CN103281022A - Double-efficiency fuzzy optimization control method for doubly-fed wind generator - Google Patents

Abstract

A double-efficiency fuzzy optimization control method for a doubly-fed wind generator is characterized in that after the wind generator is combined to a grid and before the wind generator reaches an allowed highest rotation speed, maximum wind energy tracing and controlling are adopted to control the wind generator, in the process of the maximum wind energy tracing and controlling, a control mode is switched to optimal idle work searching control at a proper moment, and accordingly double-optimization on the operation efficiency of the generator is achieved. The double-efficiency fuzzy optimization control method has the advantages that in the process of the maximum wind energy tracing and controlling, searching accuracy of a fuzzy reasoning rule list can be improved through changes of quantization factors of input variables of the fuzzy reasoning rule list, occurrence of adjusting dead bands is avoided, and a use ratio of wind energy by the wind generator is improved. In the process of optimal idle work tracing, vibration can be effectively restrained, the searching accuracy is improved, loss of the wind generator is reduced, and active output is improved.

Description

The control method of the dual efficient fuzzy optimization of double-fed wind power generator

Technical field

The present invention relates to a kind of wind-driven generator control technology, relate in particular to the control method of the dual efficient fuzzy optimization of a kind of double-fed wind power generator.

Background technology

Double-fed wind power generator is carried out dual efficient Fuzzy Optimal Control, can realize simultaneously that maximal wind-energy is followed the trail of and the dual control target of generator lowest loss operation, this is significant to the energy output that increases unit, therefore, do further to further investigate significant to operational efficiency and the overall operation level of wind-powered electricity generation unit.

Usually, with reference to carrying out indirect rotating speed control, can realize that the maximal wind-energy under not having wind speed measures is followed the trail of by dynamic setting generator unit stator active power.And the setting of generator unit stator active power is relevant with the characteristic of blower fan.In practice, parameter such as fan parameter and atmospheric density can influence fan characteristic.Generator unit stator active power in unit operation is set will depart from optimal value, the tracking effect of the wind energy that has the greatest impact all the time.Therefore, adopt fuzzy logic search blower fan optimized rotating speed, realize that the maximal wind-energy under the no wind speed measurement is followed the trail of a kind of method preferably of can yet be regarded as; On the other hand, at utmost reduce the double-fed wind power generator loss as reactive power reference qref when choosing

Principle the time because

Will relate to a plurality of parameters such as generator resistance, inductance in the calculation expression, its measurement is subjected to accuracy limitations, and resistance is subjected to Temperature Influence obvious again, so this strategy is difficult to really to reach the generator loss in actual motion minimum.And adopt fuzzy logic to search for the OPTIMAL REACTIVE POWER of generator, and can not rely on the accuracy of generator parameter, have the good conformity ability.Therefore, can adopt the fuzzy logic strategy to optimize generator speed and reactive power setting respectively, realize that no wind speed is measured maximal wind-energy tracking down and the generator lowest loss is moved double goal, thereby avoid optimum reference value setting for the dependence of image parameter accuracy, this is the advantage place that utilizes the artificial intelligence decision-making.

On the one hand, utilizing fuzzy logic to realize when putting with the approaching the best of certain step-size in search, when little deviation range, still existing one and regulating the dead band in the maximal wind-energy track-while-scan optimized rotating speed reference value process.Cause the rotating speed search is formed limit cycles oscillations near the best point, meanwhile power output also will be vibrated near maximum power point, limited the high accuracy search and location of fuzzy logic to optimized rotating speed reference value and maximum power, can bring the long-term vibration of drive shaft system thus, this will cause bigger infringement to the blower fan mechanical part.This is the defective that the fuzzy logic search exists, and therefore need take appropriate measures to improve search accuracy; Simultaneously, double-fed wind power generator is realized that maximal wind-energy is followed the trail of and lowest loss is moved as dual control target here, needed to solve the problem of coordinating between the two.Because only after having finished the maximal wind-energy track-while-scan, just can proceed the search of next step OPTIMAL REACTIVE POWER reference value.The measure of the raising search precision of therefore, taking also should have simply, characteristics fast.After finishing the maximal wind-energy search rapidly, can enter next step very soon to the search of OPTIMAL REACTIVE POWER reference value.

On the other hand, in the process to OPTIMAL REACTIVE POWER search, currently adopt the search of single fuzzy logic also to exist the search oscillation problem.How can search high-precision OPTIMAL REACTIVE POWER value fast, to reduce the search vibration, make the loss minimum of generator self, further improve generator operation efficient, this also is the problem that single fuzzy logic search need solve.

Summary of the invention

At the problem in the background technology, the present invention proposes the control method of the dual efficient fuzzy optimization of a kind of double-fed wind power generator, comprise the wind-driven generator that adopts the control of double-fed control system; When wind speed more than or equal to incision during wind speed, wind-driven generator generates electricity by way of merging two or more grid systems, increase along with wind speed, when wind-driven generator reaches maximum allowable speed, enter permanent rotating speed generating state, it is characterized in that: the back that begins to generate electricity by way of merging two or more grid systems, wind-driven generator reach before the maximum allowable speed, adopt following control method that wind-driven generator is controlled:

When wind-driven generator has just begun to generate electricity by way of merging two or more grid systems, carry out earlier the maximal wind-energy Tracing Control as follows:

1) when wind speed more than or equal to incision during wind speed, wind-driven generator generates electricity by way of merging two or more grid systems, and rotating speed and the active power of wind-driven generator are carried out continuous sampling, in each sampling period, the changing value of change in rotational speed value and active power is calculated; If the rotation speed change value that is recorded in the single sampling period is Δ ω _Ri, the active power changing value that is recorded in the single sampling period is Δ p _Ei, i is sampling number, i=1,2,3,4 ... n; If Δ ω _R1With Δ p _E1Be on the occasion of;

2) in second sampling period, with Lp _E1With Δ ω _R1Two input variables as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value, the double-fed control system is regulated the wind-driven generator rotating speed according to optimizing tachometer value; Enter step 3); Second sampling period also namely formed first control cycle;

