Fan model predictive control method and system based on Gaussian process error compensation

Abstract

The present invention proposes a wind turbine model predictive control method and system based on Gaussian process error compensation. It does not rely on complex mechanism models, but predicts the model output compensation in real time through the Gaussian model, and dynamically corrects the model accordingly, effectively overcoming the high dependence of traditional methods on model accuracy and simplifying the modeling process. The Gaussian model is used to predict future wind speeds and output compensation. The Gaussian model has excellent nonlinear approximation capabilities and can accurately predict future complex and changeable working conditions and system states, significantly improving the effectiveness of the MPC control strategy and achieving more accurate wind turbine control.

Description

Technical Field

[0001] The present invention belongs to the technical field of floating wind turbine control, and in particular relates to a wind turbine model prediction control method and system based on Gaussian process error compensation. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Unlike fixed wind turbines, floating wind turbines are installed on floating platforms at sea and are more susceptible to environmental factors such as waves and currents, showing more complex motion characteristics. This motion characteristic will have a significant impact on the aerodynamic load of the wind turbine blades, which in turn affects the operating safety and fatigue life of the wind turbine. In order to meet this challenge, variable pitch load reduction control has become one of the key technologies in the control strategy of floating wind turbines. Model predictive control (MPC) has the advantages of multi-objective optimization (floating platform motion stability, power generation, etc.) and the ability to directly incorporate system constraints (pitch angle change rate constraints, etc.) into the optimization problem. The floating wind turbine variable pitch load reduction control method based on MPC has become an advanced control method.

[0004] The floating wind turbine variable pitch load reduction control technology based on MPC is a control method specifically used to cope with the load reduction requirements of floating wind turbines in complex marine environments. The core idea of ​​MPC is to predict the dynamic behavior of the system in the future through a pre-established prediction model in each control cycle, and optimize the control input in real time based on these prediction results to achieve the predetermined control target. It can effectively cope with uncertain factors such as wind speed changes and platform movement, and achieve safe, efficient and long-life operation of wind turbines.

[0005] Although MPC-based floating wind turbine variable pitch load reduction control has shown many superior performances in complex marine environments, it also has some shortcomings and challenges in practical applications:

[0006] The performance of the model is highly dependent on the accuracy of the system model. For floating wind turbines, environmental conditions are complex and changeable, such as wind speed fluctuations, platform movement, and waves, which will affect the accuracy of the model. If the model cannot accurately reflect the dynamic behavior of the actual system, it may lead to poor control effects and even cause system instability. The prediction accuracy is limited. MPC needs to predict the system state in the future based on the current operating conditions. However, environmental factors such as wind speed and waves are inherently highly random and volatile, which poses a great challenge to accurately predicting future operating conditions. Moreover, when encountering sudden changes in operating conditions such as sudden strong winds and extreme sea conditions, the system state may change drastically, leading to system oscillations, over-limits, and even endangering the safety of the wind turbine. Summary of the invention

[0007] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a wind turbine model predictive control method and system based on Gaussian process error compensation. The present invention adopts a prediction method based on a Gaussian model to capture model errors in real time and predict the compensation amount online to correct the system model, thereby improving the model accuracy and adaptability to complex sea conditions.

[0008] In order to achieve the above object, the present invention adopts the following technical solution:

[0009] In a first aspect, the present invention provides a fan model predictive control method based on Gaussian process error compensation, comprising:

[0010] Based on the difference between the state and output of the floating wind turbine and the corresponding stable operating point, a linear model of the floating wind turbine is constructed;

[0011] According to the wave and wind speed conditions of the floating wind turbine, as well as the actual state and output of the floating wind turbine, a nonlinear model of the floating wind turbine is constructed;

[0012] Based on the Gaussian model, the output compensation between the output of the linear model and the output of the nonlinear model of the floating wind turbine and the future wind speed are predicted respectively;

[0013] constructing an objective function of floating wind turbine control according to the predicted output compensation amount and the future wind speed, and solving the objective function by minimizing the pitch angle control amount;

[0014] Based on the solution results and the output results of the unified pitch controller, the pitch angle of the floating wind turbine is controlled.

