CN106786807A - A kind of wind power station active power control method based on Model Predictive Control - Google Patents

Abstract

The invention discloses a kind of wind power station active power control method based on Model Predictive Control, wind speed is classified Wind turbines according to residing for Wind turbines, and the local linearization state-space model of every class unit is set up and based on this predictive controller that designs a model using equivalent modeling method, consider the dynamic response characteristic of inhomogeneity unit, dispatching of power netwoks value is distributed into every class unit by certain priority, revised by model predictive controller again per the active reference value obtained by class unit, revise value and be proportionately distributed to all units in such.The present invention is controlled to target with active power of wind power field, it is considered to the operation difference between unit, and based on Model Predictive Control, makes it quickly to track dispatching of power netwoks value.

Description

A kind of wind power station active power control method based on Model Predictive Control

Technical field

The invention belongs to active power of wind power field control technology field, more specifically, it is related to a kind of pre- based on model The wind power station active power control method of observing and controlling.

Background technology

Wind-power electricity generation has obtained quick development as topmost renewable energy utilization form.Traditional wind power plant Control mode is a kind of autonomous active power controller mode, it is allowed to which wind turbine change voluntarily according to wind energy is use up in wind power plant Possible generating.As wind-powered electricity generation level of interpenetration gradually increases, the active power of wind power field that this uncertainty because of wind energy causes Fluctuation bring huge challenge to the safe operation of power network.Wind power plant realizes that controllable operation will be increasingly becoming Large Scale Wind Farm Integration The development trend being incorporated into the power networks, the key technology of the controllable operation of wind power plant is the active power controller of wind power plant.

Active power of wind power field control is that wind power plant oneself can be tracked dispatching of power netwoks value to the greatest extent.In order that wind power plant is active Output pulsation is stabilized, conventional method be use large-scale energy storage device, but the equipment cost of this method, technology into This and maintenance cost are higher.Another more economical method is the method for the Collaborative Control of Wind turbines.This method utilizes wind The set effect of electric field, is given per Fans distribution power reference value, per typhoon group of motors by the control unit of wind power plant layer Regard an actuator as, every active output sum of unit is then the total active power of wind power plant.

Domestic and foreign scholars expand a series of discussions, most commonly basis to the active power cooperative control method of blower fan Deviation between the active output of wind power plant that dispatching of power netwoks value and common link point measurement are obtained, takes proportional, integral controlling party Formula, then the use ratio method of salary distribution active reference value is distributed into each unit.Because wind power plant is a strong coupling of multivariable Close, and containing the system of constraint, be especially suitable for being controlled with model predictive control method.Recently scholar proposes model prediction Control mode is applied to wind power plant control, has two kinds of sides of centralized Model Predictive Control and distributed model predictive control again respectively Case.Because farm model is a system for multiple-input and multiple-output, with blower fan increasing number, exponent number is significantly increased, distribution The more centralized Model Predictive Control scheme of formula model prediction scheme can reduce amount of calculation.Generally speaking these researchs at present are all by wind Electric field is uniformly controlled, and does not consider the difference such as diverse location wind regime difference residing for unit.But discounting for poor between unit It is different, cause power distribution improper, necessarily cause total active power tracking effect not good.And current research have ignored unit sound Between seasonable, it is believed that unit active output moment is that can reach reference value, then by research emphasis be placed on power it is traceable under wind Airborne lotus control.So the research for the purpose of quickly tracking dispatching of power netwoks is very few.

Therefore, it is a kind of to be based on Model Predictive Control mode, the Wind turbines of residing different wind regime are classified and to having Work(adjustment amount carries out reasonable distribution, and the active power controller method for enabling wind power plant quickly to track dispatching of power netwoks order has weight Want meaning.

The content of the invention

It is an object of the invention to overcome the deficiencies in the prior art, there is provided a kind of wind power plant based on Model Predictive Control has Work(Poewr control method, makes wind power plant quick response, and avoids the unnecessary action of unit, realize to dispatching of power netwoks value with Track.

For achieving the above object, wind power station active power control method of the present invention based on Model Predictive Control, its It is characterised by, comprises the following steps：

(1), dispatched from the factory parameter according to blower fan, from incision wind speed v_inTo rated wind speed v_ratedCut-out wind speed v is arrived again_outBy wind-powered electricity generation Wind regime is classified residing for unit, further according to classification results, by the active power dispatch power from grid dispatching centerWith from public affairs The real-time active power output of wind power plant that linking point measurement is obtained altogetherThe active reference value distribution of wind farm level is carried out, makes every class wind-powered electricity generation Unit commitment is to active power reference valueWherein, i=1,2 ..., n, n represent Wind turbines batch total；

(2) equivalent modeling, is carried out according to class categories to sorted each class Wind turbines, then on equivalent modeling basis On, the model predictive controller of each class Wind turbines is designed, finally using model predictive controller pairRevised, obtained To active power value

(3), by active power valuePro rate of exerting oneself according to separate unit Wind turbines in such blower fan unit is to single Typhoon group of motors, makes every typhoon group of motors be assigned to output power valuem_iRepresent the i-th class wind turbine Total number of units of group；

(4), per typhoon group of motors according to the output power value being assigned toExported, so as to complete whole wind power plant Active power controller.

