Finite element model automatic modeling method and system based on construction drawing

Automatically identify and model construction drawings through computer vision technology and large language models, solving the problem that modeling engineers need to draw finite element models and is prone to errors, and achieving efficient and high-quality automatic modeling of finite element models.

CN119720687AActive Publication Date: 2025-03-28SHANGHAI CONSTRUCTION FOURTH CONSTRUCTION GROUP CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202510005496.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-01-03
Publication Date
2025-03-28
Estimated Expiration
2045-01-03

AI Technical Summary

Technical Problem

In the prior art, modeling engineers need high professional skills to draw finite element models, which are prone to errors, easy to omissions, and low modeling efficiency.

Method used

The construction drawing recognition module based on computer vision technology automatically recognizes and labels construction drawings, combines large language models to automatically establish a three-dimensional geometric model, and gives material attributes and load information to realize automatic modeling of the finite element model.

Benefits of technology

It improves the modeling efficiency and quality of finite element models, reduces the intervention needs of professional modeling engineers, avoids errors and omissions in manual modeling, and reduces labor costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119720687A_ABST
    Figure CN119720687A_ABST
Patent Text Reader

Abstract

The invention discloses an automatic finite element model modeling method and system based on construction drawings. The method comprises the following steps: automatically outputting unlabeled to-be-modeled construction drawings as labeled to-be-modeled construction drawings by establishing a construction drawing recognition module based on a computer vision technology; the method comprises the following steps of: endowing transverse and longitudinal axis numbers and three-dimensional coordinate information on each building component marked with a to-be-modeled construction drawing by establishing an axis net size identification module; a large language model is used as an automatic modeling module to automatically identify the size information of each building component marked with the to-be-modeled construction drawing, and the given transverse and longitudinal axis numbers and three-dimensional coordinate information are fused to automatically establish a three-dimensional geometric model; performing name labeling and material attribute endowing on each building component of the three-dimensional geometric model; and giving support constraint information and load information to the three-dimensional geometric model, and automatically establishing a visual finite element model based on a construction drawing. The method has the advantages of high modeling efficiency, high quality and capability of replacing manual work.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of construction engineering modeling, and in particular to a method and system for automatically modeling a finite element model based on construction drawings. Background Art

[0002] In the field of construction, construction drawings drawn by CAD and other software are the most common type of engineering data and are widely used in construction projects. Although the dimensions, positional relationships and other information of building components can be intuitively seen through construction drawings, finite element analysis cannot be performed because most construction drawings are two-dimensional drawings. In order to perform finite element analysis, modeling engineers are required to draw finite element models, which has high professional and technical requirements. Not all engineers can draw finite element models. Errors are prone to occur in the process of assigning material properties, support constraints, load information, etc. Not only is the modeling process cumbersome, but it is also prone to errors and omissions. It is difficult to get started and the modeling efficiency is low. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for automatic modeling of finite element models based on construction drawings, so as to solve the problems that modeling engineers have high professional and technical requirements for drawing finite element models, are prone to errors, omissions, and have low modeling efficiency.

[0004] In order to solve the above technical problems, the present invention provides a method for automatically building a finite element model based on a construction drawing, comprising:

[0005] Annotate the names, sizes and material information of the building components in the construction drawings; construct a deep learning semantic segmentation network model, use the construction drawings before annotation as input data and the construction drawings after annotation as output data as model training data to establish a construction drawing recognition module based on computer vision technology; automatically identify the unannotated construction drawings to be modeled, automatically extract the names, sizes and material information of the building components, and automatically output them as annotated construction drawings to be modeled through the construction drawing recognition module;

[0006] By establishing an axis grid size recognition module, the horizontal and vertical axis numbers and three-dimensional coordinate information are assigned to each building component of the construction drawing to be modeled which is automatically output by the construction drawing recognition module;

[0007] The large language model is used as an automatic modeling module. The size information of each building component in the construction drawings to be modeled, as well as the assigned horizontal and vertical axis numbers and three-dimensional coordinate information, are automatically identified through the large language model, and the three-dimensional geometric model is automatically established in the three-dimensional drawing software; the name and material information of each building component in the construction drawings to be modeled are automatically identified, and the name of each building component in the three-dimensional geometric model is automatically labeled and material properties are assigned; and the reference specifications of the construction drawings to be modeled are automatically parsed, and the support constraint information and load information are automatically assigned to the three-dimensional geometric model, and a visual finite element model based on the construction drawings is automatically established.

