Wind speed prediction method and device

Abstract

The present invention provides a wind speed prediction method and device, wherein the wind speed prediction method comprises: obtaining meteorological simulation data of a target area under a predetermined meteorological mode; dividing the grid points of the target area into different groups according to the wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data, wherein the grid points in the same group have the same or similar wind speed variation characteristics; determining the areas where the divided groups are distributed; and predicting the actual wind speed at another point other than the point in the area where each group is distributed based on the actual wind speed observed at any point in the area where each group is distributed. The wind speed prediction method and device of the present invention can not only accurately evaluate wind resources when wind measurement is insufficient or even absent, but also effectively reduce the uncertainty of calculation of wind farm power generation.

Description

Technical Field

[0001] The present application relates to the technical field of wind power generation, and in particular to a method and device for predicting wind speed. Background Art

[0002] The accuracy of wind farm power generation calculation is a key issue that the wind power industry has been committed to solving, and the accuracy of wind farm power generation calculation depends on the accuracy of wind measurement data. However, with the shortening of wind farm investment and construction period and the lack of awareness of the importance of wind measurement data, if wind power projects start wind farm wind resource assessment without sufficient wind measurement or without wind measurement at all, it will bring great uncertainty to the calculation of wind farm power generation and the estimation of its corresponding benefits.

[0003] Therefore, there is an urgent need for a wind speed prediction method and device that can solve the above problems, so as to provide more accurate and reliable data basis for the construction of wind farms and the evaluation of their power generation. Summary of the invention

[0004] The purpose of the present invention is to provide a wind speed prediction method and device.

[0005] According to one aspect of the present invention, a wind speed prediction method is provided, the method comprising: acquiring meteorological simulation data of a target area under a predetermined meteorological mode; dividing the grid points of the target area into different groups according to wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data, wherein the grid points in the same group have the same or similar wind speed variation characteristics; determining the areas where the divided groups are distributed; and predicting the actual wind speed at another point other than the point in the area where each group is distributed based on the actual wind speed observed at any point in the area where each group is distributed.

[0006] Preferably, the step of determining the distribution areas of the divided groups includes: determining the distribution areas of the divided groups based on areas surrounded by grid points in the divided groups.

[0007] Preferably, the step of predicting the actual wind speed at another point other than the point in the area of ​​each group distribution includes: determining the deviation of the actual wind speed observed at any point in the area of ​​each group distribution relative to the simulated wind speed at the point derived from the meteorological simulation data as the deviation of the actual wind speed in the area of ​​each group distribution; and predicting the actual wind speed at another point other than the point in the area of ​​each group distribution based on the deviation of the actual wind speed in the area of ​​each group distribution and the simulated wind speed at another point other than the point in the area of ​​each group distribution derived from the meteorological simulation data.

[0008] Preferably, the predetermined meteorological pattern is a mesoscale meteorological pattern.

[0009] Preferably, the wind speed variation characteristic is a characteristic of wind speed variation over time.

[0010] According to another aspect of the present invention, a wind speed prediction device is provided, the device comprising: a meteorological acquisition unit, which acquires meteorological simulation data of a target area under a predetermined meteorological mode; a group division unit, which divides the grid points of the target area into different groups according to wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data, wherein the grid points in the same group have the same or similar wind speed variation characteristics; an area determination unit, which determines the areas in which the divided groups are distributed; and a wind speed prediction unit, which predicts the actual wind speed at another point other than the point in the area in which each group is distributed, based on the actual wind speed observed at any point in the area in which each group is distributed.

[0011] Preferably, the area determination unit determines the areas where each divided group is distributed based on areas surrounded by grid points in each divided group.

[0012] Preferably, the wind speed prediction unit includes: a deviation determination subunit, which determines the deviation of the actual wind speed observed at any point in the area of ​​each group distribution relative to the simulated wind speed at the point derived from the meteorological simulation data as the deviation of the actual wind speed in the area of ​​each group distribution; and a wind speed prediction subunit, which predicts the actual wind speed at another point in the area of ​​each group distribution other than the point based on the deviation of the actual wind speed in the area of ​​each group distribution and the simulated wind speed at another point in the area of ​​each group distribution other than the point derived from the meteorological simulation data.

[0013] Preferably, the predetermined meteorological pattern is a mesoscale meteorological pattern.

[0014] Preferably, the wind speed variation characteristic is a characteristic of wind speed variation over time.

