileage interval, the background automatically marks the abnormal interval in red on the pipe section in the three-dimensional platform, and automatically synchronizes the maintenance decision result to the abnormal interval of the pipeline section. At the same time, the evaluation results are saved and can be displayed the next time you enter the intelligent evaluation interface.

[0065] In this embodiment, it should be noted that when the network model is applied, if the signal graph or signal value length The signal graph and value will be scaled to the standard size; if the signal length Starting from the left end point of the signal array, create a Sliding window, each The signal is trimmed once until the right side of the sliding window reaches the boundary of the maximum value of the signal, and the (Round down) signal graphs or signal arrays, thereby ensuring that the potential fault area can be completely detected by the network at least once and reducing the distortion of the data signal to improve network accuracy.

[0066] For automated evaluation and repair decisions of quantified defects, this module detects batches and single defects with information such as size and type. Different defect types require different evaluation methods. Users can select an evaluation method for a specific defect type. The system then calls upon its built-in evaluation algorithm to perform batch and single evaluations, as well as repair decisions, for the same type of defect or a single defect. After a successful evaluation, the system saves the results and displays them the next time you access the evaluation page. Single defect automatic evaluation targets defects with known defect type and size. In the single evaluation module, select the defect to be evaluated, choose an evaluation method, and automatically evaluate the defect. The evaluation results are saved for each evaluated defect and displayed the next time you access the evaluation page. You can select a new evaluation method. In the batch evaluation module, users select the device, pipeline number, inspection project, and defect type. After selecting the evaluation method and entering the required information, the system uses the built-in evaluation algorithm to evaluate a specific type of defect in batches. The evaluation method can be modified and re-evaluated for already evaluated defects.

[0067] When excavation data is not available in the early stage, the intelligent maintenance decision model can be trained and tested through expert analysis data in the defect data submodule.

[0068] In this embodiment, the training process of the intelligent maintenance decision model based on the detection signal graph includes:

[0069] Obtaining sample magnetic stress images from expert analysis data in the defect management submodule; wherein the sample magnetic stress images need to cover various defect pixel locations and various repair levels;

[0070] Obtaining the mileage location and maintenance level of the pipeline abnormal area in the sample magnetic stress image from the expert analysis data in the defect management submodule;

[0071] The model uses sample magnetic stress images with predetermined defect pixel locations and repair levels as input, and uses the mileage locations and repair levels of pipeline abnormality areas in the sample magnetic stress images as output. The neural network model is trained and tested based on an intelligent learning algorithm to obtain an intelligent maintenance decision model.

[0072] It can be seen that this embodiment makes an intelligent maintenance decision based on the magnetic induction signal image data, which is highly adaptable and not affected by human subjective factors, thereby effectively improving the accuracy of the analysis of the detection results, avoiding blind maintenance, greatly improving economy and reducing labor, and at the same time improving the safety of the pipeline.

[0073] Based on the content of the above embodiment, in this embodiment, after obtaining the sample magnetic stress image from the expert analysis data in the defect management submodule, manual and intelligent image processing are also included. Among them, the manual image processing process is:

[0074] The sample magnetic stress images are uniformly scaled, specifically including:

[0075] First, count the inspection mileage X of each inspected pipe section image max , average inspection mileage Pixel length L src and the average original pixel length Average width Round to the nearest integer;

[0076] To ensure that the mileage ratio of all images is unified, firstly the pixel length L of all images is src Adjust so that the adjusted sample image length Count the average pixel value of the image length L0 Round to an integer, and then scale the image:

[0077] Case 1: For detection mileage The sample magnetic stress image is scaled to the original pixel length L0 of the image. The original width W0 is scaled to The coordinates of the pixels in the scaled image are obtained by coordinate interpolation:

[0078]

[0079]

[0080] in, is the coordinate of the pixel after transformation, L0, W0 are the length and width of the sample image;

[0081] The coordinates (x, y) of the transformed pixel point belong to the pixel point coordinate set P:

[0082]

[0083] Among them, P ij =(x i ,y j );i=1,2,…,m,j=1,2,…,n;

[0084] When the pixel coordinates (x, y) are integers, the pixel value f(x, y) of the point after processing is the same as the pixel value of the corresponding point in the original image; when the pixel coordinates (x, y) are not integers, the pixel value f(x, y) of the point after processing is obtained by bilinear interpolation with the pixels of the four points surrounding the point in the original image and then rounding.

[0085] Case 2: For detection mileage The sample magnetic stress images are copied and spliced until the sum of the spliced images’ detection mileage is Within the interval, execute case 1 again;

[0086] Case 3: For detection mileage The sample magnetic stress images are split manually or intelligently by computer. The principle of manual splitting is to ensure that the image detection mileage after splitting is within The abnormal signal area will not be split into two pictures, and the case 1 will be executed for the split pictures.

