阅读说明：本技术 基于扰动补偿的永磁直驱风力发电系统有限集模型预测电流控制方法 (Disturbance compensation-based finite set model prediction current control method for permanent magnet direct-drive wind power generation system ) 是由 王军晓 刘义宾 杨海 胡开林 徐建明 俞立 于 2021-06-07 设计创作，主要内容包括：一种基于扰动补偿的永磁直驱风力发电系统有限集模型预测电流控制方法。建立机侧、网侧的数学模型,对于内环的电流状态方程进行离散化处理；对电流、电压、速度进行采样以及坐标变换；设计内环模型预测电流控制器；根据所构建的成本函数选择出使得成本函数值最小的最优电压矢量；将扰动扩张为新的状态,引入新的状态变量,构建出新的状态空间模型；设计扩张状态观测器；设计控制器,在控制器的输出端对系统的扰动进行补偿,从而确定最终的控制律,机侧控制框图如图3所示,网侧控制框图类似。本发明内环选用的有限集模型预测控制方法有效地改善了系统的动态性能,外环引入的扩张状态观测器能够实时估计出扰动,并在控制器端进行补偿,有效地增强了系统的抗扰性能。(A disturbance compensation-based method for predicting current of a finite set model of a permanent magnet direct-drive wind power generation system. Establishing a mathematical model of a machine side and a network side, and carrying out discretization processing on a current state equation of an inner ring; sampling and coordinate transformation are carried out on current, voltage and speed; designing an inner loop model prediction current controller; selecting an optimal voltage vector which enables the cost function value to be minimum according to the constructed cost function; expanding the disturbance into a new state, introducing a new state variable, and constructing a new state space model; designing an extended state observer; designing a controller, and compensating the system disturbance at the output end of the controller to determine a final control law, wherein a machine side control block diagram is shown in fig. 3, and a network side control block diagram is similar. The method for predicting and controlling the finite set model selected by the inner ring effectively improves the dynamic performance of the system, and the extended state observer introduced by the outer ring can estimate the disturbance in real time and compensate at the controller end, thereby effectively enhancing the anti-disturbance performance of the system.)

1. A disturbance compensation-based method for predicting current of a finite set model of a permanent magnet direct-drive wind power generation system is characterized by comprising the following steps:

step 1, in order to realize maximum wind energy capture, maximum power point tracking is adopted to determine a given rotating speed value omega of a machine side speed ring^*；

Step 2, establishing a machine side permanent magnet synchronous motor mathematical model;

step 3, sampling and coordinate transformation of the machine side current and speed, and the process is as follows:

real-time acquisition machine side converter three-phase current i_a,i_b,i_cThe current value i in the two-phase stationary coordinate system (. alpha.beta.) can be obtained by Clark conversion_α,i_βOn the basis, Park conversion is carried out to obtain a current value i under a two-phase rotating coordinate system (dq)_d,i_q；

Step 4, determining a machine side current inner loop prediction model, wherein the process is as follows:

because the machine side uses a two-level converter, each bridge arm has two switch states, 8 switch states can be obtained, corresponding to 8 output voltage vectors u_i(i ═ 0.. 7.), output voltage vector u corresponding to the upper arm of the three arms of the two-level converter when all the arms are closed₇Output voltage vector u corresponding to when lower bridge arm is fully closed₀The same, so when the model prediction calculation is carried out, one is selected, but u is the final voltage vector selected₀And u₇The selection of the switch is in accordance with the principle of minimum switching loss;

wherein u is_d,u_qRepresenting two phasesSynchronously rotating the stator voltage component under the d-q coordinate system; i.e. i_d,i_qRepresenting the stator current component under a two-phase synchronous rotation d-q coordinate system; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_reRepresents an electrical angular velocity;

discrete post-processing is carried out on a current state equation by utilizing a forward Euler formula:

wherein i_d(k),i_q(k) Representing the stator current component under the two-phase synchronous rotation d-q coordinate system at the current moment; i.e. i_d(k+1)，i_q(k +1) is the stator current d, q-axis component at the next moment; u. of_d,u_qD, q axis voltage components corresponding to 8 switch states; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k) Represents the electrical angular velocity at the present moment; t is_sIs a sampling period;

due to the existence of control delay in the system, the predicted value of the motor state calculated according to the prediction model is inaccurate, and the selected voltage vector cannot guarantee the optimum, so that the control delay needs to be compensated:

wherein i_d(k),i_q(k) Representing the stator current component under the two-phase synchronous rotation d-q coordinate system at the current moment; i.e. i_d(k+1)，i_q(k +1) is a component of the d and q axes of the stator current after time delay compensation; u. of_d,u_qThe component of the optimal voltage vector acting on the system at the last moment under a d-q rotating coordinate system; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained;R_srepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k) Represents the electrical angular velocity at the present moment; t is_sIs a sampling period;

and predicting the current state at the next moment on the basis of the time delay compensation:

wherein i_d(k+1)，i_q(k +1) is a component of the d and q axes of the stator current after time delay compensation;

i_d(k+2)，i_q(k +2) are predicted values of the stator current of the q axis at the moment d and k +1 respectively; u. of_di(0,1,…7),u_qi(0,1,…7)D, q axis voltage components corresponding to 8 switch states; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k +1) is approximately equal to ω_re(k)；T_sIs a sampling period;

