Wind data interpolation method for offshore floating wind measurement device based on artificial neural network

Abstract

The application discloses a wind measurement data interpolation method and device for a marine floating wind measurement device based on an artificial neural network, an electronic product and a computer readable storage medium. The application utilizes the correlation between the wind measurement data and the nearby water body characteristic data of the motion posture data of the marine floating wind measurement device during wind measurement, and realizes the interpolation of the wind measurement data of the marine floating wind measurement device based on the artificial neural network, solves the problem that the existing wind measurement data interpolation method cannot directly interpolate the wind measurement data of the marine floating wind measurement device, guarantees the accuracy and rationality of the wind measurement data interpolation of the marine floating wind measurement device, ensures the wind energy resource evaluation accuracy, and improves the safety and economic benefits of the wind farm operation.

Description

Technical Field

[0001] This invention belongs to the field of wind power technology. Specifically, it relates to a method, apparatus, electronic product, and computer-readable storage medium for interpolating wind measurement data from a floating offshore wind measurement device based on artificial neural networks. Background Technology

[0002] Compared to traditional fixed wind measurement devices, floating offshore wind measurement equipment (such as floating offshore wind measurement towers) has advantages such as portability and low deployment costs. Currently, it is being increasingly used in offshore wind power projects, especially floating offshore wind power projects. However, due to factors such as severe weather and equipment malfunctions, the measured data from floating offshore wind measurement devices may be incomplete or invalid. Therefore, wind measurement data interpolation has become a fundamental task in wind energy resource assessment. Reasonable wind measurement data interpolation methods are of great significance for improving the accuracy of wind energy resource assessment, ensuring the safe operation of wind farms, and enhancing their economic benefits.

[0003] Existing wind measurement data interpolation methods are mainly designed for fixed wind measurement devices, and can only consider the correlation between wind speed data. Moreover, when there are no other highly correlated data sources available for reference during the missing measurement period, it may introduce significant errors in wind speed interpolation. Furthermore, existing wind measurement data interpolation methods are prone to encountering situations where there are no highly correlated data sources when applied to floating wind measurement devices at sea. Therefore, they cannot be directly used for wind measurement data interpolation of floating wind measurement devices at sea. Summary of the Invention

[0004] The purpose of this invention is to ensure the accuracy and rationality of wind measurement data interpolation by offshore floating wind measurement devices, ensure the accuracy of wind energy resource assessment, improve the safety and economic benefits of wind farm operation, and solve the problems existing in the above-mentioned background technology.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for interpolating wind measurement data from a floating marine anemometer based on an artificial neural network, comprising:

[0006] Obtain wind-related data for the period to be interpolated and perform a rationality check to remove outliers. The wind-related data includes the time series of local water body characteristic parameters and motion attitude parameters when the offshore floating wind measuring device measures wind.

[0007] The wind measurement data, after outlier removal through a rationality check, is input into a local artificial neural network interpolation model to obtain wind measurement data for the interpolation period. The local artificial neural network interpolation model is a model obtained by training and evaluating an artificial neural network, and then learning the correlation between the wind measurement data and the wind measurement data. It can calculate and output the wind measurement data for the interpolation period based on the input wind measurement data for the interpolation period. The wind measurement data includes local sea wind speed time series and local sea wind direction time series.

[0008] In some embodiments of the present invention, training and evaluating the artificial neural network includes:

[0009] Acquire the data required for training and evaluation and perform a reasonableness check to remove outliers; wherein, the data required for training and evaluation includes the time series of local meteorological characteristic parameters, local water body characteristic parameters and motion attitude parameters of the same period measured by the floating wind measuring device at sea, and the time series of local meteorological characteristic parameters includes the time series of local sea wind speed and the time series of local sea wind direction;

[0010] Construct a complete set of training and testing data; the complete set of training and testing data includes the complete set of independent variable data and the complete set of dependent variable data for the same time period. The complete set of independent variable data includes the time series of local water body characteristic parameters and motion posture parameters, and the complete set of dependent variable data includes the time series of meteorological characteristic parameters.

[0011] The entire training and detection data set is divided into a training set and a detection set. The training set is used to train the artificial neural network, and the detection set is used to evaluate the qualification of the trained artificial neural network. After the evaluation is qualified, the local artificial neural network interpolation model is obtained.

