Design support device

Abstract

To provide a design support device that quickly constructs a design model without causing rework from the estimation stage. [Solution] A design support device 2 that constructs a design model in which multiple parts are arranged on a structure, comprising: a logic design unit 233 that constructs a provisional design model in which multiple parts are arranged on the structure in accordance with predetermined design rules by inputting drawing data 23b of the structure and specification data 23c of the multiple parts; a feedback control unit 234 in which a generation AI 5 feedback controls the provisional design model so that multiple parts included in the provisional design model do not interfere with each other when the logic design unit 233 constructs the provisional design model; and an output unit 237 that outputs the provisional design model constructed by the logic design unit 233 and the feedback control unit 234 as a design model.

Description

[Technical Field]

[0001] The present disclosure relates to a design support device that constructs a design model in which a plurality of parts are arranged in a structure. [Background technology]

[0002] Conventionally, design support devices for CAD (Computer-Aided Design) systems used in BIM (Building Information Modeling) that can build design models used in the design of structures such as buildings and bridges have been known (see, for example, Patent Document 1). These CAD systems include AutoCAD and REVIT (registered trademarks, omitted below), etc.

[0003] The design support device described in Patent Document 1 includes a 3D database consisting of an existing 3D model in which the plant equipment layout and piping routes are defined, a catalog database consisting of standard product information in which representative part shapes are defined, and a storage device in which a part vendor database in which the standard product information is associated with unique identification information and in which individual part shapes for each vendor are assigned and stored. This design support device replaces the representative part shape set in the 3D model with a candidate individual part shape associated with the identification information, and when placing the replaced candidate individual part shape in the 3D model, determines whether placement is possible based on spatial interference or structural strength calculated from the piping routes. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-211736 Summary of the Invention [Problem to be solved by the invention]

[0005] The design support device described in Patent Document 1 determines whether a representative part shape can be replaced with an individual part shape based on spatial interference or structural strength calculated from piping routes, based on an already designed 3D model. In other words, since this is a technology that replaces a representative part shape placed in an already completed 3D model with an individual part shape, when building a design model in which multiple parts are placed in a new design model, the quality of the deliverables depends greatly on the skill and experience of the expert.

[0006] Furthermore, detailed consideration is difficult in the early stages of a project, especially in the estimation stage when time is limited. As a result, interference between parts or a lack of necessary space is often discovered in later processes, resulting in large-scale design changes and rework, which is one of the factors that leads to construction delays and increased construction costs. Furthermore, there are a wide variety of constraints that need to be considered, such as local ordinances and regulations and the client's required specifications, and there are limits to how much humans can comprehensively check and reflect these in the design.

[0007] Therefore, there is a demand for a design support device that can quickly build a design model without causing rework from the estimation stage. [Means for solving the problem]

[0008] A characteristic configuration of a design support device according to the present disclosure is a design support device that constructs a design model in which a plurality of parts are arranged on a structure, and includes: a logic design unit that constructs a tentative design model in which the plurality of parts are arranged on the structure in accordance with predetermined design rules by inputting drawing data of the structure and specification data of the plurality of parts; a feedback control unit that uses a generation AI to feedback control the tentative design model so that the plurality of parts included in the tentative design model do not interfere with each other when the logic design unit constructs the tentative design model; and an output unit that outputs the tentative design model constructed by the logic design unit and the feedback control unit as the design model.

[0009] In the design support device according to the present disclosure, a logic design unit constructs a provisional design model in accordance with predetermined design rules, while a feedback control unit uses a generation AI to feedback-control the provisional design model so that multiple parts included in the provisional design model do not interfere with each other. In other words, while the logic design unit quickly constructs a rough provisional design model, the generation AI performs feedback control each time to prevent multiple parts from interfering with each other, thereby improving the accuracy of automatic design of the design model.