[0008] Furthermore, the automatic finite element modeling method based on construction drawings provided by the present invention adopts an adaptive moment estimation optimization algorithm to optimize and train a deep learning semantic segmentation network model to establish a construction drawing recognition module based on computer vision technology.

[0009] Furthermore, in the automatic modeling method of finite element models based on construction drawings provided by the present invention, the construction drawing recognition module based on computer vision technology includes a plan view recognition submodule and a stereo view recognition submodule, and the size information of each building component of the plan view of the construction drawing to be modeled that is automatically output by the large language model automatic recognition plan view recognition submodule and / or the stereo view of the construction drawing to be modeled that is automatically output by the large language model automatic recognition stereo view recognition submodule, as well as the horizontal and vertical axis numbers and three-dimensional coordinate information assigned thereto are integrated to automatically establish a three-dimensional geometric model.

[0010] Furthermore, the automatic finite element modeling method based on construction drawings provided by the present invention searches and understands the reference specifications of the construction drawings to be modeled through a large language model to parse the reference specifications of the construction drawings to be modeled.

[0011] Furthermore, the method for automatically building a finite element model based on a construction drawing provided by the present invention includes:

[0012] Convert the axis grid layer into quadrant form, record the axis number and coordinates of the intersection of the horizontal axis grid and the vertical axis grid, and parse the indecomposable axis grid layer into a series of arrays consisting of horizontal and vertical axis numbers and three-dimensional coordinates.

[0013] In order to solve the above technical problems, the present invention also provides a finite element model automatic modeling system based on construction drawings, comprising:

[0014] A construction drawing recognition module based on computer vision technology is used to automatically recognize unmarked construction drawings to be modeled, automatically extract the name, size and material information of building components, and automatically output them as marked construction drawings to be modeled;

[0015] Axis grid dimension recognition module, used to assign horizontal and vertical axis numbers and three-dimensional coordinate information to each building component of the construction drawing to be modeled that is automatically output by the construction drawing recognition module;

[0016] An automatic modeling module based on a large language model automatically identifies the size information of each building component in the construction drawings to be modeled, as well as the assigned horizontal and vertical axis numbers and three-dimensional coordinate information, and integrates them into the three-dimensional drawing software to automatically establish a three-dimensional geometric model; automatically identifies the name and material information of each building component in the construction drawings to be modeled, and automatically labels the name and assigns material properties to each building component in the three-dimensional geometric model; and automatically parses the reference specifications of the construction drawings to be modeled, automatically assigns support constraint information and load information to the three-dimensional geometric model, and automatically establishes a visual finite element model based on the construction drawings.

[0017] Furthermore, in the automatic finite element modeling system based on construction drawings provided by the present invention, the automatic modeling module based on the large language model includes a plane-elevation fusion submodule, an attribute assignment submodule and a visualization submodule.

[0018] The plane-elevation fusion submodule is used to fuse the horizontal and vertical axis numbers and three-dimensional coordinate information assigned to the size information of each building component in the plane view, elevation view and combination of the construction drawings to be modeled that are automatically output by the construction drawing recognition module based on computer vision technology, and automatically establish a three-dimensional geometric model in the three-dimensional drawing software;

[0019] The attribute assignment submodule is used to automatically identify the name and material information of each building component of the construction drawing to be modeled through the large language model, automatically annotate the name of each building component of the three-dimensional geometric model and assign material attributes, and automatically parse the reference specifications of the construction drawing to be modeled through the large language model to automatically assign support constraint information and load information to the three-dimensional geometric model and automatically establish a finite element model based on the construction drawing;