Wherein, in the step (1), active power reference value is assigned to per class Wind turbinesConcretely comprise the following steps：

(2.1), power pretreatment

Calculate the active power output per class Wind turbines

Wherein,Represent exerting oneself for jth typhoon group of motors in the i-th class Wind turbines, m_iRepresent the i-th class Wind turbines Total number of units；

Calculate whole wind power plant active power output

Wherein,It is the i-th class Wind turbines active power output, n represents Wind turbines batch total；

Calculate wind power plant and treat regulation and control amount Δ P：

Wherein,It is the active power dispatch power from grid dispatching center；

(2.2), distribute per class Wind turbines active power reference value

(2.2.1), as Δ P ＞ 0, then showSo per the active power reference value of class Wind turbinesAllocation step be：

1) the power per liter ability per class Wind turbines, is calculated：

2), according to default regulation and control sequentially, add up successively per class Wind turbines power per liter ability, whenWhen, it is allocated according to equation below, obtain the active reference of every class Wind turbines Value

Wherein,The ability of the i-th class Wind turbines power per liter is represented,Represent the i-th active output of class Wind turbines Maximum, t being represented and being met regulation goal by be added to successively t class Wind turbines；

(2.2.2), as Δ P ＜ 0, showSo per the active power reference value of class Wind turbines Allocation step be：

1) the drop power capability per class Wind turbines, is calculated：

2), according to default regulation and control sequentially, add up successively per class Wind turbines drop power capability, whenWhen, it is allocated according to equation below, obtain the active of every class Wind turbines Reference value

Wherein,The ability that power drops in the i-th class unit is represented,Represent the active output minimum value of the i-th class unit.

Further, in the step (2), the method that equivalent modeling is carried out according to class categories to each class Wind turbines For：

(3.1) wind regime residing for Wind turbines is divided into n classes in, setting step (1), wherein, low wind speed area unit accounts for k-1 classes, It is expressed as T₁,T₂,…T_k-1；Close on rated wind speed area unit and account for a class, be expressed as T_k；High wind speed area unit accounts for n-k classes, is expressed as T_k+1,T_k+2,…T_n；

(3.2) equivalence, is carried out to all kinds of Wind turbines

(3.2.1), premised on capacity equivalence, equivalence is carried out to all kinds of Wind turbines respectively；

1) equivalence, is carried out to low wind speed area unit：

1.1), wind speed equivalence

According to every typhoon group of motors power in wind speed-power function the i-th class Wind turbines of calculating：P_ij=F (vi_j)

Calculate the equivalent active power of the i-th Wind turbines

The equivalent wind speed of the i-th class Wind turbines is calculated using reverse method

Wherein, j represents jth typhoon group of motors, m in the i-th class Wind turbines_iRepresent total number of units of the i-th class Wind turbines, F (v_ij) it is wind speed-power function, v_ijRepresent the wind speed of jth typhoon group of motors in the i-th class Wind turbines；

1.2), torque is equivalent

Wherein,It is the equivalent torque of the i-th class Wind turbines, its value is equivalent to certain particular rack j in the i-th class^*Turn Square；

2), equivalence is carried out to closing on rated wind speed unit and high wind speed area unit；

2.1), wind speed equivalence：

2.2), torque is equivalent：

(3.3) state-space model, is set up according to equivalence

(3.3.1), by blower fan machine torqueLaunch in the first order Taylor of given stable operating point Approximately obtain：

Wherein, δ represents the deviation of variable and its stable operating point, and ρ is atmospheric density, and R is wind wheel radius, and v is wind speed, C_p (λ, β) is power efficiency, and β is propeller pitch angle, and λ is tip speed ratio,ω_tIt is rotation speed of fan, T_tCan regard as on wind Machine rotational speed omega_t, propeller pitch angle β and wind speed v function, upper lineRepresent its steady-state operation point；

(3.3.2), set up state-space model

Y=δ P_g=Cx

Wherein,D=δ v

Wherein, ω_gIt is motor speed, T_gIt is electromagnetic torque,Subscript * represents set-point, K_sAnd B_s The equivalent coefficient of elasticity and equivalent damped coefficient of rotary shaft, J are represented respectively_tAnd J_gRepresent that blower fan rotary inertia and motor turn respectively Dynamic inertia, τ_gRepresent electromechanical time constant and propeller pitch angle time constant respectively with τ, η is electric efficiency.

