Fixed-time adaptive sliding mode control method for variable speed wind turbine based on full drive system

Abstract

The application discloses a variable-speed wind turbine fixed-time adaptive sliding mode control method based on a full-drive system, and comprises the following steps: establishing a mathematical model of a transmission system of a variable-speed wind turbine; constructing a state space equation of a variable-speed wind turbine system according to the mathematical model of the transmission system and a generator torque model; establishing a full-drive system model of the variable-speed wind turbine according to the state space equation of the variable-speed wind turbine system; establishing a sliding mode function and a time derivative function thereof according to a sliding mode control theory; approximating lumped unknown terms in the full-drive system model of the variable-speed wind turbine according to a neural network technology; designing a fixed-time adaptive sliding mode controller according to a fixed-time theory and an adaptive sliding mode control theory; and controlling rotor speed of the variable-speed wind turbine through the fixed-time adaptive sliding mode controller. The application can make tracking error converge within a fixed time, and the convergence time is weakly dependent on an initial state, so that the predictability and reliability in engineering application are improved.

Description

Technical Field

[0001] This invention relates to the field of wind power, specifically to a fixed-time adaptive sliding mode control method for variable-speed wind turbine generators based on an all-drive system. Background Technology

[0002] Wind turbines can be classified into fixed-speed wind turbines (FSWT) and variable-speed wind turbines (VSWT) based on their rotational speed characteristics. Compared to FSWT, VSWT has higher energy capture efficiency, smaller power fluctuations, and stable operation under different wind conditions, and is therefore widely used in modern wind farms. The control strategy of VSWT is typically designed according to the operating area of ​​the turbine: when the wind speed is below the rated wind speed, maximum power point tracking (MPPT) is mainly achieved by adjusting the generator torque to improve wind energy utilization while maintaining a constant pitch angle; when the wind speed exceeds the rated wind speed, pitch control is used to adjust the output power to the rated value, thereby ensuring stable system operation.

[0003] Various methods have been proposed by the academic community for MPPT control of VSWT. For example, some studies have proposed MPPT algorithms based on rotor inertial power, using proportional controllers to reduce the impact of rotational inertia on the response; others have used variable parameter nonlinear controllers to improve MPPT efficiency while mitigating transient loads on the transmission system; in addition, state feedback and feedforward gain methods based on precise output regulation (EOR) can also enhance power control performance. Although these methods have certain advantages in theory, problems such as modeling uncertainties, external disturbances, and parameter variations still exist in practical VSWT systems, which pose significant challenges to controller design.

[0004] Sliding mode control (SMC) has become an important means to improve the robustness and dynamic response of MPPT (Multi-Level Power Transmission) systems due to its insensitivity to system disturbances and parameter changes, as well as its rapid adaptive capability. In practical applications, sliding mode control can effectively improve the anti-interference capability and transient performance of VSWT (Variable Voltage Response System) under external disturbances. However, traditional SMC designs still suffer from limited control accuracy and insufficient convergence speed when dealing with complex wind power generation systems with strong nonlinearity, large disturbances, and model uncertainties.

[0005] To further improve the convergence performance of tracking errors, finite-time control has been introduced into nonlinear system control. Finite-time control can achieve state convergence within a finite time, offering advantages such as fast convergence, high accuracy, and strong robustness. In the VSWT field, finite-time control has been applied to actuator fault estimation, aerodynamic load estimation, floating platform stability, and blade vibration suppression. However, the convergence time of finite-time control typically depends on initial conditions, which limits its application scope in practical engineering.

[0006] To address this issue, fixed-time (FTS) control theory has been proposed. Its upper bound on convergence time is less dependent on initial conditions and can be preset in engineering through design parameters. Despite its theoretical advantages, the application of fixed-time control in VSWT systems remains limited. For example, while fixed-time fractional-order SMC methods have been proposed to improve power quality, these methods rely on accurate system models. Other fixed-time sliding mode control based on state observers is used for MPPT, but stability analysis depends on discontinuous Lyapunov functions, potentially leading to inaccurate stability assessments. Furthermore, adaptive fixed-time control has not been sufficiently validated in complex real-world environments when wind speed data is overly idealized.

