Visual optimization design method combining digital prototype construction and simulation data of underwater production system of offshore oil engineering

Abstract

The invention relates to the technical field of visual optimization, in particular to an offshore oil engineering underwater production system digital prototype construction and simulation data combination visual optimization design method. The method comprises the following steps: acquiring system offshore petroleum engineering project information through an interface, and creating a new project in a digital prototype platform according to the acquired project information; importing a three-dimensional geometric model of an offshore oil engineering underwater production system, and processing the three-dimensional geometric model; mounting a physical performance model on the processed geometric model, and correcting the quality factor in the mounting process; and performing CAE simulation on the three-dimensional geometric model, and importing and displaying a CAE simulation calculation result through a CAE lightweight rendering module. According to the design of the invention, a new project can be quickly created in a digital prototype platform, and comprehensive analysis and optimization of the system are carried out by utilizing a three-dimensional geometric model, a physical performance model and a CAE (Computer Aided Engineering) simulation technology.

Description

Technical Field

[0001] The present invention relates to the technical field of visualization optimization, and in particular to a visualization optimization design method combining digital prototype construction with simulation data for an underwater production system of an offshore oil project. Background Art

[0002] With my country's development and utilization of marine resources, the application of subsea production systems on offshore oil platforms is becoming increasingly widespread. However, traditional design methods lack efficiency in data integration, analysis, and utilization, making it difficult to quickly process and optimize design solutions. Due to the lack of collaborative and efficient design tools in the traditional design process, the design cycle is long and cannot quickly respond to market demand. In traditional design, each discipline's design is relatively isolated and dispersed, making it difficult to achieve efficient multi-disciplinary collaborative design. However, current offshore oil engineering is a highly integrated and complex system engineering project, requiring the deep integration of multiple disciplines such as underwater, process, electrical, and mechanical. In the field of offshore engineering, initial applications have only been found in the jack-up offshore platform lifting system and the mud intake module of the drilling system, focusing on component design. However, there is no systematic flow simulation research on the operation, start-up and shutdown, and fault conditions of the entire skid, nor is there a digital prototype of the entire skid combined with visual simulation. Therefore, a method for constructing a digital prototype of an offshore oil engineering subsea production system and combining simulation data with visual optimization design is provided. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for visual optimization design of a digital prototype of an underwater production system of an offshore oil engineering project in combination with simulation data, so as to solve the problems of low data integration efficiency, long design cycle, difficulty in multi-disciplinary collaboration and lack of visual optimization proposed in the above-mentioned background technology.

[0004] To achieve the above objectives, the present invention provides a method for visual optimization design of a digital prototype of an underwater production system for an offshore oil project by combining simulation data, comprising the following steps:

[0005] S1. Obtain offshore oil engineering project information from the system through the interface, and create a new project in the digital prototype platform based on the acquired project information;

[0006] S2. Importing a 3D geometric model of an underwater production system of an offshore oil project and processing the 3D geometric model;

[0007] S3, mounting the physical performance model on the processed geometric model, and correcting the quality factor during the mounting process;

[0008] S4. Perform CAE simulation on the 3D geometric model, and import and display the CAE simulation results through the CAE lightweight rendering module;

[0009] S5. Generate training samples based on simulation calculation results through the reduced-order model tool, and perform parameter adjustment verification on multiple working condition data to obtain the optimal design solution.

[0010] As a further improvement of the present technical solution, in S2, the three-dimensional geometric model of the offshore oil engineering underwater production system is imported and the three-dimensional geometric model is processed, including the following steps:

[0011] S1.1. Analyze the data format of the 3D geometric model of underwater production facilities, parse the STEP / IGES file, obtain the geometric entities, topological relationships, geometric construction information, material properties, and assembly information of the 3D model components, and convert different data formats into a unified data format through the prototype system format converter for import;

[0012] S1.2. Use lightweight rendering modules to optimize the display performance of 3D geometric models;

[0013] S1.3. After the 3D geometric model is imported, the coordinates and normal directions of the components of the 3D geometric model are analyzed to obtain the explosion direction of each component;

[0014] S1.4. Apply a specific explosion distance to make each component discrete relative to the view center;

[0015] S1.5. Adjust the display mode.

[0016] As a further improvement of the present technical solution, in S1.3, analyzing the coordinates and normal directions of the components of the three-dimensional geometric model to obtain the explosion direction of each component includes the following steps:

[0017] S1.31. Analyze the local coordinate system of each component in the overall model and record the coordinates of the component's center point as a reference point.