[0087] The intelligent splitting process by computer includes:

[0088] Processing situation 3: For detection mileage The original magnetic stress image of the non-abnormal area is divided from the left and right endpoints of the original magnetic stress image to the opposite side. Corresponding pixel length Get two images with a pixel length of Target magnetic stress image.

[0089] Processing situation 4: Detection mileage The original magnetic stress image of the non-abnormal area is obtained by dividing the image by the pixel length from the left end point of the original magnetic stress image. Cut, when the remaining mileage After the cropping is completed, the target magnetic stress images obtained all have a pixel length of The standard target magnetic stress picture is obtained, and the remaining parts continue to be processed according to the above-mentioned case 1-case 2, and the above-mentioned case 3.

[0090] Processing situation 5: For detection mileage The original magnetic stress image of the signal with abnormal areas, the pixel horizontal coordinate of the original magnetic stress image is [0, L], then the abnormal area set G is expressed as

[0091] G={[g i ,g i +D i ],i=1,2,…,q}

[0092] Where q is the number of abnormal regions in an original magnetic stress image, g iis the starting point of the abscissa pixel of the abnormality at the i-th location, D i is the pixel length of the abnormality at the i-th location;

[0093] For each abnormal area, divide it into two sides from the starting point and end point of the abnormal area. or If the interval completely contains other abnormal areas, the abnormal areas contained in the original magnetic stress image are merged, including:

[0094] When the abnormal area When the abnormal area is merged into the same target magnetic stress image and cropped, the same is true when When the abnormal area is merged into the same target magnetic stress image and cropped, the cropped target magnetic stress image has a pixel length of Standard target magnetic stress picture; if or The interval does not completely contain other abnormal areas. From the abnormal area [g i ,g i +D i ] are extended to both sides, and the length is interval If this interval contains the endpoint of other abnormal areas, the endpoint is used as a clipping point, and then extended to the other side to ensure that the target magnetic stress image containing the abnormal area is Wherein, the term "not completely included" includes "not included" or "half included";

[0095] For abnormal areas The original magnetic stress image is cut into a target magnetic stress image in the abnormal area, and then the target magnetic stress image of the remaining non-abnormal area after cutting is cut into pieces, and the execution of case 1, case 2, case 3 and case 4 are repeated.

[0096] When the pixel coordinates (x, y) are not integers, the pixels of the four points surrounding the point in the original image are interpolated and rounded to obtain the processed pixel value f(x, y) of the point, including:

[0097] Bilinear interpolation is:

[0098]

[0099] Among them, Q 11 , Q 12 , Q 21 , Q 22The four pixels surrounding the point in the original image are (x1, y1), (x1, y2), (x2, y1), (x2, y2), and the pixel values in the original image are f(Q 11 ), f(Q 12 ), f(Q 21 ), f(Q 22 ).

[0100] It can be understood that by uniformly scaling the magnetic induction detection signal graph, the training process of the aforementioned intelligent maintenance decision-making model based on the detection signal graph is made more accurate and effective, thereby solving the problem that the horizontal axis scale of the magnetic stress image obtained each time is different due to the different mileage measured each time, and thus the problem that the obtained magnetic stress image cannot be directly used for intelligent identification of pipeline maintenance levels based on neural networks.

[0101] In this embodiment, by first determining the relationship between the detection mileage of the magnetic stress image and the average detection mileage of the magnetic stress images of the inspected pipe section, and then uniformly scaling the magnetic stress images based on the relationship between the detection mileage of the magnetic stress image and the average detection mileage of the magnetic stress images of the inspected pipe section, the problem that the horizontal axis scale of the magnetic stress image obtained each time is different due to the different mileage measured each time, and thus the problem that the obtained magnetic stress images cannot be directly used for intelligent identification of pipeline maintenance levels based on neural networks, can be solved.

[0102] After the excavation data accumulates to a certain level, the excavation defect data can be obtained from the defect management submodule, and the intelligent maintenance decision model can be optimized through data migration.

[0103] Based on the content of the above embodiment, in this embodiment, when making intelligent pipeline maintenance decision, it also includes: a process of adjusting and migrating the intelligent maintenance decision model;

[0104] The process of adjusting and migrating the intelligent maintenance decision model includes:

[0105] After accumulating pipeline excavation data, the surrounding environmental factors that affect magnetic induction, including metal, trees, and pipeline properties, are One-Hot encoded or continuously encoded to obtain the first eigenvector. The eigenvector corresponding to the feature map obtained by the convolutional neural network of the target area selected by the region generation network RPN is spliced with the first eigenvector to obtain the second eigenvector. The second eigenvector is calculated by the fully connected neural network layer and then the Softmax function is used to calculate the probability L of the sample magnetic stress image belonging to each maintenance level. i , use the stochastic gradient descent method SGD to train the fully connected network, the network output Li In the results, the maximum L i The most likely maintenance level for the abnormality is shown in the figure below.

[0106] Among them, the probability L of the signal sample belonging to each maintenance level i The calculation method is:

[0107]

[0108] Among them, p i is the output of the fully connected network layer, and N represents the number of maintenance levels.

