Digital twin model construction method and apparatus for power device, and computer device

Abstract

A digital twin model construction method and apparatus for a power device, and a computer device, relating to the technical field of power device modeling. The method comprises: determining components comprised in a target device in a component layer and parts comprised in the target device in a part layer; classifying the parts in the part layer to obtain a visual-level part list and a simulation calculation-level part list; establishing a corresponding first part model for the parts in the visual-level part list on the basis of a first modeling mode, and establishing a corresponding second part model for the parts in the simulation calculation-level part list on the basis of a second modeling mode; on the basis of a part connection relationship, connecting the first part model and the second part model to form component models corresponding to the components; and on the basis of a component connection relationship, connecting the component models to form a device model corresponding to the target device. The method achieves the modeling of a complex power device, improves the modeling efficiency, and also improves the applicability of the model in different usage scenarios.

Description

Method, device and computer equipment for constructing digital twin model of power equipment

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on January 3, 2024, with application number 202410015373.X and invention name “Method, device and computer equipment for constructing digital twin model of power equipment”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the technical field of power equipment modeling, and in particular to a method, device, computer equipment, and storage medium for constructing a digital twin model of power equipment. Background Art

[0003] With the increasing development and application of power grid digitalization technology, the demand for more refined scene visualization and equipment simulation modeling is increasing. 3D visualization rendering technology can bring model images closer to real-world spatial representation, allowing users to immerse themselves in a virtual world composed of 3D models and quickly grasp key information. As the underlying foundation of digital visualization software and a key input for finite element simulation calculations, models play a crucial role in the initial stages of digital twin construction. Traditional digital twin models consist of only single-precision component models. If all models were modeled 1:1 to match the actual scene, the model would become cumbersome, hindering the efficiency of model rendering and data simulation.

[0004] Summary of the Invention

[0005] The purpose of this application is to solve at least one of the above-mentioned technical defects, especially the technical defect in the existing technology that the model volume is large, which affects the efficiency of model rendering and data simulation.

[0006] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0007] In a first aspect, the present application provides a method for constructing a digital twin model of an electric power device, comprising:

[0008] Determine the components included in the component layer and the parts included in the part layer of the target device;

[0009] Classify the parts in the part layer to obtain a visualization-level parts list and a simulation-level parts list;

[0010] The parts in the visualization-level parts list are used to establish corresponding first part models according to the first modeling method, and the parts in the simulation-level parts list are used to establish corresponding second part models according to the second modeling method;

[0011] Connecting the first part model and the second part model into component models corresponding to each component according to the part connection relationship;

[0012] According to the component connection relationship, the component model is connected to the device model corresponding to the target device.

[0013] In one embodiment, the parts in the visualization-level parts list are used to establish corresponding first part models according to a first modeling method, and the parts in the simulation-level parts list are used to establish corresponding second part models according to a second modeling method, including:

[0014] Constructing three-dimensional white models of the first part model and the second part model respectively according to the geometric information of the parts;

[0015] In the three-dimensional model white mold of the second part model, an element having a size smaller than a first size threshold is determined as a target element;

[0016] Simplify the target element.

[0017] Texture rendering is performed on the white model of the three-dimensional model to obtain a first part model and a second part model.

[0018] In one embodiment, the method further includes: creating corresponding first part models for the parts in the visualization-level parts list according to a first modeling method, and creating corresponding second part models for the parts in the simulation-level parts list according to a second modeling method.

[0019] The first characteristic information of the part is associated with the first part model, and the second characteristic information of the part is associated with the second part model; wherein the first characteristic information is information related to the part and visual display, and the second characteristic information is characteristic information related to the part and simulation calculation.

[0020] In one embodiment, the target element is simplified to include:

[0021] If the target element is a line, determining whether there are other lines collinear with the line within the first preset range;

[0022] If so, extend the line to merge with other collinear lines;

[0023] If not, the line is removed and the two endpoints of the line are merged.

[0024] In one embodiment, the target element is simplified, including:

[0025] If the target element is a plane, determine the shape of the plane;

[0026] If the shape of the plane is a rectangle with an aspect ratio greater than a preset aspect ratio, determining whether there is another plane coplanar with the plane within a second preset range;

[0027] If so, extend the plane to merge with other coplanar planes;

[0028] If not, remove the two short sides of the plane and merge the two long sides of the plane;

[0029] If the plane is square, removing the plane retains and merges only the four endpoints of the plane.

[0030] In one embodiment, after classifying the parts in the part layer to obtain a visualization-level parts list and a simulation-level parts list, the method further includes:

[0031] For a part in the visualization-level parts list, if the scale of the part is smaller than the second size threshold, the part is deleted from the visualization-level parts list;

[0032] For a part in the simulation calculation level list, if the scale of the part is smaller than the third size threshold and the part is located in the preset ignore area, the part will be deleted from the simulation calculation level list; the third size threshold is smaller than the second size threshold.

