CN102931906B - Method for asynchronous motor rotor flux linkage observation and rotation speed identification - Google Patents

Abstract

The invention discloses a method for asynchronous motor rotor flux linkage observation and rotation speed identification, which comprises the following steps of: according to stator voltage us and current signals is measured in real time, observing to obtain a rotor flux linkage Psi through a sliding mode observer based on a super-twisting theory and using the rotor flux linkage as a reference value; using a rotor flux linkage Psi~r calculated according to a rotor flux linkage current model as an adjustable value; constructing a model reference self-adaptive system (MRAS) based on the rotor flux linkages; and identifying the asynchronous motor rotation speed omega~r through the self-adaptive law. The rotation speed identification method provided by the invention has high robustness for the variation of stator resistance, and can provide accurate rotor flux linkages and rotation speed values, thereby being especially suitable for an asynchronous motor vector control system.

Description

The method of asynchronous machine rotor flux observation and Speed Identification

Technical field

The present invention discloses a kind of discrimination method of the parameter of electric machine, the particularly very strong rotor flux observer of a kind of robustness, and the Speedless sensor Vector Control System of Induction Motor scheme based on this observer.

Background technology

At present, it is ripe that the control mode of threephase asynchronous machine has become, and vector control and direct torque control can meet most of operating mode demand.Which kind of control mode no matter, rotating speed is all a very important controlled quentity controlled variable.But velocity transducer is installed difficulty in some cases, or sometimes in order to save this part cost, people only wish that, according to the phase current easily recording on frequency converter, phase voltage data, real-time identification rotating speed, realizes Speedless sensor and control.

Speed Sensorless Induction Motor principle can be divided into two types: a generic request rotor has asymmetry, as rotor Harmonic Method, high-frequency signal injection, these class methods need to be analyzed the frequency spectrum of signal, and program wastes time and energy, very harsh to hardware requirement during high speed; Another kind of method is the Mathematical Modeling based on asynchronous machine, with certain mathematical method identification rotating speed wherein, as the method based on state observer, relates to the method for artificial intelligence, and the method based on model reference adaptive principle (MRAS).

Nowadays, researcher has been developed a lot of Speedless sensor algorithms based on asynchronous machine Mathematical Modeling.Belong to having of state observer category: full Order Observers and reduced-order state observer, extended Kalman filter (EKF), and sliding mode observer (SMO).State observer method, to parameter of electric machine sensitive, makes algorithm complicated in order to meet Existence of Global Stable; EKF calculation of complex, a large amount of random parameters will be debugged and obtain; SMO robustness is stronger, but intrinsic shake is harmful to motor low cruise.The method that relates to artificial intelligence is one of study hotspot of the industry always, is just limited to hardware, with a certain distance from practical, also has.The explicit physical meaning of tradition MRAS, algorithm is simpler, and stable state accuracy is relatively good, but is subject to parameter of electric machine variable effect, such as stator resistance R _s.Along with the further investigation to MRAS algorithm, it is found that and select different reference model and adjustable model, can develop and the MRAS identification algorithm of different structure, as based on rotor flux, based on back-emf, based on instantaneous reactive power etc.In addition, apply sliding formwork principle, fuzzy control principle etc. under MRAS structure, can develop and a lot of different structures, the leeway of Improvement is very large.

Judging a kind of quality of Speed Identification algorithm, is mainly to see that can this algorithm keep identification precision in a wider speed adjustable range.The Mathematical Modeling of motor is always accurate not, and some parameters also can and change along with motor operation, thereby greatly affects the accuracy of identification.Under low speed, stator resistance R _svariation to rotor flux ψ _rhaving the greatest impact of observation, and then the identification precision of the Speedless sensor of impact based on flux observer.For R _simpact, have scholar in observer, to add R _sadaptive Identification improve low-speed performance, but this method has relatively high expectations to the stability Design of observer, versatility is not strong.

Summary of the invention

In order to solve the problems of the technologies described above, the present invention proposes the irrelevant rotor flux observation procedure of a kind of and stator resistance, rotor flux that can online observation asynchronous machine, Speed Identification.The method changes insensitive to stator resistance, and has stronger robustness to disturbing.

