Turbulence intensity calculation method, model training method, device, equipment and medium

Abstract

The present application provides a turbulence intensity calculation method, model training method, device, equipment, and medium, relating to the field of wind power generation. The turbulence intensity calculation method includes: obtaining measured geographic data and measured wind resource data from a remote sensing wind sensing device; inputting the measured geographic data and measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the wind tower; and calculating the turbulence intensity at the location of the remote sensing wind sensing device based on the target wind speed standard deviation of the wind tower. Utilizing the technical solution of the present application can improve the accuracy of executing business operations using turbulence intensity obtained from a remote sensing wind sensing device.

Description

Technical Field

[0001] The present application relates to the field of wind power generation, and in particular to a turbulence intensity calculation method, model training method, device, equipment and medium. Background Art

[0002] Turbulence intensity describes the degree to which wind speed varies over time and space. It reflects the relative strength of fluctuating wind speed and is one of the most important characteristics of atmospheric turbulence. Turbulence intensity can be measured and calculated through wind measurement. Turbulence intensity can be used to assess wind turbine performance and adjust wind turbine control.

[0003] Currently, there are two available wind measurement methods: wind tower measurement and remote sensing measurement. Compared to remote sensing, wind tower measurement has a lower tower height and cannot obtain wind field information at higher altitudes, resulting in greater uncertainty in capacity assessment. Due to the different principles of wind tower and remote sensing measurement, the turbulence intensity calculated from wind tower measurements differs from that calculated from remote sensing measurements. However, the current standards for using turbulence intensity for capacity assessment or adjustment control are based on the turbulence intensity calculated from wind tower measurements. Using remote sensing wind measurement to measure and calculate turbulence intensity reduces the accuracy of business operations. Summary of the Invention

[0004] The embodiments of the present application provide a turbulence intensity calculation method, model training method, device, equipment and medium, which can improve the accuracy of executing business using the obtained turbulence intensity.

[0005] In a first aspect, an embodiment of the present application provides a method for calculating turbulence intensity, including: obtaining measured geographic data of a remote sensing wind sensing device and measured wind resource data of the remote sensing wind sensing device; inputting the measured geographic data and the measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain a target wind speed standard deviation of a wind measuring tower; and calculating the turbulence intensity at the location of the remote sensing wind sensing device based on the target wind speed standard deviation of the wind measuring tower.

[0006] In some possible embodiments, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs; the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, standard deviation of wind speed corresponding to the altitude area, and wind shear index.

[0007] In some possible embodiments, the measured wind resource data includes a wind shear index, and obtaining the measured wind resource data of a remote sensing wind sensing device includes: obtaining the height area collected by the remote sensing wind sensing device and the average wind speed corresponding to the height area; using the height area and the average wind speed corresponding to the height area to perform fitting calculations to obtain the wind shear index.

[0008] In some possible embodiments, the wind speed standard deviation conversion model includes multiple sub-models, and the measured geographic data and the measured wind resource data are input into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard of the wind measurement tower, including: inputting the measured geographic data and the measured wind resource data into each sub-model respectively to obtain the wind speed standard deviation of the wind measurement tower output by each sub-model; and performing data integration calculation on the wind speed standard deviation of the wind measurement tower output by each sub-model to obtain the target wind speed standard deviation.

[0009] In a second aspect, an embodiment of the present application provides a model training method, including: obtaining multiple training data groups, one training data group includes measured geographic data of a remote sensing wind sensing device, measured wind resource data of a remote sensing wind sensing device, and measured wind speed standard deviation of a wind tower under the same conditions; using multiple training data groups to train a machine learning model to obtain a wind speed standard deviation conversion model.

[0010] In some possible embodiments, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs; the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, standard deviation of wind speed corresponding to the altitude area, and wind shear index.

[0011] In some possible embodiments, the measured wind resource data includes a wind shear index. Before obtaining multiple training data groups, the method further includes: obtaining a height area and an average wind speed value corresponding to the height area; performing a fitting calculation on the height area and the average wind speed value corresponding to the height area to obtain a wind shear index.

[0012] In some possible embodiments, the machine learning model includes multiple machine learning sub-models, and the machine learning model is trained using multiple training data groups to obtain a wind speed standard deviation conversion model, including: using multiple training data groups to train multiple machine learning sub-models separately to obtain multiple candidate conversion models; using test data groups to cross-validate the multiple candidate conversion models separately to obtain conversion performance indicators of the multiple candidate conversion models, and the conversion performance indicators are used to characterize the accuracy of the wind speed standard deviation of the wind measurement tower output by the candidate model; selecting a target candidate conversion model from the multiple candidate conversion models, the target candidate conversion model being the N candidate conversion models with the highest accuracy characterized by the conversion performance indicators, where N is an integer greater than 1; and obtaining a wind speed standard deviation conversion model based on the target candidate conversion model and the data integration calculation algorithm.

[0013] In some possible embodiments, a wind speed standard deviation conversion model is obtained based on a target candidate conversion model and a data integration calculation algorithm, including: obtaining model parameters of the target candidate conversion model; arranging and combining candidate values ​​of the model parameters to obtain multiple parameter value groups; cross-validating the target candidate conversion model using a test data group and multiple parameter value groups to obtain conversion performance indicators of each target candidate conversion model; configuring the parameter value group with the highest accuracy representing the conversion performance indicator of the target candidate conversion model as the actual value of the model parameter of the target candidate conversion model; obtaining a wind speed standard deviation conversion model based on the target candidate conversion model and the data integration calculation algorithm configured with the actual model parameters.

[0014] In a third aspect, an embodiment of the present application provides a turbulence intensity calculation device, including: an acquisition module for acquiring measured geographic data of a remote sensing wind sensing device and measured wind resource data of the remote sensing wind sensing device; a processing module for inputting the measured geographic data and the measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the wind measuring tower; and a calculation module for calculating the turbulence intensity at the location of the remote sensing wind sensing device based on the target wind speed standard deviation of the wind measuring tower.

[0015] In some possible embodiments, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs; the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, standard deviation of wind speed corresponding to the altitude area, and wind shear index.

[0016] In some possible embodiments, the measured wind resource data includes a wind shear index, and the acquisition module is further used to: obtain the height area collected by the remote sensing wind sensing device and the average wind speed corresponding to the height area; use the height area and the average wind speed corresponding to the height area to perform fitting calculations to obtain the wind shear index.

[0017] In some possible embodiments, the wind speed standard deviation conversion model includes multiple sub-models, and the processing module is specifically used to: input the measured geographic data and the measured wind resource data into each sub-model respectively to obtain the wind speed standard deviation of the wind measurement tower output by each sub-model; perform data integration calculation on the wind speed standard deviation of the wind measurement tower output by each sub-model to obtain the target wind speed standard deviation.

[0018] In a fourth aspect, an embodiment of the present application provides a model training device, including: an acquisition module for acquiring multiple training data groups, where one training data group includes measured geographic data of a remote sensing wind sensing device, measured wind resource data of a remote sensing wind sensing device, and measured wind speed standard deviation of a wind tower under the same conditions; a training module for training a machine learning model using multiple training data groups to obtain a wind speed standard deviation conversion model.

[0019] In some possible embodiments, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs; the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, standard deviation of wind speed corresponding to the altitude area, and wind shear index.

