A processor for constructing a three-dimensional model and a three-dimensional model connection early warning system

Abstract

The application relates to a processor for constructing a three-dimensional model and a three-dimensional model connection early warning system; the processor is configured to: identify coordinate information of specific components of physical objects with pointing information in a two-dimensional drawing; determine physical correlations of several three-dimensional models respectively according to the pointing information and the coordinate information; analyze specific components of mutual pairing of the several three-dimensional models which have physical correlations with each other according to the physical correlations of the three-dimensional models; and establish data intercommunication according to physical relations among entities corresponding to the three-dimensional models, so that the physical correlations of the three-dimensional models are consistent with the correlations of entity equipment. The three-dimensional model automatically generated by the application has consistent attribute data with attribute data of the entity, and the correlation relation is also consistent with the physical correlations of the entity, so that staff members do not need to adjust attribute data of the three-dimensional model and the correlation relation in the later period, and only need to check the data to be confirmed.

Description

[0001] The original basis for this divisional application is patent application No. (202311702676.0), filed on December 12, 2023, entitled "A digital twin method, system and storage medium for an engineering object". Technical Field

[0002] This invention relates to the field of digital twin data processing and construction technology, and in particular to a processor for constructing three-dimensional models and a three-dimensional model connection early warning system. Background Technology

[0003] In building a 3D model of a digital twin factory, it is necessary to construct 3D equipment data from a large amount of 2D drawings, and the equipment in the 3D model must be consistent with the equipment in the actual 3D space. However, existing 3D model generation software has a flaw: while the connection relationships of the equipment and pipelines in the 3D model appear correct, deviations in the physical connections of the actual equipment can occur during the connection process. For example, physical connections of actual equipment require sealing to prevent fluid leakage, resulting in discrepancies between the physical height of the pipeline and that in the 3D model; or the angles of the physical connections of the actual equipment may deviate from those in the 3D model. This phenomenon is caused by the lack of data communication between the 3D model and the actual space.

[0004] For example, patent application CN113919022A discloses a method for generating 3D pipeline data based on the Esri platform, including the following steps: S1, checking the original pipeline survey data, deleting empty data, and ensuring the correctness of the data coordinate system; S2, generating 2D pipeline data and 2D fitting data with z-values ​​based on the attribute table of the original pipeline survey data; S3, generating 3D pipelines and 3D fittings based on the 2D pipeline data and 2D fitting data with z-values. This technical solution can only generate 3D pipelines and fittings based on the 2D data obtained from the survey, but it is limited to reproducing related entity objects within 3D graphics modeling software and does not involve the reproduction of the attribute data of the 3D model. The attribute data of the 3D model includes, for example, connectivity, pressure, material, and the medium contained within it. This technical solution is suitable for the rapid generation of isolated 3D components, but it has no practical application value for digital twins of factories. Because there are already many existing technologies for the rapid generation of isolated 3D models, such as PDMS, which can quickly generate 3D models with attribute data based on DB files, and these 3D models can also be generated in a certain 3D space with accurate coordinates.

[0005] Rapidly generating 3D models does not mean that factory digital twins can be automated. This is because a large number of 3D models need to be digitally interconnected to achieve true physical twins, i.e., reproducing the physical connections between pipelines, the pressure data correlation between valves and flanges, and the control relationships between control units and controlled objects. For example, when using PDMS to generate two interconnected 3D models, although these two models have a consistent interface in coordinates, they do not achieve data exchange in 3D space. For instance, the inlet coordinates of the tank in the first 3D model are E389650N 320100U 854490, and the outlet coordinates of the pipeline in the second 3D model connected to the first 3D model are E 391210N 320100U 855244.5. However, the unprocessed first and second 3D models are still separate independent objects in this 3D space, and their physical properties still need to be manually set. For example, in PDMS, the physical relationships between two 3D models need to be manually set first, such as the connection relationships and directions of media flow and data flow. Here, it is used as an example to define the start and end points of the pipeline in the second 3D model. By selecting the media flow start point (i.e., outlet) of the storage tank and the start point (i.e., inlet) of the pipeline and designating them as media start points respectively, data communication between the two 3D models can be achieved, thereby realizing a digital twin of the physical relationships between the relevant 3D models. For the digital twin work of a factory, the above work means a huge investment of manual labor, which is not only inefficient but also has a high error rate, and requires complex post-processing verification.

[0006] Based on the problems mentioned above in the construction process of digital twin factories, this invention aims to provide a digital twin method for engineering objects, so that the generated three-dimensional model is an applicable three-dimensional model with physical attributes and digital connectivity, reducing the workload of staff in setting the physical attribute data of the three-dimensional model, and also reducing the workload of subsequent model verification work.

