Method, system and platform for designing fabricated residential building structure in villages and towns

Abstract

The invention discloses a village and town assembly type residential building structure design method, system and platform, and belongs to the technical field of intelligent buildings, and the method comprises the steps: obtaining the outer contour information and room requirement information of a residential base; processing the acquired information; inputting the processed information into a color block map generation model to generate a plurality of building color block maps; generating a DWG drawing and corresponding text information according to the building color block map; bIM components are compared and selected according to a DWG drawing, and proper components are selected to be automatically set up; carrying out optimization and effect simulation on the built model; and the user satisfactorily outputs a building house type image and result display information. According to the method, a plurality of building color block maps can be quickly generated according to user requirements, the interaction time is short, the safety and economy of design are ensured, the design cost is reduced, the design period is shortened, and personalized requirements can be met.

Description

Technical Field

[0001] The present invention relates to the field of intelligent building technology, and in particular to a method, system and platform for designing the structure of assembled residential buildings in villages and towns. Background Art

[0002] At present, most of my country's traditional village and town houses are built by residents themselves, blindly imitating the design style of urban houses, lacking personalized design suitable for villages and towns themselves, and generally have problems such as unreasonable layout, poor lighting, and poor ventilation. The architectural design and structural design are out of touch, and the data exchange of building information cannot be achieved. At the same time, the structural design generally adopts traditional manual experience and simple calculations for verification. It is difficult to balance the safety and economy of residential construction design.

[0003] From a technical perspective, the architectural floor plans of prefabricated houses in villages and towns are often drawn by residents based on the characteristics of the original houses, with the approximate size determined orally, or by directly using the architectural floor plans of urban buildings, lacking targeted floor plan design. If a design institute is used for design, it will face high design costs and a long design cycle, which is not suitable for the actual situation of rural houses.

[0004] Most of the related floor plans generated in the prior art are simulated and generated by obtaining distance data for existing buildings through gateway signal measurement, point cloud scanning and other means, or generating corresponding 3D floor plans based on floor plans. For example, patent CN118200936A discloses a method for generating indoor floor plans, which generates simulated floor plans based on signal measurement data and master-slave gateway device distance data. Floor plan simulation can only be performed based on existing buildings, and it cannot be applied to the new construction of village and town houses. For example, patent No. 202211496610 discloses a method and system for generating a floor plan of a rural residential building based on image recognition, which mainly filters out the homestead outline information and corresponding floor plan information in the database based on conditional information. However, it lacks diversity, innovation and practicality, and the generation quality depends entirely on the data quality of the database, which is not conducive to residents choosing their favorite floor plans from a variety of options.

[0005] On the other hand, there is a problem of information disconnection between building floor plans and building structure drawings. Traditional rural housing lacks targeted structural design, which easily leads to frequent safety accidents. Rural housing structure drawings need to be drawn on the basis of architectural drawings. The drawing process is cumbersome, with many repeated steps and is relatively time-consuming.

[0006] In the prior art, the information conversion from the floor plan of a residential building to the structural diagram is mostly focused on multiple aspects such as the architectural drawings and the user's demand text. For example, patent CN116541943A discloses an intelligent interactive architectural structure design method, which mainly obtains the initial architectural drawings and obtains the target architectural structure design drawings according to the user's demand text for the architectural structure design. However, this method requires a large language model to perform a large amount of data calculations, which is not suitable for the low-cost and efficient design and construction mode of prefabricated houses in villages and towns. Summary of the invention

[0007] The purpose of the present invention is to provide a method, system and platform for designing the structure of prefabricated residential buildings in villages and towns to solve the problems mentioned in the above background technology.