[0109] In this embodiment, through such processing, the maintenance level finally predicted by the intelligent maintenance decision model is more accurate and has greater reference significance, thereby effectively solving the problem of low accuracy and low work efficiency in manual interference signal investigation due to large differences in personnel levels and huge workload. Misjudgment and omission of signals not only cause enterprises to blindly carry out maintenance, but also cause economic losses and social impacts due to suspension of work and production.

[0110] Based on the content of the above embodiment, in this embodiment, when making a pipeline intelligent maintenance decision, the method further includes: optimizing the intelligent maintenance decision model so that the intelligent maintenance decision model outputs a defect type;

[0111] Add a multi-label Sigmod loss function after the fully connected layer of the network to obtain the probability of the sample magnetic stress image being predicted as each type of defect

[0112]

[0113] in, is the column vector output by the last fully connected layer of the network, M is the number of defect types,

[0114] Take the cross entropy loss function and use the stochastic gradient descent method SGD to train the network. The network output The defect types corresponding to all positions in the vector that are greater than the threshold;

[0115] After completing the above processing, the network outputs the defect pixel location, defect type and corresponding repair level during prediction.

[0116] In this embodiment, after further accumulation of data, the output of the network can also migrate to more accurate defect types, such as dents, metal corrosion, weld defects, etc. Through the above-mentioned processing process provided by this embodiment, when using the model for prediction, accurate defect type information can be further output, thereby providing more effective guidance for subsequent maintenance.

[0117] According to the above two embodiments, by collecting defect data and surrounding environment data of pipeline excavation and maintenance decision data obtained through empirical formulas, the accumulated actual defect repair decision results and their corresponding data signals are substituted into the training module for model tuning and further accurate prediction of defect types.

[0118] It's understandable that defect data verified through pipeline excavation can be further used to make repair decisions using standard methods for the defects identified. The accumulated actual defect repair decision results and their corresponding image signals can be fed into the training module for fine-tuning. Once the data volume reaches a certain level, the model can be directly migrated to further improve network accuracy using actual data. Furthermore, the model can be optimized by adding a multi-label Sigmod loss function, allowing the network to migrate towards more accurate defect types, such as dents, metal corrosion, and weld defects.

[0119] The following example illustrates the training and optimization process of the above model:

[0120] First, the image is normalized into vectors of uniform size based on pixel values. These vectors are then passed through a grayscale VGG convolutional neural network to extract features, resulting in a feature map vector. After the feature map is extracted by the convolutional head, it is fed into the RPN (Region Proposal Network) to identify the target region, which is then predicted.

[0121] Based on the manually annotated image anomaly locations (which can be understood as defective pixel locations) and pipeline signal anomaly types (which can be understood as maintenance levels), the ratio of abnormal to normal signal counts was calculated. The existing data was randomly divided into training and test sets based on this ratio. The training set was used to train the network model. The model input was a normalized image, and the output was the defective pixel location and repair level. The repair level of magnetic stress anomaly signals was categorized into three scenarios: immediate repair, planned repair, and monitoring and use. The network loss function was calculated for each sample. To reduce the loss function, SGD (Stochastic Gradient Descent) was used to update network parameters and train the object detection network. To accelerate network training, anchors of a single size were selected based on the shape characteristics of the magnetic induction signal images, with an aspect ratio of 1:2. The network performance was tested on the test set. As further data was collected and accumulated, additional data was fed into the training module for fine-tuning to further improve the network's detection accuracy.

[0122] After further accumulation of signal image data and pipeline excavation data, factors affecting pipelines, such as metal, trees, and pipeline attributes, can be coded using one-hot encoding or continuous numerical encoding for vectorization. This means that during excavation, factors that may affect magnetic stress, such as trees, wires, and pipeline attributes around the pipeline, are counted and recorded to generate new feature vectors. For example, pipeline attribute data (pipeline material, diameter, and wall thickness) can be categorized into one-hot encoding and continuous encoding, as follows:

[0123]

[0124] Among them, m represents the pipe type, n represents the type of pipe diameter, and r represents the different thickness values of the pipe. The feature map of the special signal area selected by the RPN network is spliced with the feature vector of the surrounding environmental factors. The spliced feature vector is calculated by the fully connected neural network layer and then the Softmax function is used to calculate the probability L of the signal sample belonging to each maintenance level. i , also use SGD method to train the network, the network output L i In the results, the maximum L i The most likely maintenance level for the abnormality is L. The probability L of the signal sample belonging to each maintenance level is i The calculation is as follows:

[0125]

[0126] Where p i It is the output of the fully connected network layer, and N represents the number of maintenance levels, such as divided into 3 levels: immediate repair, planned repair, and monitoring use.

[0127] After further accumulation of data, the output of the network can also be migrated to more accurate defect types, such as dents, metal corrosion, weld defects, etc. A multi-label Sigmod loss function is added after the fully connected layer of the network to obtain the probability of the signal sample being predicted as each type of defect

[0128]

[0129] in, is the column vector output by the last fully connected layer of the network, M is the number of abnormal types,

[0130] Adopt cross entropy loss function, SGD method to train the network, network output The defect types corresponding to all positions in the vector that are greater than a threshold (usually 0.5).

