CN111400844A

CN111400844A - Parameter scheme set generation method and wind speed forecasting method of meteorological model

Info

Publication number: CN111400844A
Application number: CN201811608828.XA
Authority: CN
Inventors: 刘钊; 王瑾; 敖娟
Original assignee: Beijing Goldwind Science and Creation Windpower Equipment Co Ltd
Current assignee: Beijing Goldwind Science and Creation Windpower Equipment Co Ltd
Priority date: 2018-12-27
Filing date: 2018-12-27
Publication date: 2020-07-10

Abstract

A parameter scheme set generation method and a wind speed forecasting method of a meteorological model are provided. The method for generating the parameter scheme set of the meteorological model comprises the following steps: selecting an initial parameter scheme from all selectable parameter schemes of the meteorological model of the target area; determining parameter scheme samples capable of normally operating from the initial parameter scheme; and setting the meteorological model by using the set of parameter schemes sampled from the parameter scheme sample to calculate a forecast result, and screening the parameter scheme set for the meteorological model according to the difference between the forecast result and the observation data of the target area.

Description

Parameter scheme set generation method and wind speed forecasting method of meteorological model

Technical Field

The present disclosure relates to a method for generating a parameter plan of a meteorological model and a method for forecasting a wind speed, and more particularly, to a method for generating a parameter plan set of a meteorological model and a method for forecasting a wind speed using the generated parameter plan set.

Background

When wind resource assessment (or wind speed forecast) is performed on a wind farm, long-term (e.g., one year) wind measurement data of a target area is often required to be able to more accurately assess the wind resources of the target area. However, in the actual development of wind farms, it is often difficult to perform long-term wind surveys in the target area due to capital investment and construction cycle limitations. Under the condition, more and more wind power plant development projects begin to evaluate wind resources of the wind power plant under the condition that the time for measuring the wind of the target area is insufficient or wind measurement is not performed at all, so that great uncertainty is brought to calculation of the generated energy of the wind power plant and prediction of income. In order to deal with the situation, developers of the wind power plant development project are encouraged to pay attention to the importance of wind measurement and select representative positions in a target area as much as possible for long-term wind measurement, and on the other hand, a new method is introduced for wind power plant development projects with insufficient wind measurement time or without wind measurement to perform wind resource assessment.

At present, under the condition that wind measuring time is insufficient or wind measuring data quality is poor, so that the wind measuring data is insufficient or no wind measuring data exists, conventional mesoscale data of a target position are generally used as wind measuring data to perform wind resource assessment, such as mesoscale data downloaded from data sources such as NNRP, ERPI, MERRA2 and the like. However, the mesoscale data is the result of a single experiment in a numerical mode, and the prediction effects of the single experiment in different areas have large differences, so that the stability of the evaluation result of wind resource evaluation by using the mesoscale data cannot be ensured, and the accuracy of the evaluation result cannot be ensured.

Disclosure of Invention

The present disclosure is proposed to solve at least the above disadvantages and to provide the advantages described below.

An aspect of the present disclosure is to provide a method and an apparatus for generating a parameter scheme set of a meteorological model, which can generate a parameter scheme library covering various weather conditions, thereby ensuring accuracy and stability of wind speed forecasting of a wind farm using the generated parameter scheme.

Another aspect of the present disclosure is to provide a wind speed forecasting method for a wind farm, which can ensure stability and accuracy of a wind speed forecasting result of the wind farm, so as to more accurately calculate the generated energy and forecast yield of the wind farm in the process of developing the wind farm, and avoid resource waste.

According to an aspect of the present disclosure, there is provided a method for generating a set of parameter solutions for a meteorological model, the method comprising: selecting an initial parameter scheme from all selectable parameter schemes of the meteorological model of the target area; determining parameter scheme samples capable of normally operating from the initial parameter scheme; setting the meteorological model to operate on a forecast result by using a set of parameter solutions sampled from the parameter solution samples; and screening to obtain a parameter scheme set for the meteorological model according to the difference between the forecast result and the observation data of the target area.

The initial parametric approach may be a parametric approach selected from the total number of selectable parametric approaches based on characteristics of the target area, wherein the characteristics of the target area may include at least one of meteorological characteristics, topographical characteristics, and wind farm characteristics.

The step of determining a parameter scenario sample that can function properly from the initial parameter scenario may comprise: bringing an initial parameter scheme into and running the meteorological model; and determining a parameter scheme sample capable of normally operating according to the operation result.

The observation data may be observation data within a date period capable of characterizing meteorological features selected from long-term observation data or reanalysis data of the target area according to meteorological elements and screening rules of the target area set by an experimenter, wherein the observation data may include at least one of meteorological tower data, radar data and meteorological station data.

The step of screening the set of parameter solutions for the meteorological model according to the difference between the forecast result and the observation data of the target area may include: sampling the parameter scheme samples to obtain a sampling sample set and a residual sample set; selecting a parameter scheme from the sampling sample set, substituting the parameter scheme selected from the sampling sample set into the meteorological model for operation to obtain a forecast result in the date period, and calculating a first difference value between the forecast result and observation data in the date period; selecting parameter schemes from the residual sample sets, substituting the parameter schemes selected from the residual sample sets into the meteorological model for operation to obtain forecast results in the date period, and calculating second difference values between the forecast results and the observation data in the date period; replacing the selected parameter scheme from the sampled sample set with the selected parameter scheme from the remaining sample set if the first difference value and the second difference value satisfy a predetermined condition; and determining the sampling sample set after replacement as a parameter scheme set for the meteorological model.

The step of selecting a parameter scheme from the sampled sample set may comprise: each parameter scheme in the sampling sample set is brought into the meteorological model to carry out operation so as to obtain a forecast result in the date period; calculating the difference value between each forecast result and the observation data in the date period; and taking the parameter scheme with the largest difference value as the parameter scheme selected from the sampling sample set.

The step of selecting a parameter scheme from the sampled sample set may comprise: a parametric scheme is randomly sampled from a sample set of samples.

The predetermined conditions may include: the second difference value is smaller than a first threshold value of the forecast deviation of the set characterization model, and the difference between the second difference value and the first difference value is larger than a second threshold value of the differentiation of the characterization parameter scheme set by an experimenter.

