CN103107534A - Double-fed induction power generation system optimization power prediction control method - Google Patents

Abstract

The invention discloses a double-fed wind power generation system coordinative optimization power prediction control method. The method includes the following steps: discretizing the mathematical models of a double-fed asynchronous wind power generator and convertors of a grid side and a rotor side of the double-fed asynchronous wind power generator based on the mathematical models of the double-fed asynchronous wind power generator and convertors of the grid side and the rotor side of the double-fed asynchronous wind power generator, predicting the change situation of controlled variable when different control behaviors are imposed, setting a target function to choose an optimum control behavior, and achieving independent and effective control of active power and reactive power of a double-fed asynchronous wind power generation system. With the consideration of control delay effect existing in a real system, delay compensation can be achieved through two-step prediction. According to the double-fed induction power generation system optimization power prediction control method, optimizing control effects of the double-fed wind power generation system can be achieved. The double-fed induction power generation system optimization power prediction control method has the following advantages which are not possessed by traditional vector control and direct power control based on a switch list: the control structure is very simple, no complex coordinate transformation or controller parameter setting is needed, no hysteresis control and section judgment is needed, and no selecting method of vectors is needed to be considered. The double-fed induction power generation system optimization power prediction control method has a fast dynamic response function and good steady-state performance.

Description

A kind of double-fed induction electricity generation system optimizing power forecast Control Algorithm

Technical field

The present invention relates to the double-fed asynchronous wind generator system net of a kind of wind power generation field internal speed-changing constant frequency side, rotor-side voltage source converter PWM control method, particularly a kind of double-fed asynchronous wind generator system coordination optimization power prediction control method.

Background technology

Modern large-scale wind powered generation syst mainly adopts two types of double-fed asynchronous generator (DFIG) and magneto alternators, for the raising generating efficiency, all adopts the variable speed constant frequency generator operational mode.Wherein, DFIG uses at most, and technology is the most ripe, is current mainstream model.The double-fed wind generating scheme can realize that variable speed constant frequency controls, and reduces the capacity of converter, also can realize gaining merit, idle decoupling zero controls, can be according to the requirement corresponding perception of output or the capacitive reactive power of electrical network, and the flexibility of this idle control is highly beneficial to electrical network.Vector control (VC) and direct Power Control (DPC) are the control strategies of dual feedback wind power generation system main flow always.Vector control can realize the independent regulation of active power and reactive power, obtains good steady-state behaviour, but needs comparatively complicated same leg speed rotating coordinate transformation and the phase information of line voltage, and the parameter tuning of pi regulator is complicated, and dynamic property is slightly poor.Tradition need not with the leg speed rotating coordinate transformation based on the direct Power Control (LUT-DPC) of switch list, control structure is simple, dynamic response is fast, but need carry out stagnant chain rate and the Stator flux linkage sectors judgement, system power fluctuates larger, and the steady-state behaviour of the method haves much room for improvement.Above control method or aspect system's dynamic response or certain defective is being arranged aspect steady operation can not reach comparatively ideal effect simultaneously, and the further investigation of therefore the double-fed induction electricity generation system being controlled has realistic meaning.

Summary of the invention

The object of the invention is to for the deficiencies in the prior art, a kind of double-fed asynchronous wind generator system coordination optimization power prediction control method is provided.The inventive method is compared traditional control method, need not to increase additional hardware, and control structure is very simple, can reach dynamic property and stable state accuracy preferably.

Technical solution of the present invention, a kind of double-fed asynchronous wind generator system coordination optimization power prediction control method comprises double-fed asynchronous wind generator system rotor side converter RSC(1) optimizing power forecast Control Algorithm and grid side converter GSC(12) the optimizing power forecast Control Algorithm; Described rotor-side converter RSC(1) optimizing power forecast Control Algorithm and grid side converter GSC(12) the optimizing power forecast Control Algorithm is within k sampling period, to in microprocessor, it be processed after the sampled signal discretization, to obtain the rotor-side converter switches signal S of k+1 sampling period planted agent output _a1(k+1), S _b1(k+1), S _c1(k+1) and grid side converter switching signal S _a2(k+1), S _b2(k+1), S _c2(k+1); If first sampling period internal rotor side converter switching signal S _a1(1), S _b1(1), S _c1(1) and grid side converter switching signal S _a2(1), S _b2(1), S _c2(1) output is 0;

Described RSC(1) optimizing power forecast Control Algorithm comprises the following steps:

(1) establishing the zero hour in this sampling period is t _k, utilize single-phase voltage Hall element (17) to gather DC side busbar voltage V _dcUtilize three voltage hall sensors (14) to gather dual-feed asynchronous wind power generator DFIG(18) threephase stator voltage signal U _SabcUtilize the first three-phase current Hall element (13-1) to gather threephase stator current signal I _Sabc, utilize the second three-phase current Hall element (13-2) to gather three-phase rotor current signal I _Rabc

(2) the threephase stator voltage signal U that collects _SabcDetect the angular frequency that obtains electrical network or stator voltage through phase-locked loop (6) _sAdopt simultaneously encoder (10) to detect the rotor position of DFIG _r, then through differentiator (7) Calculation Speed ω _rObtain the slippage angular frequency by what subtracter calculated _Slip=ω _s-ω _r

(3) with the threephase stator voltage signal U that collects _SabcWith threephase stator current signal I _SabcTo two-phase coordinate transformation module (5-1), obtain t through the first static three-phase _kStator voltage vector U under moment stator coordinate _{S α β}(k) and stator current vector I _{S α β}(k); With the three-phase rotor current signal I that collects _RabcAccording to rotor position _rRotate to two-phase static coordinate conversion module (9) through two-phase, obtain t _kRotor current vector I under moment stator coordinate _{R α β}(k);

(4) according to rotor position _rAnd the rotor converter switching signal S of output in this sampling period (k sampling period) that calculated in the upper sampling period (k-1 sampling period) _a1(k), S _b1(k), S _c1(k), calculate t through module (8) _kRotor voltage vector V under the stator coordinate that moment rotor-side converter applies _{R α β}(k);