[0020] In some possible embodiments, the measured wind resource data includes a wind shear index, and the acquisition module is further used to: obtain the height area and the average wind speed corresponding to the height area; perform fitting calculation on the height area and the average wind speed corresponding to the height area to obtain the wind shear index.

[0021] In some possible embodiments, the machine learning model includes multiple machine learning sub-models, and the training module is specifically used to: use multiple training data groups to train multiple machine learning sub-models separately to obtain multiple candidate conversion models; use test data groups to cross-validate multiple candidate conversion models separately to obtain conversion performance indicators of multiple candidate conversion models, and the conversion performance indicators are used to characterize the accuracy of the wind speed standard deviation of the wind measurement tower output by the candidate model; select a target candidate conversion model from the multiple candidate conversion models, and the target candidate conversion model is the N candidate conversion models with the highest accuracy characterized by the conversion performance indicators, where N is an integer greater than 1; and obtain a wind speed standard deviation conversion model based on the target candidate conversion model and the data integration calculation algorithm.

[0022] In some possible embodiments, the training module is specifically used to: obtain model parameters of the target candidate conversion model; arrange and combine the candidate values ​​of the model parameters to obtain multiple parameter value groups; use the test data group and the multiple parameter value groups to cross-validate the target candidate conversion model respectively to obtain the conversion performance index of each target candidate conversion model; configure the parameter value group with the highest accuracy in representing the conversion performance index of the target candidate conversion model as the actual value of the model parameter of the target candidate conversion model; obtain the wind speed standard deviation conversion model based on the target candidate conversion model configured with the actual model parameters and the data integration calculation algorithm.

[0023] In a fifth aspect, an embodiment of the present application provides a turbulence intensity calculation device, comprising: a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein when the program or instruction is executed by the processor, the turbulence intensity calculation method in the technical solution of the first aspect is implemented.

[0024] In the sixth aspect, an embodiment of the present application provides a wind tower conversion model training device, comprising: a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein when the program or instruction is executed by the processor, the model training method in the technical solution of the second aspect is implemented.

[0025] In the seventh aspect, an embodiment of the present application provides a readable storage medium, which stores a program or instruction. When the program or instruction is executed by a processor, it implements the turbulence intensity calculation method in the technical solution of the first aspect or the model training method in the technical solution of the second aspect.

[0026] The embodiments of the present application provide a turbulence intensity calculation method, model training method, device, equipment and medium, which can input the measured geographic data and measured wind resources of the remote sensing wind sensing device obtained into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the corresponding wind tower. Since the pre-established wind speed standard deviation model is obtained by training using multiple training data groups, each training data group includes the measured geographic data of the remote sensing wind sensing device, the measured wind resource data of the remote sensing wind sensing device and the measured wind speed standard deviation of the wind tower under the same conditions. Therefore, the measured geographic data of the remote sensing wind sensing device and the measured wind resource data of the remote sensing wind sensing device can be converted into the wind speed standard deviation of the wind tower through the wind speed standard deviation model. The wind speed standard deviation of the wind tower can be used to calculate the turbulence intensity at the location of the remote sensing wind sensing device that is adapted to the wind tower, so that the turbulence intensity is applicable to the business standards formulated based on the wind tower wind measurement method, thereby improving the accuracy of executing business using turbulence intensity. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present application can be better understood from the following description of the specific embodiments of the present application in conjunction with the accompanying drawings, wherein the same or similar reference numerals represent the same or similar features.

[0028] Figure 1 Schematic diagram comparing turbulence intensities measured and calculated by lidar wind measurement and wind tower wind measurement in a mountainous environment in an embodiment of the present application;

[0029] Figure 2 This is a flow chart of the model training method in one embodiment of the present application;

[0030] Figure 3 This is a flowchart of a model training method in another embodiment of the present application;

[0031] Figure 4 Schematic diagram of comparison of turbulence intensities measured and calculated in various ways under the same conditions in the embodiments of the present application;

[0032] Figure 5 This is a flow chart of a method for calculating turbulence intensity in one embodiment of the present application;

[0033] Figure 6 This is a flow chart of a method for calculating turbulence intensity in another embodiment of the present application;

[0034] Figure 7 This is a structural diagram of a model training device in one embodiment of the present application;

[0035] Figure 8 This is a schematic structural diagram of a turbulence intensity calculation device in one embodiment of the present application;

[0036] Figure 9 A schematic diagram of the structure of a model training device provided in one embodiment of the present application;

[0037] Figure 10 A schematic diagram of the structure of a turbulence intensity calculation device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0038] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In the detailed description below, many specific details are proposed to provide a comprehensive understanding of the application. However, it will be apparent to those skilled in the art that the application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the application by illustrating the examples of the present application. The application is by no means limited to any specific configuration and algorithm proposed below, but covers any modification, replacement and improvement of elements, parts and algorithms without departing from the spirit of the application. In the accompanying drawings and the following description, known structures and technologies are not shown to avoid causing unnecessary ambiguity to the application.

[0039] Wind power generation converts wind energy into electrical energy. To better monitor wind resources and implement wind power generation, it is necessary to obtain turbulence intensity, which describes the temporal and spatial variations in wind speed and reflects the relative strength of fluctuating wind speeds. Turbulence intensity can be measured and calculated using wind measurement techniques. Specifically, wind measurement methods include tower wind measurement and remote sensing wind measurement. Tower wind measurement uses a tower to measure wind parameters. A tower is a tall structure used to measure wind parameters, specifically a tower-shaped structure used to observe and record near-ground airflow. Remote sensing wind measurement uses remote wind sensing devices. Remote wind sensing devices can include lidar or ultrasonic wind measurement devices. For example, lidar wind measurement tracks a rising balloon drifting with the wind by emitting pulse waves and receiving return pulses from the target. The balloon's trajectory in space is measured to determine the wind direction and horizontal wind speed at various altitudes in the free atmosphere. Since the wind measurement principles of wind tower and remote sensing are different, the turbulence intensity calculated based on wind tower measurement technology is different from the turbulence intensity calculated based on remote sensing wind measurement. Figure 1 This is a schematic diagram comparing the turbulence intensity calculated by the laser radar wind measurement and the wind tower wind measurement in a mountainous environment in the embodiment of the present application. Figure 1 As shown, the horizontal axis represents wind speed, and the vertical axis represents turbulence intensity. The turbulence intensity calculated using lidar wind measurement and wind tower measurements in mountainous environments differs significantly. However, subsequent standards for using turbulence intensity for capacity assessment or adjustment control are based on wind tower measurements and are not applicable to turbulence intensity calculated using remote sensing wind measurements. Using business standards based on wind tower measurements and turbulence intensity calculated using remote sensing wind measurements to execute business operations will reduce business accuracy.

[0040] The embodiments of the present application provide a turbulence intensity calculation method, model training method, device, equipment and medium, which can convert the data measured by the remote sensing wind sensing device into the wind speed standard deviation of the wind tower corresponding to the data measured by the remote sensing wind sensing device through a pre-established wind speed standard deviation conversion model, and use the converted wind speed standard deviation of the wind tower to calculate the turbulence intensity, so that the turbulence intensity is applicable to the business standards formulated based on the wind tower wind measurement method, thereby improving the accuracy of executing business using turbulence intensity.