[0007] Traditional fixed-time controller design typically employs backstepping, treating the VSWT model as a strictly feedback system. However, the backstepping design process is complex, requiring recursive construction of virtual control laws and repeated differentiation, resulting in cumbersome controller structures and numerous parameters, limiting its practical application. Therefore, the Full-Drive System (FAS) method has gained attention in recent years. The FAS method emphasizes the flexibility of controller design, rather than relying solely on state-space representation and recursive design, significantly simplifying controller structure and improving performance. It has been successfully applied in multi-agent systems, spacecraft, and microgrids. However, the application of the FAS method in VSWT control remains limited, providing a technological opportunity to further improve MPPT control performance.

[0008] In summary, while existing technologies have proposed various control strategies for wind turbine power regulation, they still suffer from problems such as strong modeling dependence, insufficient robustness to external disturbances, convergence time dependence on initial values, and complex controller structures. Therefore, there is an urgent need for an improved control method that can achieve error convergence within a fixed time, reduce model dependence, and ensure robustness and control simplicity, providing technical support for efficient and reliable wind power generation. Summary of the Invention

[0009] The purpose of this invention is to provide a fixed-time adaptive sliding mode control method for variable speed wind turbines based on an all-drive system, so as to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution: A fixed-time adaptive sliding mode control method for variable-speed wind turbines based on an all-drive system includes: Step 1: Establish a mathematical model of the transmission system of the variable speed wind turbine based on the mechanical principles of the transmission system; Step 2: Construct the state-space equations of the variable-speed wind turbine system based on the mathematical model of the transmission system and the torque model of the generator; Step 3: Establish a model of the variable speed wind turbine all-drive system based on the state-space equations of the variable speed wind turbine system; Step 4: Establish the sliding mode function and its time derivative function based on sliding mode control theory; Step 5: Approximate the lumped unknowns in the variable speed wind turbine all-drive system model using neural network technology; Step 6: Design a fixed-time adaptive sliding mode controller based on fixed-time theory and adaptive sliding mode control theory; Step 7: Control the rotor speed of the variable speed wind turbine using a fixed-time adaptive sliding mode controller.

[0011] Furthermore, the mathematical model of the variable-speed wind turbine transmission system established in step 1 is expressed as follows: , Among them, Ω r It is the rotor speed. It is Ω r The derivative with respect to time, J t It is the inertial constant, K t It is the damping constant, ρ and R represent the air density and rotor radius, respectively, and C p (λ,β) is the power coefficient, where λ and β represent the tip velocity ratio and blade pitch angle, respectively, v is the wind speed, and T is the propeller angle. g d is the generator torque, and d is the comprehensive disturbance term, which includes the uncertainty in system modeling and the measurement noise of wind speed v.

[0012] Furthermore, in step 1, hypothesis 1 is constructed: the comprehensive interference term d and its time derivative. Both have an upper bound.

[0013] Furthermore, the generator torque T g The adjustment is made through the actuator of the generator, and is expressed as follows: , Where, τ g It is a time constant. Generator torque T g Time derivative, T g,des It is the generator torque T g The expected value.

[0014] Further, step 2 includes: Make the rotor speed Ω r Track its expected value Ω d Set the system state x1=Ω r -Ω d Let the tracking error be denoted as x2=T. g Fixed-time adaptive sliding mode controller u=Tg,des The state-space equations of the variable-speed wind turbine system are then obtained as follows: , in, , , x1, x2, and Ω are respectively d The time derivative of the function f(x1,Ω) d ,v,λ) is abbreviated as f.

[0015] Furthermore, in step 3, the established model of the variable-speed wind turbine all-drive system is as follows: , in, , , They are respectively , The time derivative of f The lumped unknown term in the all-drive system model of a variable-speed wind turbine is denoted as Δf, a constant. .