[0007] Furthermore, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the applicant studied a large number of documents and patents when making this invention, but due to space limitations, not all details and contents were listed in detail. However, this does not mean that the present invention does not possess the features of these prior art. On the contrary, the present invention already possesses all the features of the prior art, and the applicant reserves the right to add relevant prior art to the background art. Summary of the Invention

[0008] To quickly generate digital twins of engineering objects from a large number of 2D drawings, a novel digital twin method is needed. This method should be able to generate engineering objects in a 3D scene at least semi-automatically, thereby improving the efficiency of digital twin generation and reducing the error rate. Preferably, this digital twin method can pre-verify the physical relationships between 3D models generated from 2D drawings according to preset rules. Preferably, the pre-verification is implemented using a form for secondary manual review. In existing digital twin processes, the focus is mainly on the mapping relationship between physical and virtual objects to construct a high-fidelity model in virtual space that is consistent with the physical world. For example, patent document CN115601514A discloses an automatic association mapping method for digital twin data. This method includes acquiring building information of each physical building; standardizing and storing the building information in a building information record table based on BeiDou grid location codes; constructing a virtual space and generating virtual space location codes for each virtual building model based on BeiDou grid location codes; generating a building model information data table based on the virtual space location codes and the model IDs of the virtual building models; automatically fusing and mapping the building information record table and the building model information data table based on BeiDou grid location codes to generate a spatial association index table; optimizing the spatial association index table; and performing interaction between physical and virtual spaces based on the spatial association index table. This technical solution primarily uses a unified BeiDou grid location code to build a bridge between the building information of each physical building and each virtual building model. It then automatically fuses and maps the spatial distance between the virtual space location codes of the physical buildings and virtual building models to obtain associated data records, thereby generating a spatial association index table. This allows physical entities to directly establish associations with virtual models. However, the relationships between physical buildings and virtual building models in this technical solution can only be reflected through spatial location encoding similar to coordinate information. It cannot synchronously establish and reflect the relationships between physical objects on the virtual model. This technical solution cannot solve the problem of digital twins of engineering objects with a large number of process connections in a factory environment.

[0009] To address the shortcomings of existing technologies, this invention provides a digital twin method for engineering objects. The method includes: identifying two-dimensional drawings of physical objects involving entity relationships and generating three-dimensional model information corresponding to the physical objects; and generating a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings based on the three-dimensional model information. The three-dimensional model reconstruction data file is used to import the three-dimensional model information of the physically related physical objects into three-dimensional modeling software, thereby automatically generating a three-dimensional model corresponding to the three-dimensional model information. Compared to existing technologies that generate a single three-dimensional model based on a single two-dimensional drawing, this invention can process a large number of two-dimensional drawings in batches to form a preliminary three-dimensional representation of the engineering objects.

[0010] Existing technologies already include digital twin solutions for numerous process procedures involved in process manufacturing industrial production. For example, patent document CN115495880A discloses a method for constructing a digital twin model of a process manufacturing workshop. First, the geometric and physical attributes of various elements in the process manufacturing physical workshop are modeled using industrial modeling software to construct a static physical model (SPM). Then, based on the correlation between dynamic data in the process manufacturing physical workshop, a dynamic data relationship model (DDRM) is constructed. Subsequently, a data communication interface is established to fuse the SPM and DDRM models, completing the construction of the digital twin model. This technical solution transforms the coupling relationship between processes in the process manufacturing production process into a coupling relationship between the parameters of each twin model, thereby clarifying the correlation between each twin model. However, this technical solution still relies on industrial modeling software to model the geometric and physical attributes of various elements in the process manufacturing physical workshop to construct a static physical model of each element. This static physical model construction process cannot directly reflect the physical correlation between them, nor can it construct the directional information connection relationship between different entity mapping objects using the aforementioned method. This technical solution essentially separates static physical relationship modeling from dynamic data relationship modeling, making it impossible to synchronously construct the corresponding dynamic pointing information between static physical objects, which significantly reduces the efficiency of digital twins of engineering objects.

[0011] Compared with the prior art, the present invention can generate corresponding 3D models reflecting the interrelated physical objects on the entities based on 2D drawing information involving entity relationships. Based on the aforementioned distinguishing technical features, the problem to be solved by the present invention can include: how to quickly realize the 3D representation of engineering objects with directional information. Specifically, as with the prior art, the 3D model building process of a digital twin factory usually separates the generation of physical entity models from the connection between different virtual models. Various independent physical devices within the factory can be quickly generated using the various dimensional parameters of the devices, but the generated 3D virtual models need to be connected manually or through a separately established dynamic data relationship model to construct a complete 3D model with directional information. Conversely, the present invention can generate 3D model reconstruction data files that simultaneously involve physical relationships and directional information based on 2D drawing information with entity relationships. Therefore, only the 3D model reconstruction data files are needed to directly generate a 3D model that simultaneously reflects the physical relationships and directional information of the entity mapping objects, significantly reducing the workload of setting up and generating 3D models.

[0012] Furthermore, the analysis identifies several 3D models in the 3D model reconstruction data file, demonstrating that the entity mapping objects of these 3D models have physical relationships with each other. By identifying 3D models with physically related entity mapping objects, abnormal 3D model components can be quickly selected, prompting staff to review the physical relationships.

[0013] The starting and ending points of physically related 3D models are specified based on the pointing information of their corresponding paired components within the entity mapping objects. These 3D models are connected through data exchange based on their physical relationships. This establishes a connection of physical attributes between the 3D models, reducing the need for manual addition and verification of physical attribute information.