[0010] As a result, problems such as interference between parts or lack of necessary space being discovered in later processes are eliminated, and large-scale design changes and rework are not required, which also eliminates problems such as delays in construction time and increased construction costs.In this way, it is a design support device that quickly builds design models that do not require rework, even from the estimation stage. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a block diagram showing the overall configuration of a CAD system. [Figure 2] FIG. 1 is a block diagram of a CAD system including a design support device. [Figure 3] 1 is a flowchart showing a control method for a CAD system. [Figure 4] 10 is an example of a screen display by a design support device. [Figure 5] 10 is an example of a screen display by a design support device. [Figure 6] 10 is an example of a screen display by a design support device. [Figure 7] 10 is an example of a screen display by a design support device. [Figure 8] 10 is an example of an output list from the design support device. DETAILED DESCRIPTION OF THE INVENTION

[0012] An embodiment of a design support system according to the present disclosure will be described below with reference to the drawings. In this embodiment, a user terminal on which REVIT is installed will be used as an example of a CAD system 100. However, the present invention is not limited to the following embodiment, and various modifications are possible within the scope of the present invention.

[0013] As shown in FIG. 1, the CAD system 100 includes a BIM device 1, a design support device 2, and a user device 3. The BIM device 1 is a server owned by a vendor that manages and operates BIM. The design support device 2 is a server owned by a vendor that develops and provides add-in software for BIM. The user device 3 is a terminal on which BIM is installed, and is owned by a user who designs, constructs, and manages structures such as buildings and bridges. For convenience, the user device 3 is shown as a single terminal, but different users who design, construct, or manage structures actually own multiple terminals. These terminals may be desktop PCs, laptops, tablets, smartphones, etc.

[0014] The BIM device 1, design support device 2, and user device 3 are connected to each other so that they can communicate with each other via a network 4. The BIM device 1 includes a processor 11, a communication IF 12, a memory 13, and an input / output IF 14. Although the BIM device 1 includes other functional units, only the functional units related to this embodiment are described here.

[0015] A BIM program 13a is stored in the memory 13. The processor 11 is the central part of the computer, receiving instructions and performing calculations and data processing. The processor 11 includes a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an NPU (Neural Network Processing Unit), or other hardware for executing the BIM program 13a stored in the memory 13. In other words, the processor 11 has an arithmetic processing circuit, input ports, and output ports. The processor 11 may be an ASIC, FPGA, or SoC, and is not particularly limited.

[0016] The communication IF 12 is a communication interface that has a function of transmitting signals output from the processor 11 to the design support device 2 and the user device 3 via the network 4, and a function of sending signals received from the design support device 2 and the user device 3 via the network 4 to the processor 11. The processor 11 can transmit and receive signals between the memory 13 and the communication IF 12. The input / output IF 14 functions as an interface with input devices (e.g., keyboard, mouse, touch panel, touch pad, etc.) and output devices (e.g., display, speaker, etc.).

[0017] The memory 13 stores data within the computer and includes a main memory and a secondary memory. The main memory is a storage area for temporarily storing programs and data, and is composed of RAM or the like. The secondary memory is a non-temporary storage area for permanently storing programs and data, and is composed of an HDD, SSD or the like. The secondary memory may also be external hardware such as a cloud server or rental server.

[0018] The BIM program 13a is an application stored in the secondary memory, read into the main memory, and executed by the processor 11. The BIM program 13a is configured with REVIT as an example of a BIM. Note that the BIM program 13a may also be configured with other programs such as Archicad (registered trademark) or GLOBE (registered trademark).

[0019] The BIM program 13a is software for integrated management of digital information related to the design, construction, and management of a structure. The BIM program 13a can create a three-dimensional digital model (design model 33a). Each part of the design model 33a is associated with attribute information stored in secondary memory, such as the type of building material, specifications, performance, cost, construction procedure, and maintenance information. This allows the design model 33a to be used throughout the lifecycle of the design, construction, and management of a structure.