[0020] The visualization submodule is used to visualize the established finite element model.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] The automatic modeling method and system of finite element models based on construction drawings provided by the present invention automatically recognize unlabeled construction drawings to be modeled, automatically extract the names, sizes and material information of building components, and automatically output them as labeled construction drawings to be modeled according to a construction drawing recognition module based on computer vision technology; assign axis grid size information such as horizontal and vertical axis numbers and three-dimensional coordinates to each building component in the labeled construction drawings to be modeled through an axis grid size recognition module; automatically establish a visualized finite element model for the labeled construction drawings to be modeled with the axis grid size information through a large language model as an automatic modeling module; compared with manual modeling by professional modeling engineers, the method has high modeling efficiency and high quality, and can realize the drawing of finite element models without the intervention of professional modeling engineers, avoiding the problems of easy errors and omissions in manual modeling by professional modeling engineers, and not causing the problem of poor drawing quality and low precision of construction drawings due to artificial visual fatigue, thereby replacing manual labor and reducing labor costs.

[0023] The automatic finite element modeling method and system based on construction drawings provided by the present invention can automatically output paper versions and electronic versions of unlabeled construction drawings to be modeled into construction drawings to be modeled with the names, sizes and material information of labeled building components through a construction drawing recognition module based on computer vision technology, and has the advantage of a wide range of applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a flow chart of the automatic modeling method of the finite element model based on the construction drawing;

[0025] Figure 2 It is a structural schematic diagram of a finite element model automatic modeling system based on construction drawings;

[0026] Figure 3 It is a structural diagram of the drawing recognition module;

[0027] Figure 4 It is a structural diagram of the automatic modeling module;

[0028] As shown in the figure:

[0029] 100. Automatic finite element modeling system based on construction drawings, 110. Drawing recognition module, 111. Plan view recognition submodule, 112. Stereo view recognition submodule, 120. Axis grid dimension recognition module, 130. Automatic modeling module based on large language model, 131. Plane and elevation fusion submodule, 132. Attribute assignment submodule, 133. Visualization submodule. DETAILED DESCRIPTION

[0030] The present invention is described in detail below in conjunction with the accompanying drawings: The advantages and features of the present invention will become more apparent from the following description. It should be noted that the accompanying drawings are in very simplified form and in non-precise proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.

[0031] Please refer to Figures 1 to 3 The embodiment of the present invention provides a method for automatically building a finite element model based on a construction drawing, which may include the following steps:

[0032] Step S1, annotate the name, size and material information of the building components in the construction drawings; construct a deep learning semantic segmentation network model, use the construction drawings before annotation as input data and the construction drawings after annotation as output data as model training data to establish a construction drawing recognition module 110 based on computer vision technology; automatically identify the unannotated construction drawings to be modeled, automatically extract the name, size and material information of the building components, and automatically output them as annotated construction drawings to be modeled through the construction drawing recognition module 110. That is, the construction drawings to be modeled are annotated with the name, size and material information of each building component thereon. The building components include but are not limited to pile foundations, beams, columns, walls, slabs, etc., and the material information includes but is not limited to material grade, column concrete number, steel bar number information, etc. According to the extracted material grade information, find the matching material information in the self-built database, including the standard value of tensile strength, Poisson's ratio, Young's modulus and other information, and assign it to the corresponding building component.

[0033] In step S1, in order to establish a construction drawing recognition module 110 based on computer vision technology, an adaptive moment estimation optimization algorithm can be used to optimize and train a deep learning semantic segmentation network model to establish the construction drawing recognition module 110 based on computer vision technology.