[0041] The wind speed standard deviation model is trained using multiple training data sets. When data measured by a remote wind sensing device is input into the trained wind speed standard deviation conversion model, the model outputs the wind speed standard deviation at the wind tower corresponding to the data measured by the remote wind sensing device. The following details the method for training the wind speed standard deviation conversion model.

[0042] The present invention provides a model training method that can be executed by a model training device. The model training method can be used to train a wind speed standard deviation conversion model. Geographic data and wind resource data from a remote sensing wind sensor are input into the wind speed standard deviation conversion model, which then outputs the wind speed standard deviation of a wind tower.

[0043] Figure 2 This is a flow chart of the model training method in one embodiment of the present application. Figure 2 As shown, the model training method may include step S101 and step S102.

[0044] In step S101 , a plurality of training data sets are obtained.

[0045] The training data group is used to train a wind speed standard deviation conversion model. A training data group includes the measured geographical data of the remote sensing wind sensing device, the measured wind resource data of the remote sensing wind sensing device, and the measured wind speed standard deviation of the wind tower under the same conditions. The same conditions here specifically refer to the same environmental conditions and the same time, etc. The environmental conditions may include geographical conditions, meteorological conditions, etc., which are not limited here. That is, the measured geographical data, measured wind resource data, and the measured wind speed standard deviation of the wind tower in each training data group are all collected under the same conditions. Each training data group can be a data group collected within a predetermined time period, for example, a training data group is collected every 10 minutes. The collected multiple training data groups are collected and used to train a wind speed standard deviation conversion model. The more training data groups there are, the higher the accuracy of the wind speed standard deviation conversion model obtained through training, and as many training data groups as possible should be obtained.

[0046] The measured geographic data of the remote wind sensing device is geographic data of the location of the remote wind sensing device, and is used to characterize the geographic conditions of the location of the remote wind sensing device. In some examples, the measured geographic data may include one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step in the sector to which the wind direction belongs.

[0047] The terrain description is used to describe the terrain. Terrain roughness data is used to characterize the roughness of the terrain. The RIX (Ruggedness Index) index of the wind direction sector is used to characterize the complexity of the terrain in the sector to which the wind direction belongs. The slope of the wind direction sector is the slope of the sector to which the wind direction belongs. The maximum step of the wind direction sector is used to characterize the terrain in the sector to which the wind direction belongs. The wind direction sector is the sector to which the wind direction belongs within a predetermined time period.

[0048] The wind resource data measured by the remote sensing wind sensing device is characteristic data of the wind resource measured by the remote sensing wind sensing device, and is used to characterize the characteristics of the wind resource measured by the remote sensing wind sensing device. In some examples, the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude region, average wind speed corresponding to the altitude region, wind direction corresponding to the altitude region, standard deviation of wind speed corresponding to the altitude region, and wind shear index.

[0049] An altitude zone is defined by two altitudes, such as the 10-20 meter altitude zone. The average wind speed for an altitude zone is the average wind speed for a predetermined period within that altitude zone. The standard deviation of wind speed for an altitude zone is the standard deviation of wind speed for a predetermined period within that altitude zone. The wind shear index characterizes the cross-horizontal wind speed variation perpendicular to the wind direction.

[0050] In some examples, when the measured wind resource data includes a wind shear index, the wind shear index may be calculated before step S101. Specifically, the wind speed average values ​​corresponding to the height region and the height region may be obtained, and the wind speed average values ​​corresponding to the height region and the height region may be fitted to obtain the wind shear index. For example, the fitting calculation may be performed using the following formula (1):

[0051] U(z)=βz α (1)

[0052] Where U(z) is the average wind speed corresponding to the altitude region, z is the altitude region (i.e., the height of the altitude region), β is the intercept term in the fitting calculation, and α is the wind shear index. The wind shear index can be calculated through fitting calculations.

[0053] In the embodiment of the present application, the measured geographic data and the measured wind resource data of the preset time period can be combined into one piece of data, and the combined data is a training data set. Specifically, the training data set can be implemented in the form of a set, array or vector, which is not limited here.

[0054] In step S102, a machine learning model is trained using multiple training data sets to obtain a wind speed standard deviation conversion model.

[0055] In some examples, a machine learning model may be trained using multiple training data sets to obtain a wind speed standard deviation conversion model. In other examples, two or more machine learning models may be trained using multiple training data sets, and the wind speed standard deviation conversion model may be obtained based on the model with the highest trained output accuracy. In still other examples, two or more machine learning models may be trained using multiple training data sets, and the wind speed standard deviation conversion model may be obtained based on the two or more models with the highest trained output accuracy.

[0056] The machine learning model may specifically include a Support Vector Regression (SVR) model, a Multi-Layer Perceptron (MLP) model, a Random Forest (i.e., Random Forest) model, a Linear Regression (i.e., Linear Regression) model, an XGBoost model, a LightGBM model, or a CatBoost model, etc., which are not limited here. The XGBoost model is an optimized distributed gradient boosting library model. The LightGBM model is a fast, distributed, high-performance gradient boosting framework model based on the decision tree algorithm. The CatBoost model is a gradient boosting algorithm library model that can handle categorical features well.

[0057] Different types of machine learning models may require standardized formats for the data used to train the models. Accordingly, before training the machine learning models using multiple training data sets, the data within the multiple training data sets may be standardized. Specifically, the measured geographic data and measured wind resource data within the training data sets may be standardized separately to meet the requirements of the machine learning models. In step S102, the machine learning models may be trained using the training data sets containing the standardized data.

[0058] In some examples, character-type data can be encoded to achieve standardization. For example, when the measured geographic data includes a terrain description, the terrain description can be specifically mountainous, flat, and plateau. The terrain description can be standardized and encoded, such as using a one-hot encoding (i.e., OneHotEncoder) method to encode mountains as 100, flat land as 010, and plateau as 001, so as to facilitate the use of terrain descriptions to train machine learning models. In other examples, the z-score standardization algorithm or other standardization algorithms can be used for numerical data to perform standardization. For example, the calculation of the z-score standardization algorithm is shown in the following formula (2):

[0059]

[0060] Among them, x * is the data after standardization, x is the data before standardization, μ is the mean value of the data, and σ is the standard deviation of the data.

[0061] In an embodiment of the present application, a machine learning model is trained using multiple sets of geographical data measured by a remote sensing wind sensing device, wind resource data measured by a remote sensing wind sensing device, and wind speed standard deviation measured by a wind tower under the same conditions to obtain a wind speed standard deviation conversion model. This achieves the establishment of a corresponding relationship between the geographical data measured by a remote sensing wind sensing device, the wind resource data measured by a remote sensing wind sensing device, and the wind speed standard deviation measured by a wind tower. The trained wind speed standard deviation conversion model is capable of outputting the corresponding wind speed standard deviation of a wind tower when the geographical data of a remote sensing wind sensing device and the wind resource data of a remote sensing wind sensing device are input, thereby achieving the conversion between the geographical data of a remote sensing wind sensing device, the wind resource data of a remote sensing wind sensing device, and the wind speed standard deviation of a wind tower. The wind speed standard deviation of a wind tower output by the wind speed standard deviation conversion model trained in the embodiment of the present application is used to calculate turbulence intensity, thereby making the turbulence intensity applicable to the business standards formulated based on the wind tower wind measurement method, thereby improving the accuracy of executing business using turbulence intensity while improving the turbulence intensity. Moreover, the wind speed standard deviation conversion model trained in the embodiment of the present application also enables the remote sensing wind sensing method to be widely used, which is more in line with the application requirements of wind resource assessment in the wind power generation field.