[0020] The design support device 2 comprises a processor 21, a communication IF 22, a memory 23, an input / output IF 24, and a display 25. The processor 21, the communication IF 22, the memory 23, and the input / output IF 24 have the same basic configuration as the processor 11, the communication IF 12, the memory 13, and the input / output IF 14 of the BIM device 1, so a brief description thereof will be omitted. Note that the design support device 2 comprises other functional units, but only the functional units related to this embodiment are described here.

[0021] An add-in program 23a is stored in the memory 23. The add-in program 23a is stored in a secondary memory (storage unit 23A) of the memory 23, and is read into the main memory and executed by the processor 21. The add-in program 23a is a design support program that is sent to the BIM device 1 via the communication IF 22 and the network 4, and adds functions to the BIM program 13a. The add-in program 23a in this embodiment has an extended function that constructs a design model 33a by executing a script generated by the generation AI 5. Details of the add-in program 23a will be described later.

[0022] The display 25 is provided in the design support device 2 and is a display device that displays text, images, videos, and other visual information. This display 25 can display a BIM by executing the BIM program 13a in which the add-in program 23a is implemented by the processor 21. The display 25 is configured with an LCD display, an LED display, an OLED display, a plasma display, or the like. Note that the display 25 is not particularly limited as long as it is hardware that can execute the BIM program 13a in which the add-in program 23a is implemented by the processor 21 and display the design model 33a.

[0023] The user device 3 includes a processor 31, a communication IF 32, a memory 33, an input / output IF 34, a display 35, and an operation unit 36. The processor 31, the communication IF 32, the memory 33, the input / output IF 34, and the display 35 have the same basic configuration as the processor 21, the communication IF 22, the memory 23, the input / output IF 24, and the display 25 of the design support device 2, so a detailed description thereof will be omitted. Note that the user device 3 includes other functional units, but only the functional units related to this embodiment are described here.

[0024] A design model 33a created in the early stages of a project, particularly in the estimation stage when time is limited, is stored in the memory 33. The operation unit 36 ​​is constructed with at least one element selected from the group consisting of a touch switch, a keyboard, a mouse, a scanner, and a voice input speaker. When a user operates the operation unit 36, a signal corresponding to the operation is input to the processor 31.

[0025] (Configuration of design support device) 2 shows a block diagram of a design support device 2 that supports the construction of a design model 33a. The design support device 2 can exchange information with a generation AI (Artificial Intelligence) 5. The generation AI 5 includes any language model or large language model (LLM), such as ChatGPT, Gemini, or Claude. The design support device 2 includes a storage unit 23A and an add-in program 23a.

[0026] 2 functions as a secondary memory of the above-mentioned memory 23. The storage unit 23A stores a logic database DBa and a design model database DBb.

[0027] The logic database DBa contains design rules for constructing a provisional design model by inputting drawing data 23b of the structure and specification data 23c of multiple parts to be placed in the structure. These design rules are scripts (program code such as Dynamo Python) used by REVIT in JSON (JavaScript Object Notation) format, with DB column names for each part. The parts include structural equipment such as columns and beams, exterior and interior equipment such as windows, doors, and wall materials, accessories such as sleeves, and facility equipment such as indoor and outdoor units of air conditioners. The design model database DBb contains design models 33a generated using the generation AI5.

[0028] 2, the add-in program 23a includes a setting unit 231, an acquisition unit 232, a logic design unit 233, a feedback control unit 234, a display control unit 235, a learning unit 236, and an output unit 237. The add-in program 23a is an API (Application Programming Interface) or plug-in system that adds these functional units to the BIM program 13a.

