A wind farm power prediction method and system based on geographical environment and climate

Abstract

The present invention discloses a method and system for predicting wind farm power based on geographical environment and climate. The method comprises: S1. collecting historical power output data, geographical environment data, and climate data of the wind farm; S2. selecting geographical environment factors and climate factors that have a greater impact on the wind farm power output as characteristic variables based on the historical data; S3. determining the weight of each characteristic variable based on the degree of influence of each characteristic variable on wind power output; S4. using the weighted characteristic values ​​as input to establish a wind farm power prediction model; S5. inputting real-time or future geographical environment data and climate data into the trained power prediction model to obtain a wind farm power prediction result. By comprehensively considering geographical environment and climate conditions, especially seasonal variations, the present invention significantly improves the accuracy and reliability of wind power prediction.

Description

Technical Field

[0001] The present invention belongs to the field of data prediction, and more specifically, relates to a method for predicting wind farm power based on geographical environment and climate. The present invention also relates to a system for predicting wind farm power based on geographical environment and climate. Background Art

[0002] With the global energy transition and the rapid development of renewable energy, wind power, as a clean, renewable energy source, has gained widespread adoption in many countries. Wind farm power forecasting is crucial for grid scheduling, optimizing operations, and improving wind energy utilization. However, wind farm power output is affected by a variety of factors, including geographic environment and climate conditions. These variations lead to uncertainty and volatility in wind farm power output.

[0003] Existing wind power forecasting methods fall into two main categories: physical methods (based on numerical weather prediction models) and statistical methods (based on historical data). Physical methods rely on highly accurate meteorological data and complex physical models. While these methods offer high prediction accuracy, they are computationally intensive and place high demands on data quality and model parameters. Statistical methods, on the other hand, analyze historical data to develop forecasting models. While computationally simple and easy to implement, they often overlook the impact of geographical and climatic conditions on wind power, resulting in inaccurate forecasts.

[0004] Furthermore, existing forecasting methods typically fail to account for the impact of seasonal climate change on wind power, as well as seasonal variations in geographic environmental data. For example, climate factors such as temperature, humidity, wind speed, wind direction, and air pressure vary across seasons, and these variations can significantly impact wind power. Therefore, a wind power forecasting method that comprehensively considers both geographic and climatic factors is needed to improve forecast accuracy and reliability.

[0005] In summary, existing technologies lack a wind farm power forecasting method that comprehensively considers geographic and climatic factors, particularly a forecasting model that can adapt to seasonal variations. Therefore, developing a wind farm power forecasting method that incorporates geographic and climatic factors is crucial for improving the accuracy and reliability of wind power forecasts. Summary of the Invention

[0006] The purpose of the present invention is to address the shortcomings of the existing technology and propose a wind farm power prediction method and system based on geographical environment and climate. By comprehensively considering geographical environment and climatic conditions, especially seasonal changes, the accuracy and reliability of wind power prediction are significantly improved. At the same time, the method is easy to implement and apply, providing users with an effective tool to manage and optimize the operation of wind farms.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for predicting wind farm power based on geographical environment and climate includes the following steps:

[0009] S1. Collect historical power output data, geographical environment data, and climate data of the wind farm, wherein the geographical environment data includes topography, altitude, and surface roughness; and the climate data includes wind speed, wind direction, temperature, humidity, and air pressure;

[0010] S2. Based on historical data, select geographical and climatic factors that have a significant impact on wind farm power output as characteristic variables, and clean and standardize the collected data to eliminate the impact of different dimensions;

[0011] S3. Determine the weight of each characteristic variable based on the degree of influence of each characteristic variable on wind power output, multiply each characteristic variable by its corresponding weight to obtain a weighted characteristic value, and calculate the weight values ​​of the geographical environment and climate respectively based on the weighted characteristic values;

[0012] S4. Using the weighted eigenvalues ​​as input, establish a wind farm power prediction model, train the power prediction model using historical data, and use a portion of the data for cross-validation to evaluate the accuracy and generalization ability of the model;

[0013] S5. Input the real-time or future geographic environment data and climate data into the trained power prediction model to obtain the power prediction result of the wind farm station, and display the predicted wind farm station power output result to the user in the form of a graph or table.

[0014] Preferably, in step S1, the historical power output data of the wind farm is obtained from the operation records of the wind farm, the geographical environment data is obtained through a geographic information system or on-site measurements, and the meteorological data is obtained from a meteorological station or a numerical weather forecast model;

[0015] The step S1 also includes pre-processing the collected data, and using a method combining double quartile and clustering to eliminate abnormal data during power curtailment periods and when wind turbines fail.