(5) t that sampling is obtained _kStator voltage vector U constantly _{S α β}(k), stator current vector I _{S α β}(k), rotor current vector I _{R α β}(k), slippage angular frequency _SlipAnd t _kThe rotor voltage vector V that constantly applies _{R α β}(k), by rotor-side optimizing power prediction module (4-1) (t predict next sampling period (k+1 sampling period) zero hour _k+1Stator voltage vector U constantly) _{S α β}(k+1), rotor current vector I _{R α β}(k+1) and meritorious, the reactive power signals P of stator output _s(k+1), Q _s(k+1);

(6) according to the t that predicts in (5) step _k+1Stator voltage vector U constantly _{S α β}(k+1), rotor current vector I _{R α β}(k+1), meritorious, the reactive power signals P of stator output _s(k+1), Q _s(k+1) and 8 rotor voltage vector V that might apply in k+1 sampling period _{R α β}(k+1), predict by rotor-side second step prediction module (3-1) and applying respectively these rotor voltage vector V _{R α β}Under (the t zero hour in k+2 sampling period _k+2All possible meritorious, reactive power signals P constantly) _s(k+2), Q _s(k+2); Can obtain 8 groups of meritorious, reactive power signals P _s(k+2), Q _s(k+2);

(7) with the t that predicts in (6) step _k+2Constantly 8 groups of meritorious, reactive power signals P _s(k+2), Q _s(k+2) and t _k+2Meritorious, the reactive power reference signal P of given rotor-side of the moment _s ^*, Q _s ^*, obtain through the minimization of object function control module (2-1) the rotor-side converter switches signal S that k+1 sampling period planted agent exports _a1(k+1), S _b1(k+1), S _c1(k+1);

The rotor-side converter switches signal S of k+1 the sampling period planted agent output that (8) will obtain _a1(k+1), S _b1(k+1), S _c1(k+1) at t _k+1Constantly drive IGBT through driver module and realize RSC(1) the optimizing power PREDICTIVE CONTROL.

Described GSC(12) optimizing power forecast Control Algorithm comprises the following steps:

(1) establishing the zero hour in this sampling period is t _k, utilize the 3rd three-phase current Hall element (13-3) Gather and input GSC(12) the three phase network current signal I that flows through filter inductance (15) _Gabc

(2) the three phase network current signal I that collects _GabcTo two-phase coordinate transformation module (5-2), obtain t through the second static three-phase _kGSC(12 under moment rest frame) power network current vector I _{G α β}(k);

(3) by the interior grid side converter switching signal S of this sampling period (k sampling period) that calculated in the upper sampling period (k-1 sampling period) _a2(k), S _b2(k), S _c2(k), calculate t through module (11) _kControl voltage vector V under the rest frame that moment grid side converter applies _{C α β}(k);

(4) t that sampling is obtained _kMoment line voltage vector U _{G α β}(k), power network current vector I _{G α β}(k) and t _kThe control voltage vector V that constantly applies _{C α β}(k), by net side optimizing power prediction module (4-2) (t predict next sampling period (k+1 sampling period) zero hour _k+1Line voltage vector U constantly) _{G α β}(k+1) and the input grid side converter is meritorious, reactive power signals P _g(k+1), Q _g(k+1); Here, t _kLine voltage vector U constantly _{G α β}(k) be stator voltage vector U _{S α β}(k);

(5) t that obtains according to prediction in (4) step _k+1Line voltage vector U constantly _{G α β}(k+1), meritorious, the reactive power signals P of input grid side converter _g(k+1), Q _g(k+1) and 8 control voltage vector V that might apply in k+1 sampling period _{C α β}(k+1), applying respectively these control voltage vectors V by net side second step prediction module (3-2) prediction _{C α β}(k+1) under k+2 the zero hour in sampling period (t _k+2All possible meritorious, reactive power signals P constantly) _g(k+2), Q _g(k+2); Can obtain 8 groups of meritorious, reactive power signals P _g(k+2), Q _g(k+2);

(6) with the t that predicts in (5) step _k+2Constantly 8 groups of meritorious, reactive power signals P _g( _k+ ₂), Q _g(k+2) and t _k+2Meritorious, the reactive power reference signal P of given net side of the moment _g ^*, Q _g ^*, obtain grid side converter switching signal S in k+1 sampling period through the minimization of object function control module (2-2) _a2(k+1), S _b2(k+1), S _c2(k+1).

The grid side converter switching signal S of k+1 the sampling period planted agent output that (7) will obtain _a2(k+1), S _b2(k+1), S _c2(k+1) at t _k+1Constantly drive IGBT through driver module and realize GSC(12) the optimizing power PREDICTIVE CONTROL.

The invention has the beneficial effects as follows, can improve conventional vector control with based on each comfortable dynamic property of direct Power Control of switch list or the defective on steady-state behaviour.This method control structure is very simple, need not complicated coordinate transform and controller parameter adjusts, need not stagnant ring control and sector judgement, need not to consider the system of selection of vector, can eliminate system's control lag to controlling the impact of effect, reduce the inverter power fluctuation, reduce the harmonic content of electric current, have dynamic property fast, simultaneously the variation of net side filter inductance parameter had than strong robustness.Consider to exist in real system and control delay effect, reach delay compensation by the prediction of two steps, effectively improved the operation control performance of double-fed asynchronous wind generator system.

Description of drawings

Fig. 1 is the schematic diagram of dual-feed asynchronous wind power generator group;

Fig. 2 is the schematic diagram that double-fed asynchronous wind generator system coordination optimization power prediction is controlled;

Fig. 3 (a) is desirable control principle drawing without controlling in the time-delay situation, is (b) the actual schematic diagram that adopts two step Forecasting Methodologies to compensate in control time-delay situation that has;

Fig. 4 is that the rotor-side tradition is based on the meritorious reactive current wave simulation design sketch of the direct Power Control of switch list;

Fig. 5 is that the rotor-side taking into account system is delayed time and do not add meritorious reactive current wave simulation design sketch in the compensation of delay situation;

Fig. 6 is that the rotor-side taking into account system is delayed time and adds meritorious reactive current wave simulation design sketch in the compensation of delay situation.