[0014] According to a preferred embodiment, the method further includes: when identifying a two-dimensional drawing of a physical object involving entity relationships through text recognition, identifying the coordinate information of specific components of the physical object with directional information in the two-dimensional drawing. Compared with the prior art, the present invention can identify the coordinate information of physical objects contained in a two-dimensional drawing. Based on the above distinguishing technical features, the problem to be solved by the present invention may include: how to achieve preliminary positioning of specific components of different physical objects based on two-dimensional drawing information. Specifically, by identifying coordinate information, the positional relationship and physical association between different specific components can be preliminarily determined, avoiding the defect that the physical association between specific components needs to be manually set due to the loss of directional information.

[0015] According to a preferred embodiment, a method for automatically generating a 3D model corresponding to 3D model information includes: determining the coordinate information of specific components of a physical object with pointing information; determining the physical correlation of several 3D models based on the pointing information and coordinate information; and analyzing the specific components of several 3D models that have physical correlation with each other based on the physical correlation of the 3D models. Compared with the prior art, the present invention can determine whether there is a matching anomaly in the physical correlation between 3D model components based on pointing information and coordinate information. Based on the above distinguishing technical features, the problems to be solved by the present invention can include: how to improve the matching accuracy of 3D model components and the verification efficiency of abnormal 3D model components. Specifically, the present invention, on the one hand, can load various 2D drawings related to the abnormal 3D model components in response to the determination of abnormal 3D model components, so as to perform manual verification and avoid the phenomenon that staff have to search for 2D drawings from massive amounts of data and consume a lot of time. On the other hand, the present invention only loads the corresponding 2D drawings after the abnormal 3D model components are determined, without having to use the 2D drawings of all components as analysis objects, thereby significantly reducing the amount of data analysis and processing.

[0016] According to a preferred embodiment, a method for analyzing the physical correlation of several 3D models and identifying mutually paired components based on their physical correlation includes: comparing a first physical correlation between 3D models determined by coordinate information with a second physical correlation between 3D models determined by pointing information; identifying specific components whose first and second physical correlations are inconsistent; and identifying these inconsistent components as abnormal 3D model components. Compared with the prior art, this invention can extract information requiring manual review by considering the matching of physical attributes in two additional steps. Based on the aforementioned distinguishing technical features, the problem this invention aims to solve can include: how to determine specific abnormal 3D model components based on different information and different physical correlations between 3D models, thereby reducing the probability of errors during the generation of 3D models. Specifically, there are many possibilities for inconsistencies. It could be due to incorrect coordinate information or incorrect pointing information. Therefore, comparing the first and second physical correlations makes it easier to discover inconsistencies in coordinate or pointing information, facilitating subsequent manual review.

[0017] According to a preferred embodiment, the method further includes: when recognizing a two-dimensional drawing of a physical object involving entity relationships through text recognition, storing the coordinate information and pointing information of specific components related to the physical relationships in a mutually associated manner. During subsequent manual review, this approach facilitates the simultaneous review of the coordinate information and pointing information of specific components, avoiding one-sided review of either coordinate information or pointing information.

[0018] According to a preferred embodiment, the method further includes: establishing a data communication relationship between the interfaces of at least two specific components of the associated 3D models of entity mapping objects, whose coordinate information is consistent with each other, in a manner adapted to the pointing information, consistent with the physical relationship of the entity mapping objects. In this invention, the connection method of 3D models with physical relationships between entity mapping objects is: establishing data communication based on the physical relationship between their entities. The advantage of this setting is that it enables the physical relationship of the 3D models to be consistent with the relationship of the physical devices, avoiding the tedious work of manually reviewing each interface afterwards.

[0019] According to a preferred embodiment, the method further includes: upon identifying an abnormal 3D model component, generating a notification message to prompt manual review and indicating the location of the abnormal 3D model component. The advantage of this setup is that it simplifies the workload of manual review, allowing reviewers to directly locate the abnormal 3D model component based on the notification message.

[0020] The present invention also provides, in a second aspect, a storage medium for a digital twin method for engineering objects, used to store the coding program of the digital twin method for engineering objects. The digital twin method for engineering objects includes: identifying two-dimensional drawings of physical objects involving entity relationships and generating three-dimensional model information corresponding to the physical objects; generating a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings based on the three-dimensional model information; importing the three-dimensional model information of the physically related physical objects into three-dimensional modeling software based on the three-dimensional model reconstruction data file, thereby automatically generating a three-dimensional model corresponding to the three-dimensional model information; analyzing and determining several three-dimensional models in the three-dimensional model reconstruction data file, wherein the entity mapping objects of the three-dimensional models have physical relationships with each other; and specifying the start and end points of the physically related specific components of the mutually paired specific components of the mutually paired three-dimensional models according to their respective pointing information.

[0021] The storage medium of the present invention is capable of running the digital twin method of the engineering object of the present invention, enabling various terminals to quickly and automatically generate a three-dimensional model of the physical entity from a two-dimensional drawing through the storage medium.