Wherein, Lp _E1Be corresponding k _pWith Δ p _E1The active power input variable, Lp _E1=k _pΔ p _E1, k _pBe corresponding Δ p _EiThe input variable quantizing factor;

3) subsequent control is in the cycle, with current Lp _EiWith two input variables of the theoretical rotational speed changing value that obtains in the last control cycle as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value; The double-fed control system according to the optimization tachometer value of correspondence to the wind-driven generator rotating speed carry out continuously, dynamic adjustments; Enter step 4);

Wherein, Lp _Ei=k _pΔ p _EiLp _EiBe corresponding k _pWith Δ p _EiThe active power input variable;

4) in the control procedure of step 3), with the Δ ω of current sampling period correspondence _RiAbsolute value | Δ ω _Ri| with Δ ω _R1/ k ₁Compare in real time: if | Δ ω _Ri| greater than Δ ω _R1/ k ₁, then return step 3); If | Δ ω _Ri| be less than or equal to Δ ω _R1/ k ₁, then enter step 5);

Wherein, k ₁Be the regulatory factor that data rule of thumb obtain, k ₁Value between 2.5～3.5;

5) with current Lp _Ei ^*With two input variables of the theoretical rotational speed changing value that obtains in the last control cycle as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value; The double-fed control system according to the optimization tachometer value of correspondence to the wind-driven generator rotating speed carry out continuously, dynamic adjustments; Enter step 6);

Wherein, Lp _Ei ^*=Δ p _EiK _y, Lp _Ei ^*Be corresponding k _yWith Δ p _EiThe active power input variable, k _y=k _pK ₁k _yBe the corresponding Δ p after regulating _EiThe input variable quantizing factor;

6) with current sampling period correspondence | Δ ω _Ri| with ε _WminCompare in real time: if | Δ ω _Ri| more than or equal to ε _Wmin, then return step 5); If | Δ ω _Ri| less than ε _Wmin, then switch to OPTIMAL REACTIVE POWER search control by the maximal wind-energy Tracing Control;

Wherein, ε _WminCritical switching value for corresponding maximal wind-energy Tracing Control;

After entering OPTIMAL REACTIVE POWER search control, control as follows:

1] reactive power and the active power of wind-driven generator are carried out continuous sampling, each reactive power was calculated the changing value of reactive power and the changing value of active power in the sampling period; If the reactive power changing value that single reactive power was recorded in the sampling period is Δ Q _Sv, the active power changing value that single reactive power was recorded in the sampling period is Δ p _v, v is sampling number, v=1,2,3,4 ... n;

2] second reactive power is in the sampling period, with Δ Q _S1With Δ f ₁Two input variables as the second fuzzy inference rule table, obtain theoretical idle changing value according to the second fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system is regulated the reactive power of wind-driven generator according to optimizing reactive power value; Enter step 3]; The second reactive power sampling period also namely formed the first idle control cycle;

Wherein, Δ f ₁Be corresponding Δ p ₁The active power input variable, Δ f ₁=-Δ p ₁K _f, k _fBe corresponding Δ p _vThe input variable quantizing factor;

3] in the follow-up idle control cycle, with current Δ f _vWith two input variables of the theoretical idle changing value that obtains in the last idle control cycle as the second fuzzy inference rule table, obtain theoretical idle changing value according to the second fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system according to optimize reactive power value to the reactive power of wind-driven generator carry out continuously, dynamic adjustments; Enter step 4];

Wherein, Δ f _vBe corresponding Δ p _vThe active power input variable, Δ f _v=-Δ p _vK _f

4] in step 3] control procedure in, in each idle control cycle with current Δ Q _SvAbsolute value | Δ Q _Sv| with k ₂| Δ Q _S1| compare in real time:

If | Δ Q _Sv| greater than k ₂| Δ Q _S1|, then continue Δ p _vAbsolute value | Δ p _v| with ε _W1Compare: if | Δ p _v| greater than ε _W1, then stop OPTIMAL REACTIVE POWER search control, simultaneously, the double-fed control system is regulated the reactive power of wind-driven generator according to the reactive power rated value of setting, and switches to the maximal wind-energy Tracing Control; If | Δ p _v|≤ε _W1, then return step 3];

If | Δ Q _Sv| be less than or equal to k ₂| Δ Q _S1|, then enter step 5];

Wherein, k ₂Be the regulatory factor that data rule of thumb obtain, k ₂Value between 0.3～0.4;

5] with current Δ f _vWith two input variables of the theoretical idle changing value that obtains in the last idle control cycle as the 3rd fuzzy inference rule table, obtain theoretical idle changing value according to the 3rd fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system according to optimize reactive power value to the reactive power of wind-driven generator carry out continuously, dynamic adjustments; Enter step 6];

6] in step 5] control procedure in, in each idle control cycle will | Δ p _v| with ε _W1Compare in real time:

If satisfy | Δ p _v|＞ε _W1Condition, then stop OPTIMAL REACTIVE POWER search control, simultaneously, the double-fed control system is regulated the reactive power of wind-driven generator according to the reactive power rated value of setting, and switches to the maximal wind-energy Tracing Control; If satisfy | Δ p _v|≤ε _W1Condition, then return step 5];

Wherein, ε _W1Critical switching value for corresponding OPTIMAL REACTIVE POWER search control.

The present invention and different being of existing fuzzy efficiency optimization method:

On the one hand, in realizing the process that maximal wind-energy is followed the trail of, when when searching for the optimization tachometer value that finds generally and approach to the optimized rotating speed reference value gradually, the present invention can be under the condition that does not increase fuzzy inference rule number and amount of calculation, only by increasing input variable quantizing factor k _p, improve control system to the total meritorious input increment Delta p of generator stator-rotator _EiResolution, make that control system can be to small Δ p _EiMake a response, increase its control action, reduce the search vibration, obtain the higher optimized rotating speed reference value of precision; Simultaneously, this solution is simple, after finishing the processing that maximal wind-energy is followed the trail of rapidly, enter soon the OPTIMAL REACTIVE POWER search the processing stage, can solve better that maximal wind-energy is followed the trail of and OPTIMAL REACTIVE POWER is searched between the two coordination problem.

