Wind driven generator control method, device and system using artificial neural network

Abstract

The invention discloses a wind driven generator control method, device and system using an artificial neural network, relates to the technical field of wind power generation, and solves the problems of slow response speed, complex parameter adjustment, poor control effect and poor adaptability of a PI control algorithm in the prior art to wind speed change. The method comprises the steps that a wind driven generator is simulated, data are collected, a neural network based on a Transform model is trained through the collected data to obtain a neural network model, and the neural network model is used for predicting the state of the wind driven generator at the next time point and generating the control quantity at the current time point. The trained neural network model is subjected to optimization application and actual control over the wind driven generator, the operation state of the wind driven generator is predicted according to the environment data, the control quantity and the output quantity of operation of the wind driven generator, and the optimal control quantity is generated; and the actual operation state of the wind driven generator can reach the ideal operation state under the current condition as much as possible.

Description

Technical Field

[0001] The present application relates to the field of wind power generation technology, and in particular to a method, device and system for controlling a wind turbine generator using an artificial neural network. Background Art

[0002] Thanks to technological progress, cost reduction and policy support, wind power generation has developed rapidly around the world. Technical innovations related to wind power generation are also emerging, including efficient blade design, the application of lightweight and high-strength composite materials, and intelligent control system upgrades.

[0003] There are several sensors installed on the wind turbine tower as input parameters of the wind turbine control system. They mainly include: wind force; wind direction; hub drive shaft speed; power generation; blade tip acceleration, blade tip displacement, etc. The quantities that can be controlled by the wind turbine control system include: 1. Yaw control: adjust the yaw angle of the wind turbine; 2. Pitch control: adjust the blade pitch angle; 3. Torque control: adjust the hub torque.

[0004] The control objectives of wind turbines are: 1. Avoid safety accidents, such as blades hitting the tower, blades rotating too fast and causing damage to the blades, etc.; 2. Reduce the force and vibration of the blades during operation and extend the service life of the blades; 3. Increase power generation.

[0005] The typical feedback control algorithm is PI control (proportional integral control). The PI control algorithm has been widely used in wind power generation control, but it has some technical disadvantages, including:

[0006] 1. Slow response to wind speed changes: Since the PI control algorithm relies on the accumulation of errors for adjustment, its response speed is slow to rapidly changing wind speeds, which may cause the system to be unable to adjust in time, affecting power generation efficiency;

[0007] 2. Complex parameter adjustment: The performance of the PI controller is highly dependent on the settings of the proportional gain Kp and the integral gain Ki. Under different working conditions and load conditions, Kp and Ki need to be repeatedly debugged to obtain the best control effect, which increases the complexity of system debugging;

[0008] 3. Prone to overshoot and oscillation: If the proportional gain Kp and integral gain Ki are not set properly, the system may overshoot and oscillate, resulting in unstable output power of the wind turbine and affecting the reliability of the system;

[0009] 4. Integral saturation problem: When there is a large error for a long time, the integral term may accumulate too much, resulting in integral saturation, the controller output is out of control, and the system cannot be effectively adjusted;

[0010] 5. Poor adaptability: Traditional PI control algorithms are difficult to provide ideal control effects when facing complex nonlinear systems or multivariable control scenarios, and their adaptability is poor. This is particularly evident when dealing with changes in wind speed and direction. Summary of the invention

[0011] The purpose of this application is to overcome the problems of slow response speed of PI control algorithm to wind speed changes, complex parameter adjustment, poor control effect and poor adaptability in the prior art, and to provide a wind turbine control method, device and system using artificial neural network.

[0012] In a first aspect, a wind turbine control method using an artificial neural network is provided, comprising:

[0013] Build a wind turbine simulation model in the simulation environment and run the wind turbine;

[0014] Collecting historical operating parameters of the wind turbine, wherein the historical operating parameters include environmental data, control quantity and output quantity;

[0015] Using the historical operating parameters as a training data set to train a neural network based on a Transformer model to obtain a neural network model, wherein the neural network model is used to predict the state of the wind turbine generator at the next time point and generate the control amount at the current time point;

[0016] Using a stochastic gradient descent algorithm to tune the neural network model in a simulation environment;

[0017] The wind turbine generator is controlled using the tuned neural network model.

