Data-based non-gaussian wind energy conversion system constant power control method, device and medium

Abstract

The application discloses a constant power control method and device of a wind energy conversion system based on non-Gaussian data and a medium, establishes a performance index based on second-order information potential, and obtains optimal control input by optimizing the established performance index through a Newton method, so that the output speed and output power of the wind energy conversion system under non-Gaussian wind speed interference can track the set value, constant power control of the wind energy conversion system is realized, the fluctuation of the output power is reduced, and the service life of the wind turbine is prolonged. The application performs controller design under the condition that the interference is a non-Gaussian random variable, compared with the existing method in which noise is ignored or noise is assumed to be subject to Gaussian distribution, the application is more general and practical. Moreover, the application considers the mutual relationship between two closed loop circuits in a multivariable system, adds joint information potential in the performance index, allows variable speed during a transient period, and can obtain higher control performance.

Description

TECHNICAL FIELD

[0001] The application relates to a data-based constant-power control method for a non-Gaussian wind energy conversion system, a device and a medium, and belongs to the technical field of wind power generation system control. BACKGROUND

[0002] The main purpose of the operation of a wind energy conversion system is to convert wind energy into electric energy. Effective use of wind energy is of great significance to improving new energy utilization efficiency, reducing consumption of fossil fuels and improving the global greenhouse effect. In the whole process of operation of the wind energy conversion system, wind in nature is not only one of the main energy sources, but also the main external disturbance to the system. Wind speed is random and can be affected by factors such as seasonal changes, day and night alternation and terrain. The continuous change of wind speed will cause the output power and rotating speed of the wind energy conversion system to change with the change of wind speed, and at the same time, the service life of the wind turbine will also be seriously affected. Therefore, it is particularly important in the industry to design an effective control method to improve the conversion efficiency of the wind power generation system and ensure safe and effective operation of the whole system.

[0003] In the wind energy conversion system, wind speed is a non-Gaussian random variable, which will cause the output power of the wind energy conversion system to fluctuate randomly, so the output rotating speed and power are also non-Gaussian random variables. If the non-Gaussian nature of wind speed is not strictly considered, the control scheme will have great limitations.

[0004] At present, the control methods of the wind energy conversion system mainly include proportional-integral-derivative (PID) control, linear quadratic Gaussian (LQG) control, fuzzy logic control, sliding mode control and neural network control and the like. The above methods do not consider the non-Gaussian case of external disturbance, so the structure obtained by the above control methods is not satisfactory.

[0005] Therefore, it is necessary for those skilled in the art to propose a power control method for the wind energy conversion system under non-Gaussian disturbance. SUMMARY

[0006] Objective: In order to overcome the deficiencies in the prior art, the application provides a data-based constant-power control method for a non-Gaussian wind energy conversion system, a device and a medium, which can significantly improve the control performance of the wind energy conversion system containing non-Gaussian random disturbance.

[0007] Technical scheme: In order to solve the above technical problems, the technical scheme adopted by the application is:

[0008] In a first aspect, a data-based constant-power control method for a non-Gaussian wind energy conversion system comprises the following steps:

[0009] Obtain the computational model of the discretized non-Gaussian wind energy conversion system.

[0010] The tracking error is obtained based on the output of the calculation model and the bounded set value.

[0011] Calculate the corresponding entropy and mean of the tracking error. Based on the corresponding entropy, mean of the tracking error, and the control input vector, construct the performance index function J(u) of the non-Gaussian wind energy conversion system at time k. k ).

[0012] Constructing optimal control input The optimization problem is arg min J(u k ).

[0013] Set the precision ε and solve for the optimal control input u. k * The optimization problem arg minJ(u k ), when ▽J(u l If ε < ε holds, then the optimal control input is obtained. u l Let ▽J(u) be the input variable of the non-Gaussian wind energy conversion system at time l. l The performance index function J(u) at time l l The gradient of ).