On the other hand, in realizing the OPTIMAL REACTIVE POWER search procedure, at adopting single fuzzy logic search to have the problem of search vibration, the present invention adopts two-stage fuzzy logic way of search: with coarse adjustment fuzzy logic (being the second fuzzy reasoning table in the preamble) to accelerate search speed, with the accuracy of fine tuning fuzzy logic (being the 3rd fuzzy reasoning table in the preamble) with the raising search, solved the problem that has contradiction in the search of single fuzzy logic between the rapidity and accuracy that adopts, make deciding of generator self, the rotor copper loss minimum more is conducive to the meritorious output of generator and further improves.

Useful technique effect of the present invention is: in the maximal wind-energy tracing process, can pass through the quantizing factor of the input variable of change fuzzy inference rule table, improve the search precision of fuzzy inference rule table, avoid the dead band occurring regulating, improve wind-driven generator to utilization ratio of wind energy; In the OPTIMAL REACTIVE POWER tracing process, can effectively suppress flutter, improve search precision, reduce the own loss of wind-driven generator, improve meritorious output.

Description of drawings

Fig. 1, dual-feedback wind power generator control system principle schematic;

Fig. 2, double-fed wind power generator universe section control principle figure;

Wind energy conversion system output mechanical power and rotation speed relation figure under Fig. 3, the different wind speed;

Fig. 4, valuation functions f with

Graph of relation;

The fuzzy controller principle schematic of Fig. 5, the realization first fuzzy inference rule table;

Fig. 6, Lp _EiFuzzy membership functions;

The fuzzy controller schematic diagram of Fig. 7, the realization second fuzzy inference rule table and the 3rd fuzzy inference rule table;

The input variable Δ f of Fig. 8, the second fuzzy inference rule table _vFuzzy membership functions;

The input variable Δ Q of Fig. 9, the second fuzzy inference rule table _SvFuzzy membership functions;

The output variable Δ Q of Figure 10, the second fuzzy inference rule table _sFuzzy membership functions;

The input variable Δ f of Figure 11, the 3rd fuzzy inference rule table _vFuzzy membership functions;

The input variable Δ Q of Figure 12, the 3rd fuzzy inference rule table _SvFuzzy membership functions;

The output variable Δ Q of Figure 13, the 3rd fuzzy inference rule table _sFuzzy membership functions;

Figure 14, logic diagram of the present invention.

Embodiment

Be used for the system of double-fed wind power generator control as shown in Figure 1, this system adopts based on stator magnetic linkage oriented vector control system, realize the decoupling zero between generator electromagnetic torque and the rotor-exciting, after removing the cross-couplings item that is caused by back electromotive force through feedforward compensation again, can control generator amature magnetic linkage and electromagnetic torque respectively by d, the q axle component of regulating rotor voltage.

In the actual environment, the blade rotating speed changes with wind speed, and the running status of wind-driven generator can change with the blade rotation speed change again; When wind speed less than incision during wind speed, wind-driven generator does not generate electricity by way of merging two or more grid systems, when wind speed during more than or equal to the incision wind speed, wind-driven generator generates electricity by way of merging two or more grid systems; In the maximal wind-energy tracing Area, the rotating speed of wind-powered electricity generation unit is done corresponding the variation with wind speed, remains maximum with the power coefficient of guaranteeing wind energy conversion system.At this moment, the wind energy conversion system control subsystem is carried out fixed pitch control, and the generator control subsystem is regulated the rotating speed of unit by the power output of generator, realizes the variable speed constant frequency operation; When permanent rotating speed district, the wind-powered electricity generation unit has reached maximum speed, but the power output of wind energy conversion system does not reach the nominal operation state as yet, be protection unit nonoverload, no longer carry out the tracking of maximal wind-energy, but control to regulate propeller pitch angle by the variable pitch of wind energy conversion system control subsystem, guarantee at the permanent rotating speed generator operation that allows on the maximum (top) speed; Along with the increase wind energy conversion system output mechanical power of wind speed constantly increases, generator reaches its power limit.Need this moment the power output of control unit to make it to be no more than rated value, the wind-powered electricity generation unit is in permanent rotating speed, output-constant operation state.

The inventive method at the control section be wind-driven generator begin to generate electricity by way of merging two or more grid systems after to the maximal wind-energy tracking section that enters before the permanent rotating speed generating state, the universe section control system principle that control strategy of the present invention (being the dual efficient Fuzzy Optimal Control among Fig. 2) and permanent rotating speed control strategy and permanent power control strategy constitute can be switched by a comprehensive coordination controller corresponding to the control strategy of the different running sections of wind-driven generator as shown in Figure 2; At permanent power control strategy and permanent rotating speed control strategy, ripe scheme is all arranged in the prior art, the present invention is only at the dual efficient Fuzzy Optimal Control district among Fig. 2; Core thinking of the present invention is: among Fig. 1

Value (the optimal Generator rotating speed is with reference to set-point) adopt optimization tachometer value in the maximal wind-energy Tracing Control of the present invention, realizing the tracking to maximal wind-energy, among Fig. 1

Value (OPTIMAL REACTIVE POWER set-point) adopt optimization reactive power value in the OPTIMAL REACTIVE POWER search control of the present invention, realizing the operation of wind-driven generator lowest loss, thereby realize the control target of dual efficient fuzzy optimization.