Further, in the step (2), the design cycle of model predictive controller is：

(4.1), by state-space model discretization, obtain：

X (k+1)=A ' x (k)+B_u′u(k)+B_d′d(k)

Y (k+1)=C ' x (k)

(4.2), design optimization object function and constraints

Optimization object function is：

Constraints is：

Wherein, n_c,n_pControl time domain and prediction time domain, Q are represented respectively_P, Q_R, Q_SWeight coefficient is represented respectively, The maximum of the i-th class Wind turbines electromagnetic torque and propeller pitch angle is represented respectively, and Represent respectively the i-th class Wind turbines electromagnetic torque set-point and propeller pitch angle to Definite value, it is known that sampling time and rate of change are constrainedBoth are multiplied can then be calculated Represent that the i-th class Wind turbines electromagnetic torque changes respectively The minimum value and maximum of rate, the minimum value and maximum of pitch rate.

What goal of the invention of the invention was realized in：

Wind power station active power control method of the present invention based on Model Predictive Control, wind speed will according to residing for Wind turbines Wind turbines are classified, and are set up the local linearization state-space model per class unit using equivalent modeling method and be based on This predictive controller that designs a model, it is considered to the dynamic response characteristic of inhomogeneity unit, by certain priority by dispatching of power netwoks value Distribute to every class unit, per class unit obtained by active reference value revised by model predictive controller again, revise value by Pro rate is to all units in such.The present invention is controlled to target with active power of wind power field, it is considered to which the operation between unit is poor It is different, and based on Model Predictive Control, make it quickly to track dispatching of power netwoks value.

Meanwhile, wind power station active power control method of the present invention based on Model Predictive Control also has following beneficial effect Really：

(1), compared with existing PI control methods, classified by residing wind speed difference by Wind turbines, realized Power is allocated by unit operation difference；

(2), compared with existing PI control methods, there is priority in the distribution of power, generally reduces unit not Necessary propeller pitch angle frequent movement (especially closing on the propeller pitch angle action of rated wind speed area unit), accelerates response speed, together When be conducive to extend unit durability；

(3) model predictive controller for, using, its essence is right on the basis of the future behaviour to process is predicted Controlled quentity controlled variable is optimized；Wherein predict for the model accuracy being based on without harsh requirement, this gives wind power plant this height The certain fault tolerant workspace of the nonlinear complex system modeling of rank；Optimization is in following one section of finite time, to be referred to by a certain performance Target optimizes to determine the control action in future, and optimization is to be repeated online, different excellent with the traditional overall situation Change；And Model Predictive Control is a kind of closed loop control algorithm, taking full advantage of reality output error carries out feedback compensation, therefore Can be well controlled effect.

Brief description of the drawings

Fig. 1 is the wind power station active power control method system block diagram based on Model Predictive Control；

Fig. 2 is existing PI control methods system block diagram；

Fig. 3 is wind conditions figure in wind power plant used by simulating, verifying；

Fig. 4 is tracking dispatching of power netwoks value effect contrast figure；

Fig. 5 is certain Tai Jin rated wind speeds area unit propeller pitch angle change comparison diagram；

Fig. 6 is unit active power output change comparison diagram.

Specific embodiment

Specific embodiment of the invention is described below in conjunction with the accompanying drawings, so as to those skilled in the art preferably Understand the present invention.Requiring particular attention is that, in the following description, when known function and design detailed description perhaps When can desalinate main contents of the invention, these descriptions will be ignored herein.

Embodiment

Fig. 1 is the wind power station active power control method system block diagram based on Model Predictive Control.

In the present embodiment, as shown in figure 1, wind power station active power control method of the present invention based on Model Predictive Control System architecture mainly include：First, the active reference value distribution of wind farm level；2nd, the reference value based on Model Predictive Control is revised； 3rd, value distribution is revised in active reference between unit.Three parts are described in detail below.

First, the active reference value distribution of wind farm level

The active reference value distribution of wind farm level is premised on computer-assisted classification.Because even in a wind power plant for very little Interior, due to factors such as geographical position, wind speed can not possibly be identical suffered by per Fans, present invention wind speed as suffered by Wind turbines Corresponding to different default wind speed intervals, it is some subclasses that all Wind turbines in wind power plant are incorporated into, but is pressed over all Control mode can be divided into three major types：

Low wind speed area unit：Such Wind turbines runs in incision wind speed between rated wind speed, and propeller pitch angle is maintained at 0 °, leads to Cross the torque of control blower fan and change power coefficient to adjust rotating speed and realize active power controller, inertia time constant is small, ring Answer speed fast.

Close on rated wind speed area unit：Such unit operation in the interval of very little near rated wind speed, due in wind Near speed-power curve flex point, disturbed its working condition by the wind speed of very little and be likely to occur from compared with big change, so as to draw Being continually changing for propeller pitch angle is played, causes the mechanical action that unit is unnecessary, increase the response time, so generally not by it Regulation is included, and is always the fixed active reference value of its distribution.