[0131] After completing the above steps, the network can output the defect type and the corresponding maintenance level during prediction, such as "corrosion defect, monitoring and use".

[0132] In this embodiment, the training process of the intelligent maintenance decision model based on the detection signal strength value includes:

[0133] Obtaining sample magnetic stress detection signal strength values from expert analysis data in the defect management submodule; wherein the sample magnetic stress detection signal strength values need to cover various defective pixel locations and various repair levels;

[0134] Obtaining the pipeline abnormal area location and maintenance level corresponding to the sample magnetic stress detection signal strength value from the expert analysis data in the defect management submodule; wherein the sample data is divided into a training set and a test set;

[0135] The sample magnetic stress detection signal strength values with predetermined defective pixel locations and maintenance levels are used as the model input, and the mileage locations and maintenance levels of pipeline abnormal areas corresponding to the sample magnetic stress detection signal strength values are used as the model output. The neural network model is trained and tested based on an intelligent learning algorithm to obtain an intelligent maintenance decision-making model.

[0136] It can be seen that this embodiment makes an intelligent maintenance decision based on the magnetic induction detection signal strength data, which is highly adaptable and not affected by human subjective factors, thereby effectively improving the accuracy of the analysis of the detection results, avoiding blind maintenance, greatly improving economy and reducing labor, and at the same time improving the safety of the pipeline.

[0137] It should be noted that pipeline inspection signal data generally represents a data pattern of signal strength corresponding to pipeline mileage. This data is typically displayed and analyzed in the form of a signal graph. For example, based on the obtained magnetic stress signal intensity value, a corresponding magnetic stress signal image can be drawn, and then an intelligent algorithm can be used to predict the severity of pipeline defects based on the signal image. However, since digital signals lose some precision when converted into images, the accuracy of the algorithm has certain limitations. Processing and calculating the image requires a large amount of computing resources, resulting in slow calculation speed and very high requirements for computer hardware. To this end, this embodiment proposes a pipeline intelligent maintenance decision-making method that directly targets data signals. This method is not only applicable to the intelligent evaluation of pipeline integrity based on signal data from non-contact pipeline inspection, such as magnetic stress inspection, but can also be used for direct intelligent evaluation of other types of inspection data. On the one hand, it has a wide range of applicability. On the other hand, while improving the accuracy and efficiency of pipeline integrity evaluation, it also reduces labor costs, improves economic benefits and pipeline safety, and fills the technical gap in existing pipeline integrity intelligent evaluation.

[0138] Based on the content of the above embodiment, in this embodiment, after obtaining the sample magnetic stress detection signal strength value from the expert analysis data in the defect management submodule, the following steps are further included:

[0139] The sample magnetic stress detection signal intensity values are processed uniformly, specifically including:

[0140] The magnetic induction detection signal of each pipeline section is usually a two-dimensional array in the form of:

[0141]

[0142] Where m is the number of sampling times, and the magnetic induction detection signal obtained at each sampling point is:

[0143]

[0144] Among them, n is the dimension of magnetic induction detection signal, h ij Indicates the value of the j-th signal feature of the magnetic induction detection signal;

[0145] The sampling frequency of the magnetic induction detection signal data is unified. The length of H represents the length of the detected pipe section. The magnetic induction detection signal data is standardized so that the length of each detection signal vector is the same, which is convenient for network training.

[0146] First, the average length of the magnetic induction detection signal data of all inspected pipe sections is counted. Round to an integer, then perform Cases 1 to 5 below:

[0147] Case 1: For array length Signal data, use interpolation to scale the array length to Specifically, first get the position of the sampling points of the new array in the original numerical coordinates:

[0148]

[0149] Where X is the position information set of the new array after transformation;

[0150] According to the new position information X, the detection signal at the new position is calculated using linear interpolation The interpolation method is:

[0151]

[0152] Where j = [i] is the integer part of i, is the magnetic induction detection signal data at the corresponding position in the original data, and the new magnetic induction detection signal data H is obtained. * The length of the vector is

[0153] Case 2: For array length Signal data, for The arrays are concatenated to obtain a new signal:

[0154]

[0155] The signal length becomes N*l until Then follow the method in case 1 to scale the length to

[0156] Case 3: For array length The data without abnormal signals are divided into the opposite sides from the starting position and the end position of the array Length, get two new data, namely:

[0157]

[0158]

[0159] Case 4: For array length The non-abnormal signal data H is from the array Starting point position Start cutting at length and get a length of signal until the length of the remaining part Then, according to the method of case 1-case 3, the length of the remaining part is also changed to

[0160] Case 5: For array length There is abnormal signal data H, then the abnormal region set G is expressed as