step 5, constructing a cost function;

in order to better show the tracking effect of the predicted value on the expected value, a cost function, namely a cost function J, needs to be constructed₁The form is as follows:

wherein the content of the first and second substances,a reference value representing a stator current d, q-axis component; i.e. i_d(k+2),i_q(k +2) are predicted values of the stator current of the q axis at the moment d and k +1 respectively;

step 6, selecting an optimal voltage vector;

sequentially calculating predicted value outputs under different voltage vectors, and bringing the predicted value outputs into the constructed cost functionSelecting an optimum voltage vector u that minimizes the cost function value_{opt_1}：

u_{opt_1}＝arg min J₁

Step 7, introducing a state variable d_ωlBuilding a new state space model

In the actual speed regulation process of the motor, the outer ring of the rotating speed is easily influenced by uncertain model parameters and external disturbance, and a mechanical motion equation can be described as

Wherein J represents the rotor moment of inertia; t is_mRepresenting a drive torque; b represents a friction coefficient; ω represents a mechanical angular velocity; t is_eRepresents electromagnetic torque and satisfies

Considering the uncertainty of system parameters and the influence of external disturbance, a state variable d is introduced_ωlThe mechanical equation of motion is organized into the following form:

wherein the content of the first and second substances,representing machine-side lumped disturbances; b_ω0In relation to b_ωAn estimated value of (d);representing the control quantity after disturbance compensation at the machine side controller end; ω represents a mechanical angular velocity;

let x₁＝ω，x₂＝d_ωlThen the new state space model is represented as:

wherein h is₁Denotes d_ωlDifferentiation of (1); b_ω0In relation to b_ωAn estimated value of (d);representing the control quantity after disturbance compensation at the machine side controller end;

step 8, designing an extended state observer;

after the lumped disturbance is expanded into a new state variable, the state information of the system and the information of the lumped disturbance can be estimated in real time by means of an expanded state observer, and the representation form of the expanded state observer is as follows:

wherein the content of the first and second substances,an estimate representing ω;representing lumped disturbances d_ωlAn estimated value of (d); beta is a₁,β₂Representing the gain of the extended state observer;representing the control quantity after disturbance compensation at the machine side controller end; ω represents a mechanical angular velocity; b_ω0Is about b_ωAn estimated value of (d);

step 9, designing a control law;

compensating the disturbance on the system at the output end of the designed controller to obtain the final control law, wherein the process is as follows:

as long as the appropriate parameter beta is selected₁,β₂The estimated value of the rotating speed can be estimated in real time through the extended state observerDefining an error variable e satisfying the following expression:

the controller output is obtained in the form shown below:

and (3) carrying out disturbance compensation on the system to obtain a final control law:

wherein the content of the first and second substances,an estimate representing ω; omega^*A reference value representing an outer ring of rotational speeds; u. of₀Representing the machine side controller output;representing the control quantity after disturbance compensation at the machine side controller end; k is a radical of_pRepresenting the controller gain; b_ω0Is about b_ωAn estimated value of (d);representing machine side lumped disturbances d_ωlAn estimated value of (d);

step 10, establishing a direct current link mathematical model;

step 11, establishing a network side mathematical model;

step 12, sampling and coordinate transformation of current and voltage on the network side;

step 13, determining a network side current inner loop prediction model;

step 14, constructing a cost function;

step 15, selecting an optimal voltage vector;

step 16, introducing a state variable d_ulConstructing a new state space model;

step 17, designing an extended state observer;

and step 18, designing a control law.

2. The disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1, wherein in the step 10, the current is obtained from kirchhoff's current law:

wherein C represents a direct current filter capacitor; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor; i.e. i_LRepresenting the current delivered by the machine side; i.e. i_gRepresenting the current delivered to the net side.

3. The disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1 or 2 is characterized in that: in the step 11, a mathematical model of the grid-connected inverter of the permanent magnet direct-drive wind power generation system in a two-phase d-q rotating coordinate system is represented as follows:

in the formula: u. of_d,u_qRepresenting the output voltage component of the inverter under a two-phase synchronous rotation d-q coordinate system; e.g. of the type_d,e_qRepresenting the voltage component of the power grid voltage under a two-phase synchronous rotation d-q coordinate system; i.e. i_d,i_qRepresenting d and q axis components of the output current of the inverter under a two-phase synchronous rotation d-q coordinate system; l represents a network side filter inductor; r represents the equivalent resistance of the output end; omega_gRepresenting the grid angular frequency.