[0014] Furthermore, it also includes:

[0015] When ▽J(u l Let P ≥ ε l =-▽J(u l ), According to the optimal step size λ l and the Hessian matrix H(u l ), update u l+1 =u l +λ l P l .

[0016] Calculate ▽J(u) l+1 ), when ▽J(u l+1 If ε < ε holds, then the optimal control input is obtained. u l Let ▽J(u) be the input variable for the non-Gaussian wind energy conversion system at time l+1. l+1 ) is the performance index function J(u) at time l+1. l+1 The gradient of ).

[0017] Furthermore, the calculation model of the discretized non-Gaussian wind energy conversion system is as follows:

[0018]

[0019] where x k , u k , y k are state vector, control input vector and control output vector at time k, T s is the sampling time, f(·), p(·) represent the nonlinear functions of the non-Gaussian wind energy conversion system dynamics, v k is a non-Gaussian bounded random variable with known PDFs, x k+1 is the state vector at time k+1.

[0020] Further, the tracking error e 1,k is calculated as follows:

[0021] e 1,k = r 1,k - y 1,k

[0022] e 2,k = r 2,k - y 2,k

[0023] where r 1,k , r 2,k are the bounded setpoints at time k, y 1,k , y 2,k are the control output vectors at time k, e 1,k , e 2,k are the tracking errors at time k.

[0024] Further, the state variable x k = [x 1,k , x 2,k , x 3,k , x 4,k , x 5,k ] T = [ω t , ω g , T, β, T g ] T , the control input variable u , the control output variable y k = [y 1,k , y 2,k ] T = [ω g , P g ] T , the bounded setpoint r 1,k = ω g,ref , r 2,k = P g,ref .

[0025] where ωtis the wind turbine rotational speed, ω gis the wind turbine speed, P g is the generator output power. T is a new state variable of the internal torque, β is the pitch angle, T g is the actual value of the generator electromagnetic torque. β * is the reference value of the blade pitch angle, is the reference value of the generator electromagnetic torque. P g is the generator output power. ω g,ref and P g,ref are the rated values of the generator output speed and power, respectively.

[0026] Further, J(u k ) is calculated as follows:

[0027]

[0028] where H 1k , H 2k are the second-order Renyi entropies of the generator output speed tracking error e 1,k and the generator output power tracking error e 2,k , H 12k denotes the joint entropy of the generator output speed tracking error e 1,k and the generator output power tracking error e 2,k , E 1k , E 2k are the means of the generator output speed tracking error e 1,k and the generator output power tracking error e 2,k , u k is the control input vector at time k. R1, R2 are the weights corresponding to the second-order Renyi entropies, R3 is the weight of the joint entropy, R4, R5 are the weights corresponding to the means, and R6 is the weight of the control input vector.

[0029] Further, the calculation formulas of H 1k , H 2k , H 12k , E 1k , E 2k are as follows:

[0030]

[0031]

[0032]

[0033] where V ik (i = 1, 2) is the information potential of the tracking error e i,k (i = 1, 2), and V 12k is e 12,kthe joint information potential of γ e (·) are the PDFs of tracking error. b i , a i are the upper and lower limits of integration, respectively, γ ei,k is the tracking error e i,k are the probability density functions of tracking error e e12,k is the tracking error e i,k are the joint probability density functions of tracking error e i (i = 1, 2) represent the variables E ik (i = 1, 2) are the means of tracking error.

[0034] Further, the J(u k ) is optimized as follows:

[0035]

[0036] wherein,

[0037] Further, V 1k , V 2k , V 12k are calculated as follows:

[0038]

[0039]

[0040] wherein, G σ (x) is a Gaussian kernel function, and N is the width of the sliding window. e 1,j , e 1,m represent the generator output speed tracking errors corresponding to the jth and mth sliding windows, respectively. 2,j , e 2,m represent the generator output power tracking errors corresponding to the jth and mth sliding windows, respectively.