[0016] Further, step 4 includes: The sliding mode function s and its time derivative function Designed as follows: , , Where k1 and k2 are both positive constants, the positive constant γ1 satisfies γ1>1, and the positive constant γ2 satisfies 0<γ2<1. and It is a sign power function, a function , It is a function The derivative with respect to time.

[0017] Further, step 5 includes: The lumped unknown term Δf is approximated by a neural network in the following form: , in, This is the optimal weight vector, with the superscript T indicating transpose. For the input vector, This is represented as the approximation error of the neural network. Here is the activation function of the neural network, abbreviated as: .

[0018] Furthermore, in step 5, hypothesis 2 is constructed: and There are limits to everything.

[0019] Furthermore, in step 6, the designed fixed-time adaptive sliding mode controller u is: , Where κ, ε, A, c1, and c2 are all positive constants greater than 0, positive constant γ3 > 1, and positive constant γ4 satisfies 0 < γ4 < 1, g -1 This represents the inverse operation of the constant g. It is the hyperbolic tangent function. and It is a sign power function. yes The estimated value, The adaptive update rate is designed as follows: , in, yes The derivatives with respect to time, a1 and a2 are positive constants.

[0020] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention enables the tracking error to converge within a fixed time, with the convergence time being less dependent on the initial state, thus improving predictability and reliability in engineering applications; This invention can effectively cope with modeling uncertainties, external disturbances, and parameter changes, improving the stable operation capability of variable speed wind turbines under complex wind conditions; Compared with traditional fixed-time control methods that rely on precise mathematical models, this invention can reduce the impact of model uncertainties on control performance, making it more suitable for practical wind power systems; The use of the all-drive system method avoids the problems of recursive construction of virtual control laws and repeated differentiation in traditional backstepping methods, resulting in a simpler controller design with fewer parameters, facilitating engineering implementation; This invention has a faster response speed and better error convergence characteristics, which can reduce power fluctuations caused by wind speed fluctuations or disturbances, improving the dynamic performance of the system. Attached Figure Description

[0021] Figure 1 This is a flowchart of the present invention.

[0022] Figure 2 The effective wind speed under steady-state wind conditions.

[0023] Figure 3 This represents the rotor speed tracking error under steady-state wind speed conditions.

[0024] Figure 4 This represents the effective wind speed under gust conditions.

[0025] Figure 5 This represents the rotor speed tracking error under gust wind conditions.

[0026] Figure 6 This refers to the effective wind speed and its measured value under turbulent wind conditions.

[0027] Figure 7 This represents the rotor speed under turbulent wind conditions.

[0028] Figure 8 This represents the rotor speed tracking error under turbulent wind conditions. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Please see Figure 1 A fixed-time adaptive sliding mode control method for variable-speed wind turbines based on an all-drive system, comprising:

[0031] Step 1: Establish a mathematical model of the transmission system of the variable speed wind turbine based on the mechanical principles of the transmission system, including:

[0032] The transmission system can maintain torque balance and can be simplified to a single-mass model. The mathematical model of the transmission system of the variable-speed wind turbine is expressed as follows:

[0033] (1)

[0034] Among them, J t K is the inertial constant. t Ω is the damping constant. r It is the rotor speed. It is Ω r The time derivative of T. a and T g These represent aerodynamic torque and generator torque, respectively. J t K t T g and T a It can be represented as:

[0035] (2)

[0036] Among them, J r It is the rotor's coefficient of inertia, J g It is the generator's inertia coefficient, K. r K is the external damping coefficient of the rotor. g It is the external damping coefficient of the generator, n g and T em P represents the gear ratio and electromagnetic torque of the gearbox, respectively. a It is pneumatic power.

[0037] Considering the uncertainties in system modeling and the measurement noise in wind speed, the mathematical model of the variable speed wind turbine drive system can be reformulated as follows:

[0038] (3)

[0039] Where ρ and R represent air density and rotor radius, respectively. p (λ,β) is the power coefficient, where λ and β represent the tip velocity ratio and blade pitch angle, respectively. v is the wind speed. d is the combined disturbance term, which includes uncertainties in system modeling and measurement noise of wind speed v.