Embodiment

The present invention is further described below in conjunction with accompanying drawing, and purpose of the present invention and effect will be more obvious.

Fig. 1 is the schematic diagram of dual-feed asynchronous wind power generator group, comprises wind turbine, gear box, dual-feed asynchronous wind power generator DFIG18, rotor-side converter RSC1, grid side converter GSC12, filter inductance 15, dc-link capacitance 16 and step-up transformer.

Fig. 2 is the schematic diagram of a kind of double-fed asynchronous wind generator system coordination optimization power prediction control method that proposes of the present invention.Take a 2KW variable speed constant frequency DFIG as example, the double-fed asynchronous wind generator system coordination optimization power prediction control method that adopts the present invention to propose, being wherein the optimizing power PREDICTIVE CONTROL of RSC1 in the left-hand broken line frame, is the optimizing power PREDICTIVE CONTROL of GSC12 in the dotted line frame of right side.

Described rotor-side converter RSC1 optimizing power forecast Control Algorithm and grid side converter GSC12 optimizing power forecast Control Algorithm are within k sampling period, to in microprocessor, it be processed after the sampled signal discretization, to obtain the rotor-side converter switches signal S of k+1 sampling period planted agent output _a1(k+1), S _b1(k+1), S _c1(k+1) and grid side converter switching signal S _a2(k+1), S _b2(k+1), S _c2(k+1); If first sampling period internal rotor side converter switching signal S _a1(1), S _b1(1), S _c1(1) and grid side converter switching signal S _a2(1), S _b2(1), S _c2(1) output is 0;

Described RSC1 optimizing power forecast Control Algorithm comprises the steps:

(1) establishing the zero hour in this sampling period is t _k, utilize single-phase voltage transducer 17 to measure DC bus-bar voltage signal V _dcUtilize three-phase voltage Hall element 14 to gather the threephase stator voltage signal U of stator side _SabcUtilize the first three-phase current Hall element 13-1 to gather stator side threephase stator current signal I _SabcUtilize the second three-phase current Hall element 13-2 to gather rotor-side three-phase rotor current signal I _Rabc

(2) with the threephase stator voltage signal U that collects _SabcDetect through phase-locked loop 6 angular frequency that obtains stator voltage _sAdopt simultaneously encoder 10 to detect the rotor position of DFIG in this _r, then pass through differentiator 7 Calculation Speed ω _rAnd obtain the slippage angular frequency by what subtracter calculated _Slip=ω _s-ω _r

(3) with the threephase stator voltage signal U that collects _SabcWith threephase stator current signal I _SabcTo two-phase coordinate transformation module 5-1, obtain t through the first static three-phase _kStator voltage vector U under moment stator coordinate _{S α β}(k) and stator current vector I _{S α β}(k); With the three-phase rotor current signal I that collects _RabcAccording to rotor position _rRotate to two-phase static coordinate conversion module 9 through three-phase, obtain t _kRotor current vector I under moment stator coordinate _{R α β}(k); Take stator voltage as example, static three-phase is described suc as formula 1 to the expression formula of two-phase coordinate transform; Take rotor current as example, the expression formula that three-phase rotates to the conversion of two-phase static coordinate as shown in Equation 2;

$\left[\begin{array}{c}{U}_{\mathrm{s\α}}\left(k\right)\\ U_{\mathrm{s\β}}\left(k\right)\end{array}\right]=\frac{2}{3}\left[\begin{array}{ccc}1& -\frac{1}{2}& -\frac{1}{2}\\ 0& \frac{\sqrt{3}}{2}& -\frac{\sqrt{3}}{2}\end{array}\right]\left[\begin{array}{c}{U}_{\mathrm{sa}}\\ U_{\mathrm{sb}}\\ U_{\mathrm{sc}}\end{array}\right]---\left(1\right)$

$\left[\begin{array}{c}{I}_{\mathrm{r\α}}\left(k\right)\\ I_{\mathrm{r\β}}\left(k\right)\end{array}\right]=\frac{2}{3}\left[\begin{array}{ccc}{\cos \mathrm{\θ}}_r& \cos ({\mathrm{\θ}}_r+\frac{2}{3}\mathrm{\π})& \cos ({\mathrm{\θ}}_r-\frac{2}{3}\mathrm{\π})\\ {\sin \mathrm{\θ}}_r& \sin ({\mathrm{\θ}}_r+\frac{2}{3}\mathrm{\π})& \sin ({\mathrm{\θ}}_r-\frac{2}{3}\mathrm{\π})\end{array}\right]\left[\begin{array}{c}{I}_{\mathrm{ra}}\\ I_{\mathrm{rb}}\\ I_{\mathrm{rc}}\end{array}\right]---\left(2\right)$

In formula, U _sα(k), U _sβ(k) be respectively t _kStator voltage vector U constantly _{S α β}(k) α phase and β phase component; U _sa, U _sb, U _scBe respectively threephase stator voltage signal U _SabcA phase, b phase, c phase component; I _rα(k), I _rβ(k) be respectively t _kRotor current vector I constantly _{R α β}(k) α phase and β phase component; I _ra, I _rb, I _rcBe respectively three-phase rotor current signal I _RabcA phase, b phase, c phase component;

(4) according to rotor position _r, and the rotor-side converter switches signal S of output in this sampling period (k sampling period) that calculated in the upper sampling period (k-1 sampling period) _a1(k), S _b1(k), S _c1(k), calculate t _kRotor voltage vector V under the stator coordinate that moment rotor-side converter applies _{R α β}(k); Computing formula as shown in Equation 3;