[0062] In order to further improve the accuracy of the trained wind speed standard deviation conversion model, a machine model including multiple machine learning sub-models can be used for training to obtain a wind speed standard deviation conversion model with higher accuracy. Figure 3 This is a flowchart of a model training method in another embodiment of the present application. Figure 3 and Figure 2 The difference is that Figure 2 The step S102 shown can be specifically broken down into Figure 3Steps S1021 to S1024 are shown.

[0063] In step S1021, multiple machine learning sub-models are trained separately using multiple training data groups to obtain multiple candidate conversion models.

[0064] The multiple machine learning sub-models are different machine learning sub-models. The types of machine learning sub-models are not limited here. For example, the machine learning sub-models may include an SVR model, an MLP model, a Random Forest model, a LinearRegression model, an XGBoost model, a LightGBM model, or a CatBoost model.

[0065] Multiple training data sets are input into the machine learning sub-models, and the multiple machine learning sub-models are trained separately. The training data sets used for training different machine learning sub-models can be the same or different, and this is not limited here. After training, each machine learning sub-model can generate a corresponding candidate transformation model.

[0066] In step S1022, a plurality of candidate conversion models are cross-validated using the test data set to obtain conversion performance indicators of the plurality of candidate conversion models.

[0067] The test dataset is used to test the accuracy of the trained model. A test dataset includes geographic data from a remote wind sensor, wind resource data from the remote wind sensor, and the standard deviation of wind speeds measured by a wind tower under identical conditions. For a machine learning sub-model, the data in the test dataset differs from the data in the training dataset.

[0068] In some examples, groups of geographical data measured by remote wind sensing devices, wind resource data measured by remote wind sensing devices, and wind speed standard deviations measured by wind towers can be pre-collected. Each data group includes geographical data measured by remote wind sensing devices, wind resource data measured by remote wind sensing devices, and wind speed standard deviations measured by wind towers under identical conditions. While not limiting, some of the multiple data groups can be used as training data groups, while others can be used as testing data groups.

[0069] In some examples, cross-validation may be performed using a 5-fold cross-validation method, but is not limited thereto.

[0070] By cross-validating multiple candidate conversion models, a conversion performance index can be obtained for each candidate conversion model. The conversion performance index is used to characterize the accuracy of the standard deviation of wind speeds at the wind tower output by the candidate model. Specifically, the measured geographic data and wind resource data from the remote sensing wind sensing device in the test dataset are input into the candidate conversion model, which then outputs the converted standard deviation of wind speeds at the wind tower. The conversion performance index can be used to characterize the degree of similarity between the converted standard deviation of wind speeds at the wind tower and the measured standard deviation of wind speeds at the wind tower in the test dataset.

[0071] In some examples, the conversion performance indicator can be a mean-square error (MSE). MSE can reflect the degree of difference between the output value of the model and the actual value. The smaller the MSE value, the higher the accuracy of the model. The calculation of MSE is shown in the following formula (3):

[0072]

[0073] Where MSE is the value of MSE, m is the number of test data groups, y i is the standard deviation of wind speed of the actual wind tower in the test data set, The standard deviation of wind speed at the wind mast output by the candidate conversion model.

[0074] In step S1023 , a target candidate conversion model is selected from a plurality of candidate conversion models.

[0075] The target candidate conversion model is the N candidate conversion models with the highest conversion performance index accuracy, where N is an integer greater than 1. That is, the N candidate conversion models with the highest conversion performance index accuracy are selected from the multiple candidate conversion models as the target candidate conversion models. A higher conversion performance index accuracy indicates a higher accuracy of the candidate conversion model.

[0076] In step S1024, a wind speed standard deviation conversion model is obtained based on the target candidate conversion model and the data integration calculation algorithm.

[0077] The target candidate conversion model includes N candidate conversion models. The output data of the wind speed standard deviation conversion model may be the result of integrating the output data of the N candidate conversion models. The algorithm for integrating the output data of the N candidate conversion models is a data integration calculation algorithm. For example, the data integration calculation algorithm may be an average algorithm or a weighted algorithm, etc., although this is not limited here.

[0078] By selecting the N most accurate machine learning sub-models from multiple machine learning sub-models, and using the N most accurate machine learning sub-models and a data integration calculation algorithm to obtain a wind speed standard deviation conversion model, the accuracy of the trained wind speed standard deviation conversion model can be improved, and the accuracy of the data output by the wind speed standard deviation conversion model can be improved.

[0079] To further improve the accuracy of the trained wind speed standard deviation conversion model, the model parameters of the target candidate model can be optimized. Specifically, step S1023 can be further refined as follows: obtaining the model parameters of the target candidate conversion model; arranging and combining the candidate values ​​of the model parameters to obtain multiple parameter value groups; cross-validating the target candidate conversion model using the test data group and the multiple parameter value groups to obtain the conversion performance index of each target candidate conversion model; configuring the parameter value group with the highest accuracy of the conversion performance index of the target candidate conversion model as the actual value of the model parameters of the target candidate conversion model; and obtaining the wind speed standard deviation conversion model based on the target candidate conversion model configured with the actual model parameters and the data integration calculation algorithm.

[0080] The model parameters of the target candidate conversion model are related to the target candidate conversion model. The model parameters of different target candidate conversion models may be different and are not limited here. For example, if the target candidate conversion model includes an XGBoost model, the model parameters of the XGBoost model may include max_depth, min_child_weight, gamma, and reg_lambda. For another example, if the target candidate conversion model includes a LightGBM model, the model parameters of the LightGBM model may include num_leaves, max_depth, min_child_weight, reg_alpha, and reg_lambda. For another example, if the target candidate conversion model includes a CatBoost model, the model parameters of the CatBoost model may include deepth and l2_leaf_reg.

[0081] There can be multiple candidate values ​​for each model parameter. The candidate values ​​for each model parameter can be arranged and combined to obtain all possible combinations of candidate values ​​for the model parameters, namely parameter value groups. The target candidate conversion model with different parameter value groups is tested using the test data group to obtain conversion performance indicators for the target candidate conversion model corresponding to the different parameter value groups. The higher the accuracy of the conversion performance indicator representation, the higher the accuracy of the target candidate conversion model using the parameter value group. The target candidate conversion model configured with the parameter value group that represents the highest accuracy is the target candidate conversion model with the highest accuracy.

