Sea surface wind speed vertical extrapolation method and device based on satellite remote sensing inversion data
By constructing an artificial neural network model based on satellite remote sensing data, the problems of high cost and insufficient accuracy of wind speed measurement in offshore wind farms were solved, accurate vertical extrapolation of sea surface wind speed was achieved, and accurate wind resource assessment was provided for offshore wind farm planning.
Patent Information
- Application Number
- CN202510876005.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-27
- Publication Date
- 2025-10-14
AI Technical Summary
Wind speed measurement in offshore wind farms is costly and difficult to accurately obtain vertical wind speed data. Traditional methods rely on simplified mathematical formulas, and machine learning algorithms lack adaptability and accuracy in complex environments.
By collecting measured data from offshore wind towers and matching them with satellite remote sensing data, an artificial neural network model was constructed. Maximum and minimum normalization and feature selection were used to train the sample set to achieve vertical extrapolation of sea surface wind speed. The Adam optimizer and a learning rate of 0.01 were used for 1000 iterations to output wind speed data at different heights at sea.
Accurate vertical extrapolation of sea surface wind speed in complex offshore environments is achieved, with wind speed deviation less than 0.16m/s and wind direction deviation less than 3.92°, supporting wind resource assessment for offshore wind farms.
Smart Images
Figure CN120781091A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the field of atmospheric science, and particularly relates to a sea surface wind speed vertical extrapolation method and device based on satellite remote sensing inversion data. BACKGROUND
[0002] Offshore wind power has the advantages of high space utilization, large wind energy density, stable wind speed, etc., and is an important direction for future wind energy development. Vigorously developing offshore wind power is of great significance to energy security, green and low-carbon development, and achieving the "double carbon" goal.
[0003] However, offshore wind farms are usually located in open waters with complex environments, and it is costly and technically difficult to directly measure wind speed and direction. Traditional onshore wind farms can rely on wind measurement towers to obtain vertical wind speed data, but in the marine environment, it is very difficult and expensive to deploy and maintain wind measurement towers. With the rapid development of satellite remote sensing technology, satellite observations of sea surface wind fields have shown great advantages. Microwave sensors carried on satellites can achieve large-area long-term dynamic monitoring with quasi-real-time and high accuracy, which significantly improves the efficiency and accuracy of ocean wind field observation, providing strong support for studying marine meteorology and developing offshore wind power resources.
[0004] Currently, satellite remote sensing can only measure wind speed at 10m height, and sea surface wind speed vertical extrapolation technology is needed to extrapolate the 10m height wind speed to the target wind turbine hub height wind speed for target area wind resource assessment. There have been a large number of researches on wind speed vertical extrapolation methods for wind farms. Current researches mainly focus on fitting wind profiles, such as logarithmic linear law, logarithmic law, exponential law, etc. The above traditional wind speed extrapolation usually relies on simplified mathematical formulas. In recent years, machine learning (such as deep learning, support vector machine, etc.) has also been applied to vertical extrapolation of wind speed. Through training on a large amount of historical wind speed data, machine learning algorithms can automatically identify the complex relationship between wind speed and height, and then make more accurate wind speed predictions. This technology can greatly improve the adaptability and accuracy of the model, especially in complex environmental conditions. SUMMARY
[0005] In view of the deficiencies of the prior art, the purpose of the present application is to provide a sea surface wind speed vertical extrapolation method and device based on satellite remote sensing inversion data.
