Offshore wind resource prediction method and apparatus, and computer device and storage medium

Abstract

The present application relates to the technical field of offshore wind prediction. Disclosed are an offshore wind resource prediction method and apparatus, and a computer device and a storage medium. In the present application, a first offshore meteorological field data set is converted into a second offshore meteorological field data set, which can be identified in a weather research and forecast mode. Optionally, on the basis of a first anemometer tower wind-speed data set and a second anemometer tower wind-speed data set, advanced extreme gradient boosting is used to construct a target offshore wind resource prediction model, so that the accuracy of offshore wind resource prediction can be improved, a data fusion process can be simplified to a relatively great extent, and different pieces of data can also be compensated for on the basis of the comprehensive use of multi-source research and forecast data of a global ensemble forecast system, thereby giving full play to the advantages of various types of data, and thus a prediction result is more reliable.

Description

Offshore wind resource prediction method, device, computer equipment and storage medium Technical Field

[0001] The present application relates to the technical field of offshore wind prediction, and in particular to an offshore wind resource prediction method, apparatus, computer equipment, and storage medium. Background Art

[0002] The Weather Research and Forecast (WRF) model is a numerical model used for meteorological forecasting and research. It is a high-resolution, non-static, non-hierarchical, and non-uniform atmospheric numerical model. The WRF model simulates atmospheric phenomena by dividing the Earth's atmosphere into horizontal grids and vertical layers and solving atmospheric dynamics, thermodynamics, and hydrological cycle equations through discrete equations. This model is widely used in wind resource simulation and wind speed prediction.

[0003] In order to improve the accuracy and reliability of the model, a global combined forecast system is generally used to provide accurate and comprehensive meteorological environment driving data for the WRF model. The fusion of multi-source forecast data can maximize the advantages of each global ensemble forecast system, make up for the defects between different systems, and improve the accuracy, completeness and reliability of the forecast.

[0004] Numerous studies have shown that multi-model forecasting is significantly improved compared to a single forecasting system, but there are still some problems in the fusion process. For example, there are differences in the data formats and time resolutions of different forecasting systems, which require data format conversion and time alignment. At the same time, the physical processes and parameter settings of the forecast models of different systems are different, which often require parameter matching and data calibration. Secondly, for conflicts or inconsistencies between forecast data from different systems, if reasonable data processing and quality control are not performed, it will affect the prediction results.

[0005] Summary of the Invention

[0006] In view of this, the present application provides an offshore wind resource prediction method, apparatus, computer equipment and storage medium to solve the problem that problems exist in the existing multi-mode forecast fusion process leading to poor prediction results.

[0007] In a first aspect, the present application provides a method for predicting offshore wind resources, the method comprising:

[0008] A first offshore meteorological field dataset and a first wind tower wind speed dataset are obtained based on a global combined forecast system; a weather forecast model is driven by the global combined forecast system to convert the first offshore meteorological field dataset into a second offshore meteorological field dataset based on the weather forecast model; the second offshore meteorological field dataset is interpolated to obtain a second wind tower wind speed dataset; a target offshore wind resource prediction model is established based on the first wind tower wind speed dataset and the second wind tower wind speed dataset using a distributed gradient enhancement library; a third offshore meteorological field dataset and a third wind tower wind speed dataset of the wind field in the to-be-predicted area are obtained, and the third offshore meteorological field dataset and the third wind tower wind speed dataset are input into the target offshore wind resource prediction model to obtain a wind resource prediction result for the wind field in the to-be-predicted area.

[0009] The offshore wind resource prediction method provided in the present application converts a first offshore meteorological field dataset into a second offshore meteorological field dataset that can be recognized by a weather forecast model. Optionally, based on the first wind tower wind speed dataset and the second wind tower wind speed dataset, an advanced distributed gradient enhancement library is used to construct a target offshore wind resource prediction model. This can not only improve the accuracy of offshore wind resource prediction and greatly simplify the data fusion process, but also the comprehensive use of multi-source forecast data based on the global combined forecast system can make up for the deficiencies between different data, give full play to the advantages of various types of data, and make the prediction results more reliable.

[0010] In an optional embodiment, the global combined forecast system is used to drive the weather forecast model to convert the first offshore meteorological field dataset into a second offshore meteorological field dataset based on the weather forecast model, including: using the global combined forecast system to drive the weather forecast model to generate a target simulation area based on the weather forecast model; using a preset interpolation method to interpolate the first offshore meteorological field dataset into the target simulation area to obtain the second offshore meteorological field dataset.

[0011] The present application utilizes the global combined forecast system to drive the weather forecast model, which can convert the first offshore meteorological field data set into a second offshore meteorological field data set that can be recognized by the weather forecast model, thereby improving the accuracy of the data; at the same time, obtaining the first offshore meteorological field data set based on the global combined forecast system can make up for the deficiencies between different data, give full play to the advantages of various types of data, and make the prediction results more reliable.

[0012] In an optional embodiment, the global combined forecast system is used to drive a weather forecast model to generate a target simulation area based on the weather forecast model, including:

[0013] The global combined forecast system is used to drive a weather forecast model to generate an initial simulation area; the area length and area width of the initial simulation area are obtained; the area length is divided into multiple longitude segments, and the area width is divided into multiple latitude segments; based on the multiple longitude segments and the multiple latitude segments, the grid longitude and latitude of each grid point are determined, where the grid point is composed of the longitude segment and the latitude segment; based on the initial simulation area and the grid longitude and latitude of each grid point, a target simulation area based on the weather forecast model is generated.

[0014] This application determines the grid longitude and latitude of each grid point, so that the generated target simulation area can convert the input data into longitude and latitude data that can be recognized by the weather forecast model.

