CN103259485B - Method of improving identification precision of speedless sensor under condition of unbalanced network voltage - Google Patents

Abstract

The invention relates to a method of improving identification precision of a speedless sensor under the condition of unbalanced network voltage. The method comprises the steps that a wind power generator controlled by a double-fed control system is adopted; when the network voltage is in a balanced state, stator voltage and stator current serve as an input of an MRAS (Model Reference Adaptive System) reference model directly; and rotating speed data is obtained after PI (Proportional-Integral) adaptive rotating speed estimation. An improvement of the method is that when the network voltage is in an unbalanced state, negative sequence components of the voltage and the current due to unbalance of the network voltage are removed from the input quantity of the MRAS reference model. The method has the technical benefits that the speedless sensor meets a control requirement that the power generator works in the slight unbalanced state of the network voltage; the identification precision of the speedless sensor on the rotating speed in the unbalanced state of the network voltage is improved; an application scope of the speedless sensor is expanded; and fault toleration control of the double-fed wind power generator during a fault of a speed sensor is achieved.

Description

The method of Speedless sensor identification precision is improved under unbalanced source voltage condition

Technical field

The present invention relates to a kind of double-fed wind power generator control technology, particularly relate to a kind of method improving Speedless sensor identification precision under unbalanced source voltage condition.

Background technology

At present, double-fed wind power generator (DFIG) has become the mainstream model of MW class wind turbine group with the advantage of its variable speed constant frequency generator; In dual-feedback wind power generator control system, the rate signal of rotor is the important parameter of double-fed wind power generator vector control, play vital effect in the controls, therefore, obtain accurate rotary speed data for double-fed wind power generator control significant.

In prior art, the means for obtaining generator speed are generally had the speed sensors such as optical code disk and by account form, tach signal are carried out to the Speedless sensor two kinds of online observation; Based on high performance Hardware Arrangement, under normal operation, no matter which kind of operating state generator is in, speed sensor can get rotary speed data comparatively accurately, but also just because speed sensor is to the dependence of hardware device, cause the cost containing speed sensor in generator cost, moreover, speed sensor also needs periodic maintenance, the fault of its hardware and aging be inevitable, therefore researchers have just turned one's attention to Speedless sensor; Speedless sensor has structure advantage simple, with low cost because of it, more and more receives the concern of researchers.At present, researchers are just being devoted to improve further the accuracy of identification of Speedless sensor and are being in the adaptability under different running status to generator, so that some day, enable Speedless sensor substitute speed sensor completely.Just the dual-feedback wind power generator control system of Speedless sensor is adopted separately to analyze below;

In actual electric power system, need the long-time running of wind power generation function under the slight non-equilibrium state of line voltage, and energy short-time duty is when line voltage occur suddenly seriously to fall; At present in the failure diagnosis and control design case of transducer, all do not consider the reality (this is that the impact caused speed sensor due to unbalanced source voltage is relatively little) of unbalanced source voltage, and relative to speed sensor, the Speed identification precision based on the Speedless sensor of model reference adaptive system (MRAS) is more vulnerable to the impact of the negative sequence component in unbalanced source voltage.Based in the Speed identification of MRAS, be exactly want accurately to the basic demand of reference model, because reference model is the benchmark controlled as MRAS, only have this reference model accurate, Speedless sensor just can obtain more believable identifier, and in prior art, owing to not considering the reality of unbalanced source voltage, therefore special process is not done to the input variable of MRAS reference model, cause in the voltage of MRAS reference model, electric current input variable not only containing positive sequence component but also containing negative sequence component; In actual motion, even if less negative sequence voltage, current component, all be there is larger error (this is also a major issue of can not ignore in the research of double-fed wind power generator Speedless sensor) to tach signal online observation in the Speedless sensor caused based on MRAS.If unbalanced source voltage, double-fed wind power generator adopts separately Speedless sensor to be difficult to obtain speed measured value accurately, to directly affect the operation of whole unit, the instability of system even can be caused time serious, therefore, under unbalanced source voltage, double-fed wind power generator is in the Speed identification based on MRAS Speedless sensor, Speed identification precision how is avoided to be subject to negative sequence voltage, the current component even impact of harmonic wave interference components, improve the identification precision of Speedless sensor, this is the problem that in Speedless sensor research, letter is to be solved.

