Fan variable pitch control method and system

By predicting the wind power distribution and using the 1D-CNN network model for fault diagnosis, and adjusting the fan blade angle in advance, the problem of pitch reaction lag in the existing technology is solved, and the power generation efficiency and power of the wind turbine are improved.

CN118934445APending Publication Date: 2024-11-12HUANENG NEW ENERGY CO LTD SHANXI BRANCH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410984105.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing wind turbine pitch technology lags when wind power changes, resulting in wasting good power generation time in large areas of wind farms and failure to adjust the blade angle in time to improve power generation efficiency.

Method used

By obtaining fan operating parameters, establishing a spatial coordinate system, predicting wind power distribution, using the 1D-CNN network model for fault diagnosis and wind power prediction, adjusting the fan blade angle in advance, realizing pre-rat operation, and correcting the pitch state according to actual wind power .

Benefits of technology

It improves the power generation efficiency of wind turbines under different wind conditions, reduces the waste of pitching time, ensures that the fan is in the best state when the wind power reaches, and improves the power generation capacity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118934445A_ABST
    Figure CN118934445A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of wind power generation, in particular to a fan variable pitch control method and system.The fan variable pitch control method mainly comprises the steps that operation parameters of a fan are obtained, whether the current fan operates normally or not is judged, and if the current operation of the fan fails, operation of the fan is stopped; and obtaining a first variable pitch parameter of the fan which receives the wind speed and the wind direction first, and sending a second variable pitch parameter to the fan adjacent to the fan which receives the wind power first according to the horizontal coordinate parameter. By means of the method, parameters can be collected through the first generator receiving wind power, the parameters are processed through the position relation between the generators and then sent to all the draught fans in the whole area, the draught fans conduct pitch changing in advance before wind arrives, the pitch changing time is saved, and the efficiency of pitch changing is improved. And the fan is in the optimal variable pitch state as much as possible, so that the power generation quantity is further increased.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation, and in particular to a wind turbine pitch control method and system. Background Art

[0002] With the rapid development of my country's economy, energy consumption has increased rapidly, and the contradiction between energy supply and demand has become increasingly prominent. Establishing a clean, low-carbon, safe and reliable energy system is an inevitable choice for achieving sustainable development. While wind power installed capacity continues to rise, the capacity of wind turbines is also increasing. In order to capture high-quality wind energy at high altitudes, reduce the cost of wind power generation, and improve competitiveness, wind turbines tend to develop in the direction of large-scale, lightweight, low noise, and low failure rate.

[0003] Wind turbine pitch control technology is a technology used to adjust the angle of wind turbine blades. Wind turbines capture wind energy and convert it into mechanical energy by rotating blades, and then convert it into electrical energy. Pitch control technology can adjust the angle of blades according to changes in wind speed and direction, so that wind turbines can efficiently convert wind energy into electrical energy under different wind speed and wind direction conditions. This can improve the power generation efficiency of wind turbines and enable them to operate stably under different wind conditions.

[0004] However, the current pitch change technology generally starts to react and change the pitch when it receives wind force. Pitch change requires a process and time. Pitch change will not start until the wind arrives. In the construction of large-scale wind farms, the wind turbines are far apart. Pitch change will not start until the wind arrives, which will waste a lot of good power generation time and lose the opportunity to further increase the power generation energy. Summary of the invention

[0005] The purpose of the present invention is to provide a wind turbine pitch control method and system to solve the problem of untimely wind turbine pitch change in the prior art.

[0006] The embodiments of the present invention are implemented by the following technical solutions:

[0007] In a first aspect, the present invention provides a wind turbine pitch control method, comprising:

[0008] Obtain the operating parameters of the fan to determine whether the fan is currently operating normally. If the fan is currently operating at fault, stop the fan;

[0009] A number of wind turbines are operating normally, a spatial coordinate system is established, positioning coordinate parameters of all wind turbines are obtained, and horizontal coordinate parameters in the horizontal direction of the spatial coordinate system are selected;

[0010] Obtain a first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and send a second pitch parameter to a wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter;

[0011] The wind turbine receives the second pitch parameter, calculates and predicts the wind turbine absorption power, and performs a pre-pitch operation;

[0012] Set the power deviation threshold, and calculate the actual wind turbine absorbed power after the remaining wind turbines except the first one that receives wind power receive wind power;

[0013] If the difference between the judged wind turbine absorbed power and the actual wind turbine absorbed power is within the power deviation threshold, no change is made; if it is not within the power deviation threshold, the pitch correction is performed again according to the actual wind turbine absorbed power.