[0018] S1.32. For each component, calculate its normal vector based on the geometric characteristics of its surface;

[0019] S1.33. Based on the assembly information, clarify the connection method between each component and record the relative position and direction between adjacent components;

[0020] S1.34. Use the normal vector of the component surface as a reference for the initial explosion direction;

[0021] S1.35. Adjust the explosion direction according to the assembly relationship between components;

[0022] S1.36, transform the local explosion direction of each component into the global coordinate system;

[0023] S1.37. Use the force-directed layout algorithm to optimize the explosion direction of the components so that the components are evenly distributed in the exploded view.

[0024] As a further improvement of the present technical solution, in S1.37, a force-directed layout algorithm is used to optimize the explosion direction of the components so that the components are evenly distributed in the exploded view, including the following steps:

[0025] S1.371. Take the center point of each component as its initial position and place it at the starting point in the global coordinate system;

[0026] S1.372. Assign a virtual "mass" attribute to each component, define the connection relationship between components, and assign virtual "spring" parameters to them;

[0027] S1.373. For parts connected by assembly relationships, calculate the attractive force; for any two parts, calculate the repulsive force;

[0028] S1.374. For each component, calculate the resultant of all attractive and repulsive forces acting on it;

[0029] S1.375. Calculate the displacement of each component based on the resultant force and update the position;

[0030] S1.376. If the position changes of all components are less than the preset threshold r, the optimization is completed, and the final position of each component after optimization is recorded.

[0031] As a further improvement of the present technical solution, in S3, mounting the physical performance model on the processed geometric model includes the following steps:

[0032] S3.1. Divide the geometric model into different functional areas based on the functional requirements of the subsea production system. Determine the physical characteristics that need to be simulated in each area. Map the physical field data to the nodes of the geometric model using interpolation mapping formulas.

[0033] S3.2. For each component, identify its physical properties and record the component's working conditions;

[0034] S3.3. Bind each component to its corresponding physical model;

[0035] S3.4. According to the actual engineering data, set the material properties, initial conditions and boundary conditions for each component.

[0036] As a further improvement of the present technical solution, in S3.1, mapping the physical field data to the nodes of the geometric model through the interpolation mapping formula includes the following steps:

[0037] S3.11. Make it clear whether the physical field data is on the unit node;

[0038] S3.12, extracting node coordinates and element topology information of the geometric model;

[0039] S3.13, obtain the physical field data defined at the cell center;

[0040] S3.14. Traverse the unit topology information and record the unit list associated with each node;

[0041] S3.15. Calculate the physical field value of the node based on the physical field value of the unit to which the node belongs, and modify the weight in the physical field value calculation process;

[0042] S3.16. Update the calculated physical field values ​​to the nodes.

[0043] As a further improvement of this technical solution, in S3.15, the physical field value of the calculation node is:

[0044]

[0045] Among them, D j Represents the physical field value of the unit; w j represents weight; T i represents the interpolated physical field value; N i Indicates the unit set to which the node belongs; j indicates the index of the unit;

[0046] The weights used in the calculation of physical field values ​​are modified based on the quality factor of the element:

[0047]

[0048] Among them, T i1 Represents the physical field value after weight correction; q j Represents the quality factor of the unit; ‖P i -G i ‖ represents node P i and unit center G i The distance between i Represents a node; G i represents the unit center; p represents the smoothing parameter.

[0049] As a further improvement of the present technical solution, in S4, the CAE simulation calculation results are imported and displayed through the CAE lightweight rendering module, including the following steps:

[0050] S4.1. Import the geometric model into the CAE simulation software, analyze the CAE result data format of the simulation software, parse the result data file, and obtain information;

[0051] S4.2. Create an unstructured grid data structure based on vtk, fill the acquired information into the structure, and save the acquired information as a vtp file;

[0052] S4.3. Based on the vtk.js library, the result file processed by the data conversion program is rendered using a lightweight rendering algorithm;

[0053] S4.4. After loading the result file, parse the model and call the corresponding API for drawing and rendering.

[0054] As a further improvement of the present technical solution, in S4.3, based on the vtk.js library, the result file processed by the data conversion program is rendered using a lightweight rendering algorithm, including the following steps:

[0055] S4.31. Use the API provided by vtk.js to load the .vtp file and parse the geometry and physics data.