[0041] Further, R1 = 0.1, R2 = 0.1, R3 = 10, R4 = 0.001, R5 = 0.001,

[0042]

[0043] Further, the Hessian matrix is:

[0044]

[0045] wherein, r represents the order of the matrix, u l,i and u l,j are the ith and jth inputs at time l, respectively, and the values of i and j depend on the order r of the matrix.l,1 , u l,r respectively, the first, r input at time l.

[0046] In a second aspect, a computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the constant power control method of a data-based non-Gaussian wind energy conversion system according to any one of the first aspect.

[0047] In a third aspect, a computer device comprises:

[0048] a memory for storing instructions.

[0049] a processor for executing the instructions, causing the computer device to perform the constant power control method of a data-based non-Gaussian wind energy conversion system according to any one of the first aspect.

[0050] Beneficial effects: The constant power control method, device and medium of a data-based non-Gaussian wind energy conversion system provided by the present application have the following beneficial effects. Since the wind energy conversion system is affected by random noise brought by wind in nature during actual operation, and the noise is generally non-Gaussian, the present application designs a controller under the condition that the disturbance is a non-Gaussian random variable, which is more general and practical compared with the existing method which ignores noise or assumes that noise is subject to Gaussian distribution. Moreover, the present application considers the mutual relationship between two closed-loop circuits in a multivariable system, and adds joint information potential in the performance index. Compared with the past design of independent controllers for decentralized control, the controller designed by the present application can allow variable speed during transient period, and can obtain higher control performance. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 is a wind energy conversion system structure diagram of an embodiment of the present application.

[0052] Figure 2 is a constant power control block diagram of a data-based wind energy conversion system of an embodiment of the present application.

[0053] Figure 3 is a constant power control flowchart of a data-based wind energy conversion system of an embodiment of the present application.

[0054] Figure 4 is a combined wind speed curve diagram of an embodiment of the present application.

[0055] Figure 5 is a system output speed response curve diagram of an embodiment of the present application.

[0056] Figure 6 is a system output power response curve diagram of an embodiment of the present application.

[0057] Figure 7 This is a system control input curve diagram for a specific embodiment of the present invention.

[0058] Figure 8 The graph shows the performance indicators of a specific embodiment of the present invention.

[0059] Figure 9 This is a 3D probability density curve of the system rotation speed tracking error in a specific embodiment of the present invention.

[0060] Figure 10 This is a 3D probability density curve of the system power tracking error in a specific embodiment of the present invention. Detailed Implementation

[0061] The technical solutions 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 protection scope of the present invention.

[0062] The present invention will be further described below with reference to specific embodiments.

[0063] Wind energy conversion systems are typically modeled as nonlinear systems. Therefore, from a mechanistic analysis perspective, a wind energy conversion system can be viewed as a nonlinear dynamic system with non-Gaussian random external disturbances and uncertain parameters. Its structure diagram is shown below. Figure 1 As shown, it consists of a wind turbine, a transmission system, a permanent magnet synchronous generator, an AC-DC-AC converter, a pitch servo system, and a power grid. The main function of the wind energy conversion system is to convert wind kinetic energy into electrical energy.

[0064] The first embodiment is a constant power control method for a data-based non-Gaussian wind energy conversion system. The present invention takes the input and output of the wind energy conversion system when it is working in the full load area as an example.

[0065] In the nonlinear model, the external disturbance is represented by wind, and its speed is a non-Gaussian random variable, i.e., v. k The inputs are reference values ​​for the generator's electromagnetic torque and the blade pitch angle, i.e. and β * The output is the speed and power of the generator in the wind energy conversion system, i.e., ω. g and P g Its control block diagram and flowchart are as follows: Figure 2 and Figure 3 As shown. In Figure 2The error between the output speed and power of the wind energy conversion system and the set speed and power will be calculated, and then the estimated probability density function will be used to form a performance index through non-parametric estimation of the error probability density function, so as to design a minimum entropy controller, and the output of the controller will be used as the input of the system of the WECS (wind energy conversion system) to complete the constant power control of the WECS.