[0015] In a second aspect, the present invention provides a fan model predictive control method based on Gaussian process error compensation, comprising:

[0016] A first building module is configured to: build a linear model of the floating wind turbine based on a difference between the state and output of the floating wind turbine and a corresponding stable operating point;

[0017] A second building module is configured to: build a nonlinear model of the floating wind turbine according to the wave and wind speed conditions of the floating wind turbine and the actual state and output of the floating wind turbine;

[0018] A prediction module is configured to: predict an output compensation amount between an output of a linear model and an output of a nonlinear model of the floating wind turbine, and a future wind speed, respectively, based on a Gaussian model;

[0019] A solution module is configured to: construct an objective function of floating wind turbine control according to the predicted output compensation amount and the future wind speed, and solve the objective function to minimize the pitch angle control amount;

[0020] The control module is configured to control the pitch angle of the floating wind turbine based on the solution result and the output result of the unified pitch controller.

[0021] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method described in the first aspect is performed.

[0022] In a fourth aspect, the present invention provides a computer-readable storage medium for storing computer instructions, wherein when the computer instructions are executed by a processor, the method described in the first aspect is performed.

[0023] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method described in the first aspect.

[0024] One or more of the above technical solutions have the following beneficial effects:

[0025] In the present invention, there is no need to rely on complex mechanism models. Instead, the output compensation of the model is predicted in real time through the Gaussian model, and the model is dynamically corrected accordingly, which effectively overcomes the high dependence of traditional methods on model accuracy and simplifies the modeling process. The Gaussian model is used to predict future wind speeds and output compensation. The Gaussian model has excellent nonlinear approximation capabilities and can accurately predict future complex and changeable working conditions and system states, significantly improving the effectiveness of the MPC control strategy and achieving more accurate wind turbine control.

[0026] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0028] FIG1 (a) is a discrete time domain MPC control strategy in Embodiment 1 of the present invention;

[0029] FIG1( b ) is a GP wind speed prediction in Embodiment 1 of the present invention;

[0030] Figure 2 1 is a block diagram of a floating wind turbine variable pitch control system based on MPC in Embodiment 1 of the present invention;

[0031] Figure 3 The floating wind turbine MPC variable pitch load reduction control strategy based on GP error compensation in the first embodiment of the present invention;

[0032] Figure 4 This is the GSPI pitch controller in the first embodiment of the present invention. DETAILED DESCRIPTION

[0033] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0034] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0035] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0036] Embodiment 1

[0037] This embodiment discloses a fan model predictive control method based on Gaussian process error compensation, including:

[0038] Based on the difference between the state and output of the floating wind turbine and the corresponding stable operating point, a linear model of the floating wind turbine is constructed;

[0039] According to the wave and wind speed conditions of the floating wind turbine, as well as the actual state and output of the floating wind turbine, a nonlinear model of the floating wind turbine is constructed;

[0040] Based on the Gaussian model, the output compensation between the output of the linear model and the output of the nonlinear model of the floating wind turbine and the future wind speed are predicted respectively;

[0041] constructing an objective function of floating wind turbine control according to the predicted output compensation amount and the future wind speed, and solving the objective function by minimizing the pitch angle control amount;

[0042] Based on the solution results and the output results of the unified pitch controller, the pitch angle of the floating wind turbine is controlled.

[0043] This embodiment proposes a floating wind turbine MPC variable pitch load reduction control strategy based on Gaussian process (GP) error compensation. This strategy innovatively adopts a GP-based prediction method to correct the system model by capturing the model error in real time and predicting the compensation amount online, thereby improving the model accuracy and adaptability to complex sea conditions. By capturing the model error in real time and using the Gaussian process to predict the compensation amount at the next moment online, the system model is corrected to improve the accuracy of the wind turbine model; using machine learning-based GP prediction, selecting appropriate kernel functions and hyperparameters, real-time prediction of extreme sea conditions such as turbulent winds, and improving the model's adaptability to working conditions.

[0044] Combine the following Figure 3 A fan model predictive control method based on Gaussian process error compensation proposed in this embodiment is described in detail.

[0045] 1. MPC control strategy and GP wind speed prediction

[0046] MPC is an advanced control strategy that dynamically adjusts control inputs by predicting future system behavior to achieve the desired control objectives.