[0029] The setting unit 231 sets an input field IP (see FIG. 4 ) on the screen (display 35) of the CAD system 100 (user device 3). As shown in FIG. 4 , the setting unit 231 may automatically display the input field IP on the BIM browser screen displayed on the display 35 of the user device 3, or may provide a display button for the input field IP on the BIM browser screen. That is, the setting unit 231 inserts an add-in program 23 a for setting the input field IP into the initial screen startup program of the BIM program 13 a. In this input field IP, the user can input natural language, drawing data 23 b of the structure including images or files, specification data 23 c of multiple parts to be arranged in the structure, and constraints 23 d. Input into the input field IP is performed via an operation unit 36 ​​consisting of at least one element of a touch switch, keyboard, mouse, scanner, voice input speaker, etc. Here, natural language includes words such as keywords, word groups combining multiple words, phrases consisting of words and associated words, sentences consisting of combinations of multiple phrases, etc. The image may be a still image or a video. The files include data files (CAD drawings, vector PDFs, BIM data, etc.), project data (e.g. PowerPoint), Excel files, etc.

[0030] In this embodiment, the structure drawing data 23b, including natural language, images, or files, is a design model and a structural model composed of two-dimensional or three-dimensional data. A design model is a model designed primarily based on the structure's design, including its exterior shape, spatial configuration, materials, and finishes. A structural model is a model designed primarily based on the structure's structural components (columns, beams, walls, floors, foundations, etc.). In other words, the structure drawing data 23b is spatial data that defines the spatial space in which internal equipment, such as auxiliary equipment like sleeves and equipment like indoor and outdoor air conditioner units, can be arranged. The specification data 23c of multiple parts, including natural language, images, or files, is structured data that compiles the specifications (equipment name, model number, dimensions, diameter, location, installation conditions, constraints, etc.) of the equipment used to construct the design model 33a. In this embodiment, the drawing data 23b is a design model and a structural model composed of two-dimensional data (CAD drawings or PDF drawings) created by a user (designer) during the estimation stage, but it may also be three-dimensional data (CAD drawings or BIM data).

[0031] The laws of the area where the structure is located and the design conditions desired by the client may be input as constraints 23d in the input field IP. The laws of the area where the structure is located include text data such as national laws and ordinances established by the local government of the construction area. The design conditions desired by the client include text data such as required specifications that the client presents as constraints 23d in bidding, etc. For example, the input field IP can be input in natural language, such as "Compliant with the Building Standards Act, ceiling height of 2.7m or more, maintenance space of 600mm or more."

[0032] The acquiring unit 232 acquires attribute information about a plurality of types of parts based on drawing data 23b of the structure input to the CAD system 100 (user device 3) and specification data 23c of a plurality of parts to be arranged in the structure. In other words, the acquiring unit 232 acquires attribute information for each part stored in the memory 33 of the CAD system 100 (user device 3) as attribute information about a plurality of types of parts for constructing the design model 33a, and transmits the attribute information to the BIM apparatus 1.

[0033] The content entered in the input field IP is sent to the design support device 2 via the BIM device 1 or directly from the user device 3, and the logic design unit 233 constructs a provisional design model in accordance with predetermined design rules (scripts) based on the content of the input field IP, and the feedback control unit 234 executes feedback control on this provisional design model to construct a design model 33a. "Contents of the input field IP" includes cases where the natural language, image or file entered in the input field IP is used as is, and cases where this natural language, image or file is processed into one suitable for script generation.

[0034] The natural language, image, or file entered in the input field IP may include script modification instructions for the drawing data 23b or the specification data 23c. The script modification instructions are natural language instructions for modifying at least one of the quantity, dimensions, and materials of parts, such as "reduce the number of pillars and increase their thickness," "make the wall 1 meter higher," "add three doors," or "change the exterior wall material to concrete." The generation AI 5 generates the modification script based on the natural language instructions. In this case, the acquisition unit 232 acquires parts corresponding to the script modification instructions from data stored in the memory 33 of the CAD system 100 (user device 3) or from web data. If the script modification instructions are for BIM parts provided in advance by the BIM device 1 or acquired by the user from an external source, the acquisition unit 232 acquires the parts from the data stored in the memory 33. On the other hand, if the script modification instructions are for BIM parts that are published as web data but not stored in the memory 33, the acquisition unit 232 acquires the parts from the web data.