[0022] The present invention also provides a digital twin system for an engineering object from a third aspect, which may include a processor configured to: identify two-dimensional drawings of physical objects involving entity relationships and generate three-dimensional model information corresponding to the physical objects; generate a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings based on the three-dimensional model information; import the three-dimensional model information of physical objects that are physically related to each other based on the three-dimensional model reconstruction data file into three-dimensional model modeling software, thereby automatically generating a three-dimensional model corresponding to the three-dimensional model information; analyze and determine several three-dimensional models in the three-dimensional model reconstruction data file, wherein the entity mapping objects of the three-dimensional models are physically related to each other; and specify the start and end points of the physically related specific components of the mutually paired specific components of the mutually paired three-dimensional models with mutually related entity mapping objects according to their respective pointing information.

[0023] The system of this invention requires less hardware and operates quickly. It can pre-verify the physical relationships between 3D models generated from 2D drawings according to preset rules.

[0024] According to a preferred embodiment, the processor is further configured to: determine the coordinate information of a specific component of a physical object having pointing information; determine the physical correlation of a plurality of three-dimensional models based on the pointing information and the coordinate information; and analyze the mutually paired specific components of the plurality of three-dimensional models that have physical correlation with each other based on the physical correlation of the three-dimensional models.

[0025] The system of the present invention can respond to the determination of abnormal three-dimensional model components and load various two-dimensional drawings related to the abnormal three-dimensional model components for manual review, avoiding the phenomenon that staff have to search for two-dimensional drawings from massive amounts of data and spend a lot of time. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the three-dimensional model provided by the present invention;

[0027] Figure 2 It is a two-dimensional pipe design drawing based on existing technology;

[0028] Figure 3 This is a schematic diagram illustrating the steps of the digital twin method for engineering objects according to the present invention.

[0029] List of reference numerals

[0030] 100: First physical object; 110: First interface; 200: Second physical object; 220: Second interface; 300: Flow direction. Detailed Implementation

[0031] The following is a detailed explanation with reference to the accompanying drawings.

[0032] This invention provides a method, system, and storage medium for creating digital twins of engineering objects. It can also be a system and method for verifying the physical correlation of 3D models based on digital twin factories. Furthermore, it can be an early warning system and method for identifying physical connections between 3D models based on digital twin factories. Finally, it can be a system and method for generating 3D models showing physical correlations between entity-mapped objects.

[0033] This invention provides explanations for some terms and nouns.

[0034] The physical relationships of a 3D model include the connectivity and directionality of media flow and data flow. For example, the start and end points of a pipeline in a 3D model, the starting point (i.e., outlet) of the media flow in a storage tank, and the starting point (i.e., inlet) of a pipeline are designated as the media origin.

[0035] Directional information: refers to the directionality of the connection relationship between various specific components.

[0036] Coordinate information: refers to the coordinate data of each specific component.

[0037] 3D model reconstruction data files: These are 3D model data files containing coordinate and orientation information for specific components. They differ from ordinary 3D model information. Ordinary 3D model information is a 3D model directly extracted from 2D drawings, containing only coordinate information unrelated to physical relationships and excluding orientation information.

[0038] Example 1

[0039] To quickly generate digital twins of engineering objects from a large number of 2D drawings, a novel digital twin method is needed. This method should be able to generate engineering objects in a 3D scene at least semi-automatically, thereby improving the efficiency of digital twin generation and reducing the error rate. Preferably, this digital twin method can pre-verify the physical relationships between 3D models generated from 2D drawings according to preset rules. Preferably, the pre-verification is implemented using a form for secondary manual review.

[0040] To address the shortcomings of existing technologies, the present invention provides a digital twin method for engineering objects, such as... Figure 3 As shown, the method includes steps S1 to S2. The digital twin method for the engineering object is run by a digital twin system for the engineering object. The digital twin system for the engineering object includes at least one processor. The processor may have a display component, or it may be connected to at least one terminal with a display component to display the digital twin process of the engineering object. Preferably, the processor is capable of running the coded program for the digital twin method of the engineering object. The processor may be a processor in a computer, a dedicated integrated circuit (DIC) chip, or a combination thereof, or it may be a separately configured processor, a DIC chip, or a combination thereof.

[0041] S1: Identify the two-dimensional drawings involving entity relationships of physical objects and generate three-dimensional model information corresponding to the physical objects. Based on the three-dimensional model information, generate a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings.

[0042] A physical object may be composed of multiple specific components. For example... Figure 2 As shown, the coordinate information of the specific components of each physical object is marked on the two-dimensional drawing, which can show the position and connection relationship between the specific components. At this time, the connection relationship is not directional. The interrelationship information of each physical object is read from the two-dimensional drawing through text recognition. The interrelationship information is, for example, the first interface 110 of the first physical object 100 is connected to the second interface 220 of the second physical object 200. It should be noted that the connection relationship at this time also represents the entity connection relationship between physical objects.

[0043] After identifying the interrelationships between physical objects on a 2D drawing, a 3D model corresponding to each physical object is generated. At this stage, the 3D model information only includes coordinate information; the pointing information between the physical objects is not included. Therefore, a 3D model reconstruction data file is generated based on the 3D model information and the interrelationships between the physical objects.