In order to make those skilled in the art understand the solution of the present invention better, existing principle to the tracking of the maximal wind-energy among the present invention and OPTIMAL REACTIVE POWER tracking is explained respectively:

1) the maximal wind-energy principle of following the trail of: referring to Fig. 3, there is shown the relation of wind-driven generator output mechanical power and blade rotating speed under the different wind friction velocities; Wherein, P _mBe the mechanical output of blower fan output, ω _wBe the angular speed of blade rotation, v ₁, v ₂, v ₃Be wind speed, and v ₁＞v ₂＞v ₃As can be seen from Figure, under same wind friction velocity, when one timing of blade propeller pitch angle, the corresponding power coefficient of the different rotating speeds of blade there are differences, this species diversity makes that finally the mechanical output of blower fan output is also inequality, but under every kind of wind friction velocity, all there is the peak power point P of a best blade rotating speed of correspondence _Max, under this best blade speed conditions, blade can reach best tip speed ratio, thereby capture the ceiling capacity under this wind speed, because wind-driven generator is subjected to the blower fan transmission, if the power that can make wind-driven generator output corresponding to best blade rotating speed, just can make utilization ratio of wind energy obtain maximization.Therefore, the process of maximal wind-energy tracking can be understood as effective control wind-driven generator rotating speed

Process, be complementary by power output and the best blade rotating speed of the adjusting of wind-driven generator rotating speed being controlled wind-driven generator, improve the effect of utilizing to wind energy.

2) principle of OPTIMAL REACTIVE POWER tracking:

The calculating principle be in wind-driven generator allows range of operation, choose and can make its a certain performance evaluation function f reach optimum reactive power value, valuation functions f with

Relation can be illustrated by Fig. 4, the wind-driven generator loss characteristic of double-fed control is ignoring under the frequency converter loss situation, with the stator reactive power Q _sRelevant loss mainly is generator unit stator copper loss P _CusWith rotor copper loss P _CurSo definable valuation functions f is:

f＝P _cus+P _cur

Launch:

① $f=i_s^2{r}_s+i_r^2{r}_r=(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2)r_s+(i_{\mathrm{rd}}^2+i_{\mathrm{rq}}^2)r_r$

Wherein, r _s, r _rBe respectively the stator and rotor equivalent resistance; i _Sd, i _SqBe respectively stator d, q shaft current; i _Rd, i _RqBe respectively rotor d, q shaft current;

According to double feedback electric engine Mathematical Modeling under the d-q synchronous rotating frame, with d axle and double feedback electric engine stator magnetic linkage ψ _sOverlap, the magnetic linkage equation arranged:

2. ψ _Sd=-L _si _Sd+ L _mi _Rd=ψ _s3. ψ _Sq=-L _si _Sq+ L _mi _Rq=0

Wherein, ψ _Sd, ψ _SqBe respectively stator d, q axle magnetic linkage; L _s, L _mBe respectively stator equivalence self-induction and mutual inductance.

Active power P under the d-q synchronous rotating frame _sAnd reactive power Q _sBe respectively:

4. P _s=U _si _Sq5. Q _s=U _si _Sd

Wherein, U _sBe the stator effective voltage; Will be 2., 3., 4., 5. 1. Shi Kede of formula substitution:

$f=a{Q}_s^2+b{Q}_s+c$

Wherein, coefficient a, b, c are respectively:

$a=\frac{L_m^2{r}_s+L_s^2{r}_r}{U_s^2{L}_m^2},$ $b=\frac{2L_s{\mathrm{\ψ}}_s+U_s{r}_r}{U_s^2{L}_m^2},$ $c=\frac{P_s^2{L}_m^2{r}_s+P_s^2{L}_s^2{r}_r+{\mathrm{\ψ}}_s^2{U}_s^2{r}_r}{U_s^2{L}_m^2}$

Obviously, a, b, c are all greater than zero, at synchronous angular velocity ω ₁Following, when

When (formula is 6.), have: $f_{\min }=\frac{4\mathrm{ac}-b^2}{4a};$

Because OPTIMAL REACTIVE POWER value Permanent in negative, satisfy the idle condition of double-fed generator stable operation, 6. prior art is generally directly set according to formula

Be idle reference; But, because accuracy limitations and resistance temperature influence that generator parameter is measured are obvious, so this strategy is difficult to really to reach the loss of electric machine in actual motion minimum.At this problem, the present invention proposes the OPTIMAL REACTIVE POWER search control based on following thinking: from formula 6. as can be seen, although the shape of f will change because of parameter of electric machine perturbation,, when optimum as selecting to reduce the wind-driven generator loss

Principle the time because f and Q _sBetween have minimizing conic section relation and set up all the time, can search out by technological means and satisfy under the loss minimal condition

Thereby reduce the loss of wind-driven generator;

" the maximal wind-energy Tracing Control " of record and " OPTIMAL REACTIVE POWER search control " is the solution that proposes based on aforementioned analysis in " summary of the invention " of preamble, the first fuzzy inference rule table that wherein relates to, the second fuzzy inference rule table and the 3rd fuzzy inference rule table are for the technical tool of realizing the object of the invention, and these fuzzy inference rules are to be combined with the present invention like this:

(1) first fuzzy inference rule table

Referring to Fig. 5, provided the fuzzy reasoning schematic diagram of realizing the first fuzzy inference rule table among Fig. 5: when the blade rotation speed change, axle is that the mechanical loss variable quantity is than absorbing wind energy variation delta p _mCan ignore; And p _mOwing to not directly measure the increment Delta p so the available wind generator stator-rotator is always gained merit _eReplace Δ p _mZ among Fig. 5 ^-1Expression one step hysteresis (time delay) link; Wherein, k _pThe quantizing factor that is input variable (also is " the corresponding Δ p in the corresponding scheme _EiThe input variable quantizing factor "), to realize stator and rotor of wind power generator meritorious increment Delta p always _EiConversion from basic domain to the linguistic variable domain, the domain scope definition of linguistic variable is [1,1]; k _wIt is the output variable scale factor; Optimized rotating speed among the figure is core (namely realizing fuzzy reasoning) with reference to the fuzzy logic control module, power increment and rotation speed change in its comprehensive every bat sampling are given the rotating speed reference that makes new advances, it does not rely on parameters such as fan parameter and atmospheric density, (prior art is generally controlled with reference to carrying out indirect rotating speed by dynamic setting generator unit stator active power so this maximal wind-energy pursive strategy has very strong adaptability to environment, can realize the maximal wind-energy tracking under not having wind speed measures, and the setting of generator unit stator active power is relevant with the characteristic of blower fan, in practice, along with meteorological geographical conditions change, atmospheric density can change, unit operation is aging, situations such as dust deposit blade can influence fan characteristic equally, therefore the generator unit stator active power in unit operation is set and will be departed from optimum all the time, causes the tracking effect of maximal wind-energy relatively poor).