4. The disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1 or 2 is characterized in that: in the step 13, to implement the unit power factor grid connection, in the control of the grid-side inverter, the grid voltage vector is oriented to the d axis, and then

Wherein e is_mRepresenting the grid voltage oriented on the d-axis;

thereby, a grid-side inverter mathematical model based on grid voltage vector orientation is obtained:

wherein u is_d,u_qRepresenting the output voltage component of the inverter under a two-phase synchronous rotation d-q coordinate system; e.g. of the type_dA d-axis voltage component representing the grid voltage; i.e. i_d,i_qRepresenting d and q axis components of the output current of the inverter under a two-phase synchronous rotation d-q coordinate system; l represents a network side filter inductor; r represents the equivalent resistance of the output end; omega_gRepresenting the grid angular frequency;

because the system has control delay, the control delay needs to be compensated when the system state value is predicted, and the current state equation is subjected to discrete post-processing by utilizing a forward Euler formula:

wherein i_d(k),i_q(k) Representing the d and q axis components of the output current of the inverter at the current moment; i.e. i_d(k+1),i_q(k +1) is the d and q axis components of the inverter output current after delay compensation; u. of_d,u_qRepresenting the components of the optimal voltage vector acting on the system at the last moment in a d-q rotating coordinate system; e.g. of the type_d(k) A d-axis voltage component representing the grid voltage at the present time; omega_g(k) Representing the grid angular frequency; l represents a network side filter inductor; r represents the equivalent resistance of the output end;

and predicting the current state at the next moment on the basis of the time delay compensation:

wherein i_d(k+2),i_q(k +2) are predicted values of inverter output current d and q-axis components at the moment of k +1 respectively; i.e. i_d(k+1),i_q(k +1) is the d and q axis components of the inverter output current after delay compensation; omega_g(k +1) is approximately equal to ω_g(k)；e_d(k +1) is approximately equal to e_d(k)；T_sIs a sampling period; u. of_di(0,1,…7),u_qi(0,1,…7)And d-axis and q-axis components of 8 basic voltage space vectors under a rotating coordinate system are corresponding to 8 switching states of the inverter.

5. The disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1 or 2 is characterized in that: in the step 14, a cost function, cost function J, is constructed₂The form is as follows:

wherein the content of the first and second substances,indicating inverter outputReference values of d and q axis components of the output current; i.e. i_d(k+2),i_qAnd (k +2) are predicted values of inverter output current d and q-axis component at the moment of k +1 respectively.

6. The disturbance compensation-based finite set model prediction current control method of the permanent magnet direct-driven wind power generation system as claimed in claim 5, wherein: in the step 15, the predicted value outputs under different voltage vectors are sequentially calculated and are brought into the constructed cost function, and the optimal voltage vector u which enables the cost function value to be minimum is selected from the cost function_{opt_2}；

Wherein the content of the first and second substances,a reference value representing the d, q-axis component of the inverter output current; i.e. i_d(k+2),i_q(k +2) are predicted values of inverter output current d and q-axis components at the moment of k +1 respectively;

u_{opt_2}＝arg min J₂。

7. the disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1 or 2 is characterized in that: in the step 16, a state variable d is introduced_ulConstructing a new state space model;

from kirchhoff's current law:

wherein C represents a direct current filter capacitor; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor; i.e. i_LRepresenting the current delivered by the machine side; i.e. i_gRepresenting the current delivered to the net side;

the power on the net side is expressed as:

P＝u_dci_g

the following equation is obtained:

in a synchronous rotating coordinate system, the active power P and the reactive power Q of the grid-side converter to the power grid are respectively as follows:

therefore, it is

The formula is arranged to obtain:

wherein the content of the first and second substances,representing net side lumped disturbance; b_u0Is about b_uAn estimated value of (d);representing the control quantity after disturbance compensation of the network side controller end; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor;

selection of x₃＝u_dc；x₄＝d_ulFor a new state variable, its state space form is:

wherein h is₂Denotes d_ulDifferentiation of (1); b_u0Is about b_uAn estimated value of (d);and the control quantity after disturbance compensation is performed on the network side controller end is shown.

8. The disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1 or 2 is characterized in that: in step 17, the representation form of the network-side voltage loop extended state observer is as follows:

wherein the content of the first and second substances,represents u_dcAn estimated value of (d);representing net-side lumped disturbances d_ulAn estimated value of (d); beta is a₃,β₄Representing the gain of the extended state observer;representing the control quantity after disturbance compensation of the network side controller end; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor; b_u0Is about b_uAn estimate of (d).

9. The disturbance compensation-based method for predicting the current of the finite set model of the permanent magnet direct-drive wind power generation system according to claim 1 or 2 is characterized in that: in step 18, the grid-side voltage loop control law is designed as follows:

wherein the content of the first and second substances,represents u_dcAn estimated value of (d);represents a DC-side bus voltage reference value; u. of₁Representing the network side controller output;representing the control quantity after disturbance compensation of the network side controller end; k is a radical of_p2Representing the controller gain; b_u0Is about b_uAn estimated value of (d);representing net-side lumped disturbances d_ulAn estimate of (d).

Technical Field

The invention relates to the technical field of wind power generation, in particular to a finite set model prediction current control method of a permanent magnet direct-drive wind power generation system based on disturbance compensation of a machine side speed loop and a network side voltage loop.

Background

The permanent magnet direct-drive wind power generation system has the advantages of no gear box, high power generation efficiency at low wind speed and the like, and is widely concerned by more and more students. However, natural wind has the characteristics of randomness, time-varying property and the like, and the wind power generation system has the characteristics of nonlinearity, strong coupling, multivariable, multiple disturbances, frequent switching of working conditions and the like, so that an accurate model of the system is difficult to obtain, and the control effect of the traditional PID control method is poor. For the characteristics of the wind power generation system, a large number of advanced control strategies such as an active disturbance rejection control technology, model prediction control and the like are provided by scholars at home and abroad.