[0082] The following example illustrates this. For example, a machine learning model includes seven machine learning sub-models, namely, an SVR model, an MLP model, a Random Forest model, a Linear Regression model, an XGBoost model, a LightGBM model, and a CatBoost model. Each machine learning sub-model is trained using a training data set. The seven trained machine learning sub-models are cross-validated using the training data set to obtain the mean square error (MSE) of the seven trained machine learning sub-models. The three machine learning models with the smallest MSE values, namely, the XGBoost model, the LightGBM model, and the CatBoost model, are selected as candidate target conversion models. The candidate values ​​for the model parameters of the XGBoost model, the LightGBM model, and the CatBoost model are permuted and combined. The parameter value groups are substituted into the XGBoost model, the LightGBM model, and the CatBoost model, respectively. Based on the MSE of the XGBoost model, the LightGBM model, and the CatBoost model after the parameter value groups are substituted into the model, the optimal parameter value group is obtained, which serves as the actual values ​​for the model parameters of the XGBoost model, the LightGBM model, and the CatBoost model. The average value algorithm is used as the data integration calculation algorithm, that is, the output data of multiple target candidate conversion models are averaged as the output data of the trained wind speed standard deviation conversion model.

[0083] In order to illustrate the high accuracy of turbulence intensity calculated using the wind speed standard deviation output by the wind speed standard deviation conversion model trained in the embodiment of the present application, an example is given below. Figure 4 Schematic diagram of the comparison of turbulence intensity measured and calculated in various ways under the same conditions in the embodiment of this application. Figure 4 As shown, method one is the turbulence intensity calculated by lidar measurement; method two is the turbulence intensity calculated by converting the wind speed standard deviation output by the wind speed standard deviation converted by the XGBoost model training; method three is the turbulence intensity calculated by converting the wind speed standard deviation output by the wind speed standard deviation converted by the LightGBM model training; method four is the turbulence intensity calculated by converting the wind speed standard deviation output by the wind speed standard deviation converted by the CatBoost model training; method five is the turbulence intensity calculated by converting the wind speed standard deviation output by the wind speed standard deviation converted by the model training including the XGBoost model, the LightGBM model and the CatBoost model.

[0084] Table 1 shows the MSE values ​​corresponding to the XGBoost model, LightGBM model, CatBoost model, and the wind speed standard deviation conversion model obtained by training the model including the XGBoost model, LightGBM model, and CatBoost model under the same conditions. Figure 4 As can be seen from Table 1, the turbulence intensity corresponding to method 5 is closest to the actual turbulence intensity. The wind speed standard deviation conversion model obtained by training the models including the XGBoost model, the LightGBM model and the CatBoost model has the highest accuracy. It can be proved that the accuracy of the turbulence intensity calculated by the wind speed standard deviation output by the wind speed standard deviation conversion model including multiple sub-models is improved.

[0085] The embodiment of the present application further provides a method for calculating turbulence intensity, which can calculate turbulence intensity by using the target wind speed standard deviation of the wind tower output by the wind speed standard deviation conversion model trained in the above embodiment.

[0086] Figure 5 FIG. 1 is a flow chart of a method for calculating turbulence intensity in an embodiment of the present application. Figure 5 As shown, the turbulence intensity calculation method may include steps S201 to S203.

[0087] In step S201 , the measured geographical data and the measured wind resource data of the remote sensing wind sensing device are acquired.

[0088] The measured geographic data of the remote wind sensing device is the geographic data of the location of the remote wind sensing device. The measured geographic data differs from the measured geographic data in the above-mentioned embodiments in that the measured geographic data is the data to be measured. When the wind speed standard deviation of a wind tower is unknown, the measured geographic data is input into the wind speed standard deviation conversion model to obtain the target wind speed standard deviation for the corresponding wind tower.

[0089] In some examples, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step in the sector to which the wind direction belongs. For a detailed description of the measured geographic data, please refer to the description of the measured geographic data in the above embodiment and will not be repeated here.

[0090] The measured wind resource data of the remote sensing wind sensing device is characteristic data of the wind resource measured by the remote sensing wind sensing device and is used to characterize the characteristics of the wind resource measured by the remote sensing wind sensing device. The measured wind resource data differs from the measured wind resource data in the above-described embodiments in that the measured wind resource data is the data to be measured. When the wind speed standard deviation of a wind tower is unknown, the measured wind resource data is input into a wind speed standard deviation conversion model to obtain the target wind speed standard deviation for the corresponding wind tower.

[0091] In some examples, the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude region, average wind speed corresponding to the altitude region, wind direction corresponding to the altitude region, standard deviation of wind speed corresponding to the altitude region, and wind shear index. For a detailed description of the measured wind resource data, please refer to the description of the measured wind resource data in the above embodiment and will not be repeated here.

[0092] In some examples, when the measured wind resource data includes a wind shear index, the above-mentioned steps of obtaining the measured wind resource data of the remote sensing wind sensing device can be refined as follows: obtaining the height area collected by the remote sensing wind sensing device and the average wind speed corresponding to the height area; using the height area and the average wind speed corresponding to the height area to perform fitting calculations to obtain the wind shear index.

[0093] The wind shear index and the calculation method of the wind shear index can be found in the relevant descriptions in the above embodiments, which will not be repeated here.

[0094] In some examples, the measured geographic data and the measured wind resource data of a preset time period can be combined into one piece of data. Specifically, the combined data can be implemented in the form of a set, an array, or a vector, which is not limited here.

[0095] In step S202, the measured geographic data and the measured wind resource data are input into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the wind tower.

[0096] A pre-established wind speed standard deviation conversion model inputs measured geographic data and measured wind resource data and outputs a target wind speed standard deviation for the corresponding wind tower. This wind speed standard deviation is trained using multiple training data sets and a machine learning model. The training of the wind speed standard deviation conversion model can be found in the previous examples and will not be further described here.

[0097] The wind speed standard deviation conversion model may require standardized formats for input data. Accordingly, the measured geographic data and the measured wind resource data may be standardized before being input into the wind speed standard deviation conversion model. The specific details of the standardization process can be found in the relevant descriptions of the above embodiments and will not be repeated here.

[0098] In step S203, the turbulence intensity at the location of the remote sensing wind measuring device is calculated according to the target wind speed standard deviation of the wind measuring tower.

[0099] According to the target wind speed standard deviation of the wind tower obtained in step S202 and in combination with a turbulence intensity calculation method adapted for the wind tower, the turbulence intensity adapted for the wind tower at the location of the remote sensing wind sensing device can be calculated.

[0100] In an embodiment of the present application, the measured geographic data and measured wind resources obtained from the remote sensing wind sensing device can be input into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the corresponding wind tower. Since the pre-established wind speed standard deviation model is trained using multiple training data sets, each training data set includes the measured geographic data of the remote sensing wind sensing device, the measured wind resource data of the remote sensing wind sensing device, and the measured wind speed standard deviation of the wind tower under the same conditions. Therefore, the measured geographic data of the remote sensing wind sensing device and the measured wind resource data of the remote sensing wind sensing device can be converted into the wind speed standard deviation of the wind tower using the wind speed standard deviation model. The wind speed standard deviation of the wind tower can be used to calculate the turbulence intensity at the location of the remote sensing wind sensing device that is adapted to the wind tower, so that the turbulence intensity is applicable to the business standards formulated based on the wind tower wind measurement method, thereby improving the accuracy of executing business using turbulence intensity on the basis of improving the turbulence intensity. Moreover, the wind speed standard deviation conversion model trained in the embodiment of the present application also makes the remote sensing wind sensing method popularizable and more in line with the application requirements of wind resource assessment in the wind power generation field.

