Weather forecasting device and weather forecast model learning device

The weather prediction system uses observed and calculated meteorological data with a trained model to account for regional environments, improving forecast accuracy and enabling applications like building planning and disaster prevention.

JP7576960B2Active Publication Date: 2024-11-01TAKENAKA CORP
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
JP2020186837
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-11-09
Publication Date
2024-11-01
Estimated Expiration
2040-11-09

AI Technical Summary

Technical Problem

Existing weather prediction techniques fail to consider the surrounding environment, such as buildings and topography, when predicting weather data at a certain point or within a region, limiting their accuracy and applicability.

Method used

A weather prediction system that incorporates first and second observed meteorological data, combined with computational fluid dynamics analysis, to generate calculated weather data, which is then used with a trained model to predict weather within a region, considering its environmental factors.

Benefits of technology

Accurately predicts weather data within a region by accounting for local environmental conditions, enhancing the precision of forecasts and enabling applications like building planning and disaster prevention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007576960000001
    Figure 0007576960000001
  • Figure 0007576960000002
    Figure 0007576960000002
  • Figure 0007576960000003
    Figure 0007576960000003
Patent Text Reader

Abstract

To predict weather data considering an environment in an area in predicting the weather data in the area.SOLUTION: An acquisition unit 26 of a weather prediction device 14 acquires first observation weather data observed at an observation site and second observation weather data observed by a weather observation device installed at a predetermined site. A calculation unit 28 generates a calculation result of numerical hydrodynamics analysis associated with weather in an area by performing numerical hydrodynamics analysis associated with the weather in the area in a predetermined range including the observation site, and obtains calculation weather data being weather data obtained from the calculation result. A prediction unit 30 predicts weather data in the area by inputting the first observation weather data and the second observation weather data, and the calculation weather data for a previously learned model for predicting weather data at time being later than predetermined time from the weather data at the predetermined time.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a weather forecasting device and a weather forecast model learning device. [Background technology]

[0002] Conventionally, there are known techniques for predicting weather data. For example, there is known a technique for estimating weather data of a user located in a place other than an observation station from weather data of the observation station by using a neural network (for example, paragraph

[0010] of Patent Document 1).

[0003] Also, a technique is known for determining partial differential coefficients of atmospheric pressure in the east-west and north-south directions from the atmospheric pressure values ​​at a number of surrounding points and the atmospheric pressure value at the local point (for example, [Claim 2] of Patent Document 2). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 6-331756 [Patent Document 2] Japanese Patent Application Publication No. 9-49884 Summary of the Invention [Problem to be solved by the invention]

[0005] However, weather data such as wind speed, wind direction, temperature, humidity, precipitation, solar radiation, and atmospheric pressure at a certain point may be affected by the surrounding environment such as buildings and topography around the point. Therefore, when predicting weather data at a certain point, it is necessary to take the surrounding environment of the point into consideration. In addition, there are cases where it is desired to predict weather data within a region including a certain point, rather than to predict weather data at the certain point.

[0006] However, the techniques disclosed in the above Patent Documents 1 and 2 predict future weather data at a certain location using observed weather data, and do not take into account the surrounding environment of the location. Furthermore, the techniques disclosed in the above Patent Documents 1 and 2 predict weather data at a certain location, but do not predict weather data within a certain area.

[0007] In consideration of the above, an object of the present invention is to predict weather data in a certain area by taking into account the environment in the area. [Means for solving the problem]

[0008] A first aspect of the present invention is a weather forecasting device that includes an acquisition unit that acquires first observed weather data observed at an observation point and second observed weather data observed by a meteorological observation device installed at a predetermined point, a calculation unit that performs a computational fluid dynamics analysis on weather in an area within a predetermined range including the observation point, thereby generating calculation results of the computational fluid dynamics analysis on weather in the area, and obtains calculated weather data, which is weather data obtained from the calculation results, and a prediction unit that predicts weather data in the area by inputting the first observed weather data and the second observed weather data acquired by the acquisition unit and the calculated weather data obtained by the calculation unit into a pre-trained model for predicting weather data at a time later than a predetermined time from weather data at the predetermined time.