[0016] Preferably, in step S2, the geographical environment factors and climate factors that have a greater impact on the power output of the wind farm are identified and selected as the characteristic variables of the model, specifically:

[0017] Based on professional knowledge in the field of wind power, we determine which factors may have a significant impact on wind power output, calculate the correlation coefficient between each factor and wind power output, and select the one with the highest correlation as the feature. The expression is:

[0018] ;

[0019] Where r is the correlation coefficient, x and y are the data sets of two variables, and n is the number of samples;

[0020] The standardization process is achieved by processing outliers, specifically:

[0021] Use the interquartile range to identify outliers, as expressed in:

[0022] ;

[0023] ;

[0024] Where Q1 and Q3 are the first and third quartiles of the data set, respectively. IQR is the difference between the third and first quartiles. The definition of outliers is based on IQR. Any data point smaller than Q1 − 1.5 × IQR or greater than Q3 + 1.5 × IQR is considered an outlier.

[0025] Preferably, the data normalization is used to eliminate the influence of different dimensions and ranges so that all features contribute equally to the model training, and the expression is:

[0026] ;

[0027] Where z is the standardized value, x is the original data, μ is the mean of the data, and σ is the standard deviation of the data.

[0028] Preferably, the weighted value expression of the geographical environment data in step S3 is:

[0029] W DL =W D +W H +W C ;

[0030] Where W DL Indicates the weight value of the geographical environment, WD represents the terrain weight, WH represents the altitude weight, and WC represents the surface roughness weight;

[0031] Among them, the weight value of the geographical environment is constant and is not affected by the four seasons;

[0032] The weighted value expression of the climate data is:

[0033] W QH =aW FS +aW FX +bW WD +bW SD +cW QY；

[0034] Where W QH The weight of climate data, W FS Indicates the wind speed weight value, W FX Indicates the wind direction weight value, W WD Represents the temperature weight value, W SD Represents the humidity weight value, W QY Represents the air pressure weight value, a, b, and c are constants, and the maximum value is 1.

[0035] Preferably, due to the influence of the four seasons, temperature, humidity, wind speed, wind direction and air pressure will change in different seasons. Therefore, in different seasons, the values ​​of constants a, b, and c are changed according to actual conditions, specifically:

[0036] The constant a is 0.7 in spring, 1 in summer, 0.5 in autumn, and 0.3 in winter;

[0037] The constant b is 0.5 in spring, 0.3 in summer, 0.5 in autumn, and 1 in winter;

[0038] The constant c is 0.7 in spring, 0.3 in summer, 0.7 in autumn, and 1 in winter.

[0039] Preferably, the specific process of step S4 is:

[0040] S41. Obtain historical wind speed, wind direction, temperature, humidity, and air pressure data, as well as corresponding wind power output data, and determine the main factors affecting wind power;

[0041] S42. Weight the features according to their importance to obtain weighted feature values, and form a data set. The data set is divided into a training set and a test set, with the ratio of 70% training set and 30% test set allocated.

[0042] S43. Select a model, train the model using the training set data, and analyze the model's performance on the training set and test set to check whether there is overfitting or underfitting;

[0043] S44. Select the best performing model as the final model.

[0044] Preferably, step S5 is specifically as follows:

[0045] S51. Obtain geographic environmental data such as wind speed, wind direction, temperature, humidity, and air pressure for the current or future period, and clean and format the data to ensure that it meets the model input requirements;

[0046] S52. Using the previously determined weighting coefficient, perform weighted processing on each eigenvalue of the real-time data to obtain a weighted eigenvalue, and integrate all the weighted eigenvalues ​​into a eigenvector as the input of the model;

[0047] S53, loading the trained wind power prediction model into the system, and inputting the weighted feature vector into the model to obtain the power prediction result of the wind farm station;

[0048] S54. Use a chart to show how the predicted wind power changes over time, and organize the predicted wind power and the corresponding time points into a table to facilitate users to view specific predicted values.

[0049] A wind farm power prediction system based on geographical environment and climate, the system being used to implement the above method, comprising:

[0050] Data collection and preprocessing module: This module is responsible for collecting historical power output data, geographical environment data, and climate data of wind farms. At the same time, the collected data is cleaned and standardized to eliminate the influence of different dimensions.