[0044] S11: When identifying physical objects in a 2D drawing involving entity relationships through text recognition, the coordinate information of specific components of the physical objects with directional information in the 2D drawing is identified. By identifying the coordinate information, the positional relationships and physical associations between different specific components can be preliminarily determined, avoiding the defect that physical associations between specific components need to be manually set due to the loss of directional information.

[0045] This invention employs intelligent drawing recognition technology in the ISO drawing intelligent recognition stage, including end-to-end recognition technology. This technology has been successfully applied in actual projects. Steps S111 to S113 are as follows:

[0046] S111: Use an optical character recognition network to extract pipeline numbers and instrument tag numbers from drawings;

[0047] S112: Use detection and segmentation networks to extract symbols and pipelines from drawings;

[0048] S113: Use graph neural networks to build upstream and downstream logic pipelines for symbols and pipes.

[0049] For example, from Figure 2 The two-dimensional drawing shows that the tag number of the first physical object 100 is T01, and the tag number of the second physical object 200 is E01. The text labels "Connect to T01 N1" in the lower left corner, "Connect to E01 N1" in the upper right corner, and the arrow indicate that the connection relationship is: the first interface 110 and the second interface 220 are connected via a pipe. The textual prompt "Connection 50-OG-004-6.3A1" in the upper left corner is directional information. The flow direction of the medium is 300, following the direction indicated by the arrow. Clearly, from... Figure 2 The information identified is: the direction of the medium flow 300 is from the first interface 110 of the first physical object 100 to the second interface 220 of the second physical object 200.

[0050] according to Figure 2 The connection information of the two-dimensional drawing is used to generate Figure 1 The 3D model in [the context]. Figure 1 In this context, the connection relationships between different physical objects include not only physical connections (e.g., the first interface 110 and the second interface 220 are connected via a pipe), but also information about the direction of the medium flow. Figure 1 In the diagram, the direction of the medium flow is indicated by 300.

[0051] Preferably, when using text recognition to identify physical objects in a 2D drawing involving entity relationships, the coordinate and directional information of specific components related to the physical relationships are stored in an interconnected manner. For example, the coordinates of the inlet and outlet of media flows or data flows, as well as the flow direction of the media flows and data flows, are stored. By interconnecting the coordinate and directional information, the coordinate and directional information of the subsequently generated 3D model information is ensured to be fully utilized and can be retrieved and verified at any time.

[0052] In subsequent manual review, this method is beneficial for simultaneously reviewing the coordinate and orientation information of specific components, avoiding one-sided review of coordinate or orientation information.

[0053] Preferably, when directional information is not recorded in the 2D drawings, the directional information of a specific component is predicted based on the physical properties and functions of the physical object. This avoids the defect of unclear directional information in the subsequent 3D model generation due to the absence of directional information. The predicted directional information is displayed with a color and display method that can be distinguished from the read directional information, so that it can be quickly discovered, supplemented, and reviewed by humans during subsequent review. The display method is, for example, a flashing or other method that can attract attention.

[0054] Preferably, when identifying physical objects through text recognition in two-dimensional drawings involving entity relationships, the correctness of the pointing information is initially judged based on the physical attributes and functions of the physical objects. If a pointing information is determined to be potentially incorrect, the related three-dimensional model and three-dimensional model reconstruction data file are associated and marked with a manual review warning indicator to facilitate subsequent manual review of the pointing information.

[0055] For example, if all three interfaces of a fixed tank are identified as media inlets and there is no media outlet, the pointing information of one of the tank's interfaces may be incorrect. However, since a physically correlated 3D model has not yet been generated, the correctness of each correlation data cannot be confirmed at this time. Therefore, all three correlations of the tank's three interfaces are marked with a manual review warning tag so that the physical correlation of the later-generated 3D model can be prioritized for matching and judgment.

[0056] S2: The 3D model reconstruction data file imports the 3D model information of interconnected physical objects into 3D modeling software, thereby automatically generating a 3D model corresponding to the 3D model information. Compared with the existing technology of generating a single 3D model based on a single 2D drawing, this invention can process a large number of 2D drawings in batches to form a preliminary 3D representation of the engineering object.

[0057] 3D modeling software includes PDMS, SP3D, Autodesk AutoCAD, etc.

[0058] Preferably, the steps for automatically generating a 3D pipeline model include:

[0059] S3: Analyze and determine the physical relationships between the entity mapping objects in the 3D model reconstruction data file of several 3D models.

[0060] S31: Determine the coordinate information of the specific component with pointing information.

[0061] Preferably, the specific component whose coordinate information is associated with its pointing information is extracted, the specific component is selected based on the identifier of the pointing information, and the coordinate information of the specific component is used as the selection object.

[0062] S32: Based on the pointing information (such as...) Figure 2 The physical relationships between several 3D models are determined by both the text prompt "Connect 50-OG-004-6.3A1" in the upper left corner and the coordinate information.

[0063] For example, coordinate information can determine the connection between two physical objects. Directional information can determine the start and end points of the medium flow within this connection. Therefore, the physical relationship between these two objects is: one object transports a medium flow to the other through a pipe, and the pipe joint is a sealed connection. If the medium flow is a flammable or explosive fluid, then the physical properties of the pipe are those suitable for handling flammable or explosive fluids.