The machine side control of the permanent magnet direct-drive wind power generation system adopts double closed-loop control of a speed loop and a current loop, and if the traditional PID control is adopted, the control effect of the machine side permanent magnet synchronous generator is difficult to meet the control requirement due to the characteristics of nonlinearity, strong coupling and the like of the machine side permanent magnet synchronous generator. Compared with the traditional PID control, the finite set model predictive control (FCS-MPC) has the advantages of simple and intuitive design, no need of a modulator, quick dynamic response, multi-target synergy, easy realization of a nonlinear target and the like. The active disturbance rejection control can regard external disturbance and uncertainty factors of a system as lumped disturbance, estimate the lumped disturbance through an extended state observer, and compensate at a controller end, so that the disturbance rejection performance of the system is improved. The grid-side control of the permanent magnet direct-drive wind power generation system adopts voltage loop and current loop double closed loop control, the grid-connected inverter is an important part of energy conversion of the wind power generation system, but the improvement of the control strategy of the wind power grid-connected inverter is still a current research focus due to the characteristics of randomness, intermittence and the like of wind power generation. In order to improve the stability of the voltage at the direct current side of the grid-connected inverter of the permanent magnet direct-drive wind power generation system, a voltage outer ring controller of a first-order LADRC is designed, and a grid-side current inner ring adopts limited current collection prediction control as the same as the machine side, so that the anti-interference performance and the dynamic response of the system are improved.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention provides a disturbance compensation-based method for predicting the current of a permanent magnet direct-drive wind power generation system finite set model, wherein the outer rings of a machine side and a network side adopt an extended state observer to estimate the disturbance, and the controller end is used for compensating; the inner rings of the machine side and the network side adopt a finite set model to predict current control, thereby solving the technical problem.

In order to solve the technical problems, the invention provides the following technical scheme:

a disturbance compensation-based method for predicting current of a finite set model of a permanent magnet direct-drive wind power generation system comprises the following steps:

step 1, in order to realize maximum wind energy capture, maximum power point tracking is adopted to determine a given rotating speed value omega of a machine side speed ring^*；

Step 2, establishing a machine side permanent magnet synchronous motor mathematical model;

step 3, sampling and coordinate transformation of the machine side current and speed, and the process is as follows:

real-time acquisition machine side converter three-phase current i_a,i_b,i_cThe current value i in the two-phase stationary coordinate system (. alpha.beta.) can be obtained by Clark conversion_α,i_βOn the basis, Park conversion is carried out to obtain a current value i under a two-phase rotating coordinate system (dq)_d,i_q；

Step 4, determining a machine side current inner loop prediction model, wherein the process is as follows:

because the machine side uses a two-level converter, each bridge arm has two switch states, 8 switch states can be obtained, corresponding to 8 output voltage vectors u_i(i ═ 0.., 7). Corresponding output voltage vector u when upper bridge arms of three bridge arms of two-level converter are all closed₇Output voltage vector u corresponding to when lower bridge arm is fully closed₀The same, so when the model prediction calculation is carried out, one is selected, but u is the final voltage vector selected₀And u₇The selection of the switch is in accordance with the principle of minimum switching loss;

in the formula: u. of_d,u_qRepresenting a stator voltage component under a two-phase synchronous rotation d-q coordinate system; i.e. i_d,i_qRepresenting the stator current component under a two-phase synchronous rotation d-q coordinate system; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_reRepresents an electrical angular velocity;

discrete post-processing is carried out on a current state equation by utilizing a forward Euler formula:

wherein i_d(k),i_q(k) Representing the stator current component under the two-phase synchronous rotation d-q coordinate system at the current moment; i.e. i_d(k+1)，i_q(k +1) is the stator current d, q-axis component at the next moment; u. of_d,u_qD, q axis voltage components corresponding to 8 switch states; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k) Represents the electrical angular velocity at the present moment; t is_sIs a sampling period;

due to the existence of control delay in the system, the predicted value of the motor state calculated according to the prediction model is inaccurate, and the selected voltage vector cannot guarantee the optimum, so that the control delay needs to be compensated:

wherein i_d(k),i_q(k) Representing the stator current component under the two-phase synchronous rotation d-q coordinate system at the current moment; i.e. i_d(k+1)，i_q(k +1) is a component of the d and q axes of the stator current after time delay compensation; u. of_d,u_qThe component of the optimal voltage vector acting on the system at the last moment under a d-q rotating coordinate system; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k) Represents the electrical angular velocity at the present moment; t is_sIs a sampling period;

and predicting the current state at the next moment on the basis of the time delay compensation:

wherein i_d(k+1)，i_q(k +1) is a component of the d and q axes of the stator current after time delay compensation; i.e. i_d(k+2)，i_q(k +2) are predicted values of the stator current of the q axis at the moment d and k +1 respectively; u. of_di(0,1,…7),u_qi(0,1,…7)D, q axis voltage components corresponding to 8 switch states; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k +1) is approximately equal to ω_re(k)；T_sIs a sampling period;

step 5, constructing a cost function;

in order to better show the tracking effect of the predicted value on the expected value, a cost function, namely a cost function J, needs to be constructed₁The form is as follows:

wherein the content of the first and second substances,a reference value representing a stator current d, q-axis component; i.e. i_d(k+2),i_q(k +2) are predicted values of the stator current of the q axis at the moment d and k +1 respectively;

step 6, selecting an optimal voltage vector;

sequentially calculating predicted value outputs under different voltage vectors, bringing the predicted value outputs into the constructed cost function, and selecting the optimal voltage vector u which enables the cost function value to be minimum_{opt_1}；

u_{opt_1}＝arg min J₁

Step 7, introducing a state variable d_ωlConstructing a new state space model;

in the actual speed regulation process of the motor, the outer ring of the rotating speed is easily influenced by uncertain model parameters and external disturbance, and a mechanical motion equation can be described as

