BIM data analysis system

Abstract

To provide a BIM (Building Information Modeling) data analysis system for predicting building product data included in BIM data.SOLUTION: A system includes: a storage unit which stores BIM data and feature vectors of building product data; a model generation unit which generates, by machine learning, an estimation model which estimates feature vectors of second specific building product data in the building product data, from feature vectors of at least one piece or more of first specific building product data in the building product data, using the stored feature vectors of the building product data as training data; a BIM data receiving unit which receives an input of BIM data; a feature acquisition unit which acquires a feature vector for each building product data from the input BIM data; and a feature generation unit which generates feature vectors of the second specific building product data from the feature vectors of the at least one piece or more of first specific building product data acquired by the feature acquisition unit, using the estimation model.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to a BIM data analysis system. [Background technology]

[0002] BACKGROUND ART A system is known that utilizes and analyzes BIM (Building Information Modeling) data including building product data in order to grasp a three-dimensional model of a building on a computer (for example, Patent Document 1). [Patent Document 1] Patent No. 7129586 Summary of the Invention [Problem to be solved by the invention]

[0003] We provide a BIM data analysis system that predicts building product data contained in BIM data. [Means for solving the problem]

[0004] A first aspect of the present invention provides a BIM data analysis system for analyzing BIM data including one or more building product data. The BIM data analysis system may include a memory unit. The memory unit may store one or more BIM data and feature vectors of one or more building product data included in the one or more BIM data. The BIM data analysis system may include a model generation unit. The model generation unit may use the feature vectors of the one or more building product data stored in the memory unit as training data to generate an estimation model by machine learning that estimates a feature vector of second specific building product data among the building product data from the feature vector of at least one or more first specific building product data among the building product data. The BIM data analysis system may include a BIM data acceptance unit. The BIM data acceptance unit may accept input of BIM data. The BIM data analysis system may include a feature amount acquisition unit. The feature amount acquisition unit may acquire feature amounts for each building product data from the BIM data input to the BIM data acceptance unit. The BIM data analysis system may include a feature amount generation unit. The feature generation unit may use the estimated model generated by the model generation unit to generate a feature vector of the second specific building product data from the feature vectors of at least one or more first specific building product data acquired by the feature acquisition unit.

[0005] The BIM data analysis system may include a BIM data output unit, which may update the building product data from the feature vector of the second specific building product data generated by the feature amount generation unit, and generate BIM data including the updated building product data.

[0006] The feature vector may include a modification count, which is the number of times the building product data has been modified.

[0007] The second specific building product data may be data relating to the same building product as the first specific building product data. If the number of revisions of the first specific building product data is N, the number of revisions of the second specific building product data may be N+1.

[0008] The feature vector may include a list of building products linked to the building product associated with the building product data. The second specific building product data may be data related to building products linked to the building product associated with the first specific building product data.

[0009] The model generation unit may use the feature vectors of one or more building product data stored in the storage unit as training data and generate an estimation model by machine learning that estimates the feature vectors of the second specific building product data from the feature vectors of the plurality of first specific building product data.The feature generation unit may use the estimation model generated by the model generation unit to generate the feature vectors of the second specific building product data from the feature vectors of the plurality of first specific building product data acquired by the feature acquisition unit.

[0010] The storage unit may store the BIM data in a modifiable manner, and the BIM data receiving unit may receive the BIM data modified by the storage unit.

[0011] The building product related to the building product data may be composed of a beam, a column, a bracket, a diaphragm, a brace, or a damper.

[0012] The feature vector may include at least one of the following: feature quantities in the cross section of the beam or column; thicknesses of brackets, diaphragms, braces or dampers; lateral lengths of brackets; spacing of diaphragms; weight of the beam or column; overall length of the beam or column; overall width of the beam or column; overall height of the beam or column; and diagonal length of the beam or column.

[0013] The feature vector may include at least one of the product serial number of the building product related to the building product data, building information of the BIM data including the building product data, and product information of the building product related to the building product data. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a block diagram showing an example of the configuration of a system 100 including a BIM data analysis system 20 according to one embodiment. [Figure 2]2 is an example of a hardware configuration of the terminal device 10. [Figure 3] 1 is an example of a hardware configuration of a BIM data analysis system 20. [Figure 4] 2 is a block diagram schematically illustrating an example of functions of the terminal device 10 illustrated in FIG. 1. FIG. [Figure 5] FIG. 2 is a block diagram schematically illustrating an example of the functions of the BIM data analysis system 20 shown in FIG. [Figure 6] FIG. 10 is a diagram illustrating an example of a determination feature amount. [Figure 7] FIG. 10 is a diagram illustrating an example of feature amounts in a cross section of a beam. [Figure 8] FIG. 10 is a diagram illustrating an example of feature amounts in a cross section of a pillar. [Figure 9] FIG. 1 is a flow diagram showing prediction of building product data included in BIM data performed in the system 100. [Figure 10] FIG. 10 is a diagram illustrating an example of ST306 and ST308. [Figure 11] FIG. 10 is a diagram illustrating an example of ST302 and ST304. [Figure 12] FIG. 10 is a diagram illustrating an example of ST308, ST310, and ST312. [Figure 13] FIG. 10 is a diagram illustrating another example of ST302 and ST304. [Figure 14] FIG. 10 is a diagram illustrating another example of ST308, ST310, and ST312. [Figure 15] 10 is a diagram showing an example of a display screen that the display control unit 270 causes the display unit 110 of the terminal device 10 to display. FIG. [Figure 16] 10 is a diagram showing another example of a display screen that the display control unit 270 causes the display unit 110 of the terminal device 10 to display. FIG. [Figure 17] 10 is a diagram showing another example of a display screen that the display control unit 270 causes the display unit 110 of the terminal device 10 to display. FIG. [Figure 18] 10 is a diagram showing another example of a display screen that the display control unit 270 causes the display unit 110 of the terminal device 10 to display. FIG. [Figure 19]FIG. 1 is a flow diagram showing additional learning of a trained model performed in system 100. DETAILED DESCRIPTION OF THE INVENTION

[0015] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0016] Various embodiments of the present invention will be described below with reference to the accompanying drawings. Common components in the drawings are designated by the same reference numerals. It should also be noted that components shown in one drawing may be omitted in another drawing for clarity. It should also be noted that the accompanying drawings are not necessarily drawn to scale. Furthermore, the term "application" may refer to software or a program, and may be any set of instructions for a computer that can be combined to produce a certain result.