[0066] The method mainly comprises the following steps:

[0067] Step 1. Establishing a nonlinear model of the wind energy conversion system under non-Gaussian interference.

[0068] In the wind energy conversion system, the wind turbine can be described based on the Betz theory, and the aerodynamic power expression is:

[0069]

[0070] wherein, is the aerodynamic power of the wind turbine, ρ is the air density, ρ = 1.25 kg / m 3 , R t is the rotating radius of the wind turbine blade, R t = 33 m, v k is the wind speed, C p (λ, β) is the wind energy utilization coefficient. λ is the tip speed ratio, β is the pitch angle, and π is the circular constant.

[0071] The transmission system adopts a flexible shaft transmission model, and the expression is:

[0072]

[0073] wherein, is the rotational inertia of the wind turbine, J g is the rotational inertia of the generator, J g = 56.29 kg·m 2 , K s is the rigidity coefficient, K s = 3.18 × 10 5 Nm / rad, B s is the damping coefficient, B s = 212.2 kg·m 2 / s, i is the transmission ratio of the gearbox, i = 74.38, and η is the transmission efficiency, η = 1, is the internal torque of the wind turbine, T g is the internal torque of the generator. is the change rate of the internal torque, ω t is the rotational speed of the wind turbine, ω g is the rotational speed of the wind generator.

[0074] The mathematical model of the permanent magnet synchronous generator (PMSG) is:

[0075]

[0076] The actual value of the electromagnetic torque of the PMSG T g is:

[0077]

[0078] where i d and i q are the currents of the d-axis and q-axis, L d and L q are the stator reluctances of the d-axis and q-axis, R s is the stator resistance, φ m is the constant flux due to the permanent magnets, p is the number of pole pairs, ω g is the rotor speed of the wind turbine, u d and u q are the voltages of the d-axis and q-axis. and are the rates of change of the currents of the d-axis and q-axis.

[0079] The dynamics of the electrical subsystem are represented by a time constant τ g and a first order model with unity DC gain, as follows:

[0080]

[0081]

[0082] where T g is the actual value of the electromagnetic torque of the generator, is the reference value of the electromagnetic torque of the generator, τ g is the time constant, τ g = 20 ms, P g is the generator output power. is the rate of change of the electromagnetic torque of the generator.

[0083] The pitch actuator is modeled by a first order equivalent dynamic system with pitch amplitude and derivative saturation, as follows:

[0084]

[0085] where β is the actual value of the blade pitch angle, β * is the reference value of the blade pitch angle, τ is the time constant of the pitch actuator, τ = 0.1 s. is the rate of change of the blade pitch angle.

[0086] The state variable x = [x1, x2, x3, x4, x5] is defined T = [ω t , ω g , T, β, T g ] T The input variable The output variable y = [y1, y2] T = [ω g , P g ] T Combining equations (2), (5), (6), (7), the following nonlinear state-space model of the WECS can be obtained:

[0087]

[0088] where ω t is the wind turbine rotational speed, ω g is the wind generator rotational speed, P g is the generator output power. T is a new state variable representing the internal torque, β is the pitch angle, T g is the actual value of the generator electromagnetic torque. β * is the reference value of the blade pitch angle, is the reference value of the generator electromagnetic torque. P g is the generator output power. is the wind turbine rotational inertia, η is the transmission efficiency, i is the gear box transmission ratio, λ is the tip speed ratio, ρ is the air density, R t is the wind turbine blade rotational radius, v represents the non-Gaussian disturbance, π is the circular constant, C p is the wind energy utilization coefficient. J g is the generator rotational inertia. K s is the stiffness coefficient, B s is the damping coefficient. τ is the time constant of the pitch actuator. τ g is the time constant.

[0089] Further, its general form can be expressed as:

[0090]

[0091] where v represents the non-Gaussian disturbance suffered by the system.