[0008] To achieve the above object, the present invention provides a method for designing a structure of a prefabricated residential building in a village or town, the steps comprising:

[0009] S1. Obtain input condition information, where the input condition information includes homestead outer contour information and room requirement information;

[0010] S2, processing the input condition information, including refining the outer contour information of the homestead and standardizing the text information;

[0011] S3, inputting the processed information into a color block image generation model to generate a plurality of architectural color block images that meet the requirements, wherein the color block image generation model is trained using a stable diffusion algorithm model to calibrate color block images with color block detail annotations and corresponding text prompt information;

[0012] S4. Generate a DWG drawing and corresponding text information according to the user's needs and the architectural color block diagram selected by the user;

[0013] S5. According to the generated DWG drawings, compare and select BIM components in the component library, and select suitable components to automatically build the BIM model;

[0014] S6. Obtain the constructed BIM model information, establish a structural calculation model based on the BIM model information, and use the structural calculation model to optimize the BIM model and perform effect simulation;

[0015] S7. Obtain optimized BIM model information for user feedback, and output building floor plans and result display information when the user is satisfied.

[0016] Preferably, the processing of the input condition information in step S2 includes: using edge detection method and contour extraction algorithm to process the homestead outer contour information to obtain refined homestead outer contour information; using the forward and reverse prompt word logic in the stable diffusion algorithm to standardize the received text information.

[0017] Preferably, the information input into the color block map generation model in step S3 includes at least standardized prompt word information and a refined homestead outer contour map.

[0018] Preferably, the training to obtain the color block image generation model in step S3 also includes establishing a database, performing model training on the color block images calibrated with color block details and the corresponding text prompt information through a stable diffusion algorithm model to obtain a database model, and using different colors to represent different room types.

[0019] Preferably, the stable diffusion algorithm model in step S3 is a stable diffusion algorithm model fine-tuned using the LoRA model, and the fine-tuning method is to use the LoRA model as the base model in the stable diffusion algorithm model. The fine-tuned stable diffusion algorithm model includes four parts: LoRA model, text encoder, diffusion model and variational self-decoder.

[0020] Preferably, in step S4: according to the architectural color block diagram, the layout and size information of each room is obtained, and according to the layout and size information of each room, the axis and wall outline information corresponding to each layer is generated to form a DWG floor plan and a corresponding framework node coordinate file.

[0021] Preferably, in step S6, a structural calculation model is established using a PKPMAPI interface, the structural calculation model optimizes the BIM model through a genetic algorithm, and effect simulation is performed based on the optimized BIM model.

[0022] Preferably, the result display information output in step S7 includes: BIM model source file, structural calculation model source file, structural calculation book source file, and BIM model construction process video.

[0023] A system for designing the structure of assembled residential buildings in villages and towns, comprising: an interaction module, a data processing module, a building color block diagram generation module, a DWG conversion module, a BIM model generation module, a structural calculation model generation module and a user feedback module;

[0024] The interaction module includes:

[0025] A user requirement input unit, an information reading unit for acquiring the homestead outer contour information and room requirement information provided by the user and sending them to the data processing module;

[0026] The building color block image comparison unit presents the trained and generated multiple building color block images to the user in the form of pictures;

[0027] The building floor plan information reading unit is used to obtain the processed building floor plan DWG drawings and corresponding framework node information files, display them to users to download the source files, and send them to the BIM model generation module and the structural calculation model generation module;

[0028] Model result display unit, which displays and provides download paths for optimized BIM models and structural calculation models;

[0029] The data processing module comprises:

[0030] An information reading unit reads the information transmitted by the interactive module and divides it into a homestead outer contour part and a user demand part according to different formats;

[0031] An information processing unit, which processes text information and homestead outline information using algorithms;

[0032] The building color block diagram generation module comprises:

[0033] A color block image generating unit is used to generate a building color block image using a trained stable diffusion algorithm model according to condition information input by a user;

[0034] The color block image quality adjustment unit uses deep learning to check the generated building color block images and exclude low-quality color block images;

[0035] The DWG conversion module includes:

[0036] A conversion unit, used to convert the generated building color block image into a building unit DWG diagram and extract the corresponding framework node coordinate file;

[0037] Drawing output unit, output DWG floor plan and corresponding framework node coordinate files;

[0038] The BIM model generation module includes:

[0039] The component selection unit determines the component size required for the model from the component library according to the node coordinate file of the building plan obtained by the conversion module;

[0040] Automatically build the unit, automatically build the model based on BIM according to the size determined by the component selection module;

[0041] The structural calculation model generation module includes:

[0042] The model conversion unit uses the PKPMAPI interface to establish a structural calculation model based on the model information provided by the BIM model;

[0043] Model optimization and verification unit, which optimizes and verifies the BIM model through genetic algorithms;

[0044] Effect simulation unit, which simulates and displays the lighting and ventilation effects of the structural model based on the verified BIM model;

[0045] The user feedback module comprises:

[0046] Feedback receiving unit, receiving users’ modification opinions and satisfaction with the design plan;

[0047] The design adjustment unit makes adjustments based on user feedback and generates the final design solution.