[0015] According to another aspect of the present invention, a computer-readable storage medium storing a computer program is provided. When the computer program is executed by a processor, the wind speed prediction method as described above is implemented.

[0016] According to another aspect of the present invention, a computer device is provided, the computer device comprising: a processor; a memory storing a computer program, and when the computer program is executed by the processor, the wind speed prediction method as described above is implemented.

[0017] The wind speed prediction method and device provided by the present invention can not only accurately evaluate wind resources when wind measurement is insufficient or even without wind measurement, but also effectively reduce the uncertainty of calculation of wind farm power generation, thereby providing more accurate and reliable data basis for site selection and construction of wind farms and evaluation of their power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other objects and features of the present invention will become more apparent through the following description in conjunction with the accompanying drawings, in which:

[0019] Figure 1 is a flow chart illustrating a wind speed prediction method according to an exemplary embodiment of the present invention;

[0020] Figure 2 is a block diagram showing the structure of a wind speed prediction device according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0021] The general concept of the present invention is that, in the case of insufficient or even no wind measurement, meteorological simulation data under a predetermined meteorological mode is used as wind measurement data, and the actual wind speed at each location point in the wind farm is evaluated by clustering analysis on the wind speed variation characteristics of the meteorological simulation data and converting the clustering into clustering of spatial regional distribution, thereby realizing the evaluation of the power generation of the wind farm. In addition, the present invention also takes into account the uncertainty caused by the deviation of such meteorological simulation data from the actual wind speed in the calculation and evaluation of the power generation of the wind farm, and corrects the actual wind speed when evaluating the actual wind speed, so as to further ensure the accuracy of the calculation of the power generation of the wind farm.

[0022] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0023] Figure 1 is a flowchart illustrating a wind speed prediction method according to an exemplary embodiment of the present invention.

[0024] In step 110, weather simulation data of the target area under a predetermined weather mode may be obtained.

[0025] In specific implementation, various meteorological modes can be selected for the target area to carry out numerical simulation of wind speed within the required time period to obtain the simulated wind speed at each grid point in the target area, thereby preparing for the subsequent cluster analysis.

[0026] Considering that the mesoscale meteorological model can provide such meteorological simulation data, the meteorological simulation data is the time series data of the simulated wind speed at each point in the target area within the required time period, therefore, in an illustrative embodiment, the meteorological simulation data of the target area under the mesoscale meteorological model can be obtained, so that the obtained meteorological simulation data can better reflect the changes in the flow field and weather conditions in the target area, and is also closer to the actual situation of the wind speed change in the target area. The mesoscale meteorological model can be selected from one of the following modes: Mesoscale Weather Research and Forecasting Model (WRF), Mesoscale Non-hydrostatic Model (MM5), Regional Atmospheric Modeling System (RAMS) and other modes. In addition, the mesoscale meteorological model can also select other various climate models, ocean models, sea-air coupling models, etc. with simulation capabilities.

[0027] In step 120, the grid points of the target area may be divided into different groups according to the wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data, wherein the grid points in the same group have the same or similar wind speed variation characteristics.

[0028] The wind speed variation characteristics in step 120 depend on the meteorological simulation data derived under the meteorological mode. For example, in the mesoscale meteorological mode, the wind speed variation characteristics may be characteristics of wind speed variation over time. This is because the meteorological simulation data obtained under this mode (i.e., time series data of simulated wind speed) can better reflect changes in flow fields and weather conditions in the target area.

[0029] In an illustrative embodiment, each grid point of the target area can be divided into different groups by clustering analysis of wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data. The clustering algorithm used may include but is not limited to Timeseires KMeans, kernel KMeans, Kshape and other clustering algorithms. For example, if the target area is divided into 90×90 grids under a predetermined meteorological mode, the time series data of the simulated wind speed at each corresponding grid point is extracted from the meteorological simulation data for clustering analysis. In specific implementation, the sum of the distances from the cluster center to each time series vector can be used as a clustering index, and the time series data can be divided into n different groups based on the clustering index.

[0030] In step 130 , the areas where the divided groups are distributed may be determined.