[0017] 1. Configuration of System 100 1 is a block diagram showing an example of the configuration of a system 100 including a BIM data analysis system 20 according to one embodiment. The system 100 includes the BIM data analysis system 20 and a terminal device 10. The BIM data analysis system 20 may be capable of communicating with the terminal device 10 via a communication line 30. The communication line 30 may include, but is not limited to, a mobile phone network, a wireless LAN, a landline network, the Internet, an intranet, Ethernet (registered trademark), or the like.

[0018] In this example, the system 100 is realized by a plurality of computers (terminal devices). The terminal devices 10 may use functions provided by a BIM data analysis system 20 on a communication line 30 via the communication line 30. The BIM data analysis system 20 is, for example, a cloud web service. Although not shown, the system 100 may also be realized by a single computer (terminal device).

[0019] BIM data is a three-dimensional (solid) digital model of a building created on a computer, with added attribute data such as cost, finish, management data, and building information. BIM data includes one or more building product data. Building product data is data on the building products that make up the building related to the BIM data. A building product is a unit product shipped from a factory. A building product may be a product that can be combined to construct a building. A building product is composed of components such as beams, columns, brackets, diaphragms, braces, or dampers. A building product may be a combination of one or more of beams, columns, brackets, diaphragms, braces, and dampers.

[0020] The BIM data analysis system 20 analyzes BIM data including one or more building product data. A user can analyze the BIM data using the BIM data analysis system 20 by operating the terminal device 10. Analyzing the BIM data may mean, for example, a user checking information about the BIM data via the BIM data analysis system 20. Analyzing the BIM data may mean, for example, a user making predetermined changes to the BIM data via the BIM data analysis system 20.

[0021] When the BIM data analysis system 20 and the terminal device 10 operate as terminal devices for analyzing BIM data, they can execute an installed BIM data analysis application (which may be middleware or a combination of an application and middleware; the same applies below). This allows the BIM data analysis system 20 and the terminal device 10 to analyze BIM data.

[0022] The terminal device 10 and the BIM data analysis system 20 may be any terminal device capable of performing such operations, including, but not limited to, a smartphone, a tablet, a mobile phone (feature phone), or a personal computer.

[0023] 2. Hardware configuration of each device Next, an example of the hardware configuration of each of the terminal device 10 and the BIM data analysis system 20 will be described.

[0024] 2-1. Hardware configuration of terminal device 10 Fig. 2 is an example of the hardware configuration of the terminal device 10. Fig. 2 is a block diagram that schematically shows an example of the hardware configuration of the terminal device 10 shown in Fig. 1. In explaining the hardware configuration, the BIM data analysis system 20 and the terminal device 10 will be referred to as "each device."

[0025] 2, the terminal device 10 mainly includes a central processing unit 11, a main memory device 12, an input / output interface device 13, an input device 14, an auxiliary memory device 15, and an output device 16. These devices are connected to each other by a data bus or a control bus.

[0026] The central processing unit 11 is referred to as a "CPU" and can perform operations on instructions and data stored in the main memory 12 and store the results of the operations in the main memory 12. Furthermore, the central processing unit 11 can control an input device 14, an auxiliary memory device 15, an output device 16, etc. via an input / output interface device 13. The terminal device 10 can include one or more such central processing units 11.

[0027] The main memory device 12 is referred to as "memory" and can store instructions and data received from the input device 14, the auxiliary memory device 15, and the communication line 30 (BIM data analysis system 20) via the input / output interface device 13, as well as the results of calculations by the central processing unit 11. The main memory device 12 can include, but is not limited to, RAM (random access memory), ROM (read only memory), flash memory, etc.

[0028] The auxiliary storage device 15 is a storage device having a larger capacity than the main storage device 12. It can store instructions and data (computer programs) that constitute the specific application (BIM data analysis application) described above. Furthermore, the auxiliary storage device 15 is controlled by the central processing unit 11 to transmit these instructions and data (computer programs) to the main storage device 12 via the input / output interface device 13. The auxiliary storage device 15 can include, but is not limited to, a magnetic disk device, a magnetic tape, an optical disk device, a flash memory device, or the like.

[0029] The input device 14 is a device that inputs data from the outside and may include, but is not limited to, a touch panel, a button, a keyboard, a mouse, a sensor, etc. The sensor may include, but is not limited to, one or more cameras, one or more microphones, etc.

[0030] The output device 16 may include, but is not limited to, a display device, a touch panel, a printer device, or the like.

[0031] In such a hardware configuration, the central processing unit 11 can sequentially load instructions and data (computer programs) that constitute a specific application stored in the auxiliary storage device 15 into the main storage device 12 and perform operations on the loaded instructions and data. This allows the central processing unit 11 to control the output device 16 via the input / output interface device 13, or to send and receive various information to and from other devices (for example, the BIM data analysis system 20) via the input / output interface device 13 and the communication line 30.

[0032] It should be noted that the terminal device 10 may include one or more microprocessors or graphics processing units (GPUs) instead of or in addition to the central processing unit 11.

