Numerical prediction method and device for wind turbine power based on coupled blade element momentum monte carlo simulation

Abstract

The application discloses a wind turbine power numerical prediction method and device coupled with the bin method Monte Carlo simulation. The method statistically analyzes the historical operation data of each wind turbine of a wind farm, constructs a discrete probability density function of the theoretical power output of the whole wind farm and the power output of each wind turbine by using the bin method, and obtains a power curve by using the Monte Carlo simulation, that is, the theoretical power output of the wind farm is the abscissa and the expected power output of the wind turbine is the ordinate. Further, the curve is used as an input parameter of a wind farm parameterization model in a numerical weather prediction mode, so as to predict the power output of the wind turbine. The application realizes the online coupling of the numerical weather prediction mode and the data statistical analysis, considers the power influence of the whole wind farm regulation and control on the single wind turbine, is more accurate compared with the traditional wind farm parameterization method, can predict the power outputs of multiple wind turbines of the whole wind farm at one time, and has obvious power prediction advantages of the whole wind farm.

Description

Technical Field

[0001] This invention belongs to the field of numerical weather prediction, specifically relating to a numerical prediction method for wind turbine power using coupled Bean Monte Carlo simulation. Background Technology

[0002] Wind power, as a clean and renewable energy source, is an indispensable energy source for achieving carbon neutrality. However, the volatility of wind speed greatly affects the security of the power grid and increases the difficulty of power dispatch. Therefore, accurate wind power forecasting is of great significance for wind power grid connection. In recent years, ultra-short-term wind power forecasting has often adopted artificial intelligence methods for time series prediction of wind power. For example, application number CN202011181059.7 improves the accuracy of wind power forecasting by using a BP neural network optimized based on the Cuckoo algorithm. Application number CN201911416406.7 improves the accuracy of wind power forecasting by constructing a method that takes into account the power output fluctuation process and using various error correction methods to improve the convolutional neural network algorithm.

[0003] With the rapid development of wind power, short-term and medium-term wind power is becoming increasingly important. Traditional forecasting methods use weather forecast models and numerical simulations to predict wind speed and assess the expected power generation of wind farms. To improve forecast accuracy, more and more methods combine historical wind turbine data with weather forecasts to predict wind turbine power output. For example, application number CN201910401140.2 proposes an intelligent hybrid model method that uses a dynamic feature extraction algorithm to mine historical power time series and publicly available numerical weather forecast data to solve the problem of wind power generation being difficult to predict at different time steps. Application number CN202011409788.3 discloses a short-term power prediction method for wind farm clusters based on spatiotemporal graph convolutional neural networks, which uses a BiGRU network and a graph convolutional network to jointly train historical power vectors and sample weather forecast parameter matrix time series to improve the accuracy of wind power prediction.

[0004] However, the above methods have two main shortcomings: they do not consider the correlation of power output between wind turbines; and the artificial intelligence method and weather forecasting model are coupled offline, failing to realize their mutual influence. These shortcomings greatly limit the accuracy of power prediction. To address these shortcomings, this invention employs the Bean method Monte Carlo simulation to link the power output of the entire wind farm with the power output of a single wind turbine, thus considering the distribution relationship of different wind turbines under different power demands. Simultaneously, by coupling the above method with the Fitch wind farm parameterization method based on weather research and forecasting models, online coupling of data statistical analysis and numerical weather prediction models is achieved. Summary of the Invention

[0005] The purpose of this invention is to solve the problems existing in the current technology and to provide a numerical prediction method for wind turbine power using coupled Bean Monte Carlo simulation.

[0006] The objective of this invention is achieved through the following technical solution: a numerical prediction method for wind turbine power using coupled Bean method Monte Carlo simulation, the steps of which are as follows:

[0007] (1) Collect basic information data of each wind turbine in the target wind farm, including geographical location information, wind turbine design parameters and historical SCADA operation monitoring data of the wind turbine;

[0008] (2) Preprocess the SCADA historical operation monitoring data in step (1) to construct a new database including wind turbine wind speed, actual power output of wind turbine, and theoretical power output of wind farm;