[0033] In one embodiment, the preset ignored area includes an electrostatic shielding area, a mechanically negligible area and / or a temperature negligible area; wherein the electrostatic shielding area is an area in the target device that is electrostatically shielded, the mechanically negligible area is an area where the force applied to the target device is less than a preset force threshold when the target device is working, and the temperature negligible area is an area where the temperature is lower than the temperature threshold when the device is working.

[0034] In one embodiment, before classifying the parts in the part layer, the method further includes:

[0035] Identify duplicate parts within a part;

[0036] If a part has duplicates and a corresponding first part model or second part model has been created for the duplicate parts, the part will be ignored.

[0037] In one embodiment, identifying duplicate parts in a part includes:

[0038] Obtain the third characteristic information of the part;

[0039] Input the third feature information into the classifier to obtain the classification label of the part;

[0040] If there are parts with the same classification label, it is determined that there are duplicate parts.

[0041] In one embodiment, after connecting the component models into a device model corresponding to the target device according to the component connection relationship, the method further includes:

[0042] The corresponding traceability information is associated with the first part model, the second part model, the component module and the equipment model.

[0043] In one embodiment, after connecting the component models into a device model corresponding to the target device according to the component connection relationship, the method further includes:

[0044] Conduct modeling and evaluation of equipment models;

[0045] Repair the equipment model whose modeling evaluation results do not meet the requirements.

[0046] In one embodiment, the process of modeling and evaluating the device model includes:

[0047] Acquire images of each first part model in the device model at multiple preset viewing angles;

[0048] According to preset comparison items, the first part model is compared with an image of the corresponding part at the same preset viewing angle to obtain a score for each preset comparison item;

[0049] Based on the scores, a visualization-level modeling evaluation result of the device model is obtained.

[0050] In one embodiment, the process of modeling and evaluating the device model includes:

[0051] Sampling the annotated dimensions in the design drawings of the parts in the simulation calculation-level parts list to obtain multiple sampling standard dimensions;

[0052] Obtain the sampling model size of each sampling standard size marked position in the equipment model;

[0053] The modeling evaluation results of the simulation calculation level of the equipment model are obtained based on the error between each sampling model size and the corresponding sampling standard size.

[0054] In a second aspect, the present application provides a device for constructing a digital twin model of an electric power device, comprising:

[0055] A hierarchical division module is used to determine the components included in the component layer and the parts included in the part layer of the target device;

[0056] The classification module is used to classify the parts in the part layer to obtain the visualization level parts list and the simulation calculation level parts list;

[0057] A first modeling module is used to establish corresponding first part models for parts in the visualization-level parts list according to a first modeling method, and to establish corresponding second part models for parts in the simulation-level parts list according to a second modeling method;

[0058] A second modeling module is used to connect the first part model and the second part model into a component model corresponding to each component according to the part connection relationship;

[0059] The third modeling module is used to connect the component models into a device model corresponding to the target device according to the component connection relationship.

[0060] In a third aspect, the present application provides a computer device comprising one or more processors and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the steps of the method for constructing a digital twin model of power equipment in any of the above embodiments are executed.

[0061] In a fourth aspect, the present application provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the method for constructing a digital twin model of the power equipment in any of the above embodiments.

[0062] This solution utilizes a hierarchical modeling approach for the target device, dividing its model into the part, component, and device layers based on the assembly hierarchy. First, the component and part composition of the target device are determined. Each part is then categorized based on modeling requirements, with parts relevant only to visualization being placed in a visualization-level parts list and parts also relevant to simulation being placed in a simulation-level parts list. For the parts in these two lists, first and second part models with varying levels of accuracy are created. Based on the inter-part connectivity, these models are then connected to form component models. Finally, based on these inter-part connectivity relationships, the component models are integrated into a complete device model. This hierarchical modeling approach embodies the principle of progressive integration from the local to the global, enabling the modeling of complex power equipment and improving modeling efficiency. The hierarchical modeling approach creates models of varying resolutions based on actual modeling needs. This approach not only meets the need for rapid visualization of specific scenes but also can be used for complex, multi-physics, high-precision numerical simulations, ensuring both simulation accuracy and rendering speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present application 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0064] FIG1 is a flow chart of a method for constructing a digital twin model of power equipment provided by one embodiment of the present application;

[0065] FIG2 is a module diagram of a device for building a digital twin model of power equipment provided by one embodiment of the present application;

[0066] FIG3 is a diagram showing the internal structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0068] The present application provides a method for constructing a digital twin model of power equipment, please refer to Figure 1, including steps S102 to S108.