[0034] In step S1, in order to realize the recognition and annotation of plan views, stereo views and their combination with construction drawings, the construction drawing recognition module 110 based on computer vision technology includes a plan view recognition submodule 111 and a stereo view recognition submodule 112. The plan view of the construction drawing to be modeled that is automatically output by the plan view recognition submodule 111 and / or the stereo view of the construction drawing to be modeled that is automatically output by the stereo view recognition submodule 112 are automatically recognized by the large language model, and the size information of each building component and the horizontal and vertical axis numbers and three-dimensional coordinate information assigned thereto are integrated to automatically establish a three-dimensional geometric model.

[0035] Step S2, by establishing an axis grid size recognition module 120, assigning horizontal and vertical axis numbers and three-dimensional coordinate information to each building component of the construction drawing to be modeled that has been marked and automatically output by the construction drawing recognition module 110. The method of establishing the axis grid size recognition module 120 includes: converting the axis grid layer into a quadrant form, recording the axis number and coordinates of the intersection of the horizontal axis grid and the vertical axis grid, and parsing the indecomposable axis grid layer into a series of arrays composed of horizontal and vertical axis numbers and three-dimensional coordinates. The axis grid size recognition module 120 assigns axis grid size information such as horizontal and vertical axis numbers and three-dimensional coordinates to each building component of the construction drawing to be modeled that has been marked, identifies the axis grid size in the construction drawing, and ensures that the geometric information is included in the subsequently identified plan view and elevation view information, so as to ensure that the integration of the plan view and elevation view is carried out smoothly when establishing the three-dimensional geometric model, and improve the accuracy and coordination of the establishment of the three-dimensional geometric model.

[0036] Step S3, using the large language model as an automatic modeling module, automatically identifying the size information of each building component of the construction drawings to be modeled and the horizontal and vertical axis numbers and three-dimensional coordinate information assigned by the large language model, and automatically establishing a three-dimensional geometric model in the three-dimensional drawing software; automatically identifying the name and material information of each building component of the construction drawings to be modeled, automatically annotating the name of each building component of the three-dimensional geometric model and assigning material attributes; and automatically parsing the reference specifications of the construction drawings to be modeled, automatically assigning support constraint information and load information to the three-dimensional geometric model, and automatically establishing a visual finite element model based on the construction drawings. The established finite element model can clearly see the material information used for each building component, the information of each building node, and the load information attached to the building component.

[0037] In step S3, the reference specifications of the construction drawings to be modeled may be parsed by searching and understanding the reference specifications of the construction drawings to be modeled through the large language model.

[0038] Please refer to Figures 2 to 4 The embodiment of the present invention further provides a finite element model automatic modeling system based on construction drawings, based on the above method, including a construction drawing recognition module 110 based on computer vision technology, an axis grid size recognition module 120 and an automatic modeling module 130 based on a large language model, wherein:

[0039] The construction drawing recognition module 110 based on computer vision technology is used to automatically recognize unlabeled construction drawings to be modeled, automatically extract the names, sizes and material information of building components, and automatically output them as labeled construction drawings to be modeled.

[0040] The axis grid dimension recognition module 120 is used to assign horizontal and vertical axis numbers and three-dimensional coordinate information to each building component of the construction drawing to be modeled that is automatically output by the construction drawing recognition module 110 and has been marked.

[0041] The automatic modeling module 130 based on the large language model automatically identifies the size information of each building component in the construction drawings to be modeled, and the assigned horizontal and vertical axis numbers and three-dimensional coordinate information, and automatically establishes a three-dimensional geometric model in the three-dimensional drawing software; the automatic modeling module 130 based on the large language model automatically identifies the name and material information of each building component in the construction drawings to be modeled, and automatically labels the name and assigns material properties to each building component in the three-dimensional geometric model; the automatic modeling module 130 based on the large language model automatically parses the reference specifications of the construction drawings to be modeled, automatically assigns support constraint information and load information to the three-dimensional geometric model, and automatically establishes a visual finite element model based on the construction drawings.

[0042] Please refer to Figure 4 The finite element model automatic modeling system based on construction drawings provided in an embodiment of the present invention, the automatic modeling module 130 based on the large language model includes a plane-elevation fusion submodule 131, an attribute assignment submodule 132 and a visualization submodule 133.

