A method for obtaining effective incoming wind speed in front of a wind turbine rotor

Abstract

The present invention provides a method for detecting the effective incoming wind speed in front of a wind rotor of a wind turbine set. The method first obtains the effective incoming wind speed in front of a wind rotor in front of the wind turbine set and the detection data of a SCADA system at the same time above a cabin; secondly, the acquired data is preprocessed and a training set is constructed; then, the wind speed, wind direction, wind angle, power, generator speed and rotor speed at each moment in the training set are taken as an input feature, and each time step is taken as a sample to construct an input variable; the actual effective incoming wind speed data in front of a wind rotor of the wind turbine set at each moment in the training set are taken as an output sequence, and the mapping relationship between the input feature and the output sequence is trained by a LENET‑5 model, so that the actual effective incoming wind speed upstream of the wind turbine set can be more accurately detected without excessive reliance on a vertical wind tower type wind measuring radar in front of the wind turbine set.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation, and in particular to a method for obtaining the effective incoming wind speed in front of a wind rotor of a wind turbine. Background Art

[0002] There are many studies on the accurate acquisition of wind speed in front of the wheel. These include the method of studying the variable pitch control strategy of offshore direct-drive wind turbines based on the KELM wind speed soft sensor; and the use of laser radar wind measurement and SCADA detection data to study the effects of wake and terrain on the performance of wind turbines; but the above studies all rely on wind measurement experiments to measure the actual effective incoming wind speed in front of the wheel. Some studies have proposed a research method for SCADA detection data cleaning based on the Tukey method to improve the fitting of the wind speed power characteristic curve. Although it can improve the fitting accuracy, it cannot avoid the problem of inaccurate SCADA detection data caused by wind speed changes when the incoming wind blows through the wind wheel.

[0003] In addition to experimental methods to obtain the effective wind speed in front of the wind rotor, some researchers want to obtain a more accurate actual effective wind speed in front of the wind rotor by improving the accuracy of the algorithm. Some researchers proposed a method to use convolutional neural networks to correct the actual effective wind speed in front of the wind turbine rotor, which improved the accuracy of the actual effective wind speed in front of the wind turbine rotor; others proposed a hybrid wind speed prediction framework based on an autoencoder network to effectively predict the wind speed in front of the rotor, but the universality of the above methods is not strong.

[0004] The present invention is based on the wind speed data measured by the actual on-site wind measurement experiment and the detection data of the wind turbine SCADA system. By constructing a LENET-5 neural network model to study the transfer characteristics between the wind speed of the measuring point above the cabin and the actual effective incoming wind speed in front of the wheel, the mapping relationship between the wind measurement data above the cabin and the actual effective incoming wind speed in front of the wind rotor is found. The mapped wind speed obtained by this method is used as the actual effective incoming wind speed in front of the wind rotor and then transmitted to the main control system, thereby estimating the convertible electricity more accurately and efficiently, which is of great significance to improving the accuracy of the control system. Summary of the invention

[0005] The object of the present invention is to provide a method for detecting the effective incoming wind speed in front of a wind turbine rotor of a wind turbine, so as to achieve the purpose of more accurately detecting the effective incoming wind speed in front of a wind turbine rotor of the wind turbine.

[0006] In a first aspect, the present invention provides a method for detecting the effective incoming wind speed in front of a wind turbine rotor, comprising:

[0007] S1. Obtain the effective incoming wind speed in front of the wind turbine rotor and the SCADA system detection data above the nacelle at the same time; the detection data includes wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment;

[0008] S2. Preprocessing the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor to obtain preprocessed data; normalizing the preprocessed data and constructing a training set;

[0009] S3. The wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set are used as an input feature, and each time step is used as a sample to construct an input matrix;

[0010] S4. Taking the actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set as an output value, training the mapping relationship between the input feature and the output sequence through the LENET-5 model to obtain the corresponding mapping model;

[0011] S5. Acquire real-time detection data when the incoming wind blows through the wind rotor of the target wind turbine, and input the acquired SCADA real-time data into the mapping model to obtain the actual effective incoming wind speed in front of the wind rotor of the wind turbine.

[0012] In a second aspect, the present invention provides a system for detecting effective incoming wind speed in front of a wind turbine rotor, comprising:

[0013] An acquisition module is used to obtain the effective incoming wind speed in front of the wind rotor in front of the wind turbine and the SCADA system detection data above the nacelle at the same time; the detection data includes wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment;

[0014] A data preprocessing module, used for preprocessing the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor to obtain preprocessed data;

[0015] A data set construction module, used for normalizing the preprocessed data and constructing a training set;

[0016] An input data processing module is used to take the wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set as an input feature, and each time step as a sample to construct an input matrix;

[0017] A mapping model analysis module, used to take the actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set as an output value, train the mapping relationship between the input feature and the output sequence through the LENET-5 model, and obtain a corresponding mapping model;

[0018] The output module obtains the SCADA real-time data and inputs it into the mapping model to obtain the actual effective incoming wind speed in front of the wind turbine rotor.

