Wind evaluation device, wind evaluation method, and wind evaluation program

Abstract

To provide a wind evaluation device, a wind evaluation method and a wind evaluation program for efficiently and accurately predicting a wind distribution.SOLUTION: A control part 21 of a support server 20 records teacher information obtained by combining a first wind distribution in a first arrangement, a second arrangement changed from the first arrangement, and a second wind distribution in the second arrangement in a teacher information storage part 22. Next, a control part 21 uses the teacher information recorded in the teacher information storage part 22, generates a prediction model for predicting a secondo wind distribution in the case of changing from the first arrangement to the second arrangement by performing machine learning and records the prediction model in a learning result storage part 23. Then, the control part 21 inputs an evaluation object arrangement and the first wind distribution in the prediction model recorded in the learning result storage part 23, and predicts a wind distribution in the evaluation object arrangement in the case of acquiring the evaluation object arrangement with an evaluation object arranged and an unchanged wind distribution of the evaluation object.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a wind evaluation device, a wind evaluation method, and a wind evaluation program for estimating the wind distribution around a building or the like.

Background Art

[0002] Computational Fluid Dynamics (CFD) is performed to examine the wind conditions around a building or the like. In CFD, a building to be evaluated is modeled on a computer. Then, by applying equations related to fluid motion, an approximate solution of the wind speed is calculated. Furthermore, a technique for obtaining a wind speed distribution including wind speed information and wind direction information by using teacher information has also been studied (see Patent Document 1). The wind speed and wind pressure estimation device described in this document uses building information data representing the three-dimensional shape of a building and its surroundings, and wind direction data. Furthermore, a deep-learned trained model using, as learning data, wind speed distribution data and wind pressure distribution data in a three-dimensional space as teacher data corresponding to the building information data and the wind direction data is used. Then, the building information data and the wind direction data for which the wind speed distribution and the wind pressure distribution are to be estimated are input into the trained model to estimate the wind speed distribution and the wind pressure distribution.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In numerical fluid dynamics, every time a building is modified, reanalysis is required, resulting in a large computational load. Further, in the technique described in Patent Document 1, the prediction accuracy in areas not used during learning may decrease. For example, when there is an influence from a high-rise building far from the range of the input image, it is difficult to improve the prediction accuracy because the influence of this high-rise building cannot be considered.

Means for Solving the Problems

[0005] The reputation evaluation device for solving the above problems includes a teacher information storage unit, a learning result storage unit, and a control unit for predicting wind distribution. The control unit records teacher information combining a first wind distribution in a first arrangement, a second arrangement changed from the first arrangement, and a second wind distribution in the second arrangement in the teacher information storage unit, and by performing machine learning using the teacher information recorded in the teacher information storage unit, generates a prediction model for predicting the second wind distribution when changing from the first arrangement to the second arrangement and records it in the learning result storage unit. When an evaluation target arrangement where the evaluation target is placed and a wind distribution before change of the evaluation target are obtained, the evaluation target arrangement and the wind distribution before change are input into the prediction model recorded in the learning result storage unit to predict the wind distribution in the evaluation target arrangement.

Advantages of the Invention

[0006] The present invention can efficiently and accurately predict the wind distribution around buildings and the like.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Mode for Carrying Out the Invention

[0008] Hereinafter, an embodiment in which a reputation evaluation device, a reputation evaluation method, and a reputation evaluation program are embodied will be described with reference to FIGS. 1 to 6. In the present embodiment, a reputation evaluation device, a reputation evaluation method, and a reputation evaluation program used for predicting the wind speed ratio distribution around a building newly constructed in a block will be described. As shown in FIG. 1, the reputation evaluation device A1 of the present embodiment includes a user terminal 10 and a support server 20 that are interconnected via a network.

[0009] (Explanation of Hardware Configuration) Using FIG. 2, the hardware configuration of the information processing device H10 that constitutes the user terminal 10 and the support server 20 will be described. The information processing device H10 includes a communication device H11, an input device H12, a display device H13, a storage device H14, and a processor H15. Note that this hardware configuration is an example, and it is also possible to be realized by other hardware.