Summary of the invention

When electric power system asymmetric operation, theoretical according to Instantaneous Symmetrical Components, positive sequence, negative phase-sequence and zero-sequence component sum can be decomposed into.Consider current double-fed wind power generator industry unit all adopt three-phase three-line system and bulk power grid to be connected, namely in Circuits System without zero-sequence component path, therefore can ignore the existence of zero-sequence component.The negative phase-sequence air-gap rotating magnetic field produced due to negative-sequence current is contrary with rotor direction of rotation, induces two times of power currents, i.e. so-called frequency multiplication electric current in the magnetic pole winding of generator; The magnetic field that these two times of power currents produce induces again three times of power currents in the stator windings, mutual induction like this, so there is a series of even-order harmonic (2,4,6 in rotor windings ...) electric current, there is a series of odd harmonic (3,5,7 in the stator windings ...) electric current, the harmonic component of electric current in electric power system, voltage is increased; Mentioned effects brings the harm of two aspects: on the one hand, not only makes generator electromotive force waveform distort, and makes rotor surface and stator winding occur local overheating phenomenon, can cause usually said " negative-sequence current burning machine " time serious; Two is produce the alternating electromagnetism torque of 100Hz by negative-sequence current, and this alternating electromagnetism torque acts on rotor shaft and out frame simultaneously, causes frequency to be the vibration of 100Hz per second, likely damages the structures such as support.

According to above-mentioned analysis, and for the problem in background technology, the present invention proposes a kind of method improving Speedless sensor identification precision under unbalanced source voltage condition, comprise: the wind-driven generator adopting doubly-fed control Systematical control, the rotary speed data that described doubly-fed control system obtains according to Speedless sensor controls wind-driven generator rotating speed; Described Speedless sensor carries out the process of PI self adaptation speed estimate by MRAS reference model to stator voltage, stator current, obtains rotary speed data; When line voltage is in poised state, directly using stator voltage and stator current as the input of MRAS reference model, after the process of PI self adaptation speed estimate, obtain rotary speed data; It is characterized in that: when unbalanced source voltage, obtain rotary speed data by the following method:

1) extract real-time generator unit stator voltage, the positive sequence component of stator current under dq rotating coordinate system, obtains the positive sequence voltage under dq rotating coordinate system with the forward-order current under dq rotating coordinate system

2) will with as the input of MRAS reference model, after the process of PI self adaptation speed estimate, obtain rotary speed data;

In preceding method, due to abandoned in the reference model input variable in MRAS cause because of unbalanced source voltage voltage, electric current negative sequence component interference, make the rotary speed data of institute's identification only by the impact of positive sequence component, thus make Speedless sensor also can provide Speed Identification data accurately for doubly-fed control system when unbalanced source voltage, improve Speedless sensor to the adaptability of operation state of generator.

Aforesaid PI self adaptation speed estimate process adopts MRAS method to realize: using the stator flux observer not containing actual revolution and torque calculation as with reference to model, using the PI adaptive rate containing rotating speed to be identified as adjustable model, first acquisition speed adjuster output with the T of stator flux observer and torque calculation gained _e, ask for both error signals this error signal is sent into PI proportional integral adaptive law, its output is velocity estimation value

${\widehat{\mathrm{\ω}}}_r=(K_p+\frac{K_i}{s})({\hat{T}}_e^{*}-T_e)$

Wherein, for velocity estimation value, for pi regulator, K _p, K _ibe respectively proportionality coefficient and integral coefficient; for the output signal of speed regulator, T _eobtained by stator flux observer and torque calculation.

In the content of aforementioned schemes, relate to and need whether be in line voltage the problem that non-equilibrium state judges, based on prior art, inventors herein propose following preferred version to judge whether line voltage is in non-equilibrium state, concrete scheme is:

1] positive and negative sequence component of extract real-time generator unit stator voltage under α β rest frame, obtains α axle positive sequence voltage component β axle positive sequence voltage component α axle negative sequence voltage components with β axle negative sequence voltage components

2] positive sequence voltage amplitude is calculated according to following formula with negative sequence voltage amplitude

$u_{\mathrm{sm}}^{+}=\sqrt{{\left(u_{\mathrm{s\α}}^{+}\right)}^2+{\left(u_{\mathrm{s\β}}^{+}\right)}^2}$

$u_{\mathrm{sm}}^{-}=\sqrt{{\left(u_{\mathrm{s\α}}^{-}\right)}^2+{\left(u_{\mathrm{s\β}}^{-}\right)}^2}$

3] unbalanced source voltage degree δ is calculated according to following formula:

$\mathrm{\δ}=\frac{u_{\mathrm{sm}}^{-}}{u_{\mathrm{sm}}^{+}}\×100\%;$

4] when δ >=5%, judge that line voltage is in non-equilibrium state; As δ < 5%, judge that line voltage is in poised state.