[0051] Feature selection and weighting module: Based on historical data, this module selects geographical and climatic factors that have a significant impact on wind farm power output as feature variables, and cleans and standardizes the collected data. It then determines the weight of each feature variable based on its impact on wind power output, multiplies each feature variable by its corresponding weight, and obtains a weighted feature value. Finally, based on the weighted feature values, it calculates the weights for both the geographical environment and climate.

[0052] Model training and prediction module: This module uses weighted eigenvalues ​​as input to build a wind farm power prediction model. The power prediction model is trained using historical data and cross-validated using a subset of the data to assess the model's accuracy and generalization capabilities. Real-time or future geographic and climate data are then input into the trained power prediction model to generate wind farm power prediction results.

[0053] Result display module: This module displays the predicted wind farm power output results to the user in the form of graphs or tables; users can intuitively understand the power forecast of the wind farm in different time periods;

[0054] System management and maintenance module: This module is responsible for the management and maintenance of the entire system, including data backup, recovery, and update operations; at the same time, it can also optimize and upgrade the system according to actual needs.

[0055] Technical effects and advantages of the present invention: Compared with the prior art, the present invention provides a method and system for predicting wind farm power based on geographical environment and climate, which has the following effects:

[0056] By comprehensively considering geographical and climatic factors, this method can more accurately predict the power output of wind farms. Using weighted eigenvalues ​​and weight calculations, the model can better capture the impact of different factors on wind power, thereby improving prediction accuracy. It also considers the impact of seasonal climate change on wind power. By adjusting the values ​​of constants a, b, and c, the model can adapt to seasonal changes in temperature, humidity, wind speed, wind direction, and air pressure. Seasonal adjustments improve the model's prediction performance across different seasons.

[0057] By processing outliers and standardizing data, this method eliminates the influence of different dimensions and ranges, making all features contribute equally to model training, improving the stability and reliability of the model. It combines historical data and real-time data, uses common statistical methods and machine learning models for prediction, and is easy to implement and apply. At the same time, by displaying the prediction results in charts, users can intuitively understand the power prediction status of wind farms in different time periods. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a flow chart of the wind farm power prediction method based on geographical environment and climate of the present invention;

[0059] Figure 2 This is a flow chart of establishing a wind farm power prediction model in an embodiment of the present invention;

[0060] Figure 3 This is a flow chart of predicting the power of a wind farm using a model in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0062] The present invention provides a method and system for predicting wind farm power based on geographical environment and climate. The system includes a data collection and preprocessing module: This module is responsible for collecting historical power output data, geographical environment data, and climate data of wind farms; at the same time, the collected data is cleaned and standardized to eliminate the influence of different dimensions;

[0063] Feature selection and weighting module: Based on historical data, this module selects geographical and climatic factors that have a significant impact on wind farm power output as feature variables, and cleans and standardizes the collected data. It then determines the weight of each feature variable based on its impact on wind power output, multiplies each feature variable by its corresponding weight, and obtains a weighted feature value. Finally, based on the weighted feature values, it calculates the weights for both the geographical environment and climate.

[0064] Model training and prediction module: This module uses weighted eigenvalues ​​as input to build a wind farm power prediction model. The power prediction model is trained using historical data and cross-validated using a subset of the data to assess the model's accuracy and generalization capabilities. Real-time or future geographic and climate data are then input into the trained power prediction model to generate wind farm power prediction results.

[0065] Result display module: This module displays the predicted wind farm power output results to the user in the form of graphs or tables; users can intuitively understand the power forecast of the wind farm in different time periods;

[0066] System management and maintenance module: This module is responsible for the management and maintenance of the entire system, including data backup, recovery, and update operations; at the same time, it can also optimize and upgrade the system according to actual needs;

[0067] The above-mentioned system management and maintenance module is responsible for the management and maintenance of the entire system, including data backup, recovery, and update operations, to ensure the stable operation and continuous optimization of the system.

[0068] Based on the above system, Figure 1 As shown, the wind farm power prediction method based on geographical environment and climate includes the following steps:

[0069] S1. Collect historical power output data, geographical environment data, and climate data of the wind farm station, wherein the geographical environment data includes topography, altitude, and surface roughness; and the climate data includes wind speed, wind direction, temperature, humidity, and air pressure. In step S1, the historical power output data of the wind farm station is obtained from the operation records of the wind farm, the geographical environment data is obtained through a geographic information system or on-site measurements, and the meteorological data is obtained from a meteorological station or a numerical weather prediction model.