[0033] 2-2. Hardware configuration of BIM data analysis system 20 3 is an example of a hardware configuration of the BIM data analysis system 20. As the hardware configuration of the BIM data analysis system 20, for example, the same configuration as the hardware configuration of the terminal device 10 described above can be used.

[0034] 3, the BIM data analysis system 20 mainly includes a central processing unit 21, a main memory device 22, an input / output interface device 23, an input device 24, an auxiliary memory device 25, and an output device 26. These devices are connected to each other by a data bus or a control bus.

[0035] The central processing unit 21, main memory device 22, input / output interface device 23, input device 24, auxiliary memory device 25 and output device 26 can be approximately the same as the central processing unit 11, main memory device 12, input / output interface device 13, input device 14, auxiliary memory device 15 and output device 16 included in the terminal device 10 described above, respectively.

[0036] In such a hardware configuration, the central processing unit 21 can sequentially load instructions and data (computer programs) constituting a specific application stored in the auxiliary storage device 25 into the main storage device 22 and perform operations on the loaded instructions and data. This allows the central processing unit 21 to control the output device 26 via the input / output interface device 23, or to transmit and receive various information to and from other devices (e.g., the terminal device 10) via the input / output interface device 23 and the communication line 30.

[0037] 3. Functions of each device Next, an example of the functions of the terminal device 10 and the BIM data analysis system 20 will be described.

[0038] 3-1. Functions of the terminal device 10 An example of the functions of the terminal device 10 will be described with reference to Fig. 4. Fig. 4 is a block diagram that schematically shows an example of the functions of the terminal device 10 shown in Fig. 1.

[0039] 4, the terminal device 10 may mainly include a display unit 110, a communication unit 120, and a user interface unit 130. The terminal device 10 may also have other functions.

[0040] (1) Display section 110 The display unit 110 can display various information required for analyzing BIM data. Specifically, the display unit 110 can display a display screen based on information (screen information) transmitted from the BIM data analysis system 20. The screen information is information related to the display screen displayed by the terminal device 10. The display unit 110 may display the display screen on a display or the like provided in the terminal device 10.

[0041] (2) Communications Department 120 The communication unit 120 can communicate various information required for analyzing BIM data with the BIM data analysis system 20. For example, the communication unit 120 receives screen information related to the display screen displayed by the display unit 110 from the BIM data analysis system 20. The communication unit 120 can also transmit operation information acquired by the user interface unit 130 to the BIM data analysis system 20. The BIM data analysis system 20 can change the display screen displayed by the display unit 110 by changing the screen information related to the display screen based on the operation information.

[0042] (3) User Interface Unit 130 Various information required for analyzing BIM data is input to the user interface unit 130 through user operations. For example, operation information is input to the user interface unit 130 through user operations. The user interface unit 130 can output the operation information to the BIM data analysis system 20 via the communication unit 120.

[0043] 3-2. Functions of BIM Data Analysis System 20 An example of the functions of the BIM data analysis system 20 will be described with reference to Fig. 5. Fig. 5 is a block diagram schematically showing an example of the functions of the BIM data analysis system 20 shown in Fig. 1.

[0044] 5, the BIM data analysis system 20 can mainly include a storage unit 210, a model generation unit 220, a BIM data reception unit 230, a feature amount acquisition unit 240, a feature amount generation unit 250, a BIM data output unit 260, a display control unit 270, and a communication unit 280. The BIM data analysis system 20 may have functions other than these.

[0045] (1) Storage section 210 The storage unit 210 can store one or more BIM data and information related to one or more BIM data. The storage unit 210 may store multiple BIM data and information related to the multiple BIM data. In this example, the storage unit 210 can store one or more BIM data and feature vectors of one or more building product data included in the one or more BIM data. The BIM data and the feature vectors of the one or more building product data included in the BIM data are stored in association with each other. The information stored in the storage unit 210 is not limited to the above. The feature vector of the building product data includes one or more feature quantities of the building product data. Specific examples of the feature vector of the building product data will be described later.

[0046] The storage unit 210 may store the BIM data in a modifiable manner. That is, a user may modify the BIM data in the storage unit 210. When the BIM data is modified in the storage unit 210, the number of modifications (described later) among the feature amounts of the modified building product data included in the BIM data is updated.

[0047] (2) Model generation unit 220 The model generation unit 220 generates an estimation model by machine learning that estimates the feature vector of second specific building product data from the feature vector of one or more first specific building product data. That is, the model generation unit 220 can output the feature vector of second specific building product data as the objective variable by inputting the feature vector of one or more first specific building product data as the explanatory variable. The model generation unit 220 may generate the estimation model by machine learning using the feature vector of one or more building product data stored in the storage unit 210 as training data. Common machine learning techniques that can be used include logistic regression, random forest, decision tree, gradient boosting, support vector regression, linear regression, partial least squares (PLS) regression, Gaussian process regression, and neural network.

[0048] The first specific building product data and the second specific building product data are included in the building product data. A plurality of first specific building product data may be included in the building product data. That is, the estimation model may estimate the feature vector of the second specific building product data from the feature vectors of a plurality of first specific building product data. The estimation model may estimate the feature vector of the second specific building product data from the feature vector of one first specific building product data. The second specific building product data may be data relating to the same building product as the first specific building product data. Furthermore, the second specific building product data may be data relating to a building product different from the first specific building product data. Examples of the first specific building product data and the second specific building product data will be described later.

[0049] (3) BIM Data Reception Section 230 The BIM data accepting unit 230 accepts input of BIM data. A user may input BIM data to the BIM data analysis system 20 via the BIM data accepting unit 230. A user may upload BIM data to the BIM data accepting unit 230 (or the BIM data analysis system 20) via a terminal device 10. The input BIM data may be displayed on the user's terminal device 10 by a display control unit 270.