[0009] (3) Using the new database obtained in step (2), perform statistical analysis based on the Bean method; sort the time series data of the theoretical power output of the wind farm from low to high and divide it into equally spaced regions to obtain the theoretical output bins of the wind farm from low to high, denoted as wind farm bin; specifically: using the time series data of the actual power output of each wind turbine, count the start frequency of each wind turbine in each wind farm bin, i.e. the time when the power output is greater than zero, and calculate the start probability of each wind turbine in each wind farm bin; for each wind farm bin, sort the time series data of the actual power output of each wind turbine from low to high and divide it into equally spaced regions to obtain the actual power output bins of each wind turbine, denoted as wind turbine bin; count the operating frequency of each wind turbine in each wind turbine bin and calculate the cumulative probability in each wind turbine bin;

[0010] (4) Using the wind turbine start-up probability and cumulative probability under different wind farm bins obtained in step (3), Monte Carlo simulation is carried out to calculate the expected power output of each wind turbine under each wind farm bin. Specifically, for each wind farm bin, the first random number [0,1] is taken. For each wind turbine, if the random number is greater than the start-up probability of the wind turbine under this wind farm bin, the wind turbine is shut down and its power output is 0. Otherwise, the wind turbine is started. If the wind turbine is started, the second random number [0,1] is taken to find the wind turbine bin with the cumulative probability equal to this random number. The power output of the wind turbine is determined by linear interpolation. Then, the simulated average power output of each wind turbine is taken as the final expected power output of each wind turbine under each wind farm bin. The correspondence between the theoretical power output of the wind farm and the expected power output of each wind turbine is constructed to obtain the power curve.

[0011] (5) Using the numerical research and forecasting WRF model and the Fitch wind farm parameterization scheme built into WRF, the theoretical power output of the wind farm is predicted, and the power output of each wind turbine is predicted using the power curve obtained in step (4). Specifically, the wind farm parameterization scheme reads the wind turbine design parameters obtained in step (1) and parameterizes the interaction between the wind turbine and the atmosphere as a momentum sink and a turbulent kinetic energy source; the wind turbine incoming wind speed is calculated by simulating the atmospheric background flow field using WRF, and the theoretical power output of the wind turbine is calculated by coupling the wind farm parameterization scheme; the theoretical power output of the wind farm is obtained by summing the theoretical power output of each wind turbine; and the power output of each wind turbine is predicted by reading the new wind power curve in step (4).

[0012] Furthermore, in step (1), the geographical location information is the latitude and longitude of the wind turbine, and the wind turbine design parameters include the rotor diameter, rotor area, hub height, static thrust coefficient, power, and design power operating curve.

[0013] Further, in step (2), the preprocessing specifically involves cleaning the historical wind turbine wind speed and power output data in the SCADA historical operation monitoring data, i.e. deleting or processing unreasonable data; and calculating the theoretical power output of the wind turbine based on the historical wind speed data and the design power operation curve of the wind turbine, and summing them to obtain the theoretical power output of the wind farm.

[0014] Furthermore, the unreasonable data and processing methods are as follows: treat all values ​​with power less than zero as zero values; delete the data at the time of the missing value; delete outliers, that is, when the wind speed is less than the wind turbine's cut-in wind speed or greater than the cut-out wind speed, the power is greater than zero.

[0015] Furthermore, in step (3), the interval division specifically refers to dividing the time series data according to the time of the time series data, and the interval length is equal to the number of time points contained, that is, equal interval division.

[0016] Furthermore, in step (3), the start-up probability of each wind turbine under the wind farm Bin is calculated as the start-up frequency divided by the interval length of the wind farm Bin.

[0017] Furthermore, in step (4), the power curve is defined as follows: the horizontal axis represents the theoretical power output of the wind farm, and the vertical axis represents the expected power output of each wind turbine.

[0018] Secondly, the present invention also provides a wind turbine power numerical simulation prediction device coupled with Bean method and Monte Carlo simulation, including a memory and one or more processors, wherein the memory stores executable code, and when the processor executes the executable code, it implements the wind turbine power prediction method coupled with Bean method and Monte Carlo simulation.

[0019] Thirdly, the present invention also provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the aforementioned method for numerical prediction of wind turbine power using coupled Beansfield Monte Carlo simulation.

[0020] The beneficial effects of this invention are as follows: The wind turbine power numerical prediction method and apparatus based on coupled Bean method Monte Carlo simulation established in this invention is an innovation and improvement on traditional numerical prediction methods. It uses the Bean method to construct the probability distribution of wind power output in a wind farm, and employs the Monte Carlo method to analyze the nonlinear relationship between the power output of a single wind turbine and the entire wind farm. By coupling it to wind farm parameterization based on the WRF model, it achieves regional prediction of wind power. This invention has advantages such as higher accuracy and the ability to predict the power output of multiple wind turbines in a single operation. The improved wind power prediction method is applicable to large-scale wind farms and has a greater advantage in predicting the wind power of the entire wind farm. Attached Figure Description

[0021] Figure 1 A flowchart illustrating a method for numerical prediction of wind turbine power using coupled Bean Monte Carlo simulation provided by the present invention.