[0018] In some possible implementations, running the wind turbine simulation model in the simulation environment includes: testing the operating state of the wind turbine under different wind force and wind direction conditions in the simulation environment.

[0019] In some possible implementations, collecting historical operating parameters of the wind turbine includes: collecting historical operating parameters of the wind turbine at n consecutive time points in a simulation environment, where n is a positive integer greater than 1, and the historical operating parameters include environmental data, control quantities, and output quantities, the environmental data including wind force, wind direction, and air temperature; the control quantities including yaw angle, pitch angle, and torque; and the output quantities including blade speed, power generation, and blade bending moment.

[0020] In some possible implementations, the neural network model includes a neural network prediction model and a neural network control model, wherein the neural network prediction model is used to predict the state of the wind turbine at the next time point, and the neural network control model is used to generate a control quantity at the current time point, and the control quantity includes a pitch angle, a yaw angle, and a torque.

[0021] In some possible implementations, the neural network model is tuned using a stochastic gradient descent algorithm in a simulation environment, including:

[0022] S401, setting an ideal wind turbine generator state under different wind conditions;

[0023] S402, at a first moment, using a neural network prediction model to predict a state of a wind turbine generator at a next moment, and using a neural network control model to generate a control quantity at the first moment;

[0024] S403, in a simulation environment, inputting the generated control quantity at the first moment into a wind turbine simulation model and simulating the operation of the wind turbine;

[0025] S404, collecting the actual wind turbine state of the wind turbine simulation model at the next moment, and modifying the parameters of the neural network control model according to the gap between the actual wind turbine state at the next moment and the ideal wind turbine state;

[0026] S405, repeating steps S302 to S304 until the gap between the actual state of the wind turbine generator and the ideal state of the wind turbine generator at the next moment meets the user's needs.

[0027] Some possible implementations also include:

[0028] Using the tuned neural network model to control the first wind turbine, and selecting the second wind turbine to be controlled by the native control system, wherein the first wind turbine and the second wind turbine are located in the same area;

[0029] operating the first wind turbine and the second wind turbine for a period of time and collecting actual operating data;

[0030] The actual operation data is used to continuously tune the neural network model.

[0031] In a second aspect, a wind turbine control device using an artificial neural network is provided, comprising:

[0032] A simulation module, used to build a wind turbine simulation model and run the wind turbine in a simulation environment;

[0033] A collection module, used for collecting historical operating parameters of the wind turbine, wherein the historical operating parameters include environmental data, control quantity and output quantity;

[0034] A training module, used for training a neural network based on a Transformer model using the historical operating parameters as a training data set to obtain a neural network model, wherein the neural network model is used to predict the state of the wind turbine generator at the next time point and generate a control amount at the current time point;

[0035] A tuning module, used for tuning the neural network model using a stochastic gradient descent algorithm in a simulation environment;

[0036] The application module is used to control the wind turbine using the tuned neural network model.

[0037] In a third aspect, a wind turbine control system using an artificial neural network is provided, using a method as described in any one of the implementations of the first aspect above, characterized in that the wind turbine control system comprises:

[0038] A neural network control module is used to obtain an optimal control amount using a stochastic gradient descent algorithm so that the state of the wind turbine generator at the next time point predicted by the neural network control module is closest to the ideal wind turbine generator state;

[0039] A neural network prediction module is used to predict the state of the wind turbine at the next time point based on the state of the wind turbine at the current time point and the control amount;

[0040] The feedback optimization module is used to optimize the parameters of the neural network control module according to the difference between the state of the wind turbine generator at the next time point predicted by the neural network control module and the actual state of the wind turbine generator at the next time point.