And substituting each parameter scheme in the parameter scheme set for the meteorological model into the meteorological model for operation to obtain a forecast result in the date period, calculating a difference value between the forecast result of each parameter scheme and the observation data, and setting the weight of each parameter scheme in the parameter scheme set based on the difference value between the forecast result of each parameter scheme and the observation data.

Each parameter solution in the set of parameter solutions for the meteorological model is set to an equal weight.

The meteorological model may include at least one of a meteorological model, a mesoscale weather forecast mode, a mesoscale non-hydrostatic mode, a regional atmosphere forecast system, an atmospheric model with forecast simulation capability, a fluid model capable of forecasting, and a mathematical model capable of forecasting.

According to another aspect of the present disclosure, there is provided a wind farm wind speed forecasting method, including: generating a set of parametric solutions for the meteorological model by a parametric solution set generation method as described in any one of the above; forecasting the wind speed by using each parameter scheme in the parameter scheme set respectively to obtain a forecasting result of each parameter scheme; and obtaining a final wind speed forecasting result based on the forecasting result of each parameter scheme.

According to another aspect of the present disclosure, there is provided a parameter scheme set generating apparatus of a meteorological model, including: an initial selection module configured to select an initial parametric solution from all selectable parametric solutions of the meteorological model for the target area; a sample determination module configured to determine a parameter scheme sample capable of normal operation from the initial parameter scheme; and the set generation module is configured to set the meteorological model by using the set of parameter schemes sampled from the parameter scheme samples to calculate a forecast result, and screen the parameter scheme set for the meteorological model according to the difference between the forecast result and the observation data of the target area.

The sample determination module may be configured to: bringing an initial parameter scheme into and running the meteorological model; and determining a parameter scheme sample capable of normally operating according to the operation result.

The collection generation module may be configured to perform the following operations: sampling the parameter scheme samples to obtain a sampling sample set and a residual sample set; selecting a parameter scheme from the sampling sample set, substituting the parameter scheme selected from the sampling sample set into the meteorological model for operation to obtain a forecast result in the date period, and calculating a first difference value between the forecast result and observation data in the date period; selecting parameter schemes from the residual sample sets, substituting the parameter schemes selected from the residual sample sets into the meteorological model for operation to obtain forecast results in the date period, and calculating second difference values between the forecast results and the observation data in the date period; replacing the selected parameter scheme from the sampled sample set with the selected parameter scheme from the remaining sample set if the first difference value and the second difference value satisfy a predetermined condition; and determining the sampling sample set after replacement as a parameter scheme set for the meteorological model.

The collection generation module may be configured to: each parameter scheme in the sampling sample set is brought into the meteorological model to carry out operation so as to obtain a forecast result in the date period; calculating the difference value between each forecast result and the observation data in the date period; and taking the parameter scheme with the largest difference value as the parameter scheme selected from the sampling sample set.

The collection generation module may be configured to: a parametric scheme is randomly sampled from a sample set of samples.

The parameter scheme set generating means may further include: the weight setting module is configured to bring each parameter scheme in the parameter scheme set for the meteorological model into the meteorological model to perform operation so as to obtain a forecast result in the date period, calculate a difference value between the forecast result of each parameter scheme and the observation data, and set the weight occupied by each parameter scheme in the parameter scheme set based on the forecast result of each parameter scheme and the difference value of the observation data.

The parameter scheme set generating means may further include: a weight setting module configured to set each parameter scenario in the set of parameter scenarios for the meteorological model to an equal weight.

The meteorological model includes at least one of a meteorological model, a mesoscale weather forecast mode, a mesoscale non-hydrostatic mode, a regional atmosphere forecast system, an atmospheric mode with forecast simulation capability, a fluid mode capable of forecasting, and a mathematical mode capable of forecasting.

According to another aspect of the present disclosure, there is provided a wind farm wind speed forecasting device comprising: the parameter scheme set generating device of any one of the above, generating a parameter scheme set for a meteorological model; the forecasting module is configured to respectively forecast the wind speed by using each parameter scheme in the parameter scheme set to obtain a forecasting result of each parameter scheme; and the result output module is configured to obtain a final wind speed forecast result based on the forecast result of each parameter scheme.

According to another aspect of the present disclosure, there is provided a computer-readable storage medium storing a program configured to: the program comprises code for performing a method for parametric solution set generation of a meteorological model as described above or a method for wind farm wind speed forecasting as described above.

According to another aspect of the present disclosure, there is provided a computer comprising a readable medium storing a computer program configured to: the computer program comprises code for performing a method for parametric solution set generation of a meteorological model as described above or a method for wind farm wind speed forecasting as described above.

By the parameter scheme set generation method, more parameter schemes can be traversed step by step, so that the difference between the new and old parameter schemes is ensured to be large enough, the difference between the simulation performance and the actual measurement of the new and old parameter schemes is ensured to be small enough, the generated parameter scheme set is a parameter scheme library covering various weather conditions, and the accuracy and the stability of wind power plant wind speed forecasting by using the generated parameter schemes are ensured.

By the wind power plant wind speed forecasting method, the stability and the accuracy of the wind power plant wind speed forecasting result can be ensured, so that the generated energy and the estimated income of the wind power plant can be more accurately calculated in the process of developing the wind power plant, and the resource waste is avoided.

Drawings

The above and other objects and features of the embodiments of the present disclosure will become more apparent from the following description taken in conjunction with the accompanying drawings which illustrate, by way of example, the embodiments, wherein:

FIG. 1 is a flow chart of a method of generating a set of parameter solutions for a meteorological model according to an embodiment of the present disclosure;

FIG. 2 is a flow diagram illustrating the operation of screening a set of parameter solutions for a meteorological model by sampling parameter solution samples according to an embodiment of the present disclosure;

FIG. 3 is a flow chart diagram of a wind farm wind speed forecasting method according to an embodiment of the present disclosure;

FIG. 4 is a block diagram of a parametric approach set generation apparatus for a meteorological model according to an embodiment of the present disclosure;

FIG. 5 is a block diagram of a wind farm wind speed forecasting device according to an embodiment of the present disclosure.

Detailed Description

Reference will now be made in detail to the embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the like elements throughout. The embodiments are described below in order to explain the present invention by referring to the figures.