[0050] (4) Feature acquisition unit 240 The feature acquisition unit 240 acquires a feature vector for each piece of building product data from the BIM data input to the BIM data acceptance unit 230. The feature vector for each piece of product data may include a feature to be judged (hereinafter, judgment feature). The judgment feature is a feature checked when analyzing BIM data. The judgment feature may be automatically judged to be within a predetermined threshold value. The judgment feature may also be judged to be within a predetermined threshold value when a user inputs predetermined operation information into the user interface unit 130. If the judgment feature is not within the predetermined threshold value, the user may modify the BIM data so as to correct the judgment feature. The BIM data analysis system 20 may also be equipped with a judgment unit that judges the judgment feature.

[0051] When modifying BIM data, the user may modify the BIM data on the BIM data analysis system 20. For example, the user may modify the BIM data in the storage unit 210, and the BIM data receiving unit 230 may receive the BIM data modified in the storage unit 210. Alternatively, the user may modify the BIM data in a system other than the BIM data analysis system 20, and upload the BIM data to the BIM data receiving unit 230 (or the BIM data analysis system 20) via the terminal device 10.

[0052] 6 is a diagram showing an example of a determination feature. The determination feature includes, for example, a feature of a cross section of a beam or column included in the building product data, the thickness of a bracket, a diaphragm, a brace, or a damper, the horizontal length of a bracket included in the building product data, the spacing of a diaphragm, the material of a diaphragm or column, the clearance dimension between members, the weight of a building product (e.g., a beam or column), the total length, the total width, the total height or diagonal length, and the cross section of a beam or column. The determination feature may include other feature values.

[0053] The determination feature quantity includes, for example, a feature quantity in the cross section of a beam or column included in the building product data. The feature quantity in the cross section of a beam or column included in the building product data may include, for example, the thickness of a specific member in the cross section of the beam or column. The thickness in the cross section of the beam or column may be the thickness of a specific member in a specific direction. Furthermore, the feature quantity in the cross section of a beam or column included in the building product data may include dimensions related to the relative positions of each member present in the cross section.

[0054] 7A and 7B are diagrams showing an example of feature quantities in a cross section of a beam. Figures 7A and 7B show the cross section of the same beam. In this example, the beam is made up of an H-shaped steel beam 310 and a gusset plate 320.

[0055] 7(a), the dimension between the upper flange 312 of the H-beam 310 and the protruding portion 322 of the gusset plate 320 is indicated by L1. The dimension L1 between the upper flange 312 of the H-beam 310 and the protruding portion 322 of the gusset plate 320 is an example of a dimension relating to the relative positions of each member present in the cross section.

[0056] In Figure 7(b), the flange thickness of the H-beam 310 (the thickness of the end portion such as the upper flange 312) is represented by T1. In Figure 7(b), the web thickness of the H-beam 310 (the thickness of the central portion) is represented by T2. The flange thickness T1 and the web thickness T2 are examples of the thickness of a specific member in the cross section of a beam.

[0057] 8A, 8B, and 8C show examples of feature quantities in a cross section of a pillar. In this example, the pillar is made up of bracket 330, bracket 332, bracket 334, column 340, and diaphragm 350.

[0058] 8(a), the mounting positions of the brackets with respect to the column 340 and the diaphragm 350 are indicated by L2-a to L2-f. The mounting position L2 of each bracket is an example of a dimension relating to the relative position between each member present in the cross section.

[0059] 8(b), the dimensions by which the diaphragm 350 protrudes from the column 340 are represented by L3-a to L3-d. The dimension L3 by which the diaphragm 350 protrudes from the column 340 is an example of a dimension relating to the relative positions of the components present in the cross section.

[0060] 8(c), the thickness of the column 340 is represented by T3. The thickness T3 of the column 340 is an example of the thickness of a specific member in the cross section of the pillar.

[0061] Further, the determination feature includes, for example, the thickness of a bracket or diaphragm included in the building product data. The determination feature includes, for example, the horizontal length of a bracket included in the building product data. The determination feature includes, for example, the spacing of a diaphragm included in the building product data. The determination feature includes, for example, the material of a diaphragm or column included in the building product data. The determination feature includes, for example, the gap dimension between components included in the building product data. The determination feature includes, for example, the weight of a building product related to the building product data.

[0062] Furthermore, the determination feature quantity includes, as an example, the overall length, overall width, overall height, or diagonal length of the building product related to the building product data. In other words, the determination feature quantity includes the weight, overall length, overall width, overall height, or diagonal length of a beam or column. When orthogonal coordinate axes of the X-axis, Y-axis, and Z-axis are used, the overall length is, as an example, the length in the X-axis direction. When orthogonal coordinate axes of the X-axis, Y-axis, and Z-axis are used, the overall width is, as an example, the length in the Y-axis direction. When orthogonal coordinate axes of the X-axis, Y-axis, and Z-axis are used, the overall height is, as an example, the length in the Z-axis direction (height direction). When orthogonal coordinate axes of the X-axis, Y-axis, and Z-axis are used, if the overall width is y and the overall height is z (the overall length is x), the diagonal length L can be calculated using the following formula 1. Formula 1

[0063] TIFF2025126159000002.tif11167

[0064] Furthermore, the feature quantities for each product data may include feature quantities other than the judgment feature quantities. Feature quantities other than the judgment feature quantities are not feature quantities that are judged when analyzing BIM data. Feature quantities other than the judgment feature quantities may be modified by the user in the same way as the judgment feature quantities. The feature quantities for each product data may not include the judgment feature quantities, but may include only feature quantities other than the judgment feature quantities. The feature quantities for each product data may include only the judgment feature quantities.