[0048] A village and town prefabricated residential building structure design platform, including:

[0049] A user interaction terminal is used to input the outer contour information of the homestead and the room requirement information, and is connected to the design server via the Internet;

[0050] Design server, which processes design requirements, generates personalized building floor plans, extracts structural diagrams through image data, builds BIM models and performs structural calculation optimization verification, and finally stores the design plan in the database;

[0051] Database, storing various design plans and building materials information.

[0052] Therefore, the present invention adopts the above-mentioned method, system and platform for designing the structure of assembled residential buildings in villages and towns, which has the following beneficial effects:

[0053] (1) Through intelligent design generation and optimization process, it can effectively solve the problems of unreasonable layout, poor lighting, and poor ventilation in traditional village and town residential design;

[0054] (2) It is not limited to the data information in the database. It can innovate new building floor plans according to the information in the database under the rules. At the same time, the system realizes the integration of architectural design and structural design, ensures the safety and economy of the design, reduces the design cost, shortens the design cycle, and adapts to the actual needs of rural housing;

[0055] (3) It does not rely on large language models for data calculations, which meets the low cost and high requirements of the structural design of prefabricated houses in villages and towns.

[0056] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A method flow chart of an embodiment of a method for designing a structure of a prefabricated residential building in a village or town according to the present invention;

[0058] Figure 2 The present invention is a system flow chart of an embodiment of a system for designing the structure of assembled residential buildings in villages and towns. DETAILED DESCRIPTION

[0059] Example

[0060] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0061] Reference Figure 1 A method for designing a structure of a prefabricated residential building in a village or town, comprising the following steps:

[0062] S1. Obtain input condition information, which includes the outer contour information of the homestead and room requirement information. The outer contour information of the homestead includes the size of the homestead and picture information of the outer contour of the homestead. The room requirement information also includes information such as location, orientation, number of family members, number of bedrooms, room size, and living habits.

[0063] S2. Process the input condition information, including refining the outer contour information of the homestead and standardizing the text information.

[0064] Specifically, the edge detection method and the contour extraction algorithm are used to process the outer contour information of the homestead to obtain refined outer contour information of the homestead. The forward and reverse prompt word logic in the stable diffusion algorithm is used to standardize the received text information.

[0065] S3. Input the processed information into the color block image generation model to generate multiple building color block images that meet the requirements. One input information includes at least standardized prompt word information and a refined homestead outer contour map. The color block image generation model is trained using a stable diffusion algorithm model to calibrate color block images with color block detail annotations and corresponding text prompt information.

[0066] It also includes establishing a database, training the color block images with color block details and the corresponding text prompt information through a stable diffusion algorithm model to obtain a database model, and using different colors to represent different room types.

[0067] The stable diffusion algorithm model used in training is a stable diffusion algorithm model that is fine-tuned using the LoRA model. Use the network dimension (Network Rank & Alpha) to build a suitable LoRA model base network for the stable diffusion algorithm model to input data. The fine-tuned stable diffusion algorithm model includes four parts: LoRA model, text encoder, diffusion model, and variational self-decoder.

[0068] In the LoRA model, Rank is the rank of the network, which directly affects the size of the LoRA model. The image drawn in this real example is relatively simple, and the Rank value is set to 32. Alpha adjusts the parameters of LoRA's influence on the original model. The ratio of Alpha to Rank is the degree of LoRA's weakened weight. The closer Alpha is to Rank, the smaller the influence of LoRA on the weight of the original model. The closer it is to 0, the more significant the fine-tuning effect. Alpha is set to 16. The text encoder converts text into a text feature vector as the input of the diffusion model. The diffusion model is used to generate images. Since the generation process is carried out in the latent space after the image is compressed and reduced in dimension, the input and output of the diffusion model are both image features in the latent space. The variational self-decoder restores the image features in the latent space to images.