Wherein J represents the rotor moment of inertia; t is_mRepresenting a drive torque; b represents a friction coefficient; ω represents a mechanical angular velocity; t is_eRepresents electromagnetic torque and satisfies

Considering the uncertainty of system parameters and the influence of external disturbance, a state variable d is introduced_ωlThe mechanical equation of motion is organized into the following form:

wherein the content of the first and second substances,representing machine-side lumped disturbances; b_ω0Is about b_ωAn estimated value of (d);representing the control quantity after disturbance compensation at the machine side controller end; ω represents a mechanical angular velocity;

let x₁＝ω，x₂＝d_ωlThen the new state space model is represented as:

wherein h is₁Denotes d_ωlDifferentiation of (1); b_ω0Is about b_ωAn estimated value of (d);representing the control quantity after disturbance compensation at the machine side controller end;

step 8, designing an extended state observer;

after the lumped disturbance is expanded into a new state variable, the state information of the system and the information of the lumped disturbance can be estimated in real time by means of an expanded state observer, and the representation form of the expanded state observer is as follows:

wherein the content of the first and second substances,an estimate representing ω;representing lumped disturbances d_ωlAn estimated value of (d); beta is a₁,β₂Presentation expanderGain of the tensile state observer;representing the control quantity after disturbance compensation at the machine side controller end; b_ω0Is about b_ωAn estimated value of (d);

step 9, designing a control law;

compensating the disturbance on the system at the output end of the designed controller to obtain the final control law, wherein the process is as follows:

as long as the appropriate parameter beta is selected₁,β₂The estimated value of the rotating speed can be estimated in real time through the extended state observerDefining an error variable e satisfying the following expression:

the controller output is obtained in the form shown below:

and (3) carrying out disturbance compensation on the system to obtain a final control law:

wherein the content of the first and second substances,an estimate representing ω; omega^*A reference value representing an outer ring of rotational speeds; u. of₀Representing the machine side controller output;indicating control after disturbance compensation at the machine-side controller endPreparing quantity; k is a radical of_pRepresenting the controller gain; b_ω0Is about b_ωAn estimated value of (d);representing machine side lumped disturbances d_ωlAn estimated value of (d);

step 10, establishing a direct current link mathematical model;

step 11, establishing a network side mathematical model;

step 12, sampling and coordinate transformation of current and voltage on the network side;

step 13, determining a network side current inner loop prediction model;

step 14, constructing a cost function;

step 15, selecting an optimal voltage vector;

step 16, introducing a state variable d_ulConstructing a new state space model;

step 17, designing an extended state observer;

and step 18, designing a control law.

The invention has the beneficial effects that: the outer ring of the wind power generation system is subjected to disturbance estimation by adopting the extended state observer and compensated at the controller end, so that the dynamic performance of speed control can be effectively improved, random disturbance caused by factors such as wind speed is restrained, the current loop adopts finite set model predictive control to replace a traditional PI controller, and the dynamic response speed of the wind power generation system is accelerated due to the advantage of no modulator.

Drawings

FIG. 1 is a structural diagram of a permanent magnet direct-drive wind power generation system.

FIG. 2 is a vector diagram in α - β and d-q coordinate systems.

FIG. 3 is a block diagram of machine side first-order active disturbance rejection control based finite set model predictive current control.

Fig. 4 is a net side finite set model predictive current control block diagram based on first order active disturbance rejection control.

FIG. 5 is a flow chart of a machine side finite set model predictive current control algorithm based on first order active disturbance rejection control.

FIG. 6 is a flow chart of a finite set model predictive current control algorithm for grid-side based on first-order active disturbance rejection control.

FIG. 7 is a simulation diagram of the rotating speed waveform when the wind speed increases at 0.7 s.

FIG. 8 is a simulation of electromagnetic torque waveforms at 0.7s when wind speed increases.

Fig. 9 is a simulation diagram of a stator current component tracking waveform of the q-axis of the current loop on the side of the wind speed rise timing at 0.7 s.

FIG. 10 is a graph of voltage loop voltage waveform simulations for wind speed increase at 0.7s and grid voltage increase at 1.5 s.

FIG. 11 is a simulation diagram of the d-axis stator current component tracking waveform of the grid-side current loop when the wind speed increases at 0.7 s.

Fig. 12 is a simulation diagram of the waveform of the output voltage and current of the a phase on the network side.

Fig. 13 is a graph showing a simulated rotation speed waveform at 0.7s of disturbance at a constant wind speed.

Detailed Description

In order to make the technical solution of the present invention clearer, the following detailed description is made with reference to the accompanying drawings. The embodiments described herein are merely illustrative and are not intended to be limiting.