[0012] In some embodiments of the present invention, the local water body characteristic parameter time series includes at least one of the local Eulerian velocity time series, flow direction time series, wave height time series, and wave direction time series.

[0013] In some embodiments of the present invention, the motion attitude parameter time series includes at least one of the following: latitude and longitude position time series, movement speed time series, movement direction time series, and attitude angle time series measured locally by the marine wind measuring device.

[0014] In some embodiments of the present invention, a hierarchical random sampling method is used to divide the entire training and detection data set into a training set and a detection set.

[0015] In some embodiments of the present invention, the entire training and detection data set is divided into a training set and a detection set. Specifically, the entire training and detection data set is divided into N corresponding training and verification data subsets according to N wind speed levels. Data is extracted from the N training and verification data subsets according to the proportion α of the test set data. The extracted data forms the test set, and the remaining data from the N training and verification data subsets forms the training set, where N≥2 and 0<α≤0.5.

[0016] In some embodiments of the present invention, the wind speed level has two levels: a high wind speed level and a low wind speed level. The high wind speed is a wind speed greater than or equal to 30 m / s, and the low wind speed is a wind speed less than 30 m / s.

[0017] Secondly, the present invention provides a wind measurement data interpolation device for a floating marine wind measurement device based on an artificial neural network, comprising:

[0018] The wind measurement related data acquisition unit is used to acquire wind measurement related data for the period to be interpolated and to perform a reasonableness check to remove outliers. The wind measurement related data includes the time series of local water body characteristic parameters and motion attitude parameters when the offshore floating wind measurement device measures wind.

[0019] The wind measurement data interpolation unit is used to input the wind measurement-related data, after outlier removal through a rationality check, into a local artificial neural network interpolation model to obtain wind measurement data for the interpolation period. The local artificial neural network interpolation model is a model obtained by training and evaluating an artificial neural network to learn the correlation between the wind measurement-related data and the wind measurement data. It can calculate and output the wind measurement data for the interpolation period based on the input wind measurement-related data. The wind measurement data includes local sea wind speed time series and local sea wind direction time series.

[0020] Thirdly, the present invention provides an electronic product comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute the aforementioned wind data interpolation method for a floating marine wind measuring device based on an artificial neural network.

[0021] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned wind measurement data interpolation method for a floating marine wind measurement device based on an artificial neural network.

[0022] Beneficial effects

[0023] This invention utilizes the correlation between wind measurement data and the water feature data near the motion posture data of the offshore floating wind measurement device during wind measurement. Based on an artificial neural network, it realizes the interpolation of wind measurement data from the offshore floating wind measurement device, solving the problem that existing wind measurement data interpolation methods cannot directly interpolate wind measurement data from the offshore floating wind measurement device. This ensures the accuracy and rationality of the wind measurement data interpolation from the offshore floating wind measurement device, ensures the accuracy of wind energy resource assessment, and improves the safety and economic benefits of wind farm operation. Attached Figure Description

[0024] Figure 1 This is a flowchart of a wind measurement data interpolation method according to one embodiment of the present invention.

[0025] Figure 2 This is a flowchart of the training and evaluation of an artificial neural network in one embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram illustrating the composition principle of the wind measurement data interpolation device in one embodiment of the present invention.

[0027] Figure 4 This is a schematic diagram illustrating the composition of an electronic product according to one embodiment of the present invention. Detailed Implementation

[0028] Considering that current offshore floating wind measurement devices (such as floating wind towers) can record not only basic wind measurement data but also their motion attitude data, coupled with the wave, ocean current, and other water body characteristic data recorded by their mounted hydrological instruments, the correlation between their motion attitude data, water body characteristic data, and wind measurement data provides new possibilities for wind measurement data interpolation for offshore floating wind towers. Therefore, based on the above-mentioned inventive concept, this invention provides a method, device, electronic product, and computer-readable storage medium for wind measurement data interpolation of offshore floating wind measurement devices based on artificial neural networks. This solves the problem that existing wind measurement data interpolation methods cannot directly interpolate wind measurement data from offshore floating wind measurement devices, ensuring the accuracy and rationality of wind measurement data interpolation, ensuring the accuracy of wind energy resource assessment, and improving the safety and economic benefits of wind farm operation.