High wind speed area unit：Such Wind turbines runs in rated wind speed between cut-out wind speed, in rated speed perseverance work( Rate generating state, active power controller is realized by adjusting propeller pitch angle, and the change of propeller pitch angle is mechanical action, and inertia time is normal Number is larger, and response speed is slow；High wind speed area unit has reached rated value due to rotating speed, so suffered wind speed is higher, its output work Rate changes more sensitive to propeller pitch angle, that is, change an equal amount of angle, and the power of the assembling unit change higher of suffered wind speed is bigger.

In the present embodiment, wind regime residing for Wind turbines is divided into n classes, wherein, low wind speed area unit accounts for k-1 classes, table It is shown as T₁,T₂,…T_k-1；Close on rated wind speed area unit and account for a class, be expressed as T_k；High wind speed area unit accounts for n-k classes, is expressed as T_k+1,T_k+2,…T_n；

When wind power plant is receiving the active power dispatch power from grid dispatching centerMeasured with from common link point The real-time active power output of wind power plant for arrivingAfterwards, the deviation delta P of dispatching of power netwoks value and the real-time active power output of wind power plant, i.e. wind are obtained Electric field treats regulation and control amount, judges that wind power plant should carry out power per liter and still drop Power Control, according to the spy of every class unit response speed Levy, will by certain priorityEvery class unit is assigned to, so as to obtain the active power reference value of every class unitIn order to Accelerate wind power plant response speed, it is clear that should as few as possible adjust propeller pitch angle, then the priority for participating in regulation per class unit is suitable Sequence is first engaged in regulation for low wind speed area unit, by wind speed augment direction successively by minimum wind speed area unit, secondary low wind speed Area's unit, in low wind speed area wind speed includes active power regulation closest to certain class unit of rated wind speed, if active power Control targe is not yet reached, and high wind speed area unit is participated in regulation again, by wind speed reduction direction, successively by highest wind velocity area machine Group, secondary high wind speed area, in high wind speed area wind speed includes active power regulation closest to certain class unit of rated wind speed.Its In, after certain class unit participates in regulation, disclosure satisfy that dispatching of power netwoks requirement, then according to priority orders, unit after which has Work(output maintains initial value.

Active power reference value is assigned to every class Wind turbines belowSpecific steps be described in detail, specifically For：

(1) active power output per class Wind turbines, is calculated

Wherein,Represent exerting oneself for jth typhoon group of motors in the i-th class Wind turbines, m_iRepresent the i-th class Wind turbines Total number of units；

Calculate whole wind power plant active power output

Wherein,It is the i-th class Wind turbines active power output, n represents Wind turbines batch total；

Calculate wind power plant and treat regulation and control amount Δ P：

Wherein,It is the active power dispatch power from grid dispatching center；

(2), distribute per class Wind turbines active power reference value

(2.1), as Δ P ＞ 0, then showTracking dispatching of power netwoks value is represented, wind power plant is now needed Increase active power output, then per the active power reference value of class Wind turbinesAllocation step be：

1) the power per liter ability per class Wind turbines, is calculated：

2), according to default regulation and control sequentially, add up successively per class Wind turbines power per liter ability, whenWhen, it is allocated according to equation below, obtain the active reference of every class Wind turbines Value

Wherein,The ability of the i-th class Wind turbines power per liter is represented,Represent the i-th active output of class Wind turbines Maximum, t being represented and being met regulation goal by be added to successively t class Wind turbines；

(2.2), as Δ P ＜ 0, showTracking dispatching of power netwoks value is represented, wind power plant now needs Reduce active power output, then per the active power reference value of class Wind turbinesAllocation step be：

1) the drop power capability per class Wind turbines, is calculated：

2), according to default regulation and control sequentially, add up successively per class Wind turbines drop power capability, whenWhen, it is allocated according to equation below, obtain the active of every class Wind turbines Reference value

Wherein,The ability that power drops in the i-th class unit is represented,The active output minimum value of the i-th class unit is represented, T being represented and being met regulation goal by be added to successively t class Wind turbines.

