System and method for intelligently constructing detailed drawing of steel structure

Abstract

The invention discloses a system and method for intelligently constructing a steel structure detail drawing, and belongs to the technical field of constructional engineering. According to the method, related data are collected and preprocessed through the Internet, and an AI large model is trained by adopting an architecture combining a convolutional neural network and a recurrent neural network; after information input by a user is coded, a preliminary detail drawing is generated through a model, then AI automatic checking optimization and manual auditing are carried out, and finally collaborative sharing is carried out through the Internet. The system comprises a data acquisition and preprocessing module and an AI large model training module. Compared with the prior art, the drawing efficiency can be remarkably improved, and the drawing time can be shortened; the drawing quality is improved based on mass data training, and human errors are reduced; intellectualization and automation are realized, and dependence on professionals is reduced; by means of the internet technology, collaborative sharing is promoted, space-time limitation is broken, communication efficiency is improved, an efficient, accurate and intelligent solution is provided for drawing steel structure detailed drawings, and the method has wide application prospects.

Description

Technical Field

[0001] The present invention relates to the technical field of construction engineering, and in particular to a system and method for intelligently constructing steel structure details. Background Art

[0002] In the field of construction engineering, steel structures are increasingly being used due to their significant advantages of high strength, light weight, and fast construction speed. As the core document guiding the processing, fabrication, and on-site installation of steel structures, the drawing of steel structure details is crucial. Traditionally, the drawing of steel structure details is mainly done manually by professional technicians. These technicians need to draw detailed information on the shape, size, connection method, and material specifications of steel structure components based on the design plan and relevant standards and specifications. However, this traditional drawing method has many obvious drawbacks: First, it is inefficient. The drawing process is cumbersome, requiring technicians to invest a lot of time and energy. When faced with complex steel structure projects, the drawing cycle is lengthy, seriously hindering the overall project progress.

[0003] Second, it is prone to errors. Human errors are inevitable in manual operations, such as incorrect dimensioning and improper connection node design. These errors can lead to serious consequences in subsequent processing, manufacturing, and installation, increasing project costs and risks.

[0004] Third, poor standardization and consistency. Different technicians have different drawing habits and understanding of standards, resulting in uneven standardization and consistency in the drawings produced, which is not conducive to project collaboration and quality control.

[0005] While the recent emergence of steel structure detailing software such as AutoCAD and Tekla Structures has improved design efficiency to a certain extent, it still has many limitations. AutoCAD is suitable for projects with simple structures and a high number of repeated components. However, for complex structures, its two-dimensional drawing method is prone to high drawing errors, long drawing times, high labor intensity, and inability to effectively avoid collisions. While Tekla Structures is suitable for complex structures and can produce 3D drawings, with the ability to check the position of nodes and components during the modeling process to avoid collisions, it also has the drawbacks of a large output volume, a certain impact on production cycles, and the high price of genuine software.

[0006] With the rapid development of the construction industry, the scale of engineering projects continues to expand, and the complexity of structures continues to increase. This places higher demands on the efficiency and quality of steel structure detailing. Simultaneously, the rapid advancement of internet and AI technologies has provided innovative approaches and effective means to address the challenges of steel structure detailing. How to fully leverage these advanced technologies to achieve intelligent and efficient construction of steel structure detailing has become a key challenge in the current construction engineering field. Summary of the Invention

[0007] The present invention aims to provide a system and method for intelligently constructing steel structure details. By integrating AI large model technology and Internet technology, the efficiency and quality of steel structure detail drawing can be greatly improved, the manual workload can be significantly reduced, the incidence of human error can be effectively reduced, and the intelligent and automated generation of steel structure details can be achieved to meet the needs of the rapid development of the modern construction engineering industry.