[0040] Hypothesis 1: Combine the disturbance term d and its time derivative Both have an upper bound, that is, for t > 0, and Where d1 > 0 and d2 > 0 are both unknown positive constants. It should be noted that this assumption is a standard setting in well-known literature in the field of nonlinear control and is widely used in sliding mode control and adaptive control design. From a physical perspective, this is reasonable because disturbances and modeling uncertainties in real-world systems are typically bounded and change at a finite rate.

[0041] Generator torque T g The adjustment is made through the actuator of the generator, and is expressed as follows:

[0042] (4)

[0043] Where, τ g It is a time constant. Generator torque T g Time derivative, T g,des It is the generator torque T g The expected value.

[0044] Step 2: Construct the state-space equations of the variable-speed wind turbine system based on the mathematical model of the transmission system and the torque model of the generator, including:

[0045] The output power that the generator can provide It can be represented as:

[0046] (5)

[0047] Where μ is the efficiency of the generator.

[0048] To maximize power generation, the rotor speed Ω needs to be increased. r Track its expected value Ω d ,in , It is the optimal tip speed ratio, and the system state is x1=Ω r -Ω d Let the tracking error be denoted as x2=T. g Fixed-time adaptive sliding mode controller u=T g,des Then we can get:

[0049] (6)

[0050] in, , , x1, x2, and Ω are respectively d Time derivative, For simplicity, f(x1,Ω) d ,v,λ) is abbreviated as f.

[0051] Step 3: Establish a full-drive system model of the variable-speed wind turbine based on the state-space equations of the variable-speed wind turbine system, including:

[0052] Using the first formula in equation (6), the value of x2 can be calculated as follows:

[0053] (7)

[0054] Time derivative of x2 It can be calculated as follows:

[0055] (8)

[0056] In the formula, , , They are respectively , The time derivative of f, the variable speed wind turbine all-drive system model can be expressed as:

[0057] (9)

[0058] in, Represented as the lumped unknowns in the variable speed wind turbine all-drive system model, constant For the sake of simplicity, It is abbreviated as △f.

[0059] Step 4: Establish the sliding mode function and its time derivative function based on sliding mode control theory, including:

[0060] To achieve fixed-time convergence, and based on sliding mode control theory, the sliding mode function s can be designed as follows:

[0061] (10)

[0062] Among them, the function satisfy:

[0063] (11)

[0064] Where k1 > 0 and k2 > 0 are both positive constants. The positive constant γ1 needs to satisfy γ1 > 1, and the positive constant γ2 needs to satisfy 0 < γ2 < 1. and It is a sign power function.

[0065] time derivative of s It can be represented as:

[0066] (12)

[0067] Among them, the function , It is a function The derivative with respect to time.

[0068] Step 5: Based on the approximation of the lumped unknowns in the variable-speed wind turbine all-drive system model using neural network technology, including:

[0069] In the control of variable-speed wind turbine generators, since the variables in the lumped unknown Δf are physically restricted, Δf can be approximated using neural network technology in the following form:

[0070] (13)

[0071] in, This is the optimal weight vector, and the superscript T indicates transpose. For the input vector, and This is represented as the corresponding neural network approximation error. The activation function of a neural network is expressed in the form of:

[0072] (14)

[0073] Where m1, m2, m3, and m4 are appropriate positive constants, and exp(·) is the natural exponential function. For ease of representation, we will use... Abbreviated as .

[0074] Hypothesis 2: and There are upper limits, that is , W m and It is an unknown normal quantity.

[0075] Step 6: Design a fixed-time adaptive sliding mode controller based on fixed-time theory and adaptive sliding mode control theory, including:

[0076] By leveraging neural network technology and based on fixed-time theory and adaptive sliding mode control theory, a neural network sliding mode control law with fixed-time adaptability can be constructed:

[0077] (15)

[0078] Where κ, ε, A, c1, and c2 are all positive positive integers greater than 0. The positive integer γ3 needs to satisfy... For a positive constant γ4, 0 < γ4 < 1. -1 This represents the inverse operation of the constant g. It is the hyperbolic tangent function. and It is a sign power function.