2nd, the reference value based on Model Predictive Control is revised

For the predictive controller that designs a model, it is necessary to first equivalent modeling is carried out to every class Wind turbines, below with capacity etc. Premised on value, respectively all kinds of units are carried out with unit equivalence, equivalence method is divided into two kinds of situation explanations：

1) equivalence, is carried out to low wind speed area unit：

1.1), wind speed equivalence

According to every typhoon group of motors power P in wind speed-power function the i-th Wind turbines of calculating_ij：P_ij=F (v_ij)

Calculate the equivalent active power of the i-th Wind turbines

The equivalent wind speed of the i-th class Wind turbines is calculated using reverse method

Wherein, j represents jth typhoon group of motors, m in the i-th class Wind turbines_iRepresent total number of units of the i-th class Wind turbines, F (v_ij) it is wind speed-power function, v_ijRepresent the wind speed of jth typhoon group of motors in the i-th class Wind turbines；

As the unit that two typhoons speed is respectively 6m/s and 8m/s carries out wind speed equivalence, wind speed-power function isAtmospheric density ρ=1.2231kg/m³, wind wheel radius R=63m, because unit is located at low wind speed area, power Usage factor takes maximum C_p=C_pmax=0.482, v are wind speed.Two machines are first calculated according to wind speed-power function respectively The power of group, 0.793872MW and 1.881772MW.Ask it and obtain equivalent wind speed about further according to the anti-solution of wind speed-power function It is 9m/s.

1.2), torque is equivalent

Wherein,It is the equivalent torque of the i-th class Wind turbines, its value is equivalent to certain particular rack j in the i-th class^*Turn Square；

2), equivalence is carried out to closing on rated wind speed area unit and high wind speed area unit；

2.1), wind speed equivalence：

2.2), torque is equivalent：

3) state-space model, is set up according to equivalence

3.1) required precision, due to Model Predictive Control to forecast model is not high, so directly by blower fan machine torqueFirst order Taylor expansion is carried out in given stable operating point, by T_tThe partial line near stable operating point Propertyization is approximately obtained：

Wherein, δ represents the deviation of variable and its stable operating point, and ρ is atmospheric density, and R is wind wheel radius, and v is wind speed, C_p (λ, β) is power efficiency, and β is propeller pitch angle, and λ is tip speed ratio,ω_tIt is rotation speed of fan, T_tCan regard as on wind Machine rotational speed omega_t, propeller pitch angle β and wind speed v function, upper lineRepresent its steady-state operation point；

3.2) state-space model, is set up

Y=δ P_g=Cx

Wherein,D=δ v

Wherein, ω_gIt is motor speed, T_gIt is electromagnetic torque,Subscript * represents set-point, K_sAnd B_s The equivalent coefficient of elasticity and equivalent damped coefficient of rotary shaft, J are represented respectively_tAnd J_gRepresent that blower fan rotary inertia and motor turn respectively Dynamic inertia, τ_gRepresent electromechanical time constant and propeller pitch angle time constant respectively with τ, η is electric efficiency.

4), designed a model predictive controller further according to state-space model, its design cycle is：

4.1), by state-space model discretization, obtain：

X (k+1)=A ' x (k)+B_u′u(k)+B_d′d(k)

Y (k+1)=C ' x (k)

4.2), design optimization object function and constraints

Optimization object function is：

Constraints is：

Wherein, n_c,n_pControl time domain and prediction time domain, Q are represented respectively_P, Q_R, Q_SWeight coefficient is represented respectively, The maximum of the i-th class Wind turbines electromagnetic torque and propeller pitch angle is represented respectively, and Represent respectively the i-th class Wind turbines electromagnetic torque set-point and propeller pitch angle to Definite value in two differences of sampling instant,Represent respectively the electromagnetic torque set-point and propeller pitch angle of the i-th class Wind turbines to Definite value, it is known that sampling time and rate of change are constrainedBoth are multiplied can then be calculated Represent that the i-th class Wind turbines electromagnetic torque changes respectively The minimum value and maximum of rate, the minimum value and maximum of pitch rate.

3rd, value distribution is revised in active reference between unit

It is distributional equity, the every class machine that will be obtained by model predictive controller between the unit for ensureing similar running status That organizes active reference value revises valuePro rate of exerting oneself according to separate unit Wind turbines in such blower fan unit gives separate unit wind Group of motors, makes every typhoon group of motors be assigned to output power value

Fig. 2 is existing PI control methods system block diagram.

Deviation between the active output of wind power plant that the method is obtained according only to dispatching of power netwoks value and common link point measurement, Proportional, integral control mode is taken to be controlled, then active reference value is distributed to each unit by the use ratio method of salary distribution. The method does not account for the difference between unit, does not have priority during regulation, and the active reference value of all units can be with electricity The fluctuation of net dispatch value and fluctuate, part of generating units can be caused, especially close on the propeller pitch angle frequent movement of rated wind speed area unit, The response time certainly will be extended, be both unfavorable for that wind power plant quickly tracks dispatching of power netwoks value, also damage unit durability, increase wind power plant into This.