$\left[\begin{array}{c}{V}_{\mathrm{r\α}}\left(k\right)\\ V_{\mathrm{r\β}}\left(k\right)\end{array}\right]=\frac{2}{3}{V}_{\mathrm{dc}}\left[\begin{array}{ccc}{\cos \mathrm{\θ}}_r& \cos ({\mathrm{\θ}}_r+\frac{2}{3}\mathrm{\π})& \cos ({\mathrm{\θ}}_r-\frac{2}{3}\mathrm{\π})\\ {\sin \mathrm{\θ}}_r& \sin ({\mathrm{\θ}}_r+\frac{2}{3}\mathrm{\π})& \sin ({\mathrm{\θ}}_r-\frac{2}{3}\mathrm{\π})\end{array}\right]\left[\begin{array}{c}{S}_{a1}\left(k\right)\\ S_{b1}\left(k\right)\\ S_{c1}\left(k\right)\end{array}\right]---\left(3\right)$

In formula, V _rα(k), V _rβ(k) be respectively t _kRotor voltage vector V constantly _{R α β}(k) α phase and β phase component;

(5) by t _kStator voltage vector U constantly _{S α β}With stator current vector I _{S α β}Calculate t _kMeritorious, the reactive power signals P of stator output constantly _s(k), Q _s(k); Meritorious, reactive power is calculated formula as shown in Equation 4;

$P_s\left(k\right)=-\frac{3}{2}[U_{\mathrm{s\α}}\left(k\right)I_{\mathrm{s\α}}\left(k\right)+U_{\mathrm{s\β}}\left(k\right)I_{\mathrm{s\β}}\left(k\right)]---\left(4\right)$

$Q_s\left(k\right)=-\frac{3}{2}[-U_{\mathrm{s\α}}\left(k\right)I_{\mathrm{s\β}}\left(k\right)+U_{\mathrm{s\β}}\left(k\right)I_{\mathrm{s\β}}\left(k\right)]$

(6) with t _kMeritorious, the reactive power signals P of stator output constantly _s(k), Q _s(k), stator voltage vector U _{S α β}(k) and t _kThe rotor voltage vector V that constantly applies _{R α β}(k), (t zero hour in next sampling period of prediction (k+1 sampling period) _k+1Meritorious, the reactive power signals P of stator output constantly) _s(k+1), Q _s(k+1); The prediction expression formula as shown in Equation 5;

$\left[\begin{array}{c}{P}_s(k+1)\\ Q_s(k+1)\end{array}\right]=\left[\begin{array}{cc}1-{\mathrm{\δT}}_s{L}_r{R}_s& {-\mathrm{\ω}}_{\mathrm{slip}}{T}_s\\ {\mathrm{\ω}}_{\mathrm{slip}}{T}_s& 1-{\mathrm{\δT}}_s{L}_r{R}_s\end{array}\right]\left[\begin{array}{c}{P}_s\left(k\right)\\ Q_s\left(k\right)\end{array}\right]+$

$\frac{{3\mathrm{\δT}}_s{L}_m}{2}\left[\begin{array}{cc}{-U}_{\mathrm{s\α}}\left(k\right)& {-U}_{\mathrm{s\β}}\left(k\right)\\ {-U}_{\mathrm{s\β}}\left(k\right)& U_{\mathrm{s\α}}\left(k\right)\end{array}\right]\left[\begin{array}{c}{V}_{\mathrm{r\α}}\left(k\right)-R_r{I}_{\mathrm{r\α}}\left(k\right)\\ V_{\mathrm{r\β}}\left(k\right)-R_r{I}_{\mathrm{r\β}}\left(k\right)\end{array}\right]+---\left(5\right)$

$\frac{{3\mathrm{s\&delta;T}}_s{\mathrm{<figure-callout\; id="2"\; label="L\; r"\; filenames="HDA00002743376900012.png,HDA00002743376900021.png"\; state="\{\{state\}\}">L</figure-callout>}}_r}{2}\left[\begin{array}{c}{\left|U\right|}^2\\ 0\end{array}\right]$

In formula, R _sBe stator resistance, R _rBe rotor resistance, L _sBe stator inductance, L _rBe inductor rotor, L _mBe magnetizing inductance, T _sBe the sampling period, $\mathrm{\&delta;}=\frac{1}{{{\mathrm{\&sigma;<figure-callout\; id="2"\; label="L\; m"\; filenames="HDA00002743376900012.png,HDA00002743376900021.png"\; state="\{\{state\}\}">L</figure-callout>}}_m}^2},$ $\mathrm{\&sigma;}=1-\frac{L_s{L}_{\mathrm{<figure-callout\; id="2"\; label="r\; L\; m"\; filenames="HDA00002743376900012.png,HDA00002743376900021.png"\; state="\{\{state\}\}">r</figure-callout>}}}{{L_m}^{}},$ $\left|U\right|=\sqrt{U_{\mathrm{s\&alpha;}}{\left(k\right)}^2+U_{\mathrm{s\&beta;}}{\left(k\right)}^2};$

(7) according to desirable electrical network voltage conditions, by t _kStator voltage vector U constantly _{S α β}(k) predict t _k+1Stator voltage vector U constantly _{S α β}(k+1); The prediction expression formula is described suc as formula 6;

U _sα(k+1)=U _sα(k)cos(ω _sT _s)-U _sβ(k)sin(ω _sT _s)（6）

U _sβ(k+1)=U _sβ(k)cos(ω _sT _s)+U _sβ(k)sin(ω _sT _s)

In formula, U _sα(k+1), U _sβ(k+1) be respectively t _k+1Stator voltage vector U constantly _{S α β}(k+1) α phase and β phase component;

(8) by t _kStator voltage vector U constantly _{S α β}(k), stator current vector I _{S α β}(k), rotor current vector I _{R α β}(k) and the rotor voltage vector V _{R α β}(k), prediction t _k+1Rotor current vector I constantly _{R α β}(k+1); The prediction expression formula as shown in Equation 7;