[0022] Figure 2 A flowchart for Monte Carlo simulation;

[0023] Figure 3 A schematic diagram comparing the predicted and measured power of multiple wind turbines;

[0024] Figure 4 This is a normal distribution plot of the relative error in wind turbine power prediction.

[0025] Figure 5 A schematic diagram comparing predicted and measured power of a wind farm.

[0026] Figure 6 This is a normal distribution diagram of the relative error in wind farm power prediction.

[0027] Figure 7 The diagram shows the structure of a wind turbine power numerical prediction device coupled with Beansfield Monte Carlo simulation provided by the present invention. Detailed Implementation

[0028] The present invention will be further described and illustrated below with reference to the accompanying drawings and specific embodiments.

[0029] This invention mainly constructs a numerical prediction method for wind turbine power coupled with Bean's method Monte Carlo simulation. The method steps are as follows: Figure 1 As shown. The specific steps are as follows:

[0030] (1) Collect basic information data of each wind turbine in the target wind farm; the basic information data includes geographical location information, wind turbine design parameters and historical operation data of wind turbine, specifically including wind turbine latitude and longitude coordinates, wind turbine area, hub height, design power operation curve, SCADA database;

[0031] (2) Preprocessing the SCADA database in step (1): Data cleaning is performed on the historical wind turbine wind speed and power output data in the SCADA database. Specifically, all power values ​​less than zero are treated as zero, and the data at the time of the missing value is deleted; outliers are deleted, where data with power greater than zero are outliers when the wind speed is less than the wind turbine's cut-in wind speed or greater than the cut-out wind speed; based on the historical wind speed data and the wind turbine power design curve in step (1), the theoretical power output of the wind turbine is calculated and summed to obtain the theoretical power output of the wind farm. The calculation method is as follows:

[0032]

[0033] Among them G wf (t) represents the theoretical power output of the wind farm at time t, C P Let U(k,t) be the theoretical power coefficient of the wind turbine, U(k,t) be the wind speed at the hub height of wind turbine k at time t (referred to as wind speed), A(k) be the rotor area of ​​wind turbine k, and nt be the total number of wind turbines. Finally, step (2) constructs a new database G including wind turbine wind speed U(k,t), actual power output of wind turbine P(k,t), and theoretical power output of wind farm. wf (t).

[0034] (3) Using the new database obtained in step (2), perform statistical analysis based on the Bean method: reorder the time series data of the theoretical power output of the wind farm from low to high, and divide it into intervals to obtain the wind farm Bins from low to high. wf (i), where i is the Bin number of the wind farm. Since the probability of a wind farm operating in both low and high power output ranges is low, the low power Bin... wf and high-power Bin wf It has a longer interval length, i.e., more data; using the time series data P(k,t) of the actual power output of each wind turbine, it counts the start-up frequency of each wind turbine in each wind farm bin, i.e., the number of times P(k,t)>0, and calculates the start-up probability of each wind turbine in each wind farm bin. The calculation method is as follows:

[0035]

[0036] Among them, f wf (i,k) represents the wind turbine k in Bin. wf (i) represents the activation probability, kt(i,k) represents the number of times P(k,t)>0, and Span) represents the Bin wf The length of the interval (i);

[0037] For each Bin wf The actual power output of each wind turbine is sorted from low to high. Values ​​P(k,t) = 0 are removed and the range is divided. The reason for removing zero values ​​is that the probability of whether wind turbine k starts has already been statistically determined. wf The statistics were obtained, and the actual power output Bins of each wind turbine was finally obtained. wt (i,j,k), where i is Bin wf The number, j is Bin wt Number (k is the wind turbine number); count the number of wind turbine k in each Bin wt The frequency of operation, and calculate the number of each Bin. wt The cumulative probability of wind turbine operation under the given conditions, i.e.:

[0038]

[0039]

[0040] Where mt(m,k) is the wind turbine k in Bin wt The number of runs within (i,j,k), where Span′(i) is the Bin number. wt The length of the interval (i,j,k), f wt (i,j,k) Wind turbine k is in the interval Bin wt The probability of running (i,j,k), F wt (i,j,k) represents the cumulative probability. Finally, step (3) obtains the theoretical output of the wind farm, Bins(Bin), through the Bean method. wf (i) and the corresponding start-up probability (f) of each wind turbine under different wind farm bins. wf (i,k)), and simultaneously obtained the actual power output of each wind turbine for each wind farm Bin (Bins(Bin)). wt (i,j,k)) and the corresponding cumulative operating probability (F) of each wind turbine under different wind turbine bins. wt (i,j,k)).

