Design support device

Abstract

To support structural design of a building using structural members of an appropriate number of groups, by inputting design conditions.SOLUTION: An input unit 10 receives a building model obtained by modeling a building including a plurality of structural members, and design conditions for the building to be designed. A cross-section structure calculation unit 22 calculates, for each of the structural members of the building, based on the design conditions, cross-section structures of the structural members. A grouping processing unit 28 performs grouping to classify the structural members into multiple groups composed of structural members with the same cross-section structure, on the basis of feature quantities of the structural members, the feature quantities being obtained from a result of stress analysis on the building model including the plurality of structural members having the results of calculating the cross-section structures.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a design support device. [Background technology]

[0002] Conventionally, there is known a framework structure optimization design device that specifies a base line on the coordinate system where strength members should be arranged, and enables element members to be identified based on the positional relationship between the element members and the base line, thereby allowing a computer to perform the grouping work simply by specifying the base line, without requiring time-consuming input work such as attaching special identification tags to individual finite element data (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2008-299385 Summary of the Invention [Problem to be solved by the invention]

[0004] In the technology described in Patent Document 1, element members are grouped based on the specified base line on the coordinate system where the reinforcing members are to be installed. However, Patent Document 1 does not describe calculating the cross-sectional structure of the structural members based on the input design conditions and grouping the structural members.

[0005] In consideration of the above, the present invention aims to support the structural design of a building using an appropriate number of groups of structural members by inputting design conditions. [Means for solving the problem]

[0006] The design support device of the present invention is configured to include an input unit that accepts a building model that models a building that is a target building for design and that includes a plurality of structural members, and design conditions related to the target building; a cross-sectional calculation unit that calculates the cross-sectional structure of each of a plurality of structural members of the building based on the design conditions; and a grouping processing unit that performs grouping to classify the plurality of structural members into a plurality of groups consisting of structural members that should have the same cross-sectional structure, based on the feature values ​​of each of the plurality of structural members obtained from the results of stress analysis of the building model that includes the plurality of structural members having the calculation results of the cross-sectional structure.

[0007] In accordance with the design support system of the present invention, an input unit receives a building model of a building including multiple structural members and design conditions for the building to be designed. A cross-section calculation unit calculates the cross-sectional structure of each of the multiple structural members of the building based on the design conditions. A grouping processing unit then classifies the multiple structural members into multiple groups consisting of structural members that should have the same cross-sectional structure based on feature quantities of each of the multiple structural members obtained from results of stress analysis of the building model including the multiple structural members having the calculation results of the cross-sectional structure.

[0008] In this way, the cross-sectional structure of each of the multiple structural members of the building is calculated based on the received design conditions, and the multiple structural members are classified into groups based on the characteristic quantities of each of the multiple structural members obtained from the results of stress analysis of the building model.This makes it possible to support the structural design of a building using an appropriate number of groups of structural members by inputting design conditions.

[0009] In the design support device according to the present invention, the design conditions can include a target value for the inspection ratio, a target value for the available load capacity margin, the number of iterations, a target value for the column axial force ratio, a target value for the deformation angle of each floor, a designated rank for the column-beam brace, or a designated strength. This makes it possible to support the structural design of a building that satisfies the target value for the inspection ratio, the target value for the available load capacity margin, the number of iterations, the target value for the column axial force ratio, the target value for the deformation angle of each floor, the designated rank for the column-beam brace, or the designated strength.

[0010] In the design support system according to the present invention, the cross-section calculation unit can repeatedly change the cross-sectional structure of each of the plurality of structural members based on the results of stress analysis of the building model including the plurality of structural members having the calculation results of the cross-sectional structure until the design conditions are satisfied. This makes it possible to support the structural design of a building that satisfies the input design conditions, taking into account the results of the stress analysis of the building model.