[0069] S102: Determine the components included in the component layer and the parts included in the part layer of the target device.

[0070] It can be understood that the target device refers to the power equipment for which a digital twin model is required. Currently, there is a demand for digital modeling of complex power transmission and transformation equipment in power systems. Therefore, power equipment can include transformers, reactors, and other components. The following description uses a transformer as an example. This embodiment divides the target device model into multiple layers based on the assembly level: the parts layer, the component layer, and the device layer. The parts layer is the component unit and the smallest functional unit in the target device. The component layer is the main functional unit of the device and is typically composed of multiple parts to achieve a complete function. Of course, some components can also be made of integrally molded parts. The device layer is the complete model of the target device. In a transformer, the component layer includes windings, cores, and leads. The windings in the parts layer include high-voltage, medium-voltage, low-voltage, and voltage-regulating windings, as well as coil terminals. The core includes the core, upper yoke, lower yoke, side yoke, core plate, core plate, and spacers. The leads include high-voltage, high-voltage neutral, medium-voltage, medium-voltage neutral, low-voltage, and voltage-regulating leads. In addition, the parts layer also includes connectors and supports, etc. This step analyzes the structure of the target device in combination with various modeling requirements, clarifies the level and granularity of digital modeling, lays the foundation for subsequent modeling, helps manage model complexity, and makes modeling more targeted.

[0071] In this step, the modeler gathers design data such as the target device's structural diagram, assembly drawings, and parts drawings to understand its structure. Based on the device's operating principle, the modeler analyzes its various functional modules, determines the component layer, and generates a parts list. The modeler then further decomposes the internal structure of each component in the parts list, identifies its components, and generates a parts list. The parts list and components list are then stored in a computer.

[0072] S104, classifying the parts in the part layer to obtain a visualization-level parts list and a simulation-calculation-level parts list.

[0073] Understandably, when using digital twin models, the visual display and the part models used for precise simulation calculations have different levels of detail, resulting in different computing power requirements for rendering and simulation. Traditional techniques only model parts at a uniform level of detail. This results in low accuracy for simulation calculations using device models composed of visualization-level parts, while device models composed of simulation-level parts take a long time to render when used for visual display. However, research has revealed that not every part in the part layer is relevant for precise simulation calculations. Based on the target device's operating principle and simulation requirements, the parts list can be divided into two categories: a visualization-level parts list and a simulation-level parts list. Parts included in the visualization-level parts list are only relevant for visualization and have little impact on simulation calculations. Parts included in the simulation-level parts list are relevant for both visualization and simulation calculations. Taking a transformer as an example, among the parts corresponding to the windings at the component layer, the high-voltage and medium-voltage coils are at the simulation-level, while the low-voltage coil, voltage regulating coil, and coil outlet are at the visual-level. Among the core components in the component layer, the core, upper yoke, lower yoke, side yoke, and side column are at the precise calculation level, while the core plate, core plate, and spacer are at the visual level. Among the lead components in the component layer, the high-voltage lead, high-voltage neutral lead, medium-voltage lead, and medium-voltage neutral lead are at the precise calculation level, while the low-voltage lead and voltage regulator lead are at the visual level. Furthermore, connectors and support components in the component layer are also at the visual level.

[0074] S106 , establishing corresponding first part models for the parts in the visualization-level parts list according to the first modeling method, and establishing corresponding second part models for the parts in the simulation-level parts list according to the second modeling method.

[0075] It's understandable that the modeling of parts in a visualization-level parts list needs to intuitively display the target device's external structure. Furthermore, the simulation model, in addition to providing a visual display, needs to simulate the target device's operating conditions in multiple physical fields, and some display details may be modified to achieve higher simulation accuracy. Specifically, the simulation calculation involves inputting the target device's operating parameters into the simulation model, which then generates simulation results for the target device under physical fields such as electric and magnetic fields, temperature, and force. Therefore, for parts in different lists, first and second part models with different levels of detail need to be established.

[0076] S108 , connecting the first part model and the second part model into a component model corresponding to each component according to the component connection relationship.

[0077] It can be understood that the part connection relationship refers to the assembly relationship between different parts in the same component, including positioning constraints, isotropic limits, assembly dimensions, etc. These connection relationships follow the actual assembly requirements to determine the spatial position and limit conditions between the parts. The assembly and positioning requirements of each part in the same component can be determined based on the component design drawings of the target device, and its part connection relationship can be analyzed. In the digital modeling software, positioning constraints are added between different part models, assembly positions are adjusted, limit conditions are applied, etc. according to the determined part connection relationship to simulate physical assembly. In addition, some parts are independent parts, that is, one part can be used as a complete component, such as one-piece connectors, fixings, etc. These first part models and second part models can be marked with independent part identification when modeling, and can be directly upgraded from the part layer to the component layer in step S108 to become corresponding component models. The composition of the component model will include the first part model and / or the second part model.