[0035] It is preferable that the script modification instruction is directed only to the part selected via the operation unit 36 ​​of the user device 3, or only to the part based on the condition entered in the input field IP. The condition entered in the input field IP is a condition for modifying the part in natural language, such as "among 10 identical sleeves, please change the length of any that is 400 or more to 380."

[0036] The logic design unit 233 constructs a provisional design model including the positions of internal equipment to be placed in the design model and structural model and basic routes for piping and ducts, in accordance with the logic (script) defined by the design rules (e.g., shortest distance, use of standard components, etc.) stored in the logic database DBa, based on the contents of the drawing data 23b and specification data 23c entered into the input field IP via the operation unit 36 ​​of the user device 3.

[0037] As shown in Fig. 5, the logic design unit 233 designs a basic piping route as an initial design model based on the drawing data 23b and specification data 23c input in the input field IP, and detects interference points as indicated by dashed lines. In the example shown in Fig. 5, interference points are being detected, and when the logic design unit 233 constructs a provisional design model, points where the piping interferes with the beams are detected as indicated by dashed lines in Fig. 6. Note that the logic design unit 233 may sequentially construct provisional design models while receiving feedback control from the feedback control unit 234 during the execution of logic defined by the design rules.

[0038] The feedback control unit 234 executes feedback control on the provisional design model constructed by the logic design unit 233, and constructs the design model 33a. Specifically, the feedback control unit 234 executes a script generated by the generation AI 5 to feedback control the provisional design model so that multiple parts included in the provisional design model do not interfere with each other. This script is program code (Dynamo Python, etc.) that constructs the design model 33a.

[0039] The feedback control unit 234 may execute a script generated by the generation AI 5 based on attribute information of each part, including at least one of the quantity, dimensions, and material of the part, to perform feedback control while extracting 3D data corresponding to the attribute information of each part in the design model 33a. This script is program code (Dynamo Python, etc.) that extracts a parts group image including an assembly of each part corresponding to the attribute information (for example, a parts group image of pipes and beams as shown in FIG. 6).

[0040] The feedback control unit 234 may execute the change script generated by the generation AI 5 based on a correction instruction input into the input field IP from the operation unit 36 ​​of the user device 3. This change script is program code (such as Dynamo Python) that executes the correction, deletion, and addition of parts to the design model 33a. For example, if "Ensure a ceiling height of 2.7 m or more" is input but the ceiling height is less than 2.7 m, the generation AI 5 generates a change script based on a correction instruction input into the input field IP, such as "Please correct the ceiling height to 2.7 m or more," and executes this change script to build the corrected design model 33a.

[0041] The display control unit 235 displays the input field IP set by the setting unit 231 and the design model 33a constructed by the logic design unit 233 and the feedback control unit 234 on the display 35 (see FIGS. 4 and 7). The display control unit 235 displays the design model 33a constructed by the logic design unit 233 and the feedback control unit 234 by stacking multiple parts in order, based on the content of the natural language, image, or file input into the input field IP, on the display 35. Specifically, the display control unit 235 activates the display control function of the BIM device 1 to display the design model 33a on the display 35 (the same applies hereinafter).

[0042] The display control unit 235 may display on the display 35 a parts group image including an assembly of parts corresponding to the attribute information based on a search instruction having attribute information including at least one selected from the quantity, dimensions, and material of the parts (see FIG. 7). For example, by inputting "Please display the part where the interference has been resolved" in the input field IP, a parts group image as an assembly consisting of 3D images of beams and pipes shown in FIG. 7 is displayed on the display 35. The display control unit 235 may also switch between a 3D view and a 2D view so that the "arrangement order of parts" and the "layer structure" can be seen. Furthermore, the display control unit 235 may provide operation buttons on the display 35 to enable "redo" and "partial undo."