[0069] The steps for training a model using the stable diffusion algorithm include:

[0070] (1) Input an image and its corresponding text prompt information. The stable diffusion algorithm model decomposes the long string of prompt words into independent word units, which are the smallest semantic units.

[0071] (2) Use the “text encoder” to convert each word into a set of 768-dimensional word vectors. The word vectors are “summed up” in a certain way to obtain the pattern of the required image.

[0072] (3) Setting hyperparameters for input noise predictor, including the number of generation steps, rounds, sampling method, learning rate, etc.

[0073] (4) The stable diffusion algorithm model calculates a random noise map through random seeds. The noise predictor diffuses and removes the noise step by step under the guidance of the text encoder and adds a certain image to the image. After several cycles of denoising and adding images, the image features of the latent space are obtained.

[0074] (5) The image is converted from a “vector” format to an output image’s “bitmap” format through a variational autoencoder.

[0075] S4. According to the user's needs and the architectural color block map selected by the user, the layout and size information of each room is obtained. According to the layout and size information of each room, the format conversion plug-in is used to generate the axis and wall outline information corresponding to each layer and form a DWG floor plan and a corresponding framework node coordinate file.

[0076] S5. According to the generated DWG floor plan, select BIM components in the component library and select suitable components to automatically build the BIM model. The created component library contains most of the steel beam and steel column sizes required for prefabricated houses in villages and towns. The principle of component selection is based on the beam height-span ratio of 1 / 15. ～ 1 / 25, column slenderness ratio is 50～ 150. Automatic construction is completed using self-written BIMBase construction code.

[0077] S6. Obtain the constructed BIM model information, use the PKPMAPI interface to establish a structural calculation model based on the BIM model information, and use the structural calculation model to optimize the BIM model and perform effect simulation.

[0078] The optimization principle is to use the minimum amount of steel under safety conditions. Specifically, the BIM model is optimized using a genetic algorithm, and the structural layout parameters are optimized through the genetic algorithm. The initial design is quickly improved within a relatively small number of iterations (cooling cycles). Then, the best design obtained is used as the initial solution, and under a new set of parameters, the component size design variables are iteratively optimized until a structural design solution that meets both safety and economy is obtained. Then, the effect simulation is performed based on the optimized BIM model, including lighting and ventilation effects.

[0079] S7. Obtain optimized BIM model information for user feedback, and output building floor plans and result display information when the user is satisfied. The result display information includes: BIM model source file, structural calculation model source file, structural calculation book source file, and BIM model construction process video.

[0080] In this embodiment, when the user is not satisfied with the initially generated building color block map, adjustments can be made in two ways.

[0081] Method 1: Regenerate multiple building color block maps for customers to choose from based on the homestead outline information and room requirement information entered by the user.

[0082] Method 2: Select a satisfactory architectural color block map and generate a DWG floor plan. The operator will fine-tune and optimize the DWG floor plan according to user needs.

[0083] A system for designing the structure of assembled residential buildings in villages and towns, comprising:

[0084] Interaction Module

[0085] The user demand input unit is used to obtain the homestead outline information and room requirement information provided by the user and send it to the information reading unit of the data processing module. The homestead outline information includes the homestead size and the picture information of the homestead outline, and the room requirement information also includes the location, orientation, number of family members, number of bedrooms, room size and living habits.

[0086] The architectural color block diagram comparison unit presents multiple architectural color block diagrams generated through training to users in the form of pictures, allowing users to select their preferred architectural plan and download it locally.

[0087] The building floor plan information reading unit is used to obtain the processed building floor plan DWG format file and the corresponding framework node information file, display it to the user to download the source file and send it to the BIM model generation module and the structural calculation model generation module.

[0088] The model result display unit displays and provides the download path of the optimized BIM model and structural calculation model.

[0089] Data processing module:

[0090] The information reading unit reads the information transmitted by the interactive module and divides it into the homestead outer contour part and the user demand part according to different formats.