In the present disclosure, terms including ordinal numbers such as "first", "second", etc., may be used to describe various elements, but these elements should not be construed as being limited to only these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and vice-versa, without departing from the scope of the present disclosure.

Before setting forth the inventive concepts of the present disclosure, a related description is made of terms employed in the present disclosure.

A meteorological model: according to the actual conditions of the atmosphere, under the conditions of certain initial values and edge values, numerical calculation is carried out through a large computer, an equation set describing hydrodynamics and thermodynamics in the weather evolution process is solved, and a historical or method model for predicting the atmospheric motion state and the weather phenomenon in a certain time period is simulated. Meteorological elements considered by the meteorological model include temperature, barometric pressure, wind speed, wind direction, water vapor, and the like. The three-dimensional space within the pattern is divided into an ordered grid array, and the values of the meteorological variables at each grid point represent the current atmospheric conditions. The larger the number of grid points and the higher the mode resolution, the more the actual situation of the atmosphere can be accurately delineated, but the amount of computer computation will also be greatly increased.

WRF (Weather Research Forecast) mode: the weather forecast model is jointly developed by American scientific research institutions such as the American environmental prediction center, the American national atmospheric research center and the like. The WRF mode is a fully compressible and non-static mode and is written in the F90 language. The Arakawa C grid is adopted in the horizontal direction, the terrain following mass coordinate is adopted in the vertical direction, and a three-order or four-order Runge-Kutta algorithm is adopted in the aspect of time integration. The WRF mode can be used for individual case simulation of real weather, and can also be used as a theoretical basis for basic physical process discussion by using a module group contained in the WRF mode, and the WRF mode also has the capability of multiple nesting and convenient positioning in different geographic positions.

FIG. 1 is a flow chart of a method of generating a set of parameter solutions for a meteorological model according to an embodiment of the present disclosure.

In step 101, an initial parametric solution is selected from all selectable parametric solutions of the meteorological model for the target area.

According to the embodiment of the disclosure, the meteorological model for wind resource assessment of the wind farm is various, for example, any mode or model capable of acquiring wind farm elements of a target area, such as a meteorological mode, a mesoscale weather forecast mode (WRF), a mesoscale non-hydrostatic mode (MM5), a regional atmosphere simulation system (RAMS), an atmospheric mode with simulation capability (e.g., various climate modes, ocean coupling modes, etc.), a fluid mode capable of forecasting (e.g., WT, WindSim, OpenFOAM, Nektar + +, etc.), a mathematical model capable of forecasting (e.g., SVM, ANN, Decision Tree, etc.), and the like. An experimenter may select a meteorological model for wind resource assessment or wind power prediction of a target area based on topographical features, meteorological features, etc. of the target area or based on experience with developing wind farms in a region similar to the target area.

There are a number of options for selecting for a particular meteorological model. Since parametric solutions are solutions that parameterize physical processes such as radiation, boundary layers, micro-physics, etc., and the physical assumptions of different parametric solutions are different, it may happen that some parametric solutions are constrained against each other. Therefore, before generating the parameter scenario set of the meteorological model, it is first necessary to select a parameter scenario that does not cause mutual constraints from all the selectable parameter scenarios of the meteorological model. According to an embodiment of the present disclosure, the initial parametric approach may be selected from a total of selectable parametric approaches based on characteristics of the target area, wherein the characteristics of the target area include at least one of meteorological characteristics, topographical characteristics, and wind farm characteristics. In particular, a parametric scheme with a particular physical assumption may be selected from the total number of alternative parametric schemes based on the characteristics of the target region. Optionally, a parameter scheme already used by a topographic or topographic system or a weather system that is the same or similar to the target area may also be used as the initial parameter scheme.

In step 102, a parameter recipe sample that can function properly is determined from the initial parameter recipe.

According to the embodiments of the present disclosure, mode variables, such as spatial resolution, time step, nesting manner, etc., that conform to the target area may be set in the meteorological model of the target area. And (4) bringing the initial parameter scheme selected in the step (101) into and operating a meteorological model, and determining a parameter scheme sample capable of normally operating according to an operation result. For example, the initial parameter scheme may be brought into a meteorological model, the meteorological model may be run for a short time, and a sample of the parameter scheme that can run normally may be determined from the running results. For example, a parameter scenario capable of operating the meteorological model for 12 hours normally is labeled as 1, a parameter scenario incapable of operating the meteorological model for 12 hours normally is labeled as 0, and after each parameter scenario in the initial parameter scenario is brought into and operated on the meteorological model, the parameter scenarios all labeled as 1 are determined as parameter scenario samples. By counting the marks of the parameter schemes, the proportion P of the parameter schemes in the normally-operated parameter scheme samples in the initial parameter scheme can be counted, and the proportion can represent the proportion of the physical process parameterization schemes used in the target area. The parameter scheme sample capable of operating normally means the range of the parameter scheme sample which can be used for the target area.

In step 103, the meteorological model is set by using the set of parameter schemes sampled from the parameter scheme samples to calculate a forecast result, and the set of parameter schemes for the meteorological model is screened according to the difference between the forecast result and the observation data of the target area.

According to the embodiment of the present disclosure, the observation data of the target area is observation data within a date period capable of characterizing meteorological features selected from long-term observation data or reanalysis data of the target area according to meteorological elements and screening rules of the target area set by an experimenter. For example, observation data may be screened from observation data or reanalysis data for, e.g., 30 years for the target region.

According to embodiments of the present disclosure, an experimenter may first determine meteorological elements of interest within a target area, e.g., at least one of temperature, wind speed, wind direction, precipitation, etc., and then determine a screening principle, e.g., screening according to at least one of a mean, variance, extremum, etc., of the meteorological elements. For example, the meteorological element of interest to the experimenter in the target area may be wind speed, and the average value within a predetermined range may be determined as a screening rule, so that observation data D within a time period of a day in which a wind speed having an average value within a predetermined range occurs may be screened out in long-term (e.g., 30 years) observation data or reanalyzed data of the target area.

After the observation data in the date period in which the specific meteorological features appear is selected as described above, the parameter plan set meeting the forecast requirement can be screened according to the difference between the forecast result of the parameter plan sample and the observation data in a preset mode.