[0008] The present invention solves the above problems by adopting the following technical solutions: A method for intelligently constructing steel structure details includes the following steps: S1. Data collection and preprocessing steps: Leveraging internet technology, data related to steel structure detailing is collected from construction industry websites, design company databases, and standard specification publishing platforms. The collected raw data is cleaned to remove noise and invalid data. The data is classified according to type and purpose. Professional data annotation technology is used to annotate the data and organize it into a standard format suitable for AI model training and subsequent use. S2. AI large model training steps: Based on the preprocessed data, a large AI model for generating steel structure details is constructed using an architecture that combines a convolutional neural network and a recurrent neural network. The model is trained using a deep learning algorithm using a training set. During the training process, the model parameters are adjusted using an optimization algorithm to minimize the loss function between the model's predicted results and the true labels. The model performance is evaluated using a validation set, and regularization and network structure adjustment techniques are used based on the evaluation results to address overfitting and underfitting issues. The trained model is finally tested on a test set to ensure good generalization capabilities on unseen data. S3. Intelligent detailed drawing generation steps: Develop a user interface through which users input basic information about the building structure and specific design requirements. The system encodes the user input and converts it into a format that the AI model can understand. This encoded information is then fed into the trained AI model, which, based on the learned knowledge and patterns, generates preliminary steel structure detailed drawings, including 3D component models, 2D drawings, dimensioning, and a bill of materials. S4. Optimization and Review Steps: The AI big model automatically checks the generated steel structure detailed drawings based on the built-in steel structure design specifications and standard knowledge system to determine whether key factors such as component size, connection method, and material selection meet the specification requirements. When problems are found, the AI big model automatically provides targeted optimization suggestions based on rule-based optimization technology, case-based reasoning optimization technology, and machine learning-based optimization technology. The system has a manual review interface, through which professional technicians further review and modify the AI-generated detailed drawings. They then adopt or adjust the optimization suggestions given by the AI model based on their own professional experience and actual project conditions. S5. Internet collaboration and sharing steps: Use Internet technology to build a collaborative work platform. Project team members log in to the platform through a browser or client software to create, join, view and edit steel structure detail drawing projects online. Use cloud storage technology to store the generated steel structure details and their related data on the cloud server. The system performs version management on the detail files, records the time of each modification, the operator and the specific modification content, and users can access the cloud server to download the required detail files on demand.

[0009] Furthermore, in the data collection and preprocessing steps, web crawler technology is used to collect data from Internet channels; a deep learning-based target detection algorithm is used to process architectural design drawings, extract the geometric shape and size information of steel structure components, and convert them into digital model data; natural language processing technology is used to extract, classify and annotate keywords from text-formatted specification documents and case documents.

[0010] Furthermore, in the AI big model training step, TensorFlow or PyTorch is selected as the development framework to build the AI big model.

[0011] Furthermore, in the intelligent detailed drawing generation step, HTML, CSS, JavaScript are used in combination with Vue.js or React.js front-end frameworks to develop a user interaction interface; the numerical information input by the user is normalized, and the categorical information is converted into a vector using one-hot encoding or embedded encoding.

[0012] Furthermore, in the rule-based optimization technology, to address the problem of unreasonable component size, a rule library is established based on mechanics and structural design specifications to adjust the component geometric parameters; to address the problems existing in the connection method, the connection type is replaced according to the rules.

[0013] Furthermore, in the case-based reasoning optimization technology, a steel structure detail case library is established. When the model finds a problem, similar problem cases and corresponding solutions are retrieved from the case library, and the case solutions are adjusted and adapted according to the specific circumstances of the current problem.

[0014] Furthermore, in the machine learning-based optimization technology, a machine learning model is used to predict the performance of the steel structure, and optimization measures are taken when potential problems are found; after the optimization suggestions given by the model are applied to the actual structure, data on the actual effects are collected as feedback to further train the machine learning model.

[0015] Furthermore, in the Internet collaboration and sharing step, a collaborative work platform is built using a front-end and back-end separation architecture. The back-end is developed using Java and Python combined with Flask and Django frameworks, and the front-end uses front-end technology for interface display and user interaction; real-time collaboration between project team members is achieved through WebSocket real-time communication technology; and a well-known cloud storage service provider is selected for data storage.

[0016] A system for intelligently constructing steel structure details, comprising: Data acquisition and preprocessing module: The data acquisition and preprocessing module uses web crawler technology to collect data, uses a deep learning-based target detection algorithm to process drawing data, and uses natural language processing technology to process text data; it is used to collect data related to steel structure details from multiple channels with the help of Internet technology, and clean, classify and annotate the collected raw data, and organize it into a standard format.