[0091] The information processing unit uses the forward and reverse prompt word logic in the stable diffusion algorithm to standardize the received text information (for example, three bedrooms are needed, each with an area of ​​approximately 20 square meters) and uses the edge detection algorithm and contour extraction algorithm to obtain refined homestead outer contour information.

[0092] Building color block map generation module:

[0093] The color block map generating unit is connected to the data processing module and generates the building color block map by using the trained stable diffusion algorithm model according to the condition information input by the user.

[0094] The color block image quality adjustment unit uses a generative adversarial network to check the generated building color block images and exclude low-quality color block images.

[0095] DWG conversion module:

[0096] The conversion unit is used to convert the generated building color block image into a building unit DWG drawing and extract the corresponding framework node coordinate file for the operator to fine-tune and optimize and output the fine-tuned building drawings and corresponding framework node coordinate files through the drawing output unit.

[0097] Drawing output unit, outputs DWG floor plan and corresponding framework node coordinate files.

[0098] BIM model generation module:

[0099] The component selection unit determines the component sizes required for the model from the component library according to the node coordinate file of the building plan obtained by the conversion module.

[0100] Automatically build units, according to the size determined by the component selection module, use self-built commands to automatically build models based on BIM.

[0101] Structural calculation model generation module:

[0102] The model conversion unit establishes a structural calculation model using the PKPMAPI interface based on the model information provided by the BIM model.

[0103] The model optimization and verification unit optimizes and verifies the BIM model through genetic algorithms to ensure the safety and economy of the structure.

[0104] The effect simulation unit simulates and displays the lighting and ventilation effects of the structural model based on the verified BIM model for users' reference and selection.

[0105] User feedback module:

[0106] Feedback receiving unit receives users’ modification suggestions and satisfaction with the design plan.

[0107] The design adjustment unit makes adjustments based on user feedback and generates a final design solution that satisfies the user.

[0108] A village and town prefabricated residential building structure design platform, including:

[0109] The user interaction terminal is used to input the outer contour information of the homestead and the room requirement information, and is connected to the design server through the Internet.

[0110] The design server processes design requirements, generates personalized building floor plans, extracts structural diagrams through image data, establishes BIM models and performs structural calculation optimization verification, and finally stores the design plans in the database.

[0111] Database, storing various design plans and building materials information.

[0112] Therefore, the present invention adopts the above-mentioned method, system and platform for designing the structure of prefabricated residential buildings in villages and towns, providing a personalized, low-cost, efficient and fast multi-scheme customized prefabricated residential building structure design solution for residents in villages and towns.

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for designing the structure of a prefabricated residential building in a village or town, characterized in that the steps include: S1. Obtain input condition information, where the input condition information includes homestead outer contour information and room requirement information; S2, processing the input condition information, including refining the outer contour information of the homestead and standardizing the text information; S3, inputting the processed information into a color block image generation model to generate a plurality of architectural color block images that meet the requirements, wherein the color block image generation model is trained using a stable diffusion algorithm model to calibrate color block images with color block detail annotations and corresponding text prompt information; S4. Generate a DWG drawing and corresponding text information according to the user's needs and the architectural color block diagram selected by the user; S5. According to the generated DWG drawings, compare and select BIM components in the component library, and select suitable components to automatically build the BIM model; S6. Obtain the constructed BIM model information, establish a structural calculation model based on the BIM model information, and use the structural calculation model to optimize the BIM model and perform effect simulation; S7. Obtain optimized BIM model information for user feedback, and output building floor plans and result display information when the user is satisfied.

2. A method for designing a structure of a prefabricated residential building in a village or town according to claim 1, characterized in that: The processing of the input condition information in step S2 includes: using edge detection method and contour extraction algorithm to process the homestead outer contour information to obtain refined homestead outer contour information; using the forward and reverse prompt word logic in the stable diffusion algorithm to standardize the received text information.

3. A method for designing a structure of a prefabricated residential building in a village or town according to claim 2, characterized in that: The information input into the color block map generation model in step S3 at least includes the standardized prompt word information and the refined homestead outer contour map.