[0056] S4.32, dynamically adjust the model's level of detail based on viewpoint distance;

[0057] S4.33, divide the geometry in the scene into multiple hierarchies, each hierarchy containing a bounding box;

[0058] S4.34, use the Decimation algorithm to remove redundant vertices and faces;

[0059] S4.35. Apply the Phong lighting model to enhance rendering and map scalar data to color space.

[0060] As a further improvement of the present technical solution, in S5, a training sample is generated based on the simulation calculation results by using a reduced-order model tool, and parameter adjustment verification is performed on multiple working condition data to obtain the optimal design solution, which includes the following steps:

[0061] S5.1. In the CAE model reduction tool module, select a CAE model from the library, enter the number of samples, and obtain parameter name information. The user can edit the maximum and minimum values ​​of the parameters as needed.

[0062] S5.2. Generate a sample table based on the parameter range and sample number set by the user;

[0063] S5.3. After the sample table and sample data are set up, the reduced-order model training is performed;

[0064] S5.4. After the reduced-order model training is completed, it is compared with the corresponding CAE model and uploaded to the model library.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] 1. This offshore oil engineering underwater production system digital prototype construction and visualization optimization design method, combined with simulation data, allows for the rapid creation of new projects within the digital prototype platform and comprehensive system analysis and optimization using 3D geometric models, physical performance models, and CAE simulation technology. In particular, in step S3, by attaching the physical performance model and modifying the quality factor, the system behavior under actual operating conditions can be more accurately simulated. Furthermore, in step S5, a reduced-order model tool is used to generate training samples based on simulation results and perform multi-condition verification. This helps significantly improve design efficiency while ensuring design accuracy, thereby accelerating the product development cycle.

[0067] 2. The digital prototype construction of the offshore oil engineering underwater production system, combined with simulation data and a visual optimization design approach, uses a CAE lightweight rendering module to intuitively present complex simulation results, significantly enhancing the user's understanding and analysis of the results. Lightweight rendering and exploded view optimization not only improve the display performance of the 3D model, but also allow users to more clearly observe the assembly relationships and spatial layout of the system components. This highly visual interactive approach provides designers with a more intuitive design environment, facilitating the identification of potential issues and timely adjustments. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 is a flow chart of the overall method of the present invention;

[0069] Figure 2 This is a schematic diagram of the functional modules of the digital prototype construction system of this embodiment;

[0070] Figure 3 This is a diagram showing the operation interface of this embodiment;

[0071] Figure 4 This is a schematic diagram of the project scenario for this embodiment;

[0072] Figure 5 This is a diagram showing the cloud image results after CAE simulation calculation and mounting in this embodiment;

[0073] Figure 6 This is a schematic diagram of the model library import model in this embodiment;

[0074] Figure 7 Create a pipeline connection diagram for this example;

[0075] Figure 8 This is a cloud diagram showing the accuracy verification comparison between the order reduction algorithm and the CAE calculation results of this embodiment;

[0076] Figure 9 View different field data display diagrams for this embodiment;

[0077] Figure 10 Upload a display diagram for the reduced-order model of this embodiment. DETAILED DESCRIPTION

[0078] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0079] Example: See Figure 1-10 As shown, this embodiment provides a method for visual optimization design of a digital prototype of an offshore oil engineering underwater production system combined with simulation data, including the following steps:

[0080] S1. Obtain the system's offshore oil engineering project information through the interface, and create a new project in the digital prototype platform based on the acquired project information (including project name, person in charge, project details, geographic information code, etc.);

[0081] S2. Importing a 3D geometric model of an underwater production system of an offshore oil project and processing the 3D geometric model;

[0082] In this embodiment, the three-dimensional geometric model of the underwater production system of the offshore oil project is imported and processed, including the following steps:

[0083] S1.1. Analyze the data format of the 3D geometric model of underwater production facilities, parse the STEP / IGES file, obtain the geometric entities, topological relationships, geometric construction information, material properties, and assembly information of the 3D model components, and convert different data formats into a unified data format through the prototype system format converter for import;

[0084] S1.2. Use lightweight rendering modules to optimize the display performance of 3D geometric models (reduce the number of model polygons, retain key geometric features, compress texture map resolution, reduce memory usage, and support high-quality display of scenes, 3D objects, cameras, lighting, material textures, shadows, interactions, and other elements);

[0085] S1.3. After the 3D geometric model is imported, the coordinates and normal directions of the components of the 3D geometric model are analyzed to obtain the explosion direction of each component;

[0086] Analyzing the coordinates and normal directions of the components of the three-dimensional geometric model to obtain the explosion direction of each component includes the following steps:

[0087] S1.31. Analyze the local coordinate system of each component in the overall model, including its origin, X-axis, Y-axis, and Z-axis directions, and record the coordinates of the component's center point as a reference point.