[0015] In an optional embodiment, based on the first wind tower wind speed dataset and the second wind tower wind speed dataset, a target offshore wind resource prediction model is established using a distributed gradient boosting library, including:

[0016] The first wind tower wind speed dataset and the second wind tower wind speed dataset are processed by a data processing method to obtain a first target wind tower wind speed dataset and a second target wind tower wind speed dataset; based on the first target wind tower wind speed dataset and the second target wind tower wind speed dataset, feature engineering processing and cross-validation methods are used to generate a target feature model; based on the second target wind tower wind speed dataset and the target feature model, an initial offshore wind resource prediction model is established using a distributed gradient boosting library; and the initial offshore wind resource prediction model is adjusted using the first wind tower wind speed dataset until the target offshore wind resource prediction model is obtained.

[0017] This application uses a data processing method to process the wind speed datasets from the first and second wind towers, eliminating the impact of random annual wind speed fluctuations. Furthermore, by constructing a model using a distributed gradient boosting library and adjusting the model using the wind speed dataset from the second wind tower, the prediction accuracy of the target offshore wind resource prediction model is improved.

[0018] In an optional embodiment, based on the second target wind tower wind speed dataset and the target characteristic model, an initial offshore wind resource prediction model is established using a distributed gradient boosting library, including:

[0019] A model feature map of the target feature model is obtained; the model feature map is input into a distributed gradient enhancement library to obtain a plurality of marking points corresponding to the model feature map; a construction tree is obtained based on the plurality of marking points, and an initial offshore wind resource prediction model is established based on the construction tree.

[0020] This application uses multiple marked points corresponding to the model feature map to obtain a construction tree, and uses the construction tree to establish an initial offshore wind resource prediction model, which improves the prediction accuracy of the model to a certain extent.

[0021] In an optional embodiment, the method further includes:

[0022] The mean absolute error, mean square error and root mean square error of the target offshore wind resource prediction model are calculated; the target offshore wind resource prediction model is evaluated based on the mean absolute error, mean square error and root mean square error to obtain an evaluation result.

[0023] This application uses mean absolute error, mean square error and root mean square error to evaluate the target offshore wind resource prediction model, which can obtain the prediction performance of the target offshore wind resource prediction model and further provide data support for the prediction of offshore wind resources.

[0024] In an optional embodiment, after using the global combined forecast system to drive a weather forecast model and generate a target simulation area based on the weather forecast model, the method further includes:

[0025] The target simulation area is divided into multiple sub-areas; when computing resources are configured in the sub-areas, the sub-areas are set as nested areas.

[0026] This application uses sub-areas that require computing resource configuration as nested areas, which can achieve the use of high-resolution grids only within the nested areas, and further play a role in balancing the timeliness of calculations and the simulation accuracy of weather forecast models.

[0027] In a second aspect, the present application provides an offshore wind resource prediction device, the offshore wind resource prediction device comprising:

[0028] A data acquisition module is used to acquire a first offshore meteorological field data set and a first wind tower wind speed data set based on a global combined forecast system; a conversion module is used to use the global combined forecast system to drive a weather forecast model to convert the first offshore meteorological field data set into a second offshore meteorological field data set based on the weather forecast model; a processing module is used to interpolate the second offshore meteorological field data set to obtain a second wind tower wind speed data set; an establishment module is used to establish a target offshore wind resource prediction model based on the first wind tower wind speed data set and the second wind tower wind speed data set using a distributed gradient enhancement library; an acquisition and input module is used to acquire a third offshore meteorological field data set and a third wind tower wind speed data set of the wind field in the area to be predicted, and input the third offshore meteorological field data set and the third wind tower wind speed data set into the target offshore wind resource prediction model to obtain a wind resource prediction result for the wind field in the area to be predicted.

[0029] In a third aspect, the present application provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to execute the offshore wind resource prediction method of the first aspect or any corresponding embodiment thereof.

[0030] In a fourth aspect, the present application provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the offshore wind resource prediction method of the above-mentioned first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0032] FIG1 is a flow chart of a method for predicting offshore wind resources according to an embodiment of the present application;

[0033] FIG2 is a flow chart of another offshore wind resource prediction method according to an embodiment of the present application;

[0034] FIG3 is a flow chart of another offshore wind resource prediction method according to an embodiment of the present application;

[0035] FIG4 is a flow chart of a method for predicting offshore wind resources based on multi-source forecast data according to an embodiment of the present application;

[0036] FIG5 is a structural block diagram of an offshore wind resource prediction device according to an embodiment of the present application;

[0037] FIG6 is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0038] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of this application.

[0039] The embodiment of the present application provides a method for offshore wind resource prediction, which establishes a target offshore wind resource prediction model through the comprehensive use of multi-source forecast data and the use of a distributed gradient enhancement library, so as to improve the accuracy of offshore wind resource prediction, make up for the deficiencies between different data, give full play to the advantages of various types of data, and make the prediction results more reliable.

[0040] According to an embodiment of the present application, an embodiment of an offshore wind resource prediction method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0041] In this embodiment, a method for predicting offshore wind resources is provided. FIG1 is a flow chart of the method for predicting offshore wind resources according to an embodiment of the present application. As shown in FIG1 , the process includes the following steps:

[0042] Step S101: obtaining a first offshore meteorological field dataset and a first wind speed dataset from a wind tower based on a global combined forecast system.

[0043] Among them, the global combined forecast system can include: the National Centers for Environmental Prediction (NCEP) of the United States, the European Center for Medium-Range Weather Forecasts (ECMWF), the Japan Meteorological Agency (JMA) and the Meteorological Service of Canada (MSC), etc.

[0044] Optionally, each of the above-mentioned global combined forecast systems is configured with a corresponding database, and meteorological data is stored in the database.

[0045] The first offshore meteorological field data set may include information such as geopotential height, horizontal wind speed, temperature, air pressure, humidity, soil moisture content, and soil temperature of the offshore wind field.

[0046] The first wind tower wind speed dataset is actual measurement data of a wind tower installed in an offshore wind farm, and may include measurement values ​​corresponding to different heights, wind speeds, wind directions, surface temperatures, humidity, and air pressure.