[0101] Figure 6 Flowchart of a method for calculating turbulence intensity in another embodiment of the present application. The wind speed standard deviation conversion model in the embodiment of the present application includes multiple sub-models. Figure 6 and Figure 5 The difference is that Figure 5 Step S202 in the above example can be specifically broken down into Figure 6 Steps S2021 and S2022 in .

[0102] In step S2021, the measured geographic data and the measured wind resource data are input into each sub-model respectively to obtain the wind speed standard deviation of the wind tower output by each sub-model.

[0103] If the wind speed standard deviation conversion model includes multiple sub-models, the measured geographic data and measured wind resource data can be input into each sub-model separately. Each sub-model will output the wind speed standard deviation of the wind tower. The training methods for each sub-model can be found in the training of the target candidate conversion model in the above embodiment and will not be repeated here.

[0104] In step S2022, data integration calculation is performed on the wind speed standard deviation of the wind tower output by each sub-model to obtain the target wind speed standard deviation.

[0105] The data integration calculation is a calculation method for integrating the wind speed standard deviations of the wind towers output by each sub-model. Specifically, an average value algorithm or a weighted algorithm can be used, which is not limited here. In some examples, if the average value algorithm is used to integrate the wind speed standard deviations of the wind towers output by each sub-model, the target wind speed standard deviation output by the wind speed standard deviation conversion model is the average value of the wind speed standard deviations of the wind towers output by each sub-model. For example, the wind speed standard deviation conversion model includes three sub-models, and the wind speed standard deviations of the wind towers output by the three sub-models are A1, A2, and A3, respectively. Then the target wind speed standard deviation output by the wind speed standard deviation conversion model = 1 / 3 × (A1 + A2 + A3).

[0106] The embodiment of the present application also provides a model training device. Figure 7 This is a schematic diagram of the structure of the model training device in one embodiment of the present application. Figure 7 As shown, the model training device 300 may include an acquisition module 301 and a training module 302.

[0107] The acquisition module 301 may be used to acquire multiple training data sets.

[0108] A training data set includes geographical data measured by a remote sensing wind instrument, wind resource data measured by the remote sensing wind instrument, and standard deviation of wind speed measured by a wind tower under the same conditions.

[0109] The training module 302 may be used to train a machine learning model using multiple training data sets to obtain a wind speed standard deviation conversion model.

[0110] In an embodiment of the present application, a machine learning model is trained using multiple sets of geographical data measured by a remote sensing wind sensing device, wind resource data measured by a remote sensing wind sensing device, and wind speed standard deviation measured by a wind tower under the same conditions to obtain a wind speed standard deviation conversion model. This achieves the establishment of a corresponding relationship between the geographical data measured by a remote sensing wind sensing device, the wind resource data measured by a remote sensing wind sensing device, and the wind speed standard deviation measured by a wind tower. The trained wind speed standard deviation conversion model is capable of outputting the corresponding wind speed standard deviation of a wind tower when the geographical data of a remote sensing wind sensing device and the wind resource data of a remote sensing wind sensing device are input, thereby achieving the conversion between the geographical data of a remote sensing wind sensing device, the wind resource data of a remote sensing wind sensing device, and the wind speed standard deviation of a wind tower. The wind speed standard deviation of a wind tower output by the wind speed standard deviation conversion model trained in the embodiment of the present application is used to calculate turbulence intensity, thereby making the turbulence intensity applicable to the business standards formulated based on the wind tower wind measurement method, thereby improving the accuracy of executing business using turbulence intensity while improving the turbulence intensity. Moreover, the wind speed standard deviation conversion model trained in the embodiment of the present application also enables the remote sensing wind sensing method to be widely used, which is more in line with the application requirements of wind resource assessment in the wind power generation field.

[0111] In some examples, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step in the sector to which the wind direction belongs.

[0112] In some examples, the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude region, average wind speed corresponding to the altitude region, wind direction corresponding to the altitude region, standard deviation of wind speed corresponding to the altitude region, and wind shear index.

[0113] In some examples, the measured wind resource data includes a wind shear index. The acquisition module 301 can also be used to: acquire a height region and an average wind speed corresponding to the height region; and perform a fitting calculation on the height region and the average wind speed corresponding to the height region to obtain a wind shear index.

[0114] In some examples, the machine learning model includes multiple machine learning sub-models. The training module 302 can be specifically configured to: train the multiple machine learning sub-models using multiple training data sets to obtain multiple candidate conversion models; cross-validate the multiple candidate conversion models using a test data set to obtain conversion performance indicators for the multiple candidate conversion models, where the conversion performance indicators are used to characterize the accuracy of the wind speed standard deviation of the wind tower output by the candidate models; select a target candidate conversion model from the multiple candidate conversion models, where the target candidate conversion model is the N candidate conversion models with the highest accuracy represented by the conversion performance indicators, where N is an integer greater than 1; and obtain a wind speed standard deviation conversion model based on the target candidate conversion model and a data integration calculation algorithm.

[0115] In some examples, the training module 302 can be specifically used to: obtain model parameters of the target candidate conversion model; arrange and combine the candidate values ​​of the model parameters to obtain multiple parameter value groups; use the test data group and the multiple parameter value groups to cross-validate the target candidate conversion model respectively to obtain the conversion performance index of each target candidate conversion model; configure the parameter value group with the highest accuracy in representing the conversion performance index of the target candidate conversion model as the actual value of the model parameter of the target candidate conversion model; obtain the wind speed standard deviation conversion model based on the target candidate conversion model configured with the actual model parameters and the data integration calculation algorithm.

[0116] An embodiment of the present application also provides a turbulence intensity calculation device. Figure 8 FIG. 1 is a schematic diagram of the structure of a turbulence intensity calculation device in one embodiment of the present application. Figure 8 As shown, the turbulence intensity calculation device 400 may include an acquisition module 401, a processing module 402 and a calculation module 403.

[0117] The acquisition module 401 can be used to acquire the measured geographical data of the remote sensing wind sensing device and the measured wind resource data of the remote sensing wind sensing device;

[0118] The processing module 402 may be used to input the measured geographic data and the measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the wind tower;

[0119] The calculation module 403 may be used to calculate the turbulence intensity at the location of the remote sensing wind measuring device according to the target wind speed standard deviation of the wind measuring tower.

[0120] In an embodiment of the present application, the measured geographic data and measured wind resources obtained from the remote sensing wind sensing device can be input into a pre-established wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the corresponding wind tower. Since the pre-established wind speed standard deviation model is trained using multiple training data sets, each training data set includes the measured geographic data of the remote sensing wind sensing device, the measured wind resource data of the remote sensing wind sensing device, and the measured wind speed standard deviation of the wind tower under the same conditions. Therefore, the measured geographic data of the remote sensing wind sensing device and the measured wind resource data of the remote sensing wind sensing device can be converted into the wind speed standard deviation of the wind tower using the wind speed standard deviation model. The wind speed standard deviation of the wind tower can be used to calculate the turbulence intensity at the location of the remote sensing wind sensing device that is adapted to the wind tower, so that the turbulence intensity is applicable to the business standards formulated based on the wind tower wind measurement method, thereby improving the accuracy of executing business using turbulence intensity on the basis of improving the turbulence intensity. Moreover, the wind speed standard deviation conversion model trained in the embodiment of the present application also makes the remote sensing wind sensing method popularizable and more in line with the application requirements of wind resource assessment in the wind power generation field.