$\left[\begin{array}{c}{I}_{\mathrm{r\&alpha;}}(k+1)\\ L_{\mathrm{r\&beta;}}(k+1)\end{array}\right]=-{\mathrm{\&delta;T}}_s{L}_m\left[\begin{array}{c}{U}_{\mathrm{s\&alpha;}}\left(k\right)\\ U_{\mathrm{s\&beta;}}\left(k\right)\end{array}\right]+{\mathrm{\&delta;T}}_s{L}_s\left[\begin{array}{c}{V}_{\mathrm{r\&alpha;}}\left(k\right)\\ V_{\mathrm{r\&beta;}}\left(k\right)\end{array}\right]+\left[\begin{array}{cc}1-{\mathrm{\&delta;T}}_s{L}_s{R}_r& {-\mathrm{\&delta;T}}_s{L}_s{L}_r{\mathrm{\&omega;}}_r\\ {\mathrm{\&delta;T}}_s{L}_s{L}_r{\mathrm{\&omega;}}_r& 1-{\mathrm{\&delta;T}}_s{L}_s{R}_r\end{array}\right]\left[\begin{array}{c}{I}_{\mathrm{r\&alpha;}}\left(k\right)\\ I_{\mathrm{r\&beta;}\left(k\right)}\end{array}\right]+---\left(7\right)$

${\mathrm{\&delta;T}}_s{L}_m\left[\begin{array}{cc}{R}_s& -L_s{\mathrm{\&omega;}}_r\\ L_s{\mathrm{\&omega;}}_r& R_s\end{array}\right]\left[\begin{array}{c}{I}_{\mathrm{s\&alpha;}}\left(k\right)\\ I_{\mathrm{s\&beta;}}\left(k\right)\end{array}\right]$

In formula, I _rα(k+1), I _rβ(k+1) be respectively t _k+1Rotor current vector I constantly _{R α β}(k+1) α phase and β phase component;

(9) t that is obtained by front 3 steps prediction _k+1Stator voltage vector U constantly _{S α β}(k+1), rotor current vector I _{R α β}(k+1), meritorious, the reactive power signals P of stator output _s(k+1), Q _s(k+1) and 8 rotor voltage vector V that might apply in k+1 sampling period _{R α β}(k+1), prediction is applying respectively these rotor voltage vector V _{R α β}(the t zero hour in k+2 sampling period (k+1) _k+2All possible meritorious, reactive power signals P constantly) _s(k+2), Q _s(k+2); The prediction expression formula as shown in Equation 8;

$\left[\begin{array}{c}{P}_s(k+2)\\ Q_s(k+2)\end{array}\right]=\left[\begin{array}{cc}1-{\mathrm{\&delta;T}}_s{L}_r{R}_s& {-\mathrm{\&omega;}}_{\mathrm{slip}}{T}_s\\ {\mathrm{\&omega;}}_{\mathrm{slip}}{T}_s& 1-{\mathrm{\&delta;T}}_s{L}_r{R}_s\end{array}\right]\left[\begin{array}{c}{P}_s(k+1)\\ Q_s(k+1)\end{array}\right]+$

$\frac{{3\mathrm{\&delta;T}}_s{L}_m}{2}\left[\begin{array}{cc}{-U}_{\mathrm{s\&alpha;}}(k+1)& {-U}_{\mathrm{s\&beta;}}(k+1)\\ {-U}_{\mathrm{s\&beta;}}(k+1)& U_{\mathrm{s\&alpha;}}(k+1)\end{array}\right]\left[\begin{array}{c}{V}_{\mathrm{r\&alpha;}}(k+1)-R_r{I}_{\mathrm{r\&alpha;}}(k+1)\\ V_{\mathrm{r\&beta;}}(k+1)-R_r{I}_{\mathrm{r\&beta;}}(k+1)\end{array}\right]+---\left(8\right)$

$\frac{{3\mathrm{s\&delta;T}}_s{\mathrm{<figure-callout\; id="2"\; label="L\; r"\; filenames="HDA00002743376900012.png,HDA00002743376900021.png"\; state="\{\{state\}\}">L</figure-callout>}}_r}{2}\left[\begin{array}{c}{\left|U\right|}^2\\ 0\end{array}\right]$

In formula, the rotor voltage vector V _{R α β}(k+1) by the S of all possible rotor-side converter switches signal within k+1 sampling period _a1(k+1), S _b1(k+1), S _c1(k+1) obtain computing formula cotype 3; S _a1(k+1), S _b1(k+1), S _c1(k+1) have 2 ³Totally 8 kinds of combinations are therefore can get 8 kinds of rotor voltage vector V _{R α β}(k+1), measurablely obtain 8 t _k+2Constantly possible meritorious, reactive power signals P _s(k+2), Q _s(k+2);

(10) 8 t that prediction obtained _k+2Constantly possible meritorious, reactive power signals P _s(k+2), Q _s(k+2) with given t _k+2Constantly meritorious, reactive power reference signal P _s ^*, Q _s ^*In the substitution target function, obtain 8 target function values, relatively its size, the wherein minimum corresponding rotor voltage vector V of target function value _{R α β}(k+1), be t _k+1The rotor voltage vector V that moment rotor converter should apply _{R α β}(k+1), its switching signal S _a1(k+1), S _b1(k+1), S _c1(k+1) as t _k+1The switching signal of moment rotor-side converter; Target function as shown in Equation 9;

$g=|{P_s}^{*}-P_s(k+2)|+|{Q_s}^{*}-Q_s(k+2)|---\left(9\right)$

The rotor-side converter switches signal S of k+1 the sampling period planted agent output that (11) will obtain _a1(k+1), S _b1(k+1), S _c1(k+1) at t _k+1Constantly drive through driver module the optimizing power PREDICTIVE CONTROL that IGBT realizes RSC1.