The method of this asynchronous machine rotor flux observation and Speed Identification, according to the stator voltage u recording in real time _s, stator current i _s, by based on the theoretical sliding mode observer of ultimate attainment distortion (Super-Twisting), observe and obtain rotor flux ψ _rand as reference value, the rotor flux of being tried to achieve by rotor flux current model and as adjustable value, form the model reference adaptive system (Model Reference AdaptiveSystem, MRAS) based on rotor flux, by adaptive law, carry out identification Rotational Speed of Asynchronous Motor

By the theoretical sliding mode observer of described ultimate attainment distortion (Super-Twisting), observation obtains rotor flux ψ _rprocess as follows:

First, threephase asynchronous machine is write to following form in the Mathematical Modeling under rest frame:

$\left\{\begin{array}{c}{\stackrel{\·}{i}}_{\mathrm{s\α}}=-k_1\·i_{\mathrm{s\α}}+k_2\·(b\·{\mathrm{\ψ}}_{\mathrm{r\α}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}})+k_3\·u_{\mathrm{s\α}}\\ {\stackrel{\·}{i}}_{\mathrm{s\β}}=-k_1\·i_{\mathrm{s\β}}+k_2\·(b\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}})+k_3\·u_{\mathrm{s\β}}\\ {\stackrel{\·}{\mathrm{\ψ}}}_{\mathrm{r\α}}=a\·i_{\mathrm{s\α}}-b\·{\mathrm{\ψ}}_{\mathrm{r\α}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}}\\ {\stackrel{\·}{\mathrm{\ψ}}}_{\mathrm{r\β}}=a\·i_{\mathrm{s\β}}-b\·{\mathrm{\ψ}}_{\mathrm{r\β}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}}\end{array}\right.$

In above formula, i _{s α}and i _{s β}for stator current i _scomponent, u _{s α}and u _{s β}for stator voltage u _scomponent, ψ _{r α}and ψ _{r β}for rotor flux ψ _rcomponent;

R _sfor stator resistance, L _mfor the steady-state value of magnetizing inductance, L _sand L _rbe respectively stator, inductor rotor, T _rfor rotor time constant, ω _rfor rotor speed, other scales that occur in above formula are shown:

$\mathrm{\σ}=1-L_m^2/L_s{L}_r$

$k_1=R_s/{\mathrm{\σL}}_s+L_m^2/{\mathrm{\σL}}_s{L}_r{T}_r$

k ₂＝L _m/σL _sL _r

k ₃＝1/σL _s

a＝L _m/T _r

b＝1/T _r

Secondly, the conversion of above-mentioned model being done based on ultimate attainment distortion theory is as follows:

$\left\{\begin{array}{c}{z}_1=i_{\mathrm{s\α}}\\ z_2=i_{\mathrm{s\β}}\\ z_3=b\·{\mathrm{\ψ}}_{\mathrm{r\α}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}}\\ z_4=b\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}}\end{array}\right.$

Z wherein ₁~ z ₄for mathematics intermediate variable;

After conversion, the Mathematical Modeling under threephase asynchronous machine rest frame is written as:

$\left\{\begin{array}{c}{\stackrel{\·}{z}}_1=-k_1\·z_1+k_2\·z_3+k_3\·u_{\mathrm{s\α}}\\ {\stackrel{\·}{z}}_2=-k_1\·z_2+k_2\·z_4+k_3\·u_{\mathrm{s\β}}\end{array}\right.$

So, according to the form of ultimate attainment distortion theory, build rotor flux observer as follows:

$\left\{\begin{array}{c}{\stackrel{\·}{\hat{z}}}_1=-k_1\·z_1+k_2\·{\tilde{z}}_3+k_3\·u_{\mathrm{s\α}}+{\mathrm{\λ}}_1\·{\left|e_1\right|}^{0.5}\·\mathrm{sgn}\left(e_1\right)\\ {\stackrel{\·}{\tilde{z}}}_3={\mathrm{\δ}}_1\·\mathrm{sgn}\left(e_1\right)\\ {\stackrel{\·}{\hat{z}}}_2=-k_1\·z_2+k_2\·{\tilde{z}}_4+k_3\·u_{\mathrm{s\β}}+{\mathrm{\λ}}_2\·{\left|e_2\right|}^{0.5}\·\mathrm{sgn}\left(e_2\right)\\ {\stackrel{\·}{\tilde{z}}}_4={\mathrm{\δ}}_2\·\mathrm{sgn}\left(e_2\right)\end{array}\right.$

Wherein, e ₁and e ₂for error amount, have sgn () is-symbol function, λ ₁, λ ₂and δ ₁, δ ₂for predefined sliding formwork gain;

Finally, the measured value obtaining from above formula with they and rotor flux have following relation:

${\mathrm{\ψ}}_{\mathrm{r\α}}={\∫}_0^t(-{\tilde{z}}_3+\frac{L_m}{T_r}\·{\hat{z}}_1)\·\mathrm{dt}$

${\mathrm{\ψ}}_{\mathrm{r\β}}={\∫}_0^t(-{\tilde{z}}_4+\frac{L_m}{T_r}\·{\hat{z}}_2)\·\mathrm{dt}$

In above formula, L _mfor magnetizing inductance, T _rfor rotor time constant.