[0010] The communication device H11 is an interface that establishes a communication path with other devices and executes data transmission and reception, and is, for example, a network interface or a wireless interface.

[0011] The input device H12 is a device that receives input of various information, and is, for example, a mouse or a keyboard. The display device H13 is a display or the like that displays various information. The storage device H14 is a storage device that stores data and various programs for executing various functions of the user terminal 10 and the support server 20. Examples of the storage device H14 include a ROM, a RAM, and a hard disk.

[0012] Processor H15 controls each process in user terminal 10 and support server 20 by using programs and data stored in storage device H14. Examples of processor H15 include, for example, a CPU, an MPU, etc. This processor H15 expands a program stored in a ROM or the like to a RAM and executes various processes for each process.

[0013] Processor H15 is not limited to performing software processing for all processes it executes. For example, processor H15 may include a dedicated hardware circuit (for example, an application specific integrated circuit: ASIC) that performs hardware processing for at least a part of the processes it executes. That is, processor H15 may be configured as follows.

[0014] [1] One or more processors that operate according to a computer program (software) [2] One or more dedicated hardware circuits that execute at least a part of various processes [3] A combination thereof, including circuitry The processor includes a CPU and memories such as a RAM and a ROM, and the memories store program codes or instructions configured to cause the CPU to execute processes. The memory, that is, the computer-readable medium, includes any available medium that can be accessed by a general-purpose or dedicated computer.

[0015] (Reputation evaluation device A1) User terminal 10 is a computer terminal used by a user. Support server 20 is a computer system that predicts wind distribution. In this embodiment, the wind speed ratio distribution is predicted as the wind distribution. Support server 20 includes a control unit 21, a teacher information storage unit 22, and a learning result storage unit 23.

[0016] The control unit 21 performs processes (processes including an information acquisition stage, a learning stage, a prediction stage, etc.) to be described later. By executing a reputation evaluation program for this purpose, the control unit 21 functions as an information acquisition unit 211, a learning unit 212, a prediction unit 213, etc.

[0017] The information acquisition unit 211 executes a process of acquiring building layout information and wind information. In the present embodiment, it is assumed that a new building is newly constructed in a block where existing structures are arranged. In this case, the layout information after the new building is constructed in the block, the first wind information before the new building is constructed, and the second wind information after the new building is constructed are acquired. In the present embodiment, a wind speed ratio distribution (wind distribution) is used as the wind information.

[0018] The learning unit 212 executes a process of generating a prediction model by machine learning using teacher information. For this machine learning, for example, CNN (Convolutional Neural Network) can be used. Note that the learning method is not limited to CNN. The prediction unit 213 predicts the second wind information after the new building is constructed from the layout information after the new building is constructed and the first wind information before the new building to be constructed, using the prediction model.

[0019] Teacher information used for machine learning is recorded in the teacher information storage unit 22. In the present embodiment, it is recorded when teacher information is generated. The teacher information uses the layout information after the new building is constructed, the first wind information, and the second wind information.

[0020] The layout information is information regarding the layout of structures (subsequent second layout) after the building to be evaluated is newly constructed, for a block (preceding first layout) where a plurality of structures are newly constructed. As the layout information, for example, a three-dimensional map in which a new building model is arranged in a three-dimensional block model, or a two-dimensional map showing the presence or absence and height of each structure in a predetermined gradation (for example, 256 gradations) can be used.

[0021] The first wind information is an image showing the wind distribution (prior wind distribution) before the construction of a building. As the first wind information, for example, a map showing the wind speed ratio distribution in the block before the construction of the building can be used. As the wind information, a map showing the three-dimensional wind direction and wind speed, or a two-dimensional map showing the wind speed ratio in each direction (x direction, y direction) of the plane in a predetermined gradation (for example, 256 gradations with the wind speed ratio "0" being "127") can be used.