[0017] AI large model training module: The AI large model training module selects TensorFlow or PyTorch as the development framework to build the model; it is used to build an AI large model for steel structure detail generation based on preprocessed data, and to train, optimize and perform performance testing on the model.

[0018] Intelligent detailed drawing generation module: The intelligent detailed drawing generation module uses the front-end framework to develop a user interaction interface and normalizes or encodes the user input information; it is used to develop a user interaction interface, receive the building structure information input by the user, encode the information and input it into the AI large model to obtain the generated preliminary steel structure detailed drawings.

[0019] Optimization and Review Module: In this module, the AI big model uses rule-based, case-based reasoning, and machine learning optimization techniques to provide optimization suggestions. The AI big model automatically optimizes and reviews the generated steel structure details and provides optimization suggestions. A manual review interface is also provided for professional technicians to further review and modify the details.

[0020] Internet collaboration and sharing module: The Internet collaboration and sharing module adopts a front-end and back-end separation architecture to build a platform, realizes collaboration through real-time communication technology, and selects well-known cloud storage service providers to store data; it is used to build a collaborative work platform to realize real-time collaboration among project team members; it uses cloud storage technology to store steel structure details and related data, and perform version management.

[0021] Compared with the prior art, the present invention adopting the above technical solution has the following outstanding features: Significantly improve drawing efficiency: Leveraging the AI large model's superior computing and efficient learning capabilities, detailed steel structure drawings can be generated in record time. Compared to traditional manual drawing methods, this significantly reduces drawing time and increases efficiency by several times or even dozens of times, effectively meeting the demand for rapid drawing output in engineering projects and significantly accelerating overall project progress.

[0022] Significantly Improved Drawing Quality: AI models are trained on a vast amount of standard data and real-world examples, accurately adhering to various steel structure design codes and standards. This significantly improves the accuracy and compliance of generated detailed drawings, effectively reducing human error and significantly lowering the construction risks and costs associated with drawing errors, laying a solid foundation for high-quality project implementation.

[0023] Highly intelligent and automated: Users simply input basic structural information and design requirements, and the system automatically and seamlessly completes a complex workflow involving detailed drawing generation, optimization, and review. This process significantly reduces manual intervention, achieving intelligent and automated steel structure detailing. This not only improves work efficiency but also reduces reliance on the number and skill level of specialized technicians.

[0024] Powerfully promotes collaboration and sharing: The application of internet technology breaks down time and space constraints, enabling project team members to collaborate in real time and efficiently share detailed drawing resources. People from different disciplines and locations can collaborate on detailed drawing creation and refinement on the same platform, significantly improving project communication efficiency and collaboration, effectively avoiding errors and duplication of effort caused by information discontinuity, and facilitating the smooth progress and successful implementation of projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] FIG1 is a system architecture diagram of an embodiment of the present invention.

[0026] FIG2 is a diagram illustrating the processing flow of each module according to an embodiment of the present invention.

[0027] FIG3 is a detailed flow chart of generating steel structure details using an AI large model according to an embodiment of the present invention.

[0028] FIG4 is a block diagram of a computer program for constructing an AI large model training model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The present invention will be further described below with reference to the embodiments, the purpose of which is only to provide a better understanding of the content of the present invention. Therefore, the examples given do not limit the scope of protection of the present invention.

[0030] See also Figure 1-Figure 4 , data collection and preprocessing Data Collection: Utilizing web crawler technology, specialized program code was developed, with appropriate search rules and data capture scopes. This targeted approach involved collecting large amounts of steel structure-related data from online channels, including construction industry websites (such as Zhulong.com and Tumu.com), internal databases of design firms (accessed with legal authorization), and standard specification publication platforms (such as the official website of the Department of Standards and Norms of the Ministry of Housing and Urban-Rural Development). For example, searching for the keyword "steel structure design drawings" on well-known architectural design websites allowed them to filter and download steel structure design drawings for various types of buildings (industrial plants, commercial complexes, and residential buildings). The latest steel structure design specification documents, such as the "Standard for Steel Structure Design" (GB 50017-2017) and its related revisions, were obtained from industry standard publication platforms.