[0014] In one embodiment of the present invention, the step of determining whether the fan is currently operating normally includes:

[0015] Collect data fault samples of various pitch sensors, establish source domain data sets and target domain data sets, and perform data preprocessing and overlapping sampling;

[0016] Input the 1D-CNN network model and initialize the weights of the convolutional layer;

[0017] Obtain high-level features in the GAP network layer and predict the output of the training data set;

[0018] Calculate the errors of the cross entropy loss function and the maximum average difference loss function respectively, and construct the total loss Loss error of the two;

[0019] The model is trained through BP neural network, and it is determined whether the maximum number of iterations has been reached. If not, the training continues;

[0020] If so, a training model is obtained, data of various sensors of the current variable pitch are collected as target domain data, the target domain data is input into the training model, and a fault judgment result is obtained.

[0021] In one embodiment of the present invention, the 1D-CNN network model includes a convolutional layer, a pooling layer, an attention mechanism layer, a global average pooling layer and a softmax classifier;

[0022] The convolution layer is used to extract features through multiple networks and filters;

[0023] The pooling layer is used to reduce the dimension of the data, represent it with high-level features, and extract features;

[0024] The attention mechanism layer is used to adaptively weight different signal features, filter information, and suppress invalid features;

[0025] The global average pooling layer is used to integrate the local information with category distinction in the convolution layer or the pooling layer;

[0026] The softmax classifier is used for fault classification.

[0027] In one embodiment of the present invention, the global average pooling layer comprises:

[0028]

[0029] In the formula, That is, the value obtained by the global average aggregation layer of the I-th layer network, f1 l is the eigenvalue, and c is the number of channels.

[0030] In one embodiment of the present invention, the cross entropy loss function includes:

[0031]

[0032] Where n s is the number of training samples in the source domain, y i is the label of the source domain sample, N c is the number of samples;

[0033] The maximum average difference loss function includes:

[0034]

[0035] In the formula, and are two data sets, and the posterior probabilities of the two data sets are not equal, φ is a linear mapping function, H is the reproducing kernel space, n t The number of training samples for the target domain.

[0036] In one embodiment of the present invention, the attention mechanism layer includes:

[0037] Get the feature sequence output by each channel of the convolutional network;

[0038] Transfer the feature sequence to the fully connected network layer to obtain the attention weight of the channel;

[0039] Multiply the attention weight by the original feature sequence to obtain the feature sequence after filtering by the attention mechanism.

[0040] In one embodiment of the present invention, obtaining a first pitch parameter of a first wind turbine receiving the wind speed and wind direction comprises:

[0041] The wind speed is obtained by the following formula;

[0042]

[0043] Where P is the power absorbed by the wind turbine; ρ is the air density; R is the radius of the rotor blade; v is the current wind speed; Cp is the wind energy utilization coefficient; λ is the tip speed ratio; β is the blade pitch angle.

[0044] In one embodiment of the present invention, the sending of the second pitch parameter to a wind turbine adjacent to the first wind turbine receiving the wind force according to the horizontal coordinate parameter comprises:

[0045] Calculating a second pitch parameter by the wind speed of the first pitch parameter;

[0046]

[0047] In the formula, v1 is the wind speed of the first wind turbine that receives the wind force, x1 and y1 are the horizontal and vertical coordinate values ​​of the first wind turbine that receives the wind force, x2 and y2 are the horizontal and vertical coordinate values ​​of the wind turbine that receives the second pitch parameter, and v2 is the calculated wind speed of the wind turbine that receives the second pitch parameter.