[0070] The step S1 also includes pre-processing the collected data, and using a method combining double quartile and clustering to eliminate abnormal data during power curtailment periods and when wind turbines fail.

[0071] S2. Based on historical data, select geographical and climatic factors that have a significant impact on wind farm power output as characteristic variables, and clean and standardize the collected data to eliminate the impact of different dimensions. In step S2, identify and select those geographical and climatic factors that have a significant impact on wind farm power output as characteristic variables of the model, specifically:

[0072] Based on professional knowledge in the field of wind power, we determine which factors may have a significant impact on wind power output, calculate the correlation coefficient between each factor and wind power output, and select the one with the highest correlation as the feature. The expression is:

[0073] ;

[0074] Where r is the correlation coefficient, x and y are the data sets of two variables, and n is the number of samples;

[0075] The standardization process is achieved by processing outliers, specifically:

[0076] Use the interquartile range to identify outliers, as expressed in:

[0077] ;

[0078] ;

[0079] Where Q1 and Q3 are the first and third quartiles of the data set, respectively. IQR is the difference between the third and first quartiles. The definition of outliers is based on IQR. Any data point smaller than Q1 − 1.5 × IQR or greater than Q3 + 1.5 × IQR is considered an outlier.

[0080] Furthermore, the data normalization is used to eliminate the influence of different dimensions and ranges so that all features contribute equally to model training. The expression is:

[0081] ;

[0082] Where z is the standardized value, x is the original data, μ is the mean of the data, and σ is the standard deviation of the data.

[0083] For example, for wind farm power prediction, the following data sets were collected and their correlation coefficients with wind power output were calculated:

[0084] feature Dataset Correlation coefficient r <![CDATA[W D ]]> [0.5,0.6,0.4,0.7] 0.8 <![CDATA[W H ]]> [100,200,300,400] 0.6 <![CDATA[W C ]]> [0.1,0.2,0.3,0.4] 0.9 <![CDATA[W FS ]]> [5,6,7,8] 0.7 <![CDATA[W FX ]]> [180,190,200,210] 0.5 <![CDATA[W WD ]]> [20,22,24,26] 0.4 <![CDATA[W SD ]]> [50,55,60,65] 0.3 <![CDATA[W QY ]]> [1000,1010,1020,1030] 0.2

[0085] Use the interquartile range to identify outliers. The dataset is as follows:

[0086] feature Q1 Q3 IQR Outliers <![CDATA[W D ]]> 0.45 0.55 0.1 <0.25 or >0.85 <![CDATA[W H ]]> 150 250 100 <-50 or >350 <![CDATA[W C ]]> 0.15 0.25 0.1 <-0.05 or >0.35 <![CDATA[W FS ]]> 5.5 7.5 2.0 <1.5 or >11.5 <![CDATA[W FX ]]> 185 215 30 <155 or >245 <![CDATA[W WD ]]> 21 23 2 <17 or >29 <![CDATA[W SD ]]> 52 58 6 <46 or >64 <![CDATA[W QY ]]> 1005 1025 20 <985 or >1045

[0087] Each feature is standardized to obtain a cleaned and standardized feature dataset for subsequent model training.

[0088] S3. Determine the weight of each characteristic variable based on the degree of influence of each characteristic variable on wind power output, multiply each characteristic variable by its corresponding weight to obtain a weighted characteristic value, and calculate the weight values ​​of the geographical environment and climate respectively based on the weighted characteristic values; the weighted value expression of the geographical environment data in step S3 is:

[0089] W DL =W D +W H +W C ;

[0090] Where W DL Indicates the weight value of the geographical environment, WD represents the terrain weight, WH represents the altitude weight, and WC represents the surface roughness weight;

[0091] Among them, the weight value of the geographical environment is constant and is not affected by the four seasons;

[0092] The weighted value expression of the climate data is:

[0093] W QH =aW FS +aW FX +bW WD +bW SD +cW QY；

[0094] Where W QH The weight of climate data, W FS Indicates the wind speed weight value, W FX Indicates the wind direction weight value, W WD Represents the temperature weight value, W SD Represents the humidity weight value, WQY Represents the air pressure weight value, a, b, and c are constants, and the maximum value is 1.