In order to ensure the stability that double-fed generator runs, prior art can be combined with aforementioned schemes, building following faults-tolerant control scheme:

Adopt speed sensor and Speedless sensor to obtain generator speed data, the rotary speed data that speed sensor gets is designated as speed A, and the rotary speed data that Speedless sensor gets is designated as speed B simultaneously; During speed sensor fault-free, doubly-fed control system controls wind-driven generator rotating speed according to speed A, when speed sensor breaks down, adopts speed B to control wind-driven generator rotating speed.

Aforementioned faults-tolerant control scheme in the specific implementation, needs for speed sensor and Speedless sensor each independent configuration speed regulator.

In aforementioned faults-tolerant control scheme, relate to and need whether the problem that fault judges is existed to speed sensor, use for reference prior art, adopt following scheme to carry out method and judge whether speed sensor exists fault:

A, be calculated as follows fault residual Δ:

Δ＝|ω _A-ω _B|

Wherein, ω _afor the Current observation value of speed A, ω _bfor the Current observation value of speed B;

B, be calculated as follows failure diagnosis adaptive threshold e:

e＝0.2·ω _B

C, Δ and e to be compared: if Δ >=e, then judge that speed sensor breaks down; If Δ < is e, then judge speed sensor fault-free.

Advantageous Effects of the present invention is: make Speedless sensor be adapted to the control needs of generator operation under the slight non-equilibrium state of line voltage, improve Speedless sensor under unbalanced source voltage state to the accuracy of identification of rotating speed, the scope of application of expansion Speedless sensor, also achieves the faults-tolerant control of double-fed wind power generator when speed sensor fault simultaneously.

Accompanying drawing explanation

Fig. 1, conventional doubly-fed control systematic schematic diagram;

The doubly-fed control systematic schematic diagram of Fig. 2, employing faults-tolerant control scheme of the present invention;

In Fig. 3, faults-tolerant control scheme, judge whether speed sensor exists the principle schematic (being also the schematic diagram of fault-tolerant control module in Fig. 2) of fault;

Fig. 4, unbalanced source voltage condition adjudgement principle schematic;

Under Fig. 5, α β rest frame, positive and negative sequence component extracts principle schematic;

Fig. 6, PI self adaptation speed estimate handling principle schematic diagram;

The flow chart of Fig. 7, faults-tolerant control scheme of the present invention.

Embodiment

The present invention is understood better for those skilled in the art can be made, be necessary to make a brief description to existing doubly-fed control system: Fig. 1 is conventional doubly-fed control systematic schematic diagram, can find out from this figure, rotor-side system adopts rotor current, rotor speed double-closed-loop control, wherein, inner ring is rotor current control ring, by detection rotor phase current, through 3s/2s(three phase static coordinate/two-phase static coordinate) and 2s/2r(two-phase static coordinate/two cordic phase rotator) two step conversion, obtain the i under synchronous rotating frame _rd, i _rqtwo control channels, the set-point i of rotor current _rd ^*, i _rq ^*with i _rd, i _rqerror relatively, the current regulator ACR(of band anomalous integral output violent change adopts PI mode) regulating rear output voltage control amount, this voltage control quantity superposes front feedback voltage compensation amount △ u respectively _rd, △ u _rqobtain rotor voltage controlled quentity controlled variable, this controlled quentity controlled variable is again through 2r/2s(two cordic phase rotators/two-phase static coordinate) conversion, exciting voltage after SVPWM modulation needed for generation generator amature side reality and exciting current, realize the uneoupled control of double-fed wind power generator vector control system rotor-side.In fig. 1, the outer shroud of vector control system is rotor speed control ring, similar with current inner loop, rotary speed setting value ω _r ^*gained difference after comparing with the speed feedback value recorded by speed sensor (optical code disk), adopts PI mode through speed regulator ASR() regulate after, obtain the Driving Torque T of generator _e ^*, Driving Torque calculates through torque current again, obtains the set-point i of rotor current real component _rq ^*.And the set-point i of rotor current idle component _rd ^*, normally calculate according to the idle requirement of electrical network to wind-driven generator vector control system.