[0043] The display control unit 235 may display the script generated by the generation AI 5 based on the content of the natural language, image, or file entered in the input field IP. For example, the flow from "design instructions" to "generated script" to "preview" may be visualized. At this time, the display control unit 235 may display an explanation of the script generated by the generation AI 5. The display control unit 235 may also prepare a dashboard-like screen that displays a summary of the script or only important parameters in an easy-to-read format.

[0044] The display control unit 235 may display the script that has been changed based on the correction instructions entered in the input field IP via the operation unit 36 ​​of the user device 3. This changed script may have a comment field displayed for each line, or a screen that allows comparison of the before and after changes may be provided. Parameters such as "part name," "dimensions," and "material" may be highlighted so that it is possible to see at a glance what has changed. Then, the display control unit 235 displays the design model 33a generated based on the changed script on the display 35. At this time, an animation may be used to visually indicate which parts have been added to which positions.

[0045] The learning unit 236 inputs the revision history included in the design model 33a stored in the design model database DBb into the generation AI 5, and generates a trained model by performing deep learning using the design model 33a and a list related to the placement spaces of parts as training data. The learning unit 236 also references the revision history included in the design model 33a in real time or at predetermined intervals, allowing the generation AI 5 to learn without user intervention. Furthermore, the learning unit 236 can learn the design information of each part in the accumulated design model 33a, refer to and automatically present best practices and design examples for each learned part, and perform design optimization using previous design rules.

[0046] The learning unit 236 may learn the design rules of each part in the design model 33a stored in the design model database DBb. These design rules are stored in the logic database DBa in association with the script generated by the generation AI 5. The learning unit 236 can learn the design rules of each part in the stored design model 33a and evaluate the validity of the parameters and structure of the design model 33a.

[0047] The output unit 237 can output the list of part placement spaces and the design model 33a via the communication IF 22 and the input / output IF 24. The list of part placement spaces, as shown in FIG. 8, is composed of a list of specific dimensions and reports regarding the space to be secured, such as the dimensions of ducts and beams, ceiling height, PS (pipe space) and EPS (electrical pipe space), AHU (air conditioner), pump-related equipment, electrical panels, and server rack rows. The data output format is not limited to Excel files, and various output formats are available, such as CSV files and JSON (JavaScript Object Notation) formats. Outputting such a list allows the design model 33a along with the list of part placement spaces to be imported into other systems, making it highly versatile.

[0048] When outputting the design model 33a, the output unit 237 may output interference points, space efficiency, constructability score, maintainability, and cost efficiency as an "optimization result summary," as shown in FIG. 7. This "optimization result summary" may be calculated using a trained model generated by the learning unit 236, or may be calculated using a generated AI5 trained by the learning unit 236. In addition, the output unit 237 may display a list related to the design model 33a and the placement spaces of parts on the display 35 via the display control unit 235. For example, as shown in FIG. 7, the display control unit 235 displays details of the optimized equipment placement and the space allocation status.

[0049] (Design support method) 3 shows a flowchart of a design support method executed by the design support device 2, which can be used to add functions to the CAD system 100. The design support method shown in FIG. 3 is merely a representative example, and other flowcharts for executing each functional unit of the design support device 2 in the above-described embodiment are omitted. In this embodiment, an example is shown in which information from the user device 3 is transmitted and received to the design support device 2 via the BIM apparatus 1, but information from the user device 3 may be transmitted and received directly to the design support device 2, or information may be transmitted and received only between the user device 3 and the BIM apparatus 1 with an add-in program 23a pre-installed in the BIM program 13a of the BIM apparatus 1.

[0050] As shown in FIG. 3, the add-in program 23a of the design support device 2 is sent to the BIM device 1, the add-in program 23a is installed in the BIM program 13a, and the BIM program 13a with the installed add-in program 23a is sent to the user device 3 (#21). This allows a BIM with an input field IP to be displayed on the display 35 of the user device 3, as shown in FIG. 4. Next, the user (designer) uploads the drawing data 23b and specification data 23c for the estimate drawings to the input field IP (#31), and enters constraints 23d as necessary (#32). Based on the drawing data 23b and specification data 23c uploaded to the input field IP via the operation unit 36 ​​of the user device 3, the logic design unit 233 constructs a provisional design model including the layout of internal equipment and basic piping and duct routes in accordance with the logic (program code) defined by the design rules (e.g., shortest distance, use of standard components, etc.) stored in the logic database DBa (#22).