[0031] Data Preprocessing: Leveraging advanced image recognition technologies, such as deep learning-based object detection algorithms (FasterR - CNN, YOLO series), collected architectural design drawings are processed. Feature extraction and object recognition are performed on the drawing images to accurately extract the geometric shape (beam cross-section, column foot form) and dimensions (length, width, thickness, diameter) of steel structural components. Using professional graphics processing software and programming tools, this information is converted into digital model data and stored in formats such as OBJ and STL for easy processing. For text-based specification documents and case studies, natural language processing techniques, such as NLTK (Natural Language Toolkit) in Python, are employed to extract keywords (e.g., keywords related to "component connection," "material strength," and "seismic design"), classify (by design specification, construction process, and case study category), and annotate (by identifying the clauses and applicable scopes corresponding to the keywords). For example, for a steel structure design specification document, natural language processing technology is used to extract key clauses regarding component connection methods (specific requirements for welding connections and bolted connections) and material selection (applicable scenarios and performance indicators of different steel grades), and detailedly mark the applicable building types, structural forms, load condition scenarios and conditions.

[0032] AI large model training Model Construction: Among numerous deep learning frameworks, the mature and widely used TensorFlow or PyTorch were selected as development frameworks. Taking into account the characteristics and requirements of steel structure detail drawing generation, a composite model structure combining convolutional neural networks and recurrent neural networks was constructed. The convolutional neural network uses multiple convolutional and pooling layers to extract local and global features from steel structure detail drawing images, such as component shape features and node connection features, through convolution operations. The recurrent neural network uses a long short-term memory (LSTM) network or a gated recurrent unit (GRU) variant to effectively process component connection sequence information. For example, when processing steel structure node details, the convolutional neural network can extract the node's geometry and the features of the connected components, while the recurrent neural network can learn the connection sequence and logical relationships between the components within the node.

[0033] Data Partitioning and Training: The preprocessed data is divided into training, validation, and test sets according to a specific ratio, for example, 70% for training, 15% for validation, and 15% for testing. The model is trained using the training set. During training, the model parameters are continuously adjusted using stochastic gradient descent (SGD), Adagrad, and Adadelta optimization algorithms to minimize the loss function between the model's predictions and the true labels. The model's performance is evaluated in real time using the validation set, monitoring the model's accuracy, recall, and F1 score to promptly identify overfitting and underfitting. If overfitting occurs, L1 or L2 regularization and dropout techniques are used to mitigate it. If underfitting occurs, the model complexity is appropriately increased, the network structure is adjusted, or the training time is extended. Finally, the trained model is tested on the test set for final performance to ensure good generalization to unseen data. For example, the model achieves over 95% accuracy in recognizing steel structure joint types on the test set, and the mean absolute error for component size prediction is within the acceptable range.

[0034] Intelligent detail drawing generation User Interface Development: Utilize front-end development technologies such as HTML, CSS, and JavaScript, combined with popular front-end frameworks such as Vue.js and React.js, to develop the user interface. The interface design adheres to the principles of simplicity and ease of use, providing users with clear input fields and action buttons. For example, dedicated input boxes are set up for entering numerical information such as building height, span, and number of floors. Drop-down menus allow users to select structural type and load conditions. For special design requirements, text boxes are provided for users to describe them in detail. Furthermore, the interface provides real-time validation, providing prompts when user input does not conform to formatting requirements or logical rules.

[0035] Information Encoding and Model Input: After receiving user input, the system uses a specific encoding method to convert it into a vector form that the AI model can understand. For example, numerical information is normalized, while categorical information is converted into a vector using one-hot encoding or embedding encoding. The encoded information is then input into a fully trained large AI model. Based on the input information, the model generates preliminary steel structure details according to the learned knowledge and patterns. The generated detailed drawings are presented in the form of 3D models and 2D drawings, and include precise dimensioning and a detailed bill of materials. For example, if a user enters the steel structure design requirements for a stadium, the model generates a 3D model of the complex spatial truss structure, as well as 2D machining drawings of each component, detailing the component dimensions and angles, and listing the specifications and quantity of the required steel materials in a bill of materials.