Embodiment

In order to illustrate technique effect of the invention, active power controller method of the invention is applied to every capacity It is 5MW, the wind power plant for possessing 14 typhoon group of motors carries out simulating, verifying.And wind conditions such as Fig. 3 that false wind electric field measurement is arrived It is shown, wind speed v (is cut according to the fan parameter for emulating used_in=3m/s, rated wind speed v_rated=11.4m/s, cuts out wind Fast v_out=25m/s) wind speed is divided into 5 subclasses, as shown in table 1,14 units are incorporated into 5 by initial wind speed in wind power plant In subclass, computer-assisted classification situation is as shown in table 2.It is with dispatching of power netwoks value changes 50MW-42MW-36MW-40MW-45MW-52MW The validity of example explanation the inventive method.

Table 1 presses wind speed and divides 5 subclasses

Computer-assisted classification situation in the wind power plant of table 2

Fig. 4 is tracking dispatching of power netwoks value effect contrast figure.From fig. 4, it can be seen that wind power plant is carried out in the present inventive method Active power controller, the more existing PI control methods of response speed are faster.Active power of wind power field is defined to reach and then stablize Time in the range of the error band of dispatching of power netwoks value ± 5% is regulating time, and this index of the invention is 1.27s, and PI This of control method index is 2.4s, and the present invention improves 47.2% in rapidity than existing methods.

Tu5Shi Jin rated wind speeds area unit propeller pitch angle change comparison diagram.From fig. 5, it can be seen that wind power plant is by of the invention Method carries out active power controller, to the active reference value in every unit commitment in rated wind speed area unit is closed on all the time It is 5MW so that its propeller pitch angle remains at 0 °, is conducive to wind power plant quick response；And PI control methods do not account for such The sensivity feature that unit is disturbed to wind speed, constantly gives its distribution power reference value, causes propeller pitch angle frequently to change, and is unfavorable for Wind power plant quick response.

Fig. 6 is unit active power output change comparison diagram.As can be seen from Figure 6 under two methods, the power of five class units becomes Change trend.Described in six performance numbers of the sampling instant of stable operation with it.Six moment are respectively 5s, 15s, 25s, 35s, 45s, 55s, 65s.First three root block diagram per class describes (electricity when wind power plant tracks the dispatching of power netwoks power being gradually reduced Net dispatch value is mutated to 42MW from initial 50MW and is mutated again to 36MW), per the situation of exerting oneself of class unit.When dispatch value is from initial 50MW is mutated during to 42MW, and active power controller, the first kind unit and Equations of The Second Kind unit in low wind speed area are carried out using the present invention Take the lead in minimizing value, because not yet meeting drop quantity of power, the 5th class unit that high wind speed area is in highest wind velocity participates in tune In section, and the 5th class unit still can not meet regulated quantity after minimizing power, and subsequent high wind speed area is in time high wind speed 4th class unit also reduce exert oneself participate in regulation in, this process can be by first of every class unit to second block diagram Find out；When dispatching of power netwoks value is mutated to 36MW from 42MW, due to first and second and five class unit be all down to setting most Small-power value, only the 4th class unit also have regulation surplus, and the 4th class unit continues reduction and exerts oneself, and completes tracking target, this mistake Journey can be found out from second to the 3rd block diagram of every class unit.It is similar with drop power, in power-up process, according to this hair Bright method carries out active power controller, and every class unit is participated in regulation still according to default priority orders, i.e. the first kind, Equations of The Second Kind, the 5th class, the 4th class unit is acted successively, meets Expected Results.And had using every class unit of PI control methods Work(output can all be continually changing with the change of dispatching of power netwoks value, and the stability of a system is not good enough.

Although being described to illustrative specific embodiment of the invention above, in order to the technology of the art Personnel understand the present invention, it should be apparent that the invention is not restricted to the scope of specific embodiment, to the common skill of the art For art personnel, as long as various change is in appended claim restriction and the spirit and scope of the present invention for determining, these Change is it will be apparent that all utilize the innovation and creation of present inventive concept in the row of protection.

Claims (4)

1. a kind of wind power station active power control method based on Model Predictive Control, it is characterised in that comprise the following steps：

(1), dispatched from the factory parameter according to blower fan, from incision wind speed v_inTo rated wind speed v_ratedCut-out wind speed v is arrived again_outBy Wind turbines Residing wind regime is classified, further according to classification results, by the active power dispatch power from grid dispatching centerWith from public chain The real-time active power output of wind power plant that contact measurement is obtainedThe active reference value distribution of wind farm level is carried out, makes every class Wind turbines It is assigned to active power reference valueWherein, i=1,2 ..., n, n represent Wind turbines batch total；

(2) equivalent modeling, is carried out according to class categories to sorted each class Wind turbines, then on the basis of equivalent modeling, The model predictive controller of each class Wind turbines is designed, finally using model predictive controller pairRevised, had Work(performance number P_i ^ref；

(3), by active power value P_i ^refPro rate of exerting oneself according to separate unit Wind turbines in such blower fan unit gives separate unit wind Group of motors, makes every typhoon group of motors be assigned to output power valueJ=1,2 ..., m_i, m_iRepresent the total of the i-th class Wind turbines Number of units；

(4), per typhoon group of motors according to the output power value being assigned toExported, so as to complete having for whole wind power plant Work(Power Control.