[0047] Figure 1(a) is a diagram of the MPC control strategy in the discrete time domain. The red curve represents the reference trajectory, that is, the target signal that the MPC expects the system output to track. The reference signal is usually determined by the control target, such as maintaining the rated power output and reducing the structural load of the wind turbine. MPC predicts the behavior of the system in the future based on a mathematical model. The purple step signal in the figure represents the control input obtained by optimization at the current time t (such as the adjustment of the pitch angle). These control inputs determine how the actual output of the system will approach the reference trajectory. The core of MPC is rolling optimization, that is, in each control cycle (such as t to t+N in the figure), MPC will optimize the system output for a period of time in the future according to the prediction model and calculate a set of optimal control inputs. However, it should be noted that MPC will only perform the control action at the current moment, and then roll to the next moment to re-predict and optimize, which ensures that MPC can respond to environmental changes and system state deviations in real time. Finally, the blue curve in the figure represents the actual output of the system, which should ideally be as close to the reference trajectory as possible. MPC continuously adjusts the control input so that the system output effectively tracks the reference signal while taking into account the system constraints (such as physical limitations, pitch angle constraints, etc.).

[0048] Since wind speed changes directly affect the output power and structural load of floating wind turbines, accurate wind speed prediction is crucial to achieve efficient wind turbine predictive control. Figure 1(b) shows the process of wind speed prediction using Gaussian process. The specific steps are:

[0049] Step 1: Data training.

[0050] The left part of Figure 1(b) shows the historical data of wind speed (blue dots), which are used to train the Gaussian process model. These historical data contain the laws and trends of wind speed changes, which can help the Gaussian process model learn and capture the uncertainty and noise in the data. During the training process, the system builds a priori Gaussian process model based on historical data. The model can make a preliminary prediction of future wind speed and quantify the uncertainty of the prediction results, that is, the predicted value will have a certain variance and mean. The specific formula is as follows:

[0051] (1)

[0052] In the formula, represents any input, here represented as a time series of wind speed; Represent different random variables The corresponding function value; The covariance function (kernel function) of random functions corresponding to different inputs is used to describe the similarity between random functions and determines the properties of the GP model, such as smoothness and periodicity. The SM (Spectral Mixture) kernel function is a flexible and expressive kernel function. Based on the Fourier transform principle, it combines Gaussian kernels corresponding to specific frequencies and length scales to model various function forms in the form of:

[0053] (2)

[0054] In the formula, is the number of mixed components, Indicates The weight of the contribution of each component, Respectively The mean and variance of the components.

[0055] Step 2: Online prediction. As time goes by, new wind speed data continues to arrive, and the Gaussian process model can effectively integrate these new data and update them online. At this time, the prior model is fused with the new data to generate a posterior Gaussian process model. The model adjusts the predicted mean and variance according to the new data, thereby improving the accuracy of the prediction and more accurately predicting future wind speed changes.

[0056] Gaussian process regression implements the implicit functional relationship between the input and output of the training data set, and then obtains the entire function The joint Gaussian distribution of the function value combination can be expressed as:

[0057] (3)

[0058] In the above formula, the covariance function And the mean function Both The matrix of .

[0059] Since the goal is to predict the system's performance on new inputs based on the Gaussian process posterior distribution obtained from the training set The output of , the augmented joint Gaussian distribution can be expressed as:

[0060] (4)

[0061] in, Thus, the new input can be derived The prediction of the system output is:

[0062] (5)

[0063] In the above formula, denote the mean and variance of the prediction, respectively.

[0064] Step 3: Prediction output. The trained Gaussian process model will output the predicted value of wind speed at future moments and provide the corresponding confidence interval to quantify the uncertainty of the prediction. As shown in Figure 1(b), the red mark represents the actual mean of the model prediction results, and the root mean square value is used to evaluate the quality of the prediction results:

[0065] (6)

[0066] 2. MPC-based floating wind turbine pitch control

[0067] Figure 2 is the block diagram of the floating wind turbine pitch control system based on MPC.

[0068] In the picture The sensor detects the status of the fan , Output The difference between the corresponding steady-state operating point is used to obtain the linearized model of the fan:

[0069] (7)

[0070] In the formula, is the pitch angle control input, The wind speed disturbance. is the state matrix, is the control input matrix of state changes caused by pitch angle input, is the wind input disturbance matrix of the state change caused by wind speed disturbance, is the output state matrix, is the control input transfer matrix for output changes caused by pitch angle input, is the wind disturbance transmission matrix of the output change caused by wind disturbance. The above coefficients are obtained by performing linearization function on the wind turbine at the operating point using OpenFAST, a high-fidelity floating wind turbine simulation software developed by the National Renewable Energy Laboratory (NREL) of the United States.