[0006] The purpose of the present application is achieved by the following technical solutions: In a first aspect, the present application provides a sea surface wind speed vertical extrapolation method based on satellite remote sensing inversion data, which comprises the following steps:
[0007] 1) Collecting existing offshore wind tower measured data, the wind tower measured data comprising wind speed data above 10m height of sea surface, wind tower latitude and longitude position and wind measurement time;
[0008] 2) Collect satellite remote sensing data of the sea area, which contains wind speed data of sea surface 10m height, latitude and longitude and time information, and extract wind speed data at the location of the wind measurement tower from the satellite remote sensing data;
[0009] 3) Time match the measured data with the satellite remote sensing data to form a sample set, and perform data normalization and feature selection;
[0010] 4) Construct an artificial neural network model, which includes at least an input layer, an output layer and at least one hidden layer, wherein the input layer is used to receive the wind speed features of the sea surface 10m height in the training data set, the hidden layer is used to perform nonlinear transformation and mapping of the feature vector to the output layer, and the output layer is used to calculate the wind speed data above 10m height, realizing the vertical extrapolation of the sea surface wind speed;
[0011] 5) Train the artificial neural network model constructed in step 4) using the sample set; for the sea area to be measured, input the wind speed features of the sea surface 10m height of the area obtained from the satellite remote sensing data into the trained artificial neural network model to obtain the wind speed data at different heights of the sea area.
[0012] Further, in step 1), for a specific latitude and longitude position of the wind measurement tower, the wind speed data in the measured data includes wind speed and wind direction data at multiple different heights of the latitude and longitude position; the east-west direction wind speed and the north-south direction wind speed at different heights are calculated according to the wind speed and the wind direction.
[0013] Further, in step 2), the extraction of wind speed data at the location of the wind measurement tower from the satellite remote sensing data includes: using the nearest neighbor method to extract the east-west direction wind speed and the north-south direction wind speed of the sea surface 10m at the grid where the wind measurement tower is located from the satellite remote sensing data according to the latitude and longitude position of the wind measurement tower.
[0014] Further, in step 3), the data in the measured data obtained in step 1) which is the same in time as the satellite remote sensing data obtained in step 2) is selected, and then a sample set containing the east-west direction wind speed, the north-south direction wind speed, the hour, the month and the corresponding measured east-west direction wind speed and north-south direction wind speed at different heights of the sea surface 10m height is obtained.
[0015] Further, in step 2), the maximum and minimum normalization method is used to map each feature in the sample set to [0, 1], and the normalization formula is:
[0016]
[0017] wherein x is the original feature data in the sample set, x min is the minimum value of the feature data, xmax is the maximum value of the feature data, and x' is the normalized feature data.
[0018] Further, in step 4), the artificial neural network model comprises:
[0019] The input layer comprises 4 neurons, and the input data comprises the sea surface 10m inverted east-west wind speed and north-south wind speed, and the corresponding month and hour number.
[0020] The hidden layer 1 comprises 128 neurons, and the activation function is a rectified linear unit ReUL.
[0021] The hidden layer 2 comprises 64 neurons, and the activation function is a rectified linear unit ReUL.
[0022] The hidden layer 3 comprises 32 neurons, and the activation function is a rectified linear unit ReUL.
[0023] The output layer comprises 2 neurons, which are respectively used to calculate the sea surface 10m height above the east-west wind speed and the north-south wind speed, and the activation function is a linear activation function.
[0024] Further, in step 5), the sample set is divided into a training set, a validation set and a test set according to a preset proportion, the training set is used to train the artificial neural network model, so that the artificial neural network model can accurately learn the wind profile law of the sea surface wind speed, the validation set is used for model parameter adjustment and hyperparameter optimization, the performance of the model is evaluated on the validation machine during the model training process, and the model parameters, the learning rate and the grid structure are adjusted according to the calculated root mean square error index to optimize the model performance. The test set is used to evaluate the generalization ability of the trained artificial neural network model at different times.
[0025] Further, in step 5), the loss function of the model during the training process is a mean square error loss function, and the model weight is updated by using the back propagation and the gradient descent method.
[0026] In a second aspect, the present application also provides a sea surface wind speed vertical extrapolation device based on satellite remote sensing inversion data, comprising a memory and one or more processors, the memory stores executable code, and the processor executes the executable code to realize the sea surface wind speed vertical extrapolation method based on satellite remote sensing inversion data.
[0027] In a third aspect, the present application also provides a computer readable storage medium having a program stored thereon, and the program is executed by a processor to realize the sea surface wind speed vertical extrapolation method based on satellite remote sensing inversion data.