[0161] G={[g i ,g i +D i ],i=1,2,…,q}

[0162] Where q is the number of abnormal regions in H, g i is the starting point of the anomaly at the i-th location, D i is the length of the abnormal position at the i-th location. First, for each abnormal area, divide it from the start and end points of the abnormal signal to both sides. Length, if or If the interval completely contains other signal abnormal intervals, the abnormal intervals are merged and trimmed. Specifically: When , the abnormal areas are merged into the same signal and trimmed. Similarly, when When , the abnormal areas are merged into the same signal and trimmed. The length of the trimmed array is like or The interval does not completely contain other abnormal signals. From the abnormal signal [g i ,g u +D u ] are extended to both sides, and the length is interval If this interval contains the endpoint of other abnormal signals, then the endpoint is used as a clipping point and then extended to the other side to ensure that the length of the part containing the abnormal signal is To include endpoint g u-1 +D i-1 For example, get the clipping interval If the right endpoint falls within other abnormal intervals, it will shrink to the left to the point without abnormalities, and the final trimming interval [g i-1 +D i-1 ,g i+1 ], repeat Case 1-Case 2 for the final trimmed signal array; for The abnormal signal area is directly cropped into a signal array, and then the above steps are performed on the remaining part; for the signal-free area remaining after the above cropping, case 1 to case 4 are executed cyclically.

[0163] It can be understood that by uniformly scaling the magnetic induction detection signal data, the training process of the aforementioned intelligent maintenance decision model based on detection signal strength values is more accurate and effective, thereby resolving the problem that the horizontal scale of the magnetic induction detection signal data obtained each time due to different mileage measurements is different, and the resulting magnetic induction detection signal data cannot be directly used for intelligent identification of pipeline maintenance levels based on neural networks. In this embodiment, by first uniformly scaling the magnetic induction detection signal data based on the relationship between the array length of the magnetic induction detection signal data and the average array length of the inspected pipeline section, the problem that the horizontal scale of the magnetic induction detection signal data obtained each time due to different mileage measurements is different, and the resulting magnetic induction detection signal data cannot be directly used for intelligent identification of pipeline maintenance levels based on neural networks, can be resolved.

[0164] The following is a detailed explanation of the training and optimization process of the above data signal-based model through specific examples:

[0165] The magnetic induction detection signal data is treated as a two-dimensional array, vectorized, and the vector is passed through a grayscale VGG convolutional neural network to extract features, resulting in a feature map vector. After the feature map is extracted by the convolution head, it is fed into the RPN (Region Proposal Network) to identify the target region, which is then predicted.

[0166] Based on the location of pipeline anomalies and the repair level of abnormal signals corresponding to magnetic induction detection signal data, determined by signal analysis experts at the inspection company based on experience or excavation (the repair levels for magnetic induction abnormal signals are divided into three categories: immediate repair, planned repair, and monitoring and use), the ratio of abnormal to normal signals is calculated. The existing data is randomly divided into training and test sets based on this ratio. The training set is used to train the network model. The model input is the normalized signal data, and the output is the start and end positions of the defect location on the original signal data (i.e., the number of columns in a two-dimensional array) and the corresponding repair level. The loss function of the sample in the network is calculated. To reduce the loss function, SGD (Stochastic Gradient Descent) is used to update the network parameters and train the object detection network. To speed up network training, anchor selection is based on the characteristics of the magnetic induction signal, with a single anchor size and an aspect ratio of 1:2. The test set is used to test the network performance. After further data collection and accumulation, the network is fine-tuned by inserting new data into the training module to further improve the network's detection accuracy.

[0167] When applying the network model, for a magnetic induction signal, if the signal length Then scale the signal array to the standard size according to the method in Part 1; if the signal length is Ensure that the potential fault area can be fully detected by the network at least once and reduce the distortion of the data signal to improve the network accuracy. At this time, no scaling is performed but starting from the left end point of the signal array, a length of Sliding window, each The signal is trimmed once until the right side of the sliding window reaches the boundary of the signal array, and the result is (Round down) signal arrays. These signal arrays are input into the network for prediction, and the interval of the abnormal signal location on the data and the abnormal signal repair level are output based on the network prediction results.

[0168] After further accumulation of signal data and pipeline excavation data, factors affecting pipelines, such as metal, trees, and pipeline attributes, can be coded using one-hot encoding or continuous numerical encoding for vectorization. This means that during excavation, factors that may affect magnetic induction, such as trees, wires, and pipeline attribute data around the pipeline, are counted and recorded to generate new feature vectors. For example, pipeline attribute data (pipeline material, diameter, and wall thickness) can be categorized into one-hot encoding and continuous encoding, as follows:

[0169]

[0170] Where m represents the pipe type, n represents the type of pipe diameter, and r represents the different thickness values of the pipe. The feature map of the special signal area selected by the RPN network is spliced with the feature vector of the surrounding environmental factors. The spliced feature vector is calculated by the fully connected neural network layer and then the Softmax function is used to calculate the probability L of the signal sample belonging to each maintenance level. i , also use SGD method to train the network, the network output L i In the results, the maximum L i The most likely maintenance level for the abnormality is L. The probability L of the signal sample belonging to each maintenance level is i The calculation is as follows:

[0171]

[0172] Where p i It is the output of the fully connected network layer, and N represents the number of maintenance levels, such as divided into 3 levels: immediate repair, planned repair, and monitoring use.