[0121] In some examples, the measured geographic data includes one or more of the following: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs.

[0122] In some examples, the measured wind resource data includes one or more of the following: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, standard deviation of wind speed corresponding to the altitude area, and wind shear index.

[0123] In some examples, the measured wind resource data includes a wind shear index. Accordingly, the acquisition module 401 is further configured to acquire a height region and an average wind speed corresponding to the height region collected by a remote wind sensing device; and perform a fitting calculation using the height region and the average wind speed corresponding to the height region to obtain the wind shear index.

[0124] In some examples, the wind speed standard deviation conversion model includes multiple sub-models. Processing module 402 can be specifically configured to: input measured geographic data and measured wind resource data into each sub-model to obtain the wind speed standard deviation of the wind tower output by each sub-model; and perform data integration calculation on the wind speed standard deviation of the wind tower output by each sub-model to obtain the target wind speed standard deviation.

[0125] An embodiment of the present application also provides a model training device. Figure 9 This is a schematic diagram of the structure of the model training device provided in one embodiment of the present application. Figure 9 As shown, the model training device 500 includes a memory 501, a processor 502, and a computer program stored in the memory 501 and executable on the processor 502.

[0126] In an example, the processor 502 may include a central processing unit (CPU) or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0127] The memory 501 may include a large-capacity memory for data or instructions. For example, but not limitation, the memory 501 may include an HDD, a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 501 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 501 may be inside or outside the terminal hotspot to open the model training device 500. In a specific embodiment, the memory 501 is a non-volatile solid-state memory. In a specific embodiment, the memory 501 includes a read-only memory (ROM). In appropriate cases, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM) or a flash memory or a combination of two or more of these.

[0128] The processor 502 runs the program or instructions corresponding to the executable program code by reading the executable program code stored in the memory 501 to implement the model training method in the above embodiment.

[0129] In one example, the model training device 500 may further include a communication interface 503 and a bus 504. Figure 9 As shown, the memory 501 , the processor 502 , and the communication interface 503 are connected via a bus 504 and communicate with each other.

[0130] The communication interface 503 is mainly used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present application. Input devices and / or output devices can also be connected through the communication interface 503.

[0131] Bus 504 includes hardware, software or both, and the parts of model training device 500 are coupled to each other.For example, but not limitation, bus 504 may include accelerated graphics port (AGP) or other graphics bus, enhanced industry standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industry standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. Where appropriate, bus 504 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the application considers any suitable bus or interconnection.

[0132] An embodiment of the present application also provides a turbulence intensity calculation device. Figure 10 This is a schematic diagram of the structure of the turbulence intensity calculation device provided in one embodiment of the present application. Figure 10 As shown, the turbulence intensity calculation device 600 includes a memory 601 , a processor 602 , and a computer program stored in the memory 601 and executable on the processor 602 .

[0133] In an example, the processor 602 may include a central processing unit (CPU) or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0134] The memory 601 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 601 may include an HDD, a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 601 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 601 may be located inside or outside the turbulence intensity calculation device 600 at the terminal hotspot. In a specific embodiment, the memory 601 is a non-volatile solid-state memory. In a specific embodiment, the memory 601 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory, or a combination of two or more of these.

[0135] The processor 602 reads the executable program code stored in the memory 601 to run a program or instruction corresponding to the executable program code, so as to implement the turbulence intensity calculation method in the above embodiment.

[0136] In one example, the turbulence intensity calculation device 600 may further include a communication interface 603 and a bus 604. Figure 10 As shown, the memory 601 , the processor 602 , and the communication interface 603 are connected via a bus 604 and communicate with each other.

[0137] The communication interface 603 is mainly used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present application. Input devices and / or output devices can also be connected through the communication interface 603.

[0138] The bus 604 includes hardware, software, or both, and couples the components of the turbulence intensity calculation device 600 to each other. By way of example and not limitation, the bus 604 may include an accelerated graphics port (AGP) or other graphics bus, an enhanced industry standard architecture (EISA) bus, a front-side bus (FSB), a HyperTransport (HT) interconnect, an industry standard architecture (ISA) bus, an InfiniBand interconnect, a low pin count (LPC) bus, a memory bus, a microchannel architecture (MCA) bus, a peripheral component interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a serial advanced technology attachment (SATA) bus, a video electronics standard association local (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, the bus 604 may include one or more buses. Although the embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.

[0139] The embodiment of the present application also provides a readable storage medium. The readable storage medium stores a program or instruction, which, when executed by the processor, implements the various processes of the above-mentioned model training method embodiment or the various processes of the above-mentioned turbulence intensity calculation method embodiment, and can achieve the same technical effect. To avoid repetition, they are not described here. Among them, the readable storage medium may specifically include a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0140] It should be understood that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. For the apparatus embodiments, device embodiments, and readable storage medium embodiments, the relevant parts can be referred to the description part of the method embodiments. This application is not limited to the specific steps and structures described above and shown in the figures. Those skilled in the art can make various changes, modifications, and additions, or change the order between the steps after understanding the spirit of this application. In addition, for the sake of brevity, a detailed description of known method technologies is omitted here.

[0141] Those skilled in the art should understand that the above embodiments are illustrative rather than restrictive. Different technical features appearing in different embodiments can be combined to achieve beneficial effects. Based on a study of the drawings, the specification and the claims, those skilled in the art should be able to understand and implement other variations of the disclosed embodiments. In the claims, the term "comprising" does not exclude other devices or steps; the quantifier "one" does not exclude a plurality; the terms "first" and "second" are used to identify names rather than to indicate any specific order. Any figure marks in the claims should not be understood as limiting the scope of protection. The functions of multiple parts appearing in the claims can be implemented by a separate hardware or software module. The fact that certain technical features appear in different dependent claims does not mean that these technical features cannot be combined to achieve beneficial effects.

Claims

1. A method for calculating turbulence intensity, characterized in that: include: Obtaining measured geographic data of a remote sensing wind sensing device and measured wind resource data of the remote sensing wind sensing device, wherein the measured geographic data of the remote sensing wind sensing device is geographic data of a location where the remote sensing wind sensing device is located, and the measured wind resource data of the remote sensing wind sensing device is characteristic data of wind resources measured by the remote sensing wind sensing device; Inputting the measured geographic data and the measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain a target wind speed standard deviation of the wind tower, thereby realizing conversion between the geographic data of the remote sensing wind measuring device, the wind resource data of the remote sensing wind measuring device, and the wind speed standard deviation of the wind tower, wherein the wind speed standard deviation conversion model is obtained by training a machine learning model using multiple training data sets, each of the training data sets including the measured geographic data of the remote sensing wind measuring device, the measured wind resource data of the remote sensing wind measuring device, and the measured wind speed standard deviation of the wind tower under the same conditions; The turbulence intensity at the location of the remote sensing wind sensing device is calculated based on the target wind speed standard deviation of the wind measuring tower and in combination with a turbulence intensity calculation method adapted for the wind measuring tower.

2. The method according to claim 1, characterized in that The measured geographic data includes: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs; The measured wind resource data includes: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, and standard deviation of wind speed corresponding to the altitude area; The measured wind resource data also includes a wind shear index.