[0011] In the design support system according to the present invention, a plurality of parameter sets each consisting of parameters related to grouping are predetermined, and the grouping processing unit performs the grouping using each of the plurality of parameter sets to obtain a plurality of grouping results, thereby supporting the structural design of a building using a more appropriate number of groups of structural members. [Effects of the Invention]

[0012] As described above, according to the design support device of the present invention, the cross-sectional structure of each of the plurality of structural members of the building is calculated based on the received design conditions, and the plurality of structural members are classified into groups based on the characteristic quantities of each of the plurality of structural members obtained from the results of stress analysis of the building model, thereby achieving the effect of being able to support the structural design of a building using an appropriate number of groups of structural members by inputting the design conditions. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a block diagram showing a learning device and a design support device according to a first embodiment of the present invention. [Figure 2] 1 is a functional block diagram showing a design support apparatus according to a first embodiment of the present invention; [Figure 3] FIG. 10 is a diagram illustrating an example of a design condition input screen. [Figure 4] FIG. 10 is a diagram for explaining member information of a structural member. [Figure 5] FIG. 10 is a diagram illustrating an example of a trained model for cross-sectional structure calculation. [Figure 6A] FIG. 10 is a diagram illustrating an example of a grouping result for a grouping plan. [Figure 6B] FIG. 10 is a diagram illustrating an example of a grouping result for a grouping plan. [Figure 6C] FIG. 10 is a diagram illustrating an example of a grouping result for a grouping plan. [Figure 7] 1 is a functional block diagram showing a learning device according to a first embodiment of the present invention. [Figure 8] 3 is a flowchart showing the contents of a design support processing routine of the design support device according to the first embodiment of the present invention. [Figure 9] 5 is a flowchart showing a flow of processing for changing a cross-sectional structure in the design support apparatus according to the first embodiment of the present invention. [Figure 10] 4 is a flowchart showing the flow of grouping processing in the design support apparatus according to the first embodiment of the present invention. [Figure 11] FIG. 10 is a diagram illustrating an example of a trained model for grouping. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[0015] [First embodiment] <Configuration of the design support device according to the first embodiment of the present invention> As shown in FIG. 1, a design support device 100 according to a first embodiment of the present invention includes a CPU 12, a graphics card 13, a GPU 14, a RAM 16, an HDD 18, a communication interface 21, and a bus 23 for interconnecting these components.

[0016] The CPU 12 and GPU 14 execute various programs. The RAM 16 is used as a work area when the CPU 12 executes the various programs. The HDD 18, which serves as a recording medium, stores various programs and data, including a program for executing a design support processing routine, which will be described later.

[0017] The design support device 100 according to this embodiment is represented by functional blocks along with a program for executing a design support processing routine, as shown in Fig. 2. The design support device 100 includes an input unit 10, a calculation unit 20, and an output unit 50.

[0018] The input unit 10, operated by the designer, accepts input of a building model that is a building to be designed and that includes multiple structural members (columns, beams, walls, braces, etc.), and also accepts design conditions related to the building to be designed.

[0019] For example, the designer operates the system to place multiple structural members for each type (columns, beams, walls, braces, etc.) in the building model. Then, the designer operates the input screen in Figure 3 to accept design conditions.

[0020] For example, the design conditions include a target value for the inspection ratio, a target value for the residual strength margin, the number of repeated calculations, a target value for the column axial force ratio, a target value for the deformation angle of each floor, a rank specification for the column-beam brace, or a strength specification.

[0021] The input screen in Figure 3 shows an example of accepting the target value for the inspection ratio, the target value for the residual capacity, whether or not to adjust the eccentricity ratio, the target value for the column-beam strength ratio, the number of stories to group for steel columns, the number of iterations, whether or not to consider RC members, whether or not to consider shear walls, the target value for the τ level at the time of the primary design, and the target value for the column long-term axial force ratio. This input screen also shows an example of accepting the target value for the deformation angle, the rank specification for column-beam braces, and the strength specification for each floor.

[0022] The calculation unit 20 includes a cross-sectional structure calculation unit 22 and a grouping processing unit 28. The cross-sectional structure calculation unit 22 is an example of a cross-section calculation unit.

[0023] The cross-sectional structure calculation unit 22 calculates the cross-sectional structure of each structural member of the building model based on the member information of the structural member using a trained model for cross-sectional structure calculation trained in advance by the learning device 200 described below, and displays the calculation results. For example, as the calculation results, the structural member reflecting the calculated cross-sectional structure may be visually displayed by being superimposed on the volume of the building, or the calculation results of the quantity, weight, and cost using the calculated cross-sectional structure of the structural member may be displayed.

[0024] The trained model for cross-sectional structural calculations takes as input data component information (length L, angle θ in Figure 4, position within the building (height position, position on the plane), floor height, component density (span), bearing area, loading conditions, shear force bearing rate of the frame to which it belongs, etc.), and outputs as output data cross-sectional information representing the cross-sectional structure (component width D, component composition B, component thickness t, material strength, component weight, component performance, etc. in Figure 4) (see Figure 5). For example, as shown in Figure 5, a neural network can be used as an example of a model, and deep learning can be used as an example of a learning algorithm, and the trained model for cross-sectional structural calculations is trained so that when component information of training data is input, the cross-sectional information of the training data is output.