[0019] The beneficial effects that can be achieved by the present invention are as follows: the effective incoming wind speed detection method in front of the wind rotor of a wind turbine provided by the present invention first obtains the effective incoming wind speed in front of the wind rotor in front of the wind turbine and the detection data of the SCADA system above the cabin at the same time; secondly, the acquired data is preprocessed and a training set is constructed; then, the wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set are used as an input feature, and each time step is used as a sample to construct an input matrix; the actual effective incoming wind speed data in front of the wind rotor of the wind turbine at each moment in the training set are used as an output sequence, and the mapping relationship between the input feature and the output sequence is trained by the LENET-5 model, so that the actual effective incoming wind speed upstream of the wind turbine can be more accurately detected without excessive reliance on the wind measuring radar in front of the wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments of the present invention are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A schematic flow chart of a method for detecting the effective incoming wind speed in front of a wind turbine rotor provided by an embodiment of the present invention;

[0022] Figure 2 A schematic diagram of the topological structure of a system for detecting the effective incoming wind speed in front of a wind turbine rotor provided by an embodiment of the present invention.

[0023] Icon: 10-wind turbine upstream flow wind speed detection system; 100-acquisition module; 200-data preprocessing construction module; 300-input data processing module; 400-mapping model analysis module; 500-output module. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present invention will be described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0025] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0026] Please see Figure 1 , Figure 1 A schematic flow chart of a method for detecting the effective incoming wind speed in front of a wind turbine rotor provided by an embodiment of the present invention.

[0027] In one implementation, an embodiment of the present invention provides a method for detecting the effective incoming wind speed in front of a wind turbine rotor, the specific contents of which are described as follows.

[0028] S1. Obtain the SCADA detection data when the free wind blows through the wind rotor of the target wind turbine and the actual effective wind speed data in front of the wind turbine rotor at the same height in front of the target wind turbine; the detection data includes wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment.

[0029] For example, a laser radar anemometer can be installed above the nacelle of the target wind turbine to obtain detection data when the free wind blows through the rotor of the target wind turbine. A vertical wind tower laser radar is installed in front of the target wind turbine to replace the wind tower to measure the actual effective wind speed data in front of the rotor of the wind turbine at the same height in front of the target wind turbine, and the measurement result is more accurate.

[0030] S2. Preprocessing the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor to obtain preprocessed data.

[0031] In one embodiment, S2 includes: filtering and deleting the missing data at the same time in the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor through the LENET-5 neural network; replacing the wind speed data higher than the rated wind speed of the target wind turbine with the rated wind speed, and eliminating the wind speed data lower than the rated wind speed.

[0032] In this way, the influence of error data on the detection results can be reduced.

[0033] S3. Normalize the preprocessed data and construct a training set, take the wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set as an input feature, and use each time step as a sample to construct an input matrix.

[0034] In one implementation, after the preprocessed data is obtained, the preprocessed data may be normalized first, and then 90% of the data may be divided into a training set, and 10% of the data may be divided into a test set.

[0035] For example, the wind speed, wind direction, wind angle, power, generator speed, rotor speed and other features at each moment in the training set can be used as an input feature x, and each time step can be used as a sample m to construct a 7*m input matrix about the input feature x.

[0036] S4. The actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set is used as an output sequence, and the mapping relationship between the input features and the output sequence is trained by the LENET-5 model to obtain a corresponding mapping model.

[0037] In one embodiment, the number of fully connected layers of the LENET-5 model is set to 1, the loss function adopts the mean absolute error between the predicted value and the actual effective value of the actual effective incoming wind speed data in front of the wind turbine rotor, the optimization method adopts the stochastic gradient descent method, and the parameters of the LENET-5 model are adaptively adjusted through an iterative optimization method during the training process.

[0038] In one embodiment, the above method further comprises:

[0039] After normalizing the preprocessed data, a test set is divided, and the prediction result of the mapping model is verified by the test set. In this way, it can be checked whether the error of the model is too large.

[0040] S5. Acquire real-time detection data when free wind blows through the wind rotor of the target wind turbine, and input the real-time detection data into the mapping model to obtain the actual effective wind speed in front of the wind rotor of the target wind turbine.

[0041] Please see Figure 2 , Figure 2 A schematic diagram of the topological structure of a system for detecting the effective incoming wind speed in front of a wind turbine rotor provided by an embodiment of the present invention.