[0079] also, The estimated value is expressed as , The adaptive update rule is designed as follows:

[0080] (16)

[0081] in, yes The derivative with respect to time. a1 and a2 are positive constants.

[0082] Note 1: In equation (15), a continuous robust term is used. To reduce the approximation error of the neural network. The other three items... It is designed to achieve fixed-time stability.

[0083] Step 7: Control the rotor speed of the variable speed wind turbine using a fixed-time adaptive sliding mode controller.

[0084] The present invention also provides a verification method for the fixed-time adaptive sliding mode control method for variable-speed wind turbines as described above, comprising:

[0085] For ease of verification, the following lemma is given:

[0086] Lemma 1: Consider the following system:

[0087] (17)

[0088] Where z is the system state. It is the derivative of the system state with respect to time. Signed power function. and They can be defined as follows: ,as well as , where sgn(·) is the sign function. Furthermore, α1 and α2 are both positive constants. In this case, the equilibrium state of system (17) is fixed-time stable, and the stable time T is restricted to the following range:

[0089] (18)

[0090] Among them, T max It is the maximum value of the settling time T.

[0091] Lemma 2: For a given system If there exists a continuous function V(x) and positive constants... , making the inequality If this holds true, then the trajectory of the system is practically fixed-time stable (PFTS), and the residual set of the solution becomes... Where the positive constant ρ1 satisfies 0 < ρ1 < 1, min{·} is the minimum function, and the convergence time required to reach this residual set satisfies .

[0092] Lemma 3: For a small custom positive constant ε > 0, the hyperbolic tangent function tanh(·) has the following properties:

[0093] (19)

[0094] in, It belongs to real numbers, and .

[0095] The following theorem is designed to verify the fixed-time adaptive sliding mode control method for variable-speed wind turbines:

[0096] Theorem 1: Considering the full-drive model (9) of the variable speed wind turbine system with assumption 1, fixed-time sliding mode variables (10) and (11), fixed-time adaptive neural network sliding mode control law (15) and adaptive update law (16), it is concluded that the rotor speed tracking error x1 will converge to a certain region within a fixed time.

[0097] Proof: Select a candidate Lyapunov function V1 as follows:

[0098] (20)

[0099] in, This is expressed as an approximation error;

[0100] By differentiating V1, we can obtain:

[0101] (twenty one)

[0102] in, The time derivative of V1;

[0103] Then, combining Hypothesis 2, Lemma 3, and the adaptive update rate (16). Simplified to:

[0104] (twenty two)

[0105] According to Lemma 3, we get Based on Young's inequality and Assumption 2, we obtain , ,as well as ,Then, Represented as:

[0106] (twenty three)

[0107] Where, constant Therefore, when condition A > 1 is true, Rewritten as:

[0108] (twenty four)

[0109] Where, constant And constant In addition, there are Therefore, s and Both are bounded;

[0110] also, It can be recalculated as follows:

[0111] (25)

[0112] The boundary of the last term in equation (25) is:

[0113] (26)

[0114] Then, according to equation (26). It can be recalculated as:

[0115] (27)

[0116] Note that the following formula holds true:

[0117] (28)

[0118] Where, constant and constant Furthermore, it can be assumed that... , yes The maximum value. Then, It can be represented as:

[0119] (29)

[0120] Where, constant ,constant and constants .

[0121] According to Lemma 2, s and It can converge to the corresponding set of two regions. and middle:

[0122] (30)

[0123] Where, constant ,constant and constant Need to meet Furthermore, the fixed convergence time T1 can be calculated using the following formula:

[0124] (31)

[0125] When the fixed-time sliding mode variable s approaches At that time, we can conclude that:

[0126] (32)

[0127] The above formula can be equivalently written as:

[0128] (33)

[0129] or

[0130] (34)

[0131] Therefore, as long as k1 and k2 are chosen such that and , (33) and (34) are equivalent to the fixed-time sliding surface (10). Therefore, the tracking error x1 will converge to the following region within a fixed time:

[0132] (35)

[0133] or

[0134] (36)

[0135] From (35) and (36), we can see that the convergence region of the tracking error x1 is:

[0136] (37)

[0137] The Lyapunov function is defined in the following form:

[0138] (38)

[0139] Differentiating equation (38) yields:

[0140] (39)

[0141] Where, constant ,constant ,constant ,constant and constants According to Lemma 2 and (39), it can be concluded that the tracking error x1 will converge to the region within a fixed time T2. ,Right now:

[0142] (40)

[0143] in, .