[0048] In a second aspect, the present invention further provides a wind turbine pitch control system, comprising:

[0049] The fault diagnosis module is configured to obtain the operating parameters of the fan, determine whether the fan is currently operating normally, and stop the operation of the fan if the fan is currently operating faulty;

[0050] The calculation module is configured to establish a spatial coordinate system, obtain the positioning coordinate parameters of all wind turbines, and select the horizontal coordinate parameters in the horizontal direction of the spatial coordinate system when several wind turbines are operating normally; obtain the first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and send the second pitch parameter to the wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter; calculate and predict the wind turbine absorption power for the wind turbine that receives the second pitch parameter, and perform the pre-pitch operation;

[0051] The judgment module sets a power deviation threshold, and calculates the actual wind turbine absorbed power after the wind turbines other than the first one that receives wind power receive wind power; if the difference between the judged wind turbine absorbed power and the actual wind turbine absorbed power is within the power deviation threshold, no change is made; if it is not within the power deviation threshold, the pitch correction is performed again according to the actual wind turbine absorbed power.

[0052] In a third aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned wind turbine pitch control method is implemented.

[0053] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects:

[0054] The above method of the present invention mainly includes obtaining the operating parameters of the wind turbine, judging whether the current operation of the wind turbine is normal, and stopping the operation of the wind turbine if the current operation of the wind turbine fails; obtaining the first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and sending the second pitch parameter to the wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter. Through this method, the first generator that receives the wind force can be used to collect the parameters, and after the positional relationship between the generators, the parameters are processed and sent to all wind turbines in the entire area, so that the wind turbine can pre-pitch before the wind arrives to save the time of pitch change. If the sent parameters are correct, no further adjustment is required. If the sent parameters have errors, only a small amount of rapid adjustment is required. In this way, when the wind arrives, the wind turbine is in the best pitch state as much as possible to further increase the power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments 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.

[0056] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0058] The modules or submodules described independently may be physically separated or not: they may be implemented by software or hardware, and some modules or submodules may be implemented by software, and the processor may call the software to implement the functions of these modules or submodules, and other modules or submodules may be implemented by hardware, such as by hardware circuits. In addition, some or all of the modules may be selected according to actual needs to achieve the purpose of the present application.

[0059] Please refer to Figure 1 As shown, the present invention improves a wind turbine pitch control method, including:

[0060] S101: Obtaining the operating parameters of the fan, determining whether the fan is currently operating normally, and stopping the operation of the fan if the fan is currently operating faulty;

[0061] When performing pitch change, the present invention needs to verify in advance whether the operation of the wind turbine is normal, and the specific judgment process is described in detail below.

[0062] S102: A number of wind turbines are operating normally, a spatial coordinate system is established, positioning coordinate parameters of all wind turbines are obtained, and horizontal coordinate parameters in the horizontal direction of the spatial coordinate system are selected;

[0063] The spatial coordinate system can be obtained based on the global positioning of the wind turbine to obtain a three-dimensional coordinate axis data, but the height difference between each wind turbine has a weaker effect than the distance difference in the horizontal direction, so this embodiment only considers the distance on the horizontal plane, and does not consider the distance in height, to simplify the calculation steps and process.

[0064] S103: Acquire a first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and send a second pitch parameter to a wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter;

[0065] The first pitch parameter includes the basic parameters of the wind turbine, including the wind turbine absorbed power, air density, wind rotor blade radius, wind speed and other necessary parameters. The second pitch parameter sent is mainly the blade pitch angle. The blade pitch angle is the final rotation angle required for the pitch change, which is calculated by the formula.

[0066] S104: receiving the wind turbine of the second pitch parameter, calculating and predicting the wind turbine absorption power, and performing a pre-pitch operation;

[0067] S105: Setting a power deviation threshold, and calculating actual wind turbine absorbed power after the remaining wind turbines except the first wind turbine that receives wind power receive wind power;

[0068] S106: If the difference between the wind turbine absorbed power and the actual wind turbine absorbed power is within the power deviation threshold, no change is made; if it is not within the power deviation threshold, pitch correction is performed again according to the actual wind turbine absorbed power.

[0069] The above method of the present invention mainly includes obtaining the operating parameters of the wind turbine, judging whether the current operation of the wind turbine is normal, and stopping the operation of the wind turbine if the current operation of the wind turbine fails; obtaining the first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and sending the second pitch parameter to the wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter. Through this method, the first generator that receives the wind force can be used to collect the parameters, and after the positional relationship between the generators, the parameters are processed and sent to all wind turbines in the entire area, so that the wind turbine can pre-pitch before the wind arrives to save the time of pitch change. If the sent parameters are correct, no further adjustment is required. If the sent parameters have errors, only a small amount of rapid adjustment is required. In this way, when the wind arrives, the wind turbine is in the best pitch state as much as possible to further increase the power generation.