[0047] The wind speed dataset of the first wind tower is obtained as follows:

[0048] A measurement period is set, denoted as T, and actual measurement data within this measurement period is obtained. The amount of abnormal data in the actual measurement data, denoted as R, is also obtained. When the amount of abnormal data R exceeds the set period change value, the measurement period T is extended, and real-time monitoring of the measured data within the extended measurement period continues until the corresponding first wind tower wind speed dataset is obtained. This embodiment, by setting a measurement period and determining whether to extend the measurement period based on the amount of abnormal data, achieves the goal of eliminating the impact of random annual fluctuations in wind speed.

[0049] Optionally, by aggregating the meteorological data in the database corresponding to each global combined forecast system, multi-source meteorological data, namely the first marine meteorological field data set, can be generated, which can make up for the deficiencies between different data, give full play to the advantages of various types of data, and make the prediction results more reliable.

[0050] Step S102: using the global combined forecast system to drive the weather forecast model, the first offshore meteorological field dataset is converted into a second offshore meteorological field dataset based on the weather forecast model.

[0051] The weather forecast model (WRF model) is provided with a pre-processing process, through which data generated based on the global combined forecast system can be converted into a data format that can be directly recognized and read by the WRF model.

[0052] Specifically, when the WRF model is driven by the global combined forecast system, the pre-processing process converts the first marine meteorological field dataset into a second marine meteorological field dataset that can be directly recognized and read by the WRF model.

[0053] Step S103: interpolate the second offshore meteorological field dataset to obtain a second wind speed dataset of a wind tower.

[0054] First, the data in the second marine meteorological field data set that does not meet the pre-set data standards are filtered out;

[0055] Then, the filtered and eliminated second offshore meteorological field dataset is interpolated to the corresponding wind tower, and a refined dataset including Data-NCEP, Data-ECMWF, Data-JMA, and Data-MSC is generated, namely the wind speed dataset of the second wind tower.

[0056] Step S104 : establishing a target offshore wind resource prediction model using a distributed gradient boosting library based on the first wind tower wind speed dataset and the second wind tower wind speed dataset.

[0057] Among them, the distributed gradient boosting library (eXtreme Gradient Boosting, XGBoost) is a machine learning function library that focuses on the gradient boosting algorithm and can be used for classification, regression and sorting.

[0058] In this embodiment, the wind speed dataset of the first wind tower is used as the test set, and the wind speed dataset of the second wind tower is used as the training set. A regression model is constructed based on the XGBoost library, namely the target offshore wind resource prediction model to realize the prediction of offshore wind resources, which improves the accuracy of offshore wind resource prediction and greatly simplifies the data fusion process.

[0059] Step S105 , obtaining a third offshore meteorological field dataset and a third wind tower wind speed dataset of the wind farm in the area to be predicted, and inputting the third offshore meteorological field dataset and the third wind tower wind speed dataset into the target offshore wind resource prediction model to obtain a wind resource prediction result of the wind farm in the area to be predicted.

[0060] Specifically, the third offshore meteorological field dataset and the third wind speed dataset of the wind field in the area to be predicted are input into the constructed target offshore wind resource prediction model, and the wind resource prediction result of the wind field in the area to be predicted can be output.

[0061] The offshore wind resource prediction method provided in this embodiment converts a first offshore meteorological field dataset into a second offshore meteorological field dataset that can be recognized by a weather forecast model. Furthermore, based on the first and second wind tower wind speed datasets, an advanced distributed gradient enhancement library is used to construct a target offshore wind resource prediction model. This not only improves the accuracy of offshore wind resource prediction and greatly simplifies the data fusion process, but also, the comprehensive use of multi-source forecast data based on the global combined forecast system can compensate for the deficiencies between different data types, give full play to the advantages of various types of data, and make the prediction results more reliable.

[0062] In this embodiment, a method for predicting offshore wind resources is provided. FIG2 is a flow chart of the method for predicting offshore wind resources according to an embodiment of the present application. As shown in FIG2 , the process includes the following steps:

[0063] Step S201: Obtain a first offshore meteorological field dataset and a first wind speed dataset from a wind tower based on a global combined forecast system. For details, please refer to step S101 of the embodiment shown in FIG1 , which will not be described in detail here.

[0064] Step S202: using the global combined forecast system to drive the weather forecast model, the first offshore meteorological field dataset is converted into a second offshore meteorological field dataset based on the weather forecast model.

[0065] Specifically, the above step S202 includes:

[0066] Step S2021: Utilize the global combined forecast system to drive the weather forecast model and generate a target simulation area based on the weather forecast model.

[0067] The target simulation area is used to convert the input data into a data format that can be directly recognized and read by the WRF model.

[0068] Specifically, the corresponding target simulation area can be obtained according to the pre-processing process of the WRF model.

[0069] Step S2022: interpolate the first offshore meteorological field data set into the target simulation area using a preset interpolation method to obtain a second offshore meteorological field data set.

[0070] Among them, different interpolation dimensions are set in the target simulation area, and the first marine meteorological field dataset has corresponding interpolation characteristic coefficients.

[0071] Specifically, after the first marine meteorological field data set is input into the target simulation area, the first marine meteorological field data set corresponding to the interpolation characteristic coefficient can be matched one by one with the interpolation dimension in the target simulation area according to the correspondence of the pre-set standard interpolation correspondence table. After the matching is successful, the data in the first marine meteorological field data set can be converted into the corresponding longitude and latitude, and converted into a second marine meteorological field data set in WRF format, and the WRF mode can directly recognize the second marine meteorological field data set.

[0072] Among them, the standard interpolation correspondence table records the correspondence between the interpolation characteristic coefficients and the interpolation dimensions.

[0073] In some optional implementations, the above step S2021 includes:

[0074] Step a1: Use the global combined forecast system to drive the weather forecast model and generate the initial simulation area.