Fig. 2 is the control system figure formed after adopting faults-tolerant control scheme of the present invention, with existing control method unlike: system adopts Redundancy Design, namely adopt speed sensor and Speedless sensor mode simultaneously, obtain wind power generator rotor rotating speed, the tachometer value recorded by speed sensor is as controlling under normal circumstances to use, and Speedless sensor observes the tachometer value obtained as subsequent use; When speed sensor breaks down, then enable Speedless sensor measured value.In the design of this system, consider that Speedless sensor is subject to the negative sequence component impact under unbalanced source voltage condition, therefore, the Speedless sensor measured value obtained based on the PI self adaptation speed estimate processing mode of MRAS under Voltage unbalance condition is applied in the main scheme of the present invention in this system, that is: when Speedless sensor works separately, in wind-driven generator running, first line voltage is just being carried out, negative separation, and judge whether line voltage balances, work as unbalanced source voltage, then suppress the interference of negative sequence voltage and negative-sequence current component, to only be input in " PI self adaptation speed estimate " module containing the signal of positive sequence voltage and forward-order current, to obtain the Speedless sensor measured value of more coincideing with actual speed signal, overcome the negative sequence component impact that existing Speedless sensor is subject under unbalanced source voltage state, Speedless sensor is caused to be difficult to carry out rotating speed the defect of accurately observation.

In fig. 2, judge whether speed sensor breaks down, can be determined by " faults-tolerant control " link, see Fig. 3, this link exports two paths of signals respectively to transducer A and transducer B, for the switching controls under speed sensor fault;

When " faults-tolerant control " link judges that speed sensor is normal, rotary speed setting value ω _r ^*with the speed feedback value ω recorded by speed sensor _rthe difference of gained relatively, after speed regulator ASR1 regulates, obtains the Driving Torque T of generator _e ^*, now, 1 in transducer A and 3 are connected by " faults-tolerant control " link, and 2 and 4 connect, by Driving Torque T _e ^*export torque current to and calculate link; Meanwhile, 1 in transducer B and 3 are connected by " faults-tolerant control " link, and 2 and 4 connect, the speed feedback value ω recorded by speed sensor _rexport voltage compensation link to.

When " faults-tolerant control " link judges speed sensor fault, enable Speedless sensor speed observation value for subsequent use, rotary speed setting value ω _r ^*with the speed feedback value observed by Speedless sensor the difference of (subscript " ^ " represents and is worth by observation) relatively rear gained, after speed regulator ASR2 regulates, obtains the Driving Torque of generator this Driving Torque two-way is divided to export: a road exports PI self adaptation speed estimate to, as Speed Identification; Transducer A is directly delivered on another road; Now, 1 in transducer A and 4 are connected by " faults-tolerant control " link, and 2 and 5 connect, by Driving Torque deliver to torque current and calculate link; Meanwhile, 1 in transducer B and 4 are connected by speed probe fault tolerance judgment controlling unit, and 2 and 5 connect, the speed feedback value recorded by Speedless sensor export voltage compensation link to.

Below in conjunction with accompanying drawing, " faults-tolerant control ", " unbalanced source voltage degree estimation & disposing " addressed in above, " PI self adaptation speed estimate " this three part are introduced respectively.

(1) " faults-tolerant control "

Speed probe fault type can be divided into hard fault and soft fault two class usually according to fault degree; Wherein, hard fault generally causes by transducer component damage or by reasons such as comparatively hard pulse interference, has output amplitude change feature greatly; And the fault that soft fault general reference causes due to the reason such as part aging, null offset, there is output amplitude change not quite and more slowly feature.

Fig. 3 is the schematic diagram of the fault-tolerant control module in Fig. 2, fault residual is obtained after speed sensor output valve and Speedless sensor observer value being compared by " calculating fault residual " link in figure, simultaneously, by " calculating failure diagnosis adaptive threshold " link dynamic calculation failure diagnosis adaptive threshold, with reference to the pertinent literature that adaptive threshold conventional in fault detect is chosen, get the adaptive threshold of 0.2 times of observer output valve as failure diagnosis, this threshold parameter can the adaptive change with the change of observer rotational speed output signal; Then by " breakdown judge " link, fault residual and failure diagnosis adaptive threshold are compared, in order to judge whether speed sensor breaks down, when fault residual is greater than failure diagnosis adaptive threshold, then speed sensor fault, otherwise speed sensor is normal.