[0009] A second aspect of the present invention is a weather prediction model learning device that includes a learning unit that generates a trained model for predicting weather data at a time later than a specified time from weather data at a specified time, by learning a model for predicting weather data at a time later than a specified time from weather data at the specified time, based on learning data in which first observed weather data for learning observed at an observation point, second observed weather data for learning observed by a weather observation device installed at a specified point, calculated weather data for learning, which is weather data obtained from calculation results of a computational fluid dynamics analysis related to the weather in an area within a predetermined range including the observation point, and weather data for learning within the area. Effect of the Invention

[0010] According to the present invention, when weather data in a certain area is predicted, it is possible to obtain an effect that the weather data can be predicted taking into consideration the environment in the area. [Brief description of the drawings]

[0011] [Figure 1] 1 is a block diagram showing a weather prediction system according to an embodiment. [Diagram 2] FIG. 1 is a diagram for explaining an overview of an embodiment. [Diagram 3] FIG. 4 is a diagram for explaining learning data according to the embodiment; [Figure 4] FIG. 2 is a diagram for explaining a model of the first embodiment. [Diagram 5] FIG. 1 is a diagram of a hypothetical building and terrain model within an area in which a CFD simulation is performed. [Figure 6] FIG. 11 is a diagram for explaining correction of meteorological data. [Figure 7] 5 is a flowchart showing the contents of a learning process routine of the weather prediction system according to the embodiment. [Figure 8] 4 is a flowchart showing the contents of a weather forecast processing routine of the weather forecasting system according to the embodiment. [Figure 9]FIG. 11 is a diagram for explaining a model of the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0013] [First embodiment]

[0014] <Configuration of the Weather Forecasting System of the Present Embodiment>

[0015] Fig. 1 is a block diagram showing an example of the configuration of a weather prediction system 10 according to the first embodiment. Functionally, the weather prediction system 10 can be represented as including an input unit 12, a weather prediction device 14, and a display unit 16, as shown in Fig. 1.

[0016] The input unit 12 receives data input by a user, etc. The input unit 12 is realized by, for example, a mouse, a keyboard, etc.

[0017] The weather forecasting device 14 is realized by a computer including a CPU (Central Processing Unit), a ROM (Read Only Memory) that stores programs for realizing each processing routine, a RAM (Random Access Memory) that temporarily stores data, a memory as a storage means, and a network interface, etc.

[0018] As shown in FIG. 1, the weather forecasting device 14 functionally comprises a learning data storage unit 20, a learning unit 22, a trained model storage unit 24, an acquisition unit 26, a calculation unit 28, and a prediction unit 30.

[0019] FIG. 2 is a diagram for explaining the outline of this embodiment. As shown in FIG. 2, when predicting weather data in a certain area, the weather forecasting device 14 of the first embodiment predicts weather data in the area A using weather data observed by a weather observation device I1 at an observation point in the area (hereinafter simply referred to as "first observed weather data"), weather data observed by a weather observation device I2 installed at a predetermined point (hereinafter simply referred to as "second observed weather data"), and weather data obtained from the calculation result of a computational fluid dynamics analysis on the weather in the area (hereinafter simply referred to as "calculated weather data"). The weather data includes data such as wind speed, wind direction, temperature, humidity, precipitation, solar radiation, and atmospheric pressure. The second observed weather data is, for example, data observed by a weather observation device installed throughout the country for a prediction system such as so-called AMeDAS.

[0020] When forecasting weather data in a certain area, it is necessary to take into account the environment of the area, such as buildings and topography, etc. For example, if the area is a city block, it is necessary to take into account the presence of buildings, etc.

[0021] Therefore, the weather forecasting device 14 of this embodiment predicts weather data within the region using first observed weather data observed at an observation point within the region, second observed weather data observed by a weather observation device installed at a specified point, and also calculated weather data obtained from the calculation results of a computational fluid dynamics analysis of the weather within the region. By using the calculation results of a computational fluid dynamics analysis of the weather within the region, the environment such as buildings and topography within the region is taken into consideration, making it possible to accurately predict weather data within the region.

[0022] Weather forecast information for a region can be used when planning buildings in that region. When planning a building, it is necessary to perform an evaluation of wind around the building, an evaluation of habitability due to wind shaking, and an evaluation of natural ventilation. At this time, weather data is required for these evaluations.

[0023] If the weather forecasting device 14 of this embodiment can obtain forecast information of weather data within the area in which the planned building is located, the forecast information of weather data within the area can be used for building wind evaluation, livability evaluation, natural ventilation evaluation, etc. of the building.