[0028] The present application has the beneficial effects that: the present application obtains good extrapolation results by performing sea surface wind speed vertical extrapolation based on satellite remote sensing inversion data, and overcomes the problem of wind resource assessment in the early stage of offshore wind farm planning. BRIEF DESCRIPTION OF DRAWINGS
[0029] Figure 1 The method logic diagram of the present application is shown in the figure.
[0030] Figure 2 The extrapolation results are compared with the measured results.
[0031] Figure 3 The device structure diagram provided by the present application is shown in the figure. DETAILED DESCRIPTION
[0032] The present application will be further described and explained in conjunction with the accompanying drawings and specific embodiments.
[0033] The present application mainly constructs a sea surface wind speed vertical extrapolation method based on satellite remote sensing inversion data, aiming to reasonably estimate the wind speed at different heights by using the sea surface wind field information obtained by remote sensing, and to provide support for marine meteorological research and offshore wind power and other applications. The specific steps are as follows:
[0034] 1) Collect measured data. Collect the measured data of the existing offshore wind tower. The physical quantity measured by the wind tower is the wind speed at 110m height, and the wind direction. According to the wind speed and wind direction, the east-west direction and north-south direction wind speed at 110m height are calculated.
[0035] 2) Collect satellite remote sensing data products. Download the Global Ocean Hourly Sea Surface Wind and Stress from Scatterometer and Model provided by Copernicus Marine Service, with a time interval of 1h. The data set contains sea surface 10m east-west direction and north-south direction wind speed variables. According to the latitude and longitude of the wind tower, the data at the grid where the wind tower is located in the data set is extracted using the nearest neighbor method.
[0036] 3) Data preprocessing. The measured data obtained in step (1) and the satellite remote sensing data obtained in step (2) are filtered to obtain a sample set containing sea surface 10m height east-west direction wind speed, north-south direction wind speed, hour, month and corresponding measured 110m height east-west direction wind speed, north-south direction wind speed. The maximum and minimum normalization method is used to normalize each feature. The maximum and minimum normalization formula is:
[0037]
[0038] wherein x is the original feature data, x min is the minimum value of the feature, x max is the maximum value of the feature, and x' is the normalized feature data.
[0039] 4) Building an artificial neural network model. An artificial neural network model is built using the pytorch framework, which includes an input layer, 3 hidden layers, and an output layer. The input layer contains 4 neurons, and the input data includes the east-west direction wind speed and the north-south direction wind speed of the sea surface 10m, as well as the corresponding month and hour number. Hidden layer 1 contains 128 neurons, and the activation function is the rectified linear unit ReUL; hidden layer 2 contains 64 neurons, and the activation function is the rectified linear unit ReUL. Hidden layer 3 contains 32 neurons, and the activation function is the rectified linear unit ReUL. The output layer contains 2 neurons, which are used to calculate the east-west direction wind speed and the north-south direction wind speed of the sea surface 10m, and the activation function is the linear activation function.
[0040] 5) Model training and testing. The training set and the test set are divided in a ratio of 14:1, and the validation set is divided in a ratio of 0.1 in the training set. In the model training process, the optimizer used is Adam, the learning rate is 0.01, and the number of iterations is 1000.
[0041] The following examples demonstrate the specific implementation effects of the above method.
[0042] Example
[0043] In this example, a certain offshore sea area in Zhejiang Province is taken as the research object, and the specific implementation effect and advantage of the sea surface wind speed vertical extrapolation method based on satellite remote sensing inversion data are explored. The relevant parameter settings are as follows:
[0044] The measured data of a certain offshore wind tower from January 2023 to May 2024 is collected, with a time interval of 15 minutes.
[0045] The Global Ocean Hourly Sea Surface Wind and Stress from Scatterometer and Model data set provided by the Copernicus Marine Service is downloaded, with a time interval of 1 hour. The nearest neighbor method is used to extract the data at the grid where the wind tower is located in the data set.
[0046] The measured data of the wind tower and the satellite remote sensing data are screened for the same time, and there are 12284 pieces of data. Then, the sample set containing the 10m height of the sea surface, the east-west direction wind speed, the north-south direction wind speed, the hour, the month and the corresponding measured 110m height of the east-west direction wind speed, the north-south direction wind speed is obtained. The maximum and minimum normalization method is used for normalization of each feature.