[0095] In addition, due to the influence of the four seasons, the temperature, humidity, wind speed, wind direction and air pressure will change in different seasons. Therefore, the values ​​of constants a, b and c will change according to the actual situation in different seasons. Specifically:

[0096] The constant a is 0.7 in spring, 1 in summer, 0.5 in autumn, and 0.3 in winter;

[0097] The constant b is 0.5 in spring, 0.3 in summer, 0.5 in autumn, and 1 in winter;

[0098] The constant c is 0.7 in spring, 0.3 in summer, 0.7 in autumn, and 1 in winter.

[0099] For example, when predicting the power of a wind farm in spring, the weighted value W of the climate data is calculated. QH , the parameters are as follows:

[0100] Wind speed weight value W FS is 0.2, the wind direction weight value W FX is 0.3, the temperature weight value W WD is 0.1, the humidity weight value W SD is 0.2, the air pressure weight value W QY is 0.2;

[0101] Substitute these values ​​into the weighted value expression W of climate data QH =aW FS +aW FX +bW WD +bW SD +cW QY In the equation, we get:

[0102] W QH =0.7×0.2+0.7×0.3+0.5×0.1+0.5×0.2+0.7×0.2;

[0103] W QH =0.14+0.21+0.05+0.1+0.14;

[0104] W QH =0.64;

[0105] In the above example, the weighted value W of the climate data QH The final output is 0.64.

[0106] By dynamically adjusting the weights of climate factors (the values ​​of constants a, b, and c) based on seasonal variations, this approach can more accurately reflect the impact of seasonal climate conditions on wind power output. This dynamic adjustment mechanism makes the forecasting model more flexible and adaptable, better capturing the impact of seasonal climate change on wind farm performance.

[0107] By weighting the feature variables, we can ensure that the model pays more attention to those factors that have a greater impact on wind power output during training. This approach helps improve the model's prediction accuracy because it allows the model to adjust the contribution of each feature to the prediction results based on its importance.

[0108] By weighting the feature variables, the interference of unimportant features can be reduced, thereby enhancing the generalization ability of the model; this means that the model not only performs well on the training data, but also maintains good predictive performance on unseen test data;

[0109] In addition to climate factors, the model also considers the impact of geographical environmental factors (such as topography, altitude and surface roughness) on wind power output; this comprehensive consideration makes the prediction more comprehensive and accurate, as geographical environmental factors also play an important role in the performance of wind farms.

[0110] S4. Use the weighted eigenvalues ​​as input to establish a wind farm power prediction model, use historical data to train the power prediction model, and use a portion of the data for cross-validation to evaluate the accuracy and generalization ability of the model; Figure 2 As shown, the specific process of step S4 is:

[0111] S41. Obtain historical wind speed, wind direction, temperature, humidity, and air pressure data, as well as corresponding wind power output data, and determine the main factors affecting wind power;

[0112] S42. Weight the features according to their importance to obtain weighted feature values, and form a data set. The data set is divided into a training set and a test set, with the ratio of 70% training set and 30% test set allocated.

[0113] S43. Select a model, train the model using the training set data, and analyze the model's performance on the training set and test set to check whether there is overfitting or underfitting;

[0114] S44. Select the best performing model as the final model.

[0115] Through the above steps, a wind farm power prediction model with weighted features was established, and the accuracy and generalization ability of the model were evaluated through cross-validation, ensuring that our model can effectively learn from historical data and provide reliable results in future forecasts.

[0116] S5. Input the real-time or future geographical environment data and climate data into the trained power prediction model to obtain the power prediction results of the wind farm station, and display the predicted wind farm station power output results to the user in the form of a graph or table; Figure 3 As shown, step S5 is specifically as follows:

[0117] S51. Obtain geographic environmental data such as wind speed, wind direction, temperature, humidity, and air pressure for the current or future period, and clean and format the data to ensure that it meets the model input requirements;

[0118] S52. Using the previously determined weighting coefficient, perform weighted processing on each eigenvalue of the real-time data to obtain a weighted eigenvalue, and integrate all the weighted eigenvalues ​​into a eigenvector as the input of the model;

[0119] S53, loading the trained wind power prediction model into the system, and inputting the weighted feature vector into the model to obtain the power prediction result of the wind farm station;

[0120] S54. Use a chart to show how the predicted wind power changes over time, and organize the predicted wind power and the corresponding time points into a table to facilitate users to view specific predicted values.