[0022] The second wind information is an image showing the wind distribution (subsequent wind distribution) after the construction of a building. As the second wind information, for example, a map showing the wind speed ratio distribution in the block after the construction of the building can be used. The first wind information and the second wind information can be calculated, for example, by performing an analysis using computational fluid dynamics, but the calculation method of the wind information is not limited to the analytical method. For example, actually measured wind information or wind information measured in a wind tunnel experiment or the like may be used.

[0023] In the learning result memory unit 23, a prediction model for predicting wind information is recorded. This prediction model is a model that predicts the second wind information of the block after the construction of the building by taking as inputs the arrangement information (information regarding the second arrangement) of the buildings after the construction in the block and the first wind information of the block before the construction of the building.

[0024] (Wind evaluation process) Next, with reference to FIGS. 3 and 4, the wind evaluation process will be described. This evaluation process of the wind speed distribution is composed of a learning-time process (FIG. 3) and a prediction-time process (FIG. 4).

[0025] (Learning-time process) The learning process will be described with reference to FIG. 3. Here, first, the control unit 21 of the support server 20 executes a process for specifying the first arrangement (step S11). Specifically, the information acquisition unit 211 of the control unit 21 acquires the first arrangement information from the user terminal 10. As this first arrangement information, a three-dimensional map (first arrangement map) showing the arrangement of existing structures before the building to be evaluated is newly constructed is used. In this three-dimensional map, the shape and position (coordinates) of the existing buildings are set. Then, the information acquisition unit 211 temporarily stores the acquired first arrangement information in the memory.

[0026] Next, the control unit 21 of the support server 20 executes the acquisition process of the first wind information in the first arrangement (step S12). Specifically, the information acquisition unit 211 of the control unit 21 acquires the first wind information in the block where the existing structure was arranged before the building to be evaluated was newly constructed, from the user terminal 10. Here, as the first wind information, the wind speed ratio distribution information (first wind distribution) obtained by mapping the wind speed ratio in the first arrangement map is used. This first wind information can be calculated by numerical fluid dynamics using the first arrangement map. Then, the information acquisition unit 211 temporarily stores the acquired first wind information in the memory. Specifically, as shown in FIG. 5, for the first arrangement map 300 where the existing structure is arranged, the first wind distribution 301 obtained by mapping the wind speed ratio distribution is used.

[0027] Next, the control unit 21 of the support server 20 executes the identification process of the second arrangement with the first arrangement changed (step S13). Specifically, the information acquisition unit 211 of the control unit 21 acquires the second arrangement information specified by the user from the user terminal 10. This second arrangement information uses a three-dimensional map (second arrangement map) showing the arrangement after the building to be evaluated is newly constructed, in addition to the arrangement of the existing structures in the block. The shape and position (coordinates) of the building to be evaluated are set in this three-dimensional map together with the existing buildings. Then, the information acquisition unit 211 temporarily stores the acquired second arrangement information in the memory.

[0028] Next, the control unit 21 of the support server 20 executes the acquisition process of the second wind information in the second arrangement (step S14). Specifically, the information acquisition unit 211 of the control unit 21 acquires the second wind information (second wind distribution) in the second arrangement map where the building to be evaluated is arranged, from the user terminal 10. Here, as the second wind information, the wind speed ratio distribution information obtained by mapping the wind speed ratio in the second arrangement map is used. This second wind information can also be calculated by numerical fluid dynamics using the second arrangement map. Specifically, as shown in FIG. 5, for the second layout map 310 in which the building T0 to be evaluated is arranged with respect to the existing structure, the second wind distribution 311 obtained by mapping the wind speed ratio distribution is used.

[0029] Next, the control unit 21 of the support server 20 executes a teacher information generation process (step S15). Specifically, the information acquisition unit 211 of the control unit 21 records a data set including the first wind information, the second wind information, and the second layout information in the teacher information storage unit 22 as teacher information.