[0047] The data from January 2023 to March 2024 is taken as the training set and the validation set, and the validation set accounts for 0.1, and the data from April 2024 to May 2024 is taken as the test set. In the model training process, the adopted optimizer is Adam, the learning rate is 0.01, and the iteration number is 1000.
[0048] In order to intuitively show the extrapolation effect, the extrapolation result of the model on the test set and the measured result are shown in the attached Figure 2 The comparison chart of the extrapolation result of the model on the test set and the measured result is shown. The deviation of the wind speed is-0.16m / s, and the deviation of the wind direction is-3.92°, which shows good extrapolation result.
[0049] The above examples are only exemplary, and in the practical application of the present application, the historical wind speed data of different heights of the existing wind tower can be further obtained to further enrich the sample data, so that the artificial neural network model finally obtained can obtain the wind speed data of different heights in the offshore area.
[0050] Corresponding to the foregoing embodiment of the vertical extrapolation method of the sea surface wind speed based on satellite remote sensing inversion data, the present application also provides an embodiment of a vertical extrapolation device of the sea surface wind speed based on satellite remote sensing inversion data.
[0051] Referring to Figure 3 The vertical extrapolation device of the sea surface wind speed based on satellite remote sensing inversion data provided by the embodiment of the present application comprises a memory and one or more processors, the memory stores executable codes, and the processor executes the executable codes to implement the vertical extrapolation method of the sea surface wind speed based on satellite remote sensing inversion data in the foregoing embodiment.
[0052] The embodiment of the vertical extrapolation device of the sea surface wind speed based on satellite remote sensing inversion data provided by the present application can be applied to any device with data processing capability, which can be a device or apparatus such as a computer. The device embodiment can be realized by software, or by hardware or a combination of software and hardware. Taking software realization as an example, as a logical device, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory and running by the processor of the device with data processing capability. From the hardware level, as shown in the accompanying drawings, the device comprises a processor 1 and a memory 2. Figure 3As shown in the figure, it is a hardware structure diagram of a device provided by the application for a satellite remote sensing inversion data-based sea surface wind speed vertical extrapolation device on any data processing capable device, in addition to Figure 3 In addition to the processor, the memory, the network interface, and the non-volatile memory shown in the figure, any data processing capable device on which the device in the embodiment is generally also provided with other hardware according to the actual functions of the data processing capable device, and no further description is given.
[0053] The implementation process of the functions and roles of each unit in the above device is specifically described in the implementation process of the corresponding steps in the above method, and no further description is given here.
[0054] For the device embodiment, since it basically corresponds to the method embodiment, the relevant part is described in the part of the method embodiment. The device embodiment described above is only schematic, and the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the application scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.
[0055] The embodiment of the application also provides a computer readable storage medium, which stores a program, and the program is executed by a processor to implement the machine learning wind speed correction method for coupling vertical wind profile.
[0056] The computer readable storage medium can be an internal storage unit of any data processing capable device, such as a hard disk or a memory. The computer readable storage medium can also be an external storage device of any data processing capable device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. Further, the computer readable storage medium can include both the internal storage unit and the external storage device of any data processing capable device. The computer readable storage medium is used to store the computer program and other programs and data required by the data processing capable device, and can also be used to temporarily store data that has been output or will be output.
[0057] The above embodiments are used to explain and illustrate the application, rather than limit the application, and any modifications and changes made to the application within the spirit and protection scope of the claims of the application fall within the protection scope of the application.