[0071] For convenience, the following simplification is made to formula (7):

[0072] (8)

[0073] In MPC, the linear model is used to predict the state and output of the system at N future moments. The control target is to meet the following constraints: ① Considering the constraints of the action range of the variable pitch actuator, let ; ② Considering the inertia of the pitch actuator, constraints are imposed on the pitch angle change speed , minimize the deviation between the predicted system output and the reference value, and avoid frequent actions of the pitch angle control mechanism, that is, to achieve the maximum benefit with the minimum control amount. Therefore, the objective function of the optimization problem can be described as:

[0074] (9)

[0075] in, Q , R Respectively represent the output weight matrix and input weight matrix of the system. By adjusting the weight matrix, the optimization effect is balanced between reducing the load and reducing the frequent action of the pitch angle. Q elements in the matrix, MPC will pay more attention to the reduction of output error, making the load reduction effect faster. Q The matrix may cause drastic changes in the pitch angle input action, increasing the difficulty of operating the system's pitch angle actuator. R The matrix constrains the size of the pitch angle control input to prevent the pitch angle from changing too frequently or drastically. R The elements in the matrix can reduce the controller's action range and make the system more stable, but too large R The matrix may result in inaccurate state tracking and slow system response.

[0076] For the above optimization problem, the optimal control quantity is solved , giving the reference adjustment of the wind turbine IPC pitch angle, that is, When the time step moves to the next moment, the system will update the prediction of the system state and output according to the latest sensor measurement value, and recalculate the optimal control quantity at the next moment based on the rolling optimization strategy, thereby realizing real-time dynamic adjustment of the wind turbine pitch angle.

[0077] 3. MPC variable pitch load reduction control strategy based on GP error compensation

[0078] The linearized model represented by equation (8) has defects, such as the influence of wave information is not considered in the model, and the model accuracy needs to be improved, which greatly reduces the control effect of directly applying the model to MPC. Therefore, the model needs to be corrected. Considering that the optimization goal is that the system output predicted by the model can better track the reference value, the model output equation is corrected to offset the model error caused by various factors. The corrected model is shown in equation (10).

[0079] (10)

[0080] In the formula is the output of the linearized model represented by equation (8): Output of the actual OpenFAST nonlinear model The error between them is:

[0081] (11)

[0082] Similar to the GP-based wind speed forecast, the wind speed prediction based on the current time t is The historical data is used to establish the Gaussian model prior distribution, select appropriate hyperparameters, and update the data and prior model in real time to obtain the Gaussian model's posterior distribution to achieve the output compensation of the predicted linear model. , thereby achieving effective correction of the linearized model.

[0083] The complete control system block diagram of this method is shown in Figure 3. The unified pitch control (CPC) adopts a variable gain PI (GSPI) controller to adjust the wind turbine speed and thus control the output power; the independent pitch control (IPC) adopts the MPC pitch control strategy based on GP error compensation of the present invention to make a more precise adjustment of the pitch angle. Respectively represent the amount of compensation used to predict With wind speed The GP model established.

[0084] The block diagram of the GSPI controller is shown in Figure 4. The input is the actual speed output of the generator. With reference rated speed The difference between the fan speed and rated speed is realized by the GSPI controller, which can be specifically expressed as:

[0085] (12)

[0086] The controller parameters are expressed as output pitch angle The function of offsets the influence of small disturbances of pitch angle on the sensitivity of power under different wind speeds, that is:

[0087] (13)

[0088] in, are the controller parameters when no pitch change is performed, For gain:

[0089] (14)

[0090] in, is the experimental fitting coefficient, which is approximately .

[0091] The specific working steps of the fan model predictive control method based on Gaussian process error compensation proposed in this embodiment are:

[0092] Step 1: Calculate the output of the linearized model based on the wind, wave, control signal and other data obtained by the sensor The output of the actual nonlinear model , and calculate the error between the two according to formula (11).

[0093] Step 2: Use the established GP model Online prediction of the compensation output of the linearized model and future wind speed .