3. The method according to claim 1, characterized in that The measured wind resource data includes wind shear index, Acquiring the measured wind resource data of the remote sensing wind sensing device includes: Obtaining a height region collected by the remote sensing wind sensing device and an average wind speed corresponding to the height region; The wind shear index is obtained by performing a fitting calculation using the height region and the average wind speed value corresponding to the height region.

4. The method according to claim 1, wherein The wind speed standard deviation conversion model includes multiple sub-models. The step of inputting the measured geographic data and the measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain a target wind speed standard of a wind tower includes: Inputting the measured geographic data and the measured wind resource data into each sub-model respectively, and obtaining the wind speed standard deviation of the wind tower output by each sub-model; The target wind speed standard deviation is obtained by performing data integration calculation on the wind speed standard deviation of the wind tower output by each sub-model.

5. A model training method, characterized in that: include: Acquire multiple training data sets, one of the training data sets including geographical data measured by a remote sensing wind sensor, wind resource data measured by the remote sensing wind sensor, and a standard deviation of wind speed measured by a wind tower under the same conditions; A machine learning model is trained using multiple training data groups to obtain a wind speed standard deviation conversion model. The wind speed standard deviation conversion model realizes the conversion between the geographical data of the remote sensing wind sensing device, the wind resource data of the remote sensing wind sensing device and the wind speed standard deviation of the wind measuring tower. The measured geographical data of the remote sensing wind sensing device and the measured wind resource data of the remote sensing wind sensing device are input into the wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the wind measuring tower. The target wind speed standard deviation is combined with a turbulence intensity calculation method adapted for the wind measuring tower to calculate the turbulence intensity at the location of the remote sensing wind sensing device. The measured geographical data of the remote sensing wind sensing device is the geographical data of the location of the remote sensing wind sensing device, and the measured wind resource data of the remote sensing wind sensing device is the characteristic data of the wind resources measured by the remote sensing wind sensing device.

6. The method according to claim 5, characterized in that The measured geographic data includes: altitude, terrain description, terrain roughness, RIX index of the sector to which the wind direction belongs, slope of the sector to which the wind direction belongs, and maximum step of the sector to which the wind direction belongs; The measured wind resource data includes: humidity, temperature, pressure, altitude area, average wind speed corresponding to the altitude area, wind direction corresponding to the altitude area, and standard deviation of wind speed corresponding to the altitude area; The measured wind resource data also includes a wind shear index.

7. The method according to claim 5, characterized in that The measured wind resource data includes wind shear index, Before acquiring the plurality of training data sets, the method further includes: Obtaining a height region and an average wind speed value corresponding to the height region; The wind shear index is obtained by performing a fitting calculation on the height region and the average wind speed value corresponding to the height region.

8. The method according to claim 5, characterized in that The machine learning model includes multiple machine learning sub-models, The method of training a machine learning model using a plurality of training data sets to obtain a wind speed standard deviation conversion model includes: Using the plurality of training data groups to train the plurality of machine learning sub-models respectively to obtain a plurality of candidate conversion models; Cross-validating the plurality of candidate conversion models using a test data set to obtain a conversion performance index for each candidate conversion model, wherein the conversion performance index is used to characterize the accuracy of the standard deviation of the wind speed of the wind measurement tower output by the candidate conversion model; Selecting a target candidate conversion model from the plurality of candidate conversion models, the target candidate conversion model being the N candidate conversion models with the highest accuracy represented by the conversion performance indicator, where N is an integer greater than 1; The wind speed standard deviation conversion model is obtained according to the target candidate conversion model and the data integration calculation algorithm. The data integration calculation algorithm is an algorithm for integrating and processing the output data of N candidate conversion models.

9. The method according to claim 8, characterized in that The wind speed standard deviation conversion model is obtained based on the target candidate conversion model and the data integration calculation algorithm, including: Obtaining model parameters of the target candidate conversion model; Arrange and combine the candidate values ​​of the model parameters to obtain multiple parameter value groups; Cross-validating the target candidate conversion models using the test data group and the plurality of parameter value groups to obtain the conversion performance index of each target candidate conversion model; configuring the parameter value group with the highest conversion performance indicator representation accuracy of the target candidate conversion model as the actual values ​​of the model parameters of the target candidate conversion model; The wind speed standard deviation conversion model is obtained according to the actual values ​​of the model parameters of the target candidate conversion model.

10. A turbulence intensity calculation device, characterized in that: include: an acquisition module, configured to acquire measured geographic data of a remote sensing wind sensing device and measured wind resource data of the remote sensing wind sensing device, wherein the measured geographic data of the remote sensing wind sensing device is geographic data of a location where the remote sensing wind sensing device is located, and the measured wind resource data of the remote sensing wind sensing device is characteristic data of wind resources measured by the remote sensing wind sensing device; a processing module, configured to input the measured geographic data and the measured wind resource data into a pre-established wind speed standard deviation conversion model to obtain a target wind speed standard deviation of a wind tower, thereby realizing conversion between the geographic data of the remote sensing wind measuring device, the wind resource data of the remote sensing wind measuring device, and the wind speed standard deviation of the wind tower, wherein the wind speed standard deviation conversion model is obtained by training a machine learning model using a plurality of training data sets, each of the training data sets including the measured geographic data of the remote sensing wind measuring device, the measured wind resource data of the remote sensing wind measuring device, and the measured wind speed standard deviation of the wind tower under the same conditions; The calculation module is used to calculate the turbulence intensity at the location of the remote sensing wind sensing device according to the target wind speed standard deviation of the wind measuring tower and a turbulence intensity calculation method adapted for the wind measuring tower.

11. A model training device, characterized in that: include: An acquisition module is configured to acquire a plurality of training data sets, wherein one of the training data sets includes geographical data measured by a remote sensing wind sensing device, wind resource data measured by the remote sensing wind sensing device, and a standard deviation of wind speed measured by a wind tower under the same conditions; A training module is used to train a machine learning model using multiple training data groups to obtain a wind speed standard deviation conversion model, wherein the wind speed standard deviation conversion model realizes the conversion between the geographical data of the remote sensing wind sensing device, the wind resource data of the remote sensing wind sensing device and the wind speed standard deviation of the wind measuring tower. The measured geographical data of the remote sensing wind sensing device and the measured wind resource data of the remote sensing wind sensing device are input into the wind speed standard deviation conversion model to obtain the target wind speed standard deviation of the wind measuring tower. The target wind speed standard deviation is combined with a turbulence intensity calculation method adapted for the wind measuring tower to calculate the turbulence intensity at the location of the remote sensing wind sensing device. The measured geographical data of the remote sensing wind sensing device is the geographical data of the location of the remote sensing wind sensing device, and the measured wind resource data of the remote sensing wind sensing device is the characteristic data of the wind resources measured by the remote sensing wind sensing device.

12. A turbulence intensity calculation device, characterized in that: include: The method comprises a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the turbulence intensity calculation method according to any one of claims 1 to 4.

13. A wind tower conversion model training device, characterized in that: include: It includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the model training method as described in any one of claims 5 to 9.

14. A readable storage medium, characterized in that The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the turbulence intensity calculation method according to any one of claims 1 to 4 or the model training method according to any one of claims 5 to 9 is implemented.

Images