CN101240775B - Control Method of Hydraulic Pitch System of Wind Turbine - Google Patents

Abstract

The control method of the wind turbine hydraulic pitch system aims at the shortcomings of the traditional minimum variance control method, and proposes an improved minimum variance control method that extends the prediction step and combines the time-varying limiting strategy and the soft control function to solve the problem. For non-minimum phase systems and systems with a stable zero point close to the unit circle, the traditional minimum variance control method has problems such as unbounded control input and excessive control effect. The system control method adopts the minimum variance control, and the controller records all the input control signals {u(k-1), u(k-2),...} and the output pitch of the controlled object up to the current sampling time k in real time Angle observation data {θ(k), θ(k-1),…}, and according to the expected pitch angle signal θ _m (k), the control variable u(k) of the control system is obtained.

Description

Control Method of Hydraulic Pitch System of Wind Turbine

technical field

The invention is a method for adjusting the hydraulic pitch variable pitch system of the wind turbine by improving the minimum variance control strategy, so that the pitch angle of the blade can track the set value quickly, stably and without deviation.

Background technique

The wind turbine pitch control system completes the adjustment of the blade pitch angle through the pitch regulator, which ensures the maximum wind energy capture efficiency below the rated wind speed, and limits the power output when the wind speed is high. The hydraulic variable pitch pitch regulator is realized through a proportional valve. The controller gives a control voltage of -10~+10V according to the signal of the position sensor, and converts it into a certain range of current signals through the proportional valve controller to control the direction and size of the output flow of the proportional valve. The size of the current signal is directly proportional to determine the opening of the proportional valve and adjust the flow of the oil circuit. The positive and negative of the current signal determines the direction of the oil circuit. When the oil reaches the hydraulic cylinder, the piston rod is pushed forward or backward according to the direction and flow output by the proportional valve, and the pitch angle of the blade is controlled to move within a certain range.

Due to the randomness and uncertainty of wind speed, load disturbance caused by tower shadow, wind shear, yaw rotation, etc., the reciprocating action of the variable pitch switch propeller, and the inertial link of the hydraulically driven large-mass impeller load, the pitch control The system has technical characteristics such as parameter nonlinearity, parameter time-varying, and hysteresis.

The minimum variance control method is very beneficial for the control of large-delay systems, and it also has certain practical value for nonlinear and time-varying systems. The traditional minimum variance control method requires that the object must be a minimum phase system, which is very sensitive to the stable zero point close to the unit circle. For the non-minimum phase system, it is impossible to design the control system due to the non-boundedness of the control input due to the unstable zero point. At the same time, the control effect may be too large, and the inaccurate initial conditions and the uncertainty of the delay time may significantly affect the optimal prediction and cause the control process to diverge.

Contents of the invention

Technical problem: The purpose of this invention is to provide a control method for the hydraulic pitch control system of wind turbines, which is used in the hydraulic pitch control system of wind turbines, that is, to solve the problems of nonlinear pitch control system parameters, time-varying, hysteresis, etc. .

Technical solution: In order to overcome the above problems, the improved minimum variance control method is used to make up for the shortcomings of the traditional method, and it is used in the hydraulic pitch control system, so that the system responds quickly, the dynamic tracking performance is good, and there is no adjustment error. At the same time, no Identify the object model online and make full use of the robustness of the control method.

The control method of the wind turbine hydraulic pitch control system of the present invention adopts the minimum variance control, and the realization steps are as follows:

Step 1: According to the pure delay time of the controlled object of the hydraulic pitch system, extend the prediction step size, and select a certain moment after the pure delay time as the prediction step size L, so that L is greater than the pure delay time of the object;

Step 2: From the prediction step size L and the parameter polynomial of the discrete model of the plant $A\left(q^{-1}\right)=1+\underset{i=1}{\overset{{\mathrm{no}}_{\mathrm{no}}}{\mathrm{\Σ}}}{a}_i{q}^{-i}$ and $B\left(q^{-1}\right)=\underset{i=0}{\overset{{\mathrm{no}}_b}{\mathrm{\Σ}}}{b}_i{q}^{-i},$ And relation 1=E(q ^-1 )A(q ^-1 )Δ+q ^-L G(q ^-1 ) and F(q ^-1 )=E(q ^-1 )B(q ^-1 ), we get polynomial $\mathrm{E.}\left(q^{-1}\right)=1+\underset{i=0}{\overset{L-1}{\mathrm{\Σ}}}{e}_i{q}^{-i},$ $G\left(q^{-1}\right)=\underset{i=0}{\overset{{\mathrm{no}}_a}{\mathrm{\Σ}}}{g}_i{q}^{-i},$ $f\left(q^{-1}\right)=\underset{i=0}{\overset{{\mathrm{no}}_b+L-1}{\mathrm{\Σ}}}{f}_i{q}^{-i},$ Stored in the controller module, used to obtain the control amount of the control system,