[0094] Step 3: Based on the compensation model (10) and the objective function (9), the output of the system at future moments is predicted and optimized in a rolling manner to solve the optimal control quantity. .

[0095] Step 4: Calculate the output of the CPC part of GSPI according to equations (12)-(14) , and The superposition forms a complete control signal, which is transmitted to the wind turbine as a pitch angle reference to form a control closed loop.

[0096] This embodiment aims to improve the accuracy of the control model, enhance the robustness of the system, achieve more effective load reduction control of floating wind turbines, and improve their operating safety and efficiency under complex sea conditions by combining Gaussian process prediction and model predictive control technology.

[0097] The present invention can ensure the accuracy of the model in the entire time domain by predicting the error between the linearized model and the actual wind turbine simulation system and correcting the linearized model. In terms of result verification, the method proposed by the present invention can significantly reduce power fluctuations and the load on the wind turbine, including not only the load on the tower, but also the torque on the wind turbine blades, the main movement of the floating platform and other load forms. Moreover, since the model error compensation prediction of the present invention is an online prediction, it can adapt to different working conditions and therefore has stronger robustness and adaptability.

[0098] This embodiment can achieve a more accurate model, stronger applicability of the model, and higher prediction accuracy in the MPC control of floating wind turbines, and has the following advantages:

[0099] 1. No need for accurate mechanism model: This embodiment does not rely on complex mechanism models, but uses the GP model to predict the model compensation in real time and dynamically correct the model accordingly, effectively overcoming the high dependence of traditional methods on model accuracy and simplifying the modeling process.

[0100] 2. Higher prediction accuracy: The GP model used in this embodiment has excellent nonlinear approximation ability and can accurately predict future complex and changeable working conditions and system states, significantly improving the effectiveness of the MPC control strategy and achieving more accurate fan control.

[0101] 3. Stronger adaptability; the GP model used in this embodiment can be updated online according to new data, and can continuously learn and adapt to changes in the marine environment and system characteristics, thereby enhancing the robustness and adaptability of the control system and making it perform better in a complex and changing marine environment.

[0102] Embodiment 2

[0103] The purpose of this embodiment is to provide a fan model predictive control system based on Gaussian process error compensation, including:

[0104] A first building module is configured to: build a linear model of the floating wind turbine based on a difference between the state and output of the floating wind turbine and a corresponding stable operating point;

[0105] A second building module is configured to: build a nonlinear model of the floating wind turbine according to the wave and wind speed conditions of the floating wind turbine and the actual state and output of the floating wind turbine;

[0106] A prediction module is configured to: predict an output compensation amount between an output of a linear model and an output of a nonlinear model of the floating wind turbine, and a future wind speed, respectively, based on a Gaussian model;

[0107] A solution module is configured to: construct an objective function of floating wind turbine control according to the predicted output compensation amount and the future wind speed, and solve the objective function to minimize the pitch angle control amount;

[0108] The control module is configured to control the pitch angle of the floating wind turbine based on the solution result and the output result of the unified pitch controller.

[0109] In further embodiments, there is also provided:

[0110] An electronic device includes a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method described in Embodiment 1 is performed. For the sake of brevity, no further description is given here.

[0111] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0112] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0113] A computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the method described in embodiment 1 is completed.

[0114] The method in the first embodiment can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor. The software module can be located in a mature storage medium in the field 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 the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.

[0115] A computer program product includes a computer program, and when the computer program is executed by a processor, the method described in the first embodiment is implemented.

[0116] The present invention also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer executable instructions, such as instructions included in a program module, which are executed in a device on a real or virtual processor of the target to perform the process / method as described above. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or implement specific abstract data types. In various embodiments, the functions of program modules can be combined or divided between program modules as needed. Machine executable instructions for program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0117] The computer program code for implementing the method of the present invention can be written in one or more programming languages. These computer program codes can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the computer or other programmable data processing device, causes the function / operation specified in the flow chart and / or block diagram to be implemented. The program code can be executed completely on a computer, partially on a computer, as an independent software package, partially on a computer and partially on a remote computer or completely on a remote computer or server.

[0118] In the context of the present invention, computer program codes or related data may be carried by any appropriate carrier to enable a device, apparatus or processor to perform the various processes and operations described above. Examples of carriers include signals, computer readable media, etc. Examples of signals may include electrical, optical, radio, acoustic or other forms of propagation signals, such as carrier waves, infrared signals, etc.