[0078] This solution utilizes a hierarchical modeling approach for the target device, dividing its model into the part, component, and device layers based on the assembly hierarchy. First, the component and part composition of the target device are determined. Each part is then categorized based on modeling requirements, with parts relevant only to visualization being placed in a visualization-level parts list and parts also relevant to simulation being placed in a simulation-level parts list. For the parts in these two lists, first and second part models with varying levels of accuracy are created. Based on the inter-part connectivity, these models are then connected to form component models. Finally, based on these inter-part connectivity relationships, the component models are integrated into a complete device model. This hierarchical modeling approach embodies the principle of progressive integration from the local to the global, enabling the modeling of complex power equipment and improving modeling efficiency. The hierarchical modeling approach creates models of varying resolutions based on actual modeling needs. This approach not only meets the need for rapid visualization of specific scenes but also can be used for complex, multi-physics, high-precision numerical simulations, ensuring both simulation accuracy and rendering speed.

[0079] In one embodiment, the parts in the visualization-level parts list are used to establish corresponding first part models according to a first modeling method, and the parts in the simulation-level parts list are used to establish corresponding second part models according to a second modeling method, including:

[0080] (1) Construct three-dimensional white models of the first part model and the second part model according to the geometric information of the parts.

[0081] The reference data for building a part model is the part's geometric information. Geometric information describes spatial information such as the part's shape and dimensions. Generally, it can be point cloud data obtained by scanning the part. While both the first and second part models are constructed based on geometric information, different fidelity and simplification strategies are employed to meet the requirements of visualization and accurate calculation, respectively. A 3D model white mold contains only 3D geometric shape information, without rendering information such as material and color. Based on the geometric information of the same part, a 3D model white mold is first created that reflects the outline of the target device. During the 3D model white mold construction process, since both levels of the model need to display their appearance, the constructed 3D model white molds can be identical. Alternatively, the 3D model white molds can be differentiated based on the computing power of the computer. For example, the 3D model white mold corresponding to the first part model can be constructed by performing appropriate surface fitting and simplification while maintaining the main outline to reduce the complexity of the 3D surface. The 3D model white mold corresponding to the second part model can be constructed to preserve the various details of the target device as much as possible.

[0082] (2) In the three-dimensional model white mold of the second part model, elements with a scale smaller than a first size threshold are determined as target elements.

[0083] It can be understood that the scale is used to represent the size of the elements of the three-dimensional model. The scale of the surface element can be represented by the length of its smallest side. The scale of the line element can be directly represented by its side length. In the process of generating the three-dimensional model white mold, there may be some elements with smaller scales that will affect the finite element simulation analysis. These elements need to be simplified to eliminate the impact on the simulation accuracy. Therefore, before rendering, the elements (lines and planes) in the three-dimensional model white mold are traversed and checked according to the first size threshold, and the elements with a scale smaller than the first size threshold are determined as elements that need to be simplified. The first size threshold here can generally be set to 0.1mm.

[0084] (3) Simplify the target elements.

[0085] After the target element is determined, corresponding simplified means can be used to process it according to the type and shape of the target element.

[0086] (4) Texture rendering is performed on the white model of the three-dimensional model to obtain the first part model and the second part model.

[0087] It is understood that after the 3D model white mold is constructed, materials, colors, etc. can be added to the surface of the 3D model white mold to achieve rendering of the 3D model white mold. The texture data used for rendering can be selected and generated in the rendering software based on the material of the target device.

[0088] In one embodiment, a first part model is created for each part in the visualization-level parts list using a first modeling method, and a second part model is created for each part in the simulation-level parts list using a second modeling method. The method further includes associating first feature information of the part with the first part model, and associating second feature information of the part with the second part model. The first feature information is information related to the visualization of the part, and the second feature information is feature information related to the simulation.