(2) " unbalanced source voltage degree estimation & disposing "

Fig. 4 is the schematic diagram of " unbalanced source voltage degree estimation & disposing " module in Fig. 2; First voltage signal and the positive and negative sequence component of current signal under α β rest frame is extracted respectively with " the positive and negative sequence of current signal is separated " link, then by " positive sequence voltage amplitude by " separation of voltage signal positive and negative sequence " link in figure " link and " negative sequence voltage amplitude " link calculates positive sequence voltage amplitude and negative sequence voltage amplitude; " calculating unbalanced source voltage degree δ " link calculates unbalanced source voltage degree δ according to positive sequence voltage amplitude and negative sequence voltage crest meter; and control according to the action of size to transducer C and transducer D of voltage unbalance factor δ; while aforementioned processing is carried out; the positive sequence voltage after being separated and not separated voltage are all carried out 2s/2r(two-phase static coordinate/two cordic phase rotator) conversion, and export the two paths of signals after converting to transducer C; In like manner, current signal is also done the coordinate system transformation process similar with voltage, export the two paths of signals after conversion to transducer D; Consider and can bear transient state by Wind turbines to reach 5%(even higher) unbalance voltage and do not trip, using the separation whether δ=5% balances as line voltage, when calculating the δ tried to achieve and being less than this value, can think that line voltage is balance, 1 of transducer C and 3 connect, 2 and 4 connect, and transducer C exports the voltage u under dq rotating coordinate system _sdq; And when calculating the δ that tries to achieve and being more than or equal to 5%, unbalanced source voltage can be thought, connect 1 of transducer C and 4,2 and 5 connect, and transducer C exports the positive sequence voltage under dq rotating coordinate system current signal is also done similar process, and transducer D is the alternative current i exported under line voltage balance also _sdqor the forward-order current of line voltage under imbalance

Fig. 5 is the principle schematic of " the positive and negative sequence of voltage signal is separated " link in Fig. 4, the positive and negative sequence of voltage signal is separated existing " T/4 time delay (T is the electrical network fundamental voltage the cycle) " method of employing and realizes: under α β rest frame, the positive and negative sequence voltage after separation is shown below respectively.

$u_{\mathrm{s\&alpha;}}^{+}=\frac{1}{2}[u_{\mathrm{s\&alpha;}}\left(t\right)-u_{\mathrm{s\&beta;}}(t-T/4)]---\left(1\right)$

$u_{\mathrm{s\&alpha;}}^{-}=\frac{1}{2}[u_{\mathrm{s\&alpha;}}\left(t\right)+u_{\mathrm{s\&beta;}}(t-T/4)]---\left(2\right)$

$u_{\mathrm{s\&beta;}}^{+}=\frac{1}{2}[u_{\mathrm{s\&beta;}}\left(t\right)-u_{\mathrm{s\&alpha;}}(t-T/4)]---\left(3\right)$

$u_{\mathrm{s\&beta;}}^{-}=\frac{1}{2}[u_{\mathrm{s\&beta;}}\left(t\right)-u_{\mathrm{s\&alpha;}}(t-T/4)]---\left(4\right)$

In formula, parameters is respectively: α axle positive sequence voltage component β axle positive sequence voltage component α axle negative sequence voltage components β axle negative sequence voltage components α shaft voltage component u _{s α}(t), β shaft voltage component u _{s β}t (), T are the electrical network fundamental voltage cycle, t is the time.

Positive sequence voltage amplitude after separation can be tried to achieve by following formula:

$u_{\mathrm{sm}}^{+}=\sqrt{{\left(u_{\mathrm{s\&alpha;}}^{+}\right)}^2+{\left(u_{\mathrm{s\&beta;}}^{+}\right)}^2}---\left(5\right)$

Negative sequence voltage amplitude after separation can be tried to achieve by following formula:

$u_{\mathrm{sm}}^{-}=\sqrt{{\left(u_{\mathrm{s\&alpha;}}^{-}\right)}^2+{\left(u_{\mathrm{s\&beta;}}^{-}\right)}^2}---\left(6\right)$

Unbalanced source voltage degree is the percentage value of the ratio of negative sequence voltage and positive sequence voltage, namely

$\mathrm{\&delta;}=\frac{u_{\mathrm{sm}}^{-}}{u_{\mathrm{sm}}^{+}}\&times;100\%---\left(7\right)$

(3) " PI self adaptation speed estimate "

Fig. 6 is the schematic diagram of " PI self adaptation speed estimate " module in Fig. 2.PI self adaptation speed estimate make use of self adaptation thought, model reference adaptive (MRAS) mode is adopted to obtain speed measured value, it is a kind of respond well speed estimation method, its basic thought makes full use of control system existing structure exactly, using not containing the stator flux observer of actual revolution and torque calculation as with reference to model, using the PI adaptive rate containing rotating speed to be identified as adjustable model.Reference model is wherein for stator flux observer and torque T _ecalculating, be below the derivation to its formula and explanation.