[0042] In one implementation, an embodiment of the present invention provides a system for detecting effective incoming wind speed in front of a wind turbine rotor, comprising:

[0043] An acquisition module is used to acquire detection data when free wind flows through the wind rotor of a target wind turbine and actual effective wind speed data in front of the wind turbine rotor at the same height in front of the target wind turbine; the detection data includes wind speed, wind direction, angle to wind, power, generator speed, and rotor speed at each moment;

[0044] A data preprocessing module, used for preprocessing the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor to obtain preprocessed data;

[0045] A data set construction module, used for normalizing the preprocessed data and constructing a training set;

[0046] An input data processing module is used to take the wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set as an input feature, and each time step as a sample to construct an input matrix;

[0047] A mapping model analysis module, used to use the actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set as an output sequence, train the mapping relationship between the input features and the output sequence through the LENET-5 model, and obtain a corresponding mapping model;

[0048] The output module obtains real-time detection data when the free flow wind blows through the wind rotor of the target wind turbine, and inputs the real-time detection data into the mapping model to obtain the actual effective flow wind speed in front of the wind rotor of the target wind turbine.

[0049] The above-mentioned effective incoming wind speed detection system in front of the wind turbine rotor can more accurately detect the actual incoming wind speed upstream of the wind turbine without overly relying on the vertical wind tower type wind measuring radar in front of the wind turbine.

[0050] In summary, an embodiment of the present invention provides a method for obtaining the effective incoming wind speed in front of a wind turbine rotor. The method first obtains the detection data when the free incoming wind blows through the target wind turbine rotor and the actual effective incoming wind speed data in front of the wind turbine rotor at the same height in front of the target wind turbine; secondly, the acquired data is preprocessed and a training set is constructed; then, the wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set are used as an input feature, and each time step is used as a sample to construct an input matrix; the actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set is used as an output value, and the mapping relationship between the input feature and the output sequence is trained by the LENET-5 model, so that the actual incoming wind speed upstream of the wind turbine can be more accurately detected without excessive reliance on the vertical wind tower type wind measuring radar in front of the wind turbine.

[0051] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for detecting the effective incoming wind speed in front of a wind turbine rotor, characterized in that: include: S1. Obtain the effective incoming wind speed in front of the wind turbine rotor and the SCADA system detection data above the nacelle at the same time; the detection data includes wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment; S2. Preprocessing the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor to obtain preprocessed data; Normalizing the preprocessed data and constructing a training set; S3. The wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set are used as an input feature, and each time step is used as a sample to construct an input matrix; S4. The actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set is used as an output sequence, and the mapping relationship between the input features and the output sequence is trained by the LENET-5 model to obtain a corresponding mapping model; the number of fully connected layers of the LENET-5 model is set to 1, the loss function adopts the mean absolute error between the predicted value and the actual effective value of the actual effective incoming wind speed data in front of the wind turbine rotor, the optimization method adopts the stochastic gradient descent method, and the parameters of the LENET-5 model are adaptively adjusted by the iterative optimization method during the training process; S5. Acquire real-time SCADA detection data when free wind blows through the wind rotor of the target wind turbine, and use the acquired real-time SCADA detection data to input into the mapping model to obtain the actual effective wind speed in front of the wind rotor of the wind turbine.

2. The method according to claim 1, characterized in that The method further comprises: The preprocessed data is normalized and then divided into a test set, and the prediction result of the mapping model is verified by the test set.

3. A wind turbine upstream wind speed detection system, characterized in that: include: An acquisition module is used to acquire detection data when free wind flows through the wind rotor of a target wind turbine and actual effective wind speed data in front of the wind turbine rotor at the same height in front of the target wind turbine; the detection data includes wind speed, wind direction, angle to wind, power, generator speed, and rotor speed at each moment; A data preprocessing module, used for preprocessing the detection data and the actual effective incoming wind speed data in front of the wind turbine rotor to obtain preprocessed data; A data set construction module, used for normalizing the preprocessed data and constructing a training set; An input data processing module is used to take the wind speed, wind direction, wind angle, power, generator speed, and rotor speed at each moment in the training set as an input feature, and each time step as a sample to construct an input matrix; A mapping model analysis module is used to use the actual effective incoming wind speed data in front of the wind turbine rotor at each moment in the training set as an output sequence, and train the mapping relationship between the input features and the output sequence through the LENET-5 model to obtain a corresponding mapping model; the number of fully connected layers of the LENET-5 model is set to 1, the loss function adopts the mean absolute error between the predicted value and the actual effective value of the actual effective incoming wind speed data in front of the wind turbine rotor, the optimization method adopts the stochastic gradient descent method, and the parameters of the LENET-5 model are adaptively adjusted through an iterative optimization method during the training process; The output module obtains real-time detection data when the free flow wind blows through the wind rotor of the target wind turbine, and inputs the real-time detection data into the mapping model to obtain the actual effective flow wind speed in front of the wind rotor of the target wind turbine.

Images