[0173] After further accumulation of data, the output of the network can also be migrated to more accurate defect types, such as dents, metal corrosion, weld types, etc. A multi-label Sigmod loss function is added after the fully connected layer of the network to obtain the probability of the signal sample being predicted as each type of defect

[0174]

[0175] in, is the column vector output by the last fully connected layer of the network, M is the number of abnormal types,

[0176] Adopt cross entropy loss function, SGD method to train the network, network output The defect types corresponding to all positions in the vector that are greater than a threshold (usually 0.5).

[0177] After completing the above steps, the network can output the defect type and the corresponding repair level during prediction, such as "weld defect, monitoring and use".

[0178] As can be seen, this embodiment addresses the technical blind spots of existing methods by providing a data-signal-based intelligent pipeline integrity assessment method. This embodiment standardizes and preprocesses historically accumulated magnetic induction detection data signals, then trains a target detection convolutional neural network to make decisions based on the processing results. The decision-making process also incorporates environmental factors to continuously improve recognition accuracy. Specifically, the technical solution employed in this embodiment includes: collecting magnetic induction detection signal data, abnormal locations in the signal data, and maintenance decision levels determined by experts based on experience or excavation; labeling abnormal points in the signal data with maintenance level classification; constructing a neural network model, standardizing and preprocessing the collected data to a uniform size, and then randomly dividing the data into training and test sets in proportion. The model is trained using the training set data and tested using the test set data, outputting an intelligent integrity assessment result; further collecting pipeline excavation defect data and surrounding environmental data, as well as maintenance decision data obtained through empirical formulas, and substituting the accumulated actual defect integrity intelligent assessment results and their corresponding data signals into the training module for model tuning and further accurate defect type prediction.

[0179] This demonstrates that this embodiment, based on magnetic induction detection signal data, can objectively and intelligently evaluate pipeline integrity. This improves the accuracy of test result analysis, avoids blind repairs, and significantly improves economic efficiency and work efficiency while reducing labor. This embodiment preprocesses historically accumulated magnetic induction detection signal data, then trains a target detection convolutional neural network to make decisions based on the processed data. This decision-making process also incorporates environmental factors to continuously improve recognition accuracy.

[0180] In the present invention, descriptions with reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples and features of different embodiments or examples described in this specification without mutual contradiction.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A pipeline integrity intelligent evaluation visualization system, characterized by: include: 3D visualization module and pipeline intelligent evaluation module; The three-dimensional visualization module includes a three-dimensional model visualization display submodule, a supporting environment submodule, an evaluation module data interface submodule, a pipeline information directory tree submodule, and a pipeline engineering data management submodule; The pipeline intelligent evaluation module includes a basic data management submodule, a detection management submodule, a defect management submodule, and an integrity intelligent evaluation and maintenance decision submodule; The evaluation module data interface submodule is used to provide a data interface for the basic data management submodule, the detection management submodule, the defect management submodule and the integrity intelligent evaluation and maintenance decision submodule; The integrity intelligent evaluation and maintenance decision submodule includes intelligent maintenance decisions for pipeline abnormal signals whose defect size and type cannot be quantified, and automatic evaluation and maintenance decisions for quantified defects. For pipeline abnormal signals whose defect size and type cannot be quantified, the detection signal graph and detection signal strength value of the pipe section are first obtained from the detection management submodule, and an intelligent maintenance decision model based on the signal graph or signal data pre-trained by deep learning is used to make intelligent maintenance decisions. Specifically, for pipeline abnormal signals whose defect size and type cannot be quantified, the detection signal graph and detection signal strength value of the pipe section are first obtained from the detection management submodule. For the detection signal graph and detection signal strength value, a pre-trained intelligent maintenance decision model based on the detection signal graph and an intelligent maintenance decision model based on the detection signal strength value are used respectively. According to the signal graph and signal strength data of the pipeline, a signal graph or signal data value of a standard length is determined. If the signal graph or signal value length The signal graph and value will be scaled to the standard size; if the signal length Starting from the left end point of the signal array, create a Sliding window, each The signal is trimmed once until the right side of the sliding window reaches the boundary of the maximum value of the signal, and the signal diagrams or signal arrays; these signal diagrams or signal arrays are input into the intelligent maintenance decision model for prediction, and the mileage interval where the abnormal signal location is located and the maintenance level of the abnormal signal are output according to the prediction results; For the automatic evaluation and repair decision-making of quantitative defects, batch and single defects with size and type information are detected. Different evaluation methods are used for different defect types. After selecting an evaluation method for a certain type of defect, the built-in evaluation algorithm is called to perform batch evaluation, single evaluation and repair decision-making on defect points of the same type or a certain defect using the same evaluation method; the three-dimensional model visualization display submodule performs three-dimensional dynamic interactive visualization of the intelligent evaluation and repair decision results.