[0041] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable medium stores a program code for execution by a device, wherein the program code includes steps for executing the method in any one of the implementations of the first aspect described above.

[0042] In a fifth aspect, an electronic device is provided, comprising 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 a method as in any one of the implementations in the first aspect above.

[0043] The present application has the following beneficial effects: the present application realizes real-time control of wind turbines in a data-driven manner, predicts the operating state of the wind turbine according to the environmental data, control quantity and output quantity of the wind turbine operation, and generates the optimal control quantity, so that the actual operating state of the wind turbine can reach the ideal operating state under the current conditions as much as possible, thereby improving the control performance of the wind turbine, increasing power generation, reducing the stress on the wind turbine blades and extending the service life. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application.

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 is a flow chart of a wind turbine control method using an artificial neural network according to Embodiment 1 of the present application;

[0047] Figure 2 is a structural block diagram of a wind turbine control device using an artificial neural network according to Embodiment 2 of the present application;

[0048] Figure 3 is a framework diagram of a wind turbine control system using an artificial neural network according to Example 3 of the present application;

[0049] Figure 4 It is a schematic diagram of the internal structure of the electronic device of Example 5 of the present application.

[0050] Reference numerals:

[0051] 100, simulation module; 200, acquisition module; 300, training module; 400, tuning module; 500, application module; 600, neural network control module; 700, neural network prediction module; 800, feedback optimization module; 900, controlled system. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] Example 1

[0054] like Figure 1 As shown, a wind turbine control method using an artificial neural network according to Embodiment 1 of the present application includes:

[0055] S100, constructing a wind turbine simulation model in a simulation environment and operating the wind turbine;

[0056] First, the model predictive control system is developed in a simulation environment. In this embodiment, the simulation environment uses Blade wind turbine simulation software, and the Blade simulation model is used instead of the actual wind turbine.

[0057] S200, collecting historical operating parameters of the wind turbine, wherein the historical operating parameters include environmental data, control quantity and output quantity;

[0058] Run the wind turbine in a simulation environment, test the operating status of the wind turbine under different wind forces and wind directions, and collect historical operating parameters of the wind turbine at n consecutive time points in the simulation environment, where n can be 9, 12, 36, etc., and the intervals between time points can be customized according to needs, for example: the intervals between time points are 0.5 seconds, 1 second, 3 seconds, 6 seconds, etc., and the historical operating parameters include environmental data, control quantities and output quantities. The environmental data include wind force, wind direction and temperature; the control quantities include yaw angle, pitch angle and torque; the output quantities include blade speed, power generation and blade bending moment.

[0059] S300, using the historical operating parameters as a training data set to train a neural network based on a Transformer model to obtain a neural network model, wherein the neural network model is used to predict the state of the wind turbine generator at the next time point and generate a control amount at the current time point;

[0060] Specifically, the neural network model includes a neural network prediction model and a neural network control model. Both the neural network prediction model and the neural network control model are neural networks based on the Transformer model. The input of the neural network prediction model is the state and control quantity of the wind turbine in the previous period of time, and the output of the neural network prediction model is the predicted value of the wind turbine state at the next moment. The input of the neural network control model includes constraints, the wind turbine state data of the previous 10-20 sampling points and the expected state output of the wind turbine. The output of the neural network control model is the wind turbine control quantity at the current moment, including pitch angle, yaw angle, and torque.

[0061] Using the neural network prediction model, the wind turbine state y at the next moment can be predicted m (k+1). From the collected historical operating parameters, select the relevant data at each time point, including wind force, wind direction, temperature, wind turbine yaw angle, pitch angle, torque, blade speed, power generation and blade bending moment. The sequence of state historical data and control values ​​at 9 consecutive time points is used as input and input into the neural network prediction model to calculate the state of the wind turbine at the next time point.

[0062] In this embodiment, the neural network control model does not use the method of solving the mathematical model to calculate the control quantity, but uses the neural network to fit the control quantity under different states. The neural network controller is developed by training with historical data and using supervised learning methods to minimize the error between the actual output and the expected output.