[0036] Optimization and Audit Automatic Optimization and Review: The AI large model has comprehensive and detailed knowledge of steel structure design codes and standards pre-installed within it, such as the relevant provisions of the "Steel Structure Design Standard" and the "Building Seismic Design Code." The model automatically checks the generated steel structure details, comparing the component dimensions, connection methods, and material selection information in the details with the requirements of the standards to determine compliance. For example, it checks whether the cross-sectional dimensions of steel beams meet strength and stability requirements, and whether the number and arrangement of bolts at connection nodes comply with seismic standards. If any problems are identified, the model automatically provides targeted suggestions based on built-in optimization strategies, such as adjusting component cross-sectional dimensions or changing connection node configurations.

[0037] Optimization strategy selection includes component size optimization strategy, connection method optimization strategy, and material selection optimization strategy. Specific optimization techniques can be addressed from the following three aspects: Rule-based optimization technology Geometric parameter adjustment: To address unreasonable component sizing, a rule library is established based on mechanics and structural design specifications. For example, for bending components, if the calculated stress exceeds the allowable stress, the cross-sectional dimensions are proportionally increased according to the rules, such as increasing the flange width or web thickness. For compressive components, if the slenderness ratio does not meet stability requirements, the cross-sectional area of the component is increased or the calculated length factor is reduced according to the rules.

[0038] Connection type replacement: If problems are found in the connection method, such as poor fatigue performance of welded nodes, some welded nodes will be replaced with bolted connection nodes according to the rules, and the better toughness and replaceability of bolted connections will be used to improve structural performance; if the shear capacity of bolted connection nodes is insufficient, the number of bolts will be increased or high-strength bolts will be used according to the rules.

[0039] Case-based reasoning optimization technology Similar Case Search: Build a case library for steel structure details. When a problem is discovered in the model, search the case library for similar cases and corresponding solutions. For example, when encountering a stress concentration problem at a complex node, search to find successful cases of similar node construction in the past and refer to their optimization solutions.

[0040] Case Study Adjustment and Application: Adjust and adapt the retrieved case study solutions to the specific circumstances of the current problem. For example, if the node dimensions in the reference case differ from the overall dimensions of the current structure, adjust the dimensions of the node components proportionally and then apply the adjusted solution as an optimization suggestion.

[0041] Machine learning-based optimization techniques Predictive optimization: Machine learning models are used to predict the performance of steel structures. For example, by mapping the finite element model of the structure to actual performance, the structure's response under different operating conditions can be predicted. When potential problems are discovered, such as predicted stress concentration in a specific area, optimization measures can be taken in advance, such as adding reinforcements or adjusting component layout.

[0042] Feedback Optimization: After applying the model's optimization recommendations to the actual structure, data on the actual results is collected as feedback to further train the machine learning model, enabling it to continuously optimize the accuracy and effectiveness of the recommendations. For example, based on the actual stress distribution data of the optimized structure, the model's parameters can be adjusted to provide more accurate optimization recommendations for similar problems in the future.

[0043] Manual Review and Adjustment: The system provides a manual review interface. After logging in to the review interface, professional technicians can view the detailed drawings generated by the AI model and the optimization suggestions provided. Using their professional knowledge and practical engineering experience, technicians can further review and modify the detailed drawings. In the review interface, technicians can directly annotate and modify dimensions on 2D drawings, and rotate and slice 3D models to observe the structure from different angles. For example, a technician may find that the weld design at a certain node meets the specifications, but considering the operability of on-site construction, they decide to adjust the weld form and record the reasons and content of the modification in the system.

[0044] Internet collaboration and sharing The collaborative work platform utilizes a separate front-end and back-end architecture. The back-end, developed in Java and Python (Flask and Django frameworks), handles business logic, data storage, and interface provision. The front-end utilizes the aforementioned technologies for interface display and user interaction. WebSocket real-time communication technology enables real-time collaboration among project team members. For example, when a designer makes changes to a detailed drawing in the office, these changes are synchronized in real time to the mobile device of a technician on the construction site, allowing the technician to immediately review and provide feedback. The platform also provides permission management, allowing for different operational permissions based on member roles (e.g., project manager, designer, and construction worker) to ensure data security and standardized operations.