The inference rule of the first fuzzy inference rule table can be set by following thinking:

1, in the last control cycle, if Δ ω _RiWith Δ p _EiBe on the occasion of, the working point is described near extreme point (namely optimizing tachometer value), then new generator speed increment Delta ω _r(being the theoretical rotational speed changing value) just also should be, and need carry out forward lookup;

2, in the last control cycle, if Δ ω _RiBe Δ p just _EiWhen negative, the working point is described away from extreme point, the generator speed increment Delta ω that this is stylish _rShould be negative, need carry out reverse search;

Add aforementioned 1,2 inference rule and set the inverse process of condition, can set up with Lp accordingly _EiWith Δ ω _RiBe input variable, with theoretical rotational speed changing value Δ ω _rBe the fuzzy rule of output variable, as shown in table 1 below:

Table 1

Last table is the first fuzzy inference rule table, and this table adopts dual input-single output mode, input variable Lp _EiFuzzy domain comprises 7 fuzzy subsets, the language value on domain get NB, NM, NS, ZE, PS, PM, PB}, namely negative big, and negative in, negative little, zero, just little, the center, honest }.Accompanying drawing 6 is input variable Lp _EiFuzzy membership functions.The triangular function of membership function employing uneven distribution its objective is that when variable approached zero, the sensitiveness of membership function increased, so that the instant step-size in search of adjusting is with the raising search efficiency when search approaches best point.The Δ ω of input variable _RiFuzzy domain comprises 3 fuzzy subsets, the language value on domain get N, ZE, P}, namely negative, zero, just }.Output variable Δ ω _rFuzzy domain on language value and input variable Lp _EiIdentical.

The step 4), 5 of maximal wind-energy Tracing Control scheme) in, relates to input variable quantizing factor k _pAdjustment, its principle is: though at defeated people's variables L p _EiWith Δ ω _RiMembership function design in, considered to choose the triangular function of uneven distribution, so that variable improves the sensitiveness of membership function when approaching zero, but fuzzy controller itself is not owing to have integral element, thereby search precision is not high, it is static surplus poor to exist, this is the defective that the fuzzy logic search exists, when putting with the approaching the best of certain step-size in search, when little deviation range, still exist one and regulate the dead band, cause search near the best point, to form limit cycles oscillations, the long-term vibration that can bring drive shaft system thus, this will cause bigger infringement to the wind-driven generator mechanical part, therefore need to adopt further step to improve search precision; During adjusting dead band in the prior art in solving fuzzy reasoning, generally by the shelves of fuzzy reasoning table are got the precision that more carefully improves fuzzy control, but this processing mode also can make regular number and system-computed amount also increase thereupon, thereby cause overhead and time of delay significantly to increase, the real-time of influence control, and can cause subsequent control time delay to occur.Therefore during the adjusting dead-time problem in solving fuzzy reasoning, should adopt more simply, way of search fast, the present invention has namely adopted step 4), 5) in scheme realize simply, search fast:

Step 4), 5) scheme in makes the present invention to output variable Δ ω _rSearch for the search that forms a kind of variable step generally, i.e. early stage | Δ ω _r| be worth bigger, the later stage | Δ ω _r| be worth less.When searching for generally when approaching to the optimized rotating speed reference value gradually, the fuzzy division on the bigger input domain of initial given range just seems coarse, and control precision is not high; For further segmenting zero territory, with reference to increment search effect among a small circle, this solution is from the quantizing factor k of input variable to system's rotating speed for the increasing fuzzy controller _pStart with, the method that is fixed value with the selected input variable quantizing factor of common fuzzy controller is different, approaches in the best process of putting in the search later stage to increase k _p, be equivalent to dwindle input variable Lp _EiBasic domain, improved fuzzy controller to input variable Lp _EiResolution, make that fuzzy controller can be to small Lp _EiMake a response, increased its control action, thereby can reduce the search vibration, improve search performance.

In maximal wind-energy Tracing Control process, such scheme can improve search precision under the condition that does not increase overhead and delay, solve the problem of regulating the dead band; Complete maximal wind-energy Tracing Control scheme is as follows:

1), starts wind-driven generator and generate electricity by way of merging two or more grid systems during wind speed more than or equal to incision when wind speed, rotating speed and the active power of wind-driven generator are carried out continuous sampling, in each sampling period, the changing value of change in rotational speed value and active power is calculated; If the rotation speed change value that is recorded in the single sampling period is Δ ω _Ri, the active power changing value that is recorded in the single sampling period is Δ p _Ei, i is sampling number, i=1,2,3,4 ... n; If Δ ω _R1With Δ p _E1Be on the occasion of;

2) in second sampling period, with Lp _E1With Δ ω _R1Two input variable (Δ ω as the first fuzzy inference rule table _R1It is the rotation speed change value that collects in first sampling period, it also is the maximum in the rotation speed change value), obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value, the double-fed control system is regulated the wind-driven generator rotating speed according to optimizing tachometer value; Enter step 3); Second sampling period also namely formed first control cycle;

Wherein, Lp _E1Be corresponding k _pWith Δ p _E1The active power input variable, Lp _E1=k _pΔ p _E1, k _pBe corresponding Δ p _EiThe input variable quantizing factor;

3) subsequent control is in the cycle, with current Lp _EiWith two input variables of the theoretical rotational speed changing value that obtains in the last control cycle as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value; The double-fed control system according to the optimization tachometer value of correspondence to the wind-driven generator rotating speed carry out continuously, dynamic adjustments; Enter step 4);

Wherein, Lp _Ei=k _pΔ p _EiLp _EiBe corresponding k _pWith Δ p _EiThe active power input variable;