According to one embodiment of the present disclosure, a parameter scenario set for a meteorological model may be filtered by sampling parameter scenario samples and gradually updating the sampled parameter scenario samples.

One embodiment of obtaining a set of parameter solutions for a meteorological model by sampling parameter solution samples and screening according to a difference between a forecast result of the sampled parameter solution samples and observation data will be described in detail below with reference to fig. 2.

In general, sampling the parameter scheme samples to obtain a sampled sample set and a residual sample set;

selecting a parameter scheme from the sampling sample set, substituting the parameter scheme selected from the sampling sample set into the meteorological model for operation to obtain a forecast result in the date period, and calculating a first difference value between the forecast result and observation data in the date period;

selecting parameter schemes from the residual sample sets, substituting the parameter schemes selected from the residual sample sets into the meteorological model for operation to obtain forecast results in the date period, and calculating second difference values between the forecast results and the observation data in the date period;

replacing the selected parameter scheme from the sampled sample set with the selected parameter scheme from the remaining sample set if the first difference value and the second difference value satisfy a predetermined condition;

and determining the sampling sample set after replacement as a parameter scheme set for the meteorological model.

It should be understood that the sample screening approach described below is only one embodiment for implementing a screening parameter scheme set. Any other sampling method may be used by those skilled in the art to screen the parameter scheme set.

Referring to fig. 2, first, in step 201, a parameter scheme sample is sampled to obtain a sample set and a residual sample set.

According to an embodiment of the present disclosure, a predetermined number of parameter schemes may be extracted from parameter scheme samples as a sample set by a sampling method such as random sampling, whole group sampling, hierarchical sampling, system sampling, and the like, and the parameter schemes remaining in the parameter scheme samples may be used as a remaining sample set. According to the embodiment of the present disclosure, the number of parameter schemes in the sampling sample set may be determined according to the number of parameter schemes in the parameter scheme set ultimately desired by the experimenter, but the parameter schemes in the sampling sample set should occupy at least 5% -10% of the parameter scheme samples.

For example, the number of parameter schemes in the parameter scheme samples determined to be capable of operating normally is 1000 from the initial parameter schemes, and 100 parameter schemes are sampled from the parameter scheme samples to form a sampling sample set S, wherein the sampling sample set S comprises the parameter schemes S₁、S₂、…、S₁₀₀. The remaining parameter schemes in the parameter scheme samples other than the 100 parameter schemes in the sample set constitute a residual sample set R, wherein the residual sample set R includes the parameter schemes R₁、R₂、…、R₉₀₀。

In step 202, a parameter scheme is selected from the sampling sample set, the parameter scheme selected from the sampling sample set is brought into the meteorological model for operation to obtain a forecast result within a date period of the observation data, and a first difference value between the forecast result and the observation data within the date period is calculated.

According to the embodiment of the disclosure, one parameter scheme can be selected from sampling samples in a random sampling mode. Optionally, all parameter schemes in the sampling sample can be brought into the meteorological model for operation to obtain a forecast result in a date period where the observation data are located, and then the parameter schemes are selected according to the difference value between the forecast result and the observation data of each parameter scheme in the date period. The difference value may be, for example, one of a correlation coefficient, a root mean square error, a mean absolute error, or the like, or a combination of two or more of a correlation coefficient, a root mean square error, a mean absolute error, or the like. For example, the parameter scheme with the largest root mean square error between the forecast results and the observed data may be selected from the sampled sample set.

For example, the parameter scheme S may be selected from the sample set S by random sampling₂In this case, the parameter scheme S may be used₂Carry out operation by introducing into meteorological modelObtaining the forecast result P in the date period of the observation data D₂. Alternatively, the parameter scheme S in the sample set S may be sampled₁、S₂、…、S₁₀₀Respectively carrying the data into a meteorological model to carry out operation so as to obtain a forecast result P in the date period of the observation data D₁、P₂、…、P₁₀₀And calculating a difference value (e.g., root mean square error) I between each predictor and the observed data D₁、I₂、…、I₁₀₀. If the difference value is I₁、I₂、…、I₁₀₀Has a maximum value of I₄Selecting a parameter scheme S from the sample set S₄。

In step 203, a parameter scheme is selected from the remaining sample set and the remaining sample set is updated.

According to embodiments of the present disclosure, one parameter scheme may be selected from the remaining sample set by a random sampling method or an experimenter's active selection. After selecting one of the parameter schemes, the selected sample may be removed from the remaining sample set or marked as selected, thereby updating the remaining sample set. In case the selected sample is marked as selected, when subsequently selecting a parameter scheme from the remaining sample set, a selection will be made from the remaining parameter schemes in the remaining sample set, except for the parameter scheme marked as selected.

For example, a parameter scheme R may be selected from the remaining sample set R₁And removing the parameter solution R from the remaining sample set R₁The updated remaining samples include 899 parameter schemes, i.e., R₂、R₃、…、R₉₀₀. Alternatively, the parameter scheme R may be₁Marked as selected, when a parameter scheme is subsequently selected from the remaining sample set, R may be divided₁And selecting other parameter schemes.

In step 204, a parameter scheme selected from the remaining sample set is substituted into the meteorological model for operation to obtain a forecast result in the date period, and a second difference value between the forecast result and the observation data in the date period is calculated.

According to an embodiment of the present disclosure, the first and second disparity values may be characterized using correlation coefficients, root mean square errors, average absolute errors, and the like.

In step 205, it is determined whether the first difference value and the second difference value satisfy a predetermined condition, and if the first difference value and the second difference value satisfy the predetermined condition, one parameter scheme selected from the sample set is replaced with one selected from the remaining sample set, and if the first difference value and the second difference value do not satisfy the predetermined condition, the replacement is not performed.

According to an embodiment of the present disclosure, if the second difference value is smaller than a first threshold value set for the prediction deviation of the characterization model, and the difference between the second difference value and the first difference value is larger than a second threshold value set by an experimenter for the differentiation of the characterization parameter scheme, one parameter scheme selected from the sample set is replaced with one selected from the remaining sample set.