[0029] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0030] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0031] In one embodiment of the present invention, a method for interpolating wind measurement data from a floating marine wind measurement device based on an artificial neural network is provided. For example... Figure 1 As shown, the method includes:

[0032] Obtain wind-related data for the period to be interpolated and perform a rationality check to remove outliers. The wind-related data includes the time series of local water body characteristic parameters and motion attitude parameters when the wind is measured by the floating wind measuring device at sea.

[0033] The wind measurement data, after outlier removal through a rationality check, is input into a local artificial neural network interpolation model to obtain wind measurement data for the interpolation period. The local artificial neural network interpolation model is a model obtained by training and evaluating an artificial neural network, and then learning the correlation between the wind measurement data and the wind measurement data. It can calculate and output the wind measurement data for the interpolation period based on the input wind measurement data for the interpolation period. The wind measurement data includes local sea wind speed time series and local sea wind direction time series.

[0034] In some embodiments of the present invention, such as Figure 2 As shown, the training and evaluation of the artificial neural network includes:

[0035] Acquire the data required for training and evaluation and perform a reasonableness check to remove outliers; wherein, the data required for training and evaluation includes the time series of local meteorological characteristic parameters, local water body characteristic parameters and motion attitude parameters of the same period measured by the floating wind measuring device at sea, and the time series of local meteorological characteristic parameters includes the time series of local sea wind speed and the time series of local sea wind direction.

[0036] Construct a complete set of training and testing data; wherein, the complete set of training and testing data includes a complete set of independent variable data and a complete set of dependent variable data for the same time period, the complete set of independent variable data includes time series of local water body characteristic parameters and time series of motion attitude parameters, the complete set of dependent variable data includes time series of meteorological characteristic parameters, and the time series of local meteorological characteristic parameters includes time series of local sea wind speed.

[0037] The entire training and detection data set is divided into a training set and a detection set. The training set is used to train the artificial neural network, and the detection set is used to evaluate the qualification of the trained artificial neural network. After the evaluation is qualified, the local artificial neural network interpolation model is obtained.

[0038] In some preferred embodiments of the present invention, the local water body characteristic parameter time series includes at least one of the local Eulerian velocity time series, flow direction time series, wave height time series, and wave direction time series.

[0039] In some preferred embodiments of the present invention, the motion attitude parameter time series includes at least one of the following: latitude and longitude position time series, movement speed time series, movement direction time series, and attitude angle time series measured locally by the marine wind measuring device.

[0040] In some embodiments of the present invention, a hierarchical random sampling method is used to divide the entire training and detection data set into a training set and a detection set.

[0041] In some embodiments of the present invention, the entire training and testing data set is divided into a training set and a testing set. Specifically, the entire training and testing data set is divided into N corresponding training and validation data subsets according to N wind speed levels. Data is extracted from each of the N training and validation data subsets according to a test set data ratio α. The extracted data forms the test set, and the remaining data from the N training and validation data subsets forms the training set, where N ≥ 2 and 0 < α ≤ 0.5. Preferably, there are two wind speed levels: high wind speed level and low wind speed level. High wind speed is defined as wind speed greater than or equal to 30 m / s, and low wind speed is defined as wind speed less than 30 m / s.

[0042] In some embodiments of the present invention, specifically, the floating wind measuring device is a floating wind measuring tower, which is equipped with hydrological measuring instruments to measure the time series of local water body characteristic parameters, and also equipped with motion attitude measuring devices, such as GPS, to measure the time series of motion attitude parameters.

[0043] In some embodiments of the present invention, the artificial neural network is a BiGRU artificial neural network.

[0044] The present invention provides a wind measurement data interpolation method for a floating marine wind measurement device based on an artificial neural network in one embodiment. This method utilizes the correlation between the floating wind measurement tower's motion attitude data, wind and wave velocity data, and wind measurement data, employing a matrix calculation method. Specifically:

[0045] Step 1: Obtain the data required for training evaluation.

[0046] The required training and evaluation data include: (1) time series of meteorological parameters, including local wind speed and wind direction, observed and recorded during the wind measurement cycle of the floating wind tower; (2) time series of water body characteristics, including local Euler flow velocity, flow direction, wave height and wave direction; and (3) time series of floating wind tower motion attitude, including latitude and longitude location, moving speed, moving direction and attitude angle data.