[0024] In addition, the weather forecasting device 14 of this embodiment uses a trained model generated in advance by machine learning when predicting weather data in the region from the first observed weather data, the second observed weather data, and the calculation target data. Since the trained model reflects learning data corresponding to the results collected in advance, it is possible to accurately predict weather data in the region.

[0025] The specific details will be explained below.

[0026] A plurality of pieces of learning data are stored in the learning data storage unit 20. The learning data in this embodiment is data in which first observed weather data for learning, second observed weather data for learning, calculated weather data for learning, and weather data within the region for learning are associated with each other. The learning data is, for example, data based on past performance.

[0027] A diagram for explaining the learning data of this embodiment is shown in Fig. 3. As shown in Fig. 3, the learning data is stored in such a manner that the first observed weather data for learning, the second observed weather data for learning, the calculated weather data for learning, and the weather data for learning within the region are associated with each other.

[0028] 3 indicates that when the first observed meteorological data for learning is "A1,...", the second observed meteorological data for learning is "D1,...", and the calculated meteorological data for learning is "F1,...", the meteorological data for learning in a certain region is "Z1,...". Each meteorological data includes data such as wind speed, wind direction, temperature, humidity, precipitation, solar radiation, and air pressure.

[0029] The learning unit 22 generates a trained model for predicting weather data at a time later than a specified time from weather data at a specified time by learning a model for predicting weather data at a time later than a specified time from weather data at a specified time based on multiple learning data stored in the learning data memory unit 20.

[0030] Fig. 4 shows a diagram for explaining the model of the first embodiment. As shown in Fig. 4, the model of the first embodiment is a neural network. For this reason, a trained neural network M1 is generated by the learning unit 22. When the first observed weather data, the second observed weather data, and the calculated weather data are input to the trained neural network M1, weather data within the area to be predicted is output.

[0031] For this reason, the learning unit 22 trains the neural network by known supervised machine learning based on the multiple pieces of learning data stored in the learning data storage unit 20, and generates a trained neural network. Note that any method of supervised machine learning may be used.

[0032] Then, the learning unit 22 stores the generated trained model in the trained model storage unit 24.

[0033] The trained model storage unit 24 stores the trained model generated by the training unit 22.

[0034] The acquisition unit 26 acquires data of a prediction target. Specifically, the acquisition unit 26 acquires first observed weather data observed at a certain observation point and second observed weather data observed by a meteorological observation device installed at a predetermined point.

[0035] Calculation unit 28 performs a meteorological computational fluid dynamics (CFD) analysis (hereinafter, simply referred to as a "CFD simulation") in an area within a predetermined range including the observation point of the first observed weather data. Specifically, calculation unit 28 performs a CFD simulation using wind speed, temperature, humidity, solar radiation, precipitation, air pressure, and the like represented by the first observed weather data and the second observed weather data as input data. Moreover, the input data to calculation unit 28 may be arbitrarily set in addition to using the first observed weather data and the second observed weather data.

[0036] When a CFD simulation is performed within a region, the buildings, topography, etc. within the region are modeled. Specifically, for example, the materials, etc. of the buildings are set in advance, and the CFD simulation within the region is performed under those conditions.

[0037] A model of virtual buildings and terrain in an area when a CFD simulation is performed is shown in Fig. 5. As shown in Fig. 5, the model V of virtual buildings and terrain in an area includes, for example, buildings B and trees t. The model V of virtual buildings and terrain in an area is realized in a computer, and the CFD simulation is performed according to the model V of virtual buildings and terrain.

[0038] The calculation unit 28 generates a calculation result of a computational fluid dynamics analysis on the weather in the area by executing a CFD simulation. Then, the calculation unit 28 obtains computed meteorological data, which is meteorological data obtained from the calculation result. Note that any method of computational fluid dynamics analysis may be used.

[0039] The prediction unit 30 predicts weather data within the area by inputting the first observation weather data and the second observation weather data acquired by the acquisition unit 26 and the calculated weather data obtained by the calculation unit 28 to the learned model stored in the learned model storage unit 24. Specifically, for example, the prediction unit 30 predicts weather data for multiple points within the area. Alternatively, the prediction unit 30 may predict weather data for one point within the area. Note that the data input to the learned model is weather data at a specified time, and the data output from the learned model is weather data at a time later than the specified time.