[0119] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0120] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A fan model predictive control method based on Gaussian process error compensation is characterized in that: include: Based on the difference between the state and output of the floating wind turbine and the corresponding stable operating point, a linear model of the floating wind turbine is constructed; According to the wave and wind speed conditions of the floating wind turbine, as well as the actual state and output of the floating wind turbine, a nonlinear model of the floating wind turbine is constructed; Based on the Gaussian model, the output compensation between the output of the linear model and the output of the nonlinear model of the floating wind turbine and the future wind speed are predicted respectively; constructing an objective function of floating wind turbine control according to the predicted output compensation amount and the future wind speed, and solving the objective function by minimizing the pitch angle control amount; Based on the solution results and the output results of the unified pitch controller, the pitch angle of the floating wind turbine is controlled.

2. The wind turbine model predictive control method based on Gaussian process error compensation according to claim 1, characterized in that: According to the error between the linear model and the nonlinear model of the floating wind turbine, a Gaussian model prior distribution is established, appropriate hyperparameters are selected, and the error data and the prior model are updated in real time to obtain the posterior distribution of the Gaussian model and realize the prediction of the output compensation amount of the linearized model.

3. The wind turbine model predictive control method based on Gaussian process error compensation according to claim 1, characterized in that: The unified variable pitch controller uses a variable gain PI controller to adjust the wind turbine speed and control the output power of the floating wind turbine. Specifically: Characterize the parameters of the variable gain PI controller as a function of the output pitch angle; The gain is determined based on the output pitch angle, and the parameters of the variable gain PI controller are then determined.

4. The wind turbine model predictive control method based on Gaussian process error compensation according to claim 1, characterized in that: The objective function of the constructed floating wind turbine control is as follows: Among them, Q and R represent the output weight matrix and input weight matrix of the system respectively, |u (k) |≤u max represents the constraint condition of the action range of the pitch actuator, |Δu (k) |≤Δu max It represents the constraint condition of the pitch angle change speed, and the superscript T represents transposition.

5. The wind turbine model predictive control method based on Gaussian process error compensation according to claim 1, characterized in that: A Gaussian prior model is constructed based on historical wind speed data, and the real-time wind speed data is fused with the Gaussian prior model to obtain a Gaussian posterior model; the Gaussian posterior model adjusts the predicted mean and variance according to the real-time wind speed data to obtain a trained Gaussian process model; and future wind speeds are predicted based on the trained Gaussian process model.

6. The wind turbine model predictive control method based on Gaussian process error compensation according to claim 1, characterized in that: The linear model of the floating wind turbine is corrected by using the output compensation between the output of the linear model of the floating wind turbine and the output of the nonlinear model. The corrected linear model of the floating wind turbine is: Where Dy is the error between the output of the linearized model and the output of the actual OpenFAST nonlinear model, A is the state matrix, B is the control input matrix of the state change caused by the pitch angle input, and B d is the wind input disturbance matrix of state change caused by wind speed disturbance, C is the output state matrix, and D is the control input transfer matrix of output change caused by pitch angle input.

7. A fan model predictive control system based on Gaussian process error compensation, characterized in that: include: A first building module is configured to: build a linear model of the floating wind turbine based on a difference between the state and output of the floating wind turbine and a corresponding stable operating point; A second building module is configured to: build a nonlinear model of the floating wind turbine according to the wave and wind speed conditions of the floating wind turbine and the actual state and output of the floating wind turbine; A prediction module is configured to: predict an output compensation amount between an output of a linear model and an output of a nonlinear model of the floating wind turbine, and a future wind speed, respectively, based on a Gaussian model; A solution module is configured to: construct an objective function of floating wind turbine control according to the predicted output compensation amount and the future wind speed, and solve the objective function to minimize the pitch angle control amount; The control module is configured to control the pitch angle of the floating wind turbine based on the solution result and the output result of the unified pitch controller.

8. An electronic device, characterized in that: The method comprises a memory and a processor and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method according to any one of claims 1 to 6 is completed.

9. A computer-readable storage medium, characterized in that: Used to store computer instructions, which, when executed by a processor, complete the method described in any one of claims 1 to 6.

10. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images