[0088] S1.32. For each component, calculate its normal vector based on the geometric characteristics of its surface (triangular mesh or surface);

[0089] S1.33. Based on the assembly information, clearly identify the connection method between each component (e.g., by bolts, welding, or other means), and record the relative position and orientation of adjacent components.

[0090] S1.34. Use the normal vector of the component surface as a reference for the initial explosion direction. If the component has multiple normal directions (complex curved surfaces), use the average normal direction as a representative.

[0091] S1.35. Adjust the explosion direction based on the assembly relationship between parts to avoid overlapping parts in the exploded view. If two parts are connected by bolts, set their explosion directions in opposite directions to separate them along the connection axis.

[0092] S1.36. Transform the local explosion direction of each component into the global coordinate system to ensure that the explosion directions of all components are consistent throughout the model.

[0093] S1.37. Use the force-directed layout algorithm to optimize the explosion direction of components so that the components are evenly distributed in the exploded view for easy observation and analysis.

[0094] Furthermore, a force-directed layout algorithm is used to optimize the explosion direction of the components so that the components are evenly distributed in the exploded view, including the following steps:

[0095] S1.371. Take the center point of each component as its initial position and place it at a starting point in the global coordinate system (usually near the origin). The initial position is roughly distributed based on the local coordinates or assembly relationship of the component.

[0096] S1.372. Assign a virtual "mass" attribute to each component (set according to the actual volume or importance of the component), define the connection relationship (assembly relationship) between the components, and assign virtual "spring" or "repulsion force" parameters to them;

[0097] S1.373. For parts connected by assembly relationships, calculate the attractive force; for any two parts, calculate the repulsive force;

[0098] Among them, the attraction is:

[0099] F a =k a ·(d AB-d id );

[0100] Where k a represents the attraction coefficient, d AB Indicates the current distance between components A and B, d id Indicates the ideal assembly distance;

[0101] The repulsive force is:

[0102]

[0103] Among them, k r represents the repulsion coefficient;

[0104] S1.374. For each component, calculate the resultant of all attractive and repulsive forces acting on it;

[0105] Combined Force F t For: F t =∑F a +∑F r ;

[0106] S1.375. Calculate the displacement of each component based on the resultant force and update the position;

[0107] The displacement Δx is: Where m represents the mass of the component; Δt represents the time step;

[0108] S1.376. If the position changes of all components are less than the preset threshold r, the optimization is completed, and the final position of each component after optimization is recorded.

[0109] S1.4. Apply a specific explosion distance to make each component discrete relative to the view center to achieve an explosion effect;

[0110] S1.5. Adjust the display mode (the transparency of the model can be modified, and the display mode of the model can be switched by switching the data type (point, surface) drawn by the model. It supports overall modification of the model and individual modification of components. By controlling the position and observation angle of the scene camera, the model view can be switched. The axonometric view is displayed by default, and can be quickly switched to the front view, rear view, top view, bottom view, left view, and right view. It provides fast adaptive window and default view switching. By dragging the mouse and sliding the wheel, the model can be rotated, translated, and scaled to adjust to any view).

[0111] S3, mounting a physical performance model on the processed geometric model;

[0112] In this embodiment, the physical performance model is mounted on the processed geometric model, including the following steps:

[0113] S3.1. Divide the geometric model into different functional areas (such as pipelines, valves, pumps, sensors, etc.) based on the functional requirements of the subsea production system. Determine the physical properties that need to be simulated in each area (such as fluid flow, heat transfer, structural strength, vibration, etc.). Then, map the physical field data to the nodes of the geometric model using interpolation mapping formulas.

[0114] The physical field data is mapped to the nodes of the geometric model through the interpolation mapping formula, including the following steps:

[0115] S3.11. Make sure that the physical field data is at the element nodes (if the physical field data is defined at the element center (such as the element average), it needs to be mapped to the node);

[0116] S3.12, extracting node coordinates and element topology information of the geometric model;

[0117] S3.13, obtain the physical field data defined at the cell center;

[0118] S3.14. Traverse the unit topology information and record the unit list associated with each node;

[0119] S3.15. Calculate the physical field value of the node based on the physical field value of the unit to which the node belongs;

[0120] The physical field values ​​of the calculation node are:

[0121]