[0065] An example of a feature other than the determination feature includes a product serial number of a building product related to the building product data. An example of a feature other than the determination feature includes building information of BIM data including building product data. An example of a feature other than the determination feature includes building information of BIM data including building product data. Building information of BIM data is information such as the square meters (unit of planar area) and number of floors of the building related to the BIM data. An example of a feature other than the determination feature includes product information of the building product related to the building product data. Product information of the building product related to the building product data is information such as the material, size, shape, position, weight, paint, and welding leg length of the building product. An example of a feature other than the determination feature includes information regarding a list of building products (connection parts) connected to the building product related to the building product data. An example of a feature other than the determination feature includes the number of revisions. The number of revisions is the number of times the building product data has been revised. The number of revisions includes not only the number of revisions made on the BIM data analysis system 20 but also the number of revisions made outside the BIM data analysis system 20.

[0066] (5) Feature generation unit 250 The feature generation unit 250 generates a feature vector of the building product data using the estimated model generated by the model generation unit 220. In this example, the feature generation unit 250 generates a feature vector of the second specific building product data from the feature vectors of at least one or more pieces of first specific building product data acquired by the feature acquisition unit 240 using the estimated model generated by the model generation unit 220.

[0067] (6) BIM data output unit 260 The BIM data output unit 260 generates BIM data. In this example, the building product data is updated from the feature vector of the second specific building product data generated by the feature amount generation unit 250, and BIM data including the updated building product data is generated. The BIM data generated by the BIM data output unit 260 may be downloadable from the BIM data analysis system 20. The output BIM data may be displayed on the user's terminal device 10 by the display control unit 270.

[0068] (7) Display control unit 270 The display control unit 270 causes the display screen to be displayed on the user's terminal device 10. The display control unit 270 may transmit screen information relating to the display screen to be displayed on the user's terminal device 10 via the communication unit 280. The display screen that the display control unit 270 causes the terminal device 10 to display will be described later.

[0069] (8) Communications Department 280 The communication unit 280 can communicate various information required for analyzing BIM data with the terminal device 10. For example, the communication unit 280 transmits to the terminal device 10 screen information related to a display screen to be displayed on the terminal device 10.

[0070] 4. Operation of BIM Data Analysis System 20 Next, the operation executed in the BIM data analysis system 20 having the above-described configuration will be described with reference to Fig. 9. Fig. 9 is a flow diagram showing prediction of building product data included in BIM data performed in the system 100. Before the flow is executed, an application is launched by the terminal device 10.

[0071] First, referring to FIG. 9, in step (hereinafter referred to as "ST") 302, the BIM data analysis system 20 (storage unit 210) stores feature vectors of BIM data and one or more pieces of building product data included in the BIM data.

[0072] Next, in ST304, the BIM data analysis system 20 (model generation unit 220) uses the feature vectors of the building product data as training data and generates an estimation model through machine learning that estimates the feature vectors of the second specific building product data from the feature vectors of the first specific building product data.

[0073] Furthermore, in ST306, the BIM data analysis system 20 (BIM data receiving unit 230) receives input of BIM data.

[0074] Then, in ST308, the BIM data analysis system 20 (feature amount acquisition unit 240) acquires a feature vector for each piece of building product data from the received BIM data.

[0075] Next, in ST310, the BIM data analysis system 20 (feature amount generation unit 250) generates a feature vector of the second specific building product data from the feature vector of the first specific building product data using the estimation model.

[0076] Then, in ST312, the BIM data analysis system 20 (BIM data output unit 260) updates the building product data from the generated feature vector of the second specific building product data, and generates BIM data including the updated building product data.

[0077] 10 is a diagram illustrating an example of ST306 and ST308. The storage unit 210 stores the feature vector acquired by the feature amount acquisition unit 240 as information related to the BIM data. Therefore, ST306 and ST308 may be executed before ST302. Note that the storage unit 210 may store, as information related to the BIM data, a feature vector directly input to the storage unit 210, rather than the feature vector acquired by the feature amount acquisition unit 240.

[0078] The storage unit 210 may store the BIM data received by the BIM data receiving unit 230. In this case, the storage unit 210 acquires features from the stored BIM data via the feature acquisition unit 240 and outputs the features to the model generation unit 220.

[0079] In FIG. 10, the memory unit 210 stores a group of feature vectors of building product data for property A (a property related to specific BIM data). The group of feature vectors of building product data for property A may include feature vectors of multiple building product data. In this example, the group of feature vectors of building product data for property A includes a feature vector of building product data a (building product data related to a specific building product, where building product data related to building product a is referred to as building product data a) and a feature vector of building product data b (building product data related to a specific building product, where building product data related to building product b is referred to as building product data b). The group of feature vectors of building product data for property A may include feature vectors of three or more building product data. As shown in FIG. 10, the feature vector of the building product data may include, for example, a product serial number, building information, product information, determination features (an example of which is shown in FIG. 6), a list of linked building products, and the number of revisions. Each building product data may include, as features, a product serial number, building information, product information, determination features, a list of linked building products, and the number of revisions.

[0080] (1) First Example The first embodiment will be described with reference to Fig. 11 and Fig. 12. Fig. 11 is a diagram illustrating an example of ST302 and ST304. In Fig. 11, the model generation unit 220 generates an estimation model by machine learning using a group of feature vectors of the building product data of property A as training data.

[0081] In the example shown in FIG. 11, each feature vector of the building product data of property A includes the number of revisions (the number of revisions is N). For example, in the group of feature vectors of the building product data of property A before revision, N=0. For example, in the group of feature vectors of the building product data of property A after the first revision, N=1. For example, in the group of feature vectors of the building product data of property A after the first revision, N=2. In the first example, by using the group of feature vectors of the building product data of property A before and after revision, which include the number of revisions as features, as training data, the estimation model generated by the model generation unit 220 can predict the group of feature vectors of the building product data of property A with different revisions (the estimation model that predicts the feature vectors of the building product data of property A with different revisions is referred to as estimation model 1). In other words, when the feature vectors of the building product data of property A are input, estimation model 1 outputs the feature vectors of the building product data of property A that reflect the expected revisions.