[0070] In one embodiment of the present invention, the step of determining whether the fan is currently operating normally includes:

[0071] Collect data fault samples of various pitch sensors, establish source domain data sets and target domain data sets, and perform data preprocessing and overlapping sampling; input the 1D-CNN network model to initialize the weights of the convolution layer; obtain high-level features in the GAP network layer and predict the output of the training data set; calculate the errors of the cross entropy loss function and the maximum average difference loss function respectively, and construct the total loss Loss error of the two; train the model through the BP neural network and determine whether the maximum number of iterations has been reached. If not, repeat the above steps; if so, obtain the training model, collect the data of various current pitch sensors as the target domain data, input the target domain data into the training model, and obtain the fault judgment result.

[0072] Specifically, in this embodiment, the 1D-CNN network model includes a convolution layer, a pooling layer, an attention mechanism layer, a global average pooling layer and a softmax classifier;

[0073] Convolutional layer, used to extract features through multiple networks and filters;

[0074] x l =σ(w l x l-1 +b l )

[0075] In the formula, x l is the vector of the output layer I, σ is the activation function, x l-1 is the input vector to layer I, w i is the matrix composed of the weights of all connection units between layer I-1 and layer I, b l is bias.

[0076] The pooling layer is used to reduce the information redundancy of features, reduce the dimension of data, represent it with high-level features, and then extract features. Usually, the maximum pooling operation is used, the window size and step size are set, and the local maximum value obtained is used as the retained feature;

[0077] p l(i,i) =max{a l(i,j)} (j-1)w+1≤t≤jw

[0078] In the formula, p l(i,j) is the correlation value after the pooling operation of the I layer neurons, in a (i,j) is the value of the tth neuron in the ith structure of layer I, (j-1)w+1≤t≤jw is the threshold of t, and w is the pooling width.

[0079] The attention mechanism layer is used to adaptively weight different signal features, filter information, and suppress invalid features;

[0080] Specifically, the attention mechanism layer includes: obtaining the feature sequence V = [v1, v2, v3..., vN] output by each channel of the convolutional network; N is the length of the sequence, transmitting the feature sequence V to the fully connected network layer, and obtaining the channel attention weight S = [s1, s2, s3..., s N ]; Multiply the attention weight S by the original feature sequence to obtain the feature sequence after the attention mechanism is used for filtering.

[0081] The global average pooling layer is used to integrate the local information with category discrimination in the convolutional layer or pooling layer;

[0082]

[0083] In the formula, That is, the value obtained by the global average aggregation layer of the I-th layer network, f1 l is the eigenvalue, and c is the number of channels.

[0084] In this embodiment, Softmax is a classifier for multi-category classification tasks. In this article, faults are classified according to the size of the Softmax value.

[0085] In one embodiment of the present invention, the minimum empirical classification error of source label data is achieved through a cross entropy loss function, and the cross entropy loss function includes:

[0086]

[0087] Where n s is the number of training samples in the source domain, y i is the label of the source domain sample, N c is the number of samples;

[0088] The maximum average difference loss function is used to measure the difference between the source domain and the target domain. The closer the value is to 0, the higher the matching degree of the two data distributions is, including;

[0089]

[0090] In the formula, and are two data sets, and the posterior probabilities of the two data sets are not equal, φ is a linear mapping function, H is the reproducing kernel space, n t The number of training samples for the target domain.

[0091] The following is a practical example to explain the above neural network module:

[0092] A deep learning framework based on CUDA11.3 and pytorch1.10.1 is used to implement fault diagnosis. The model uses the SGD optimizer that minimizes the cost function to update the training parameters. The learning efficiency parameter is 0.01, the number of iteration batches is 50, and the size of each batch is 30.