[0122] Among them, D j Represents the physical field value of the unit; w j represents weight; T i represents the interpolated physical field value; N i Indicates the unit set to which the node belongs; j indicates the index of the unit;

[0123] In finite element analysis or meshing, the shape and size of the elements may vary greatly. Some elements are slender, distorted, or significantly larger or smaller than other elements; these irregular elements will lead to increased interpolation errors because their physical field data may not reflect the actual value of the node well; the quality factor of the element (such as area, volume, shape factor, etc.) is introduced to correct the weight, and higher-quality elements (such as elements with regular shape and moderate area) contribute more to the interpolation results, while the contribution of poorer-quality elements (such as elements with distorted shape and too small area) is weakened; irregular elements (such as slender triangles, flat quadrilaterals, etc.) lead to distorted interpolation results. The physical field value of a very slender element at its center is greatly different from the actual value of the node. If the weights of these elements are not corrected, the interpolation results will be excessively affected by these abnormal elements; the interpolation results are more stable, avoiding the interference of abnormal elements on the overall results;

[0124] The weights used in the calculation of physical field values ​​are modified based on the quality factor (area, volume) of the element to reduce interpolation errors caused by irregular element shapes:

[0125]

[0126] Among them, T i1 Represents the physical field value after weight correction; q j Represents the quality factor of the unit; ‖P i -G i ‖ represents node P i and unit center G i The distance between i Represents a node; G i represents the unit center; p represents the smoothing parameter, which controls the locality of the interpolation.

[0127] S3.16. Update the calculated physical field values ​​to the nodes.

[0128] S3.2. For each component, identify its physical properties (e.g., material density, elastic modulus, thermal conductivity, friction coefficient, etc.) and record the component's operating conditions (e.g., pressure, temperature, flow rate, etc.) to facilitate subsequent definition of boundary conditions.

[0129] S3.3. Bind each component to its corresponding physical model;

[0130] S3.4. According to the actual engineering data, set the material properties for each component and set the initial conditions and boundary conditions (initial conditions: specify the state of the system at the initial moment (such as initial pressure, temperature, speed, etc.); boundary conditions: define how the system interacts with the outside world).

[0131] S4. Perform CAE simulation on the three-dimensional geometric model, and import and display the CAE simulation calculation results through a CAE (Computer Aided Engineering) lightweight rendering module;

[0132] In this embodiment, CAE (Computer Aided Engineering) lightweight rendering module is used to import and display CAE simulation calculation results, including the following steps:

[0133] Click "Add CAE Model" under the geometry model node and select "Model Library Selection" or "Local Upload";

[0134] S4.1. Import the geometric model into the CAE simulation software, analyze the CAE result data format of the simulation software, parse the result data file, and obtain the node coordinates, element type, element node index, node value and other information;

[0135] S4.2. Create an unstructured mesh data structure based on VTK, fill the acquired information (node ​​coordinates, element type, element node index, etc.) into the structure, and save the acquired information as a vtp (surface information) or vtu (volume information) file;

[0136] S4.3. Based on the vtk.js library, render the result file (.vtp) after the data conversion program using a lightweight rendering algorithm (introduce the vtk.js library into the front-end application and write the corresponding JavaScript code to load the previously generated .vtp file);

[0137] Among them, based on the vtk.js library, the result file (.vtp) processed by the data conversion program is rendered using a lightweight rendering algorithm, including the following steps:

[0138] S4.31. Use the API provided by vtk.js to load the .vtp file and parse the geometry and physics data (extracting information such as points, cells, value sizes, and value ranges).

[0139] S4.32, dynamically adjust the model's level of detail based on viewpoint distance (when the user is far away from the model, reduce the number of polygons to improve rendering efficiency, and dynamically switch between mesh representations of different resolutions);

[0140] S4.33. Divide the geometry in the scene into multiple hierarchies, each containing a bounding box. Only render the geometry within the visible area, ignoring the occluded parts.

[0141] S4.34, use the Decimation algorithm to remove redundant vertices and faces, and compress texture map resolution to reduce memory usage;

[0142] S4.35. Apply the Phong lighting model to enhance rendering and map scalar data to color space.

[0143] S4.4. After loading the result file, parse the model's points, units, numerical values, numerical ranges, data sets, etc., and call the corresponding API for drawing and rendering;

[0144] It supports manual import of any vtp file and loading and rendering in sequence. By parsing the vtm file, it automatically determines the vtp file that needs to be loaded. The color set can be modified to adjust the color display of all models / components. Currently, it supports a variety of common color schemes. The transparency of the model can be adjusted to achieve CAD perspective effects and highlight CAE result data.