Where Δ=1-q ^-1 , q ^-1 is the backward shift operator, n _a , n _b are the order of the object discrete model parameter polynomials A(q ^-1 ), B(q ^-1 ) respectively, a _i ( i=1...n _a ), b _i (i=0...n _b ), e _i (i=1...L-1),

g _i (i=0...n ₀ ), f _i (i=0...n _b +L-1) are the coefficients of the corresponding polynomial respectively; Step 3: the control system records online and real-time all the parameters of the controller up to the current sampling time k control letter

number {u(k-1), u(k-2),…} and the pitch angle observation data {θ(k), θ(k-1),…} output by the pitch control system, and record the control quantity at the same time Increments of {Δu(k-1), Δu(k-2),…}, and receive the expected pitch angle signal θ _m (k) given by the main controller of the wind turbine, using the polynomial G stored in the control module (q ^-1 ) and F(q ^-1 ), the control quantity of the control system can be obtained according to the following relationship $u\left(k\right):\left\{\begin{array}{c}\mathrm{\Δ\; u}\left(k\right)=\frac{1}{\mathrm{\α}{f}_0}[{\mathrm{\θ}}_m\left(k\right)-\underset{i=0}{\overset{{\mathrm{no}}_a}{\mathrm{\Σ}}}{g}_i\mathrm{\θ}(k-i)-\underset{i=1}{\overset{{\mathrm{no}}_b+L-1}{\mathrm{\Σ}}}{f}_i\mathrm{\Δ\; u}(k-i)]\\ u\left(k\right)=u(k-1)+\mathrm{\Δ\; u}\left(k\right)\end{array}\right.,$ Among them: Δu(k) is the control increment of the control output at the current moment compared with the control output at the previous moment, and α is any normal constant;

Step 4: The control quantity u(k) of the control system is converted into a control voltage signal in the range of -10 ~ +10V after output, and converted into a certain range of current signal through the proportional valve controller. The current signal is according to the positive and negative poles The proportional control valve is connected to different circuits, and the opening of the valve position is determined in proportion to the current, and the flow of pressure oil is controlled so that the pressure oil flows into the front or rear end of the hydraulic cylinder from the connection circuit, thereby pushing the piston of the hydraulic cylinder Move backward or forward, control the pitch angle to change within a certain range, and track the desired pitch angle signal.

The prediction step size L can be predicted at any time between the pure delay time and the steady state time.

In the calculation of the control increment Δu(k), the time-varying limiting strategy is adopted, that is, a larger value of αf ₀ is taken at the initial stage of control to limit the control range, and the value of αf ₀ is gradually reduced in the later stage of control to speed up the time-varying αf ₀ of convergence. design method.

When selecting a certain moment after the pure delay time as the prediction step size L, you can also choose the prediction step size L ₁ and L ₂ between the time after the pure delay time and the steady state moment, so that L ₂ > L ₁ , where L ₁ is close to In the response stage after the pure delay time of the object, L ₂ is close to the steady state stage of the object, and a weighted compromise is made on the control quantities u ₁ (k) and u ₂ (k) obtained by the two prediction steps: u(k)=(1 -β)u ₁ (k)+βu ₂ (k), β is the input softening coefficient, and 0≤β≤1, the obtained u(k) is output as the control quantity of the control system.

Beneficial effects: The traditional minimum variance control is improved by extending the prediction step size, adopting time-varying limiting strategy and softening control effect, etc., and it is used in the hydraulic pitch control system of wind turbines, so that the system responds quickly and has no dynamic deviation. At the same time, for the time-varying nature of the object model, making full use of the robustness of the method, it is not necessary to identify the object model online, and still obtain good control results.

Description of drawings

Figure 1 is a schematic diagram of a hydraulic pitch control system for a wind turbine.

Detailed ways

The control method of the wind turbine hydraulic pitch control system of the present invention adopts minimum variance control, and the controller records all input control signals {u(k-1), u(k-2) of the controlled object up to the current sampling time k in real time online ),…} and the output pitch angle observation data {θ(k), θ(k-1),…}, and according to the desired pitch angle signal θ _m (k), the control system can be obtained according to the following relationship Control quantity u(k):

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; \&Delta;u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; \&Delta;u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ \mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Delta;u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\end{array}\right.$

The output control quantity u(k) of the controller is converted into a control voltage signal within the range of -10 ~ +10V, which is converted into a current signal within a certain range through the proportional valve controller, and the current signal controls the proportional valve to be connected according to the positive and negative polarity. Different circuits, and the opening of the valve position is determined in proportion to the current, and the flow of pressure oil is controlled so that the pressure oil flows into the front or rear end of the hydraulic cylinder from the connected circuit, thereby pushing the piston of the hydraulic cylinder to move backward or forward , control the pitch angle to change within a certain range, and track the expected pitch angle signal.