[0025] Then, the cross-sectional structure calculation unit 22 changes the cross-sectional structure of each of the structural members of the building model so as to satisfy the accepted design conditions.

[0026] Specifically, the cross-sectional structure calculation unit 22 performs stress analysis on a building model including multiple structural members having the calculation results of the structural cross section, and changes the cross-sectional structure of each structural member based on the results of the stress analysis so as to satisfy the design conditions. This process is repeated until the design conditions are satisfied.

[0027] The grouping processing unit 28 performs grouping to classify multiple structural members into multiple groups consisting of structural members that should have the same cross-sectional structure, based on the characteristic quantities of each of the multiple structural members obtained from the results of stress analysis of a building model including multiple structural members having calculation results of structural cross-sections.

[0028] Specifically, for each type of structural member, grouping is performed based on the distribution of the feature values ​​of the structural members to be grouped. As an example of a grouping algorithm, a clustering method can be used.

[0029] For example, a stress analysis is performed on a building model including multiple structural members with calculated structural cross-sections, feature quantities for the structural member groups are determined based on the results of the stress analysis, the structural member groups are clustered for each type of structural member based on the distribution of feature quantities for the structural member groups, the cross-sectional structures are modified so as to unify the cross-sectional structures of structural members in the same cluster, and a stress analysis is performed on the building model including multiple structural members with the modified structural cross-sections, and the results of the stress analysis are output. The feature quantities include long-term axial force, short-term moment, short-term axial force, column length, column coordinates, etc. obtained as a result of the stress analysis.

[0030] Furthermore, a plurality of parameter sets made up of parameters related to grouping are determined in advance, and a plurality of grouping results are obtained by performing grouping using each of the plurality of parameter sets (FIGS. 6A to 6C).

[0031] The parameter set includes, for example, the value of K related to clustering and a weight vector consisting of weights for each feature amount.

[0032] 6A to 6C show examples of displaying three grouping results for three parameter sets. The lower parts of Figures 6A to 6C show enlarged rectangular frame portions of the grouping results shown on the upper side.

[0033] <Configuration of the learning device according to the first embodiment of the present invention> As shown in FIG. 1 above, the learning device 200 according to the first embodiment of the present invention, like the design support device 100, includes a CPU 12, a graphics card 13, a GPU 14, a RAM 16, an HDD 18, a communication interface 21, and a bus 23 for interconnecting these.

[0034] The CPU 12 and GPU 14 execute various programs. The RAM 16 is used as a work area when the CPU 12 executes the various programs. The HDD 18, which serves as a recording medium, stores various programs, including a program for executing learning processing, and various data.

[0035] Learning device 200 in this embodiment is represented by functional blocks in accordance with a program for executing the learning process, as shown in Figure 7. Learning device 200 includes input unit 110, calculation unit 120, and output unit 150.

[0036] The input unit 110 receives as input learning data including a combination of component information, including position information of the structural components, obtained for each structural component (columns, beams, walls, braces, etc.) from the building's performance information, and the cross-sectional structure of the structural components.

[0037] The calculation unit 120 includes a learning unit 122 .

[0038] The learning unit 122 obtains a learned model for cross-sectional structural calculation for each type of structural member based on the plurality of learning data received by the input unit 10.

[0039] In this embodiment, a trained model for cross-sectional structural calculation is generated for each type of structural member (column, beam, wall, brace, etc.), and output to the design support device 100 by the output unit 150.

[0040] <Learning device operation> Next, the operation of learning device 200 according to the first embodiment of the present invention will be described.

[0041] The input unit 110 receives as input learning data including a combination of member information, including position information of the structural members, obtained for each structural member (columns, beams, walls, braces, etc.) from the building's performance information, and the cross-sectional structures of the structural members. The learning unit 122 then obtains a trained model for cross-sectional structure calculation based on the multiple pieces of learning data received by the input unit 110.

[0042] <Operation of the design support system> Next, the operation of the design support device 100 according to the first embodiment of the present invention will be described.

[0043] The input unit 10 receives input of a building model of the building to be designed, which is a model of a building including multiple structural members (columns, beams, walls, braces, etc.), through operation by the designer, and also receives design conditions related to the building to be designed. Then, the design support device 100 executes the design support processing routine shown in FIG. 8.