[0145] S5. Generate training samples based on simulation results using a reduced-order model tool, and perform parameter adjustment and verification on multiple operating conditions to obtain the optimal design solution.

[0146] In this embodiment, a reduced-order model tool is used to generate training samples based on simulation calculation results, and multiple operating condition data parameter adjustment verification is performed to obtain the optimal design solution, including the following steps:

[0147] S5.1. Select a CAE model from the library in the CAE drop-down menu in the Reduced Order Model Tool module, enter the number of samples, and obtain parameter name information. The user can edit the maximum and minimum values ​​of the parameters as needed.

[0148] S5.2. Generate a sample table based on the parameter range and number of samples set by the user. Click "Import Sample Points" and the table will automatically update accordingly. Compress the "data" folder into a zip format and upload the imported data.

[0149] S5.3. After setting up the sample table and sample data, perform the reduced-order model training. Select the "SVD" algorithm, set the "Orthogonal decomposition retention order", "Verification sample ratio", and "Maximum number of rows in matrix partitioning", and click "Train".

[0150] S5.4. After the reduced-order model training is completed, compare it with the corresponding CAE model and click "Upload to Reduced-Order Model Library" to upload it to the model library.

[0151] Furthermore, the digital prototype construction system function in this embodiment includes the following steps (such as Figure 2 shown):

[0152] Log in to the platform and enter the system (the system operation interface is as follows Figure 3As shown), by default, you enter the "Portal Dashboard Interface", where you can view the number of "Published Projects", the number of "Projects Under Construction", the number of "Geometry Models", and the number of "Performance Models". You can also directly view the "Published" and "Under Construction" projects and related information (running project scenarios such as Figure 4 shown);

[0153] Click "Create Project", fill in "Project Name", "Person in Charge", "Project Details", and upload a project cover. After completion, click "Save". You can also obtain external information from the system through the interface to complete the entry of project information. A new project will be created in "Build Project";

[0154] Import models. Two import methods are supported:

[0155] Import via the model library:

[0156] In the model library on the right, click the model name according to the node where the model is located to directly import the corresponding model;

[0157] By local import:

[0158] In the Upload Model tab, click "Import model from local" (import model from library such as Figure 6 As shown), you can select the corresponding CAD model in the pop-up interface. Currently, stp, iges, obj, stl, etc. are supported;

[0159] After uploading is complete, you can click on the model, and the model will be automatically loaded into the display interface and displayed in the tree node below the root node;

[0160] After the CAD model is imported, the following operations can be performed on it: move, rotate, scale, create pipelines (create pipeline connections such as Figure 7 shown), copy, delete;

[0161] In the "Performance Mount" interface, you can mount the performance model to the geometric model. The performance model includes "Add CAE Model", "Add Reduced Order Model", "Add Business Algorithm", and "Add Monitoring Data". Click "Add CAE Model" under the geometric model node, and you can choose "Model Library Selection" or "Local Upload". If the user selects in the model library, you can find the corresponding name by pulling down and click "Submit" to complete the mounting. After clicking "Complete", it will prompt "Performance Mounting Successful" (the cloud map result after CAE simulation calculation is mounted is shown as follows) Figure 5 If the user selects "Local Upload", they need to fill in the "Name", select the calculation software from the drop-down list, and upload the calculation script;

[0162] The function of the reduced-order model tool module is to train the reduced-order model through the user's CAE model batch calculation or existing data as training samples, and can verify the accuracy of the reduced-order model and upload it to the library. Reduced-order model training requires the generation of training samples first. The training samples can come from the CAE batch calculation results. Select the CAE model in the library in the CAE drop-down option and enter the number of samples. The tool will automatically obtain the parameter name information. The user can edit the maximum and minimum values ​​of the parameters as needed, and can add or delete parameters on this basis. Click "Generate Sample Points" to generate a sample table based on the parameter range and number of samples set by the user. If the user has already generated a sample table, click "Import Sample Points" and the table will be automatically updated accordingly. To import data, you can compress the "data" folder into a zip format and upload it;