[0030] After that, when it is detected that a predetermined number or more of teacher information has been recorded in the teacher information storage unit 22, the control unit 21 of the support server 20 executes a machine learning process (step S16). Specifically, the learning unit 212 of the control unit 21 uses the teacher information recorded in the teacher information storage unit 22, takes the second layout information after the building is newly constructed and the first wind information before the building is newly constructed as inputs, and generates a prediction model for predicting the second wind information after the building is newly constructed, and records it in the learning result storage unit 23. Here, as shown in FIG. 5, a prediction model M1 is generated using the first wind distribution 301, the second layout map 310, and the second wind distribution 311 as teacher information.

[0031] (Processing at the time of prediction) The processing at the time of prediction will be described with reference to FIG. 4. This process is performed when a new building to be evaluated is newly constructed.

[0032] Here, first, the control unit 21 of the support server 20 executes a process for specifying the evaluation target layout (step S21). Specifically, the information acquisition unit 211 of the control unit 21 acquires an evaluation target map (evaluation target layout) from the user terminal 10. This evaluation target map is a three-dimensional map including the three-dimensional model of the building to be evaluated in the three-dimensional model of the surrounding block where the building to be evaluated is newly constructed.

[0033] Next, the control unit 21 of the support server 20 executes the acquisition process of the first wind information before the new construction (step S22). Specifically, the information acquisition unit 211 of the control unit 21 acquires the wind distribution before the change before the new construction of the building to be evaluated from the user terminal 10. This wind distribution before the change is a map showing the wind speed ratio distribution obtained by mapping the wind speed ratio in the three-dimensional model of only the existing buildings around the building to be evaluated.

[0034] Next, the control unit 21 of the support server 20 executes the prediction process of the second wind information after the new construction (step S23). Specifically, the information acquisition unit 211 of the control unit 21 calculates the second wind distribution (wind distribution after the change) after the new construction of the building to be evaluated by inputting the map to be evaluated and the wind distribution before the change into the prediction model recorded in the learning result storage unit 23.

[0035] As shown in FIG. 6, the map 400 to be evaluated and the first wind distribution 410 are input into the prediction model M1 to obtain the second wind distribution 420 in the map 400 to be evaluated. In this case, the influence of the wind by the building to be evaluated is confirmed by the second wind distribution 420. When wind countermeasures are necessary in the second wind distribution 420, a wind countermeasure model for wind countermeasures is arranged on the map 400 to be evaluated. Examples of this wind countermeasure model include structures such as windbreak plantings, windbreak nets / fences, and walls, and structures such as eaves. Also, the building shape may be changed. Then, the map to be evaluated with the wind countermeasures and the first wind distribution 410 are re-input into the prediction model to confirm the effect of the countermeasures.

[0036] (Operation of the Embodiment) Since the first wind information before the new construction of the new building is generated according to the arrangement information of the existing structures in the block, the second wind information of the new building considering the arrangement of the existing structures is calculated from the second arrangement information after the new construction of the new building.

[0037] (Effect of the Embodiment) (1) In this embodiment, the control unit 21 of the support server 20 executes the specific process of the first arrangement (step S11) and the acquisition process of the first wind information in the first arrangement (step S12). Thereby, the first wind distribution before the new construction of the structure to be evaluated can be grasped.

[0038] (2) In this embodiment, the control unit 21 of the support server 20 executes the specific process of the second arrangement (step S13) in which the first arrangement is changed, and the acquisition process of the second wind information in the second arrangement (step S14). Thereby, the second wind distribution after the new construction of the structure to be evaluated can be grasped.

[0039] (3) In this embodiment, the control unit 21 of the support server 20 executes the generation process of the teacher information (step S15). Thereby, the teacher information used for machine learning can be generated.

[0040] (4) In this embodiment, the control unit 21 of the support server 20 executes the machine learning process (step S16). Thereby, a prediction model for predicting the second wind information after the new construction of the building can be generated by using the second arrangement information after the new construction of the building and the first wind information before the new construction of the building as inputs.