For example, a parameter scheme S may be selected from a sample set of samples₂And calculating a parameter scheme S₂Forecast result P of₂A first difference value (e.g., correlation coefficient) I from the observation data D₂. The parameter scheme R may be selected from the remaining sample set₁And calculate R₁Forecast result P of₁' second difference value (e.g., correlation coefficient) I from observation data D₁'. If I₁’<A first threshold value α, and l'₁-I₂>Second threshold β, parameter scheme R in the remaining sample set is used₁Replacing S in a sample set₂。

At step 206, it is determined the number of times that no replacement has been performed in succession. If the number of times of continuous non-replacement is less than the predetermined number of times, the operation is returned to step 202, that is, the steps 202 to 206 are repeated, and if the number of times of continuous non-replacement is greater than or equal to the predetermined number of times, the sample set after replacement is determined as the parameter scheme set for the meteorological model in step 207.

That is, by setting the predetermined number of times, it is possible to continue to replace the parameter scheme in the sample set with the parameter scheme in the remaining sample set until the number of times that the replacement is not continuously performed reaches the predetermined number of times. For example, the experimenter may set the predetermined number of times to 10 times, and when the first difference value between the parameter scheme selected from the sample set and the observation data and the second difference value between the parameter scheme selected from the remaining sample set and the observation data do not satisfy the predetermined condition for 10 consecutive times, that is, the parameter scheme selected from the sample set is not replaced with the parameter scheme selected from the remaining sample set for 10 consecutive times, the parameter scheme selected from the sample set is not continuously selected, but the sample set after the last replacement is used as the parameter scheme set for the meteorological model.

According to an embodiment of the present disclosure, when a set of parameter solutions for a meteorological model is derived, a weight may be set for each parameter solution in the set of parameter solutions. For example, each parameter scheme may be brought into the meteorological model to perform an operation to obtain a forecast result within a date period in which the observation data is located, a difference value between the forecast result and the observation data of each parameter scheme is calculated, and a weight occupied by each parameter scheme in the parameter scheme set is set according to the difference value between the forecast result and the observation data of each parameter scheme. For example, the weight of each parameter scheme in the parameter scheme set may be set according to a combination of the prediction result of each parameter scheme and one or at least one of the correlation coefficient, the root mean square error, and the average absolute value error of the observation data. For example, the parameter scheme set includes three parameter schemes S₁、S₂And S₃The difference (e.g., correlation coefficient) between the forecast result and the observed data of each parameter scheme is I₁、I₂And I₃If I is₁、I₂And I₃A ratio of 1:2:3 with respect to each other, S may be determined based on the ratio₁Is set to 1/6, S is set₂Is set to 1/3, S₃Is set to 1/2. Alternatively, each parameter scheme in the parameter scheme set may also be directly set to be equally weighted.

FIG. 3 is a flow chart of a wind farm wind speed forecasting method according to an embodiment of the present disclosure.

In step 301, a set of parameter solutions for the meteorological model is generated by the above-described method for generating a set of parameter solutions for the meteorological model.

According to the embodiment of the present disclosure, an initial parameter scheme may be selected from all selectable parameter schemes of the meteorological model of the target area where the wind speed forecast is performed, and a parameter scheme set may be generated from the initial parameter scheme, and since the process of generating the parameter scheme set is described in detail with reference to fig. 1 and 2, the detailed description will not be provided herein. Alternatively, if a forecast result is expected to be obtained quickly, a parameter scheme set of a certain area similar to the physical characteristics of the target area can be selected from a sample library storing parameter scheme sets as an initial parameter scheme by performing similar analysis on the physical characteristics (e.g., terrain characteristics, weather processes, etc.) of the target area, and then the parameter scheme set is generated from the initial parameter scheme by the parameter scheme generation method.

In step 302, wind speed forecasting is performed by using each parameter scheme in the parameter scheme set respectively, and a forecasting result of each parameter scheme is obtained.

According to the embodiment of the disclosure, each parameter scheme in the parameter scheme set can be brought into the meteorological model of the target area for wind speed forecasting, so that the forecasting result of each parameter scheme is obtained.

In step 303, a final wind speed forecast is obtained based on the forecast results for each parameter scenario.

According to the embodiment of the disclosure, the weight of each parameter scheme in the parameter scheme set can be set according to the difference value between the forecast result and the observation data of each parameter scheme. Alternatively, each parameter scheme may also be directly set to be equally weighted. For example, if the difference between the prediction results of each parameter solution is not large, each parameter solution may be directly set to an equal weight. And carrying out weighted average on the forecast result of each parameter scheme according to the weight set for each parameter scheme in the parameter scheme set so as to obtain a final wind speed forecast result.

FIG. 4 is a block diagram of a parametric approach set generation apparatus for a meteorological model according to an embodiment of the present disclosure.

The parameter solution set generating apparatus 400 includes an initial selection module 401, a sample determination module 402, and a set generating module 403.

The initial selection module 401 may select an initial parametric solution from all selectable parametric solutions of the meteorological model for the target area.

There are a number of options for selecting for a particular meteorological model. Since parametric solutions are solutions that parameterize physical processes such as radiation, boundary layers, micro-physics, etc., and the physical assumptions of different parametric solutions are different, it may happen that some parametric solutions are constrained against each other. Therefore, before generating the parameter scenario set of the meteorological model, it is first necessary to select a parameter scenario that does not cause mutual constraints from all the selectable parameter scenarios of the meteorological model. According to an embodiment of the present disclosure, the initial selection module 401 may select an initial parametric plan from all selectable parametric plans according to characteristics of the target area, wherein the characteristics of the target area include at least one of meteorological characteristics, topographic characteristics, and wind farm characteristics. Specifically, the initial selection module 401 may select a parameter scheme having a specific physical hypothesis from all the selectable parameter schemes according to the characteristics of the target region. Optionally, the initial selection module 401 may also use as the initial parameter scheme a parameter scheme already used by the same or similar terrain and terrain or weather system as the target area.

The sample determination module 402 may determine parameter recipe samples from the initial parameter recipes that are capable of functioning properly.