4) in the control procedure of step 3), with the Δ ω of current sampling period correspondence _RiAbsolute value | Δ ω _Ri| with Δ ω _R1/ k ₁Compare in real time: if | Δ ω _Ri| greater than Δ ω _R1/ k ₁, illustrate that the wind-driven generator rotating speed still has bigger rising surplus, need not to regulate k _p, should return step 3); If | Δ ω _Ri| be less than or equal to Δ ω _R1/ k ₁, illustrate that the rising surplus of wind-driven generator rotating speed is less, should be to k _pRegulate, enter step 5);

Wherein, k ₁Be the regulatory factor that data rule of thumb obtain, k ₁Value between 2.5～3.5;

5) with current Lp _Ei ^*With two input variables of the theoretical rotational speed changing value that obtains in the last control cycle as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value; The double-fed control system according to the optimization tachometer value of correspondence to the wind-driven generator rotating speed carry out continuously, dynamic adjustments; Enter step 6);

Wherein, Lp _Ei ^*=Δ p _EiK _y, Lp _Ei ^*Be corresponding k _yWith Δ p _EiThe active power input variable, k _y=k _pK ₁k _yBe the corresponding Δ p after regulating _EiThe input variable quantizing factor;

6) when wind speed enters plateau, having little significance of maximal wind-energy Tracing Control proceeded near rotating speed fluctuation best power point basically, should therefore can pass through to set the less ε of a numerical value with diversion to reducing in the wind-driven generator loss _Wmin, come in time control strategy to be switched to OPTIMAL REACTIVE POWER search control: with current sampling period correspondence | Δ ω _Ri| with ε _WminCompare in real time: if | Δ ω _Ri| more than or equal to ε _Wmin, then return step 5); If | Δ ω _Ri| less than ε _Wmin, then switch to OPTIMAL REACTIVE POWER search control by the maximal wind-energy Tracing Control;

Wherein, ε _WminBe the critical switching value of corresponding maximal wind-energy Tracing Control, its value can be by experiment or empirical data determine;

(2) second fuzzy inference rule tables and the 3rd fuzzy inference rule table

Referring to Fig. 4, the basic design philosophy of fuzzy inference rule table that is used for OPTIMAL REACTIVE POWER search control is: by searching for mode generally, real time altering generator reactive increment, detect variation and the last one idle variation with reference to increment constantly of input loss increment simultaneously, thereby determine new idle with reference to increment, by such search, can make the working point finally be stabilized in the smallest point f of valuation functions _MinNear.

When the fuzzy controller that is used for the OPTIMAL REACTIVE POWER tracking is put with the approaching the best of certain step-size in search, also can there be the adjusting dead-time problem the same with prior art, therefore, when being designed for the fuzzy controller of realizing OPTIMAL REACTIVE POWER search control, in order to obtain effect preferably, should consider how to take into account the quick search in early stage and the accurate orientation problem in later stage;

From the switching condition of the maximal wind-energy Tracing Control of front and OPTIMAL REACTIVE POWER search control as can be seen, the wind-driven generator running status the when section of OPTIMAL REACTIVE POWER search control performance control action is comparatively stablized corresponding to wind-force.In such cases, the complexity that the wind-driven generator running status changes is more simple with respect to the maximal wind-energy Tracing Control stage, at this moment required fuzzy inference rule quantity is less relatively, therefore, even with the further refinement of the gear of fuzzy controller, the fuzzy inference rule quantity of Zeng Jiaing is also less relatively thereupon, and is very limited with the negative effect that time-delay causes to overhead; So in OPTIMAL REACTIVE POWER search control, the present invention is referred from and handles the means of regulating the dead band in the prior art, design two fuzzy controllers (also namely two fuzzy inference rule tables) and come respectively corresponding two time domains, be used for satisfying the pinpoint requirement during near OPTIMAL REACTIVE POWER value point of the rapidity of search in early stage and later stage respectively.

The system configuration of two fuzzy controllers all adopts structure shown in Figure 7, the Z among Fig. 7 ^-1Expression one step hysteresis (time delay) link, two input variable Δ f among the figure _vWith Δ Q _SvBe respectively variable quantity and the reactive power changing value of valuation functions f, k _fBe the input variable quantizing factor, the domain scope definition of linguistic variable is [1,1]; k _qBe the output variable scale factor, wherein OPTIMAL REACTIVE POWER is the core with reference to fuzzy logic control, and it imports Δ f _vWith the theoretical idle changing value that obtains in the last idle control cycle, export new theoretical idle changing value, this mode does not rely on the accuracy of generator parameter, has good adaptive capacity.

To use multiple parameter values such as stator and rotor resistance parameters and inductance owing to directly calculate the variable quantity of valuation functions f, this can cause problem complicated, consider that this moment, wind-driven generator was operated in the wind speed stabilization sub stage, therefore, the variable quantity of valuation functions can replace with the anti-value of active power variable quantity, namely-and Δ p _v, so Δ f is arranged _v=-Δ p _vK _f

The inference rule of the second fuzzy inference rule table can be set by following thinking:

1, as Δ Q _SvWith Δ f _vBe timing, present working point be described away from OPTIMAL REACTIVE POWER value point, then new Δ Q _s(being theoretical idle changing value) should be negative, needs reverse search;

2, as Δ Q _SvFor just, and Δ f _vFor negative, present working point close OPTIMAL REACTIVE POWER value point then is described, then should continue to search for by the current search direction;

Add aforementioned 1,2 inference rule and set the inverse process of condition, can set up with Δ Q _SvWith Δ f _vFor input variable, theoretical idle changing value (are the Δ Q among Fig. 7 _s) be the second fuzzy inference rule table of output variable, as shown in table 2 below:

Table 2,

The second fuzzy inference rule table adopts dual input-single output mode, and Fig. 8,9,10 is respectively Δ f _v, Δ Q _SvFuzzy membership functions with output variable.Δ f _vFuzzy domain comprise 5 fuzzy subsets, the language value on domain is got { NB, NS, ZE, PS, PB}, i.e. { negative big, negative little, zero, just little, honest }.The triangular function of membership function employing uneven distribution its objective is that when variable approached zero, the sensitiveness of membership function increased, in order in time adjust step-size in search with the raising search efficiency when search approaches best point.Δ Q _SvFuzzy domain comprise 2 fuzzy subsets, the language value on domain get N, P}, namely negative, just }.Language value on the fuzzy domain of output variable and input variable Δ f _vIdentical.