[0041] (4) Using the wind turbine start-up probability and cumulative operating probability obtained in step (3) under different bins, perform Monte Carlo simulation to calculate the expected power output of each wind turbine under each wind farm bin: such as Figure 2 As shown, for each wind farm Bin, a random number α is taken that is uniformly distributed in the range [0,1]. If α > f wf If (i,k), then wind turbine k is in a stopped state, i.e., P(i,k)=0; otherwise, wind turbine k is started. If the wind turbine is started, take a random number β uniformly distributed in [0,1] to find F. wt (i,j+1,k)≤β≤F wt (i,j,k), determine the value of k, which is the measured power output Bin of the wind turbine. wt The numbering is determined by linear interpolation to obtain the expected power output P(i,k) of wind turbine k, i.e.:

[0042]

[0043] Repeat the first two steps 10,000 times, taking the simulated average power output of each wind turbine as the final expected power output of each wind turbine in each wind farm bin, i.e.:

[0044]

[0045] Where ntest represents the total number of Monte Carlo simulations. After multiple tests, the expected value tends to stabilize after ntest exceeds 5000. In this study, ntest = 10000 is used. Finally, step (4) constructs the correspondence between the theoretical power output of the wind farm and the power output of each wind turbine, that is, a new type of power curve with the theoretical power output of the wind farm on the horizontal axis.

[0046] (5) Using the numerical weather prediction model WRF and the Fitch wind farm parameterization scheme, the theoretical power output of the wind farm is predicted, and the power output of each wind turbine is predicted using the power curve obtained in step (4): The Fitch wind farm parameterization scheme reads the wind turbine design parameters obtained in step (1) and models the wind turbine into momentum sink and turbulent energy terms; the wind turbine inflow velocity is calculated by simulating the atmospheric background flow field using WRF, and the theoretical power output of the wind turbine is calculated by coupling with the Fitch wind farm parameterization scheme; the theoretical power output G of the wind farm is obtained by summing the theoretical power output of each wind turbine. wf ; By reading the new Bin in step (4) wf -P wt The wind power curve predicts the power output of each wind turbine.

[0047] The specific implementation effect of the above method is demonstrated below with reference to the embodiments.

[0048] Example

[0049] In this embodiment, a large offshore wind farm located in Hangzhou Bay, Zhejiang Province, is used as the research object to explore the specific implementation effect and advantages of a numerical prediction method for wind turbine power coupled with Bean's method Monte Carlo simulation. The historical operating data consists of SCADA data from each wind turbine, with a data duration approximately covering a full year of 2022 and a time step of 10 minutes. Based on preliminary data quality analysis, wind speed and power data from November, which has relatively good data quality, are specifically used, resulting in a total data length of 6 × 24 × 30 = 4320. This data is used to predict the wind power for the first three days of December, specifically the power output from December 1st to December 3rd.

[0050] This embodiment is divided into 63 Bins. wf (i), i.e., i = 1, 2, 3, ..., 63. Since the occurrence frequency is low at the two ends of the interval, meaning the probability of wind farms operating in both low and high power output ranges is low, the first three and last three intervals need to be divided into larger data sets. Specifically, the data length G is 4320. wf The data is divided into two halves, each containing 6 bins, for 20% of the time intervals. The total interval length is 4320 × 20% / 6 = 144. Therefore, the remaining 57 bins correspond to an interval length of 61. Next, the data is divided into 20 bins of equal length. wt (i,j,k). During the Monte Carlo simulation, through multiple tests, the expected value tended to stabilize after ntest was greater than 5000. In this embodiment, ntest = 10000 was chosen. Bin was constructed using the Monte Carlo method. wf -P wt Curve. WRF reads the Bin... wf -P wt The curve is used to calculate the power P of each wind turbine.