[0051] The process of constructing this provisional design model is sent in real time from the design support device 2 to the user device 3, and the display control unit 235 activates the display control function of the BIM device 1 to display the provisional design model on the display 35 of the user device 3. As shown in Figure 5, the logic design unit 233 designs a basic piping route as an initial design model based on the drawing data 23b and specification data 23c entered in the input field IP, and detects interference points as shown by the dashed lines. In the example shown in Figure 5, interference points are being detected, and when the logic design unit 233 constructs the provisional design model, points where the piping interferes with the beams are detected as shown by the dashed lines in Figure 6.

[0052] Next, the feedback control unit 234 executes the script generated by the generation AI 5, taking into account the constraints 23d, to prevent multiple parts included in the provisional design model from interfering with each other, and constructs a design model 33a by feedback-controlling the provisional design model (#23). The construction process of this design model 33a is sent in real time from the BIM device 1 to the user device 3, and the display control unit 235 activates the display control function of the BIM device 1 to display the design model 33a on the display 35 of the user device 3 (#33). In the example shown in FIG. 7, the design model 33a is displayed together with interference points, space efficiency, constructability score, maintainability, and cost efficiency as an "optimization result summary," and details of the optimized equipment layout and the space allocation status are also displayed. Next, the learning unit 236 stores the script (program code) associated with the finally completed design model 33a in the logic database DBa and learns it as design rules or feedback control logic for each part of the design model 33a stored in the design model database DBb (#24).

[0053] To display the required parts in the design model 33a constructed in this way, the user enters natural language, an image, or a file in the input field IP to issue a search command including attribute information (#34). Next, the natural language, image, or file entered in the input field IP is sent from the user device 3 to the design support device 2 via the BIM device 1 or directly. The design support device 2 then executes the script generated by the generation AI 5 based on the attribute information of each part, including at least one of the part's quantity, dimensions, and material (#25). For example, as shown in FIG. 7, by entering "Please display the part where the interference has been resolved" in the input field IP "Fuzzy Search," a parts group image is displayed as an aggregate consisting of 3D images of beams and pipes. Next, the modified design model 33a is sent from the BIM device 1 to the user device 3, and the display control unit 235 activates the display control function of the BIM device 1 to display the design model 33a consisting of the parts group image on the display 35 of the user device 3 (#35).

[0054] In the above-described embodiment, the following configurations are envisioned. (1) The design support device 2 according to this embodiment is a design support device 2 that constructs a design model 33a in which a plurality of parts are arranged on a structure, and includes: a logic design unit 233 that constructs a provisional design model in which a plurality of parts are arranged on a structure in accordance with predetermined design rules by inputting drawing data 23b of the structure and specification data 23c of the plurality of parts; a feedback control unit 234 that uses a generation AI 5 to feedback control the provisional design model so that the plurality of parts included in the provisional design model do not interfere with each other when the logic design unit 233 constructs the provisional design model; and an output unit 237 that outputs the provisional design model constructed by the logic design unit 233 and the feedback control unit 234 as a design model 33a.

[0055] In the design support device 2 according to this embodiment, the logic design unit 233 constructs a provisional design model in accordance with predetermined design rules, while the feedback control unit 234 feedback-controls the provisional design model using the generation AI 5 so that multiple parts included in the provisional design model do not interfere with each other. In other words, while the logic design unit 233 quickly constructs a rough provisional design model, the generation AI 5 feedback-controls each time so that multiple parts do not interfere with each other, thereby improving the accuracy of the automatic design of the design model 33a.

