Wind state estimation device and wind state estimation method

Abstract

To make it possible to accurately predict a wind state during a non-observation period from wind state observation data in a short period in a wind state estimation target place.SOLUTION: A wind state estimation device 100 includes: a correction wind state estimation model generation part 112 for generating a wind state estimation model on the basis of a wind state prediction value in an estimation target place in a first period, a wind state observation value in the estimation target place in the first period, a wind state observation value in a surrounding place of the estimation target place in the first period, a wind state prediction value in the estimation target place in a second period, and a wind state observation value in the surrounding place in the second period; and an estimation part for estimating a wind state in the estimation target place in the second period by using the wind state estimation model. The wind state estimation model is a model which is used in estimating a wind state in the estimation target place in the second period on the basis of the wind state prediction value in the estimation target place in the second period.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a wind condition estimation device and a wind condition estimation method for estimating wind conditions at an estimation target site. [Background technology]

[0002] To decide on the construction site of a wind power plant, it is necessary to estimate (predict) the annual wind conditions (distribution of wind direction and speed at the height of the wind turbine) at the construction site and calculate the amount of power generation. To estimate the wind conditions, it is necessary to observe wind direction and speed. For observation, it is desirable to install an observation tower (mast, wind condition observation tower) and observe (measure) the annual wind conditions, but this is expensive. For this reason, it is necessary to estimate the annual wind conditions based on actual measurement data from a short observation period.

[0003] The wind condition estimation device described in Patent Document 1 comprises a memory unit storing calibration data indicating wind conditions measured near the installation location of a wind power generation facility and reference data indicating wind conditions obtained by a measurement method or analysis method that is homogeneous and whose observation location does not change during an observation period, a regression model learning unit that learns a regression model based on the calibration data and the reference data, where a first period is a period included in the calibration data, a second period is a period included in the reference data, and a third period is a period obtained by subtracting the first period from the second period, and a wind condition estimation unit that estimates wind conditions based on the reference data and the regression model for the third period, including the uncertainty of the estimation result. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 7114162 Summary of the Invention [Problem to be solved by the invention]

[0005] The wind condition estimation device described in Patent Document 1 can estimate wind conditions during non-observation periods, including their uncertainties. However, it is desirable to reduce errors contained in the estimation results of wind conditions in seasons different from the observation period. The present invention has been made in consideration of the above-mentioned background, and has an objective of providing a wind condition estimation device and a wind condition estimation method that enable highly accurate prediction of wind conditions during non-observation periods from short-term wind condition observation data at a location where wind conditions are to be estimated. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, the wind condition estimation device of the present invention comprises a corrected wind condition estimation model generation unit that generates a wind condition estimation model based on predicted wind conditions at an estimation target location for a first period, observed wind conditions at the estimation target location for the first period, observed wind conditions at the estimation target location for the first period, predicted wind conditions at the estimation target location for a second period, and observed wind conditions at the surrounding areas of the estimation target location for the second period, and an estimation unit that estimates wind conditions at the estimation target location for the second period using the wind condition estimation model, and the wind condition estimation model is a model used when estimating wind conditions at the estimation target location for the second period based on predicted wind conditions at the estimation target location for the second period. Effect of the Invention

[0007] According to the present invention, it is possible to provide a wind condition estimation device and a wind condition estimation method that enable highly accurate prediction of wind conditions during non-observation periods from short-term wind condition observation data at a target location for wind condition estimation. Problems, configurations and effects other than those described above will become clear from the description of the following embodiments. [Brief description of the drawings]

[0008] [Figure 1] 1 is a functional block diagram of a wind condition estimation device according to an embodiment of the present invention. [Diagram 2] 4 is a diagram for explaining a wind condition estimation model generated by a wind condition estimation model generating unit according to the embodiment; FIG. [Diagram 3] 1 is a diagram for explaining estimation of wind conditions at an estimation target site using a wind condition estimation model according to the present embodiment. FIG. [Figure 4] 4 is a flowchart of a wind condition estimation process according to the embodiment. [Diagram 5] 13 is a diagram for explaining a wind condition estimation model generated by a wind condition estimation model generating unit according to a modified example of the present embodiment. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] A wind condition estimation device in a form (embodiment) for carrying out the present invention will be described below. The wind condition estimation device generates a first wind condition estimation model, which is a regression model, based on the predicted wind conditions and observed wind conditions at the estimation target location in a first period. The wind condition estimation device also generates a second wind condition estimation model, which is a regression model, based on the predicted wind conditions at the estimation target location in a first period and the observed wind conditions at the surrounding locations of the estimation target location in the first period. The wind condition estimation device also generates a third wind condition estimation model, which is a regression model, based on the predicted wind conditions at the estimation target location in a second period and the observed wind conditions at the surrounding locations in the second period.