[0064] S33: Analyze the specific components of several 3D models that have physical relationships with each other based on the determined physical relationships.

[0065] The first physical relationship between 3D models, determined based on coordinate information, is compared with the second physical relationship between 3D models, determined based on pointing information, to identify specific components that are inconsistent between the two. These inconsistent components are identified as abnormal 3D model components.

[0066] There are many possibilities for "inconsistencies". For example, (1) incorrect coordinate or pointing information may occur during the identification process; (2) contradictory pointing information may appear in several two-dimensional drawings; (3) corresponding information may be missing due to failure to collect pointing or coordinate information during the identification process, etc.

[0067] Based on the above possibilities, preferably, in response to the identification of abnormal 3D model components, all related 2D drawings are loaded and displayed in parallel with the 3D model to facilitate manual review. This is because review sometimes requires re-examining the 2D drawings, which means a significant time consumption for massive amounts of data. This invention retrieves all related 2D drawings of the abnormal 3D model components, reducing the time spent by staff searching for files and tracing back information. The parallel display method allows staff to compare the 2D drawings with the 3D model, facilitating the identification of the causes of inconsistencies.

[0068] Preferably, when displaying, specific components with inconsistent physical relationships due to different reasons are represented in different display methods, such as through color differences, flashing differences, and marking differences, so that staff can quickly verify based on the cause.

[0069] Preferably, the step of determining the first physical correlation between 3D models based on coordinate information includes:

[0070] Based on coordinate information, the relationship between specific components of a physically related object can be preliminarily determined.

[0071] If a physical object has at least two associated physical objects, verify whether the association relationships and the physical attributes of the corresponding physical objects are related. If the association relationships and the physical attributes of the corresponding physical objects are related, determine the association relationship between the specific components of the physical object as the first physical association. If the association relationships and the physical attributes of the corresponding physical objects are not related (or violate the association sample pattern), mark the association relationship as the first physical association to be confirmed and add a manual review warning.

[0072] For example, if the physical property of the medium flowing between tanks is high-temperature gas, can the material category parameter in the physical properties of the connecting pipe be applied to high-temperature gas? If it can, the first physical correlation is considered normal. If it cannot, the first physical correlation is considered a first physical correlation to be confirmed.

[0073] The steps for determining the second physical relationship between three-dimensional models based on the pointing information include: preliminarily determining the relationship between specific components of physical objects that have a physical relationship based on the pointing information.

[0074] If a physical object has at least two associated physical objects, verify whether the pointing direction and the physical attributes of its corresponding physical object are related. If the pointing direction and the physical attributes of its corresponding physical object are related, determine the association between the specific components of the physical object as the second physical association. If the pointing direction and the physical attributes of its corresponding physical object are not related (or violate the association sample pattern), mark the association as a second physical association to be confirmed and indicate it with a manual review warning.

[0075] For example, if the medium flowing between two tanks is a high-temperature gas, can the physical properties of the first tank used to transport the high-temperature gas and the second tank used to receive the high-temperature gas meet the requirements for storing the gas? For instance, if the materials of the first and second tanks are high-temperature resistant metal tanks, then their physical properties meet the requirements for storing high-temperature gas. In this case, the pointing direction is related to the physical properties of the physical object. The resulting physical correlation is the second physical correlation. Otherwise, the resulting physical correlation is marked as a second physical correlation pending confirmation and a manual review warning is issued.

[0076] This invention takes into account physical properties in two steps, which can extract information to be confirmed for staff, thereby reducing the probability of errors in the process of generating 3D models.

[0077] S4: Specify the start and end points of the physically related paired components based on the pointing information of the corresponding specific components of the 3D models that are associated with the entity mapping objects.

[0078] Establish a data communication relationship between the interfaces of at least two specific components of a 3D model whose coordinate information is consistent with that of the entity mapping object, in a manner that is adapted to the pointing information, and consistent with the physical relationship of the entity mapping object.

[0079] In this invention, the connection method for 3D models where entity mapping objects have physical relationships with each other is to establish data communication based on the physical relationships between their entities. The advantage of this setup is that it ensures that the physical relationships between the 3D models are consistent with the relationships between the physical devices, avoiding the tedious work of manually reviewing each interface later.

[0080] For example, such as Figure 2 As shown, the first physical object 100 has a tag number of T01 and has a first interface 110. The second physical object 200 has a tag number of E01 and has a second interface 220. The first interface 110 and the second interface 220 are connected by a pipe.

[0081] The flow direction 300 of the medium in the pipe between the first interface 110 and the second interface 220 is first interface 110 → second interface 220. After the three-dimensional model is generated, the medium flow data of the first interface 110 of the first physical object 100 is automatically changed into output data, and the medium flow data of the second interface 220 of the second physical object 200 is automatically changed into input data.

[0082] The 3D model automatically generated by this invention has attribute data consistent with the attribute data of its entities, and the association relationship is also consistent with the physical association relationship of the entities. Therefore, staff do not need to adjust the attribute data and association relationship of the 3D model in the later stage; they only need to check the data to be confirmed.