Theoretical according to Electrical Motor, in synchronous rotating frame, double-fed generator Mathematical Modeling is:

$\left.\begin{array}{c}{u}_{\mathrm{sd}}=-R_s{i}_{\mathrm{sd}}+{\mathrm{\&omega;}}_1{\mathrm{\&psi;}}_{\mathrm{sq}}-p{\mathrm{\&psi;}}_{\mathrm{sd}},u_{\mathrm{sq}}=-R_s{i}_{\mathrm{sq}}-{\mathrm{\&omega;}}_1{\mathrm{\&psi;}}_{\mathrm{sd}}-{\mathrm{p\&psi;}}_{\mathrm{sq}}\\ u_{\mathrm{rd}}=R_r{i}_{\mathrm{rd}}-{\mathrm{\&omega;}}_s{\mathrm{\&psi;}}_{\mathrm{rq}}+p{\mathrm{\&psi;}}_{\mathrm{rd}},u_{\mathrm{rq}}=R_r{i}_{\mathrm{rq}}+{\mathrm{\&omega;}}_s{\mathrm{\&psi;}}_{\mathrm{rd}}+p{\mathrm{\&psi;}}_{\mathrm{rq}}\end{array}\right\}---\left(8\right)$

$\left.\begin{array}{c}{\mathrm{\&psi;}}_{\mathrm{sd}}=l_s{i}_{\mathrm{sd}}-l_m{i}_{\mathrm{rd}},{\mathrm{\&psi;}}_{\mathrm{sq}}=l_s{i}_{\mathrm{sq}}-l_m{i}_{\mathrm{rq}}\\ {\mathrm{\&psi;}}_{\mathrm{rd}}=-l_m{i}_{\mathrm{sd}}+l_r{i}_{\mathrm{rd}},{\mathrm{\&psi;}}_{\mathrm{rq}}=-l_m{i}_{\mathrm{sq}}+l_r{i}_{\mathrm{rq}}\end{array}\right\}---\left(9\right)$

$T_e=\frac{3}{2}{p}_N{l}_m(i_{\mathrm{sd}}{i}_{\mathrm{rq}}-i_{\mathrm{sq}}{i}_{\mathrm{rd}})---\left(10\right)$

In formula, R _sfor stator resistance, R _rfor rotor resistance, l _sfor stator inductance, l _rrotor resistance inductance; l _mfor mutual inductance between rotor; ω ₁for synchronous rotary angular speed, ω _rfor rotor velocity, ω _sfor slip angular velocity, ω _s=ω ₁-ω _r; P is differential operator, p _nfor number of pole-pairs, u _sd, u _sq, i _sd, i _sq, ψ _sd, ψ _sqbe respectively the voltage of stator dq axle under synchronous rotating frame, electric current and magnetic flux; u _rd, u _rq, i _rd, i _rq, ψ _rd, ψ _rqbe respectively the voltage of synchronous rotating frame lower rotor part dq axle, electric current and magnetic flux; T _efor electromagnetic torque.

For the Shunt-connected Wind Power Generation System, the impact of stator resistance can be ignored.When selecting the d axle of coordinate system along ψ ₁time directed, then have:

ψ _sd＝ψ ₁，ψ _sq＝0 （11）

u _sd＝0，u _sq＝-u ₁（12）

In formula, u ₁for line voltage effective value.