[0010] Next, the wind condition estimation device generates a wind condition estimation model, which is a regression model that estimates wind conditions from the predicted wind conditions at the estimation target location, based on the parameters of the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model. Finally, the wind condition estimation device estimates wind conditions based on the predicted wind conditions at the estimation target location for the second period using the wind condition estimation model.

[0011] Such a wind condition estimation device can predict wind conditions in a second period in which there are no wind condition observation values, based on wind condition observation values ​​in a first period. In actual operation, the second period is, for example, one year, and the first period is several months (for example, winter) of the year. Based on wind condition observation values ​​for a short period, annual wind conditions can be estimated with reduced seasonally dependent errors, which in turn allows wind conditions to be estimated at low cost. The first and second periods can be set arbitrarily, such as daily, monthly, or yearly.

[0012] <Configuration of the wind condition estimation device> 1 is a functional block diagram of a wind condition estimation device 100 according to this embodiment. The wind condition estimation device 100 is a computer, and includes a control unit 110, a storage unit 120, and an input / output unit 180. User interface devices such as a display, keyboard, and mouse are connected to the input / output unit 180. The input / output unit 180 may include a communication device, and may be capable of transmitting and receiving data to and from other devices. A media drive may also be connected to the input / output unit 180, and data may be exchanged using a recording medium.

[0013] <Wind condition estimation device: memory section> The storage unit 120 includes storage devices such as a Read Only Memory (ROM), a Random Access Memory (RAM), a Solid State Drive (SSD), etc. A wind condition database 130, a wind condition estimation model database 140, and a program 128 are stored in the storage unit 120. The program 128 includes a description of the wind condition estimation process (see FIG. 4 described later).

[0014] The wind condition database 130 stores observed and predicted wind conditions at the estimation target location and the surrounding areas of the estimation target location. In more detail, the wind condition database 130 includes predicted wind conditions at the estimation target location in a first period, observed wind conditions at the estimation target location in the first period, observed wind conditions in the surrounding areas in the first period, predicted wind conditions at the estimation target location in a second period, and observed wind conditions in the surrounding areas in the second period.

[0015] Here, wind condition observation values ​​(also simply referred to as observed values) are the results of measuring wind conditions (wind speed and direction) using masts, AMeDAS, Doppler lidar, etc. Wind condition forecast values ​​(also simply referred to as forecast values) are forecast values ​​of wind conditions obtained using a numerical weather forecast model. Surrounding areas are points close to the estimation target area where wind condition observation values ​​exist. The first period and the second period may be different periods or may be periods that overlap, the second period may include a period that is not included in the first period, or the second period may include the first period.

[0016] The wind condition estimation model database 140 stores wind condition estimation models that estimate wind conditions based on predicted wind condition values. The wind condition estimation models are regression models expressed as linear monomials. There are two types of wind condition estimation models: one that is generated based on observed wind condition values ​​and predicted wind condition values, and one that uses values ​​calculated based on parameters (coefficients) of the generated model as parameters.

[0017] <Wind condition estimation device: control unit> The control unit 110 includes a CPU (Central Processing Unit) and is provided with a wind condition estimation model generation unit 111, a corrected wind condition estimation model generation unit 112, and an estimation unit 113. The control unit 110 may include a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), etc.

[0018] <Control unit: Wind condition estimation model generation unit> The wind condition estimation model generating unit 111 generates a wind condition estimation model, which is a regression model expressed by a linear monomial, based on the wind condition observation values ​​and the wind condition prediction values. Fig. 2 is a diagram for explaining the wind condition estimation models 431 to 433 generated by the wind condition estimation model generating unit 111 according to this embodiment.