[0083] This invention utilizes automated recognition of ISO drawings and 3D pipeline model reconstruction technology, employing a unified end-to-end deep learning framework to minimize manual recognition and intervention in the drawing recognition and reconstruction process. This system requires a large amount of drawing data for symbol, relationship, and pipeline annotation. Through supervised learning, the accuracy of each network is optimized, allowing them to interconnect and cooperate for fully automated recognition and reconstruction.

[0084] Preferably, by identifying 3D models in which entity mapping objects have physical relationships with each other, abnormal 3D model components can be quickly selected to remind staff to review the physical relationships.

[0085] When an abnormal 3D model component is identified, a notification message is generated to prompt manual review and indicate the location of the abnormal 3D model component. The advantage of this setup is that it simplifies the workload of manual review, allowing reviewers to directly locate the abnormal 3D model component based on the notification message.

[0086] Example 2

[0087] This embodiment is a further improvement on embodiment 1, and repeated content will not be described again.

[0088] The present invention also provides, in a second aspect, a storage medium for a digital twin method for engineering objects, used to store the coding program of the digital twin method for engineering objects. The digital twin method for engineering objects includes: identifying two-dimensional drawings of physical objects involving entity relationships and generating three-dimensional model information corresponding to the physical objects; generating a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings based on the three-dimensional model information; importing the three-dimensional model information of the physically related physical objects into three-dimensional modeling software based on the three-dimensional model reconstruction data file, thereby automatically generating a three-dimensional model corresponding to the three-dimensional model information; analyzing and determining several three-dimensional models in the three-dimensional model reconstruction data file, wherein the entity mapping objects of the three-dimensional models have physical relationships with each other; and specifying the start and end points of the physically related specific components of the mutually paired specific components of the mutually paired three-dimensional models according to their respective pointing information.

[0089] The storage medium of the present invention is capable of running the digital twin method of the engineering object of the present invention, enabling various terminals to quickly and automatically generate a three-dimensional model of the physical entity from a two-dimensional drawing through the storage medium.

[0090] The present invention also provides a digital twin system for an engineering object in a third aspect, which may include a processor configured to:

[0091] S1: Identify the two-dimensional drawings involving entity relationships of physical objects and generate three-dimensional model information corresponding to the physical objects. Based on the three-dimensional model information, generate a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings.

[0092] S2: The data file based on the 3D model reconstruction imports the 3D model information of the physical objects that are related to each other into the 3D model modeling software, thereby automatically generating a 3D model corresponding to the 3D model information.

[0093] Methods for automatically generating 3D models corresponding to 3D model information include:

[0094] S3: Analyze and identify several 3D models in the 3D model reconstruction data file, where the entity mapping objects of the 3D models have physical relationships with each other.

[0095] S31: Determine the coordinate information of the specific component with pointing information.

[0096] Preferably, the specific component whose coordinate information is associated with its pointing information is extracted, the specific component is selected based on the identifier of the pointing information, and the coordinate information of the specific component is used as the selection object.

[0097] S32: Based on the pointing information (such as...) Figure 2The physical relationships between several 3D models are determined by both the text prompt "Connect 50-OG-004-6.3A1" in the upper left corner and the coordinate information.

[0098] For example, coordinate information can determine the connection between two physical objects. Directional information can determine the start and end points of the medium flow within this connection. Therefore, the physical relationship between these two objects is: one object transports a medium flow to the other through a pipe, and the pipe joint is a sealed connection. If the medium flow is a flammable or explosive fluid, then the physical properties of the pipe are those suitable for handling flammable or explosive fluids.

[0099] S33: Analyze the specific components of several 3D models that have physical relationships with each other based on the determined physical relationships.

[0100] S4: Specify the start and end points of the physically related paired components based on the pointing information of the corresponding specific components of the 3D models that are associated with the entity mapping objects.

[0101] The system of this invention requires less hardware and operates quickly. It can pre-verify the physical relationships between 3D models generated from 2D drawings according to preset rules.

[0102] The system of the present invention can respond to the determination of abnormal three-dimensional model components and load various two-dimensional drawings related to the abnormal three-dimensional model components for manual review, avoiding the phenomenon that staff have to search for two-dimensional drawings from massive amounts of data and spend a lot of time.

[0103] It should be noted that the specific embodiments described above are exemplary. Those skilled in the art can devise various solutions inspired by the disclosure of this invention, and these solutions all fall within the scope of this invention and its protection. Those skilled in the art should understand that this specification and its accompanying drawings are illustrative and not intended to limit the scope of the claims. The scope of protection of this invention is defined by the claims and their equivalents. This specification contains multiple inventive concepts; terms such as "preferredly," "according to a preferred embodiment," or "optionally" indicate that the corresponding paragraph discloses an independent concept. The applicant reserves the right to file divisional applications based on each inventive concept.