[0082] In the first embodiment, the first specific building product data is building product data before correction (for example, building product data a before correction), and the second specific building product data is building product data after correction (for example, building product data related to building product a after correction). For example, the second specific building product data is data related to the same building product as the first specific building product data. Furthermore, if the number of times the first specific building product data has been corrected is N, the number of times the second specific building product data has been corrected is N+1.

[0083] Note that the training data is not limited to the feature vectors of the building product data of a single property; feature vectors of the building product data of various properties can be used as training data. By using feature vectors of the building product data of various properties before and after modification, including the number of modifications as a feature, as training data, the estimation model 1 generated by the model generation unit 220 can predict feature vectors of the building product data of specific properties with different numbers of modifications. In other words, when the feature vectors of the building product data of a specific property are input, the estimation model 1 outputs the feature vectors of the building product data of the specific property that reflect the expected modifications.

[0084] 12 is a diagram illustrating an example of ST308, ST310, and ST312. The feature amount acquisition unit 240 acquires a feature vector of the building product data from the received BIM data. In this example, the feature amount acquisition unit 240 acquires a feature vector of the building product data related to building product a.

[0085] In this example, the feature generation unit 250 generates a feature vector of the corrected building product data a from the feature vector of the pre-correction building product data a using the estimation model 1. Then, the BIM data output unit 260 updates the building product data using the generated feature vector of the corrected building product data a, and generates BIM data including the updated building product data. As described above, by using the estimation model 1, it is possible to predict the next BIM data to be corrected.

[0086] (2) Second Example A second example will be described with reference to FIGS. 13 and 14. FIG. 13 is a diagram illustrating another example of ST302 and ST304. In the second example, the model generation unit 220 uses, as training data, feature vectors of one or more pieces of building product data stored in the storage unit 210, and generates, through machine learning, an estimation model that estimates feature vectors of second specific building product data from feature vectors of multiple pieces of first specific building product data. In FIG. 13, the model generation unit 220 generates, through machine learning, an estimation model using a group of feature vectors of building product data at the time of completion of multiple properties as training data. In this example, the model generation unit 220 generates, through machine learning, an estimation model using a group of feature vectors of building product data at the time of completion of property A, a group of feature vectors of building product data at the time of completion of property B, and a group of feature vectors of building product data at the time of completion of property C as training data.

[0087] Each feature value of the construction product data at the time of completion of a property includes a list of linked construction products as a feature value. In the second embodiment, by using a group of feature vectors of construction product data at the time of completion of multiple properties, which includes a list of linked construction products as a feature value, as training data, the estimation model generated by the model generation unit 220 can predict a group of feature vectors of construction product data that are expected to be linked (the estimation model that predicts the feature vectors of linked construction product data is referred to as estimation model 2). In other words, estimation model 2 receives feature vectors of multiple construction product data and outputs feature vectors of construction product data that are expected to be linked.

[0088] In the second embodiment, the second specific building product data is data relating to a building product linked to the building product related to the first specific building product data. In this embodiment, the first specific building product data is a plurality of building product data (for example, building product data relating to building product b, building product c, building product d, and building product e), and the second specific building product data is building product data relating to a building product linked to a plurality of building products (for example, building product data a).

[0089] 14 is a diagram illustrating another example of ST308, ST310, and ST312. The feature amount acquisition unit 240 acquires feature vectors of multiple pieces of building product data from the received BIM data. In this example, the feature amount acquisition unit 240 acquires feature vectors of building product data related to building product b, building product c, building product d, and building product e.

[0090] Furthermore, the feature quantity generation unit 250 uses the estimation model 2 to generate a feature vector of second specific building product data from the feature vectors of the multiple first specific building product data acquired by the feature quantity acquisition unit 240. In this example, the feature quantity generation unit 250 uses the estimation model 2 to generate a feature vector of building product data related to building product a that is connected to building product b, building product c, building product d, and building product e. Here, building product a is a building product that is connected to building product b, building product c, building product d, and building product e, and is not included in the BIM data accepted by the BIM data acceptance unit 230 in this example. Note that the feature quantity generation unit 250 may also use the estimation model 2 to generate feature vectors of building product data related to building product a, building product b, building product c, building product d, and building product e (building product data a, building product data b, building product data c, building product data d, and building product data e). The feature vectors of the building product data (building product data b, building product data c, building product data d, building product data e) relating to building product b, building product c, building product d, and building product e generated by the feature generation unit 250 may be identical to the feature vectors of the building product data relating to building product b, building product c, building product d, and building product e acquired by the feature acquisition unit 240.

[0091] Then, the BIM data output unit 260 updates the building product data using the feature vectors of the generated building product data a, building product data b, building product data c, building product data d, and building product data e, and generates BIM data including the updated building product data. Note that the building product data for building products b, c, d, and e do not need to be updated. As described above, by using estimation model 2, it is possible to predict BIM data including building product data for building products that are expected to be added.

[0092] 5.Example of display screen Next, specific examples of the display screen of the terminal device 10 will be described with reference to Figures 15 to 18. In this example, the terminal device 10 will be described as a personal computer.

[0093] FIG. 15 is a diagram showing an example of a display screen that the display control unit 270 causes to be displayed on the display unit 110 of the terminal device 10. BIM data is displayed on the display screen. The displayed BIM data may be BIM data received by the BIM data receiving unit 230. The displayed BIM data may be BIM data generated by the BIM data output unit 260. In FIG. 15(a), the display control unit 270 displays cross-sectional views of multiple pieces of building product data. In this example, cross-sectional views of 10 pieces of building product data are displayed in FIG. 15(a). Displaying cross-sectional views of multiple pieces of building product data by the display control unit 270 enables the user to check and analyze the BIM data.