[0019] The wind condition estimation model generation unit 111 generates a wind condition estimation model 431 using learning data in which the wind condition predicted value 412 (referred to as "predicted value" in FIG. 2) at the estimation target location (referred to as "destination" in FIG. 2) in the first period is used as an explanatory variable, and the wind condition observed value 411 (referred to as "observed value" in FIG. 2) is used as a target variable. In addition, the wind condition estimation model generation unit 111 generates a wind condition estimation model 432 using learning data in which the wind condition prediction value 412 at the estimation target location in the first period is used as an explanatory variable and the wind condition observation value 413 at the surrounding location is used as a target variable.

[0020] Furthermore, the wind condition estimation model generation unit 111 generates a wind condition estimation model 433 using learning data in which the predicted wind condition value 414 at the estimation target location in the second period is used as an explanatory variable and the observed wind condition value 415 at the surrounding location is used as a target variable. The wind condition estimation models 431 to 433 are represented by formulas (1) to (3), respectively, with an explanatory variable X and a target variable Y. Note that the explanatory variable X may be continuous data of instantaneous values ​​(time-series data, for example, observed values ​​acquired at 10-minute intervals) or may be a period average value (for example, an hourly average value).

[0021] Y=A 1 X (1) Y=A 2 X (2) Y=A 3 X (3)

[0022] As described above, the wind condition estimation device 100 includes a wind condition estimation model generation unit 111 that generates a first wind condition estimation model (see wind condition estimation model 431) using learning data in which the predicted wind condition value 412 at the estimation target location in the first period is used as an explanatory variable and the observed wind condition value 411 at the estimation target location in the first period is used as a target variable. In addition, the wind condition estimation model generation unit 111 generates a second wind condition estimation model (see wind condition estimation model 432) using learning data in which the predicted wind condition value 412 at the estimation target location in the first period is used as an explanatory variable and the observed wind condition value 413 at the surrounding location in the first period is used as a target variable. Furthermore, the wind condition estimation model generation unit 111 generates a third wind condition estimation model (see wind condition estimation model 433) using learning data in which the predicted wind condition value 414 at the estimation target location in the second period is used as an explanatory variable and the observed wind condition value 415 at the surrounding location in the second period is used as a target variable.

[0023] <Control unit: Corrected wind condition estimation model generation unit> The corrected wind condition estimation model generation unit 112 corrects the wind condition estimation model 431 based on the wind condition estimation models 432 and 433. 2 , A 3Based on A 1 to generate wind condition estimation model 434 (see FIG. 3 described later) expressed by equation (4). It can be said that the corrected wind condition estimation model generation unit 112 generates the wind condition estimation model 434 based on the wind condition estimation models 431-433.

[0024] Y=(A 1 -A 2 +A 3 )X (4)

[0025] As described above, wind condition estimation model 434, the first wind condition estimation model (see wind condition estimation model 431), the second wind condition estimation model (see wind condition estimation model 432), and the third wind condition estimation model (see wind condition estimation model 433) are regression models represented by a first-order monomial with the predicted wind condition values ​​as the explanatory variables and the predicted wind condition values ​​as the objective variables. The wind condition estimation device 100 includes a corrected wind condition estimation model generation unit 112 that calculates the coefficients of a linear monomial of a regression model corresponding to the wind condition estimation model 434 by adding or subtracting coefficients of linear monomials corresponding to the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model. The corrected wind condition estimation model generation unit 112 generates a wind condition estimation model 434 based on the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model.

[0026] In other words, the corrected wind condition estimation model generation unit 112 generates a wind condition estimation model 434 based on a wind condition prediction value 412 at the estimation target location in a first period, a wind condition observation value 411 at the estimation target location in a first period, a wind condition observation value 413 at the surrounding area of ​​the estimation target location in a first period, a wind condition prediction value 414 at the estimation target location in a second period, and a wind condition observation value 415 at the surrounding area in a second period. The wind condition estimation model 434 is a model used when estimating the wind conditions (see wind condition estimation value 416) at the estimation target location in a second period based on the wind condition prediction value 414 at the estimation target location in a second period.

[0027] <Control unit: Estimation unit> The estimation unit 113 calculates a wind condition estimated value 416 (see FIG. 3 described later) based on the wind condition predicted value 414 at the estimation target location for the second time period using the wind condition estimation model 434. FIG. 3 is a diagram for explaining the estimation of wind conditions at the estimation target location using the wind condition estimation model 434 according to this embodiment. The estimation unit 113 calculates a wind condition estimated value 416 (referred to as "estimated value" in FIG. 3) based on the wind condition predicted value 414 at the estimation target location for the second time period using the wind condition estimation model 434.