[0075] Step a2: Obtain the area length and area width of the initial simulation area.

[0076] Step a3: Divide the area length into multiple longitude segments, and divide the area width into multiple latitude segments.

[0077] Step a4: determining the grid longitude and latitude of each grid point based on the multiple longitude segments and the multiple latitude segments.

[0078] Step a5: generating a target simulation area based on a weather forecast model based on the initial simulation area and the grid longitude and latitude of each grid point.

[0079] Specifically, when the global combined forecast system drives the WRF model, the corresponding offshore measurement area can be located, denoted as S 海Optionally, the offshore measurement area S 海 Marked as the initial simulation area, marked with S 模 .

[0080] Then, obtain the initial simulation area S 模 The area length and area width of are marked as w and h respectively.

[0081] Optionally, the area length w is divided into multiple longitude segments in equal proportion according to the longitude, and recorded as w1, w2, w3, ..., w N , where N is a natural number and N is an integer. Similarly, the area width h is divided into multiple latitude segments in proportion to the latitude, denoted as h1, h2, h3, ..., h M , where M is a natural number and M is an integer.

[0082] Optionally, obtain the grid point longitude and latitude.

[0083] Specifically, every two adjacent longitude segments and every two adjacent latitude segments form a square grid, and the intersection of the longitude segment and the latitude segment is called a grid point, that is, the initial simulation area consists of multiple square grids, and each square grid includes multiple grid points.

[0084] Optionally, the grid point is recorded as X, X = <longitude, latitude>, then the grid longitude and latitude corresponding to each grid point is X =<w,h> , the value range of w is w1-w N , the value range of h is h1-h M .

[0085] Finally, based on the initial simulation area and the grid longitude and latitude of each grid point, the corresponding target simulation area based on the weather forecast model can be determined.

[0086] In some optional implementations, after the above step a1, the above step S2021 further includes:

[0087] Step a6: Divide the target simulation area into multiple sub-areas.

[0088] Step a7: When computing resources are configured in the sub-area, the sub-area is set as a nested area.

[0089] Specifically, multiple longitude segments and multiple latitude segments are aggregated, and the longitude segments and latitude segments are crossed to form an equal longitude and latitude grid, and the simulation area is covered by the entire equal longitude and latitude grid.

[0090] Alternatively, the equal longitude and latitude grid does not have the ability to be locally encrypted, and the equal longitude and latitude grid is a high-resolution grid, which consumes a large amount of computing resources.

[0091] Therefore, the target simulation area is divided into multiple sub-areas, each of which can be used to configure computing resources. When configuring computing resources in a sub-area, the sub-area's location number is obtained and the sub-area with the location number is set as a nested area.

[0092] In this embodiment, the sub-area that requires computing resource configuration is used as a nested area, and high-resolution grids are only used in the nested area. The part that belongs to the target simulation area but does not belong to the sub-area where computing resources are being configured is divided into a parent area. The parent area provides coarse resolution, which plays a role in balancing calculation timeliness and simulation accuracy.

[0093] Optionally, the second marine meteorological field dataset obtained through the above steps can be called downscaled data because of its low spatial resolution.

[0094] Step S203: interpolate the second offshore meteorological field dataset to obtain a second wind tower wind speed dataset. For details, please refer to step S103 of the embodiment shown in FIG1 , which will not be described in detail here.

[0095] Step S204: Based on the first and second wind tower wind speed datasets, a target offshore wind resource prediction model is established using a distributed gradient boosting library. For details, please refer to step S104 of the embodiment shown in FIG1 , which will not be described in detail here.

[0096] Step S205: Obtain a third offshore meteorological dataset and a third wind tower wind speed dataset for the wind farm in the target area. These datasets are then input into the target offshore wind resource prediction model to obtain a wind resource prediction result for the wind farm in the target area. For details, refer to step S105 in the embodiment shown in FIG1 , and will not be further described here.

[0097] Step S206 , calculating the mean absolute error, mean square error, and root mean square error of the target offshore wind resource prediction model.

[0098] Specifically, the mean absolute error is calculated as shown in the following equation (1): MAE = (1 / n) × ∑ |y_pred - y_true| (1)

[0099] Where: MAE represents the mean absolute error; n represents the number of samples of the target offshore wind resource prediction model; y_pred represents the predicted value of the target offshore wind resource prediction model; y_true represents the true target value of the target offshore wind resource prediction model.

[0100] The mean square error is the average of the squares of the errors between the predicted values ​​and the true values. The prediction error of the model is measured by calculating the square of the difference between the predicted values ​​and the true values, as shown in the following equation (2): MSE = (1 / n) × ∑ (y_pred - y_true) 2 (2)

[0101] Where: MSE represents mean square error.

[0102] The root mean square error is the square root of the mean square error, which is used to measure the average error between the predicted value and the true value, as shown in the following relationship (3): RMSE = sqrt(MSE)

[0103] Where: RMSE represents the root mean square error.

[0104] Step S207 : evaluating the target offshore wind resource prediction model based on the mean absolute error, the mean square error, and the root mean square error to obtain an evaluation result.

[0105] Specifically, the smaller the values ​​of mean absolute error (MAE), mean square error (MSE), and root mean square error (RMSE), the smaller the difference between the prediction results of the target offshore wind resource prediction model and the true value, and the better the prediction performance of the model.

[0106] The offshore wind resource prediction method provided in this embodiment, by determining the grid longitude and latitude of each grid point, can enable the generated target simulation area to convert the input first offshore meteorological field dataset into a second offshore meteorological field dataset that can be recognized by the weather forecast model, thereby improving data accuracy. Furthermore, obtaining the first offshore meteorological field dataset based on the global combined forecast system can compensate for the deficiencies between different data types, fully leveraging the advantages of various data types and making the prediction results more reliable. Optionally, the mean absolute error, mean square error, and root mean square error are used to evaluate the model prediction performance of the target offshore wind resource prediction model, providing data support for offshore wind resource prediction.