2. the wind power station active power control method based on Model Predictive Control according to claim 1, it is characterised in that In the step (1), active power reference value is assigned to per class Wind turbinesConcretely comprise the following steps：

(2.1), power pretreatment

Calculate the active power output P per class Wind turbines_i ^out：

Wherein,Represent exerting oneself for jth typhoon group of motors in the i-th class Wind turbines, m_iRepresent the head station of the i-th class Wind turbines Number；

Calculate whole wind power plant active power output

Wherein, P_i ^outIt is the i-th class Wind turbines active power output, n represents Wind turbines batch total；

Calculate wind power plant and treat regulation and control amount Δ P：

Wherein,It is the active power dispatch power from grid dispatching center；

(2.2), distribute per class Wind turbines active power reference value

(2.2.1), as Δ P ＞ 0, then showSo per the active power reference value of class Wind turbines's Allocation step is：

1) the power per liter ability per class Wind turbines, is calculated：ΔP_i ^up=P_i ^max-P_i ^out

2), according to default regulation and control sequentially, add up successively per class Wind turbines power per liter ability, when When, it is allocated according to equation below, obtain the active reference value of every class Wind turbines

$\stackrel{\&OverBar;}{P_i^{ref}}=\left\{\begin{array}{cc}{P}_i^{\max },& 1\&le;i<t\\ P_t^{out}+\mathrm{\&Delta;}P-\left({\mathrm{\&Delta;P}}_{t-1}^{up}+{\mathrm{\&Delta;P}}_{t-2}^{up}+\mathrm{...}{\mathrm{\&Delta;P}}_1^{up}\right),& i=t\\ P_i^{out},& t<i\&le;n\end{array}\right.;$

Wherein, Δ P_i ^upRepresent the ability of the i-th class Wind turbines power per liter, P_i ^maxRepresent that the active output of the i-th class Wind turbines is maximum Value, t being represented and being met regulation goal by be added to successively t class Wind turbines；

(2.2.2), when Δ P >=0, showSo per the active power reference value of class Wind turbinesPoint It is with step：

1) the drop power capability per class Wind turbines, is calculated：ΔP_i ^down=P_i ^out-P_i ^min

2), according to default regulation and control sequentially, add up successively per class Wind turbines drop power capability, when When, it is allocated according to equation below, obtain the active reference value of every class Wind turbines

$\stackrel{\&OverBar;}{P_i^{ref}}=\left\{\begin{array}{cc}{P}_i^{\min },& 1\&le;i<t\\ P_t^{down}-\&lsqb;\left|\mathrm{\&Delta;}P\right|-\left({\mathrm{\&Delta;P}}_{t-1}^{down}+{\mathrm{\&Delta;P}}_{t-2}^{down}+\mathrm{...}{\mathrm{\&Delta;P}}_1^{down}\right)\&rsqb;,& i=t\\ P_i^{out},& t<i\&le;N\end{array}\right.;$

Wherein, Δ P_i ^downRepresent the ability that power drops in the i-th class unit, P_i ^minRepresent the active output minimum value of the i-th class unit.

3. the wind power station active power control method based on Model Predictive Control according to claim 1, it is characterised in that In the step (2), it is according to the method that class categories carry out equivalent modeling to each class Wind turbines：

(3.1) wind regime condition residing for Wind turbines is divided into n classes in, setting step (1), wherein, low wind speed area unit accounts for k-1 classes, table It is shown as T₁,T₂,…T_k-1；Close on rated wind speed unit and account for a class, be expressed as T_k；High wind speed area unit accounts for n-k classes, is expressed as T_k+1, T_k+2,…T_n；

(3.2) equivalence, is carried out to all kinds of Wind turbines

(3.2.1), premised on capacity equivalence, equivalence is carried out to all kinds of Wind turbines respectively；

1) equivalence, is carried out to low wind speed area unit：

1.1), wind speed equivalence

According to every typhoon group of motors power in wind speed-power function the i-th Wind turbines of calculating：P_ij=F (v_ij)

Calculate the equivalent active-power P of the i-th Wind turbines_i ^eq：

The equivalent wind speed of the i-th class Wind turbines is calculated using reverse method

Wherein, j represents jth typhoon group of motors, m in the i-th class Wind turbines_iRepresent total number of units of the i-th class Wind turbines, F (v_ij) It is wind speed-power function, v_ijRepresent the wind speed of jth typhoon group of motors in the i-th class Wind turbines；

1.2), torque is equivalent

$T_{i,eq}^{wt}=T_{ij*}^{wt}$

Wherein,It is the equivalent torque of the i-th class Wind turbines, its value is equivalent to certain particular rack j in the i-th class^*Torque；