[0031] In an illustrative embodiment, the areas where each divided group is distributed can be determined based on the areas surrounded by the grid points in each divided group. In addition, in order to facilitate the distinction between the areas where different groups are distributed, different colors can be used to mark the areas where different groups are distributed in Google Earth (a virtual globe software developed by Google) (areas with the same color represent the same group / category), thereby drawing a spatial regional clustering effect diagram of the wind speed distribution status of the target area. In this way, the clustering according to wind speed change characteristics can be converted into spatial regional clustering of wind speed distribution conditions, and the time series data of the simulated wind speed at each grid point included in the area where each group is distributed has the same / similar wind speed change characteristics. Based on this, the wind speed change characteristics at each position in the target area can be spatially classified.

[0032] It should be understood that the above method of determining the group distribution area is only exemplary, and the present invention is not limited thereto. Any other suitable method that helps to determine the group distribution area can be applied to the present invention.

[0033] In step 140 , the actual wind speed at another point other than the point in the area where each group is distributed may be predicted based on the actual wind speed observed at any point in the area where each group is distributed.

[0034] In an optional embodiment, the deviation of the actual wind speed observed at any point in the area of ​​each group distribution relative to the simulated wind speed at the point derived from the meteorological simulation data can be determined as the deviation of the actual wind speed in the area of ​​each group distribution; then, the actual wind speed at another point other than the point in the area of ​​each group distribution is predicted based on the deviation of the actual wind speed in the area of ​​each group distribution and the simulated wind speed at another point other than the point in the area of ​​each group distribution derived from the meteorological simulation data. For example, when the deviation is positive, the result obtained by adding the simulated wind speed at the other point to the deviation is determined as the simulated wind speed at the other point; when the deviation is negative, the result obtained by subtracting the deviation from the simulated wind speed at the other point is determined as the simulated wind speed at the other point. In this way, not only the correction of the predicted wind speed can be achieved, but also the influence of the deviation between the meteorological simulation data and the actual wind speed on the evaluation of the power generation of the wind farm can be minimized.

[0035] In a further embodiment, all data points observed during the same period in the target area can also be marked on the above-mentioned clustered regional distribution map to determine the available observation points in the target area, and then determine the distribution area with the same / similar wind speed change characteristics as these observation data points, so as to provide more data basis for predicting the actual wind speed at each location point in the target area.

[0036] It should be understood that the above-mentioned method of predicting the actual wind speed is only exemplary, and the present invention is not limited thereto. Any other suitable method that helps to improve the accuracy of wind speed prediction can be applied to the present invention.

[0037] In addition, before performing step 140, the actual wind speed observed at any point (e.g., a wind tower) in the area where each group is distributed can be filtered according to a predetermined criterion (e.g., IEC standard) to filter out abnormal data beyond the normal range. As an example, data including but not limited to wind speed greater than 50m / s, wind speed less than 0m / s, temperature less than -45°C, temperature greater than 50°C, air pressure less than 60kPa, and air pressure greater than 120kPa can be regarded as data beyond the normal range, and data that remains constant within a certain time range can also be determined as frozen data, and data with inconsistent wind speed trends and low values ​​in a certain channel can also be regarded as abnormal data. In this way, these abnormal data can be screened out from the observed actual wind speed for removal, thereby further ensuring the validity and accuracy of the data.

[0038] According to the above method, the actual wind speed of each group distribution area in the target area can be predicted, and the wind speed prediction model of the target area can be constructed based on the predicted actual wind speed. In order to make the predicted results of the constructed wind speed prediction model more accurate, various methods can also be used to test the effect and train the wind speed prediction model.

[0039] In an illustrative embodiment, first, the actual wind speed Y1 observed at point A in an area S of any group distribution in the target area and the concurrent simulated wind speed X1 for point A derived from the meteorological simulation data can be obtained, and an initial wind speed prediction model for the area S is constructed based on the actual wind speed Y1 and the concurrent simulated wind speed X1, and the initial wind speed prediction model is trained to obtain a wind speed prediction model M for the area S; then, the actual wind speed Y2 observed at another point (point B) in the area S and the concurrent simulated wind speed X2 for point B derived from the meteorological simulation data can be obtained, and the actual wind speed Y2 and the concurrent simulated wind speed X2 are used to test the effectiveness of the wind speed prediction model M. Specifically, the simulated wind speed X2 for the same period can be input into the wind speed prediction model M to output the actual wind speed P2 predicted for point B, and then the predicted actual wind speed P2 is compared with the acquired actual wind speed Y2. If the comparison index reaches a preset threshold value T0 (the threshold value can be set based on experience), it indicates that the effect of the wind speed prediction model M has passed the test, and the wind speed prediction model M is recorded and saved and determined as the wind speed prediction model for the area to provide a calculation reference for future predictions; if the comparison index cannot reach the preset threshold value T0, it indicates that the effect of the wind speed prediction model M is poor, and the wind speed prediction model can be reconstructed by re-acquiring spatial clustering features or modifying wind speed prediction model parameters or replacing the wind speed prediction model.