[0063] S400, optimizing the neural network model using a stochastic gradient descent algorithm in a simulation environment;

[0064] Specifically, under different wind conditions, the objective function of the wind turbine control system is set (i.e., ideal wind turbine state). For example: During the startup phase, the control goal is to rotate the wind turbine blades as fast as possible. During the rated power operation phase, the control goal is to keep the wind turbine blades rotating at a stable speed. During the protection operation phase, the control goal is to minimize the bending of the wind turbine blades.

[0065] The function of the neural network controller is to use the stochastic gradient descent algorithm to solve the optimal control variable u(k) at each moment k, so that the model predicts the state y of the wind turbine at the next moment. m (k+1), as close as possible to the ideal wind turbine state

[0066] The control variable u(k) at time t(k) is input into the wind turbine model of the Blade software to simulate the operation of the wind turbine.

[0067] At the next moment t(k+1), the actual state y(k+1) of the wind turbine is obtained from the sensor of the simulation software, the actual state value is input into the wind turbine prediction model, and the stochastic gradient descent algorithm is used to calculate the control quantity u(k+1) at t(k+1).

[0068] If the actual state of the wind turbine generator y(k+1) is different from the ideal state of the wind turbine generator If the gap between them is large, the parameters of the neural network controller, that is, the parameters of the stochastic gradient descent algorithm, can be modified.

[0069] In the simulation environment, the operating status of the wind turbine is tested under different wind speed and wind direction conditions to ensure that the wind turbine does not have accidents such as blade over-rotation and blade hitting the tower, and the power generation of the wind turbine is tested. If a failure occurs or the power generation of the wind turbine does not meet the requirements, the neural network controller must be modified, boundary restrictions must be added, and the parameters of the gradient descent algorithm must be modified.

[0070] S500: Controlling a wind turbine using the optimized neural network model.

[0071] In step S400, if the neural network model meets the requirements in the simulation environment, the wind turbine control system can be actually deployed to control the wind turbine.

[0072] In order to verify and further improve the control performance of the neural network model, the first wind turbine is controlled by using the tuned neural network model, and the second wind turbine is selected to be controlled by a native control system, wherein the first wind turbine and the second wind turbine are located in the same area; the first wind turbine and the second wind turbine are operated for a period of time and actual operation data are collected; by comparing the actual operation data of the first wind turbine and the second wind turbine in terms of blade speed, power generation and blade bending moment, the control performance of the neural network model in this embodiment can be verified, and the neural network model can be continuously tuned using the actual operation data.

[0073] For example, two wind turbines in close proximity are selected, and a neural network model is deployed in the control system of one of them, while the other maintains the original feedback control system. The two wind turbines are run for a period of time to collect the actual operation data of the wind turbines. The neural network controller is retrained with the data collected during the actual operation of the wind turbines, and its actual control effect is improved by continuously updating the model.

[0074] In this embodiment, real-time control of the wind turbine is achieved in a data-driven manner based on the Transformer model. The operating state of the wind turbine is predicted and the optimal control quantity is generated according to the environmental data, control quantity and output quantity of the wind turbine operation, so that the actual operating state of the wind turbine can reach the ideal operating state under the current conditions as much as possible, thereby improving the control performance of the wind turbine, increasing power generation, reducing the stress on the wind turbine blades and extending its life.

[0075] Example 2

[0076] like Figure 2 As shown, a wind turbine control device using an artificial neural network according to Embodiment 2 of the present application includes:

[0077] The simulation module 100 is used to construct a wind turbine simulation model in a simulation environment and operate the wind turbine;

[0078] The collection module 200 is used to collect historical operating parameters of the wind turbine, wherein the historical operating parameters include environmental data, control quantity and output quantity;

[0079] A training module 300, for training a neural network based on a Transformer model using the historical operating parameters as a training data set to obtain a neural network model, wherein the neural network model is used to predict the state of the wind turbine generator at the next time point and generate a control amount at the current time point;

[0080] A tuning module 400, used to tune the neural network model using a stochastic gradient descent algorithm in a simulation environment;

[0081] The application module 500 is used to control the wind turbine using the tuned neural network model.