Formula (11), (12) substitute into formula (8):

${\mathrm{\&psi;}}_1=\frac{u_1}{{\mathrm{\&omega;}}_1}---\left(13\right)$

Formula (9) is substituted into formula (10):

$T_e=\frac{3}{2}{p}_N{i}_{\mathrm{sq}}{\mathrm{\&psi;}}_1---\left(14\right)$

Can be found out by (13), (14) formula: in MRAS system, calculate obtained electromagnetic torque T by reference model _edetermined by the physical quantity such as voltage, electric current of stator side completely.And when unbalanced source voltage, three-phase alternating current system includes voltage, the electromagnetic quantities such as electric current and magnetic linkage of positive and negative sequence composition.Based in the Speed identification of MRAS, be exactly want accurately to the basic demand of reference model.Because reference model is the benchmark controlled as MRAS, only have this reference model accurate, Speedless sensor just can obtain more believable identifier.Therefore, when electric power system asymmetric operation, by negative-sequence current cause electric current in electric power system, voltage harmonic component increase when, how to obtain stator flux observer value and electromagnetic torque Te value comparatively accurately and just seem extremely important.The present invention is exactly by positive and negative sequence component quick separating in unbalanced voltages, current system, with get rid of negative phase-sequence even harmonic component on the impact of Speedless sensor identification precision, using positive sequence voltage under two-phase rotating coordinate system, electric current as the input of MRAS reference model, to obtaining stator flux observer value and the calculating torque of more standard, thus improve Speedless sensor identification precision.

First acquisition speed adjuster ASR2 exports and by the above-mentioned electromagnetic torque T obtained by reference model _evalue, asks for both error signals this error signal is sent into proportional integral adaptive law (pi regulator), and its output is velocity estimation value that is:

${\widehat{\mathrm{\&omega;}}}_r=(K_p+\frac{K_i}{s})({\hat{T}}_e^{*}-T_e)---\left(15\right)$

Wherein, for velocity estimation value, for pi regulator, K _p, K _ibe respectively proportionality coefficient and integral coefficient; for speed regulator ASR2 outputs signal, T _eobtained by stator flux observer and torque calculation.

The velocity estimation value of gained namely the operation for double-fed wind power generator controls.

Fig. 7 gives the flow chart of faults-tolerant control scheme of the present invention.

For verifying effect of the present invention, breaking down and under unbalanced source voltage degree is the condition of δ=6%, emulation experiment having been carried out to the speed observation of double-fed speed-variable frequency-constant wind-driven generator being set with velocity transducer; Test wind turbine generator parameter used as follows: fan blade number 3, rotor diameter D=5m, optimum tip-speed ratio λ=9.5, gear box speed increasing ratio N=7.56, generator rating power P _n=4kW, number of pole-pairs p _n=2, moment of inertia J=0.87kg.m ².

Given and the Dynamic Speed of experiment point steady-state speed switches two kinds of situations and carries out, with the Identification Errors of Speedless sensor when more original scheme (situation when namely not rejected from the input voltage and electric current of MRAS reference model by negative sequence component) and employing the present invention program.Table 1 and table 2 are respectively under given different rotating speeds, adopt original scheme and adopt the present invention program to survey the table of comparisons of rotating speed and Speed Identification error; Table 3 and table 4 are by metasynchronism speed respectively when generator the 3rd second state is switched to supersynchronous speed state, adopts original scheme and adopts the present invention program to survey the table of comparisons of rotating speed and Speed Identification error.

Rotating speed and the Speed Identification error table of comparisons (original scheme) is surveyed under table 1 given rotating speed

Rotational speed setup (rpm)	1290	1340	1390	1440	1490	1540	1590
								Actual measurement rotating speed (rpm)	1289.3	1338.1	1391.5	1441.8	1488.5	1539.6	1588.3
Speed Identification (rpm)	1244.2	1303.3	1417.9	1502.4	1546.6	1565.8	1624.8
								Identification and the difference of surveying rotating speed	-45.13	-34.8	26.44	60.6	58.1	26.2	36.5
Percentage error (%)	-3.5	-2.6	1.9	4.2	3.9	1.7	2.3

Rotating speed and the Speed Identification error table of comparisons (the present invention program) is surveyed under table 2 given rotating speed

Rotational speed setup (rpm)	1290	1340	1390	1440	1490	1540	1590
								Actual measurement rotating speed (rpm)	1289.3	1338.1	1391.5	1441.8	1488.5	1539.6	1588.3
Speed Identification (rpm)	1277.7	1331.4	1409.6	1466.3	1504.8	1548.8	1596.2
								Identification and the difference of surveying rotating speed	-11.6	-6.7	18.1	24.5	16.4	9.2	7.9
Percentage error (%)	-0.9	-0.5	1.3	1.7	1.1	0.6	0.5

Table 3 rotating speed switches lower actual measurement rotating speed and the Speed Identification error table of comparisons (original scheme)

Table 4 rotating speed switches lower actual measurement rotating speed and the Speed Identification error table of comparisons (the present invention program)

Can be found out by the contrast of above-mentioned table, after adopting the present invention program, due to get rid of voltage, electric current negative phase-sequence even harmonic component to the interference of MRAS reference model, be no matter given at steady-state speed or switch in dynamic process at rotating speed, all improve the identification precision of Speedless sensor, make Speedless sensor measured value and actual speed value more identical.Therefore, the method is, under dual feedback wind power generation system realizes faults-tolerant control under unbalanced power grid condition, provide more accurate Speedless sensor identification precision, can improve control performance and the stability of whole generating set.