[0028] As described above, the wind condition estimation device 100 includes the estimation unit 113 that uses the wind condition estimation model 434 to estimate wind conditions (see the wind condition estimate value 416) at the estimation target location in the second time period.

[0029] <Wind condition estimation processing> 4 is a flowchart of the wind condition estimation process according to this embodiment. At the start of the wind condition estimation process, the wind condition database 130 stores the predicted wind conditions at the estimation target location in the first period, the observed wind conditions at the estimation target location in the first period, the observed wind conditions at the surrounding locations in the first period, the predicted wind conditions at the estimation target location in the second period, and the observed wind conditions at the surrounding locations in the second period.

[0030] In step S11, the wind condition estimation model generating unit 111 generates the wind condition estimation models 431-433. In step S12, the corrected wind condition estimation model generation unit 112 generates the wind condition estimation model 434 based on the wind condition estimation models 431-433. In step S13, the estimation unit 113 uses the wind condition estimation model 434 to calculate the wind condition estimation value 416 based on the wind condition prediction value 414 at the estimation target location for the second time period.

[0031] <Features of the wind condition estimation device>

[0032] The wind condition estimation device 100 generates wind condition estimation models 431-433 based on the wind condition observation values ​​411, 413, 415 and the wind condition prediction values ​​412, 414. The wind condition estimation device 100 generates a wind condition estimation model 434 based on the wind condition estimation models 431-433 and estimates the wind conditions in the second period at the estimation target location. The wind condition observation value 411 at the estimation target location is only for the first period, but the wind condition estimation model 431 is corrected based on the wind condition observation value 413 for the first period and the wind condition observation value 415 for the second period at the surrounding location to estimate the wind condition estimated value 416 for the second period at the estimation target location. The wind condition estimated value 416 is calculated using the wind condition estimation model 434 that corrects the difference between the wind condition estimation model 431 that differs depending on the season, and is a highly accurate estimate.

[0033] <Modification: Generalized wind condition estimation model> In the above embodiment, the wind condition estimation models 431-434 are regression models of linear monomials. They are not limited to linear monomials, but may be linear expressions or polynomials of second or higher order, or may be regression models using machine learning techniques. It is assumed that the wind condition estimation models 431-433 represented by such generalized regression models are represented by equations (5)-(7). The corrected wind condition estimation model generation unit 112 then generates the wind condition estimation model 434 represented by equation (8).

[0034] Y=f 1 (X) (5) Y=f 2 (X) (6) Y=f 3 (X) (7) Y=f 1 (X)-f 2 (X)+f 3 (X) (8)

[0035] As explained above, wind condition estimation model 434, the first wind condition estimation model (see wind condition estimation model 431), the second wind condition estimation model (see wind condition estimation model 432), and the third wind condition estimation model (see wind condition estimation model 433) are regression models represented by functions with the predicted wind condition values ​​as explanatory variables and the predicted wind condition values ​​as objective variables. The corrected wind condition estimation model generation unit 112 calculates a function corresponding to the wind condition estimation model by adding and subtracting functions corresponding to the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model.

[0036] <<Variation: Wind condition estimation model using wind condition prediction values ​​for surrounding areas>> In the above embodiment, the corrected wind condition estimation model generation unit 112 corrects the wind condition estimation model 431 based on wind condition estimation models 432, 433 generated using surrounding area wind condition observation values ​​413, 415. The corrected wind condition estimation model generation unit 112 may correct the wind condition estimation model 431 based on wind condition estimation models 435, 436 generated using surrounding area wind condition prediction values ​​417, 418 (see FIG. 5 described later).

[0037] 5 is a diagram for explaining wind condition estimation models 431-433, 435, 436 generated by the wind condition estimation model generation unit 111 according to the modified example of this embodiment. The wind condition estimation model generation unit 111 generates a wind condition estimation model 435 using learning data in which a predicted wind condition value 417 in the surrounding area for a first time period is used as an explanatory variable and an observed wind condition value 413 is used as a target variable.