Claims
1. A method for vertical extrapolation of sea surface wind speed based on satellite remote sensing inversion data, characterized in that: The steps of this method are as follows: 1) Collecting measured data from existing wind towers at sea, including wind speed data above 10 m above sea level, the tower's longitude and latitude, and wind measurement time; 2) Collecting satellite remote sensing data of the offshore area, wherein the satellite remote sensing data includes wind speed data at an altitude of 10 meters above the sea surface, longitude and latitude, and time information, and extracting wind speed data at the location of the wind tower from the satellite remote sensing data; 3) Time-matching the measured data with satellite remote sensing data to form a sample set, and performing data normalization and feature selection; 4) constructing an artificial neural network model, wherein the network structure includes at least an input layer, an output layer, and at least one hidden layer, wherein the input layer is used to receive the wind speed characteristics at a height of 10 meters above the sea surface in the training data set, the hidden layer is used to perform a nonlinear transformation on the feature vector and map it to the output layer, and the output layer is used to calculate the wind speed data above the height of 10 meters to realize vertical extrapolation of the sea surface wind speed; 5) The artificial neural network model constructed in step 4) is trained using the sample set; for an offshore area to be measured without a wind tower, the wind speed characteristics at a height of 10 m above the sea surface in the area obtained by satellite remote sensing data are input into the trained artificial neural network model to obtain wind speed data at different heights in the area to be measured.
2. The method for vertical extrapolation of sea surface wind speed based on satellite remote sensing inversion data according to claim 1, characterized in that: In step 1), for a wind tower at a specific latitude and longitude, the wind speed data in the measured data includes wind speed and wind direction data at multiple different heights at the latitude and longitude position; the east-west wind speed and north-south wind speed at different heights are calculated based on the wind speed and wind direction.
3. The method for vertical extrapolation of sea surface wind speed based on satellite remote sensing inversion data according to claim 1, characterized in that: In step 2), extracting the wind speed data at the location of the wind tower from the satellite remote sensing data product includes: extracting the east-west wind speed and the north-south wind speed at 10 m above the sea surface at the grid where the wind tower is located from the satellite remote sensing data using the nearest neighbor method based on the latitude and longitude of the wind tower.
4. The method for vertical extrapolation of sea surface wind speed based on satellite remote sensing inversion data according to claim 1, characterized in that: In step 3), the measured data obtained in step 1) are screened for data with the same time as the satellite remote sensing data obtained in step 2), thereby obtaining a sample set including the east-west wind speed, north-south wind speed, hour, month, and the corresponding east-west wind speed and north-south wind speed at different measured heights at a height of 10 m above the sea surface; Use the maximum and minimum normalization method to map each feature in the sample set to [0, 1]. The normalization formula is: Among them, x is the original feature data in the sample set, x min is the minimum value of the characteristic data, x max is the maximum value of the feature data, and x' is the normalized feature data.
5. The method for vertical extrapolation of sea surface wind speed based on satellite remote sensing inversion data according to claim 1, characterized in that: The artificial neural network model described in step 4) includes: Input layer: Contains 4 neurons, and the input data includes the east-west wind speed and north-south wind speed inverted at 10 meters above the sea surface, as well as the corresponding months and hours; Hidden layer 1: contains 128 neurons, and the activation function is the rectified linear unit ReUL; Hidden layer 2: contains 64 neurons, and the activation function is the rectified linear unit ReUL; Hidden layer 3: contains 32 neurons, and the activation function is the rectified linear unit ReUL; Output layer: Contains two neurons, one for calculating the east-west wind speed and the other for calculating the north-south wind speed above 10 m above the sea surface. The activation function is a linear activation function.
6. The method for vertical extrapolation of sea surface wind speed based on satellite remote sensing inversion data according to claim 1, characterized in that: The sample set described in step (5) is used to train the artificial neural network model so that it can accurately learn the wind profile law of the sea surface wind speed; the loss function of the model is the mean square error loss function, and the model weights are updated using back propagation and gradient descent method; During the model training process, its performance is evaluated regularly, and the model parameters, learning rate, and grid structure are adjusted according to the calculated root mean square error index to optimize the model performance.
7. A device for vertically extrapolating sea surface wind speed based on satellite remote sensing inversion data, comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that: When the processor executes the executable code, the vertical extrapolation method of sea surface wind speed based on satellite remote sensing inversion data according to any one of claims 1 to 6 is implemented.
8. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the vertical extrapolation method of sea surface wind speed based on satellite remote sensing inversion data according to any one of claims 1 to 6 is implemented.