The rotor flux being obtained by above-mentioned observer is expressed as [ψ under rest frame _{r α}ψ _{r β}] ^t, as the reference value of model reference adaptive method.

Described rotor flux current model, as adjustable model, carrys out identification Rotational Speed of Asynchronous Motor by adaptive law process as follows:

First, build the observer based on rotor flux current model,

${\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}=\frac{1}{1+T_r\·p}\·(L_m\·i_{\mathrm{s\α}}-{\widehat{\mathrm{\ω}}}_r\·T_r\·{\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}})$

${\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}=\frac{1}{1+T_r\·p}\·(L_m\·i_{\mathrm{s\β}}+{\widehat{\mathrm{\ω}}}_r\·T_r\·{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}})$

In above formula, ${\left[\begin{array}{cc}{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}& {\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\end{array}\right]}^T$ Be the rotor flux value of estimating, p is differential operator, the Rotational Speed of Asynchronous Motor that identification obtains, T _rfor rotor time constant;

Secondly, by the output valve [ψ based on the theoretical sliding mode observer of ultimate attainment distortion (Super-Twisting) _{r α}ψ _{r β}] ^tas reference value, ${\left[\begin{array}{cc}{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}& {\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\end{array}\right]}^T$ As adjustable value, according to model reference adaptive (MRAS) principle, adaptive error value is tried to achieve by following formula:

$\mathrm{\ϵ}={\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\·{\mathrm{\ψ}}_{\mathrm{r\α}}$

Finally, according to the adaptive law of stability principle design speed, with the equivalence of PI controller, so the output valve of PI controller is Identification of Rotational Speed of Asynchronous Motor result

Beneficial effect of the present invention is, has adopted the sliding mode observer based on ultimate attainment distortion theory, and this observer does not need stator resistance R _svalue, can in full speed range, guarantee rotor flux ψ _raccuracy of observation.Using its output valve as with reference to value, and the MRAS Speedless sensor of structure based on rotor flux, can effectively improve the identification precision of motor speed in low speed situation.

Accompanying drawing explanation

Fig. 1 Vector Control System for Asynchronous Machine schematic diagram;

The structural representation of Fig. 2 asynchronous machine rotor flux observation and Speed Identification method;

The structural representation of the rotor flux sliding mode observer of Fig. 3 based on ultimate attainment distortion theory;

Fig. 4 flux observation and Speed Identification result figure-rotating speed contrast;

Fig. 5 flux observation and Speed Identification result figure-magnetic linkage contrast.

Embodiment

Below in conjunction with drawings and Examples, the invention will be further elaborated.

Referring to Fig. 1, strong power part, three-phase alternating-current supply obtains DC bus-bar voltage U through not controlling rectification _dc, service voltage source type inverter, then obtain supplying with the three phase mains of asynchronous machine.

Weak current part, adopt vector control mode, comprise voltage, current sensor, 3 phase/2 phase static coordinate conversion modules, 2 mutually static/2 same pace coordinate transformation modules, Type New Observer module, rotor flux amplitude judge module, speed ring PI controller module, electric current loop PI controller module, 2 same pace/2 phase static coordinate conversion modules, Realization of pulse width modulation based on voltage space vector module.

The present invention relates generally to Type New Observer module, and other modules are the required functional module of Vector Control System of Induction Motor, is general knowledge known in this field.

The workflow of whole system is described below, to introduce the annexation of each module.

1. by transducer, recorded each phase current and the voltage of threephase asynchronous machine, input " 3 phase/2 phase static coordinate conversion module ", obtains stator current i _scomponent i _{s α}and i _{s β}, stator voltage u _scomponent u _{s α}and u _{s β}.

2. utilize stator voltage, current signal, by Type New Observer of the present invention, obtain real-time rotate speed and synchronous speed angle.Type New Observer comprises: (a) the sliding mode observer module based on ultimate attainment distortion theory, (b) rotor flux space bit angle setting computing module, (c) rotor flux current model block, (d) error amount computing module, (e) rotating speed adaptation module.Concrete details as shown in Figure 2.