[0051] This invention employs both the original wind farm parameterization scheme (Fitch) wind power prediction method and the Bean method Monte Carlo simulation coupled with wind farm numerical parameterization (Control) wind power prediction method based on historical operating data. The effectiveness of the new model is demonstrated through comparison with measured data and two sets of experiments. The prediction period was selected as the week ending September 2022, i.e., September 23-30. The parameter settings corresponding to WRF version 4.3 used are as follows:

[0052] Table 1 WRF Parameter Settings

[0053]

[0054]

[0055] To demonstrate the predictive capability of this invention, three prediction methods were compared: (1) Physical model (PHY): high-precision WRF model + theoretical power curve prediction, that is, using high-resolution WRF to predict wind speed and predicting power based on the established power curve of the wind turbine; (2) Statistical model (STA): high-precision WRF model + ANN, that is, using WRF to predict wind speed as output and using ANN machine learning model to output predicted power; (3) This invention (Monte-WRF): a numerical prediction method for wind turbine power coupled with Bean method Monte Carlo simulation. Figure 3 The paper presents a comparison between the power prediction results and measured data of four wind turbines under different prediction methods (Obs). Figure 4 This represents the normal distribution result of the relative error in wind turbine power prediction. Figure 5 This study compares the predicted results of wind farms under different power prediction methods with measured data. Figure 6 This represents the normal distribution of the relative error in wind power prediction for wind farms. By comparing this with measured data, we can... Figure 3 and Figure 4 As can be seen, this invention efficiently captures the power control phenomenon of wind turbines during high-power output in wind farms, effectively controlling the relative predicted power error within [-0.5 to 0.5], with an average relative error of only 0.07, i.e., an accuracy rate as high as 93%. Compared with traditional models, its accuracy is improved by more than 10%. Meanwhile, because this invention has the characteristic of regional prediction, that is, considering the correlation of wind turbines within the wind farm, it is suitable for the power output of the entire wind farm. Figure 5 and Figure 6 It can be found that the error confidence interval of its wind farm power prediction is reduced by 17% compared with the wind turbine power prediction, and by more than 20% compared with the error interval of the traditional model.

[0056] Corresponding to the aforementioned embodiment of a wind turbine power numerical prediction method coupled with Beansfield Monte Carlo simulation, the present invention also provides an embodiment of a wind turbine power numerical prediction device coupled with Beansfield Monte Carlo simulation.

[0057] See Figure 7 The present invention provides a wind turbine power numerical prediction device coupled with Beansfield Monte Carlo simulation, comprising a memory and one or more processors. The memory stores executable code, and when the processor executes the executable code, it is used to implement a wind turbine power numerical prediction method coupled with Beansfield Monte Carlo simulation as described in the above embodiment.

[0058] An embodiment of the wind turbine power numerical prediction device coupled with Beansfield Monte Carlo simulation provided by this invention can be applied to any device with data processing capabilities, such as a computer. The device embodiment can be implemented in software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it is formed by the processor of any data processing device loading the corresponding computer program instructions from non-volatile memory into memory for execution. From a hardware perspective, such as... Figure 7 The diagram shown is a hardware structure diagram of any device with data processing capabilities, which is a numerical prediction device for wind turbine power coupled with the Beansfield Monte Carlo simulation provided by this invention. (Except for...) Figure 7 In addition to the processor, memory, network interface, and non-volatile memory shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0059] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0060] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the present invention according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0061] This invention also provides a computer-readable storage medium storing a program that, when executed by a processor, implements a method for numerical prediction of wind turbine power using coupled Beansfield Monte Carlo simulation as described in the above embodiments.

[0062] The computer-readable storage medium can be an internal storage unit of any data processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device of any data processing device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units and external storage devices of any data processing device. The computer-readable storage medium is used to store the computer program and other programs and data required by the data processing device, and can also be used to temporarily store data that has been output or will be output.