Referring to fig. 1 to 13, a disturbance compensation-based method for controlling a current predicted by a finite set model of a permanent magnet direct-drive wind power generation system includes the following steps:

step 1, in order to realize maximum wind energy capture, maximum power point tracking is adopted to determine a given rotating speed value omega of a machine side speed ring^*；

Step 2, establishing a machine side mathematical model, wherein the process is as follows:

mathematical model of PMSG under two-phase synchronous rotation d-q coordinate system:

in the formula: u. of_d,u_qRepresenting stator voltage components on a two-phase synchronous rotation d-q coordinate system; i.e. i_d,i_qRepresents twoRotating the stator current component on the d-q coordinate system in phase synchronization; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; omega_reRepresents an electrical angular velocity; ω represents a mechanical angular velocity; psi_fRepresents the permanent magnet flux; j represents moment of inertia; p is a radical of_nRepresenting the number of pole pairs; b represents a friction coefficient; t is_mRepresents a driving torque; t is_eRepresenting an electromagnetic torque.

Step 3, sampling and coordinate transformation of machine side current and speed

Real-time acquisition machine side converter three-phase current i_a,i_b,i_cThe current value i in the two-phase stationary coordinate system (. alpha.beta.) can be obtained by Clark conversion_α,i_βOn the basis, Park conversion is carried out to obtain a current value i under a two-phase rotating coordinate system (dq)_d,i_q。

Clark transformation:

park transformation:

wherein, theta_reTo rotate the electrical angle, satisfy

Step 4, determining a machine side current inner loop prediction model, wherein the process is as follows:

because the machine side uses a two-level converter, each bridge arm has two switch states, 8 switch states can be obtained, corresponding to 8 output voltage vectors u_i(i ═ 0.., 7). Corresponding output voltage vector u when upper bridge arms of three bridge arms of two-level converter are all closed₇Output voltage vector u corresponding to when lower bridge arm is fully closed₀Same, so model prediction is being performedDuring calculation, one voltage vector is selected, and the voltage vector u is finally selected₀And u₇The selection of (1) follows the principle of minimum switching loss.

In the formula: u. of_d,u_qRepresenting stator voltage components on a two-phase synchronous rotation d-q coordinate system; i.e. i_d,i_qRepresenting stator current components on a two-phase synchronous rotation d-q coordinate system; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; omega_reRepresents an electrical angular velocity; ω represents a mechanical angular velocity; psi_fRepresents the permanent magnet flux; j represents moment of inertia; p is a radical of_nRepresenting the number of pole pairs; b represents a friction coefficient; t is_mRepresents a driving torque; t is_eRepresenting an electromagnetic torque.

Discrete post-processing is carried out on a current state equation by utilizing a forward Euler formula:

wherein i_d(k),i_q(k) Representing the stator current component under the two-phase synchronous rotation d-q coordinate system at the current moment; i.e. i_d(k+1)，i_q(k +1) is the stator current d, q-axis component at the next moment; u. of_d,u_qD, q axis voltage components corresponding to 8 switch states; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k) Represents the electrical angular velocity at the present moment; t is_sIs the sampling period.

Due to the control delay of the system, the motor state prediction value calculated according to the prediction model is inaccurate, and the selected voltage vector cannot be guaranteed to be optimal. For this reason, the control delay needs to be compensated:

wherein i_d(k),i_q(k) Representing the stator current component under the two-phase synchronous rotation d-q coordinate system at the current moment; i.e. i_d(k+1)，i_q(k +1) is a component of the d and q axes of the stator current after time delay compensation; u. of_d,u_qThe component of the optimal voltage vector acting on the system at the last moment under a d-q rotating coordinate system; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k) Represents the electrical angular velocity at the present moment; t is_sIs the sampling period.

And predicting the current state at the next moment on the basis of the time delay compensation:

wherein i_d(k+1)，i_q(k +1) is a component of the d and q axes of the stator current after time delay compensation; i.e. i_d(k+2)，i_q(k +2) are predicted values of the stator current of the q axis at the moment d and k +1 respectively; u. of_di(0,1,…7),u_qi(0,1,…7)D, q axis voltage components corresponding to 8 switch states; l is_sThe inductance of a stator under a d-q coordinate system in a surface-mounted permanent magnet synchronous motor is obtained; r_sRepresenting the stator resistance; psi_fRepresents a permanent magnet flux linkage; omega_re(k +1) is approximately equal to ω_re(k)；T_sIs the sampling period.

Step 5, constructing a cost function;

in order to better show the tracking effect of the predicted value on the expected value, a cost function, namely a cost function J, needs to be constructed₁The form is as follows:

wherein the content of the first and second substances,a reference value representing a stator current d, q-axis component; i.e. i_d(k+2),i_qAnd (k +2) are predicted values of the stator current of the q axis at the moment d of k +1 respectively.

Step 6, selecting an optimal voltage vector;

sequentially calculating predicted value outputs under different voltage vectors, bringing the predicted value outputs into the constructed cost function, and selecting the optimal voltage vector u which enables the cost function value to be minimum_{opt_1}。

u_{opt_1}＝arg min J₁

Step 7, introducing a state variable d_ωlConstructing a new state space model;

in the actual speed regulation process of the motor, the outer ring of the rotating speed is easily influenced by uncertain model parameters and external disturbance, and a mechanical motion equation can be described as

Wherein J represents the rotor moment of inertia; t is_mRepresenting a drive torque; b represents a friction coefficient; ω represents a mechanical angular velocity; t is_eRepresents electromagnetic torque and satisfies

Considering the uncertainty of system parameters and the influence of external disturbance, a state variable d is introduced_ωlThe mechanical equation of motion can be organized into the following form:

wherein the content of the first and second substances,representing machine-side lumped disturbances; b_ω0Is about b_ωAn estimated value of (d);representing the control quantity after disturbance compensation at the machine side controller end; ω represents the mechanical angular velocity.