2), equivalence is carried out to closing on rated wind speed unit and high wind speed area unit；

2.1), wind speed equivalence：

2.2), torque is equivalent：

(3.3) state-space model, is set up according to equivalence

(3.3.1), by blower fan machine torqueLaunch approximate in the first order Taylor of given stable operating point Obtain：

${\mathrm{\&delta;T}}_t=K_{\mathrm{\&omega;}}({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}){\mathrm{\&delta;\&omega;}}_t+K_v({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}})\mathrm{\&delta;}v+K_{\mathrm{\&beta;}}({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}})\mathrm{\&delta;}\mathrm{\&beta;}$

$\left.\begin{array}{c}{K}_{\mathrm{\&omega;}}\left({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}\right){\mathrm{\&delta;\&omega;}}_t=\frac{\&part;T_t}{\&part;{\mathrm{\&omega;}}_t}{|}_{\left({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}\right)}\\ K_v\left({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}\right)\mathrm{\&delta;}v=\frac{\&part;T_t}{\&part;v}{|}_{\left({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}\right)}\\ K_{\mathrm{\&beta;}}\left({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}\right)\mathrm{\&delta;}\mathrm{\&beta;}=\frac{\&part;T_t}{\&part;\mathrm{\&beta;}}{|}_{\left({\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_t,\stackrel{\&OverBar;}{v},\stackrel{\&OverBar;}{\mathrm{\&beta;}}\right)}\end{array}\right\}$

Wherein, δ represents the deviation of variable and its stable operating point, and ρ is atmospheric density, and R is wind wheel radius, and v is wind speed, C_p(λ, β) It is power efficiency, β is propeller pitch angle, and λ is tip speed ratio,ω_tIt is rotation speed of fan, T_tCan regard as and turn on blower fan Fast ω_t, propeller pitch angle β and wind speed v function, upper lineRepresent its steady-state operation point；

(3.3.2), set up state-space model

$\stackrel{\&CenterDot;}{x}=Ax+B_{u}u+B_{d}d$

Y=δ P_g=Cx

Wherein,D=δ v

$A=\left[\begin{array}{ccccc}0& 1& -1& 0& 0\\ -K_S/J_t& (K_{{\mathrm{\&omega;}}_t}-B_S)/J_t& B_S/J_t& 0& K_{\mathrm{\&beta;}}/J_t\\ K_S/J_g& B_S/J_g& -B_S/J_g& -1/J_g& 0\\ 0& 0& 0& -1/{\mathrm{\&tau;}}_g& 0\\ 0& 0& 0& 0& -1/\mathrm{\&tau;}\end{array}\right]$

$\begin{array}{cc}{B}_u=\left[\begin{array}{cc}0& 0\\ 0& 0\\ 0& 0\\ 1/{\mathrm{\&tau;}}_g& 0\\ 0& 1/\mathrm{\&tau;}\end{array}\right]& B_d=\left[\begin{array}{c}0\\ K_v/J_t\\ 0\\ 0\\ 0\end{array}\right]\end{array}$

$C=\left[\begin{array}{ccccc}0& 0& \mathrm{\&eta;}{\stackrel{\&OverBar;}{T}}_g& \mathrm{\&eta;}{\stackrel{\&OverBar;}{\mathrm{\&omega;}}}_g& 0\end{array}\right]$

Wherein, ω_gIt is motor speed, T_gIt is electromagnetic torque,Subscript * represents set-point, K_sAnd B_sRespectively Represent the equivalent coefficient of elasticity and equivalent damped coefficient of rotary shaft, J_tAnd J_gRepresent that blower fan rotary inertia and motor rotate used respectively Amount, τ_gRepresent electromechanical time constant and propeller pitch angle time constant respectively with τ.

4. the wind power station active power control method based on Model Predictive Control according to claim 1, it is characterised in that In the step (2), the design cycle of model predictive controller is：

(4.1), by state-space model discretization, obtain：

X (k+1)=A ' x (k)+B_u′u(k)+B_d′d(k)

Y (k+1)=C ' x (k)

(4.2), design optimization object function and constraints

Optimization object function is：

Constraints is：

Wherein, n_c,n_pControl time domain and prediction time domain, Q are represented respectively_P, Q_R, Q_SWeight coefficient is represented respectively,Respectively The maximum of the i-th class Wind turbines electromagnetic torque and propeller pitch angle is represented, and Represent respectively the i-th class Wind turbines electromagnetic torque set-point and propeller pitch angle to Definite value in two differences of sampling instant,The electromagnetic torque set-point of the i-th class Wind turbines is represented respectively, it is known that during sampling Between and rate of change constraintBoth are multiplied can then be calculated The minimum value and maximum of the i-th class Wind turbines electromagnetic torque, the minimum of propeller pitch angle are represented respectively Value and maximum.