2. The pipeline integrity intelligent evaluation visualization system according to claim 1 is characterized in that: The 3D visualization module is based on the principle of human-computer interaction. It determines the geographic information or relative position of the modeling object by conducting on-site surveys of refinery equipment and devices, ground markers, and ground and underground pipeline objects connected to the equipment and devices. It then builds a model of the object in the 3D model and adds, checks, deletes, and modifies the established engineering information related to pipelines, devices, and equipment through the 3D model or directory tree. The pipeline information serves as the data basis for the integrity evaluation of buried pipelines.

3. The pipeline integrity intelligent evaluation visualization system according to claim 1 is characterized in that: The basic data management submodule is used to obtain basic information of the pipeline where the pipeline is located from the three-dimensional visualization module, and is used to add, check, delete and modify the pipe diameter, wall thickness, pipe material, conveying medium and connection equipment information.

4. The pipeline integrity intelligent evaluation visualization system according to claim 1 is characterized in that: The detection management submodule includes three functions: detection pipeline management, detection project management, and detection data management, which are aimed at the logical management of detection operations, detection signals, and detection results; Among them, the inspected pipeline management is used to manage the inspection pipeline and its related inspection projects. The inspected pipeline management has an inspection scope for the inspected pipeline. Based on the built-in pipeline number naming rules, the pipeline number of the newly added inspection pipeline is automatically generated based on the pipeline where the inspected pipeline is located. The information of the newly added inspected pipeline includes: inspection length, inspection location, and remarks. The information is maintained, modified, and viewed. At the same time, after a new inspection pipeline is created, the inspected pipeline is highlighted on the 3D model by the pipeline number. In addition, the inspection project and related information of the inspected pipeline can be added and modified under the added inspected pipeline. Check, delete, and modify. After a pipeline inspection, manually locate the inspection marker information on the pipeline 3D model and click on the pipeline to be inspected to determine the inspection range of the inspection pipeline. The system automatically generates the inspection pipeline number based on the pipeline where the inspection pipeline is located. At the same time, based on the inspection pipeline number, the inspection pipeline is highlighted and located in the 3D model. The information of the inspected pipeline is maintained and modified, including the entry and modification of the inspection length, inspection positioning point, design coefficient information, and the deletion of the inspected pipeline. Among them, after the inspected pipeline is deleted, the inspection project, defect point location, and defect data associated with it are also deleted. Inspection project management: This module is used to establish an inspection project after conducting an inspection on a pipeline section. This includes creating a new inspection project under the newly inspected pipeline and managing the addition, query, deletion, modification, and maintenance of the inspection project and its related information, including the inspection project name, inspection method, inspection equipment, inspection time, and inspection unit. Deleting an inspection project also deletes the associated defect location and defect data. This module also allows users to view a list of all inspection projects under the inspected pipeline by inspected pipeline number. Detection data management: used for the management of detection data for a certain detection project, with functions including detection signal diagram management, detection signal data management, and detection result data management; among them, detection signal diagram management is the management of the detection signal diagram of the pipeline; detection signal data management is in the form of a form, and the strength values of the detection signal along the detection mileage are imported into the system in the form of a form according to the detector sampling results; the detection result data is also managed in the form of importing the system into the form; it also includes detection signal, result template download, data import, and data addition, query, deletion, and modification.

5. The pipeline integrity intelligent evaluation visualization system according to claim 1 is characterized in that: The defect management submodule is used to manage defect data obtained after a pipeline inspection through expert signal analysis, excavation verification, or intelligent maintenance decision-making. Defect data can interact with the three-dimensional model to enable the addition, query, deletion, and modification of defect information on the inspected pipeline. At the same time, defect information is managed separately according to different defect types. Among them, the corrosion rate of corrosion defects is directly calculated based on the defect depth and commissioning time using a built-in background algorithm. New defects are handled differently: For individual defects identified through excavation verification and expert analysis, the defect location is directly added to the 3D pipeline model by measuring with a ruler using markers on the inspected pipeline, and then the defect data is entered. For defects obtained through intelligent maintenance decisions, the intelligent maintenance decision model outputs the mileage location and maintenance level of the pipeline abnormal area, automatically marks the abnormal interval on the corresponding pipe section on the 3D platform, and automatically synchronizes the maintenance level to the abnormal interval of the corresponding pipeline. At the same time, the defects and maintenance decision results are saved to the defect management submodule so that the maintenance decision results can be displayed when entering the intelligent evaluation interface again. Among them, the defects obtained through intelligent decision-making can be saved together with the defects obtained through expert analysis and excavation verification through the defect addition method field; the defect addition methods include: expert analysis, excavation verification or intelligent decision-making; For batch defects, the batch defects obtained through intelligent decision-making are intelligently added and located in the same way. For other defects, batch defect data is imported through data import. The specific operations are as follows: download the defect information template, edit the defect information, and then import it into the system. After the batch defect data is successfully imported, the defect location is manually entered for each defect data, and the newly added defects are associated and displayed in the 3D model, or the defects to be added and located are selected, and the defect is added to the 3D pipeline through the Add button of the defect location, and the defect information is automatically updated to the 3D defect point.