Let x₁＝ω，x₂＝d_ωlThen the new state space model is:

wherein h is₁Denotes d_ωlDifferentiation of (1); b_ω0Is about b_ωAn estimated value of (d);and the control quantity after disturbance compensation is carried out on the machine side controller end is shown.

Step 8, designing an extended state observer;

after the lumped disturbance is expanded into a new state variable, the state information of the system and the information of the lumped disturbance can be estimated in real time by means of an expanded state observer, and the representation form of the expanded state observer is as follows:

wherein the content of the first and second substances,an estimate representing ω;representing lumped disturbances d_ωlAn estimated value of (d); beta is a₁,β₂Representing the gain of the extended state observer;representing the control quantity after disturbance compensation at the machine side controller end;ω represents a mechanical angular velocity; b_ω0Is about b_ωAn estimate of (d).

Step 9, designing a control law;

compensating the disturbance on the system at the output end of the designed controller to obtain the final control law, wherein the process is as follows:

as long as the appropriate parameter beta is selected₁,β₂The estimated value of the rotating speed can be estimated in real time through the extended state observerDefining an error variable e satisfying the following expression:

controller output can be obtained in the form shown below:

and (3) carrying out disturbance compensation on the system to obtain a final control law:

wherein the content of the first and second substances,an estimate representing ω; omega^*A reference value representing an outer ring of rotational speeds; u. of₀Representing the machine side controller output;representing the control quantity after disturbance compensation at the machine side controller end; k is a radical of_pRepresenting the controller gain; b_ω0Is about b_ωAn estimated value of (d);representing machine side lumped disturbances d_ωlAn estimate of (d).

Step 10, establishing a direct current link mathematical model;

from kirchhoff's current law:

wherein C represents a direct current filter capacitor; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor; i.e. i_LRepresenting the current delivered by the machine side; i.e. i_gRepresenting the current delivered to the net side.

Step 11, establishing a network side mathematical model;

the method comprises the following steps of (1) performing mathematical model of a grid-connected inverter of the wind power generation system under a two-phase d-q rotating coordinate system:

wherein u is_d,u_qRepresenting the output voltage component of the inverter under a two-phase synchronous rotation d-q coordinate system; e.g. of the type_d,e_qRepresenting the voltage component of the power grid voltage under a two-phase synchronous rotation d-q coordinate system; i.e. i_d,i_qRepresenting d and q axis components of the output current of the inverter under a two-phase synchronous rotation d-q coordinate system; l represents a network side filter inductor; r represents the equivalent resistance of the output end; omega_gRepresenting the grid angular frequency.

Step 12, sampling and coordinate transformation of current and voltage on the network side;

in order to simplify the design of the control system, it is necessary to convert the grid voltage in the three-phase stationary coordinate system, the inverter output current, and the voltages of the respective phases output by the inverter into dc values in the two-phase rotating coordinate system.

Clark transformation:

park transformation:

wherein, theta_gIs the spatial angle of the power grid.

Step 13, determining a network side current inner loop prediction model;

wherein u is_d,u_qRepresenting the output voltage component of the inverter under a two-phase synchronous rotation d-q coordinate system; e.g. of the type_d,e_qRepresenting the voltage component of the power grid voltage under a two-phase synchronous rotation d-q coordinate system; i.e. i_d,i_qRepresenting d and q axis components of the output current of the inverter under a two-phase synchronous rotation d-q coordinate system; l represents a network side filter inductor; r represents the equivalent resistance of the output end; omega_gRepresenting the grid angular frequency.

In the control of the grid-side inverter, the grid voltage vector is usually oriented on the d-axis, and

wherein e is_mRepresenting the grid voltage oriented on the d-axis.

Then a grid-side inverter mathematical model based on grid voltage vector orientation can be obtained:

wherein u is_d,u_qRepresenting the output voltage component of the inverter under a two-phase synchronous rotation d-q coordinate system; e.g. of the type_dA d-axis voltage component representing the grid voltage; i.e. i_d,i_qD-q coordinate representing synchronous rotation of inverter output current in two phasesThe d, q axis components of the system; l represents a network side filter inductor; r represents the equivalent resistance of the output end; omega_gRepresenting the grid angular frequency.

Discrete post-processing is carried out on a current state equation by utilizing a forward Euler formula:

wherein i_d(k),i_q(k) Representing the d and q axis components of the output current of the inverter at the current moment; i.e. i_d(k+1),i_q(k +1) is a d, q axis component of the inverter output current at the next moment; u. of_d,u_qRepresenting the inverter output voltage component under the d-q coordinate system corresponding to 8 switch states; e.g. of the type_d(k) A d-axis voltage component representing the grid voltage at the present time; l represents a network side filter inductor; r represents the equivalent resistance of the output end; omega_gRepresenting the grid angular frequency.