[0045] Cloud storage and version management: Choose a well-known cloud storage service provider, such as Alibaba Cloud or Tencent Cloud, to store the generated steel structure details and their related data on the cloud server. Cloud storage has high reliability, high scalability, and efficient data access performance. The system performs version management on the detail files. After each file modification, a new version is automatically generated, and the version number, modification time, operator, and modification content information are recorded. When downloading a file, the user selects a specific version to download. For example, during the project construction process, due to design changes, it is necessary to view a previous version of the steel structure details. The user can find the corresponding version file in the version list stored in the cloud for download and viewing, which facilitates the tracing of project historical information and ensures the smooth progress of the project.

[0046] In terms of technical solutions, the present invention covers multiple key modules. The data acquisition and preprocessing module uses the internet to widely collect data and, through cleaning, classification, and labeling, provides high-quality data for subsequent model training. The AI large model training module adopts an architecture that combines convolutional neural networks and recurrent neural networks, and uses advanced deep learning algorithms for training, continuously optimizing model parameters and structure to improve its accuracy and generalization capabilities. The intelligent detailed drawing generation module develops a convenient user interaction interface to achieve efficient input and encoding of user information. The trained model quickly generates detailed drawings that include 3D models, 2D drawings, dimension annotations, and bills of materials. The optimization and review module uses the built-in normative knowledge of the AI large model for automatic optimization and review. At the same time, it combines optimization techniques based on rules, case reasoning, and machine learning to provide targeted suggestions and set up a manual review interface to ensure the quality of detailed drawings. The Internet collaboration and sharing module builds a collaborative work platform to achieve real-time collaboration among members, and uses cloud storage and version management to ensure data security and project traceability.

[0047] The present invention has brought significant beneficial effects. In terms of drawing efficiency, compared with traditional manual drawing, the time is greatly shortened, increasing by several times or even dozens of times, meeting the needs of rapid drawing production in engineering projects. In terms of drawing quality, the AI model is trained based on massive data and strictly adheres to specifications and standards, effectively reducing human errors and lowering construction risks and costs. It has a high degree of intelligence and automation. Users simply input information, and the system can automatically complete complex processes, reducing dependence on the number and skills of professionals. At the same time, Internet technology promotes collaboration and sharing among project teams, breaking the limitations of time and space, improving communication efficiency, and avoiding work errors and duplication of work. Through innovative technology integration and system construction, it provides an efficient, accurate, and intelligent solution for steel structure detailed drawing. It has extremely high practical value and broad application prospects, and is expected to promote major changes in steel structure detailed drawing technology in the construction engineering field.

[0048] The above description is only a preferred embodiment of the present invention and does not limit the scope of the present invention. Any equivalent changes made using the contents of the present invention specification and its drawings are included in the scope of the present invention.

Claims

1. A method for intelligently constructing steel structure details, characterized in that: The following steps are involved: S1. Data collection and preprocessing steps: With the help of Internet technology, data related to steel structure details are collected from construction industry websites, design company databases, and standard specification publishing platforms; Clean the collected raw data to remove noise and invalid data; classify the data by type and purpose; use professional data annotation technology to annotate the data and organize it into a standard format suitable for AI model training and subsequent use; S2. AI large model training steps: Based on the preprocessed data, a large AI model for generating steel structure details is constructed using an architecture that combines a convolutional neural network and a recurrent neural network. The model is trained using a deep learning algorithm using a training set. During the training process, the model parameters are adjusted using an optimization algorithm to minimize the loss function between the model's predicted results and the true labels. The model performance is evaluated using a validation set, and regularization and network structure adjustment techniques are used based on the evaluation results to address overfitting and underfitting issues. The trained model is finally tested on a test set to ensure good generalization capabilities on unseen data. S3. Intelligent detailed drawing generation steps: Develop a user interface through which users input basic information about the building structure and specific design requirements. The system encodes the user input and converts it into a format that the AI model can understand. This encoded information is then fed into the trained AI model, which, based on the learned knowledge and patterns, generates preliminary steel structure detailed drawings, including 3D component models, 2D drawings, dimensioning, and a bill of materials. S4. Optimization and review steps: The AI large model automatically checks the generated steel structure details based on the built-in steel structure design specifications and standard knowledge system to determine whether key factors such as component size, connection method, and material selection meet the specification requirements; When problems are discovered, the AI model automatically provides targeted optimization suggestions based on rule-based optimization techniques, case-based reasoning optimization techniques, and machine learning-based optimization techniques. The system has a manual review interface, through which professional technicians can further review and modify the AI-generated detailed drawings. They can also adopt or adjust the optimization suggestions given by the AI model based on their own professional experience and actual project conditions. S5. Internet collaboration and sharing steps: Use Internet technology to build a collaborative work platform. Project team members log in to the platform through a browser or client software to create, join, view and edit steel structure detail drawing projects online. Use cloud storage technology to store the generated steel structure details and their related data on the cloud server. The system performs version management on the detail files, records the time of each modification, the operator and the specific modification content, and users can access the cloud server to download the required detail files on demand.