[0041] (5) In this embodiment, the control unit 21 of the support server 20 executes the prediction process of the second wind information after the new construction (step S23). Thereby, if the preceding first wind information and the subsequent second arrangement information can be acquired, the subsequent second wind information can be predicted. The first wind information includes the wind situation before the arrangement of the structure to be evaluated. For example, in the area where the building exists, the wind speed becomes "0", so the first arrangement is incorporated into the first wind information, and thus the wind situation after the arrangement can be accurately predicted. In addition, in order to predict the change in the wind distribution based on the change in the arrangement, the calculation load can be reduced and the prediction can be efficiently performed.

[0042] This embodiment can be implemented with the following modifications. This embodiment and the following modification examples can be implemented in combination with each other within a technically non - conflicting range. ·In the above-described embodiment, the user terminal 10 and the support server 20 are used, but the hardware configuration is not limited thereto.

[0043] ·In the above-described embodiment, the wind speed ratio distribution is used as the wind information, but the wind information is not limited thereto. For example, a rank that is an index for evaluating the wind environment around a building may be used.

[0044] ·In the above-described embodiment, the first wind information and the second arrangement information are input to predict the second wind information. In this case, when wind countermeasures are taken in the second arrangement information, the first wind information and the second arrangement map with wind countermeasures are input to predict the second wind information. Here, the input of machine learning is not limited to the first wind information and the second arrangement information. For example, position information (latitude and longitude information) for specifying the building arrangement may be used. Alternatively, by adding the first arrangement information, the first arrangement information, the first wind information, and the second arrangement information may be input, and the second wind information may be output. In particular, when the building height is changed in the first arrangement information, it cannot be specified by a two-dimensional map indicating the presence or absence of a building, but the changed content can be specified from the difference between the first arrangement information and the second arrangement information. Also, only the difference from the first arrangement information may be used as the second arrangement information.

[0045] ·In the above-described embodiment, the first wind information and the second arrangement information are input to predict the second wind information. Here, the calculated second wind information (second wind information) may be used as the wind information before the change and the second arrangement information with wind countermeasures as input to predict the wind information after the change (third wind information). As shown in FIG. 7, when the wind distribution 501 is generated in the arrangement map 500, the map 510 in which the evaluation target T1 is arranged and the wind distribution 501 are input to the prediction model M1 to predict the wind distribution 511 (second wind information).

[0046] Next, when arranging the evaluation target T2, the arrangement map 520 (third arrangement) in which the evaluation target T2 is arranged and the wind distribution 511 are input to the prediction model M1 to predict the wind distribution 521 (third wind information). Furthermore, when placing the evaluation target T3, the placement map 530 (fourth placement) where the evaluation target T3 is placed and the wind distribution 521 are input into the prediction model M1 to predict the wind distribution 531 (fourth wind information). Thereby, when a part of the placement is changed, prediction can be efficiently performed.

[0047] · In the above embodiment, the control unit 21 of the support server 20 executes machine learning processing (step S16). Here, the second placement information after the new construction of the building is used. The change from the first placement in the second placement is not limited to the new construction of the building. It may be applied to building renovation, planting, installation of windbreak fences, etc. In this case, machine learning may be performed for each type of change (new installation, renovation, planting, etc.) to generate a prediction model according to the type.

[0048] · In the above embodiment, the control unit 21 of the support server 20 executes the acquisition process of the first wind information in the first placement (step S12) and the acquisition process of the second wind information in the second placement (step S14). In this case, the wind information is acquired from the user terminal 10. As long as the wind information can be acquired, it is not limited to the acquisition method. For example, when the support server 20 acquires the first placement and the second placement, the first and second wind information may be calculated by numerical fluid dynamics. Also, the first and second wind information may be acquired by on-site wind measurement. Further, teacher information may be generated by machine learning described in Patent Document 1.

[0049] · In the above embodiment, a new structure was placed with respect to the block map where the existing structure was placed. The placement is not limited to the case of newly constructing the evaluation target structure when the placement is changed. For example, it may be applied to terrain changes (changes in placement). Also, when planting trees, changes due to the growth of the trees (changes in placement) may be used as changes in placement. In this case, the wind distribution is calculated using the placement information obtained by simulating the shape change due to the growth of the trees.