[0092] Equation (9) can be discretized by numerical integration, and the discretized nonlinear system has the following form:

[0093]

[0094] where x k = [x 1,k , x2,k ,x 3,k ,x 4,k ,x 5,k ] T ,u k =[u 1,k ,u 2,k ] T ,y k =[y 1,k ,y 2,k ] T is the discrete-time state vector at time k, the control input vector and the output vector, the sampling time T s = 1 s, f(·), p(·) represent the nonlinear functions of the dynamic characteristics of the wind energy conversion system, v k ∈ R q is a non-Gaussian bounded random variable with known PDFs, R q is a vector of q columns in 1 row. The combined wind speed profile is shown in Figure 4 .

[0095] Step 2. The tracking error e i,k is obtained according to the input and output of the wind energy conversion system, i is a constant of 1 and 2.

[0096] Further, define the bounded setpoint r 1,k = ω g,ref , r 2,k = P g,ref , then the tracking error e 1,k , e 2,k can be expressed as:

[0097] e 1,k = r 1,k - y 1,k (11)

[0098] e 2,k = r 2,k - y 2,k (12)

[0099] where ω g,ref and P g,ref are the rated values of the generator output speed and power, ω g,ref = 220 rad / s and P g,ref = 2 MW.

[0100] Step 3. Performance index establishment based on statistical information.

[0101] The uncertainty of the output speed and power tracking error in the nonlinear state-space model of WECS is characterized by entropy. By minimizing its entropy, the probability density functions (PDFs) of the speed and power tracking error can be made narrower and sharper, which means minimizing the randomness of the tracking error. At the same time, it is also necessary to minimize the mean of the tracking error and the control input constraints. Therefore, in this invention, the performance index function at time k is selected as follows:

[0102]

[0103] Among them, H 1k H 2k Tracking error e 1,k e 2,k The second-order Renyi entropy, H 12k E represents the joint entropy of the two tracking errors. 1k E 2k The tracking error e is respectively 1,k e 2,k The mean of the vectors, corresponding to the entropy, and the weights of the mean and control input vector are R1 = 0.1, R2 = 0.1, R3 = 10, R4 = 0.001, and R5 = 0.001, respectively. The expressions for the second-order Renyi entropy, joint entropy, and mean are as follows:

[0104]

[0105]

[0106]

[0107] Among them, V ik (i = 1, 2) is the tracking error e i,k The information potential of (i = 1, 2), V 12k It is e 12,k The joint information potential, γ e (·) represents PDFs of the tracking error. i a i These are the upper and lower limits of the points, respectively. For tracking error e i,k The probability density function of (i = 1, 2), For tracking error e i,k The joint probability density function of (i = 1, 2), z i (i = 1, 2) represents the variable E within the integral. ik (i = 1, 2) represents the mean of the tracking error.

[0108] Since the second order Renyi entropy is a monotonically decreasing function of the information potential, minimizing the Renyi entropy of tracking error is equivalent to maximizing its information potential. Therefore, equation (13) can be expressed as:

[0109]

[0110] For convenience, let

[0111] J(u k ) = -0.1V 1k -0.1V 2k -10V 12k + 0.001E 1k + 0.001E 2k (18)

[0112] The performance index function is obtained as:

[0113]

[0114] Step 4. Obtain the probability density functions (PDFs) of the system closed-loop tracking error e i,k (i = 1, 2) by the method of non-parametric estimation.

[0115] In the wind energy conversion system, the PDFs of tracking error are estimated from samples by the method of non-parametric estimation. The expression of Parzen window estimation method is:

[0116]

[0117]

[0118] where, is a Gaussian kernel function, whose shape and size are determined by the parameter σ, and N is the width of the sliding window.

[0119] The estimated PDFs are brought into equations (14) and (15) to obtain the non-parametric expression of the second order Renyi entropy:

[0120]

[0121]

[0122] The corresponding information potential is:

[0123]

[0124]

[0125] where, the window width N = 100.

[0126] At this point, the establishment of the performance index function has been completed. The next step is to optimize the performance index function to obtain the optimal control input.