Claims (4)

1. under a unbalanced source voltage condition, improve the method for Speedless sensor identification precision, comprise: the wind-driven generator adopting doubly-fed control Systematical control, the rotary speed data that described doubly-fed control system obtains according to Speedless sensor controls wind-driven generator rotating speed; Described Speedless sensor carries out the process of PI self adaptation speed estimate by MRAS reference model to stator voltage and stator current, obtains rotary speed data; When line voltage is in poised state, directly using stator voltage and stator current as the input of MRAS reference model, after the process of PI self adaptation speed estimate, obtain rotary speed data; It is characterized in that: when unbalanced source voltage, obtain rotary speed data by the following method:

1) extract real-time generator unit stator voltage, the positive sequence component of stator current under dq rotating coordinate system, obtains the positive sequence voltage under dq rotating coordinate system with the forward-order current under dq rotating coordinate system

2) will with as the input of MRAS reference model, after the process of PI self adaptation speed estimate, obtain rotary speed data;

In preceding method, due to abandoned in the input variable of MRAS reference model cause because of unbalanced source voltage voltage, electric current negative sequence component interference, make the rotary speed data of institute's identification only by the impact of positive sequence component, thus make Speedless sensor also can provide Speed Identification data accurately for doubly-fed control system when unbalanced source voltage.

2. improve the method for Speedless sensor identification precision under unbalanced source voltage condition according to claim 1, it is characterized in that: judge whether line voltage is in non-equilibrium state according to following method:

1] positive and negative sequence component of extract real-time generator unit stator voltage under α β rest frame, obtains α axle positive sequence voltage component β axle positive sequence voltage component α axle negative sequence voltage components with β axle negative sequence voltage components

2] positive sequence voltage amplitude is calculated according to following formula with negative sequence voltage amplitude

$u_{\mathrm{sm}}^{+}=\sqrt{{\left(u_{\mathrm{s\&alpha;}}^{+}\right)}^2+{\left(u_{\mathrm{s\&beta;}}^{+}\right)}^2}$

$u_{\mathrm{sm}}^{-}=\sqrt{{\left(u_{\mathrm{s\&alpha;}}^{-}\right)}^2+{\left(u_{\mathrm{s\&beta;}}^{-}\right)}^2}$

3] unbalanced source voltage degree δ is calculated according to following formula:

$\mathrm{\&delta;}=\frac{u_{\mathrm{sm}}^{-}}{u_{\mathrm{sm}}^{+}}\&times;100\%;$

4] when δ >=5%, judge that line voltage is in non-equilibrium state; As δ < 5%, judge that line voltage is in poised state.

3. improve the method for Speedless sensor identification precision under unbalanced source voltage condition according to claim 1, it is characterized in that: adopt and with the following method faults-tolerant control is carried out to wind-driven generator:

Adopt speed sensor and Speedless sensor to obtain generator speed data, the rotary speed data that speed sensor gets is designated as speed A, and the rotary speed data that Speedless sensor gets is designated as speed B simultaneously; During speed sensor fault-free, doubly-fed control system controls wind-driven generator rotating speed according to speed A, when speed sensor breaks down, adopts speed B to control wind-driven generator rotating speed.

4. improve the method for Speedless sensor identification precision under unbalanced source voltage condition according to claim 3, it is characterized in that: judge whether speed sensor exists fault according to following method:

A, be calculated as follows fault residual Δ:

Δ＝|ω _A-ω _B|

Wherein, ω _afor the Current observation value of speed A, ω _bfor the Current observation value of speed B;

B, be calculated as follows failure diagnosis adaptive threshold e:

e＝0.2·ω _B

C, Δ and e to be compared: if Δ >=e, then judge that speed sensor breaks down; If Δ < is e, then judge speed sensor fault-free.