[0040] As an example, the Pearson correlation coefficient, root mean square error, mean absolute error, etc. may be used as the above comparison indexes, including but not limited to the Pearson correlation coefficient, root mean square error, mean absolute error, etc., and the Pearson correlation coefficient, root mean square error, and mean absolute error may be calculated using the following formulas, respectively:

[0041]

[0042]

[0043]

[0044] In the above formulas, r represents the Pearson correlation coefficient, RMSE represents the root mean square error, MAE represents the mean absolute error, n represents the number of samples, and X i and Y i They represent the predicted actual wind speed and the observed actual wind speed at point B, respectively. and They represent the mean of the predicted actual wind speed and the mean of the observed actual wind speed at point B, respectively. X and S Y They represent the variance of the predicted actual wind speed and the variance of the observed actual wind speed at point B respectively.

[0045] It should be pointed out that the various steps described above in the present invention may be split into more steps according to the needs of implementation, and two or more steps or partial operations of the steps may be combined into a new step to implement the present invention.

[0046] It can be seen that by adopting the above implementation process, not only can the wind resources be accurately evaluated when wind measurement is insufficient or even absent, but the uncertainty in the calculation of wind farm power generation can also be effectively reduced, thereby providing more accurate and reliable data basis for the site selection and construction of wind farms and the evaluation of their power generation.

[0047] Figure 2 is a block diagram showing the structure of a wind speed prediction device according to an exemplary embodiment of the present invention.

[0048] Reference Figure 2 , the wind speed prediction device may include a meteorological acquisition unit 210, a group division unit 220, an area determination unit 230 and a wind speed determination unit 240. The meteorological acquisition unit 210 may be used to obtain meteorological simulation data of the target area under a predetermined meteorological mode. The group division unit 220 may be used to divide the grid points of the target area into different groups according to the wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data, wherein the grid points in the same group have the same or similar wind speed variation characteristics. The area determination unit 230 may be used to determine the areas where the divided groups are distributed. The wind speed prediction unit 240 may be used to predict the actual wind speed at another point other than the point in the area where each group is distributed based on the actual wind speed observed at any point in the area where each group is distributed.

[0049] As mentioned above, in the wind speed prediction device, the meteorological model can adopt a mesoscale meteorological model, so that the acquired meteorological simulation data can better reflect the changes in the flow field and weather conditions in the target area, and is also closer to the actual situation of wind speed changes in the target area. Accordingly, in the mesoscale meteorological model, the wind speed change feature can be a feature of wind speed changes over time.

[0050] As described above, in the wind speed prediction device, the area determination unit 230 may determine the areas where each divided group is distributed based on the areas surrounded by the grid points in each divided group.

[0051] As described above, in the wind speed prediction device, the wind speed prediction unit 240 may further include a deviation determination subunit (not shown) and a wind speed prediction subunit (not shown). The deviation determination subunit may be used to determine the deviation of the actual wind speed observed at any point in the area of ​​each group distribution relative to the simulated wind speed at the point derived from the meteorological simulation data as the deviation of the actual wind speed in the area of ​​each group distribution. The wind speed prediction subunit may be used to predict the actual wind speed at another point in the area of ​​each group distribution other than the point based on the deviation of the actual wind speed in the area of ​​each group distribution and the simulated wind speed at another point in the area of ​​each group distribution other than the point derived from the meteorological simulation data.

[0052] It can be seen that by adopting the above implementation process, not only can the wind resources be accurately evaluated when wind measurement is insufficient or even absent, but the uncertainty in the calculation of wind farm power generation can also be effectively reduced, thereby providing more accurate and reliable data basis for the site selection and construction of wind farms and the evaluation of their power generation.