Described GSC12 optimizing power forecast Control Algorithm comprises the following steps:

(1) establishing the zero hour in this sampling period is t _kConstantly, utilize the 3rd three-phase current Hall element 13-3 to gather the three phase network current signal I of net side _Gabc

(2) with the three phase network current signal I that collects _GabcTo two-phase coordinate transformation module 5-2, obtain t through the second static three-phase _kPower network current vector I under moment rest frame _{G α β}(k); Conversion expression formula cotype 1;

(3) by the grid side converter switching signal S in this sampling period (k sampling period) that calculated in the upper sampling period (k-1 sampling period) _a2(k), S _b2(k), S _c2(k) calculate t _kThe control voltage vector V that moment grid side converter applies _{C α β}(k); Computing formula as shown in Equation 10;

$\left[\begin{array}{c}{V}_{\mathrm{c\&alpha;}}\left(k\right)\\ V_{\mathrm{c\&beta;}}\left(k\right)\end{array}\right]=\frac{2}{3}{V}_{\mathrm{dc}}\left[\begin{array}{ccc}1& -\frac{1}{2}& -\frac{1}{2}\\ 0& \frac{\sqrt{3}}{2}& -\frac{\sqrt{3}}{2}\end{array}\right]\left[\begin{array}{c}{S}_{a2}\left(k\right)\\ S_{b2}\left(k\right)\\ S_{c2}\left(k\right)\end{array}\right]---\left(10\right)$

In formula, V _cα(k), V _cβ(k) be respectively t _kControl voltage vector V constantly _{C α β}(k) α phase and β phase component;

(4) by t _kLine voltage vector U constantly _{G α β}With power network current vector I _{G α β}Calculate t _kConstantly input meritorious, the reactive power signals P of grid side converter _g(k), Q _g(k); Meritorious, reactive power is calculated formula as shown in Equation 11;

$P_g\left(k\right)=\frac{3}{2}[U_{\mathrm{g\&alpha;}}\left(k\right)I_{\mathrm{g\&alpha;}}\left(k\right)+U_{\mathrm{g\&beta;}}\left(k\right)I_{\mathrm{g\&beta;}}\left(k\right)]---\left(11\right)$ $Q_g\left(k\right)=\frac{3}{2}[-U_{\mathrm{g\&alpha;}}\left(k\right)I_{\mathrm{g\&beta;}}\left(k\right)+U_{\mathrm{g\&beta;}}\left(k\right)I_{\mathrm{g\&alpha;}}\left(k\right)]$

In formula, t _kLine voltage vector U in moment rest frame _{G α β}(k) be stator voltage vector U _{S α β}(k); U _gα(k), U _gβ(k) be respectively t _kLine voltage vector U constantly _{G α β}(k) α phase and β phase component;

(5) with t _kMeritorious, reactive power signals P constantly _g(k), Q _g(k), line voltage vector U _{G α β}And t (k), _kThe control voltage vector V that constantly applies _{C α β}(k), (t zero hour in next sampling period of prediction (k+1 sampling period) _k+1Input grid side converter constantly) is meritorious, reactive power signals P _g(k+1), Q _g(k+1), the prediction expression formula as shown in Equation 12;

$\left[\begin{array}{c}{P}_g(k+1)\\ Q_g(k+1)\end{array}\right]=\left[\begin{array}{cc}1& {-\mathrm{\&omega;}}_s{T}_s\\ {\mathrm{\&omega;}}_s{T}_s& 1\end{array}\right]\left[\begin{array}{c}{P}_g\left(k\right)\\ Q_g\left(k\right)\end{array}\right]-\frac{{3T}_s}{{2L}_g}\left[\begin{array}{cc}{-U}_{\mathrm{g\&alpha;}}\left(k\right)& {-U}_{\mathrm{g\&beta;}}\left(k\right)\\ {-U}_{\mathrm{g\&beta;}}\left(k\right)& U_{\mathrm{g\&alpha;}}\left(k\right)\end{array}\right]\left[\begin{array}{c}{V}_{\mathrm{c\&alpha;}}\left(k\right)\\ V_{\mathrm{c\&beta;}}\left(k\right)\end{array}\right]----\left(12\right)$

$\frac{{3T}_s}{{2L}_g}\left[\begin{array}{c}{\left|U\right|}^2\\ 0\end{array}\right]$

In formula, L _gBe the grid side converter filter inductance;

(6) t that is obtained by prediction before _k+1Line voltage vector U constantly _{G α β}, the input grid side converter meritorious, reactive power signals P _g(k+1), Q _g(k+1) and 8 control voltage vector V that might apply in k+1 sampling period _{C α β}(k+1), prediction is applying respectively these control voltage vectors V _{C α β}(the t zero hour in k+2 sampling period (k+1) _k+2All possible meritorious, reactive power signals P constantly) _g(k+2), Q _g(k+2); Here, t _k+1Line voltage vector U constantly _{G α β}Be t _k+1Stator voltage vector U constantly _{S α β}(k+1); The prediction expression formula as shown in Equation 13;

$\left[\begin{array}{c}P(k+2)\\ Q(k+2)\end{array}\right]=\left[\begin{array}{cc}1& {-\mathrm{\&omega;}}_s{T}_s\\ {\mathrm{\&omega;}}_s{T}_s& 1\end{array}\right]\left[\begin{array}{c}P(k+1)\\ Q(k+1)\end{array}\right]-\frac{{3T}_s}{{2L}_g}\left[\begin{array}{cc}{-U}_{\mathrm{g\&alpha;}}(k+1)& {-U}_{\mathrm{g\&beta;}}(k+1)\\ {-U}_{\mathrm{g\&beta;}}(k+1)& U_{\mathrm{g\&alpha;}}(k+1)\end{array}\right]\left[\begin{array}{c}{V}_{\mathrm{c\&alpha;}}(k+1)\\ V_{\mathrm{c\&beta;}}(k+1)\end{array}\right]----\left(13\right)$

$\frac{{3T}_s}{{2L}_g}\left[\begin{array}{c}{\left|U\right|}^2\\ 0\end{array}\right]$

In formula, control voltage vector V _{C α β}(k+1) by all possible grid side converter switching signal S within k+1 sampling period _a2(k+1), S _b2(k+1), S _c2(k+1) obtain computing formula cotype 10; S _a2(k+1), S _b2(k+1), S _c2(k+1) have 2 ³Totally 8 kinds of combinations are controlled voltage vector V therefore can get 8 kinds _{C α β}(k+1), measurablely obtain 8 t _k+2Constantly possible meritorious, reactive power signals P _g(k+2), Q _g(k+2);

(7) 8 t that prediction obtained _k+2Constantly possible meritorious, reactive power signals P _g(k+2), Q _g(k+2) with given meritorious, reactive power reference signal