[0089] It can be understood that establishing an association here refers to adding and storing one-to-one corresponding information to the part models. After the first and second part models are established, corresponding feature information needs to be added to the models to meet the needs of their respective applications. Geometric information only describes the shape of the parts; geometric features alone are insufficient to support the different uses of the models. Adding various types of feature information can enrich the semantic description of the models, enabling the models to support different functional applications. Therefore, it is necessary to add targeted feature information for the visualization-level model and the simulation-level model. Specifically, the first feature information can include geometric information and material properties. The second feature information includes geometric information, spatial relationships, material properties, number of models, operating conditions, precise calculation drawings, test plans, insulation criteria, etc. Since the primary purpose of the visualization model is to display the product's appearance and structure, its feature information only needs to include geometric information and material properties related to the appearance for real-life rendering. Simulation models, on the other hand, require precise numerical analysis, so their feature information needs to be as complete and detailed as possible. This includes geometric information, spatial relationships, material properties, number of models, operating conditions, precise calculation drawings, test plans, insulation criteria, and other information that are crucial for simulation calculations. In some embodiments, the naming method for storing feature information can adopt a specific coding method, for example, XY.YY-AN, where X is the first letter of the English name of the part, Y is the first letter of the English name of the part with which it has a part connection relationship, A is the first letter of the English name of the material of the part, and N is the number of the part model under the same type of model.

[0090] In one embodiment, the target element is simplified, including:

[0091] (1) If the target element is a line, determine whether there are other lines collinear with the line within the first preset range.

[0092] It can be understood that the first preset range is the search range set when judging whether there are collinear lines with the target element as the center. This step is for the lines that need to be simplified in the three-dimensional model. If there are other collinear lines nearby, it means that the line may be part of other lines, but it is disconnected due to an error during rendering. Therefore, if other collinear lines can be searched within the first preset range, the line is directly extended to merge it with the other collinear lines into a complete line. Otherwise, it means that the line is a redundant line and can be removed directly. The two endpoints left after the line is removed can be merged into one endpoint to prevent point loss.

[0093] (2) If so, extend the line to merge with other collinear lines.

[0094] (3) If not, remove the line and merge the two endpoints of the line.

[0095] In one embodiment, the target element is simplified, including:

[0096] (1) If the target element is a plane, determine the shape of the plane.

[0097] It is understandable that for plane elements, different shapes have different simplification strategies. Therefore, it is necessary to first determine the specific shape of the plane.

[0098] (2) If the shape of the plane is a rectangle with an aspect ratio greater than a preset aspect ratio, it is determined whether there is another plane coplanar with the plane within a second preset range.

[0099] It can be understood that a rectangle with an aspect ratio greater than the preset aspect ratio is a narrow facet. This means that the facet may be part of another plane, disconnected due to a rendering error, or that the line was incorrectly copied to form an extra plane. Therefore, if other coplanar planes can be found within the second preset range, the plane is directly extended and merged with the other planes to form a complete plane. Otherwise, the line is a superfluous plane, and the two short sides are removed and the two long sides of the plane are merged.

[0100] (3) If so, extend the plane to merge with other coplanar planes.

[0101] (4) If not, remove the two short sides of the plane and merge the two long sides of the plane.

[0102] (5) If the shape of the plane is a square, remove the plane and only retain and merge the four endpoints of the plane.

[0103] It can be understood that the influence of the small square surface on the simulation calculation is negligible. In order to simplify the modeling, it can be directly removed and only the four endpoints of the plane are retained and merged.

[0104] In one embodiment, in order to simplify the modeling process and speed up the modeling efficiency, some parts with less impact in the part layer can be ignored. After classifying the parts in the part layer and obtaining the visualization level parts list and the simulation calculation level parts list, the following is also included:

[0105] (1) For a part in the visualization-level parts list, if the scale of the part is smaller than the second size threshold, the part is deleted from the visualization-level parts list.

[0106] It's understandable that the first part model is a visualization-level model. During visualization, parts that are too small generally don't require attention during display, and simplifying them won't affect the presentation. Therefore, during the traversal and modeling of parts at the part layer, if a traversed part is found to be smaller than the second size threshold, that part can be omitted from modeling. Generally, the second size threshold can be set to 10mm. By specifying the second size threshold, it's possible to automatically determine which smaller parts can be excluded from modeling, thereby simplifying the visualization modeling process and improving modeling efficiency.

[0107] (2) For a part in the simulation calculation level list, if the size of the part is smaller than the third size threshold and the part is located in the preset ignore area, the part is deleted from the simulation calculation level list. The third size threshold is smaller than the second size threshold.

[0108] It can be understood that the second part model is a precise simulation-level model. During simulation, in addition to being sufficiently small, the part must also be within a pre-set ignore region that has no impact on the finite element simulation. This ignore region is determined by the physical fields to be simulated by the simulation-level model. For example, a transformer's operation involves multiple physical fields, including electric, magnetic, temperature, and mechanical fields. Therefore, regions that do not affect the distribution of these physical fields can be designated as ignore regions. Specifically, the pre-set ignore regions include electrostatic shielding regions, mechanical ignore regions, and / or temperature ignore regions. The electrostatic shielding region is an area of ​​the target device that is electrostatically shielded. Due to the electrostatic shielding, components within the electrostatic shielding region have no impact on the electric field distribution or the device's insulation state, and therefore can be ignored. The mechanical ignore region is an area of ​​the target device where forces below a pre-set threshold are applied during operation. These areas experience no stress concentration, or even if stress concentration occurs, it is significantly below the material damage threshold, so simplifying the details will not affect the device's mechanical performance verification. The temperature ignore region is an area where the device's operating temperature is below a threshold. The temperature in the negligible temperature area is low and has little impact on the temperature distribution. After simplification, it will not affect the temperature field distribution verification calculation of the equipment.