[0044] First, in step S100, the cross-sectional structural calculation unit 22 acquires a building model including a plurality of input structural members.

[0045] In step S102, the cross-sectional structure calculation unit 22 calculates the cross-sectional structure of each structural member of the building model using a trained model for cross-sectional structure calculation based on the member information of the structural member, and displays the calculation results.

[0046] In step S104, the cross-sectional structure calculation unit 22 changes the cross-sectional structure of each of the structural members of the building model calculated in step S102 so as to satisfy the accepted design conditions.

[0047] In step S106, the grouping processing unit 28 performs grouping to classify the structural members into multiple groups consisting of structural members that should have the same cross-sectional structure, based on the feature values ​​of each of the structural members obtained from the results of stress analysis of the building model including the structural members having the calculation results of the structural cross-section. This grouping is performed for each parameter set related to grouping, and multiple grouping proposals are obtained.

[0048] In step S108, the grouping results for the plurality of grouping plans are displayed by the output unit 150, and the design support processing routine is terminated.

[0049] The above step S104 is realized by the processing routine shown in FIG.

[0050] In step S110, the cross-sectional structure calculation unit 22 changes the cross-sectional structure of each structural member.

[0051] In step S112, the cross-sectional structure calculation unit 22 performs stress analysis on the building model including a plurality of structural members having the calculation results or change results of the structural cross section.

[0052] In step S114, the cross-sectional structure calculation unit 22 determines whether the input design conditions are satisfied based on the results of the stress analysis. If the input design conditions are not satisfied, the process returns to step S110. On the other hand, if the input design conditions are satisfied, the process determines that the cross-sectional structures of the structural members that satisfy the design conditions have been obtained, and ends the processing routine.

[0053] The above step S106 is realized by the processing routine shown in Fig. 10. This processing routine is repeatedly executed for each parameter set related to grouping.

[0054] In step S120, the grouping processing unit 28 performs stress analysis on the building model including a plurality of structural members having the calculation results of the structural cross sections.

[0055] In step S122, the grouping processing unit 28 acquires the feature quantities of the structural member group based on the results of the stress analysis.

[0056] In step S124, the grouping processing unit 28 performs clustering of the structural member groups for each type of structural member based on the distribution of the feature amounts of the structural member groups.

[0057] In step S126, the grouping processing unit 28 changes the cross-sectional structure so as to unify the cross-sectional structures of structural members in the same cluster.

[0058] In step S128, the grouping processing unit 28 performs stress analysis on the building model including a plurality of structural members having the changed structural cross section, outputs the results of the stress analysis, and ends the processing routine.

[0059] As described above, the design support device according to the first embodiment of the present invention calculates the cross-sectional structure of each of the structural members of a building based on the received design conditions, and performs grouping to classify the structural members based on the feature values ​​of each of the structural members obtained from the results of stress analysis of the building model. This makes it possible to support the structural design of a building using an appropriate number of groups of structural members simply by inputting the design conditions.

[0060] [Second embodiment] Next, a second embodiment of the present invention will be described. Note that the configurations of the design support device and learning device of the second embodiment are similar to those of the first embodiment, so the same reference numerals are used and the description will be omitted.

[0061] The second embodiment differs from the first embodiment in that grouping is performed using supervised learning.

[0062] <Configuration of a learning device according to a second embodiment of the present invention> Similar to the first embodiment, the input unit 110 of the learning device 200 according to the second embodiment of the present invention receives as input learning data for cross-sectional structural calculations. The input unit 110 also receives as input learning data including the structural member information of each of the two structural members, and a determination result for each structural member pair consisting of two structural members out of all structural members obtained from the building performance information about the structural members, as to whether or not the structural members are grouped based on the structural member information of each of the two structural members.

[0063] Specifically, from the building's performance information, for a structural member pair consisting of two structural members from among all structural members, learning data is created that includes the member information of the two structural members (length L, angle θ, position within the building (height position, position on the plane), floor height, member density (span), bearing area, and other information that characterizes the member (member width D, member thickness B, member thickness t, etc. in Figure 4 above)) and the result of determining whether or not the two structural members are grouped based on the structural member information for each of the two structural members. Then, for each structural member pair, learning data is received that includes a combination of the member information of the two structural members and the result of determining whether or not they are grouped.

[0064] In this embodiment, this learning data is received for each type of structural member (column, beam, wall, brace, etc.).