[0121] In summary, compared with the prior art, the present invention has the following effects:

[0122] By comprehensively considering geographical and climatic factors, this method can more accurately predict the power output of wind farms. Using weighted eigenvalues ​​and weight calculations, the model can better capture the impact of different factors on wind power, thereby improving prediction accuracy. It also considers the impact of seasonal climate change on wind power. By adjusting the values ​​of constants a, b, and c, the model can adapt to seasonal changes in temperature, humidity, wind speed, wind direction, and air pressure. Seasonal adjustments improve the model's prediction performance across different seasons.

[0123] By processing outliers and standardizing data, this method eliminates the influence of different dimensions and ranges, making all features contribute equally to model training, improving the stability and reliability of the model. It combines historical data and real-time data, uses common statistical methods and machine learning models for prediction, and is easy to implement and apply. At the same time, by displaying the prediction results in charts, users can intuitively understand the power prediction status of wind farms in different time periods.

[0124] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting wind farm power based on geographical environment and climate, characterized in that: The steps include: S1. Collect historical power output data, geographical environment data, and climate data of the wind farm, wherein the geographical environment data includes topography, altitude, and surface roughness; and the climate data includes wind speed, wind direction, temperature, humidity, and air pressure; S2. Based on historical data, select geographical and climatic factors that have a significant impact on wind farm power output as characteristic variables, and clean and standardize the collected data to eliminate the impact of different dimensions; S3. Based on the degree of influence of the characteristic variables of terrain, altitude, surface roughness, wind speed, wind direction, temperature, humidity and air pressure on wind power output, obtain the characteristic value of each of terrain, altitude, surface roughness, wind speed, wind direction, temperature, humidity and air pressure, and calculate the weight value of the geographical environment and climate factors respectively based on each characteristic value; specifically: The weighted value expression of geographic environment data is: W DL =W D +W H +W C Where W DL Represents the weight value of the geographical environment, WD represents the terrain weight, WH represents the altitude weight, and WC represents the surface roughness weight. The weight value of the geographical environment is constant and is not affected by the four seasons. The weighted value expression of climate data is: W QH =aW FS +aW FX +bW WD +bW SD +cW QY Where W QH Represents the weight value of climate data, W FS Indicates the wind speed weight value, W FX Indicates the wind direction weight value, W WD Represents the temperature weight value, W SD Represents the humidity weight value, W QY Indicates the air pressure weight value, a, b, and c are constants, and the maximum value is 1; S4. Using the weighted values ​​of geographic environment data and climate data as input, establish a wind farm power prediction model, use historical data to train the power prediction model, and use a portion of the data for cross-validation to evaluate the accuracy and generalization ability of the model; S5. Input the real-time or future geographic environment data and climate data into the trained power prediction model to obtain the power prediction result of the wind farm station, and display the predicted wind farm station power output result to the user in the form of a graph or table.

2. A method for predicting wind farm power based on geographical environment and climate according to claim 1, characterized in that: In step S1, the historical power output data of the wind farm is obtained from the operation records of the wind farm, the geographical environment data is obtained through a geographic information system or on-site measurements, and the climate data is obtained from a meteorological station or a numerical weather prediction model; The step S1 also includes pre-processing the collected data, and using a method combining double quartile and clustering to eliminate abnormal data during power curtailment periods and when wind turbines fail.

3. The method for predicting wind farm power based on geographical environment and climate according to claim 1, characterized in that: In step S2, the geographical environment factors and climate factors that have a greater impact on the power output of the wind farm are identified and selected as the characteristic variables of the model, specifically: Determine the factors that affect wind power, calculate the correlation coefficient between each factor and wind power output, and select the one with the highest correlation as the feature. The expression is: ； Where r is the correlation coefficient, x and y are the data sets of two variables, and n is the number of samples; The standardization process is achieved by processing outliers, specifically: Use the interquartile range to identify outliers, as expressed in: ； ； Where Q1 and Q3 are the first and third quartiles of the data set, respectively, and IQR is the difference between the third quartile and the first quartile; The definition of outliers is based on IQR. Any data point smaller than Q1−1.5×IQR or larger than Q3+1.5×IQR is considered an outlier.

4. A method for predicting wind farm power based on geographical environment and climate according to claim 3, characterized in that: Data normalization is used to eliminate the effects of different dimensions and ranges so that all features contribute equally to model training. The expression is: ； Where z is the standardized value, x is the original data, μ is the mean of the data, and σ is the standard deviation of the data.

Images