[0109] In one embodiment, before classifying the parts in the part layer, the method further includes:

[0110] (1) Identify duplicate parts in a part.

[0111] It is understood that a duplicate part refers to a part that is exactly the same as the current part.

[0112] (2) If there are duplicate parts and the corresponding first part model or second part model has been established for the duplicate parts, the parts will be ignored.

[0113] During part modeling, duplicate parts may require repeated modeling. First, check whether the part is a duplicate of a previous part. If so, further check whether corresponding visualization and computational models have already been established for the duplicate part. If so, there's no need to create duplicate models. Instead, simply reuse the existing models for the first and second parts. Skipping modeling avoids duplication and significantly improves efficiency.

[0114] In one embodiment, identifying duplicate parts in a part includes:

[0115] (1) Obtain the third characteristic information of the part.

[0116] It is understood that the third feature information is feature information used for part classification and identification, and may include images of the part from multiple perspectives, the part number and model number when purchased, and the like.

[0117] (2) Input the third feature information into the classifier to obtain the classification label of the part.

[0118] It can be understood that a classifier is a machine learning model that maps feature information to category output.

[0119] (3) If there are parts with the same classification label, it is determined that there are duplicate parts. For the case where the third feature information includes an image, the model can be based on a convolutional neural network. The entire network architecture consists of several convolutional layers, pooling layers, and fully connected layers. The input image passes through multiple convolutional layers, and after the pooling layer, the final result is output through the fully connected layer. The multiple convolutional layers and pooling layers can extract the deep abstract features of the image layer by layer. Finally, all the abstract features are combined through the fully connected layer for calculation, and a complex nonlinear mapping relationship between the abstract features and the target output is established to realize the classification function. As for the serial numbers when purchasing parts, the corresponding relationship can be directly established using semantics, etc., which will not be elaborated here.

[0120] In one embodiment, after connecting the component model to the device model corresponding to the target device according to the component connection relationship, the method further includes associating the corresponding traceability information with the first part model, the second part model, the component module and the device model.

[0121] It can be understood that the traceability information of a product reflects various historical data of the entire life cycle of the product, from design, manufacturing to use. Traceability information is very important for ensuring product quality and process control. Digitizing the traceability information of the target device and associating it with the first device model and the second device model can achieve persistent storage and effective utilization of the traceability information. Traceability information can include digital QR codes, processing technology, material sources, production batches, responsible persons, etc. Since the production of equipment is carried out in units of parts, the traceability information needs to be associated with the first device model and the second device model here, and the traceability information needs to be associated with each part model. The traceability information is injected after the equipment model is established because some parts are ignored and defaulted during the modeling process. Executing it after the final equipment layer modeling is completed can reduce the workload of traceability information statistics.

[0122] In one embodiment, after connecting the component models into a device model corresponding to the target device according to the component connection relationship, the method further includes:

[0123] (1) Modeling and evaluation of equipment models.

[0124] It is understood that a special evaluation is required for the constructed first device model and the second device model to determine whether the constructed models meet the use requirements. Different evaluation methods can be used for the two models with different accuracy levels.

[0125] (2) Repair the equipment model whose modeling evaluation results do not meet the requirements.

[0126] Based on the evaluation feedback, if the model is found to have issues that do not meet the requirements, targeted model repair is required. Repair methods may include model reconstruction, parameter adjustment, and supplementary information. The goal of repair is to eliminate the issues found in the evaluation and meet the quality requirements of the model.

[0127] In one embodiment, the process of modeling and evaluating the device model includes:

[0128] (1) Obtain images of each first part model in the equipment model at multiple preset viewing angles.

[0129] It's understandable that for visualization-level models, the quality of the model is determined by its visual relevance to the target device. By setting multiple viewpoints to cover the main portion of the part, we captured images of the part and the real scene from the same preset viewpoints, and took screenshots of the first part model to construct an image dataset for comparison.

[0130] (2) According to the preset comparison items, the first part model is compared with the image of the corresponding part at the same preset viewing angle to obtain the score of each preset comparison item.