[0107] In this embodiment, a method for predicting offshore wind resources is provided. FIG3 is a flow chart of the method for predicting offshore wind resources according to an embodiment of the present application. As shown in FIG3 , the process includes the following steps:

[0108] Step S301: Obtain a first offshore meteorological field dataset and a first wind speed dataset from a wind tower based on a global combined forecast system. For details, please refer to step S101 of the embodiment shown in FIG1 , which will not be described in detail here.

[0109] Step S302: Using the global combined forecast system to drive the weather forecast model, the first offshore meteorological field dataset is converted into a second offshore meteorological field dataset based on the weather forecast model. For details, please refer to step S202 of the embodiment shown in FIG2 , which will not be described in detail here.

[0110] Step S303: interpolate the second offshore meteorological field dataset to obtain a second wind tower wind speed dataset. For details, please refer to step S103 of the embodiment shown in FIG1 , which will not be described in detail here.

[0111] Step S304 : establishing a target offshore wind resource prediction model using a distributed gradient boosting library based on the first wind tower wind speed dataset and the second wind tower wind speed dataset.

[0112] Specifically, the above step S304 includes:

[0113] Step S3041: Process the first wind tower wind speed dataset and the second wind tower wind speed dataset using a data processing method to obtain a first target wind tower wind speed dataset and a second target wind tower wind speed dataset.

[0114] Among them, data processing methods can include data evaluation, missing value processing, duplicate value processing, outlier processing and variable conversion.

[0115] (1) Data evaluation: The integrity of the test set and training set is checked through data evaluation to filter out missing values, duplicate values, and outliers in the test set and training set.

[0116] (2) Missing value processing: Missing value processing is used to identify the content of missing values ​​and the row and column positions of missing values, and to perform supplementary operations on missing values.

[0117] (3) Duplicate value processing: Duplicate data in the test set (wind speed data set of the second wind tower) and the training set (wind speed data set of the first wind tower) are detected through duplicate value processing. The data in the test set and the training set are traversed in turn. The first time the data is traversed is marked as A. The second and subsequent times the data is traversed, it is marked as A`. The number of times the data is obtained after the traversal is completed is obtained, recorded as Y, and Y-1 data are eliminated.

[0118] (4) Outlier processing: Outlier processing is used to detect abnormal data values ​​in the training set and test set, and to retain, delete, and modify the abnormal data values.

[0119] (5) Variable conversion: The training set and the data set are converted into a modeling format. The converted modeling format can be used to build a data model.

[0120] Step S3042 : Based on the first target wind tower wind speed dataset and the second target wind tower wind speed dataset, a target feature model is generated by subjecting the dataset to feature engineering and cross-validation.

[0121] Among them, feature engineering processing can include feature extraction, feature conversion and feature construction.

[0122] Specifically, after the modeling format is obtained, feature extraction, feature conversion, and feature construction are performed on the first target wind tower wind speed dataset and the second target wind tower wind speed dataset corresponding to the modeling format:

[0123] First, multiple valid feature points and invalid noise points in the wind speed dataset of the first target wind tower and the wind speed dataset of the second target wind tower are obtained through feature extraction, and the invalid noise points are eliminated.

[0124] Secondly, the acquired valid feature points are converted into marking points of feature map information through feature conversion, and multiple marking points are aggregated to generate the model feature map.

[0125] Finally, the model feature map is modeled into the corresponding feature model through feature construction.

[0126] Optionally, in the process of constructing the feature model, cross-validation is also required on the second target wind tower wind speed dataset (training set):

[0127] First, the K-fold cross-validation method is used to divide the training set into K equal subset folds, traverse the K subset folds, and take each subset fold being traversed as the validation set in turn, and take the other K-1 subset folds as sub-training sets.

[0128] Secondly, the feature model is trained K-1 times using the sub-training set. During each K-1 training, the validation set is used to evaluate the performance of the feature model. The training of the feature model using the sub-training set and the evaluation of the feature model performance using the test set are repeated until each subset fold has served as a validation set once and the optimal model parameters are obtained.

[0129] Finally, based on the optimal model parameters and the corresponding feature model constructed through feature engineering, the corresponding target feature model can be determined.

[0130] Step S3043 : Based on the second target wind tower wind speed dataset and the target characteristic model, an initial offshore wind resource prediction model is established using a distributed gradient boosting library.

[0131] Specifically, the target feature model is trained using the XGBoost library and the second target wind tower wind speed dataset to obtain the corresponding initial offshore wind resource prediction model.

[0132] Step S3044: Use the first wind tower wind speed data set to adjust the initial offshore wind resource prediction model until a target offshore wind resource prediction model is obtained.

[0133] Specifically, the model parameters and hyperparameters of the initial offshore wind resource prediction model are adjusted using the wind speed dataset of the first wind tower until the model parameters and hyperparameters are optimal, and the corresponding target offshore wind resource prediction model is output.

[0134] Among them, the model parameters can select a set of initial parameters as a benchmark; hyperparameters can include learning rate, number of constructed trees, maximum depth of constructed trees, and column sampling rate.

[0135] Optionally, grid search or random search can be used for hyperparameter tuning: any two different hyperparameters are selected for combination, and the performance of the initial offshore wind resource prediction model when using different hyperparameters is evaluated. The performance of the initial offshore wind resource prediction model is measured by the performance coefficient, and the hyperparameter combination with the largest performance coefficient is selected as the optimal hyperparameter combination.

[0136] Optionally, hyperparameters can be kept constant while model parameters are adjusted for optimization, so that the performance of the initial offshore wind resource prediction model is continuously improved. Model parameter optimization can be performed by adjusting the learning rate to different set learning coefficient levels, the maximum depth of the tree to different set tree depth levels, and increasing the column sampling rate.

[0137] In some optional implementations, step S3043 includes:

[0138] Step b1: Obtain a model feature graph of the target feature model.