[0127] Step 5. Use Newton's method to optimize the performance index and solve for the optimal control input.

[0128] Solving for optimal control input The optimization problem can be expressed as:

[0129]

[0130] For a given accuracy ε > 0, the approximate solution of the optimal control input of the minimum entropy controller can be expressed as:

[0131] u l+1 =u l +λ l P l (27)

[0132] in, To approximate the optimal step size, P l =-▽J(u l ), u l+1 and u l These are the control input for the next moment and the current control input, respectively.

[0133] The Hessian matrix is:

[0134]

[0135] Its main steps are as follows:

[0136] Step (1): Select initial value u0 = [u 0,1 ,…,u 0,r ] T Set an appropriate precision ε > 0, denoted as l: = 0.

[0137] Step (2): Calculate the gradient of the performance metric at this moment. If ▽J(u) l If ε < ε is true, then stop the calculation and obtain the optimal control input. Otherwise proceed to step (3).

[0138] Step (3): Let P l =-▽J(u l ), According to the optimal step size λ l and the Hessian matrix H(u l ), update u l+1 =u l +λ l Pl ; wherein

[0139]

[0140] Step (4): Calculate ∇J(u l+1 ), denoted as l:=l+1, and then return to step (2) to continue iteration.

[0141] At this point, the solution to the optimal control input is completed, and the variation curve of the optimal control input over time is as shown in Figure 7 , the response curves of the generator speed and power are as shown in Figure 5 and Figure 6 , and the settings of the conventional PI controller are and As can be seen from the figure, under the condition of continuously changing wind speed, the generator speed and the generator output power can effectively track the target speed and power set value by using the control method of the application. Compared with the conventional PI controller, the controller designed in the application produces smaller fluctuations in the generator speed and output power, and the tracking set value time is shorter. In actual operation, the mechanical load of the wind turbine generator can be effectively reduced, and the electric power can be more stable. The performance index function variation curve is as shown in Figure 8 , from which it can be seen that the performance index of the system is continuously decreasing over time, and eventually tends to be stable. The three-dimensional PDFs of the generator speed and power error are as shown in

[0142] The second embodiment is a computer readable storage medium, which stores a computer program. When the computer program is executed by a processor, the constant power control method of the data-based non-Gaussian wind energy conversion system is realized as any one of the first embodiment.

[0143] The third embodiment is a computer device, which comprises:

[0144] A memory for storing instructions.

[0145] A processor for executing the instructions, so that the computer device executes the constant power control method of the data-based non-Gaussian wind energy conversion system as any one of the first embodiment.

[0146] Embodiment:

[0147] The application uses the second-order Renyi entropy to characterize the uncertainty of the wind energy conversion system, establishes a performance index based on the second-order information potential, and obtains the optimal control input by optimizing the established performance index through the Newton method, so that the output speed and output power of the wind energy conversion system under the non-Gaussian wind speed interference can track the set value, thereby realizing the constant power control of the wind energy conversion system, reducing the fluctuation of the output power, and prolonging the service life of the wind turbine.

[0148] As Figure 9 , Figure 10 shown, the PDFs of the generator speed error and power error become narrower and sharper over time, indicating that the uncertainty of the two outputs in the WECS is decreasing. It can be found that the control strategy designed in the present application can effectively reduce the influence of randomness on the WECS.

[0149] The above only describes the preferred embodiments of the present application, and it should be noted that those of ordinary skill in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements should also be considered within the protection scope of the present application.