[0082] It should be noted that other specific implementations of the wind turbine control device using an artificial neural network in this embodiment can refer to the specific implementations of the wind turbine control method using an artificial neural network mentioned above, and will not be repeated here to avoid redundancy.

[0083] Example 3

[0084] like Figure 3 As shown, a wind turbine control system using an artificial neural network according to Example 3 of the present application adopts the method described in any one of the implementations of Example 1 above, characterized in that the wind turbine control system includes:

[0085] The neural network control module 600 is used to obtain the best control quantity by using the stochastic gradient descent algorithm according to the current state of the system and the prediction result of the neural network prediction module 700, and the control quantity is output to the controlled system 900 to actually control the wind turbine generator according to the control quantity, wherein the control quantity includes the pitch angle, the yaw angle and the torque, so that the state of the wind turbine generator at the next time point predicted by the neural network control module is closest to the ideal wind turbine generator state;

[0086] A neural network prediction module 700 is used to predict the state of the wind turbine at the next time point based on the state of the wind turbine at the current time point and the control amount. The neural network prediction module 700 is trained based on input data and can capture the dynamic characteristics of the system, thereby improving the control accuracy;

[0087] The feedback optimization module 800 is used to optimize the parameters of the neural network control module according to the difference between the state of the wind turbine at the next time point predicted by the neural network control module and the actual state of the wind turbine at the next time point, so as to adjust the control amount output by the neural network control module 600 according to the actual feedback, to ensure that the system can still maintain stable control when there is external interference or the model is inaccurate, that is, the system status can be monitored in real time and necessary adjustments can be made.

[0088] It should be noted that other specific implementations of the wind turbine control system using an artificial neural network in this embodiment can refer to the specific implementations of the wind turbine control method using an artificial neural network mentioned above, and will not be repeated here to avoid redundancy.

[0089] Example 4

[0090] A computer-readable storage medium according to Embodiment 4 of the present application, wherein the computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing a method in any one of the implementations in Embodiment 1 of the present application;

[0091] Among them, the computer-readable storage medium can be a read-only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM); the computer-readable storage medium can store program code, and when the program stored in the computer-readable storage medium is executed by the processor, the processor is used to execute the steps of the method in any one of the implementation methods in Example 1 of the present application.

[0092] Example 5

[0093] like Figure 4As shown, an electronic device involved in Embodiment 5 of the present application includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, and when the program or instruction is executed by the processor, a method in any one of the implementations in Embodiment 1 of the present application is implemented;

[0094] Among them, the processor can adopt a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), a graphics processing unit (GPU) or one or more integrated circuits to execute relevant programs to implement the method in any one of the implementation methods in Example 1 of the present application.

[0095] The processor may also be an integrated circuit electronic device with signal processing capability. In the implementation process, each step of the method in any implementation of Embodiment 1 of the present application may be completed by an integrated logic circuit of hardware in the processor or by instructions in software form.

[0096] The above-mentioned processor may also be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed by a hardware decoding processor, or may be executed by a combination of hardware and software modules in a decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory, and completes the functions required to be performed by the unit included in the data processing device of the embodiment of the present application in combination with its hardware, or executes the method in any one of the implementation modes in the embodiment 1 of the present application.

[0097] The above are only preferred specific implementations of the present application; however, the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solution and its improved ideas of the present application within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A wind turbine control method using an artificial neural network, characterized in that: include: Build a wind turbine simulation model in the simulation environment and run the wind turbine; Collecting historical operating parameters of the wind turbine, wherein the historical operating parameters include environmental data, control quantity and output quantity; Using the historical operating parameters as a training data set to train a neural network based on a Transformer model to obtain a neural network model, wherein the neural network model is used to predict the state of the wind turbine generator at the next time point and generate the control amount at the current time point; Using a stochastic gradient descent algorithm to tune the neural network model in a simulation environment; The wind turbine generator is controlled using the tuned neural network model.