[0047] Step 2: Perform a reasonableness check on the collected data and remove outliers.

[0048] Specifically, the reasonable range for wind speed is between 0 and 40 m / s; the reasonable range for current velocity is between 0 and 10 m / s; the reasonable range for wave height is between 0 and 10 m; the reasonable range for the moving speed of the floating wind tower is between 0 and 20 m / s; and the wind direction, current direction, and wave direction are between 0 and 360°. The standard deviation of each variable must not be 0 for two consecutive hours; otherwise, the instrument measuring that variable is considered frozen or damaged.

[0049] Step 3: Reconstruct the data after the rationality test into a complete set of training and evaluation data.

[0050] Specifically, in this embodiment, a matrix is ​​used to organize and store data and to perform data operations.

[0051] The time series of water body characteristics (including local Eulerian current velocity, flow direction, wave height, and wave direction) and the time series of floating anemometer status (including moving speed, moving direction, and attitude angle data) together constitute the independent variable matrix X.

[0052]

[0053] Where, x i,j For a valid data value, i represents the position in the time series, and j represents the variable to which it belongs. Considering that different data may be removed after reasonableness checks, it can be a null value NaN.

[0054] The time series of meteorological parameters (wind speed and wind direction) constitute the dependent variable (target variable) matrix Y.

[0055]

[0056] Among them, y i,1 Represents the wind speed time series, y i,2 Let i represent the wind direction time series, and i represent the position in the time series. To ensure the reliability of the model, the wind speed data at any time point i0 is used. It cannot be a null value (NaN).

[0057] Step 4: Obtain the sample data matrix required for training and evaluation, namely the training set and the detection set, by performing stratified random sampling on the entire training and evaluation data set.

[0058] Specifically, based on the wind speed vector y in the dependent variable matrix Y i,1 The size of the wind speed determines the independent variable data matrices X and Y, which are divided into two categories: low wind speed (wind speed less than 30 m / s) and high wind speed (wind speed greater than or equal to 30 m / s), denoted as X. n X e With Y n Y e Where the subscript n represents the low wind speed category and the subscript e represents the high wind speed category; respectively for X n With X e Perform simple random sampling, with the sample size being 10% of the total population, denoted as . and The remaining 90% of the training and evaluation dataset that was not sampled is denoted as... and Based on the sampling results of the independent variable matrix X, four corresponding matrices are obtained from the dependent variable matrix Y, denoted as follows: Will and Combined into independent variable test data matrix X c ,Will and Combined into a dependent variable test data matrix Y c ,Right now

[0059]

[0060]

[0061] Similarly, obtain the training data matrix X of the independent variables. t With the dependent variable training data matrix Y t ,Right now

[0062]

[0063]

[0064] Step 5: Train the BiGRU artificial neural network model using the training set.

[0065] Specifically, the training data matrix X obtained by hierarchical random sampling and recombination t As feature data, its corresponding Y t As the target variable, it is provided to the BiGRU (Bidirectional Gated Recurrent Units) artificial neural network model for supervised learning; BiGRU consists of a forward GRU and a backward GRU, and the formula is:

[0066]

[0067]

[0068]

[0069] in, For forward output, For reverse output, For the model results, This is a bitwise summation.

[0070] BiGRU fills in missing information by bidirectionally connecting past and future states. Considering that GRU uses techniques such as update gates and reset gates to preserve information and gradients, it is faster and easier to deploy. In this embodiment, the BiGRU model uses a 3-layer network with 128, 64, and 32 neurons respectively, and the sigmoid activation function is chosen for each layer. During training, MSE is selected as the cost function, and Adam is selected as the optimizer. The training endpoint is determined by setting an error threshold ε; in this embodiment, ε = 0.0002.

[0071] Step 6: Use cross-validation with the test set X. c With Y c Evaluate the supervised learning model.

[0072] Specifically, the optimal hyperparameter tuning and the training dataset matrix (X) are selected. t With Y t The trained BiGRU artificial neural network model will be used to reconstruct the test data matrix X obtained by hierarchical random sampling. c The optimal hyperparameters are then input into the BiGRU neural network model for simulation testing, and the simulation result matrix is ​​obtained. Compare it with the test dataset matrix Y c To compare the models, their simulation performance in wind speed and wind direction was evaluated. The model's skill score (SS) and root mean square error (RMSE) were also assessed using the following formulas:

[0073]

[0074]

[0075] in, Y represents the simulation results (wind speed or wind direction) of the BiGRU artificial neural network model. i cThe test dataset contains real observation data (wind speed or wind direction), and N is the size of the data.