[0144] Therefore, the tracking error x1 will settle in the following total time T. total Converging to the region ,Right now:

[0145] (41)

[0146] To verify the effectiveness of the proposed control method, simulations were performed on NREL's 5-MW variable-speed wind turbine using the OpenFAST platform developed by the National Renewable Energy Laboratory (NREL). Specifically, simulation comparisons were presented for three operating scenarios (steady-state wind speed, gusts, and turbulent wind) using the proposed fixed-time adaptive neural network sliding mode control method (FTANSMC), the adaptive sliding mode controller (ASMC), and the baseline controller with integral control (BLC+IC). Figure 1 This is a flowchart of the present invention.

[0147] Under steady-state wind speed conditions, the initial rotor speed was set to 1.27 rad / s. The steady-state wind speed curve used is as follows: Figure 2 As shown, the effective wind speed is 8 m / s. Figure 3 This demonstrates a comparison of rotor speed tracking errors among three different controllers. From... Figure 3It can be seen that all three control methods can effectively make the rotor speed track its reference value. Figure 3 The small figure in the diagram is a magnified view of the rotor speed tracking error within the time range of 0–50s, used to more clearly illustrate the error convergence process and the differences in convergence speed between different control methods in the initial stage. However, the control method proposed in this invention achieves a faster tracking error convergence speed. Table I shows the tracking error convergence time under different initial rotor speeds when using FTANSMC. Table I demonstrates that the control method proposed in this invention makes the tracking error convergence time unaffected by initial conditions.

[0148] Table I. Convergence Time of Rotor Speed ​​Tracking Error (Steady-State Wind Condition)

[0149] Gust data such as Figure 4 As shown, in this scenario, the initial rotor speed is set to 1.27 rad / s. A comparison of rotor speed tracking errors under the three controllers is presented. Figure 5 As shown. Figure 5 The smaller image is a magnified view of the rotor speed tracking error within a time range of 100s to 200s, used to more clearly illustrate the error fluctuations and steady-state tracking performance differences of different control methods during gusts. Figure 5 It can be seen that under FTANSMC control, the rotor speed tracking error exhibits better convergence performance. This indicates that the proposed control algorithm has stronger robustness under extreme wind conditions. Table II shows the convergence time of the tracking error under different initial rotor speeds when using FTANSMC. As can be seen from Table II, the control method proposed in this patent ensures that the convergence time of the tracking error is independent of the initial conditions.

[0150] Table II Convergence Time of Rotor Speed ​​Tracking Error (Gust Condition)

[0151] The turbulent wind speed data used in this patent was generated by the TurbSim platform developed by the U.S. National Renewable Energy Laboratory. Actual effective wind speed v o and its measured value v, such as Figure 6 As shown, the average wind speed at the hub height is 8 m / s, the turbulence intensity is 10%, and the effective wind speed measurement noise Δv follows a uniform distribution within the interval [-0.2, 0.2]. In this scenario, the initial value of the rotor speed is set to 1.27 rad / s. When using FTANSMC control, the rotor speed Ω... r and its expected value Ω d like Figure 7 As shown. The rotor speed tracking error under the three controllers is as follows. Figure 8 As shown. From Figure 8 It can be seen that although turbulent wind speed is random and intermittent, and the measurement of effective wind speed is noisy, the proposed method can still enable the rotor speed tracking error to converge quickly to a region closer to zero. Table III shows the convergence time of the tracking error under different initial rotor speeds when using FTANSMC. It is clear from Table III that the control method proposed in this patent ensures that the convergence time of the tracking error is independent of the initial conditions.