[0065] The learning unit 122 obtains a learned model for cross-sectional structure calculation, similarly to the first embodiment.

[0066] Furthermore, the learning unit 122 obtains a learned model for grouping based on the learning data.

[0067] Specifically, as shown in Figure 11, the trained model for grouping uses the feature quantities of two structural members as input data and the degree to which the two structural members should be grouped as output data. For example, a neural network can be used as an example of the model, and deep learning can be used as an example of the learning algorithm. A trained model for grouping is generated for each type of structural member (column, beam, wall, brace, etc.).

[0068] <Configuration of the design support device according to the second embodiment of the present invention> The grouping processing unit 28 of the design support device 100 of the second embodiment calculates the degree of grouping for each structural member pair consisting of two structural members out of all the generated structural members for each type of structural member (column, beam, wall, brace, etc.) based on the feature values ​​of each of the two structural members and the trained model for grouping, and groups the structural member pairs in descending order of the degree of grouping so as to achieve the specified number of groups.

[0069] Specifically, for each type of structural member (column, beam, wall, brace, etc.), the grouping processing unit 28 calculates the degree of grouping for each structural member pair consisting of two structural members out of all structural members, based on the structural member information of each of the two structural members and the trained model for grouping the type of structural member.

[0070] For example, for each pair of structural members, the component information of the two structural members (length, angle, position within the building (vertical position, position on the plane), floor height, component density (span), load area, and other information that characterizes the components) is input into a trained model for grouping to determine the degree to which they should be grouped.

[0071] Then, the grouping processing unit 28 repeatedly groups two structural members of a structural member pair into the same group in descending order of the degree of grouping, for each type of structural member (column, beam, wall, brace, etc.), until the specified number of groups is reached. This allows for grouping results for the specified number of groups. The grouping processing unit 28 also sequentially changes the specified number of groupings and similarly groups the structural member pairs. This allows for grouping results for multiple grouping proposals.

[0072] Other configurations and operations of the design support device 100 and the learning device 200 according to the second embodiment are the same as those of the first embodiment, and therefore description thereof will be omitted.

[0073] As described above, the design support device according to the second embodiment of the present invention calculates the cross-sectional structure of each of the multiple structural members of a building based on the received design conditions, calculates the degree of grouping based on the feature values ​​of each of the multiple structural members obtained from the results of stress analysis of the building model, and performs grouping to classify the multiple structural members. This makes it possible to support the structural design of a building using an appropriate number of groups of structural members simply by inputting the design conditions.

[0074] The present invention is not limited to the above-described embodiment, and various modifications and applications are possible without departing from the spirit and scope of the present invention.

[0075] For example, in the above-described embodiment, an example was described in which the learning device and the design support device are configured as separate devices, but this is not limited to this, and the learning device and the design support device may also be configured as a single device.

[0076] The program of the present invention may be provided in a form stored on a storage medium. [Explanation of symbols]

[0077] 10, 110 Input section 12 CPU 20, 120 calculation section 22 Cross-sectional structure calculation section 28 Grouping processing section 100 Design support equipment 122 Learning Department 200 Learning Device

Claims

1. an input unit that receives a building model of a building to be designed, the building model including a plurality of structural members, and design conditions related to the building to be designed; a cross-section calculation unit that calculates a cross-sectional structure of each of a plurality of structural members of the building based on the design conditions; a grouping processing unit that performs grouping to classify the plurality of structural members into a plurality of groups each consisting of structural members that should have the same cross-sectional structure, based on feature quantities of each of the plurality of structural members obtained from a result of stress analysis of the building model including the plurality of structural members having the calculation result of the cross-sectional structure; Including, The design conditions include a target value of the deformation angle, a designation of the rank of the column-beam brace, and a designation of the strength for each floor, A design support device in which the feature quantities include long-term axial force, short-term moment, short-term axial force, length, and coordinates of the structural member.

2. The cross section calculation unit 2. The design support device according to claim 1, wherein the cross-sectional structure of each of the plurality of structural members is changed based on the results of stress analysis of the building model including the plurality of structural members having the calculation results of the cross-sectional structure, and the change is repeated until the design conditions are satisfied.

3. a plurality of parameter sets each including a parameter related to grouping are determined in advance; 3. The design support system according to claim 1, wherein the grouping processing unit performs the grouping using the parameter set for each of the plurality of parameter sets, thereby obtaining a plurality of grouping results.

Images