According to an embodiment of the present disclosure, the sample determination module 402 may set a mode variable, e.g., spatial resolution, time step size, nesting manner, etc., conforming to the target area in the meteorological model of the target area. The sample determination module 402 may bring the initial parameter scheme selected by the initial selection module 401 into and operate the meteorological model, and determine a parameter scheme sample capable of operating normally according to an operation result. For example, the initial parameter scheme may be brought into a meteorological model, the meteorological model may be run for a short time, and a sample of the parameter scheme that can run normally may be determined from the running results. For example, a parameter scenario capable of operating the meteorological model for 12 hours normally is labeled as 1, a parameter scenario incapable of operating the meteorological model for 12 hours normally is labeled as 0, and after each parameter scenario in the initial parameter scenario is brought into and operated on the meteorological model, the parameter scenarios all labeled as 1 are determined as parameter scenario samples. By counting the marks of the parameter schemes, the proportion P of the parameter schemes in the normally-operated parameter scheme samples in the initial parameter scheme can be counted, and the proportion can represent the proportion of the physical process parameterization schemes used in the target area. The parameter scheme sample capable of operating normally means the range of the parameter scheme sample which can be used for the target area.

The set generating module 403 may set the meteorological model by using a set of parameter schemes sampled from a sample of parameter schemes to operate to obtain a forecast result, and filter a set of parameter schemes for the meteorological model according to a difference between the forecast result and observation data of a target area.

After selecting the observation data within the date period in which the specific meteorological features appear as described above, the set generation module 403 may screen the parameter scenario set meeting the forecast requirement according to the difference between the forecast result of the parameter scenario sample and the observation data in a predetermined manner.

According to an embodiment of the present disclosure, the set generation module 403 may obtain the parameter scheme set for the meteorological model by sampling the parameter scheme samples and gradually updating the sampled parameter scheme samples.

In general, the set generating module 403 is configured to sample the parameter scheme samples, to obtain a sample set and a residual sample set;

One embodiment of obtaining a set of parameter solutions for a meteorological model by sampling parameter solution samples and screening according to a difference between a forecast result and observation data of the sampled parameter solution samples will be described in detail below. It should be understood that the sample screening approach described below is only one embodiment for implementing a screening parameter scheme set. Any other sampling method may be used by those skilled in the art to screen the parameter scheme set.

First, the set generating module 403 samples the parameter scheme samples to obtain a sample set and a residual sample set.

According to an embodiment of the present disclosure, the set generating module 403 may extract a predetermined number of parameter schemes from the parameter scheme samples as a sample set by a sampling method such as random sampling, whole group sampling, hierarchical sampling, system sampling, and the like, and may take the remaining parameter schemes in the parameter scheme samples as a remaining sample set. According to the embodiment of the present disclosure, the number of parameter schemes in the sampling sample set may be determined according to the number of parameter schemes in the parameter scheme set ultimately desired by the experimenter, but the parameter schemes in the sampling sample set should occupy at least 5% -10% of the parameter scheme samples.

The set generating module 403 may select a parameter scheme from the sampling sample set, bring the parameter scheme selected from the sampling sample set into the meteorological model for operation to obtain a forecast result within a date period of the observation data, and calculate a first difference value between the forecast result and the observation data within the date period.

According to an embodiment of the present disclosure, the set generation module 403 may select a parameter scheme from the sampled samples by means of random sampling. Optionally, the set generating module 403 may also bring all parameter schemes in the sampling sample into the meteorological model for operation to obtain the forecast result within the date period of the observation data, and then select a parameter scheme according to the difference between the forecast result and the observation data of each parameter scheme within the date period. The difference value may be, for example, one of a correlation coefficient, a root mean square error, a mean absolute error, or the like, or a combination of two or more of a correlation coefficient, a root mean square error, a mean absolute error, or the like. For example, the set generating module 403 may select a parameter scheme from the sampled sample set in which the root mean square error between the forecast results and the observed data is the largest.

For example, the set generating module 403 may select the parameter scheme S from the sampling sample set S by random sampling₂In this case, the set generation module 403 may apply the parameter scheme S₂Bringing into a meteorological model to perform an operation to obtain forecasts within a date period of the observation data DResults P₂. Alternatively, the set generating module 403 may select the parameter scheme S in the sample set S₁、S₂、…、S₁₀₀Respectively carrying the data into a meteorological model to carry out operation so as to obtain a forecast result P in the date period of the observation data D₁、P₂、…、P₁₀₀And calculating a difference value (e.g., root mean square error) I between each predictor and the observed data D₁、I₂、…、I₁₀₀. If the difference value is I₁、I₂、…、I₁₀₀Has a maximum value of I₄The set generating module 403 selects a parameter scheme S from the sample set S₄. The set generation module 403 may select one parameter scheme from the remaining sample set and update the remaining sample set.

According to an embodiment of the present disclosure, the set generation module 403 may select one parameter scheme from the remaining sample set by a random sampling method or an active selection by an experimenter. Upon selecting one of the parameter solutions, the set generation module 403 may update the remaining sample set by removing the selected sample from the remaining sample set or marking the selected sample as having been selected. In case the selected sample is marked as selected, when subsequently selecting a parameter scheme from the remaining sample set, a selection will be made from the remaining parameter schemes in the remaining sample set, except for the parameter scheme marked as selected.

For example, the set generation module 403 may select the parameter scheme R from the remaining sample set R₁And removing the parameter solution R from the remaining sample set R₁The updated remaining samples include 899 parameter schemes, i.e., R₂、R₃、…、R₉₀₀. Optionally, the set generation module 403 may apply the parameter scheme R₁Marked as selected, when a parameter scheme is subsequently selected from the remaining sample set, R may be divided₁And selecting other parameter schemes.

The set generating module 403 may bring one parameter scheme selected from the remaining sample sets into the meteorological model for operation to obtain a forecast result in the date period, and calculate a second difference value between the forecast result and the observation data in the date period.

The set generating module 403 may determine whether the first difference value and the second difference value satisfy a predetermined condition, replace one parameter scheme selected from the sampled sample set with one parameter scheme selected from the remaining sample set if the first difference value and the second difference value satisfy the predetermined condition, and not perform the replacement if the first difference value and the second difference value do not satisfy the predetermined condition.

According to an embodiment of the present disclosure, if the second difference value is smaller than a first threshold value of the set prediction deviation of the characterization model, and the difference between the second difference value and the first difference value is larger than a second threshold value of the differentiation of the characterization parameter scheme set by the experimenter, the set generating module 403 replaces one parameter scheme selected from the sampled sample set with one parameter scheme selected from the remaining sample set.