The 3rd fuzzy inference rule table is as shown in table 3 below.

Table 3

The 3rd fuzzy inference rule table adopts dual input-single output mode, and accompanying drawing 11,12,13 is respectively Δ f _v, Δ Q _SvFuzzy membership functions with output variable.Δ f _vFuzzy domain increase by 4 fuzzy subsets than table 2, be 9 fuzzy subsets, the language value on domain get NVB, NB, NM, NS, ZE, PS, PM, PB, PVB}, namely negative very big, negative big, and negative in, negative little, zero, just little, the center, honest, just very big }.Membership function is chosen the triangular function of uneven distribution, and its purpose is the same.Δ Q _SvFuzzy domain comprise 2 fuzzy subsets, the language value on domain get N, P}, namely negative, just }.Language value on the fuzzy domain of output variable and Δ f _vIdentical.

Complete OPTIMAL REACTIVE POWER search control scheme is as follows:

1] reactive power and the active power of wind-driven generator are carried out continuous sampling, each reactive power was calculated the changing value of reactive power and the changing value of active power in the sampling period; If the reactive power changing value that single reactive power was recorded in the sampling period is Δ Q _Sv, the active power changing value that single reactive power was recorded in the sampling period is Δ p _v, v is sampling number, v=1,2,3,4 ... n;

2] second reactive power is in the sampling period, with Δ Q _S1With Δ f ₁Two input variables as the second fuzzy inference rule table, obtain theoretical idle changing value according to the second fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system is regulated the reactive power of wind-driven generator according to optimizing reactive power value; Enter step 3]; The second reactive power sampling period also namely formed the first idle control cycle;

Wherein, Δ f ₁Be corresponding Δ p ₁The active power input variable, Δ f ₁=-Δ p ₁K _f, k _fBe corresponding Δ p _vThe input variable quantizing factor;

3] in the follow-up idle control cycle, with current Δ f _vWith two input variables of the theoretical idle changing value that obtains in the last idle control cycle as the second fuzzy inference rule table, obtain theoretical idle changing value according to the second fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system according to optimize reactive power value to the reactive power of wind-driven generator carry out continuously, dynamic adjustments; Enter step 4];

Wherein, Δ f _vBe corresponding Δ p _vThe active power input variable, Δ f _v=-Δ p _vK _f

4] in step 3] control procedure in, in each idle control cycle with current Δ Q _SvAbsolute value | Δ Q _Sv| with k ₂| Δ Q _S1| compare in real time:

If | Δ Q _Sv| greater than k ₂| Δ Q _S1|, illustrate that the reactive power of wind-driven generator also with the big unidirectional reactive power point that approaches corresponding lowest loss of step-length, then continues Δ p _vAbsolute value | Δ p _v| with ε _W1Compare: if | Δ p _v| greater than ε _W1, illustrate that the wind speed variation having caused meritorious sudden change, should follow the trail of maximal wind-energy again, at this moment, stop OPTIMAL REACTIVE POWER search control, simultaneously, the double-fed control system is regulated the reactive power of wind-driven generator according to the reactive power rated value of setting, and switches to the maximal wind-energy Tracing Control; If | Δ p _v|≤ε _W1, then return step 3];

If | Δ Q _Sv| be less than or equal to k ₂| Δ Q _S1|, the reactive power point that the reactive power that wind-driven generator is described has been positioned at corresponding lowest loss need carry out comparatively accurate localization to the OPTIMAL REACTIVE POWER power points nearby, then enters step 5];

Wherein, k ₂Be the regulatory factor that data rule of thumb obtain, k ₂Value between 0.3～0.4;

5] with current Δ f _vWith two input variables of the theoretical idle changing value that obtains in the last idle control cycle as the 3rd fuzzy inference rule table, obtain theoretical idle changing value according to the 3rd fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system according to optimize reactive power value to the reactive power of wind-driven generator carry out continuously, dynamic adjustments; Enter step 6];

6] in step 5] control procedure in, in each idle control cycle will | Δ p _v| with ε _W1Compare in real time:

If satisfy | Δ p _v|＞ε _W1Condition, illustrate that wind speed changes and having caused meritorious sudden change, should follow the trail of maximal wind-energy again, should stop OPTIMAL REACTIVE POWER search control, simultaneously, the double-fed control system is regulated the reactive power of wind-driven generator according to the reactive power rated value of setting, and switches to the maximal wind-energy Tracing Control; If satisfy | Δ p _v|≤ε _W1Condition, illustrate that wind speed is steady relatively, then return step 5];

Wherein, ε _W1Be the critical switching value of corresponding OPTIMAL REACTIVE POWER search control, its value can be determined according to experiment or empirical data.

Logic diagram of the present invention as shown in figure 14.

Claims (1)

1. the control method of the dual efficient fuzzy optimization of double-fed wind power generator comprises the wind-driven generator that adopts the control of double-fed control system; When wind speed more than or equal to incision during wind speed, wind-driven generator generates electricity by way of merging two or more grid systems, increase along with wind speed, when wind-driven generator reaches maximum allowable speed, enter permanent rotating speed generating state, it is characterized in that: the back that begins to generate electricity by way of merging two or more grid systems, wind-driven generator reach before the maximum allowable speed, adopt following control method that wind-driven generator is controlled:

When wind-driven generator has just begun to generate electricity by way of merging two or more grid systems, carry out earlier the maximal wind-energy Tracing Control as follows:

1) when wind speed more than or equal to incision during wind speed, wind-driven generator generates electricity by way of merging two or more grid systems, and rotating speed and the active power of wind-driven generator are carried out continuous sampling, in each sampling period, the changing value of change in rotational speed value and active power is calculated; If the rotation speed change value that is recorded in the single sampling period is Δ ω _Ri, the active power changing value that is recorded in the single sampling period is Δ p _Ei, i is sampling number, i=1,2,3,4 ... n; If Δ ω _R1With Δ p _E1Be on the occasion of;