Due to the control delay of the system, the motor state prediction value calculated according to the prediction model is inaccurate, and the selected voltage vector cannot be guaranteed to be optimal. For this reason, the control delay needs to be compensated:

wherein i_d(k),i_q(k) Representing the d and q axis components of the output current of the inverter at the current moment; i.e. i_d(k+1),i_q(k +1) is the d and q axis components of the inverter output current after delay compensation; u. of_d,u_qRepresenting the components of the optimal voltage vector acting on the system at the last moment in a d-q rotating coordinate system; e.g. of the type_d(k) A d-axis voltage component representing the grid voltage at the present time; omega_g(k) Representing the grid angular frequency; l represents a network side filter inductor; and R represents the equivalent resistance of the output end.

And predicting the current state at the next moment on the basis of the time delay compensation:

wherein i_d(k+2),i_q(k +2) are predicted values of inverter output current d and q-axis components at the moment of k +1 respectively; i.e. i_d(k+1),i_q(k +1) is the d and q axis components of the inverter output current after delay compensation; omega_g(k +1) is approximately equal to ω_g(k)；e_d(k +1) is approximately equal to e_d(k)；T_sIs a sampling period; u. of_di(0,1,…7),u_qi(0,1,…7)And d-axis and q-axis components of 8 basic voltage space vectors under a rotating coordinate system are corresponding to 8 switching states of the inverter.

Step 14, constructing a cost function;

in order to better show the tracking effect of the predicted value on the expected value and the stability of the direct current bus voltage, a cost function, namely a cost function J, needs to be constructed₂The form is as follows:

wherein the content of the first and second substances,a reference value representing the d, q-axis component of the inverter output current; i.e. i_d(k+2),i_qAnd (k +2) are predicted values of inverter output current d and q-axis component at the moment of k +1 respectively.

Step 15, selecting an optimal voltage vector;

sequentially calculating predicted value outputs under different voltage vectors, bringing the predicted value outputs into the constructed cost function, and selecting the optimal voltage vector u which enables the cost function value to be minimum_{opt_2}。

u_{opt_2}＝arg min J₂

Step 16, introducing a state variable d_ulConstructing a new state space model;

from kirchhoff's current law:

wherein C represents a direct current filter capacitor; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor; i.e. i_LRepresenting the current delivered by the machine side; i.e. i_gRepresenting the current delivered to the net side.

The power on the net side can be expressed as:

P＝u_dci_g

the following equation can be obtained:

in a synchronous rotating coordinate system, the active power P and the reactive power Q of the grid-side converter to the power grid are respectively

Therefore, it is

The formula is arranged to obtain:

wherein the content of the first and second substances,representing net side lumped disturbance; b_u0Is about b_uAn estimated value of (d);representing the control quantity after disturbance compensation of the network side controller end; u. of_dcTo representAnd voltage at two ends of the direct current bus capacitor.

Selection of x₃＝u_dc；x₄＝d_ulFor a new state variable, its state space form is:

wherein h is₂Denotes d_ulDifferentiation of (1); b_u0Is about b_uAn estimated value of (d);and the control quantity after disturbance compensation is performed on the network side controller end is shown.

Step 17, designing an extended state observer;

after the lumped disturbance is a new state variable, the information of the state of the system and the information of the lumped disturbance can be estimated in real time by means of an extended state observer, and the representation form of the extended state observer is as follows:

wherein the content of the first and second substances,represents u_dcAn estimated value of (d);representing net-side lumped disturbances d_ulAn estimated value of (d); beta is a₃,β₄Representing the gain of the extended state observer;representing the control quantity after disturbance compensation of the network side controller end; u. of_dcRepresenting the voltage at two ends of the direct current bus capacitor; b_u0Is about b_uAn estimate of (d).

Step 18, designing a control law;

wherein the content of the first and second substances,represents u_dcAn estimated value of (d);represents a DC-side bus voltage reference value; u. of₁Representing the network side controller output;representing the control quantity after disturbance compensation of the network side controller end; k is a radical of_p2Representing the controller gain; b_u0Is about b_uAn estimated value of (d);representing net-side lumped disturbances d_ulAn estimate of (d).

Finally, the algorithm is realized in Matlab-simulink software, and the simulation results are shown in FIGS. 7-13.

As shown in fig. 7 and 8, when the wind speed increases at 0.7s, the rotating speed and the electromagnetic torque of the permanent magnet synchronous motor quickly reach a new steady state; as shown in fig. 9, the actual q-axis current can well track the q-axis current reference given by the voltage outer loop controller; as shown in fig. 10 and 11, when the wind speed increases at 0.7s, the voltage quickly returns to the set value of 600V, and when the wind speed increases at 1.5s, the grid voltage increases, and it can be seen that the voltage at both ends of the dc filter capacitor quickly returns to the set value. The actual d-axis current at these two times can then track well to the d-axis current reference given by the voltage outer loop controller. From fig. 12 it can be seen that the full power factor grid connection can still be maintained in case of increased wind speed and increased grid; fig. 13 shows that the rotational speed can be quickly restored to the reference value when the disturbance is applied at the time 0.7s with a constant wind speed. Simulation results show that the designed rotating speed loop controller can well restrain external disturbance; the designed voltage loop controller can well restrain the disturbance under the conditions of wind speed change and grid voltage change, and can quickly recover to a set value.