The prediction step size L can be predicted at any time between the pure delay time and the steady state time.

In the calculation of the control increment Δu(k), the time-varying limiting strategy is adopted, that is, a larger value of αf ₀ is taken at the initial stage of control to limit the control range, and the value of αf ₀ is gradually reduced in the later stage of control to speed up the time-varying αf ₀ of convergence. design method.

When selecting a certain moment after the pure delay time as the prediction step size L, you can also choose the prediction step size L ₁ and L ₂ between the time after the pure delay time and the steady state moment, so that L ₂ > L ₁ , where L ₁ is close to In the response stage after the pure delay time of the object, L ₂ is close to the steady state stage of the object, and a weighted compromise is made on the control quantities u ₁ (k) and u ₂ (k) obtained by the two prediction steps: u(k)=(1 -β)u ₁ (k)+βu ₂ (k), β is the input softening coefficient, and 0≤β≤1, the obtained u(k) is output as the control quantity of the control system.

The invention is an improved minimum variance control method aimed at the characteristics of non-linearity, time-varying and hysteresis of the parameters of the hydraulic variable pitch system of the wind turbine, so that the control system responds quickly, has no dynamic deviation, and does not need an online identification object model.

The specific implementation process of the improved minimum variance control method is as follows:

1. Identify the model of the controlled object offline, select the sampling time, discretize the object model, and use the controlled autoregressive integral sliding average model $A\left(q^{-1}\right)\mathrm{\&theta;}\left(k\right)=q^{-d}B\left(q^{-1}\right)u\left(k\right)+\frac{\mathrm{\&xi;}\left(k\right)}{\mathrm{\&Delta;}}$ Describe the controlled object, k is the sampling time, θ(k) is the pitch angle signal at time k, u(k) is the output control signal of the controller at time k, ξ(k) is the disturbance signal received by the hydraulic system at time k, d is the pure delay of the controlled object.

From the traditional minimum variance control control law, it can be known that the current input control effect will take effect at the next sampling time of d, and there is a large error in the optimal prediction of the future d steps for the object with a d step delay, which may make the control effect The range is too large. In view of this situation, according to the object's pure delay d, the prediction step length is appropriately extended, and L ₁ step and L ₂ step are taken as the prediction step size, where L ₁ step is close to the response stage after the object's pure delay d, and L ₂ steps are close to the object's steady state. state stage.

和

2. Solve the Diophantine equation 1=E(q ^-1 )A(q ^-1 )Δ+q ^-L G(q ^-1 ) and the relationship F from the prediction step size and the parameter polynomial of the discrete model of the controlled object (q ^-1 )=E(q ^-1 )B(q ^-1 ), thus obtaining the polynomial corresponding to L ₁ step and L ₂ step

and

和

由如下关系式： $\left\{\begin{array}{c}\mathrm{\&Delta;u}\left(k\right)=\frac{1}{\mathrm{\&alpha;}{f}_0}[{\mathrm{\&theta;}}_m\left(k\right)-\underset{i=0}{\overset{n_a}{\mathrm{\&Sigma;}}}{g}_i\mathrm{\&theta;}(k-i)-\underset{i=1}{\overset{n_b+L-1}{\mathrm{\&Sigma;}}}{f}_i\mathrm{\&Delta;u}(k-i)]\\ u\left(k\right)=u(k-1)+\mathrm{\&Delta;u}\left(k\right)\end{array}\right.$ 分别得到L₁步和L₂步对应的控制量u₁(k)和u₂(k)。3. Online real-time recording of all the control signals {u(k-1), u(k-2),...} of the controller up to the current sampling time k and the pitch angle observation data {θ(k ), θ(k-1),…}, while recording the increment {Δu(k-1), Δu(k-2),…} of the control variable, and receiving the expected pitch given by the main controller of the wind turbine Angle signal θ _m (k), using the obtained in the second step

and

By the following relationship: $\left\{\begin{array}{c}\mathrm{\&Delta;\; u}\left(k\right)=\frac{1}{\mathrm{\&alpha;}{f}_0}[{\mathrm{\&theta;}}_m\left(k\right)-\underset{i=0}{\overset{{\mathrm{no}}_a}{\mathrm{\&Sigma;}}}{g}_i\mathrm{\&theta;}(k-i)-\underset{i=1}{\overset{{\mathrm{no}}_b+L-1}{\mathrm{\&Sigma;}}}{f}_i\mathrm{\&Delta;u}(k-i)]\\ u\left(k\right)=u(k-1)+\mathrm{\&Delta;u}\left(k\right)\end{array}\right.$ The control quantities u ₁ (k) and u ₂ (k) corresponding to L ₁ step and L ₂ step are obtained respectively.