[0094] When 1C1A:Plate(4) (an example of a building product) is selected in FIG. 15(a), the display control unit 270 displays FIGS. 15(b), 15(c), and 15(d). FIG. 15(b) is a side view of BIM data including the building product. FIG. 15(c) is a top view of BIM data including the building product. FIG. 15(d) is an oblique view (perspective drawing) of BIM data including the building product. In FIGS. 15(b), 15(c), and 15(d), the selected building product is highlighted by being enclosed in parentheses. Examples of highlighting are not limited to brackets, but may also include hatching or a different color. Because the selected portion can be highlighted as described above, it is possible to confirm where the building product is located within the overall BIM data. For example, if a building product is modified in the BIM data analysis system 20, the corresponding building product may also be highlighted.

[0095] FIG. 16 is a diagram showing another example of a display screen that the display control unit 270 displays on the display unit 110 of the terminal device 10. FIG. 16(a) is an overall view of the input BIM data. FIG. 16(b) is an enlarged view of a portion of FIG. 16(a). FIG. 16(c) is an overall view of the BIM data generated by the BIM data output unit 260 after the feature generation unit 250 generates features using the estimation model 1. FIG. 16(d) is an enlarged view of a portion of FIG. 16(c). Comparing FIG. 16(b) and FIG. 16(d), a diaphragm has been added to FIG. 16(d). As shown in FIG. 16, the user can confirm the corrections made to the BIM data generated by the BIM data output unit 260.

[0096] FIG. 17 is a diagram showing another example of a display screen that the display control unit 270 displays on the display unit 110 of the terminal device 10. FIG. 17(a) is an overall view of the input BIM data. FIG. 17(b) is an overall view of the BIM data generated by the BIM data output unit 260 after the feature generation unit 250 generates features using the estimation model 2. Comparing FIG. 17(a) and FIG. 17(b), FIG. 17(a) displays only building product b, building product c, building product d, and building product e, whereas FIG. 17(b) adds building product a. As shown in FIG. 17, the user can confirm the corrections made to the BIM data generated by the BIM data output unit 260.

[0097] FIG. 18 is a diagram showing another example of a display screen that the display control unit 270 displays on the display unit 110 of the terminal device 10. FIG. 18(a) is a diagram of a portion of the input BIM data. FIG. 18(b) is a diagram of a portion of the BIM data generated by the BIM data output unit 260 after the feature generation unit 250 generates features using the estimation model 1. FIG. 18(c) highlights a portion of FIG. 18(b). In FIG. 18(c), the corrections made to the BIM data generated by the BIM data output unit 260 are highlighted. This makes it easier for the user to check the corrections made to the BIM data generated by the BIM data output unit 260.

[0098] 6. Other Examples of Feature Vectors Another embodiment of the feature vector acquired by the feature acquisition unit 240 will be described with reference to Fig. 19. Fig. 19 is a flow diagram showing additional learning of a trained model performed in the system 100. Before this flow is executed, an application is launched by the terminal device 10.

[0099] First, in ST402, the BIM data analysis system 20 (BIM data receiving unit 230) receives input of BIM data. ST402 corresponds to ST306 in Fig. 9. Note that steps corresponding to ST302 and ST304 may be performed before ST402.

[0100] Then, in ST404, the BIM data analysis system 20 (feature amount acquisition unit 240) acquires a feature vector for each piece of building product data from the received BIM data. ST404 corresponds to ST308 in FIG.

[0101] Next, in ST406, the BIM data analysis system 20 uses a decoder to convert the feature vectors of the acquired first specific building product data into text data. The decoder may be included in the BIM data analysis system 20, or the BIM data analysis system 20 may access an external decoder. By converting the feature vectors of the acquired first specific building product data into text data, natural language describing the first specific building product data is acquired. The decoder may be an application, function, etc. that inputs the feature vectors and generates new data. In this example, the decoder converts the feature vectors into text data, but it may also convert the feature vectors into image data or audio data. For example, the decoder is a trained model that has trained the correspondence between the feature vectors and text data describing the building product data in natural language. A Transformer model or the like can be used as the trained model. The decoder may also be a program that mechanically converts feature vectors based on a preset conversion method. When the BIM data analysis system 20 uses a decoder, the BIM data analysis system 20 may include the decoder itself, or it may refer to a decoder stored on a separate server device as needed.

[0102] Then, in ST408, the BIM data analysis system 20 additionally trains the acquired text data into the trained model. Any learning method, such as fine tuning or RAG (Retrieval-Augmented Generation), can be used for the additional training. Here, the trained model may be included in the BIM data analysis system 20, or the BIM data analysis system 20 may access an external trained model.

[0103] The trained model may be a generative trained model such as a large-scale language model (LLM) that is generated by deep learning using a large amount of text data, etc. Examples of such trained models include various models such as Bidirectional Encoder Representations from Transformers (BERT), Generative Pre-trained Transformer (GPT), Gemini, DALL-E, Midjourney, or combinations of these.

[0104] In another example, the "trained model" may be any model that outputs answer information when given input information is input. Such a trained model is a model that has undergone arbitrary machine learning. One example of a trained model is a trained model that is generated by preparing multiple combinations of training data and correct labels assigned to the training data, inputting the combinations into a learning device, and learning a correct pattern through machine learning. Examples of the learning device include neural network-based methods such as neural networks, convolutional neural networks, multi-layer Herceptrons (MLPs), long short-term memories (LSTMs), gated recurrent units (GRUs), graph neural networks (GNNs), and transformers; gradient boosting decision trees (GBDTs) such as LightGBM (Light Gradient Boosting Machine), XGBoost, and CatBoost; ridge regression, logistic regression, support vector regression (SVR), nearest neighbor algorithms, decision trees, regression trees, random forests, and various combinations thereof.