2) in second sampling period, with Lp _E1With Δ ω _R1Two input variables as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value, the double-fed control system is regulated the wind-driven generator rotating speed according to optimizing tachometer value; Enter step 3); Second sampling period also namely formed first control cycle;

Wherein, Lp _E1Be corresponding k _pWith Δ p _E1The active power input variable, Lp _E1=k _pΔ p _E1, k _pBe corresponding Δ p _EiThe input variable quantizing factor;

3) subsequent control is in the cycle, with current Lp _EiWith two input variables of the theoretical rotational speed changing value that obtains in the last control cycle as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value; The double-fed control system according to the optimization tachometer value of correspondence to the wind-driven generator rotating speed carry out continuously, dynamic adjustments; Enter step 4);

Wherein, Lp _Ei=k _pΔ p _EiLp _EiBe corresponding k _pWith Δ p _EiThe active power input variable;

4) in the control procedure of step 3), with the Δ ω of current sampling period correspondence _RiAbsolute value | Δ ω _Ri| with Δ ω _R1/ k ₁Compare in real time: if | Δ ω _Ri| greater than Δ ω _R1/ k ₁, then return step 3); If | Δ ω _Ri| be less than or equal to Δ ω _R1/ k ₁, then enter step 5);

Wherein, k ₁Be the regulatory factor that data rule of thumb obtain, k ₁Value between 2.5～3.5;

5) with current Lp _Ei ^*With two input variables of the theoretical rotational speed changing value that obtains in the last control cycle as the first fuzzy inference rule table, obtain the theoretical rotational speed changing value according to the first fuzzy inference rule table, calculate the optimization tachometer value of wind-driven generator according to the theoretical rotational speed changing value; The double-fed control system according to the optimization tachometer value of correspondence to the wind-driven generator rotating speed carry out continuously, dynamic adjustments; Enter step 6);

Wherein, Lp _Ei ^*=Δ p _EiK _y, Lp _Ei ^*Be corresponding k _yWith Δ p _EiThe active power input variable, k _y=k _pK ₁k _yBe the corresponding Δ p after regulating _EiThe input variable quantizing factor;

6) with current sampling period correspondence | Δ ω _Ri| with ε _WminCompare in real time: if | Δ ω _Ri| more than or equal to ε _Wmin, then return step 5); If | Δ ω _Ri| less than ε _Wmin, then switch to OPTIMAL REACTIVE POWER search control by the maximal wind-energy Tracing Control;

Wherein, ε _WminCritical switching value for corresponding maximal wind-energy Tracing Control;

After entering OPTIMAL REACTIVE POWER search control, control as follows:

1] reactive power and the active power of wind-driven generator are carried out continuous sampling, each reactive power was calculated the changing value of reactive power and the changing value of active power in the sampling period; If the reactive power changing value that single reactive power was recorded in the sampling period is Δ Q _Sv, the active power changing value that single reactive power was recorded in the sampling period is Δ p _v, v is sampling number, v=1,2,3,4 ... n;

2] second reactive power is in the sampling period, with Δ Q _S1With Δ f ₁Two input variables as the second fuzzy inference rule table, obtain theoretical idle changing value according to the second fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system is regulated the reactive power of wind-driven generator according to optimizing reactive power value; Enter step 3]; The second reactive power sampling period also namely formed the first idle control cycle;

Wherein, Δ f ₁Be corresponding Δ p ₁The active power input variable, Δ f ₁=-Δ p ₁K _f, k _fBe corresponding Δ p _vThe input variable quantizing factor;

3] in the follow-up idle control cycle, with current Δ f _vWith two input variables of the theoretical idle changing value that obtains in the last idle control cycle as the second fuzzy inference rule table, obtain theoretical idle changing value according to the second fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system according to optimize reactive power value to the reactive power of wind-driven generator carry out continuously, dynamic adjustments; Enter step 4];

Wherein, Δ f _vBe corresponding Δ p _vThe active power input variable, Δ f _v=-Δ p _vK _f

4] in step 3] control procedure in, in each idle control cycle with current Δ Q _SvAbsolute value | Δ Q _Sv| with k ₂| Δ Q _S1| compare in real time:

If | Δ Q _Sv| greater than k ₂| Δ Q _S1|, then continue Δ p _vAbsolute value | Δ p _v| with ε _W1Compare: if | Δ p _v| greater than ε _W1, then stop OPTIMAL REACTIVE POWER search control, simultaneously, the double-fed control system is regulated the reactive power of wind-driven generator according to the reactive power rated value of setting, and switches to the maximal wind-energy Tracing Control; If | Δ p _v|≤ε _W1, then return step 3];

If | Δ Q _Sv| be less than or equal to k ₂| Δ Q _S1|, then enter step 5];

Wherein, k ₂Be the regulatory factor that data rule of thumb obtain, k ₂Value between 0.3～0.4;

5] with current Δ f _vWith two input variables of the theoretical idle changing value that obtains in the last idle control cycle as the 3rd fuzzy inference rule table, obtain theoretical idle changing value according to the 3rd fuzzy inference rule table, calculate the optimization reactive power value of wind-driven generator according to the idle changing value of theory, the double-fed control system according to optimize reactive power value to the reactive power of wind-driven generator carry out continuously, dynamic adjustments; Enter step 6];

6] in step 5] control procedure in, in each idle control cycle will | Δ p _v| with ε _W1Compare in real time:

If satisfy | Δ p _v|＞ε _W1Condition, then stop OPTIMAL REACTIVE POWER search control, simultaneously, the double-fed control system is regulated the reactive power of wind-driven generator according to the reactive power rated value of setting, and switches to the maximal wind-energy Tracing Control; If satisfy | Δ p _v|≤ε _W1Condition, then return step 5];

Wherein, ε _W1Critical switching value for corresponding OPTIMAL REACTIVE POWER search control.

Images