[0056] As a result, problems such as interference between parts or lack of necessary space being discovered in later processes are eliminated, large-scale design changes and rework are not required, and problems such as delays in construction time and increased construction costs are also eliminated. In this way, the design support device 2 is able to quickly build a design model 33a that does not require rework from the estimation stage.

[0057] (2) In the design support device 2 of (1), it is preferable that the output unit 237 outputs a list regarding the arrangement spaces of the parts.

[0058] As in this embodiment, by outputting a list of part placement spaces, it becomes possible to record an optimized space list, which can be used as a design basis or manually corrected if there are any errors.

[0059] (3) It is preferable that the design support device 2 of (2) further comprises a learning unit 236 that trains the generated AI 5 based on the design model 33 a and the list output by the output unit 237.

[0060] As in this embodiment, if the generated AI 5 is trained based on the design model 33a and the list output by the output unit 237, the accuracy of the automatic design of the design model 33a can be further improved.

[0061] (4) In the design support device 2 of any one of (1) to (3), the feedback control unit 234 preferably inputs the laws of the area where the structure is located and the design conditions desired by the client to the generation AI 5.

[0062] In this embodiment, the generation AI 5 analyzes and feedback-controls the constraints 23d that need to be taken into consideration, such as the laws and regulations of the construction site and the client's required specifications, so large-scale design changes and rework are not required, eliminating problems such as delays in construction time and increased construction costs.

[0063] (5) In the design support device 2 of any one of (1) to (4), it is preferable that the drawing data 23b is two-dimensional data and the design model 33a is three-dimensional data.

[0064] In this embodiment, if the drawing data 23b input to the logic design unit 233 is two-dimensional data and the design model 33a output by the output unit 237 is three-dimensional data, an accurate design model 33a can be constructed from, for example, a two-dimensional drawing created by an expert.

[0065] [Other embodiments] (a) In the above-described embodiment, the design support device 2 is described as being added to the BIM device 1, but the design support device 2 may be added to AutoCAD or open source CAD. (b) Some of the steps in the above-described embodiments may be omitted, or may be combined appropriately to implement the functions. [Industrial Applicability]

[0066] The present disclosure can be used in a design support device that builds a design model in which a plurality of parts are arranged in a structure. [Explanation of symbols]

[0067] 2: Design support device, 233: Logic design unit, 234: Feedback control unit, 236: Learning unit, 237: Output unit, 33a: Design model, 5: Generative AI

Claims

1. A design support device for constructing a design model in which a plurality of parts are arranged in a structure, a logic design unit that executes a script of predetermined design rules by inputting a plurality of drawing data including a design model relating to the design of the structure and a structural model relating to the structural parts of the structure as separate drawings, and specification data of a plurality of the parts, to construct a provisional design model in which a plurality of the parts are arranged on the structure; a feedback control unit that, when the logic design unit constructs the provisional design model, causes a generation AI to feedback-control the provisional design model so that the plurality of parts included in the provisional design model do not interfere with each other; and an output unit that outputs the provisional design model constructed by the logic design unit and the feedback control unit as the design model.

2. A design support device for constructing a design model in which a plurality of parts are arranged in a structure, a logic design unit that constructs a provisional design model in which the parts are arranged on the structure in accordance with predetermined design rules by inputting drawing data of the structure and specification data of the parts; a feedback control unit that, when the logic design unit constructs the provisional design model, causes a generation AI to feedback-control the provisional design model so that the plurality of parts included in the provisional design model do not interfere with each other; and an output unit that outputs the provisional design model constructed by the logic design unit and the feedback control unit as the design model and also outputs a list regarding the placement spaces of the parts; A design support device comprising: a learning unit that trains the generating AI based on the design model and the list output by the output unit.

3. 3. The design support device according to claim 1, wherein the feedback control unit inputs the laws of the area where the structure is located and the design conditions desired by the client to the generation AI.

4. 3. The design support device according to claim 1, wherein the drawing data is two-dimensional data, and the design model is three-dimensional data.

Images