[0038] The wind condition estimation model generation unit 111 also generates a wind condition estimation model 436 using learning data in which the predicted wind condition value 418 in the surrounding area for the second period is an explanatory variable and the observed wind condition value 415 is a response variable. The wind condition estimation models 435 and 436 are assumed to be expressed by the following equations (9) to (10), with X as an explanatory variable and Y as a response variable.

[0039] Y=A 4 X (9) Y=A 5 X (10)

[0040] The corrected wind condition estimation model generation unit 112 corrects the wind condition estimation model 431 based on the wind condition estimation models 432, 433, 435, and 436. 2 , A3 , A 4 , A 5 Based on A 1 to generate the wind condition estimation model 434 expressed by equation (11). In other words, the corrected wind condition estimation model generation unit 112 generates the wind condition estimation model 434 based on the wind condition estimation models 431-433, 435, and 436.

[0041] Y=(A 1 -A 2 +A 3 -A 4 +A 5 )X (11)

[0042] The same is true for generalized wind condition estimation models, where the corrected wind condition estimation model generation unit 112 generates wind condition estimation model 434 based on wind condition estimation models 431-433, 435, and 436. It is assumed that wind condition estimation models 435 and 436 are expressed by equations (12) and (13). Then, the corrected wind condition estimation model generation unit 112 generates wind condition estimation model 434 expressed by equation (14).

[0043] Y=f 4 (X) (12) Y=f 5 (X) (13) Y=f 1 (X)-f 2 (X)+f 3 (X)-f 4 (X)+f 5 (X) (14)

[0044] <<Variation: Multiple Surrounding Areas>> In the above embodiment, there is one surrounding area, but there may be multiple surrounding areas. For example, assume that there are three surrounding areas, and the wind condition estimation model 432 is expressed by equations (15) to (17), and the wind condition estimation model 433 is expressed by equations (18) to (20). Then, the wind condition estimation model 434 is expressed by equation (21). Here, α 1 , α 2 , α 3 is a weighting coefficient whose sum equals 1, and the closer it is to the estimated location, the larger the value becomes.

[0045] Y=A 21 X (15) Y=A 22 X (16) Y=A 23 X (17) Y=A 31 X (18) Y=A 32 X (19) Y=A 33 X (20) Y=(A 1 -α 1 A 21 -α 2 A 22 -α 3 A 23 +α 1 A 31 +α 2 A 32 +α 3 A 33 )X (21)

[0046] As described above, the wind condition estimation model generation unit 111 generates a fourth wind condition estimation model (see wind condition estimation model 435) using learning data in which the predicted wind condition value 417 in the surrounding area in the first time period is used as an explanatory variable and the observed wind condition value 413 in the surrounding area in the first time period is used as a target variable. In addition, the wind condition estimation model generation unit 111 generates a fifth wind condition estimation model (see wind condition estimation model 436) using learning data in which the predicted wind condition value 418 in the surrounding area in the second time period is used as an explanatory variable and the observed wind condition value 415 in the surrounding area in the second time period is used as a target variable. The corrected wind condition estimation model generation unit 112 generates a wind condition estimation model 434 based on the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model, as well as the fourth wind condition estimation model and the fifth wind condition estimation model.

[0047] <Other wind condition estimation models> The wind condition estimation model generated by the corrected wind condition estimation model generation unit 112 is in the form of a linear expression of the base wind condition estimation model as shown in equations (8), (11), and (21), but is not limited to this. For example, instead of equation (8), a general equation such as the following equation (22) may be used.

[0048] Y = F(f 1 (X),f 2 (X),f 3 (X)) (22)

[0049] Other variations Although several embodiments of the present invention have been described above, these embodiments are merely illustrative and do not limit the technical scope of the present invention. The present invention can take various other embodiments, and various modifications such as omissions and substitutions can be made without departing from the gist of the present invention. These embodiments and their modifications are included in the scope and gist of the invention described in this specification, etc., and are included in the scope of the invention described in the claims and their equivalents. [Explanation of symbols]