Claims

1. A constant power control method for a data-based non-Gaussian wind energy conversion system, characterized in that: Includes the following steps: Obtain the computational model of the discretized non-Gaussian wind energy conversion system; The tracking error is obtained based on the output of the calculation model and the bounded set value. Calculate the corresponding entropy and mean of the tracking error. Based on the corresponding entropy, mean of the tracking error, and the control input vector, construct the performance index function J(u) of the non-Gaussian wind energy conversion system at time k. k ); Constructing optimal control input The optimization problem argminJ(u k ); Set the precision ε and solve for the optimal control input. The optimization problem argminJ(u k ),when This establishes the optimal control input. u l Let l be the input variable for the non-Gaussian wind energy conversion system. ▽J(u l The performance index function J(u) at time l l The gradient of ). The calculation model for the discretized non-Gaussian wind energy conversion system is as follows: Where, x k u k ,y k Let T be the state vector, control input vector, and control output vector at time k, respectively. s Let f(·) and p(·) represent the sampling time, respectively, and let v be a nonlinear function representing the dynamic characteristics of the non-Gaussian wind energy conversion system. k Let x be a non-Gaussian bounded random variable of known PDFs. k+1 Let be the state vector at time k+1, where PDFs are the probability density functions.

2. The constant power control method for a data-based non-Gaussian wind energy conversion system according to claim 1, characterized in that: Also includes: when make According to the optimal step size λ l and the Hessian matrix H(u l ), update u l+1 =u l +λ l P l ; calculate when This establishes the optimal control input. u l For the non-Gaussian wind energy conversion system at time l+1, The performance index function J(u) at time l+1 l+1 The gradient of ).

3. The constant power control method for a data-based non-Gaussian wind energy conversion system according to claim 1, characterized in that: The tracking error is calculated using the following formula: e 1,k =r 1,k -y 1,k e 2,k =r 2,k -y 2,k Where r1,k and r2,k are the bounded setpoints at time k, respectively, and y 1,k y 2,k These are the control output vectors at time k, e 1,k e 2,k denoted as k, respectively, representing the tracking error at time k.

4. The constant power control method for a data-based non-Gaussian wind energy conversion system according to claim 3, characterized in that: J(u k The calculation formula is as follows: Among them, H 1k For generator output speed tracking error e 1,k The second-order Renyi entropy, H 2k For generator output power tracking error e 2,k The second-order Renyi entropy, H 12k Indicates the generator output speed tracking error e 1,k Generator output power tracking error e 2,k The joint entropy, E 1k For generator output speed tracking error e 1,k The mean, E 2k For generator output power tracking error e 2,k The mean, u k R1 is the control input vector at time k; R2 and R3 are the weights of the corresponding second-order Renyi entropy, R4 and R5 are the weights of the corresponding mean, and R6 is the weight of the control input vector.

5. The constant power control method for a data-based non-Gaussian wind energy conversion system according to claim 4, characterized in that: H 1k H 2k H 12k E 1k E 2k The calculation formula is as follows: Among them, V ik (i = 1, 2) is the tracking error e i,k The information potential of (i = 1, 2), V 12k It is e 12,k The joint information potential, γ e (·) represents PDFs of the tracking error; b i a i These are the upper and lower limits of integration, γ. ei,k For tracking error e i,k The probability density function of (i = 1, 2), γ e12,k For tracking error e i,k The joint probability density function of (i = 1, 2), z i E represents the variable in the integral. ik Let be the mean of the tracking error (i = 1, 2).

6. The constant power control method for a data-based non-Gaussian wind energy conversion system according to claim 5, characterized in that: The J(u) k The expression is optimized as follows: in, 7. The constant power control method for a data-based non-Gaussian wind energy conversion system according to claim 6, characterized in that: V 1k V 2k V 12k The calculation formula is as follows: Among them, G σ (x) is the Gaussian kernel function, and N is the width of the sliding window; e 1,j e 1,m e represents the generator output speed tracking error corresponding to the j-th and m-th sliding windows, respectively. 2,j e 2,m These represent the generator output power tracking errors corresponding to the j-th and m-th sliding windows, respectively.

8. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, implements a constant power control method for a data-based non-Gaussian wind energy conversion system as described in any one of claims 1-7.

9. A computer device, characterized in that: include: Memory, used to store instructions; A processor for executing the instructions, causing the computer device to perform a constant power control method for a data-based non-Gaussian wind energy conversion system as described in any one of claims 1-7.

Images