[0063] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A numerical prediction method for wind turbine power coupled with Bean's method Monte Carlo simulation, characterized in that, The steps of this method are as follows: (1) Collect basic information data of each wind turbine in the target wind farm, including geographical location information, wind turbine design parameters and historical SCADA operation monitoring data of the wind turbine; (2) Preprocess the SCADA historical operation monitoring data in step (1) to construct a new database including wind turbine wind speed, actual power output of wind turbine, and theoretical power output of wind farm; (3) Using the new database obtained in step (2), perform statistical analysis based on the Bean method; sort the time series data of the theoretical power output of the wind farm from low to high and divide it into equally spaced intervals to obtain the theoretical output Bins of the wind farm from low to high, denoted as wind farm Bin; specifically: using the time series data of the actual power output of each wind turbine, count the number of times each wind turbine starts in each wind farm Bin, that is, the time when the power output is greater than zero, and calculate the starting probability of each wind turbine in each wind farm Bin; For each wind farm Bin, the time series data of the actual power output of each wind turbine is sorted from low to high and divided into equally spaced regions to obtain the actual power output Bins of each wind turbine, denoted as wind turbine Bin; Count the operating frequency of each wind turbine in each wind turbine bin, and calculate the cumulative probability under each wind turbine bin; (4) Using the wind turbine start-up probability and cumulative probability under different wind farm bins obtained in step (3), Monte Carlo simulation is carried out to calculate the expected power output of each wind turbine under each wind farm bin. Specifically, for each wind farm bin, the first random number [0, 1] is taken. For each wind turbine, if the random number is greater than the start-up probability of the wind turbine under this wind farm bin, the wind turbine is shut down and its power output is 0. Otherwise, the wind turbine is started. If the wind turbine is started, the second random number [0, 1] is taken to find the wind turbine bin with the cumulative probability equal to this random number. The power output of the wind turbine is determined by linear interpolation. Then, the simulated average power output of each wind turbine is taken as the final expected power output of each wind turbine under each wind farm bin. The correspondence between the theoretical power output of the wind farm and the expected power output of each wind turbine is constructed to obtain the power curve. (5) Using the numerical research and forecasting WRF model and the Fitch wind farm parameterization scheme built into WRF, the theoretical power output of the wind farm is predicted, and the power output of each wind turbine is predicted using the power curve obtained in step (4). Specifically, the wind farm parameterization scheme reads the wind turbine design parameters obtained in step (1) and parameterizes the interaction between the wind turbine and the atmosphere as a momentum sink and a turbulent kinetic energy source; the wind turbine incoming wind speed is calculated by simulating the atmospheric background flow field using WRF, and the theoretical power output of the wind turbine is calculated by coupling the wind farm parameterization scheme; the theoretical power output of the wind farm is obtained by summing the theoretical power output of each wind turbine; and the power output of each wind turbine is predicted by reading the new wind power curve in step (4).

2. The method for numerical prediction of wind turbine power using coupled Bean method Monte Carlo simulation as described in claim 1, characterized in that, In step (1), the geographical location information is the latitude and longitude of the wind turbine, and the wind turbine design parameters include the rotor diameter, rotor area, hub height, static thrust coefficient, power, and design power operating curve.

3. The method for numerical prediction of wind turbine power using coupled Bean method Monte Carlo simulation as described in claim 2, characterized in that, In step (2), the preprocessing specifically involves cleaning the historical wind turbine wind speed and power output data in the SCADA historical operation monitoring data, i.e. deleting or processing unreasonable data; and calculating the theoretical power output of the wind turbine based on the historical wind speed data and the design power operation curve of the wind turbine, and summing them to obtain the theoretical power output of the wind farm.

4. The method for numerical prediction of wind turbine power using coupled Bean's method Monte Carlo simulation as described in claim 3, characterized in that, The unreasonable data and processing methods are as follows: treat all values ​​with power less than zero as zero values; delete the data at the time of the missing value; delete outlier values, that is, values ​​with power greater than zero when the wind speed is less than the wind turbine's cut-in wind speed or greater than the cut-out wind speed.

5. The method for numerical prediction of wind turbine power using coupled Bean's method Monte Carlo simulation as described in claim 1, characterized in that, In step (3), the interval division is specifically based on the time of the time series data. The interval length is equal to the number of time points contained, which is an equal interval division.

6. The method for numerical prediction of wind turbine power using coupled Bean method Monte Carlo simulation as described in claim 1, characterized in that, In step (3), the start-up probability of each wind turbine under the Bin of the wind farm is calculated as the start-up frequency divided by the interval length of the Bin of the wind farm.

7. The method for numerical prediction of wind turbine power using coupled Bean method Monte Carlo simulation as described in claim 1, characterized in that, In step (4), the power curve is represented by the theoretical power output of the wind farm on the horizontal axis and the expected power output of each wind turbine on the vertical axis.

8. A numerical simulation and prediction device for wind turbine power coupled with Beansfield Monte Carlo simulation, comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that, When the processor executes the executable code, it implements a wind turbine power prediction method that combines the Bean method with Monte Carlo simulation as described in any one of claims 1-7.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements a method for numerical prediction of wind turbine power using coupled Bean Monte Carlo simulation as described in any one of claims 1-7.

Images