[0093] Firstly, the method provided in this scheme is used for multi-sensor fault diagnosis of variable pitch system under different switching modes and compared with traditional signal processing methods SVM+EMD+Hibert envelope spectrum, BPNN+EMD+Hibert, envelope spectrum, ResNet and traditional methods to verify its fault diagnosis performance in different modes without noise influence. The scheme reached 99% under the switching of constant gain-constant deviation and constant deviation-constant gain modes, which is greatly improved in accuracy compared with traditional diagnosis methods. However, traditional signal processing methods have low diagnostic accuracy in different modes, that is, fault diagnosis under different mode switching cannot be realized. The fault diagnosis accuracy is higher than that of the method provided in this article, and it has better diagnostic effect.

[0094] In one embodiment of the present invention, obtaining a first pitch parameter of a first wind turbine receiving the wind speed and wind direction comprises:

[0095] The wind speed is obtained by the following formula;

[0096]

[0097] Where P is the power absorbed by the wind turbine; ρ is the air density; R is the radius of the rotor blade; v is the current wind speed; Cp is the wind energy utilization coefficient; λ is the tip speed ratio; β is the blade pitch angle.

[0098] In the above formula, the current wind speed can be directly calculated. Secondly, the second pitch parameter is sent to the wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter, including:

[0099] Calculating a second pitch parameter by the wind speed of the first pitch parameter;

[0100]

[0101] In the formula, v1 is the wind speed of the first wind turbine that receives the wind force, x1 and y1 are the horizontal and vertical coordinate values ​​of the first wind turbine that receives the wind force, x2 and y2 are the horizontal and vertical coordinate values ​​of the wind turbine that receives the second pitch parameter, and v2 is the calculated wind speed of the wind turbine that receives the second pitch parameter.

[0102] The wind force component of the wind turbine that accepts the second pitch parameter can be calculated based on the positional relationship between the wind turbines, so as to approximately calculate the second pitch parameter of the wind turbine that accepts the second pitch parameter.

[0103] After actually receiving the wind force, it is determined whether the generated power is within the threshold and whether to adjust the current pitch state. At this time, after the pre-pitch change, even if the pitch needs to be changed again later, only a small fine-tuning is required.

[0104] In a second aspect, the present invention further provides a wind turbine pitch control system, comprising:

[0105] The fault diagnosis module is configured to obtain the operating parameters of the fan, determine whether the fan is currently operating normally, and stop the operation of the fan if the fan is currently operating faulty;

[0106] The calculation module is configured to establish a spatial coordinate system, obtain the positioning coordinate parameters of all wind turbines, and select the horizontal coordinate parameters in the horizontal direction of the spatial coordinate system when several wind turbines are operating normally; obtain the first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and send the second pitch parameter to the wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter; calculate and predict the wind turbine absorption power for the wind turbine that receives the second pitch parameter, and perform the pre-pitch operation;

[0107] The judgment module sets a power deviation threshold, and calculates the actual wind turbine absorbed power after the wind turbines other than the first one that receives wind power receive wind power; if the difference between the judged wind turbine absorbed power and the actual wind turbine absorbed power is within the power deviation threshold, no change is made; if it is not within the power deviation threshold, the pitch correction is performed again according to the actual wind turbine absorbed power.

[0108] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0109] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A wind turbine pitch control method, characterized in that: include; Obtain the operating parameters of the fan to determine whether the fan is currently operating normally. If the fan is currently operating at fault, stop the fan; A number of wind turbines are operating normally, a spatial coordinate system is established, positioning coordinate parameters of all wind turbines are obtained, and horizontal coordinate parameters in the horizontal direction of the spatial coordinate system are selected; Obtain a first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and send a second pitch parameter to a wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter; The wind turbine receives the second pitch parameter, calculates and predicts the wind turbine absorption power, and performs a pre-pitch operation; Set the power deviation threshold, and calculate the actual wind turbine absorbed power after the remaining wind turbines except the first one that receives wind power receive wind power; If the difference between the judged wind turbine absorbed power and the actual wind turbine absorbed power is within the power deviation threshold, no change is made; if it is not within the power deviation threshold, the pitch correction is performed again according to the actual wind turbine absorbed power.