[0040] In this way, the weather forecasting device 14 of the first embodiment uses calculated weather data obtained from a CFD simulation in the area when forecasting weather data in the area. This allows buildings, topography, and the like in the area to be forecasted to be taken into account, making it possible to forecast weather data in the area to be forecasted with high accuracy.

[0041] Next, the prediction unit 30 statistically corrects the weather data within the region output from the trained model. For example, as shown in FIG. 6, the prediction unit 30 corrects the weather data P1 within the region by adding a predetermined value to the weather data P1 within the region output from the trained model. The weather data within the region output from the trained model may deviate from the actual measured value. Therefore, the prediction unit 30 generates corrected weather data P2 within the region by statistically correcting the weather data P1 within the region output from the trained model. Any specific correction method may be used.

[0042] The display unit 16 displays the information output from the weather prediction device 14. The display unit 16 is realized by, for example, a display. The display unit 16 displays the prediction result obtained by the prediction unit 30. The user checks the prediction result displayed on the display unit 16 and uses it, for example, for building planning.

[0043] <The Function of Weather Forecasting Systems>

[0044] Next, the operation of the weather prediction system 10 of the first embodiment will be described with reference to the drawings. When the weather prediction device 14 of the weather prediction system 10 receives learning data via the input unit 12, the weather prediction device 14 stores the learning data in the learning data storage unit 20. Then, when the weather prediction device 14 of the weather prediction system 10 receives an instruction signal to start a learning process, the weather prediction device 14 executes a learning process routine shown in FIG.

[0045] <Learning processing routine>

[0046] In step S100 , the learning unit 22 acquires a plurality of pieces of learning data stored in the learning data storage unit 20 .

[0047] In step S102, the learning unit 22 generates a trained neural network for predicting weather data by training a neural network, which is an example of a model, based on the multiple learning data acquired in step S100.

[0048] In step S104, the learning unit 22 stores the trained model generated in step S102 in the trained model storage unit 24, and ends the learning process routine.

[0049] Next, when the weather forecasting device 14 of the weather forecasting system 10 receives the first observed weather data and the second observed weather data of an observation point within the area to be predicted input from the input unit 12, it executes the weather forecast processing routine shown in FIG. 8.

[0050] <Weather forecast processing routine>

[0051] In step S200, the acquisition unit 26 acquires the first observed weather data input from the input unit 12.

[0052] In step S202, the acquisition unit 26 acquires the second observed weather data input from the input unit 12.

[0053] In step S204, the calculation unit 28 executes a CFD simulation in an area including the observation point of the first observation weather data based on the first observation weather data acquired in step S200 and the second observation weather data acquired in step S202. Then, the calculation unit 28 obtains calculated weather data from the calculation result of the CFD simulation.

[0054] In step S206, the prediction unit 30 reads out the trained neural network, which is the trained model stored in the trained model storage unit 24.

[0055] In step S208, the prediction unit 30 predicts weather data within the region by inputting the first observed weather data acquired in step S200, the second observed weather data acquired in step S202, and the calculated weather data obtained in step S204 into the trained neural network read out in step S206.

[0056] In step S210, the prediction unit 30 statistically corrects the weather data within the area obtained in step S208 above.

[0057] In step S212, the prediction unit 30 outputs the corrected weather data within the area obtained in step S210 as a result.

[0058] The display unit 16 displays information output from the weather forecasting device 14. The user checks the information displayed on the display unit 16 and uses it for, for example, building planning.

[0059] As described above in detail, the weather forecasting device of the first embodiment acquires the first observed weather data observed at the observation point and the second observed weather data observed by a weather observation device installed at a predetermined point. Then, the weather forecasting device executes a CFD simulation on the weather in an area within a predetermined range including the observation point to generate a calculation result of the CFD simulation on the weather in the area, and obtains calculated weather data, which is weather data obtained from the calculation result. The weather forecasting device predicts weather data in the area by inputting the acquired first observed weather data and second observed weather data and the calculated weather data to a trained model trained in advance for predicting weather data at a time after a predetermined time from weather data at a predetermined time. This makes it possible to predict weather data in a certain area by taking into account the environment in the area.