In the process of control convergence, the size of αf ₀ directly affects the size of the control action, and then affects the speed of convergence. Usually, the maximum control range often appears in the early stage of the sudden change process, so the design method of time-varying αf ₀ is adopted to limit the control range in the early stage of control and accelerate the convergence in the late stage of control.

In order to limit the large change of the control action and accelerate the convergence speed, the control action is compromised between u ₁ (k) and u ₂ (k): u(k)=(1-β)u ₁ (k)+βu ₂ (k), β is the input softening coefficient, and 0≤β≤1. u(k) is the output of the controller.

4. The control quantity u(k) of the control system is converted into a control voltage signal in the range of -10 ~ +10V after output, and converted into a certain range of current signal through the proportional valve controller. The current signal is according to the positive and negative polarity The control proportional valve is connected to different circuits, and the valve position opening is determined in proportion to the current, and the flow of pressure oil is controlled so that the pressure oil flows into the front or rear end of the hydraulic cylinder from the connection circuit, thereby pushing the piston of the hydraulic cylinder to Move backward or forward, control the pitch angle to change within a certain range, and track the desired pitch angle signal.

Claims (4)

1. A control method of a wind turbine hydraulic variable pitch system, characterized in that the system control method adopts minimum variance control, and the realization steps are as follows: Step 1: According to the pure delay time of the controlled object of the hydraulic variable pitch system, prolong the prediction step size, and select a certain moment after the pure delay time as the prediction step size L, so that L is greater than the pure delay time of the controlled object; Step 2: From the prediction step size L and the parameter polynomial of the discrete model of the plant

and

And relation 1=E(q ^-1 )A(q ^-1 )Δ+q ^-L G(q ^-1 ) and F(q ^-1 )=E(q ^-1 )B(q ^-1 ), we get polynomial

Stored in the controller module, it is used to obtain the control quantity of the control system, where Δ=1-q ^-1 , q ^-1 is the backward shift operator, n _a , n _b are the object discrete model parameter polynomial A(q ^{- 1} ), the order of B(q ^-1 ), a _i (i=1…n _a ), b _i (i=0…n _b ), e _i (i=1…L-1), g _i ( i=0...n _a ), f _i (i=0...n _b +L-1) are the coefficients of the corresponding polynomial respectively; Step 3: The control system records all the control signals {u(k-1), u(k-2),...} of the controller up to the current sampling time k in real time and the pitch angle observation data output by the pitch control system { θ(k), θ(k-1),…}, while recording the increment {Δu(k-1), Δu(k-2),…} of the control quantity, and receiving the information given by the main controller of the wind turbine The desired pitch angle signal θ _m (k), using the polynomials G(q ^-1 ) and F(q ^-1 ) stored in the control module, is used to obtain the control variable u(k) of the control system according to the following relationship:

Among them: Δu(k) is the control increment of the control output at the current moment compared with the control output at the previous moment, and α is any normal constant; Step 4: The control quantity u(k) of the control system is converted into a control voltage signal in the range of -10 ~ +10V after output, and converted into a certain range of current signal through the proportional valve controller. The current signal is according to the positive and negative poles The proportional control valve is connected to different circuits, and the opening of the valve position is determined in proportion to the current, and the flow of pressure oil is controlled so that the pressure oil flows into the front or rear end of the hydraulic cylinder from the connection circuit, thereby pushing the piston of the hydraulic cylinder Move backward or forward, control the pitch angle to change within a certain range, and track the desired pitch angle signal. 2. The control method of the wind turbine hydraulic pitch control system according to claim 1, characterized in that the predicted step size L is predicted at any time between the pure delay time and the steady state time. 3. The control method of the wind turbine hydraulic pitch control system according to claim 1, characterized in that in the calculation of the control increment Δu(k), a time-varying limiting strategy is adopted, that is, a larger αf ₀ is taken at the initial stage of control value, limit the control range, gradually reduce the value of αf ₀ in the later stage of control, and speed up the design method of time-varying αf ₀ of convergence. 4. The control method of the wind turbine hydraulic pitch variable pitch system according to claim 2, characterized in that when a certain moment after the pure delay time is selected as the prediction step L, it is also between the pure delay time and the steady-state moment Choose between L ₁ and L ₂ prediction steps, so that L ₂ > L ₁ , where L ₁ is close to the response stage after the pure delay time of the object, and L ₂ is close to the steady state stage of the object, the control amount obtained for these two prediction steps u ₁ (k) and u ₂ (k) make a weighted compromise: u(k)=(1-β)u ₁ (k)+βu ₂ (k), β is the input softening coefficient, and 0≤β≤1 , the obtained u(k) is output as the control quantity of the control system.