2. A wind turbine pitch control method according to claim 1, characterized in that: The step of judging whether the fan is currently operating normally includes: S collects data fault samples of various pitch sensors, establishes source domain data sets and target domain data sets, and performs data preprocessing and overlapping sampling; Input the 1D-CNN network model and initialize the weights of the convolutional layer; Obtain high-level features in the GAP network layer and predict the output of the training data set; Calculate the errors of the cross entropy loss function and the maximum average difference loss function respectively, and construct the total loss Loss error of the two; The model is trained through BP neural network, and it is determined whether the maximum number of iterations has been reached. If not, the training continues; If so, a training model is obtained, data of various sensors of the current variable pitch are collected as target domain data, the target domain data is input into the training model, and a fault judgment result is obtained.

3. A wind turbine pitch control method according to claim 2, characterized in that: The 1D-CNN network model includes a convolution layer, a pooling layer, an attention mechanism layer, a global average pooling layer and a softmax classifier; The convolution layer is used to extract features through multiple networks and filters; The pooling layer is used to reduce the dimension of the data, represent it with high-level features, and extract features; The attention mechanism layer is used to adaptively weight different signal features, filter information, and suppress invalid features; The global average pooling layer is used to integrate the local information with category distinction in the convolution layer or the pooling layer; The softmax classifier is used for fault classification.

4. A wind turbine pitch control method according to claim 3, characterized in that: The global average pooling layer comprises: In the formula, That is, the value obtained by the global average aggregation layer of the l-th layer network, f1 l is the eigenvalue, and c is the number of channels.

5. A wind turbine pitch control method according to claim 3, characterized in that: The cross entropy loss function includes: Where n s is the number of training samples in the source domain, y i is the label of the source domain sample, N c is the number of samples; The maximum average difference loss function includes: In the formula, and are two data sets, and the posterior probabilities of the two data sets are not equal, φ is a linear mapping function, H is the reproducing kernel space, n t The number of training samples for the target domain.

6. A wind turbine pitch control method according to claim 3, characterized in that: The attention mechanism layer includes: Get the feature sequence output by each channel of the convolutional network; Transfer the feature sequence to the fully connected network layer to obtain the attention weight of the channel; Multiply the attention weight by the original feature sequence to obtain the feature sequence after filtering by the attention mechanism.

7. A wind turbine pitch control method according to claim 1, characterized in that: The obtaining of a first pitch parameter of a first wind turbine receiving the wind speed and wind direction comprises: The wind speed is obtained by the following formula; Where P is the power absorbed by the wind turbine; ρ is the air density; R is the radius of the rotor blade; v is the current wind speed; Cp is the wind energy utilization coefficient; λ is the tip speed ratio; β is the blade pitch angle.

8. A wind turbine pitch control method according to claim 7, characterized in that: The sending of the second pitch parameter to the wind turbine adjacent to the first wind turbine receiving the wind force according to the horizontal coordinate parameter comprises: Calculating a second pitch parameter by the wind speed of the first pitch parameter; In the formula, v1 is the wind speed of the first wind turbine that receives the wind force, x1 and y1 are the horizontal and vertical coordinate values ​​of the first wind turbine that receives the wind force, x2 and y2 are the horizontal and vertical coordinate values ​​of the wind turbine that receives the second pitch parameter, and v2 is the calculated wind speed of the wind turbine that receives the second pitch parameter.

9. A wind turbine pitch control system, characterized in that: include; The fault diagnosis module is configured to obtain the operating parameters of the fan, determine whether the fan is currently operating normally, and stop the operation of the fan if the fan is currently operating faulty; The calculation module is configured to establish a spatial coordinate system, obtain the positioning coordinate parameters of all wind turbines, and select the horizontal coordinate parameters in the horizontal direction of the spatial coordinate system when several wind turbines are operating normally; obtain the first pitch parameter of the wind turbine that first receives the wind speed and wind direction, and send the second pitch parameter to the wind turbine adjacent to the first wind turbine that receives the wind force according to the horizontal coordinate parameter; calculate and predict the wind turbine absorption power for the wind turbine that receives the second pitch parameter, and perform the pre-pitch operation; The judgment module sets a power deviation threshold, and calculates the actual wind turbine absorbed power after the wind turbines other than the first one that receives wind power receive wind power; if the difference between the judged wind turbine absorbed power and the actual wind turbine absorbed power is within the power deviation threshold, no change is made; if it is not within the power deviation threshold, the pitch correction is performed again according to the actual wind turbine absorbed power.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, a wind turbine pitch control method according to any one of claims 1 to 8 is implemented.