For example, the set generation module 403 may select a parameter scheme S from the sample set of samples₂And calculating a parameter scheme S₂Forecast result P of₂A first difference value (e.g., correlation coefficient) I from the observation data D₂. The set generation module 403 may select the parameter scheme R from the remaining sample set₁And calculate R₁Forecast result P of₁' second difference value (e.g., correlation coefficient) I from observation data D₁'. If I₁’<A first threshold value α, and l'₁-I₂>Second threshold β, the set generation module 403 uses the parameter scheme R in the remaining sample set₁Replacing S in a sample set₂。

The collection generation module 403 may determine the number of times that no replacement has been made consecutively. If the number of times of continuous non-replacement is less than the predetermined number of times, the operations of selecting the parameter scheme from the sampling sample set, selecting the parameter scheme from the remaining sample set, and determining whether to perform replacement according to the difference value between the selected parameter scheme and the observation data are repeated, and if the number of times of continuous non-replacement is greater than or equal to the predetermined number of times, the set generation module 403 may determine the sampling sample set after replacement as the parameter scheme set for the meteorological model.

According to an embodiment of the present disclosure, the parameter solution combination generation module further includes a weight setting module 404 (not shown), and when a set of parameter solutions for the meteorological model is obtained, the weight setting module 404 may set a weight for each parameter solution in the set of parameter solutions. For example, the weight setting module 404 may bring each parameter scheme into the meteorological model for operation to obtain a forecast result in a date period of the observation data, calculate a difference value between the forecast result and the observation data of each parameter scheme, and set the weight occupied by each parameter scheme in the parameter scheme set according to the difference value between the forecast result and the observation data of each parameter scheme. For example, the weight setting module 404 may set the weight of each parameter scheme in the parameter scheme set according to one of the prediction result of each parameter scheme and the observed data, or a combination of two or more of the correlation coefficient, the root mean square error, and the average absolute value error. For example, the parameter scheme set includes three parameter schemes S₁、S₂And S₃The difference between the forecast and observed data for each parameter scheme (e.g.E.g., correlation coefficient) are each I₁、I₂And I₃If I is₁、I₂And I₃A ratio of 1:2:3, the weight setting module 404 may set S based on the ratio₁Is set to 1/6, S is set₂Is set to 1/3, S₃Is set to 1/2. Alternatively, the weight setting module 404 may also directly set each parameter solution in the parameter solution set to be equal weight.

Through the parameter scheme set generation device, more parameter schemes can be traversed step by step, so that the difference between the new and old parameter schemes is ensured to be large enough, the difference between the simulation performance and the actual measurement of the new and old parameter schemes is ensured to be small enough, the generated parameter scheme set is a parameter scheme library covering various weather conditions, and the accuracy and the stability of wind power plant wind speed forecasting by using the generated parameter schemes are ensured.

Wind farm wind speed forecasting device 500 may include a parameter solution set generation device 400, a forecasting module 501, and a result output module 502.

As described above, the parameter scenario set generation apparatus 400 generates a set of parameter scenarios for the meteorological model.

The forecasting module 501 may respectively use each parameter scheme in the parameter scheme set to perform wind speed forecasting, so as to obtain a forecasting result of each parameter scheme.

According to the embodiment of the present disclosure, the forecasting module 501 may bring each parameter scheme in the parameter scheme set into the meteorological model of the target area for wind speed forecasting, so as to obtain a forecasting result of each parameter scheme.

The result output module 502 may obtain a final wind speed forecast result based on the forecast result of each parameter scenario.

According to the embodiment of the present disclosure, the result output module 502 may set the weight of each parameter scheme in the parameter scheme set according to the difference between the forecast result and the observation data of each parameter scheme. Alternatively, the result output module 502 may also directly set each parameter scheme to be equally weighted. For example, if the difference between the forecast results for each parameter solution is not large, the result output module 502 may directly set each parameter solution to an equal weight. The result output module 502 performs weighted average on the forecast result of each parameter solution according to the weight set for each parameter solution in the parameter solution set, so as to obtain the final wind speed forecast result.

By the wind power plant wind speed forecasting device, the stability and the accuracy of the wind power plant wind speed forecasting result can be ensured, so that the generated energy and the estimated income of the wind power plant can be more accurately calculated in the process of developing the wind power plant, and the resource waste is avoided.

There is also provided, in accordance with an embodiment of the present disclosure, a computer-readable storage medium storing a computer program. The computer readable storage medium stores a computer program which, when executed by a processor, causes the processor to perform a method for generating a set of parameter solutions for a meteorological model or a method for wind speed forecasting for a wind farm as described above. The computer readable storage medium is any data storage device that can store data which can be read by a computer system. Examples of computer-readable storage media include: read-only memory, random access memory, read-only optical disks, magnetic tapes, floppy disks, optical data storage devices, and carrier waves (such as data transmission through the internet via wired or wireless transmission paths).

There is also provided, in accordance with an embodiment of the present disclosure, a computing device. The computing device includes a processor and a memory. The memory is for storing a computer program. The computer program is executed by a processor causing the processor to execute a method for generating a set of parametric solutions for a meteorological model or a method for wind speed forecasting at a wind farm as described above.

Although a few exemplary embodiments of the present invention have been shown and described, it would be appreciated by those skilled in the art that changes may be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the claims and their equivalents.

Claims

1. A method for generating a parameter scheme set of a meteorological model comprises the following steps:

selecting an initial parameter scheme from all selectable parameter schemes of the meteorological model of the target area;

determining parameter scheme samples capable of normally operating from the initial parameter scheme;

setting the meteorological model to operate on a forecast result by using a set of parameter solutions sampled from the parameter solution samples;

and screening to obtain a parameter scheme set for the meteorological model according to the difference between the forecast result and the observation data of the target area.

2. The method of generating a set of parametric solutions according to claim 1, wherein the initial parametric solution is a parametric solution selected from the total number of selectable parametric solutions according to a characteristic of a target area,

wherein the characteristics of the target area include at least one of meteorological characteristics, topographical characteristics, and wind farm characteristics.