[0076] In this embodiment, the criteria for a qualified model are SS greater than 0.6, RMSE of wind speed less than 0.5 m / s, and RMSE of wind direction less than 10°.

[0077] Step 7: Use a validated BiGRU artificial neural network model to simulate and interpolate the wind measurement data for the interpolation period.

[0078] Specifically, local Eulerian current velocity, direction, wave height, and wave direction data, as well as the latitude and longitude location, moving speed, moving direction, and attitude angle data of the floating anemometer tower, are collected during the interpolation period. After the rationality check described in step 2 to remove invalid data, the independent variable matrix X to be interpolated is formed. * X * The validated BiGRU artificial neural network model is input for simulation to obtain the dependent variable matrix Y. * This refers to the interpolated wind measurement data.

[0079] like Figure 3 As shown, in one embodiment of the present invention, the wind measurement data interpolation device for a floating marine wind measurement device based on an artificial neural network includes:

[0080] The wind measurement related data acquisition unit is used to acquire wind measurement related data for the period to be interpolated and to perform a reasonableness check to remove outliers. The wind measurement related data includes the time series of local water body characteristic parameters and motion attitude parameters when the offshore floating wind measurement device measures wind.

[0081] The wind measurement data interpolation unit is used to input the wind measurement-related data, after outlier removal through a rationality check, into a local artificial neural network interpolation model to obtain wind measurement data for the interpolation period. The local artificial neural network interpolation model is a model obtained by training and evaluating an artificial neural network to learn the correlation between the wind measurement-related data and the wind measurement data. It can calculate and output the wind measurement data for the interpolation period based on the input wind measurement-related data. The wind measurement data includes local sea wind speed time series and local sea wind direction time series.

[0082] In one embodiment, the present invention provides an electronic product, such as... Figure 4 As shown, the electronic device includes at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute the above-described wind measurement data interpolation method for a floating marine wind measurement device based on an artificial neural network.

[0083] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripherals, voltage regulators, and power management circuits, via interfaces, as is well known in the art. Interfaces provide a connection between the bus and the transceiver, such as communication interfaces or user interfaces. The transceiver can be a single component or multiple components, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0084] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0085] In one embodiment of the present invention, a computer-readable storage medium is provided, storing a computer program that, when executed by a processor, implements the above-described embodiment of the wind measurement data interpolation method for a floating marine wind measurement device based on an artificial neural network.

[0086] Those skilled in the art will understand from the foregoing description that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes, but is not limited to, various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic storage devices, and optical storage devices.

[0087] This invention utilizes the correlation between wind measurement data and the water feature data near the motion posture data of the offshore floating wind measurement device during wind measurement. Based on an artificial neural network, it realizes the interpolation of wind measurement data from the offshore floating wind measurement device, solving the problem that existing wind measurement data interpolation methods cannot directly interpolate wind measurement data from the offshore floating wind measurement device. This ensures the accuracy and rationality of the wind measurement data interpolation from the offshore floating wind measurement device, ensures the accuracy of wind energy resource assessment, and improves the safety and economic benefits of wind farm operation.

[0088] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0089] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0090] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0091] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0092] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for interpolating wind measurement data from a floating marine anemometer based on an artificial neural network, characterized in that, Obtain wind-related data for the period to be interpolated and perform a rationality check to remove outliers. The wind-related data includes the time series of local water body characteristic parameters and motion attitude parameters when the offshore floating wind measuring device measures wind. The wind measurement data, after outlier removal through a rationality check, is input into a local artificial neural network interpolation model to obtain wind measurement data for the interpolation period. The local artificial neural network interpolation model is a model obtained by training and evaluating an artificial neural network to learn the correlation between the wind measurement data and the wind measurement data. It can calculate and output the wind measurement data for the interpolation period based on the input wind measurement data. The wind measurement data includes local sea wind speed time series and local sea wind direction time series. The training and evaluation of the artificial neural network includes: Acquire the data required for training and evaluation and perform a reasonableness check to remove outliers; wherein, the data required for training and evaluation includes the time series of local meteorological characteristic parameters, local water body characteristic parameters and motion attitude parameters of the same period measured by the floating wind measuring device at sea, and the time series of local meteorological characteristic parameters includes the time series of local sea wind speed and the time series of local sea wind direction; Construct a complete set of training and testing data; the complete set of training and testing data includes the complete set of independent variable data and the complete set of dependent variable data for the same time period. The complete set of independent variable data includes the time series of local water body characteristic parameters and motion posture parameters, and the complete set of dependent variable data includes the time series of meteorological characteristic parameters. A stratified random sampling method is used to divide the entire training and testing data set into a training set and a testing set. Specifically, the entire training and testing data set is divided into N corresponding training and validation data subsets according to N wind speed levels. Data is extracted from each of the N training and validation data subsets according to the proportion α of the testing set data. The extracted data forms the testing set, and the remaining data from the N training and validation data subsets forms the training set, where N≥2 and 0<α≤0.