[0050] 100 Wind condition estimation device 111 Wind condition estimation model generation unit 112 Corrected wind condition estimation model generation unit 113 Estimation Department 128 Programs 130 Wind Database 140 Wind condition estimation model database 411 Wind observation values ​​(wind observation values ​​at the estimated target area during the first period) 412 Wind forecast value (wind forecast value at the estimated target area in the first period) 413 Wind observation values ​​(wind observation values ​​in the surrounding areas of the estimated target area during the first period) 414 Wind forecast value (wind forecast value at the estimated target area in the second period) 415 Wind observation values ​​(wind observation values ​​in surrounding areas during the second period) 417 Wind forecast (wind forecast in surrounding areas during the first period) 418 Wind forecast (wind forecast in surrounding areas during the second period) 431 Wind Condition Estimation Model (First Wind Condition Estimation Model) 432 Wind Condition Estimation Model (Second Wind Condition Estimation Model) 433 Wind Condition Estimation Model (Third Wind Condition Estimation Model) 434 Wind Condition Estimation Model (Wind Condition Estimation Model) 435 Wind Condition Estimation Model (4th Wind Condition Estimation Model) 436 Wind Condition Estimation Model (5th Wind Condition Estimation Model)

Claims

1. a corrected wind condition estimation model generation unit that generates a wind condition estimation model based on a predicted wind condition value at the estimation target location in a first time period, an observed wind condition value at the estimation target location in the first time period, an observed wind condition value at a surrounding area of ​​the estimation target location in the first time period, a predicted wind condition value at the estimation target location in a second time period, and an observed wind condition value at the surrounding area in the second time period; an estimation unit that estimates wind conditions in the estimation target location during the second period using the wind condition estimation model, The wind condition estimation model is a model used when estimating wind conditions at the estimation target location in the second period based on a wind condition prediction value at the estimation target location in the second period. Wind condition estimation device.

2. a first wind condition estimation model using learning data in which a predicted value of wind conditions at the estimation target location in the first period is an explanatory variable and an observed value of wind conditions at the estimation target location in the first period is a target variable; A second wind condition estimation model using learning data in which the predicted wind conditions at the estimation target location in the first period are explanatory variables and the observed wind conditions at the surrounding location in the first period are target variables; and a wind condition estimation model generation unit that generates a third wind condition estimation model using learning data in which a predicted wind condition value at the estimation target location in the second time period is an explanatory variable and an observed wind condition value at the surrounding location in the second time period is a target variable; The corrected wind condition estimation model generation unit generating the wind condition estimation model based on the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model; The wind condition estimating device according to claim 1 .

3. the wind condition estimation model, the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model are regression models represented by linear monomials with the wind condition predicted value as an explanatory variable and the wind condition predicted value as a response variable, The corrected wind condition estimation model generation unit Calculating coefficients of a linear monomial of a regression model corresponding to the wind condition estimation model by adding and subtracting coefficients of linear monomials corresponding to the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model. The wind condition estimating device according to claim 2.

4. the wind condition estimation model, the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model are regression models represented by functions having the wind condition prediction value as an explanatory variable and the wind condition prediction value as a response variable, The corrected wind condition estimation model generation unit A function corresponding to the wind condition estimation model is calculated by adding and subtracting functions corresponding to the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model. The wind condition estimating device according to claim 2.

5. The wind condition estimation model generation unit A fourth wind condition estimation model using learning data in which the predicted wind condition value in the surrounding area during the first period is an explanatory variable and the observed wind condition value in the surrounding area during the first period is a target variable; and Further generating a fifth wind condition estimation model using learning data in which the predicted wind condition value in the surrounding area during the second time period is an explanatory variable and the observed wind condition value in the surrounding area during the second time period is a target variable; The corrected wind condition estimation model generation unit The wind condition estimation model is generated based on the fourth wind condition estimation model and the fifth wind condition estimation model in addition to the first wind condition estimation model, the second wind condition estimation model, and the third wind condition estimation model. The wind condition estimating device according to claim 2.

6. The second period includes a period not included in the first period. The wind condition estimating device according to claim 1 .

7. The first period is included in the second period The wind condition estimating device according to claim 6.

8. The wind condition estimation device, generating a wind condition estimation model based on a predicted wind condition value at the estimation target location in a first time period, an observed wind condition value at the estimation target location in the first time period, an observed wind condition value at a surrounding area of ​​the estimation target location in the first time period, a predicted wind condition value at the estimation target location in a second time period, and an observed wind condition value at the surrounding area in the second time period; estimating wind conditions at the estimation target location for the second period using the wind condition estimation model; The wind condition estimation model is a model used when estimating wind conditions at the estimation target location in the second period based on a wind condition prediction value at the estimation target location in the second period. Wind estimation methods.

Images