6. The pipeline integrity intelligent evaluation visualization system according to claim 5 is characterized in that: There are three situations for removing defects: A. Delete the defect point location to retain the defect data. When you add the pipeline abnormal area mileage location with the same defect number again, the defect data will be automatically updated. B. Delete the defect data but keep the 3D defect point location. After adding the defect point information at that location again, it will be automatically associated with the 3D defect point. C. Delete the defect data and the mileage position of the abnormal area of the pipeline on the 3D model pipeline at the same time; The defect management submodule realizes the highlight positioning of defect points in three dimensions by numbering the defect points; On the defect management page, select the year, device, pipeline number, or inspection project from the drop-down list to view all defect points and defect information for the inspection project.

7. The pipeline integrity intelligent evaluation visualization system according to claim 1 is characterized in that: If the signal includes ultrasonic guided waves and magnetic stress signals, the results of both are output simultaneously. The mileage interval of the abnormal signal is the union of the two, and the maintenance decision level is the result of the higher level. The abnormal signal mileage interval of the pipeline and the corresponding maintenance decision result are output; or the results of the two are fused according to the weight and then output; Based on the output abnormal mileage interval, the background automatically marks the abnormal interval in red on the pipe section in the three-dimensional platform, and automatically synchronizes the maintenance decision result to the abnormal interval of the pipeline section. At the same time, the evaluation result is saved and can be displayed when entering the intelligent evaluation interface next time.

8. The pipeline integrity intelligent evaluation visualization system according to claim 7 is characterized in that: For the automatic evaluation and repair decision of quantitative defects, the system can obtain batch and single defects with size and type information through detection. Different defect types have different defect evaluation methods. After selecting the evaluation method for a certain type of defect, the system can call the built-in evaluation algorithm to perform batch evaluation, single evaluation and repair decision for the same type of defect points or a single defect using the same evaluation method. After the system evaluation is successful, the evaluation results are saved and displayed the next time you enter the evaluation page. Single defect automatic evaluation is for defect points with known defect types and sizes. In the single evaluation module, select the defect point to be evaluated, select the evaluation method, and automatically evaluate the defect point. The defect evaluation results are saved for the evaluated defect point. The evaluation results are displayed when you enter the page again and you can select the evaluation method to re-evaluate. Automatic batch defect evaluation is a process of selecting the device, pipeline number, inspection project, and defect type on the batch evaluation page. After selecting the evaluation method and entering the required evaluation information for the defect points to be evaluated, the system's built-in evaluation algorithm is used to evaluate a certain type of defect in batches. The evaluation method of the evaluated defect points can be modified and re-evaluated.

9. The pipeline integrity intelligent evaluation visualization system according to claim 7 is characterized in that: When excavation data is not available in the early stage, the intelligent maintenance decision model is trained and tested through expert analysis data in the defect management submodule; After the degree of excavation data accumulation meets the preset conditions, the excavation defect data is obtained from the defect management submodule, and the intelligent maintenance decision model is optimized through data migration.

10. The pipeline integrity intelligent evaluation visualization system according to claim 9 is characterized in that: The training process of the intelligent maintenance decision model based on the detection signal graph includes: Obtaining sample magnetic stress images from expert analysis data in the defect management submodule; wherein the sample magnetic stress images need to cover various defect pixel locations and various repair levels; Obtaining the mileage location and maintenance level of the pipeline abnormal area in the sample magnetic stress image from the expert analysis data in the defect management submodule; The model uses sample magnetic stress images with predetermined defect pixel locations and repair levels as input, and uses the mileage locations and repair levels of pipeline abnormality areas in the sample magnetic stress images as output. The neural network model is trained and tested based on an intelligent learning algorithm to obtain an intelligent maintenance decision model. The training process of the intelligent maintenance decision model based on the detection signal strength value includes: Obtaining sample magnetic stress detection signal strength values from expert analysis data in the defect management submodule; wherein the sample magnetic stress detection signal strength values need to cover various defective pixel locations and various repair levels; Obtaining the pipeline abnormal area location and maintenance level corresponding to the sample magnetic stress detection signal strength value from the expert analysis data in the defect management submodule; wherein the sample data is divided into a training set and a test set; The sample magnetic stress detection signal strength values with predetermined defective pixel locations and maintenance levels are used as the model input, and the mileage locations and maintenance levels of pipeline abnormal areas corresponding to the sample magnetic stress detection signal strength values are used as the model output. The neural network model is trained and tested based on an intelligent learning algorithm to obtain an intelligent maintenance decision-making model.

Images