5. The training set is used to train the artificial neural network, and the detection set is used to evaluate the qualification of the trained artificial neural network. After the evaluation is qualified, the local artificial neural network interpolation model is obtained.

2. The wind measurement data interpolation method as described in claim 1, characterized in that, The local water body characteristic parameter time series includes at least one of the local Euler velocity time series, flow direction time series, wave height time series, and wave direction time series.

3. The wind measurement data interpolation method as described in claim 1, characterized in that, The motion attitude parameter time series includes at least one of the following: latitude and longitude position time series, movement speed time series, movement direction time series, and attitude angle time series measured locally by the marine wind measuring device.

4. The wind measurement data interpolation method as described in claim 1, characterized in that, There are two levels of wind speed: high wind speed and low wind speed. High wind speed is defined as wind speed greater than or equal to 30 m / s, and low wind speed is defined as wind speed less than 30 m / s.

5. A wind measurement data interpolation device for a floating marine wind measurement device based on an artificial neural network, comprising: The wind measurement related data acquisition unit is used to acquire wind measurement related data for the period to be interpolated and to perform a reasonableness check to remove outliers. The wind measurement related data includes the time series of local water body characteristic parameters and motion attitude parameters when the offshore floating wind measurement device measures wind. The wind measurement data interpolation unit is used to input the wind measurement-related data, after outlier removal through a rationality check, into a local artificial neural network interpolation model to obtain wind measurement data for the interpolation period. The local artificial neural network interpolation model is a model obtained by training and evaluating an artificial neural network to learn the correlation between the wind measurement-related data and the wind measurement data. It can calculate and output the wind measurement data for the interpolation period based on the input wind measurement-related data for the interpolation period. The wind measurement data includes local sea wind speed time series and local sea wind direction time series. The training and evaluation of the artificial neural network includes: Acquire the data required for training and evaluation and perform a reasonableness check to remove outliers; wherein, the data required for training and evaluation includes the time series of local meteorological characteristic parameters, local water body characteristic parameters and motion attitude parameters of the same period measured by the floating wind measuring device at sea, and the time series of local meteorological characteristic parameters includes the time series of local sea wind speed and the time series of local sea wind direction; Construct a complete set of training and testing data; the complete set of training and testing data includes the complete set of independent variable data and the complete set of dependent variable data for the same time period. The complete set of independent variable data includes the time series of local water body characteristic parameters and motion posture parameters, and the complete set of dependent variable data includes the time series of meteorological characteristic parameters. A stratified random sampling method is used to divide the entire training and testing data set into a training set and a testing set. Specifically, the entire training and testing data set is divided into N corresponding training and validation data subsets according to N wind speed levels. Data is extracted from each of the N training and validation data subsets according to the proportion α of the testing set data. The extracted data forms the testing set, and the remaining data from the N training and validation data subsets forms the training set, where N≥2 and 0<α≤0.

5. The training set is used to train the artificial neural network, and the detection set is used to evaluate the qualification of the trained artificial neural network. After the evaluation is qualified, the local artificial neural network interpolation model is obtained.

6. An electronic product comprising: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the wind data interpolation method for a floating marine wind measuring device based on an artificial neural network as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, wherein, When the computer program is executed by the processor, it implements the wind measurement data interpolation method for a floating marine wind measurement device based on an artificial neural network, as described in any one of claims 1 to 4.

Images