[0043] The plan-elevation fusion submodule 131 is used to fuse the horizontal and vertical axis numbers and three-dimensional coordinate information assigned to the size information of each building component in the plan view, elevation view and combination of the marked construction drawings to be modeled, which are automatically output by the construction drawing recognition module 110 based on computer vision technology, into the three-dimensional drawing software to automatically establish a three-dimensional geometric model.

[0044] Among them, the attribute assignment submodule 132 is used to automatically identify the name and material information of each building component in the construction drawings to be modeled through the large language model, automatically label the name of each building component in the three-dimensional geometric model and assign material attributes, and automatically parse the reference specifications of the construction drawings to be modeled through the large language model, automatically assign support constraint information and load information to the three-dimensional geometric model, and automatically establish a finite element model based on the construction drawings.

[0045] The visualization submodule 133 is used to visualize the established finite element model.

[0046] The automatic modeling method and system of finite element models based on construction drawings provided in the embodiments of the present invention automatically recognize the unlabeled construction drawings to be modeled, automatically extract the names, sizes and material information of the building components, and automatically output them as labeled construction drawings to be modeled according to the construction drawing recognition module 110 based on computer vision technology. The axis grid size information such as horizontal and vertical axis numbers and three-dimensional coordinates is assigned to each building component of the labeled construction drawings to be modeled through the axis grid size recognition module 120. The labeled construction drawings to be modeled with the axis grid size information are automatically established into a visualized finite element model through the large language model as an automatic modeling module. Compared with the manual modeling of professional modeling engineers, the method has high modeling efficiency and high quality, and the finite element model can be drawn without the intervention of professional modeling engineers, avoiding the problems of easy errors and omissions in manual modeling by professional modeling engineers, and will not cause artificial visual fatigue leading to poor quality and low precision of construction drawings. It replaces manual labor and reduces labor costs.

[0047] The automatic finite element modeling method and system based on construction drawings provided in the embodiments of the present invention can automatically output paper versions and electronic versions of unlabeled construction drawings to be modeled into construction drawings to be modeled that are labeled with the names, sizes and material information of labeled building components through a construction drawing recognition module 110 based on computer vision technology, and has the advantage of a wide range of applications.

[0048] The present invention is not limited to the above-mentioned specific implementation modes. Obviously, the above-mentioned embodiments are only some embodiments of the embodiments of the present invention, but not all embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention. Those skilled in the art can make other levels of modifications and changes to the present invention. In this way, if these modifications and changes of the present invention fall within the scope of the claims of the present invention, the present invention is also intended to include these changes and changes.

Claims

1. A method for automatically building a finite element model based on construction drawings, characterized in that: include: Annotate the names, sizes and material information of the building components in the construction drawings; construct a deep learning semantic segmentation network model, use the construction drawings before annotation as input data and the construction drawings after annotation as output data as model training data to establish a construction drawing recognition module based on computer vision technology; automatically identify the unannotated construction drawings to be modeled, automatically extract the names, sizes and material information of the building components, and automatically output them as annotated construction drawings to be modeled through the construction drawing recognition module; By establishing an axis grid size recognition module, the horizontal and vertical axis numbers and three-dimensional coordinate information are assigned to each building component of the construction drawing to be modeled which is automatically output by the construction drawing recognition module; The large language model is used as an automatic modeling module. The size information of each building component in the construction drawings to be modeled, as well as the assigned horizontal and vertical axis numbers and three-dimensional coordinate information, are automatically identified through the large language model, and the three-dimensional geometric model is automatically established in the three-dimensional drawing software; the name and material information of each building component in the construction drawings to be modeled are automatically identified, and the name of each building component in the three-dimensional geometric model is automatically labeled and material properties are assigned; and the reference specifications of the construction drawings to be modeled are automatically parsed, and the support constraint information and load information are automatically assigned to the three-dimensional geometric model, and a visual finite element model based on the construction drawings is automatically established.