2. The wind turbine control method using an artificial neural network according to claim 1, characterized in that: Running the wind turbine simulation model in the simulation environment includes: testing the operation status of the wind turbine under different wind force and wind direction conditions in the simulation environment.

3. The wind turbine control method using an artificial neural network according to claim 1 or 2, characterized in that: Collecting historical operating parameters of the wind turbine includes: collecting historical operating parameters of the wind turbine at n consecutive time points in a simulation environment, wherein n is a positive integer greater than 1, and the historical operating parameters include environmental data, control quantities and output quantities, the environmental data include wind force, wind direction and air temperature; the control quantities include yaw angle, pitch angle and torque; the output quantities include blade speed, power generation and blade bending moment.

4. The wind turbine control method using an artificial neural network according to claim 3, characterized in that: The neural network model includes a neural network prediction model and a neural network control model, wherein the neural network prediction model is used to predict the state of the wind turbine at the next time point, and the neural network control model is used to generate the control quantity at the current time point, and the control quantity includes the pitch angle, yaw angle and torque.

5. The wind turbine control method using an artificial neural network according to claim 4, characterized in that: The neural network model is tuned using a stochastic gradient descent algorithm in a simulation environment, including: S401, setting an ideal wind turbine generator state under different wind conditions; S402, at a first moment, using a neural network prediction model to predict a state of a wind turbine generator at a next moment, and using a neural network control model to generate a control quantity at the first moment; S403, in a simulation environment, inputting the generated control quantity at the first moment into a wind turbine simulation model and simulating the operation of the wind turbine; S404, collecting the actual wind turbine state of the wind turbine simulation model at the next moment, and modifying the parameters of the neural network control model according to the gap between the actual wind turbine state at the next moment and the ideal wind turbine state; S405, repeating steps S302 to S304 until the gap between the actual state of the wind turbine generator and the ideal state of the wind turbine generator at the next moment meets the user's needs.

6. The wind turbine control method using an artificial neural network according to any one of claims 1, 2, 4 and 5, characterized in that: Also includes: Using the tuned neural network model to control the first wind turbine, and selecting the second wind turbine to be controlled by the native control system, wherein the first wind turbine and the second wind turbine are located in the same area; operating the first wind turbine and the second wind turbine for a period of time and collecting actual operating data; The actual operation data is used to continuously tune the neural network model.

7. A wind turbine control device using an artificial neural network, characterized in that: include: A simulation module, used to build a wind turbine simulation model and run the wind turbine in a simulation environment; A collection module, used for collecting historical operating parameters of the wind turbine, wherein the historical operating parameters include environmental data, control quantity and output quantity; A training module, used for training a neural network based on a Transformer model using the historical operating parameters as a training data set to obtain a neural network model, wherein the neural network model is used to predict the state of the wind turbine generator at the next time point and generate a control amount at the current time point; A tuning module, used for tuning the neural network model using a stochastic gradient descent algorithm in a simulation environment; The application module is used to control the wind turbine using the tuned neural network model.

8. A wind turbine control system using an artificial neural network, using the method according to any one of claims 4 to 6, characterized in that: The wind turbine control system comprises: A neural network control module is used to obtain an optimal control amount using a stochastic gradient descent algorithm so that the state of the wind turbine generator at the next time point predicted by the neural network control module is closest to the ideal wind turbine generator state; A neural network prediction module is used to predict the state of the wind turbine at the next time point based on the state of the wind turbine at the current time point and the control amount; The feedback optimization module is used to optimize the parameters of the neural network control module according to the difference between the state of the wind turbine generator at the next time point predicted by the neural network control module and the actual state of the wind turbine generator at the next time point.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing the method according to any one of claims 1 to 6.

10. An electronic device, characterized in that: The electronic device comprises a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction implements the method as claimed in any one of claims 1 to 6 when executed by the processor.

Images