[0060] More specifically, the weather forecasting device of the first embodiment performs a CFD simulation of the weather in an area, and by using calculated weather data, which is the weather data obtained from the results, it is possible to predict weather data taking into account the environment in the area.

[0061] In addition, by using a trained model instead of a complex physical model, computational costs and time can be reduced.

[0062] In addition, the weather forecasting device of the first embodiment can change the weather data to be input to the trained model in various ways. For example, if weather data of the AMeDAS data, which is an example of the second observed weather data, that shows a 5 degree increase in temperature and a 20% decrease in humidity is input to the trained model, the weather data in a certain area under those conditions can be predicted. That is, the weather data in a certain area under weather conditions that may occur in the future can be predicted. In addition, past weather can be reproduced by inputting past weather data into the trained model. When reproducing past weather, weather data at a certain time in the past is predicted from the weather data at the certain time, and the past weather is reproduced.

[0063] Furthermore, in conventional techniques (for example, Patent Document 1 above), predictions are made based on weather data observed at a certain time at a location different from the location to be predicted. Therefore, the time of prediction of the weather data at the location to be predicted is limited to the time when the weather data was observed at the different location. In contrast, the weather prediction device of the first embodiment can predict future weather data (for example, about 10 minutes to several hours from now) within the region to be predicted from data obtained at existing observation locations.

[0064] In addition, by using the calculated weather data obtained by CFD simulation, it is possible to accurately predict not only the weather data for a certain location, but also the weather data for the area that includes that location.

[0065] In addition, it will be possible to pinpoint weather forecasts for the location of the work site at the construction site, and if strong winds and heavy rain are predicted, it will be possible to take immediate measures in advance, thereby reducing the occurrence of disasters at construction sites.

[0066] In addition, it is possible to design buildings to be wind-resistant based on the vertical wind speed in the predicted weather data, which allows for more rational design and reduces costs.

[0067] In addition, evaluation of the habitability of the planned building and prediction of wind pressure across the building were conventionally performed based on meteorological data observed at existing observation points. By contrast, the use of the weather forecasting device of the present embodiment makes it possible to accurately predict meteorological data within an area that includes a certain point, thereby further improving the safety of the building.

[0068] In addition, natural ventilation prediction was also evaluated and predicted based on meteorological data observed at existing observation points in the past. In contrast, by using the weather prediction device of the present embodiment, it is possible to accurately predict meteorological data within an area including a certain point, enabling efficient energy management in architectural planning.

[0069] In addition, heat load calculations have traditionally been evaluated and predicted based on meteorological data observed at existing observation points. In contrast, the use of the weather forecasting device of this embodiment makes it possible to accurately predict meteorological data within an area that includes a certain point, enabling rational facility design. This allows the cost of building facilities to be reduced.

[0070] Furthermore, by using the weather forecasting device of the first embodiment, it is possible to reduce costs since there is no need to set up a large number of observation points.

[0071] [Second embodiment]

[0072] Next, a second embodiment will be described. The second embodiment differs from the first embodiment in that the trained model is a decision tree model. Note that the configuration of the weather forecasting system according to the second embodiment is the same as that of the first embodiment, so the same reference numerals are used and the description is omitted.

[0073] An example of a model according to the second embodiment is shown in Fig. 9. As shown in Fig. 9, the model according to the second embodiment is a decision tree model. The decision tree model according to this embodiment is generated for each piece of meteorological data. In the example of Fig. 9, a decision tree model M2-1 for predicting wind speed, a decision tree model M2-2 for predicting temperature, and a decision tree model M2-3 for predicting precipitation are shown.

[0074] For example, when each weather data and time information of the forecast target are input to the wind speed forecasting decision tree model M2-1, the following is generated: 2 The judgment is made at the node "Is it more than 600W / m2?" 2 If it is more than this, it will move to the node on the left. On the other hand, if the solar radiation is 600W / m2 If it is less than 1, the process moves to the node on the right. Finally, a predicted value such as "wind speed xx m / s" is output from the wind speed prediction decision tree model M2-1. Note that decision tree models for predicting other meteorological data are also generated in the learning unit 22, which will be described later.