[0163] After the sample table and sample data are set up, you can train the reduced-order model. Select the "SVD" algorithm, set the "Orthogonal decomposition retention order", "Verification sample ratio", and "Maximum number of rows in the matrix block". If there are no special requirements, you can use the default values ​​and click "Train". The training time of the reduced-order model varies according to the dimension of the data, the number of nodes, and the number of nodes. For a node number of about 20,000 and a sample number of 100, the reduced-order model training based on 5 field data takes about 10 minutes. Based on the above data, you can estimate the actual training time required. After the training is completed, the sample list in the interface is updated and divided into "internal inspection samples" and "external inspection samples". "Internal inspection samples" refer to samples that participated in the training of the reduced-order model, and "external inspection samples" refer to samples that did not participate in the training of the reduced-order model. Therefore, it is relatively more reasonable to use "external inspection samples" for accuracy verification. Click on any row of data, and the platform will call the reduced-order model according to the sample parameters to predict the field data. After the prediction is completed, the "reduction result", "comparison interpolation result", and "original cloud map" will be displayed at the same time (comparison of the accuracy verification of the reduced-order algorithm and the CAE calculation results). Figure 8 Select the field data and click “CAE result view” (to view different field data as shown in the following example). Figure 9 After the reduced-order model training is completed, it can be compared with the corresponding CAE model, and you can click "Upload to Reduced-Order Model Library" to upload it directly to the model library (the reduced-order model upload is shown as Figure 10 shown).

[0164] The basic principles, main features, and advantages of the present invention are shown and described above. It should be understood by those skilled in the art that the present invention is not limited to the above-described embodiments. The above-described embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention, and such changes and modifications fall within the scope of the invention claimed.

Claims

1. A method for visual optimization design combining digital prototype construction and simulation data for underwater production systems of offshore oil projects, characterized in that: The following steps are involved: S1. Obtain offshore oil engineering project information from the system through the interface, and create a new project in the digital prototype platform based on the acquired project information; S2. Importing a 3D geometric model of an underwater production system of an offshore oil project and processing the 3D geometric model; S3, mounting the physical performance model on the processed geometric model, and correcting the quality factor during the mounting process; S4. Perform CAE simulation on the 3D geometric model, and import and display the CAE simulation results through the CAE lightweight rendering module; S5. Generate training samples based on simulation calculation results through the reduced-order model tool, and perform parameter adjustment verification on multiple working condition data to obtain the optimal design solution.

2. The method for combining digital prototype construction and simulation data with visual optimization design of an offshore oil engineering underwater production system according to claim 1 is characterized by: In the above S2, the three-dimensional geometric model of the underwater production system of the offshore oil project is imported and the three-dimensional geometric model is processed, which includes the following steps: S1.

1. Analyze the data format of the 3D geometric model of underwater production facilities, parse the STEP / IGES file, obtain the geometric entities, topological relationships, geometric construction information, material properties, and assembly information of the 3D model components, and convert different data formats into a unified data format through the prototype system format converter for import; S1.

2. Use lightweight rendering modules to optimize the display performance of 3D geometric models; S1.

3. After the 3D geometric model is imported, the coordinates and normal directions of the components of the 3D geometric model are analyzed to obtain the explosion direction of each component; S1.

4. Apply a specific explosion distance to make each component discrete relative to the view center; S1.

5. Adjust the display mode.

3. The method for combining digital prototype construction and simulation data with visual optimization design of an offshore oil engineering underwater production system according to claim 2 is characterized by: In said S1.3, analyzing the coordinates and normal directions of the components of the three-dimensional geometric model to obtain the explosion direction of each component includes the following steps: S1.

31. Analyze the local coordinate system of each component in the overall model and record the coordinates of the component's center point as a reference point. S1.

32. For each component, calculate its normal vector based on the geometric characteristics of its surface; S1.

33. Based on the assembly information, clarify the connection method between each component and record the relative position and direction between adjacent components; S1.

34. Use the normal vector of the component surface as a reference for the initial explosion direction; S1.

35. Adjust the explosion direction according to the assembly relationship between components; S1.36, transform the local explosion direction of each component into the global coordinate system; S1.

37. Use the force-directed layout algorithm to optimize the explosion direction of the components so that the components are evenly distributed in the exploded view.

4. The method for combining digital prototype construction and simulation data with visual optimization design of an offshore oil engineering underwater production system according to claim 3 is characterized by: In S1.37, a force-directed layout algorithm is used to optimize the explosion direction of components so that the components are evenly distributed in the exploded view, including the following steps: S1.

371. Take the center point of each component as its initial position and place it at the starting point in the global coordinate system; S1.

372. Assign a virtual "mass" attribute to each component, define the connection relationship between components, and assign virtual "spring" parameters to them; S1.

373. For parts connected by assembly relationships, calculate the attractive force; for any two parts, calculate the repulsive force; S1.