[0139] In step b2, the model feature map is input into the distributed gradient enhancement library to obtain multiple marker points corresponding to the model feature map.

[0140] Step b3: obtaining a constructed tree based on the multiple marker points, and establishing an initial offshore wind resource prediction model based on the constructed tree.

[0141] Specifically, the model feature graph of the obtained target feature model is input into the XGBoost library, and the XGBoost library can obtain multiple marking points included in the feature graph information deconstructed from the model feature graph.

[0142] Then, get the build tree by the marked point.

[0143] Optionally, the constructed tree includes multiple tree nodes, different tree nodes have corresponding split contributions, the split contributions have corresponding feature importances, and the feature importances have different importance scores. In this embodiment, the importance score with the highest value is selected as the final feature subset, the feature subset is obtained, and an initial offshore wind resource prediction model is constructed.

[0144] Among them, the importance score can be obtained through the feature_import attribute of the XGBoost library.

[0145] Step S305: Obtain a third offshore meteorological dataset and a third wind tower wind speed dataset for the wind farm in the target area. These datasets are then input into the target offshore wind resource prediction model to obtain a wind resource prediction result for the wind farm in the target area. For details, refer to step S105 in the embodiment shown in FIG1 , and will not be further described here.

[0146] The offshore wind resource prediction method provided in this embodiment uses a data processing method to process the first and second wind tower wind speed datasets, thereby eliminating the impact of random annual wind speed fluctuations. Furthermore, a tree is constructed using multiple marker points corresponding to the model feature map obtained from a distributed gradient boosting library. The first wind tower wind speed dataset and the tree are then used to establish a target offshore wind resource prediction model, thereby improving the prediction accuracy of the target offshore wind resource prediction model.

[0147] In one example, a method for predicting offshore wind resources based on multi-source forecast data is provided, as shown in FIG4 , including the following steps:

[0148] Step S1: Connect to the global ensemble forecast system for data-driven WRF model, define the simulation area and nested areas, calculate the grid longitude and latitude, obtain land surface data and meteorological field data and interpolate them into the simulation area;

[0149] Step S2: Obtain meteorological field data of the target area wind field, interpolate the downscaled data generated by the WRF model to the wind tower to obtain refined wind speed data, obtain the measured data from the wind tower, use the refined wind speed data and the measured data as the training set and test set respectively, perform data preprocessing and feature engineering operations, and cross-validate the training set to select the optimal model parameters;

[0150] Step S3: Use the XGBoost library to build a regression model, train the regression model using the training set, and adjust the model parameters and hyperparameters of the regression model based on the test set. Calculate the mean absolute error, mean square error, and root mean square error of the regression model to evaluate the prediction performance of the regression model, and use the trained regression model to predict the wind speed at the wind farm location.

[0151] Optionally, the definition of simulation regions and nested regions includes:

[0152] The offshore measurement area located by the data-driven WRF model is obtained and marked as a simulation area. The length and width of the offshore measurement area are obtained and divided into several longitude and latitude segments, respectively. The longitude and latitude segments intersect to form an equal longitude and latitude grid. The simulation area is divided into several sub-areas, each of which allows for computing resource configuration. When computing resource configuration is performed on a sub-area, the sub-area is set as a nested area.

[0153] Optionally, the process of obtaining the grid point longitude and latitude includes:

[0154] Obtain several square grids consisting of every two adjacent longitude segments and every two adjacent latitude segments in the simulation area. The intersection of the longitude segments and the latitude segments constitutes a grid point. The grid point has a corresponding longitude and latitude. The longitude and longitude of the grid point of the entire simulation area are obtained according to the longitude and latitude of each grid point.

[0155] Optionally, the interpolation process includes:

[0156] After obtaining the land surface data and meteorological field data, the simulation area is divided into a land surface layer and a meteorological field layer. The land surface data and the meteorological field data have corresponding interpolation characteristic coefficients, and the land surface layer and the meteorological field layer are set with different interpolation dimensions. When the land surface data and the meteorological field data are input into the simulation area, the land surface data and meteorological field data associated with the interpolation characteristic coefficients are matched with the interpolation dimensions corresponding to the land surface layer and the meteorological field layer.

[0157] Optionally, the process of generating the training set and the test set includes:

[0158] The wind field area where the wind speed is to be detected is set as the target area, the meteorological field data of the target area is obtained, the downscaled data is interpolated to the set wind measurement tower, and the refined data is generated after the downscaled data is interpolated to the wind measurement tower. The measured data actually measured at the wind measurement tower is obtained, and the obtained refined data is used as the training set, and the obtained measured data is used as the test set.

[0159] Optionally, the process of data preprocessing and feature engineering operations includes:

[0160] Data preprocessing includes data evaluation, missing value processing, duplicate value processing, outlier processing and variable conversion; data evaluation is used to screen out missing values, duplicate values ​​and outliers in the test set and training set; missing value processing identifies the content of missing values ​​and the row and column position of the missing values ​​and supplements the missing values; duplicate value processing is used to detect duplicate data in the training set and test set, and remove the duplicate data after traversal; outlier processing is used to detect abnormal data values ​​in the training set and test set, and retain, delete and modify the abnormal data values; variable conversion is used to convert the modeling format of the training set and data set; feature extraction is performed on the training set and data set corresponding to the modeling format to obtain several valid feature points, and the obtained valid feature points are converted into marking points of feature map information through feature conversion, and several marking points are summarized to generate a model feature map, and the model feature map is modeled into a corresponding feature model through feature construction.

[0161] Optionally, the process of cross-validation on the training set includes:

[0162] After obtaining the training set, divide the training set into K equal subset folds, traverse the K subset folds, and use each subset fold being traversed as a validation set in turn, and use the other K-1 subset folds as sub-training sets; use the sub-training set to train the feature model K-1 times, and use the validation set to evaluate the performance of the feature model each time. Repeat the training of the feature model with the sub-training set and the evaluation of the feature model performance with the test set until each subset fold has served as a validation set once. After cross-validation, the optimal model parameters are generated.