[0075] The learning unit 22 of the weather forecasting device 14 of the second embodiment, like the first embodiment, trains a decision tree model using a supervised machine learning algorithm for a known decision tree model based on a plurality of learning data, and generates trained decision tree models M2-1, M2-2, M2-3, etc. Then, the learning unit 22 stores the trained decision tree models M2-1, M2-2, M2-3, etc. in the trained model storage unit 24.

[0076] As in the first embodiment, the prediction unit 30 of the weather prediction device 14 of the second embodiment predicts weather data within an area by inputting the first observed weather data, the second observed weather data, and the calculated weather data to the trained decision tree models M2-1, M2-2, M2-3, etc. Note that the decision tree model M2-1 for predicting wind speed outputs a predicted value of wind speed, the decision tree model M2-2 for predicting temperature outputs a predicted value of temperature, and the decision tree model M2-3 for predicting precipitation outputs a predicted value of precipitation.

[0077] From the trained decision tree models M2-1, M2-2, M2-3, etc., it is possible to obtain the importance that indicates how much the division of the feature at the node contributes to the prediction. 2 This corresponds to the "amount of solar radiation" in the division of "more than 100%?". The importance is calculated, for example, from known Gini impurity.

[0078] The user can check the importance of the features and which features contribute to the prediction of the weather data, which can be useful for, for example, building planning.

[0079] Other configurations and operations of the weather prediction system according to the second embodiment are similar to those of the first embodiment, and therefore description thereof will be omitted.

[0080] As described above in detail, the weather forecasting device of the weather forecasting system of the second embodiment predicts weather data in a region using a decision tree model as a trained model. When predicting weather data in a certain region, the weather data can be predicted taking into account the environment in the region. In addition, the importance of features when predicting weather data can be obtained.

[0081] 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.

[0082] For example, in each of the above embodiments, a neural network or a decision tree model is used as an example of a trained model, but the trained model is not limited to this. Any model obtained by machine learning may be used.

[0083] In addition, although the above describes an embodiment in which the program according to the present invention is pre-stored (installed) in a memory unit (not shown), the program according to the present invention can also be provided in a form in which it is recorded on a recording medium such as a CD-ROM, a DVD-ROM, or a micro SD card. [Explanation of symbols]

[0084] 10. Weather Forecasting System 12 Input section 14 Weather forecasting equipment 16 Display 20 Learning data storage unit 22 Learning Department 24 Trained model memory 26 Acquisition Department 28 Calculation section 30 Prediction Department

Claims

1. an acquisition unit that acquires first observed weather data observed at an observation point and second observed weather data observed by a meteorological observation device installed at a predetermined point; a calculation unit that performs a computational fluid dynamics analysis on meteorology in a region within a predetermined range including the observation point, thereby generating a computation result of the computational fluid dynamics analysis on meteorology in the region, and obtains computed meteorological data that is meteorological data obtained from the computation result; a prediction unit that predicts weather data in the area by inputting the first observed weather data and the second observed weather data acquired by the acquisition unit and the calculated weather data obtained by the calculation unit into a trained model that has been trained in advance to predict weather data at a time after a predetermined time from weather data at the predetermined time; A weather forecasting device comprising:

2. At least one building exists within the area that is the subject of the computational fluid dynamics analysis of the meteorological phenomena. The weather forecasting device according to claim 1 .

3. A learning unit that generates a trained model for predicting weather data at a time later than a predetermined time from weather data at a predetermined time by learning a model for predicting weather data at a time later than a predetermined time from weather data at the predetermined time based on learning data in which first observed weather data for learning observed at an observation point, second observed weather data for learning observed by a weather observation device installed at a predetermined point, calculated weather data for learning, which is weather data obtained from a calculation result of a computational fluid dynamics analysis related to weather in an area within a predetermined range including the observation point, and weather data for learning in the area are associated with each other. A weather forecast model learning device comprising:

4. At least one building exists within the area that is the subject of the computational fluid dynamics analysis of the meteorological phenomena. The weather forecast model learning device according to claim 3.

Citation Information

Patent Citations

  • Forecast system

    JP1994331756A

  • Predicting method for local weather using neural network

    JP1997049884A

  • Method for weather forecasting for narrow area, method for distribution of weather forecast for narrow area, method for commodity sales promotion, weather forecasting device, and weather forecasting information providing system

    JP2003021687A

  • Wind condition prediction system and wind condition prediction method

    JP2020159725A

  • Weather forecasting system

    US20170192128A1