Claims

1. A processor for constructing three-dimensional models, characterized in that, The processor has a display component and is configured as follows: The system identifies two-dimensional drawings of physical objects involving entity relationships and generates three-dimensional model information corresponding to the physical objects. Based on the three-dimensional model information, it generates a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings. Specifically, it uses a text recognition network to extract pipeline numbers and instrument tag numbers from the two-dimensional drawings, uses a detection and segmentation network to extract symbols and pipelines from the two-dimensional drawings, and uses a graph neural network to establish upstream and downstream logical pipelines for the symbols and pipelines. It also identifies the coordinate information of specific components of physical objects with directional pointing information that has connection relationships in the two-dimensional drawings, and stores the coordinate information and the pointing information in a mutually associated manner. The starting and ending points of the physically related 3D models are specified based on the pointing information of the corresponding specific components of the associated 3D models. The 3D model reconstruction data file imports the 3D model information of the physical objects that are related to each other into the 3D model modeling software, thereby automatically generating a 3D model corresponding to the 3D model information. The physical relationships between several three-dimensional models are determined based on the pointing information and the coordinate information. Based on the physical correlation of the three-dimensional models, analyze the specific components of the mutual pairing of several three-dimensional models that have physical correlation with each other, including: comparing the first physical correlation between the three-dimensional models determined by coordinate information with the second physical correlation between the three-dimensional models determined by pointing information, determining the specific components that are inconsistent between the first physical correlation and the second physical correlation, and determining the specific components that are inconsistent between the first physical correlation and the second physical correlation as abnormal three-dimensional model components. Data communication is established based on the physical relationships between entities corresponding to the 3D model, so that the physical correlation of the 3D model is consistent with the correlation of the physical devices.

2. The processor for constructing a three-dimensional model according to claim 1, characterized in that, The processor is also configured to: Establish a data communication relationship between the interfaces of at least two specific components of a 3D model whose coordinate information is consistent with that of the entity mapping object, in a manner that is adapted to the pointing information, and consistent with the physical relationship of the entity mapping object.

3. The processor for constructing a three-dimensional model according to claim 2, characterized in that, The processor is also configured to: If a physical object has at least two associated physical objects, verify whether the association relationships and the physical attributes of the corresponding physical objects are related. When the association relationship and the physical attributes of the corresponding physical objects are related, the association relationship between the specific components of the physical objects is determined as the first physical association.

4. The processor for constructing a three-dimensional model according to claim 2, characterized in that, The processor is also configured to: If a physical object has at least two associated physical objects, check whether the pointing direction and the physical properties of the corresponding physical objects are related. When the direction of pointing and the physical properties of the corresponding physical object are related, the relationship between the specific components of the physical object is determined as the second physical relationship.

5. A three-dimensional model connection early warning system, comprising a processor, characterized in that, The processor is configured to: The system identifies two-dimensional drawings of physical objects involving entity relationships and generates three-dimensional model information corresponding to the physical objects. Based on the three-dimensional model information, it generates a three-dimensional model reconstruction data file corresponding to the two-dimensional drawings. Specifically, it uses a text recognition network to extract pipeline numbers and instrument tag numbers from the two-dimensional drawings, uses a detection and segmentation network to extract symbols and pipelines from the two-dimensional drawings, and uses a graph neural network to establish upstream and downstream logical pipelines for the symbols and pipelines. It also identifies the coordinate information of specific components of physical objects with directional pointing information that has connection relationships in the two-dimensional drawings, and stores the coordinate information and the pointing information in a mutually associated manner. The starting and ending points of the physically related 3D models are specified based on the pointing information of the corresponding specific components of the associated 3D models. The 3D model reconstruction data file imports the 3D model information of the physical objects that are related to each other into the 3D model modeling software, thereby automatically generating a 3D model corresponding to the 3D model information. The physical relationships between several three-dimensional models are determined based on the pointing information and the coordinate information. Based on the physical correlation of the three-dimensional models, analyze the specific components of the mutual pairing of several three-dimensional models that have physical correlation with each other; The process involves comparing the first physical correlation between three-dimensional models determined based on coordinate information with the second physical correlation between three-dimensional models determined based on pointing information, and identifying specific components that are inconsistent between the two. This includes comparing the first physical correlation between three-dimensional models determined based on coordinate information with the second physical correlation between three-dimensional models determined based on pointing information, identifying specific components that are inconsistent between the first physical correlation and the second physical correlation, and identifying specific components that are inconsistent between the first physical correlation and the second physical correlation as abnormal three-dimensional model components. It also includes a warning sign for manual review.

6. The three-dimensional model connection early warning system according to claim 5, characterized in that, It also includes terminals with display components. In response to the identification of the abnormal 3D model component, the processor loads the various 2D drawings related to the abnormal 3D model component and displays the 2D drawings and 3D model in parallel on the terminal to facilitate manual review.

7. The three-dimensional model connection early warning system according to claim 5 or 6, characterized in that, The processor is configured to: In cases where directional information is not recorded in two-dimensional drawings, the directional information of specific components is predicted based on the physical properties and functions of the physical object to avoid ambiguity in the directional information generated by the three-dimensional model.

8. The three-dimensional model connection early warning system according to claim 7, characterized in that, The processor is configured to: When using text recognition to identify physical objects in two-dimensional drawings involving entity relationships, the correctness of the pointing information is initially judged based on the physical attributes and functions of the physical objects. If the pointing information is determined to be incorrect, the 3D model and 3D model reconstruction data file related to the pointing information will be associated and marked with a manual review warning icon.

Images