[0053] According to an exemplary embodiment of the present invention, a computer-readable storage medium storing a computer program is also provided. The computer-readable storage medium stores a computer program that, when executed by a processor, causes the processor to perform a method for determining wind speed according to the present invention. The computer-readable recording medium is any data storage device that can store data read out by a computer system. Examples of computer-readable recording media include: read-only memory, random access memory, read-only optical disk, magnetic tape, floppy disk, optical data storage device, and carrier wave (such as data transmission through the Internet via a wired or wireless transmission path).

[0054] According to an exemplary embodiment of the present invention, a computer device is also provided. The computer device includes a processor and a memory. The memory is used to store a computer program. The computer program is executed by the processor so that the processor executes the computer program of the method for determining wind speed according to the present invention.

[0055] While the present application has been shown and described with reference to preferred embodiments, it will be understood by those skilled in the art that various modifications and variations may be made to these embodiments without departing from the spirit and scope of the present application as defined by the appended claims.

Claims

1. A wind speed prediction method, characterized in that: The method comprises: Acquire weather simulation data of the target area under a predetermined weather mode; According to wind speed variation characteristics of simulated wind speed at each grid point of the target area derived from the meteorological simulation data, the grid points of the target area are divided into different groups, wherein the grid points in the same group have the same or similar wind speed variation characteristics, the wind speed variation characteristics being characteristics of wind speed variation over time and depending on time series data of simulated wind speed derived under a predetermined meteorological mode; Determine the distribution areas of each divided group, so as to convert the clustering based on wind speed change characteristics into spatial area clustering of wind speed distribution conditions; Based on the actual wind speed observed at any point in the area where each group is distributed, predicting the actual wind speed at another point other than the point in the area where each group is distributed; The step of predicting the actual wind speed at another point other than the point in the area where each group is distributed comprises: determining a deviation of an actual wind speed observed at any point in the area of ​​each group distribution from a simulated wind speed at the point derived from the meteorological simulation data as a deviation of the actual wind speed in the area of ​​each group distribution; Based on the deviation of the actual wind speed in the area of ​​each group distribution and the simulated wind speed at another point other than the point in the area of ​​each group distribution derived from the meteorological simulation data, the actual wind speed at another point other than the point in the area of ​​each group distribution is predicted.

2. The wind speed prediction method according to claim 1, characterized in that: The step of determining the distribution areas of the divided groups includes: The areas where each of the divided groups is distributed are determined based on the areas surrounded by the grid points in each of the divided groups.

3. The wind speed prediction method according to claim 1, characterized in that: The predetermined meteorological model is a mesoscale meteorological model.

4. A wind speed prediction device, characterized in that: The device comprises: A meteorological acquisition unit, which acquires meteorological simulation data of a target area under a predetermined meteorological mode; a group division unit, dividing each grid point of the target area into different groups according to wind speed variation characteristics of the simulated wind speed at each grid point of the target area derived from the meteorological simulation data, wherein the grid points in the same group have the same or similar wind speed variation characteristics, the wind speed variation characteristics being characteristics of wind speed variation over time and depending on time series data of the simulated wind speed derived under a predetermined meteorological mode; An area determination unit determines the area where each divided group is distributed, so as to convert the clustering according to the wind speed change characteristics into the spatial area clustering of the wind speed distribution status; a wind speed prediction unit for predicting an actual wind speed at another point other than the point in the area where each group is distributed, based on the actual wind speed observed at any point in the area where each group is distributed; Wherein, the wind speed prediction unit comprises: a deviation determination subunit, which determines a deviation of an actual wind speed observed at any point in the area of ​​each group distribution relative to a simulated wind speed at the point derived from the meteorological simulation data as a deviation of the actual wind speed in the area of ​​each group distribution; A wind speed prediction subunit predicts the actual wind speed at another point other than the point in the area of ​​each group distribution based on the deviation of the actual wind speed in the area of ​​each group distribution and the simulated wind speed at another point other than the point in the area of ​​each group distribution derived from the meteorological simulation data.

5. The wind speed prediction device according to claim 4, characterized in that: The area determination unit determines the areas where each of the divided groups is distributed based on areas surrounded by grid points in each of the divided groups.

6. The wind speed prediction device according to claim 4, characterized in that: The predetermined meteorological model is a mesoscale meteorological model.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for wind speed prediction according to any one of claims 1 to 3 is implemented.

8. A computer device, characterized in that: The computer device comprises: processor; A memory storing a computer program, which, when executed by a processor, implements the method for wind speed prediction as described in any one of claims 1 to 3.

Images