2. The method for intelligently constructing steel structure details according to claim 1, wherein: In the data collection and preprocessing steps, web crawler technology is used to collect data from Internet channels; a deep learning-based target detection algorithm is used to process architectural design drawings, extract the geometric shape and size information of steel structure components, and convert them into digital model data; Natural language processing technology is used to extract, classify and annotate keywords from text-based regulatory documents and case documents.

3. The method for intelligently constructing steel structure details according to claim 1, wherein: In the AI large model training step, TensorFlow or PyTorch is selected as the development framework to build the AI large model.

4. The method for intelligently constructing steel structure details according to claim 1, wherein: In the intelligent detailed drawing generation step, a user interaction interface is developed using HTML, CSS, and JavaScript in combination with the Vue.js or React.js front-end framework; numerical information input by the user is normalized, and categorical information is converted into a vector using one-hot encoding or embedded encoding.

5. The method for intelligently constructing steel structure details according to claim 1, wherein: In the rule-based optimization technology, to address the problem of unreasonable component size, a rule library is established based on mechanics and structural design specifications to adjust the component geometric parameters; to address problems with connection methods, the connection type is replaced according to the rules.

6. The method for intelligently constructing steel structure details according to claim 1, wherein: In the case-based reasoning optimization technology, a steel structure detailing case library is established. When the model finds a problem, similar problem cases and corresponding solutions are retrieved from the case library, and the case solutions are adjusted and adapted according to the specific circumstances of the current problem.

7. The method for intelligently constructing steel structure details according to claim 1, wherein: In the machine learning-based optimization technology, a machine learning model is used to predict the performance of steel structures, and optimization measures are taken when potential problems are found. After the optimization suggestions given by the model are applied to the actual structure, data on the actual effects are collected as feedback to further train the machine learning model.

8. The method for intelligently constructing steel structure details according to claim 1, wherein: In the Internet collaboration and sharing steps, a collaborative work platform is built using a front-end and back-end separation architecture. The back-end is developed using Java and Python combined with Flask and Django frameworks, and the front-end technology is used for interface display and user interaction. Real-time collaboration between project team members is achieved through WebSocket real-time communication technology. A well-known cloud storage service provider is selected for data storage.

9. A system for intelligently constructing steel structure details, characterized in that: include: Data acquisition and preprocessing module: This module uses web crawler technology to collect data, a deep learning-based target detection algorithm to process drawing data, and natural language processing technology to process text data. It is used to collect data related to steel structure details from multiple channels using Internet technology, and clean, classify, and annotate the collected raw data, and organize it into a standard format. AI large model training module: This module uses TensorFlow or PyTorch as a development framework to build a model. It is used to build an AI large model for generating steel structure details based on preprocessed data, and to train, optimize, and perform performance testing on the model. Intelligent detailed drawing generation module: The intelligent detailed drawing generation module uses the front-end framework to develop a user interaction interface and normalize or encode the user input information; Used to develop a user interaction interface, receive building structure information input by the user, encode the information and input it into the AI large model to obtain the generated preliminary steel structure details; Optimization and Review Module: In this module, the AI big model uses rule-based, case-based reasoning, and machine learning optimization techniques to provide optimization suggestions. The AI big model automatically optimizes and reviews the generated steel structure details and provides optimization suggestions. A manual review interface is also provided for professional technicians to further review and modify the details. Internet collaboration and sharing module: The Internet collaboration and sharing module adopts a front-end and back-end separation architecture to build a platform, realizes collaboration through real-time communication technology, and selects well-known cloud storage service providers to store data; it is used to build a collaborative work platform to realize real-time collaboration among project team members; it uses cloud storage technology to store steel structure details and related data, and perform version management.

Images