[0163] Optionally, the process of building a regression model and tuning model parameters and hyperparameters includes:

[0164] Obtain the model feature graph and input the model feature graph into the XGBoost library. The XGBoost library obtains several marking points included in the feature graph information deconstructed by the model feature graph, and obtains the construction tree through the marking points. The construction tree includes several tree nodes. Different tree nodes have corresponding split contributions. The split contributions have different feature importances. The feature importances have different importance scores. The one with the highest importance score is selected as the feature subset, and a regression model is constructed based on the feature subset. The regression model has corresponding model parameters and hyperparameters. The model parameters and hyperparameters are passed to the regression model, and the regression model is trained using the training set. The model parameters and hyperparameters are tuned according to the performance of the test set.

[0165] Optionally, the process of calculating the mean absolute error, mean square error and root mean square error of the regression model refers to the above-mentioned equations (1) to (3).

[0166] Compared with the existing technology, the offshore wind resource prediction method based on multi-source forecast data provided in this example has the following benefits:

[0167] 1. Based on the WRF model prediction results of the target area wind speed and measured data, the advanced XGBoost regression fitting algorithm is used for fitting and prediction. This not only improves the accuracy of offshore wind resource prediction and greatly simplifies the data fusion process, but also the comprehensive use of multi-source forecast data can make up for the deficiencies between different data, give full play to the advantages of various types of data, and make the prediction results more reliable.

[0168] 2. The equal longitude and latitude grid does not have the ability to be locally encrypted, and the equal longitude and latitude grid is a high-resolution grid. The high-resolution grid will consume a lot of computing resources. The sub-area that requires computing resource configuration is used as a nested area, and the high-resolution grid is only used in the nested area. The part that belongs to the simulation area but does not belong to the sub-area where the computing resources are being configured is divided into the parent area. The parent area provides coarse resolution, which plays a role in balancing the timeliness of calculation and simulation accuracy. By setting the measurement period and deciding whether to extend the measurement period based on the amount of abnormal data, the purpose of eliminating the impact of the randomness of the annual fluctuation of wind speed is achieved.

[0169] This embodiment also provides an offshore wind resource prediction device for implementing the above-described embodiments and optional implementations. Details already described are omitted for clarity. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. While the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0170] This embodiment provides an offshore wind resource prediction device, as shown in FIG5 , comprising:

[0171] The data acquisition module 501 is configured to acquire a first offshore meteorological field dataset and a first wind speed dataset from a wind tower based on a global combined forecast system.

[0172] The conversion module 502 is configured to utilize the global combined forecast system to drive the weather forecast model and convert the first offshore meteorological field dataset into a second offshore meteorological field dataset based on the weather forecast model.

[0173] The processing module 503 is configured to perform interpolation processing on the second offshore meteorological field dataset to obtain a second wind speed dataset of a wind tower.

[0174] The establishing module 504 is configured to establish a target offshore wind resource prediction model using a distributed gradient boosting library based on the first wind tower wind speed dataset and the second wind tower wind speed dataset.

[0175] The acquisition and input module 505 is used to obtain the third offshore meteorological field dataset and the third wind speed dataset of the wind farm in the area to be predicted, and input the second offshore meteorological field dataset and the third wind speed dataset of the wind farm into the target offshore wind resource prediction model to obtain the wind resource prediction result of the wind farm in the area to be predicted.

[0176] In some optional implementations, the conversion module 502 includes:

[0177] The generation unit is used to drive the weather forecast model by using the global combined forecast system to generate a target simulation area based on the weather forecast model.

[0178] The interpolation unit is used to interpolate the first offshore meteorological field data set into the target simulation area using a preset interpolation method to obtain a second offshore meteorological field data set.

[0179] In some optional embodiments, the generating unit includes:

[0180] The first generating subunit is used to drive the weather forecast model using the global combined forecast system to generate an initial simulation area.

[0181] The first acquisition subunit is used to acquire the area length and area width of the initial simulation area.

[0182] The division subunit is used to divide the area length into multiple longitude segments and divide the area width into multiple latitude segments.

[0183] The determination subunit is used to determine the grid longitude and latitude of each grid point based on multiple longitude segments and multiple latitude segments, where the grid point is composed of a longitude segment and a latitude segment.

[0184] The second generating subunit is used to generate a target simulation area based on a weather forecast model based on the initial simulation area and the grid longitude and latitude of each grid point.

[0185] In some optional implementations, the establishing module 504 includes:

[0186] The first processing unit is configured to process the first wind speed data set and the second wind speed data set using a data processing method to obtain a first target wind speed data set and a second target wind speed data set.

[0187] The second processing unit is configured to generate a target feature model based on the first target wind speed dataset and the second target wind speed dataset through feature engineering and cross-validation.

[0188] The unit is used to establish an initial offshore wind resource prediction model based on the second target wind tower wind speed data set and the target characteristic model using a distributed gradient enhancement library.

[0189] The adjustment unit is configured to adjust the initial offshore wind resource prediction model using the first wind speed data set of the wind measurement tower until a target offshore wind resource prediction model is obtained.

[0190] In some optional embodiments, the establishing unit includes:

[0191] The second acquisition subunit is used to acquire a model feature map of the target feature model.

[0192] The input subunit is used to input the model feature map into the distributed gradient enhancement library to obtain multiple markers corresponding to the model feature map.

[0193] A subunit is established for obtaining a construction tree based on a plurality of marking points, and establishing an initial offshore wind resource prediction model based on the construction tree.

[0194] In some optional embodiments, the offshore wind resource prediction device further includes:

[0195] The calculation module is used to calculate the mean absolute error, mean square error and root mean square error of the target offshore wind resource prediction model.