[0050] In the above embodiment, the control unit 21 of the support server 20 executes a process of identifying a first arrangement (step S11) and a process of identifying a second arrangement obtained by modifying the first arrangement (step S13). Here, a map with hierarchical layers according to attributes may be used as the arrangement information. In this case, a different prediction model is generated for each layer by machine learning. This makes it possible to predict wind distribution according to attributes.

[0051] Next, the technical ideas that can be understood from the above embodiment and other examples will be described below. 2. The wind evaluation device according to claim 1, wherein: (a) the control unit acquires a wind speed ratio distribution as the first wind distribution and the second wind distribution.

[0052] (b) The wind evaluation device according to claim 1, characterized in that the control unit obtains a wind rank distribution as the first wind distribution and the second wind distribution. (c) the control unit When the first arrangement is acquired, the first wind distribution is calculated; When the second arrangement is acquired, the second wind distribution is calculated; 2. The wind evaluation device according to claim 1, wherein the calculated first and second wind distributions are used as the teaching information.

[0053] (d) the control unit, obtaining a third arrangement by further modifying the second arrangement, The wind distribution predicted by inputting it into the prediction model is used as the first wind distribution, and the third arrangement obtained by changing the evaluation target arrangement is input into the prediction model; 2. The wind evaluation device according to claim 1, further comprising: a wind evaluation device for predicting a wind distribution in the third arrangement. [Explanation of symbols]

[0054] A1...wind evaluation device, 10...user terminal, 20...support server, 21...control unit, 211...information acquisition unit, 212...learning unit, 213...prediction unit, 22...teacher information storage unit, 23...learning result storage unit.

Claims

1. A wind evaluation device comprising a teacher information storage unit, a learning result storage unit, and a control unit for predicting a wind distribution, wherein the control unit records teacher information combining a first wind distribution in a first arrangement, a second arrangement changed from the first arrangement, and a second wind distribution in the second arrangement in the teacher information storage unit, generates a prediction model for predicting the second wind distribution when changing from the first arrangement to the second arrangement by performing machine learning using the teacher information recorded in the teacher information storage unit, and records the prediction model in the learning result storage unit, When an evaluation target arrangement where an evaluation target is arranged and a wind distribution before change of the evaluation target are obtained, the evaluation target arrangement and the wind distribution before change are input into the prediction model recorded in the learning result storage unit to predict the wind distribution in the evaluation target arrangement. A wind evaluation device characterized by this.

2. A method for performing wind evaluation using a wind evaluation device comprising a teacher information storage unit, a learning result storage unit, and a control unit for predicting a wind distribution, wherein the control unit records teacher information combining a first wind distribution in a first arrangement, a second arrangement changed from the first arrangement, and a second wind distribution in the second arrangement in the teacher information storage unit, generates a prediction model for predicting the second wind distribution when changing from the first arrangement to the second arrangement by performing machine learning using the teacher information recorded in the teacher information storage unit, and records the prediction model in the learning result storage unit, When an evaluation target arrangement where an evaluation target is arranged and a wind distribution before change of the evaluation target are obtained, the evaluation target arrangement and the wind distribution before change are input into the prediction model recorded in the learning result storage unit to predict the wind distribution in the evaluation target arrangement. A wind evaluation method characterized by this.

3. A program for performing wind evaluation using a wind evaluation device comprising a teacher information storage unit, a learning result storage unit, and a control unit for predicting a wind distribution, wherein the control unit records teacher information combining a first wind distribution in a first arrangement, a second arrangement changed from the first arrangement, and a second wind distribution in the second arrangement in the teacher information storage unit, generates a prediction model for predicting the second wind distribution when changing from the first arrangement to the second arrangement by performing machine learning using the teacher information recorded in the teacher information storage unit, and records the prediction model in the learning result storage unit, A wind evaluation program, characterized in that when an evaluation target arrangement in which an evaluation target is arranged and a wind distribution before change of the evaluation target are obtained, the evaluation target arrangement and the wind distribution before change are input into a prediction model recorded in the learning result storage unit, and the program functions as means for predicting the wind distribution in the evaluation target arrangement.

Images