374. For each component, calculate the resultant of all attractive and repulsive forces acting on it; S1.

375. Calculate the displacement of each component based on the resultant force and update the position; S1.

376. If the position changes of all components are less than the preset threshold r, the optimization is completed, and the final position of each component after optimization is recorded.

5. The method for combining digital prototype construction with simulation data for visual optimization design of an offshore oil engineering underwater production system according to claim 1 is characterized by: In S3, the physical performance model is mounted on the processed geometric model, including the following steps: S3.

1. Divide the geometric model into different functional areas based on the functional requirements of the subsea production system. Determine the physical characteristics that need to be simulated in each area. Map the physical field data to the nodes of the geometric model using interpolation mapping formulas. S3.

2. For each component, identify its physical properties and record the component's working conditions; S3.

3. Bind each component to its corresponding physical model; S3.

4. According to the actual engineering data, set the material properties, initial conditions and boundary conditions for each component.

6. The method for combining digital prototype construction and simulation data with visual optimization design of an offshore oil engineering underwater production system according to claim 5 is characterized by: In S3.1, mapping the physical field data to the nodes of the geometric model through the interpolation mapping formula includes the following steps: S3.

11. Make it clear whether the physical field data is on the unit node; S3.12, extracting node coordinates and element topology information of the geometric model; S3.13, obtain the physical field data defined at the cell center; S3.

14. Traverse the unit topology information and record the unit list associated with each node; S3.

15. Calculate the physical field value of the node based on the physical field value of the unit to which the node belongs, and modify the weight in the physical field value calculation process; S3.

16. Update the calculated physical field values ​​to the nodes.

7. The method for combining digital prototype construction with simulation data for visual optimization design of an offshore oil engineering underwater production system according to claim 6 is characterized by: In S3.15, the physical field value of the calculation node is: Among them, D j Represents the physical field value of the unit; w j represents weight; T i represents the interpolated physical field value; N i Indicates the unit set to which the node belongs; j indicates the index of the unit; The weights used in the calculation of physical field values ​​are modified based on the quality factor of the element: Among them, T i1 Represents the physical field value after weight correction; q j Represents the quality factor of the unit; ‖P i -G i ‖ represents node P i and unit center G i The distance between i Represents a node; G i represents the unit center; p represents the smoothing parameter.

8. The method for combining digital prototype construction with simulation data for visual optimization design of an offshore oil engineering underwater production system according to claim 1 is characterized by: In S4, CAE simulation calculation results are imported and displayed through the CAE lightweight rendering module, including the following steps: S4.

1. Import the geometric model into the CAE simulation software, analyze the CAE result data format of the simulation software, parse the result data file, and obtain information; S4.

2. Create an unstructured grid data structure based on vtk, fill the acquired information into the structure, and save the acquired information as a vtp file; S4.

3. Based on the vtk.js library, the result file processed by the data conversion program is rendered using a lightweight rendering algorithm; S4.

4. After loading the result file, parse the model and call the corresponding API for drawing and rendering.

9. The method for combining digital prototype construction with simulation data for visual optimization design of an offshore oil engineering underwater production system according to claim 8, characterized in that: In S4.3, based on the vtk.js library, the result file processed by the data conversion program is rendered using a lightweight rendering algorithm, including the following steps: S4.

31. Use the API provided by vtk.js to load the .vtp file and parse the geometry and physics data. S4.32, dynamically adjust the model's level of detail based on viewpoint distance; S4.33, divide the geometry in the scene into multiple hierarchies, each hierarchy containing a bounding box; S4.34, use the Decimation algorithm to remove redundant vertices and faces; S4.

35. Apply the Phong lighting model to enhance rendering and map scalar data to color space.

10. The method for combining digital prototype construction with simulation data for underwater production system of offshore oil projects according to claim 1, characterized in that: In S5, a training sample is generated based on the simulation calculation results by using a reduced-order model tool, and parameter adjustment and verification of multiple working condition data are performed to obtain the optimal design solution, which includes the following steps: S5.

1. In the CAE model reduction tool module, select a CAE model from the library, enter the number of samples, and obtain parameter name information. The user can edit the maximum and minimum values ​​of the parameters as needed. S5.

2. Generate a sample table based on the parameter range and sample number set by the user; S5.

3. After the sample table and sample data are set up, the reduced-order model training is performed; S5.

4. After the reduced-order model training is completed, it is compared with the corresponding CAE model and uploaded to the model library.

Images