[0196] The evaluation module is used to evaluate the target offshore wind resource prediction model based on the mean absolute error, mean square error and root mean square error to obtain an evaluation result.

[0197] In some optional implementations, the conversion module 502 further includes:

[0198] The division unit is used to divide the target simulation area into multiple sub-areas.

[0199] The setting unit is used to set the sub-area as a nested area when the sub-area performs computing resource configuration.

[0200] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0201] The offshore wind resource prediction device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0202] An embodiment of the present application also provides a computer device having the offshore wind resource prediction device shown in FIG5 .

[0203] Please refer to Figure 6, which is a structural diagram of a computer device provided by an optional embodiment of the present application. As shown in Figure 6, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed within the computer device, including instructions stored in or on the memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 6 takes a processor 10 as an example.

[0204] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may also optionally include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0205] The memory 20 stores instructions that can be executed by at least one processor 10, so as to enable at least one processor 10 to execute the method shown in the above embodiment.

[0206] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0207] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0208] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0209] The embodiments of the present application also provide a computer-readable storage medium. The above-mentioned method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; optionally, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0210] Although the embodiments of the present application have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations shall fall within the scope defined by the appended claims.

Claims

1. A method for predicting offshore wind resources, characterized in that: The method comprises: Based on the global combined forecast system, the first offshore meteorological field dataset and the first wind speed dataset of a wind tower are obtained; Using the global combined forecast system to drive a weather forecast model, the first offshore meteorological field data set is converted into a second offshore meteorological field data set based on the weather forecast model; Interpolating the second offshore meteorological field data set to obtain a second wind tower wind speed data set; Based on the first wind tower wind speed data set and the second wind tower wind speed data set, a target offshore wind resource prediction model is established using a distributed gradient enhancement library; A third offshore meteorological field data set and a third wind speed data set of a wind farm in the area to be predicted are obtained, and the third offshore meteorological field data set and the third wind speed data set of a wind farm are input into the target offshore wind resource prediction model to obtain a wind resource prediction result of the wind farm in the area to be predicted.

2. The method according to claim 1, characterized in that Using the global combined forecast system to drive a weather forecast model, converting the first offshore meteorological field data set into a second offshore meteorological field data set based on the weather forecast model, comprising: Using the global combined forecast system to drive a weather forecast model to generate a target simulation area based on the weather forecast model; The first offshore meteorological field data set is interpolated into the target simulation area using a preset interpolation method to obtain the second offshore meteorological field data set.

3. The method according to claim 2, characterized in that Using the global combined forecast system to drive a weather forecast model to generate a target simulation area based on the weather forecast model includes: Using the global combined forecast system to drive a weather forecast model to generate an initial simulation area; Obtaining the area length and area width of the initial simulation area; Dividing the length of the area into a plurality of longitude segments, and dividing the width of the area into a plurality of latitude segments; Determine the grid point longitude and latitude of each grid point based on the plurality of longitude segments and the plurality of latitude segments, the grid point being composed of a longitude segment and a latitude segment; Based on the initial simulation area and the grid longitude and latitude of each of the grid points, the target simulation area based on the weather forecast model is generated.

4. The method according to claim 1, characterized in that Based on the first wind tower wind speed data set and the second wind tower wind speed data set, a target offshore wind resource prediction model is established using a distributed gradient enhancement library, including: Processing the first wind tower wind speed data set and the second wind tower wind speed data set by a data processing method to obtain a first target wind tower wind speed data set and a second target wind tower wind speed data set; Based on the first target wind tower wind speed data set and the second target wind tower wind speed data set, a target feature model is generated through feature engineering and cross-validation methods; Based on the second target wind tower wind speed data set and the target characteristic model, establishing an initial offshore wind resource prediction model using the distributed gradient enhancement library; The initial offshore wind resource prediction model is adjusted using the first wind tower wind speed data set until the target offshore wind resource prediction model is obtained.

5. The method according to claim 4, characterized in that Based on the first target wind tower wind speed data set and the target characteristic model, an initial offshore wind resource prediction model is established using the distributed gradient enhancement library, including: Obtaining a model feature graph of the target feature model; Inputting the model feature map into the distributed gradient enhancement library to obtain a plurality of marking points corresponding to the model feature map; A constructed tree is obtained based on the multiple marking points, and the initial offshore wind resource prediction model is established based on the constructed tree.

6. The method according to claim 1, characterized in that The method further comprises: Calculate the mean absolute error, mean square error and root mean square error of the target offshore wind resource prediction model; The target offshore wind resource prediction model is evaluated based on the mean absolute error, the mean square error and the root mean square error to obtain an evaluation result.

7. The method according to claim 2, characterized in that: After using the global combined forecast system to drive the weather forecast model and generate a target simulation area based on the weather forecast model, the method further includes: Dividing the target simulation area into a plurality of sub-areas; When the sub-region performs computing resource configuration, the sub-region is set as a nested region.

8. An offshore wind resource prediction device, characterized in that: The device comprises: A data acquisition module, used for acquiring a first offshore meteorological field data set and a first wind speed data set of a wind tower based on a global combined forecast system; A conversion module, used for using the global combined forecast system to drive the weather forecast model, and converting the first offshore meteorological field data set into a second offshore meteorological field data set based on the weather forecast model; A processing module, used for interpolating the second offshore meteorological field data set to obtain a second wind tower wind speed data set; An establishment module is used to establish a target offshore wind resource prediction model using a distributed gradient enhancement library based on the first wind tower wind speed data set and the second wind tower wind speed data set; The acquisition and input module is used to acquire a third offshore meteorological field data set and a third wind speed data set of a wind farm in the area to be predicted, and input the two offshore meteorological field data sets and the third wind speed data set of a wind farm into the target offshore wind resource prediction model to obtain the wind resource prediction result of the wind farm in the area to be predicted.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the offshore wind resource prediction method according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the offshore wind resource prediction method according to any one of claims 1 to 7.

Images