[0152] Table III Convergence Time of Rotor Speed ​​Tracking Error (Turbulent Wind Condition)

[0153] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system, characterized in that, include: Step 1: Establish a mathematical model of the transmission system of the variable speed wind turbine based on the mechanical principles of the transmission system; Step 2: Construct the state-space equations of the variable-speed wind turbine system based on the mathematical model of the transmission system and the torque model of the generator; Step 3: Establish a model of the variable speed wind turbine all-drive system based on the state-space equations of the variable speed wind turbine system; Step 4: Establish the sliding mode function and its time derivative function based on sliding mode control theory; Step 5: Approximate the lumped unknowns in the variable speed wind turbine all-drive system model using neural network technology; Step 6: Design a fixed-time adaptive sliding mode controller based on fixed-time theory and adaptive sliding mode control theory; Step 7: Control the rotor speed of the variable speed wind turbine using a fixed-time adaptive sliding mode controller.

2. The fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 1, characterized in that, The mathematical model of the variable-speed wind turbine transmission system established in step 1 is expressed as follows: ， Among them, Ω r It is the rotor speed. It is Ω r The derivative with respect to time, J t It is the inertial constant, K t It is the damping constant, ρ and R represent the air density and rotor radius, respectively, and C p (λ,β) is the power coefficient, where λ and β represent the tip velocity ratio and blade pitch angle, respectively, v is the wind speed, and T is the propeller angle. g d is the generator torque, and d is the comprehensive disturbance term, which includes the uncertainty in system modeling and the measurement noise of wind speed v.

3. The fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 2, characterized in that, In step 1, hypothesis 1 is constructed: the comprehensive interference term d and its time derivative. Both have an upper bound.

4. The fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 2, characterized in that, The generator torque T g The adjustment is made through the actuator of the generator, and is expressed as follows: ， Where, τ g It is a time constant. Generator torque T g Time derivative, T g,des It is the generator torque T g The expected value.

5. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 4, characterized in that, Step 2 includes: Make the rotor speed Ω r Track its expected value Ω d Set the system state x1=Ω r -Ω d Let the tracking error be denoted as x2=T. g Fixed-time adaptive sliding mode controller u=T g,des The state-space equations of the variable-speed wind turbine system are then obtained as follows: ， in, , , x1, x2, and Ω are respectively d The time derivative of the function f(x1,Ω) d ,v,λ) is abbreviated as f.

6. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 5, characterized in that, In step 3, the established model of the variable speed wind turbine all-drive system is as follows: ， in, , , They are respectively , The time derivative of f The lumped unknown term in the all-drive system model of a variable-speed wind turbine is denoted as Δf, a constant. .

7. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 6, characterized in that, Step 4 includes: The sliding mode function s and its time derivative function Designed as follows: ， ， Where k1 and k2 are both positive constants, the positive constant γ1 satisfies γ1>1, and the positive constant γ2 satisfies 0<γ2<1. and It is a sign power function, a function , It is a function The derivative with respect to time.

8. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 7, characterized in that, Step 5 includes: The lumped unknown term Δf is approximated by a neural network in the following form: ， in, This is the optimal weight vector, with the superscript T indicating transpose. For the input vector, This is represented as the approximation error of the neural network. Here is the activation function of the neural network, abbreviated as: .

9. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 8, characterized in that, In step 5, hypothesis 2 is constructed: and There are limits to everything.

10. A fixed-time adaptive sliding mode control method for a variable-speed wind turbine based on an all-drive system according to claim 9, characterized in that, In step 6, the designed fixed-time adaptive sliding mode controller u is: ， Where κ, ε, A, c1, and c2 are all positive constants greater than 0, positive constant γ3 > 1, and positive constant γ4 satisfies 0 < γ4 < 1, g -1 This represents the inverse operation of the constant g. It is the hyperbolic tangent function. and It is a sign power function. yes The estimated value, The adaptive update rate is designed as follows: ， in, yes The derivatives with respect to time, a1 and a2 are positive constants.

Images