[0131] As you can understand, the preset comparison items refer to the items that require comparison between the screenshot of the visual model and the corresponding real-life image. In addition to the standard visual similarity, additional checks can be added as needed, such as perspective correctness, component interference, and completeness. For each preset comparison item, a quantitative indicator is generated to reflect the performance of the first part model in that item. Finally, the scores for each preset comparison item are combined to obtain the modeling evaluation results for the visualization-level model.

[0132] (3) Based on the scores, the visualization-level modeling evaluation results of the equipment model are obtained.

[0133] In one embodiment, the process of modeling and evaluating the device model includes:

[0134] (1) Sampling the marked dimensions in the design drawings of the parts in the simulation calculation level parts list to obtain multiple sampling standard dimensions.

[0135] Sampling standard dimensions refer to the standard dimensions selected from the dimensions marked on the design drawings of the target equipment parts according to the sampling method. Super Latin cube sampling can be used for sampling.

[0136] (2) Obtain the sampling model dimensions at the locations marked with each sampling standard dimension in the equipment model.

[0137] It can be understood that in order to compare whether the dimensions on the equipment model are consistent with the dimensions required on the actual drawing, it is necessary to calculate the dimensions of these locations in the model at the corresponding positions in the equipment model and form data pairs with the standard dimensions.

[0138] (3) Based on the error between each sampling model size and the corresponding sampling standard size, the modeling evaluation result of the simulation calculation level of the equipment model is obtained.

[0139] Finally, the error of each pair of data can be calculated and the error distribution can be analyzed to judge the geometric accuracy of the model based on the error. The error can be calculated in the form of residuals.

[0140] This application provides a device for constructing a digital twin model of an electric power device, as shown in FIG2 , including:

[0141] The hierarchical division module 210 is used to determine the components included in the component layer and the parts included in the part layer of the target device.

[0142] The classification module 220 is used to classify the parts in the part layer to obtain a visualization-level parts list and a simulation-calculation-level parts list.

[0143] The first modeling module 230 is used to establish corresponding first part models for the parts in the visualization-level part list according to a first modeling method, and to establish corresponding second part models for the parts in the simulation-level part list according to a second modeling method.

[0144] The second modeling module 240 is used to connect the first part model and the second part model into a component model corresponding to each component according to the part connection relationship.

[0145] The third modeling module 250 is used to connect the component models into a device model corresponding to the target device according to the component connection relationship.

[0146] For the specific definition of the digital twin model construction device of the electric power equipment, please refer to the definition of the digital twin model construction method of the electric power equipment mentioned above. Each module in the digital twin model construction device of the above-mentioned electric power equipment can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0147] The present application provides a computer device comprising one or more processors and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the steps of the method for constructing a digital twin model of power equipment in any of the above embodiments are executed.

[0148] Schematically, as shown in FIG3 , FIG3 is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. Referring to FIG3 , the computer device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by a memory 301 for storing instructions that can be executed by the processing component 302, such as an application. The application stored in the memory 301 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to perform the steps of the method for constructing a digital twin model of power equipment in any of the above embodiments.

[0149] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300 , a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305 .

[0150] Those skilled in the art will understand that the structure shown in FIG3 is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0151] The present application provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the method for constructing a digital twin model of power equipment in any of the above embodiments.

[0152] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referenced to each other.

[0153] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a digital twin model of a power device, characterized in that, Including: Determine the components included in the target device at the component level and the parts included at the part level; Classify the parts in the part layer to obtain a visual-level part list and a simulation calculation-level part list; Build corresponding first part models for the parts in the visual-level part list according to the first modeling method, and build corresponding second part models for the parts in the simulation calculation-level part list according to the second modeling method; Connect the first part models and the second part models into component models corresponding to each component according to the part connection relationship; Connect the component models into a device model corresponding to the target device according to the component connection relationship.

2. The method for constructing a digital twin model of a power device according to claim 1, wherein The step of building corresponding first part models for the parts in the visual-level part list according to the first modeling method and building corresponding second part models for the parts in the simulation calculation-level part list according to the second modeling method includes: Construct three-dimensional model white models of the first part model and the second part model respectively according to the geometric information of the parts; In the three-dimensional model white model of the second part model, determine the elements with a scale smaller than the first size threshold as target elements; Simplify the target elements; Perform texture rendering on the three-dimensional model white model to obtain the first part model and the second part model.

3. The method for constructing a digital twin model of the power equipment according to claim 2, wherein The step of building corresponding first part models for the parts in the visual-level part list according to the first modeling method and building corresponding second part models for the parts in the simulation calculation-level part list according to the second modeling method further includes: Associate the first feature information of the part with the first part model, and associate the second feature information of the part with the second part model; wherein, the first feature information is the information related to visual display of the part, and the second feature information is the feature information related to simulation calculation of the part.