4. A method for designing a structure of a prefabricated residential building in a village or town according to claim 3, characterized in that: The training of the color block image generation model in step S3 also includes establishing a database, performing model training on the color block images calibrated with color block details and the corresponding text prompt information through a stable diffusion algorithm model to obtain a database model, and using different colors to represent different room types.

5. A method for designing a structure of a prefabricated residential building in a village or town according to claim 4, characterized in that: The stable diffusion algorithm model in step S3 is a stable diffusion algorithm model fine-tuned using the LoRA model, and the fine-tuning method is to use the LoRA model as the base model in the stable diffusion algorithm model. The fine-tuned stable diffusion algorithm model includes four parts: the LoRA model, the text encoder, the diffusion model and the variational self-decoder.

6. A method for designing a structure of a prefabricated residential building in a village or town according to claim 5, characterized in that: In step S4: according to the architectural color block diagram, the layout and size information of each room is obtained, and according to the layout and size information of each room, the axis and wall outline information corresponding to each layer is generated to form a DWG floor plan and a corresponding framework node coordinate file.

7. A method for designing a structure of a prefabricated residential building in a village or town according to claim 6, characterized in that: In step S6, the PKPMAPI interface is used to establish a structural calculation model, the structural calculation model optimizes the BIM model through a genetic algorithm, and the effect simulation is performed based on the optimized BIM model.

8. A method for designing a structure of a prefabricated residential building in a village or town according to claim 7, characterized in that: The result display information output in step S7 includes: BIM model source file, structural calculation model source file, structural calculation book source file, and BIM model construction process video.

9. A system for designing the structure of assembled residential buildings in villages and towns, using the method for designing the structure of assembled residential buildings in villages and towns as described in claim 8, characterized in that: include: Interaction module, data processing module, building color block diagram generation module, DWG conversion module, BIM model generation module, structural calculation model generation module and user feedback module; The interaction module includes: A user requirement input unit, an information reading unit for acquiring the homestead outer contour information and room requirement information provided by the user and sending them to the data processing module; The building color block image comparison unit presents the trained and generated multiple building color block images to the user in the form of pictures; The building floor plan information reading unit is used to obtain the processed building floor plan DWG drawings and corresponding framework node information files, display them to users to download the source files, and send them to the BIM model generation module and the structural calculation model generation module; Model result display unit, which displays and provides download paths for optimized BIM models and structural calculation models; The data processing module comprises: An information reading unit reads the information transmitted by the interactive module and divides it into a homestead outer contour part and a user demand part according to different formats; An information processing unit, which processes text information and homestead outline information using algorithms; The building color block diagram generation module comprises: A color block image generating unit is used to generate a building color block image using a trained stable diffusion algorithm model according to condition information input by a user; The color block image quality adjustment unit uses deep learning to check the generated building color block images and exclude low-quality color block images; The DWG conversion module includes: A conversion unit, used to convert the generated building color block image into a building unit DWG diagram and extract the corresponding framework node coordinate file; Drawing output unit, output DWG floor plan and corresponding framework node coordinate files; The BIM model generation module includes: The component selection unit determines the component size required for the model from the component library according to the node coordinate file of the building plan obtained by the conversion module; Automatically build the unit, automatically build the model based on BIM according to the size determined by the component selection module; The structural calculation model generation module includes: The model conversion unit uses the PKPMAPI interface to establish a structural calculation model based on the model information provided by the BIM model; Model optimization and verification unit, which optimizes and verifies the BIM model through genetic algorithms; Effect simulation unit, which simulates and displays the lighting and ventilation effects of the structural model based on the verified BIM model; The user feedback module comprises: Feedback receiving unit, receiving users’ modification opinions and satisfaction with the design plan; The design adjustment unit makes adjustments based on user feedback and generates the final design solution.

10. A village and town prefabricated residential building structure design platform, using the village and town prefabricated residential building structure design system described in claim 9, characterized in that: include: A user interaction terminal is used to input the outer contour information of the homestead and the room requirement information, and is connected to the design server via the Internet; Design server, which processes design requirements, generates personalized building floor plans, extracts structural diagrams through image data, builds BIM models and performs structural calculation optimization verification, and finally stores the design plan in the database; Database, storing various design plans and building materials information.

Images