In the substitution target function, obtain 8 target function values, relatively its size, the wherein minimum corresponding control voltage vector of target function value V _{C α β}(k+1), be t _k+1The control voltage vector V that moment grid side converter should apply _{C α β}(k+1), its switching signal S _a2(k+1), S _b2(k+1), S _c2(k+1) as t _k+1Moment grid side converter switching signal; Target function as shown in Equation 14;

$g=|{P_g}^{*}-P_g(k+2)|+|{Q_g}^{*}-Q_g(k+2)|---\left(14\right)$

K+1 sampling period planted agent output that (8) will obtain grid side converter switching signal S _a2(k+1), S _b2(k+1), S _c2(k+1) at t _k+1Constantly drive through driver module the optimizing power PREDICTIVE CONTROL that IGBT realizes GSC12.。

With reference to Fig. 3, the actual value of S (t) expression power, U _i(i=1 ... 8) represent 8 kinds of different voltage vectors of inverter output, T _sThe sampling period of representative system, S (t) ^*Represent the reference value of power, S _pi(t _k+1) (i=1 ... 8) for adopting voltage vector U _iThe time power predicted value.Principle is, according to the predicted value S of power _pi(t _k+1) (i=1 ... 8) select to make the control behavior of target function minimum, can make actual power near given power.Can be found out at t by figure (a) _k+1Moment S _p3(t _k) near S (t _k+1) ^*So, at t _kConstantly select to apply U ₃, in like manner, at t _k+1Constantly select to apply U ₂, the rest may be inferred.In real system, the sampling of system, calculating can't instantaneously be completed, and time-delay certainly exists, and need to carry out compensation of delay.Suppose at t _kConstantly, the prediction by back has determined to apply U ₃, can be according to U ₃And the S (t that records _k) to S _p3(t _k+1) carry out first step prediction, then according to S _p3(t _k+1) and U _iTo S _pi(t _k+2) carry out second step prediction, visible t _k+2Moment S _pi(t _k+2) near set-point, so at t _k+1Constantly select to apply U ₂, the rest may be inferred, that is: the voltage vector of selecting next control cycle to apply in current control cycle.Through the prediction of two steps, as long as complete control logic in a control cycle, the control of system time-delay can be eliminated.

With reference to Fig. 4,5,6, can find out through the optimizing power forecast Control Algorithm after the prediction lag compensation of two steps and can effectively eliminate fluctuation meritorious, reactive power, reduce the harmonic content of electric current, have with based on the identical dynamic response performance fast of the direct Power Control of switch list.

In sum, method control structure disclosed by the invention is very simple, need not to carry out complicated coordinate transform and controller parameter adjusts, need not stagnant ring controller and the vector sector is judged, also need not to consider according to the sector system of selection of vector, simultaneously the variation of net side filter inductance parameter had than strong robustness, can realize that dual-feed asynchronous wind power generator is meritorious, independent, effective control of reactive power, reach dynamic response and steady-state behaviour control effect preferably.

Claims (3)

1. double-fed asynchronous wind generator system coordination optimization power prediction control method, it is characterized in that, the method comprises double-fed asynchronous wind generator system rotor side converter RSC(1) optimizing power forecast Control Algorithm and grid side converter GSC(12) the optimizing power forecast Control Algorithm; Described rotor-side converter RSC(1) optimizing power forecast Control Algorithm and grid side converter GSC(12) the optimizing power forecast Control Algorithm be within k sampling period, to process in microprocessor after the sampled signal discretization, to obtain the rotor-side converter switches signal S of k+1 sampling period planted agent output _a1(k+1), S _b1(k+1), S _c1(k+1) and grid side converter switching signal S _a2(k+1), S _b2(k+1), S _c2(k+1); If first sampling period internal rotor side converter switching signal S _a1(1), S _b1(1), S _c1(1) and grid side converter switching signal S _a2(1), S _b2(1), S _c2(1) output is 0;

Described RSC(1) optimizing power forecast Control Algorithm comprises the following steps:

1) establishing this zero hour in sampling period is t _k, utilize single-phase voltage Hall element (17) to gather DC side busbar voltage V _dcUtilize three voltage hall sensors (14) to gather dual-feed asynchronous wind power generator DFIG(18) threephase stator voltage signal U _SabcUtilize the first three-phase current Hall element (13-1) to gather threephase stator current signal I _Sabc, utilize the second three-phase current Hall element (13-2) to gather three-phase rotor current signal I _Rabc

2) the threephase stator voltage signal U that collects _SabcDetect the angular frequency that obtains electrical network or stator voltage through phase-locked loop (6) _sAdopt simultaneously encoder (10) to detect the rotor position of DFIG _r, then through differentiator (7) Calculation Speed ω _rObtain the slippage angular frequency by what subtracter calculated _Slip=ω _s-ω _r

3) with the threephase stator voltage signal U that collects _SabcWith threephase stator current signal I _SabcTo two-phase coordinate transformation module (5-1), obtain t through the first static three-phase _kStator voltage vector U under moment stator coordinate _{S α β}(k) and stator current vector I _{S α β}(k); With the three-phase rotor current signal I that collects _RabcAccording to rotor position _rRotate to two-phase static coordinate conversion module (9) through two-phase, obtain t _kRotor current vector I under moment stator coordinate _{R α β}(k);

4) according to rotor position _rAnd the rotor converter switching signal S of output in this sampling period (k sampling period) that calculated in the upper sampling period (k-1 sampling period) _a1(k), S _b1(k), S _c1(k), calculate t through module (8) _kRotor voltage vector V under the stator coordinate that moment rotor-side converter applies _{R α β}(k);

5) t that sampling is obtained _kStator voltage vector U constantly _{S α β}(k), stator current vector I _{S α β}(k), rotor current vector I _{R α β}(k), slippage angular frequency _SlipAnd t _kThe rotor voltage vector V that constantly applies _{R α β}(k), by rotor-side optimizing power prediction module (4-1) (t predict next sampling period (k+1 sampling period) zero hour _k+1Stator voltage vector U constantly) _{S α β}(k+1), rotor current vector I _{R α β}(k+1) and meritorious, the reactive power signals P of stator output _s(k+1), Q _s(k+1);