2. The automatic finite element modeling method based on construction drawings according to claim 1 is characterized in that: An adaptive moment estimation optimization algorithm is used to optimize the training of the deep learning semantic segmentation network model to establish a construction drawing recognition module based on computer vision technology.

3. The automatic finite element modeling method based on construction drawings according to claim 1 is characterized in that: The construction drawing recognition module based on computer vision technology includes a plan view recognition submodule and a stereo view recognition submodule. The plan view of the construction drawing to be modeled that is automatically output by the plan view recognition submodule through the large language model and / or the stereo view of the construction drawing to be modeled that is automatically output by the stereo view recognition submodule through the large language model are automatically recognized. The size information of each building component and the horizontal and vertical axis numbers and three-dimensional coordinate information assigned to them are integrated to automatically establish a three-dimensional geometric model.

4. The automatic finite element modeling method based on construction drawings according to claim 1 is characterized in that: The reference specifications of the construction drawings to be modeled are searched and understood through a large language model to parse the reference specifications of the construction drawings to be modeled.

5. The automatic finite element modeling method based on construction drawings according to claim 1 is characterized in that: The method for establishing the axis grid size identification module comprises: Convert the axis grid layer into quadrant form, record the axis number and coordinates of the intersection of the horizontal axis grid and the vertical axis grid, and parse the indecomposable axis grid layer into a series of arrays consisting of horizontal and vertical axis numbers and three-dimensional coordinates.

6. A finite element model automatic modeling system based on construction drawings, characterized in that: include: A construction drawing recognition module based on computer vision technology is used to automatically recognize unmarked construction drawings to be modeled, automatically extract the name, size and material information of building components, and automatically output them as marked construction drawings to be modeled; Axis grid dimension recognition module, used to assign horizontal and vertical axis numbers and three-dimensional coordinate information to each building component of the construction drawing to be modeled that is automatically output by the construction drawing recognition module; An automatic modeling module based on a large language model automatically identifies the size information of each building component in the construction drawings to be modeled, as well as the assigned horizontal and vertical axis numbers and three-dimensional coordinate information, and integrates them into the three-dimensional drawing software to automatically establish a three-dimensional geometric model; automatically identifies the name and material information of each building component in the construction drawings to be modeled, and automatically labels the name and assigns material properties to each building component in the three-dimensional geometric model; and automatically parses the reference specifications of the construction drawings to be modeled, automatically assigns support constraint information and load information to the three-dimensional geometric model, and automatically establishes a visual finite element model based on the construction drawings.

7. The automatic finite element modeling system based on construction drawings according to claim 6 is characterized in that: The automatic modeling module based on the large language model includes a plane-elevation fusion submodule, an attribute assignment submodule and a visualization submodule. The plane-elevation fusion submodule is used to fuse the horizontal and vertical axis numbers and three-dimensional coordinate information assigned to the size information of each building component in the plane view, elevation view and combination of the construction drawings to be modeled that are automatically output by the construction drawing recognition module based on computer vision technology, and automatically establish a three-dimensional geometric model in the three-dimensional drawing software; The attribute assignment submodule is used to automatically identify the name and material information of each building component of the construction drawing to be modeled through the large language model, automatically annotate the name of each building component of the three-dimensional geometric model and assign material attributes, and automatically parse the reference specifications of the construction drawing to be modeled through the large language model to automatically assign support constraint information and load information to the three-dimensional geometric model and automatically establish a finite element model based on the construction drawing; The visualization submodule is used to visualize the established finite element model.

Citation Information

Patent Citations

  • Automatic identification and application method of axis net of building and structure CAD drawing

    CN113901539A

  • Bridge pier drawing recognition automatic modeling method based on computer vision and medium

    CN115238368A

  • Building information modeling method and device and electronic equipment

    CN116861507A

  • Building information model intelligent extraction method and system based on large language model

    CN117332069A

  • Rapid bridge modeling method

    CN118350187A