[0105] Because the BIM data analysis system 20 additionally trains the trained model with text data related to the first specific building product data, the trained model can output answer information related to the first specific building product data. For example, if a user inputs input information (prompt information) inquiring about details and progress information of the first specific building product data into the trained model, answer information related to the details and progress information of the first specific building product data can be output. The user can input input information into the trained model in chat format and obtain answer information.

[0106] Furthermore, the trained model may perform a search process in another server device when acquiring response information. For example, the trained model may perform a search process in a progress information management server device when acquiring progress information. The progress information management server device stores, as production progress information for each piece of building product data, information on the phase (process), information on the scheduled start date, information on the scheduled end date, information on the start date, information on the end date, information on which step in each phase, etc. Phases (processes) include, for example, material delivery, component check, marking, marking inspection, processing, product inspection, welding, painting, packaging, shipping, on-site acceptance, installation, and completion. The trained model can output progress information by searching the progress information management server device based on the product matching number of the first specific building product data acquired from the text data.

[0107] As another example, the trained model may execute a search process in a market sales information management server device when acquiring market sales information regarding the sales information in the market for each building material used in a building product. The market sales information management server device stores, for each sales site (EC site) for each building material used in a building product, information on inventory quantity, price information, location information, sales company information, shipping company information, delivery fee information, etc. The trained model can output the market sales information by searching the market sales information management server device based on the product matching number of the first specific building product data acquired from the text data.

[0108] It should be noted that the embodiments of the present disclosure are presented as examples and are not intended to limit the scope of the present disclosure. The present embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the present disclosure. These embodiments and their modifications are included within the scope and spirit of the invention, and are also included in the inventions and their equivalents as set forth in the claims. [Explanation of symbols]

[0109] 10 Terminal device, 11 Central processing unit, 12 Main memory device, 13 Input / output interface device, 14 Input device, 15 Auxiliary memory device, 16 Output device, 20 BIM data analysis system, 21 Central processing unit, 22 Main memory device, 23 Input / output interface device, 24 Input device, 25 Auxiliary memory device, 26 Output device, 30 Communication line, 100 System, 110 Display unit, 120 Communication unit, 130 User interface unit, 210, memory unit, 220, model generation unit, 230, BIM data reception unit, 240, feature acquisition unit, 250, feature generation unit, 260, BIM data output unit, 270, display control unit, 280, communication unit, 310, H-beam, 312, upper flange, 320, gusset plate, 322, protrusion, 330, bracket, 332, bracket, 334, bracket, 340, column, 350, diaphragm

Claims

1. 1. A BIM data analysis system for analyzing BIM data including one or more building product data, comprising: A storage unit that stores feature vectors of one or more of the BIM data and one or more of the building product data included in the one or more BIM data; a model generation unit that uses the feature vectors of one or more of the building product data stored in the storage unit as training data and generates, by machine learning, an estimation model that estimates the feature vector of a second specific building product data among the building product data from the feature vector of at least one or more first specific building product data among the building product data; A BIM data receiving unit that receives input of the BIM data; A feature acquisition unit that acquires a feature vector for each of the building product data from the BIM data input to the BIM data acceptance unit; a feature amount generation unit that generates a feature vector of the second specific building product data from the feature vector of at least one of the first specific building product data acquired by the feature amount acquisition unit, using the estimation model generated by the model generation unit; A BIM data analysis system comprising:

2. The BIM data analysis system according to claim 1, further comprising a BIM data output unit that updates the building product data from the feature vector of the second specific building product data generated by the feature amount generation unit and generates the BIM data including the updated building product data.

3. The BIM data analysis system of claim 2 , wherein the feature vector includes a modification count, which is the number of times the building product data has been modified.

4. The second specific building product data is data relating to the same building product as the first specific building product data, When the number of revisions of the first specific building product data is N, the number of revisions of the second specific building product data is N+1. The BIM data analysis system of claim 3 .

5. the feature vector includes a list of building products that are linked to the building product associated with the building product data; The second specific building product data is data relating to a building product linked to the building product relating to the first specific building product data, The BIM data analysis system of claim 2 .

6. the model generation unit uses the feature vectors of one or more of the building product data stored in the storage unit as training data, and generates an estimation model by machine learning that estimates the feature vector of the second specific building product data from the feature vectors of the plurality of first specific building product data; the feature quantity generation unit generates a feature vector of the second specific building product data from the feature vectors of the plurality of first specific building product data acquired by the feature quantity acquisition unit, using the estimation model generated by the model generation unit; The BIM data analysis system of claim 5.

7. The storage unit stores the BIM data in a modifiable manner, The BIM data reception unit receives the BIM data modified by the storage unit, The BIM data analysis system of claim 2 .

8. The building product related to the building product data is composed of a beam, a column, a bracket, a diaphragm, a brace, or a damper. The BIM data analysis system according to any one of claims 1 to 7.

9. the feature vector includes at least one of a feature amount in a cross section of the beam or the column, a thickness of the bracket, the diaphragm, the brace or the damper, a lateral length of the bracket, an interval of the diaphragm, a weight of the beam or the column, a total length of the beam or the column, a total width of the beam or the column, a total height of the beam or the column, and a diagonal length of the beam or the column; The BIM data analysis system of claim 8.

10. The feature vector includes at least one of the product serial number of the building product related to the building product data, the building information of the BIM data including the building product data, and the product information of the building product related to the building product data, The BIM data analysis system of claim 9.

Images