6) according to the t that predicts in (5) step _k+1Stator voltage vector U constantly _{S α β}(k+1), rotor current vector I _{R α β}(k+1), meritorious, the reactive power signals P of stator output _s(k+1), Q _s(k+1) and 8 rotor voltage vector V that might apply in k+1 sampling period _{R α β}(k+1), predict by rotor-side second step prediction module (3-1) and applying respectively these rotor voltage vector V _{R α β}Under (the t zero hour in k+2 sampling period _k+2All possible meritorious, reactive power signals P constantly) _s(k+2), Q _s(k+2); Can obtain 8 groups of meritorious, reactive power signals P _s(k+2), Q _s(k+2);

7) with the t that predicts in (6) step _k+2Constantly 8 groups of meritorious, reactive power signals P _s(k+2), Q _s(k+2) and t _k+2Meritorious, the reactive power reference signal P of given rotor-side of the moment _s ^*, Q _s ^*, obtain through the minimization of object function control module (2-1) the rotor-side converter switches signal S that k+1 sampling period planted agent exports _a1(k+1), S _b1(k+1), S _c1(k+1);

The rotor-side converter switches signal S of k+1 the sampling period planted agent output that 8) will obtain _a1(k+1), S _b1(k+1), S _c1(k+1) at t _k+1Constantly drive IGBT through driver module and realize RSC(1) the optimizing power PREDICTIVE CONTROL.

Described GSC(12) optimizing power forecast Control Algorithm comprises the following steps:

1) establishing the zero hour in this sampling period is t _k, utilize the 3rd three-phase current Hall element (13-3) Gather and input GSC(12) the three phase network current signal I that flows through filter inductance (15) _Gabc

2) the three phase network current signal I that collects _GabcTo two-phase coordinate transformation module (5-2), obtain t through the second static three-phase _kGSC(12 under moment rest frame) power network current vector I _{G α β}(k);

3) by the interior grid side converter switching signal S of this sampling period (k sampling period) that calculated in the upper sampling period (k-1 sampling period) _a2(k), S _b2(k), S _c2(k) calculate t through module (11) _kControl voltage vector V under the rest frame that moment grid side converter applies _{C α β}(k);

4) t that sampling is obtained _kMoment line voltage vector U _{G α β}(k), power network current vector I _{G α β}(k) and t _kThe control voltage vector V that constantly applies _{C α β}(k), by net side optimizing power prediction module (4-2) (t predict next sampling period (k+1 sampling period) zero hour _k+1Line voltage vector U constantly) _{G α β}(k+1) and the input grid side converter is meritorious, reactive power signals P _g(k+1), Q _g(k+1); Here, t _kLine voltage vector U constantly _{G α β}(k) be stator voltage vector U _{S α β}(k);

5) t that obtains according to prediction in (4) step _k+1Line voltage vector U constantly _{G α β}(k+1), meritorious, the reactive power signals P of input grid side converter _g(k+1), Q _g(k+1) and 8 control voltage vector V that might apply in k+1 sampling period _{C α β}(k+1), applying respectively these control voltage vectors V by net side second step prediction module (3-2) prediction _{C α β}(k+1) under k+2 the zero hour in sampling period (t _k+2All possible meritorious, reactive power signals P constantly) _g(k+2), Q _g(k+2); Can obtain 8 groups of meritorious, reactive power signals P _g(k+2), Q _g(k+2);

6) with the t that predicts in (5) step _k+2Constantly 8 groups of meritorious, reactive power signals P _g(k+2), Q _g(k+2) and t _k+2Meritorious, the reactive power reference signal P of given net side of the moment _g ^*, Q _g ^*, obtain grid side converter switching signal S in k+1 sampling period through the minimization of object function control module (2-2) _a2(k+1), S _b2(k+1), S _c2(k+1).

The grid side converter switching signal S of k+1 the sampling period planted agent output that 7) will obtain _a2(k+1), S _b2(k+1), S _c2(k+1) at t _k+1Constantly drive IGBT through driver module and realize GSC(12) the optimizing power PREDICTIVE CONTROL.

2. a kind of double-fed asynchronous wind generator system coordination optimization power prediction control method according to claim 1, is characterized in that, RSC(1) the rotor-side optimizing power prediction module (4-1) described in the optimizing power forecast Control Algorithm can be by t _kThe baseband signal of moment rotor-side is as rotor-side stator voltage vector U _{S α β}(k), stator current vector I _{S α β}(k), rotor current vector I _{R α β}(k), slippage angular frequency _Slip, and t _kControl signal constantly is as sided rotor voltage vector V _{R α β}(k), predict next t constantly _k+1Meritorious, the reactive power signals P of stator output _s(k+1), Q _s(k+1); GSC(12) the net side optimizing power prediction module (4-2) described in the optimizing power forecast Control Algorithm can be by t _kThe baseband signal of moment net side is as net side line voltage vector U _{G α β}(k), power network current vector I _{G α β}(k), reach t _kControl signal is constantly controlled voltage vector V as the net side _{C α β}(k), predict next t constantly _k+1Meritorious, the reactive power signals P of input grid side converter _g(k+1), Q _g(k+1).

3. a kind of double-fed asynchronous wind generator system coordination optimization power prediction control method according to claim 1, it is characterized in that RSC(1) optimizing power forecast Control Algorithm and GSC(12) the minimization of object function control module (2-1), (2-2) described in the optimizing power forecast Control Algorithm, target function can be set according to the actual requirements, have certain flexibility and can reach different control effects, generally choose g=|P ^*-P|+|Q ^*-Q| after calculating applies target function under different control signals, chooses the control signal that makes the target function value minimum, and the space vector as optimum obtains t _k+1Moment rotor-side converter switches signal S _a1(k+1), S _b1(k+1), S _c1(k+1) and grid side converter switching signal S _a2(k+1), S _b2(k+1), S _c2(k+1).

Images