4. The method for constructing a digital twin model of a power device according to claim 2, wherein The step of simplifying the target elements includes: If the target element is a line, determine whether there are other lines collinear with the line within the first preset range; If so, extend the line to merge with other collinear lines; If not, remove the line and merge the two endpoints of the line.

5. The method for constructing a digital twin model of a power device according to claim 2, wherein The step of simplifying the target elements includes: If the target element is a plane, determine the shape of the plane; If the shape of the plane is a rectangle with an aspect ratio greater than the preset aspect ratio, determine whether there are other planes coplanar with the plane within the second preset range; If so, extend the plane to merge with other coplanar planes; If not, remove the two short sides of the plane and merge the two long sides of the plane; If the shape of the plane is a square, remove the plane and only retain and merge the four endpoints of the plane.

6. The method for constructing a digital twin model of a power device according to claim 1, wherein After classifying the parts in the part layer to obtain a visual-level part list and a simulation calculation-level part list, it further includes: For the parts in the visual-level part list, if the scale of a part is smaller than the second size threshold, delete the part from the visual-level part list; For the parts in the simulation calculation-level list, if the scale of a part is smaller than the third size threshold and the part is located in a preset ignored area, delete the part from the simulation calculation-level list; the third size threshold is smaller than the second size threshold.

7. The method for constructing a digital twin model of a power device according to claim 6, characterized in that, The preset ignored area includes an electrostatic shielding area, a mechanical negligible area, and / or a temperature negligible area; wherein, the electrostatic shielding area is the area shielded by static electricity in the target device, the mechanical negligible area is the area where the force received during the operation of the target device is less than the preset force threshold, and the temperature negligible area is the area where the temperature of the device is lower than the temperature threshold during operation.

8. The method for constructing a digital twin model of a power device according to claim 1, wherein Before classifying the parts in the part layer, it further includes: Identifying duplicate parts among the parts; If there are duplicate parts among the parts and the corresponding first part model or second part model has been established for the duplicate parts, ignore the parts.

9. The method for constructing a digital twin model of the power equipment according to claim 8, wherein, The identifying duplicate parts among the parts includes: Obtaining the third feature information of the parts; Inputting the third feature information into a classifier to obtain the classification labels of the parts; If there are parts with the same classification labels, determine that there are duplicate parts among the parts.

10. The method for constructing a digital twin model of a power device according to claim 1, characterized in that, After connecting the component models into the device model corresponding to the target device according to the component connection relationship, it further includes: Associating the corresponding traceability information with the first part model, the second part model, the component module, and the device model.

11. The method for constructing a digital twin model of a power device according to claim 1, characterized in that, After connecting the component models into the device model corresponding to the target device according to the component connection relationship, it further includes: Performing a modeling evaluation on the device model; Repairing the device model whose modeling evaluation result does not meet the requirements.

12. The method for constructing a digital twin model of a power device according to claim 11, wherein, The process of performing a modeling evaluation on the device model includes: Obtaining images of each of the first part models in the device model from multiple preset perspectives; Comparing the first part model with the image of the corresponding part from the same preset perspective according to the preset comparison items to obtain the scores of each preset comparison item; Obtaining the modeling evaluation result at the visual level of the device model according to each score.

13. The method for constructing a digital twin model of a power device according to claim 11, wherein The process of performing a modeling evaluation on the device model includes: Sampling the marked dimensions in the design drawings of the parts in the simulation calculation-level part list to obtain multiple sampled standard dimensions; Obtaining the sampled model dimensions at the positions marked by each of the sampled standard dimensions in the device model; Obtaining the modeling evaluation result at the simulation calculation level of the device model according to the error between each sampled model dimension and the corresponding sampled standard dimension.

14. A digital twin model construction device for a power equipment, characterized in that, It includes: A hierarchical division module for determining the components included in the component layer of the target device and the parts included in the part layer; A classification module for classifying the parts in the part layer to obtain a visual-level part list and a simulation calculation-level part list; The first modeling module is used to establish corresponding first part models for the parts in the visualization-level part list according to a first modeling method, and establish corresponding second part models for the parts in the simulation calculation-level part list according to a second modeling method; The second modeling module is used to connect the first part model and the second part model into component models corresponding to the components according to the part connection relationship; The third modeling module is used to connect the component models into a device model corresponding to the target device according to the component connection relationship.

15. A computer device, characterized in that, It includes one or more processors and a memory. Computer-readable instructions are stored in the memory. When the computer-readable instructions are executed by the one or more processors, the steps of the method for constructing a digital twin model of the power device according to any one of claims 1-13 are executed.

16. A storage medium, characterized in that, Computer-readable instructions are stored in the storage medium. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the method for constructing a digital twin model of the power device according to any one of claims 1-13.

Images