3. The method of generating a set of parametric solutions according to claim 1, wherein the step of determining the parametric solution samples that can function properly from the initial parametric solution comprises:

bringing an initial parameter scheme into and running the meteorological model;

and determining a parameter scheme sample capable of normally operating according to the operation result.

4. The parameter scenario set generation method of claim 1, wherein the observation data is observation data in a time period of a day that is capable of characterizing meteorological features selected from long-term observation data or reanalysis data of the target area according to meteorological elements and filtering rules of the target area set by an experimenter,

wherein the observation data includes at least one of anemometer tower data, radar data, and meteorological station data.

5. The method of claim 4, wherein the step of screening the set of parameter solutions for the meteorological model based on the difference between the forecast results and the observed data of the target area comprises:

sampling the parameter scheme samples to obtain a sampling sample set and a residual sample set;

6. The method of generating a set of parametric schemes of claim 5, wherein the step of selecting a parametric scheme from the sampled sample set comprises:

each parameter scheme in the sampling sample set is brought into the meteorological model to carry out operation so as to obtain a forecast result in the date period;

calculating the difference value between each forecast result and the observation data in the date period;

and taking the parameter scheme with the largest difference value as the parameter scheme selected from the sampling sample set.

7. The method of generating a set of parametric schemes of claim 5, wherein the step of selecting a parametric scheme from the sampled sample set comprises:

a parametric scheme is randomly sampled from a sample set of samples.

8. The method of generating a set of parameter solutions according to claim 5, wherein the predetermined condition comprises: the second difference value is smaller than a first threshold value of the forecast deviation of the set characterization model, and the difference between the second difference value and the first difference value is larger than a second threshold value of the differentiation of the characterization parameter scheme set by an experimenter.

9. The method as claimed in claim 4, wherein each parameter solution in the parameter solution set for the meteorological model is substituted into the meteorological model to perform operation to obtain the forecast result in the date period, the difference between the forecast result and the observation data of each parameter solution is calculated, and the weight of each parameter solution in the parameter solution set is set based on the difference between the forecast result and the observation data of each parameter solution.

10. The method of generating a set of parameter solutions of claim 1 wherein each parameter solution in the set of parameter solutions for the meteorological model is set to equal weight.

11. The parametric approach set generation method of claim 1, wherein the meteorological model comprises at least one of a meteorological model, a mesoscale weather forecast mode, a mesoscale non-hydrostatic mode, a regional atmosphere forecast system, an atmospheric model with forecast simulation capability, a fluid model capable of forecasting, and a mathematical model capable of forecasting.

12. A wind power plant wind speed forecasting method comprises the following steps:

generating a set of parametric solutions for a meteorological model by a parametric solution set generation method according to any one of claims 1 to 11;

forecasting the wind speed by using each parameter scheme in the parameter scheme set respectively to obtain a forecasting result of each parameter scheme;

and obtaining a final wind speed forecasting result based on the forecasting result of each parameter scheme.

13. A parametric solution set generation apparatus for a meteorological model, comprising:

an initial selection module configured to select an initial parametric solution from all selectable parametric solutions of the meteorological model for the target area;

a sample determination module configured to determine a parameter scheme sample capable of normal operation from the initial parameter scheme;

and the set generation module is configured to set the meteorological model by using the set of parameter schemes sampled from the parameter scheme samples to calculate a forecast result, and screen the parameter scheme set for the meteorological model according to the difference between the forecast result and the observation data of the target area.

14. The apparatus according to claim 13, wherein the initial parameter scheme is a parameter scheme selected from the total selectable parameter schemes according to the characteristics of the target area,

15. The apparatus according to claim 13, wherein the sample determination module is configured to:

bringing an initial parameter scheme into and running the meteorological model;

16. The parameter scenario set generation apparatus of claim 13, wherein the observation data is observation data in a time period of day that is capable of characterizing meteorological features selected from long-term observation data or reanalysis data of the target area according to meteorological elements and filtering rules of the target area set by an experimenter,

17. The apparatus according to claim 16, wherein the set generation module is configured to perform the following operations:

18. The parameter solution set generating apparatus of claim 17, wherein the set generating module is configured to:

19. The parameter solution set generating apparatus of claim 17, wherein the set generating module is configured to:

a parametric scheme is randomly sampled from a sample set of samples.

20. The apparatus for generating a set of parameter solutions according to claim 17, wherein the predetermined condition comprises: the second difference value is smaller than a first threshold value of the forecast deviation of the set characterization model, and the difference between the second difference value and the first difference value is larger than a second threshold value of the differentiation of the characterization parameter scheme set by an experimenter.

21. The set of parameter solutions generating apparatus of claim 16, further comprising: the weight setting module is configured to bring each parameter scheme in the parameter scheme set for the meteorological model into the meteorological model to perform operation so as to obtain a forecast result in the date period, calculate a difference value between the forecast result of each parameter scheme and the observation data, and set the weight occupied by each parameter scheme in the parameter scheme set based on the forecast result of each parameter scheme and the difference value of the observation data.

22. The parameter solution set generating apparatus of claim 13, further comprising: a weight setting module configured to set each parameter scenario in the set of parameter scenarios for the meteorological model to an equal weight.

23. The parametric solution set generating device of claim 13, wherein the meteorological model comprises at least one of a meteorological model, a mesoscale weather forecast mode, a mesoscale non-hydrostatic mode, a regional atmosphere forecast system, an atmospheric model with forecast simulation capability, a fluid model capable of forecasting, and a mathematical model capable of forecasting.

24. A wind farm wind speed forecasting device comprising:

the set of parametric solutions generating device of any one of claims 13-23, generating a set of parametric solutions for a meteorological model;

the forecasting module is configured to respectively forecast the wind speed by using each parameter scheme in the parameter scheme set to obtain a forecasting result of each parameter scheme;

and the result output module is configured to obtain a final wind speed forecast result based on the forecast result of each parameter scheme.

25. A computer-readable storage medium storing a program configured to: the program comprises code for performing a method for generating a set of parametric solutions for a meteorological model according to any one of claims 1-11 or a method for wind speed forecasting for a wind farm according to claim 12.

26. A computer comprising a readable medium storing a computer program configured to: the computer program comprises code for performing a method for generating a set of parametric solutions for a meteorological model according to any one of claims 1-11 or a method for wind speed forecasting for a wind farm according to claim 12.