(a) first, the Mathematical Modeling under threephase asynchronous machine rest frame is write to following form:

$\left\{\begin{array}{c}{\stackrel{\·}{i}}_{\mathrm{s\α}}=-k_1\·i_{\mathrm{s\α}}+k_2\·(b\·{\mathrm{\ψ}}_{\mathrm{r\α}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}})+k_3\·u_{\mathrm{s\α}}\\ {\stackrel{\·}{i}}_{\mathrm{s\β}}=-k_1\·i_{\mathrm{s\β}}+k_2\·(b\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}})+k_3\·u_{\mathrm{s\β}}\\ {\stackrel{\·}{\mathrm{\ψ}}}_{\mathrm{r\α}}=a\·i_{\mathrm{s\α}}-b\·{\mathrm{\ψ}}_{\mathrm{r\α}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}}\\ {\stackrel{\·}{\mathrm{\ψ}}}_{\mathrm{r\β}}=a\·i_{\mathrm{s\β}}-b\·{\mathrm{\ψ}}_{\mathrm{r\β}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}}\end{array}\right.$

In above formula, under rest frame, i _{s α}and i _{s β}for stator current i _scomponent, u _{s α}and u _{s β}for stator voltage u _scomponent, ψ _{r α}and ψ _{r β}for rotor flux ψ _rcomponent;

R _sfor stator resistance, L _mfor the steady-state value of magnetizing inductance, L _sand L _rbe respectively stator, inductor rotor, T _rfor rotor time constant, ω _rfor rotor speed, other scales that occur in above formula are shown:

$\mathrm{\σ}=1-L_m^2/L_s{L}_r$

$k_1=R_s/{\mathrm{\σL}}_s+L_m^2/{\mathrm{\σL}}_s{L}_r{T}_r$

k ₂＝L _m/σL _sL _r

k ₃＝1/σL _s

a＝L _m/T _r

b＝1/T _r

Secondly, to above-mentioned model, do based on the theoretical conversion of ultimate attainment distortion (Super-Twisting) as follows:

$\left\{\begin{array}{c}{z}_1=i_{\mathrm{s\α}}\\ z_2=i_{\mathrm{s\β}}\\ z_3=b\·{\mathrm{\ψ}}_{\mathrm{r\α}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}}\\ z_4=b\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}}\end{array}\right.$

Z wherein ₁~ z ₄for mathematics intermediate variable;

After conversion, the Mathematical Modeling under threephase asynchronous machine rest frame is written as:

$\left\{\begin{array}{c}{\stackrel{\·}{z}}_1=-k_1\·z_1+k_2\·z_3+k_3\·u_{\mathrm{s\α}}\\ {\stackrel{\·}{z}}_2=-k_1\·z_2+k_2\·z_4+k_3\·u_{\mathrm{s\β}}\end{array}\right.$

So the form theoretical according to ultimate attainment distortion (Super-Twisting) builds rotor flux observer:

$\left\{\begin{array}{c}{\stackrel{\·}{\hat{z}}}_1=-k_1\·z_1+k_2\·{\tilde{z}}_3+k_3\·u_{\mathrm{s\α}}+{\mathrm{\λ}}_1\·{\left|e_1\right|}^{0.5}\·\mathrm{sgn}\left(e_1\right)\\ {\stackrel{\·}{\tilde{z}}}_3={\mathrm{\δ}}_1\·\mathrm{sgn}\left(e_1\right)\\ {\stackrel{\·}{\hat{z}}}_2=-k_1\·z_2+k_2\·{\tilde{z}}_4+k_3\·u_{\mathrm{s\β}}+{\mathrm{\λ}}_2\·{\left|e_2\right|}^{0.5}\·\mathrm{sgn}\left(e_2\right)\\ {\stackrel{\·}{\tilde{z}}}_4={\mathrm{\δ}}_2\·\mathrm{sgn}\left(e_2\right)\end{array}\right.$

Wherein, e ₁and e ₂for error amount, have sgn () is-symbol function, λ ₁, λ ₂and δ ₁, δ ₂for predefined sliding formwork gain.

By above formula discretization, obtain following formula statement:

$\left\{\begin{array}{c}{\hat{z}}_1\left(k\right)={\hat{z}}_1(k-1)+T\·[-k_1\·z_1(k-1)+k_2\·{\tilde{z}}_3(k-1)\\ +k_3\·u_{\mathrm{s\α}}(k-1)+{\mathrm{\λ}}_1\·{\left|e_1(k-1)\right|}^{0.5}\·\mathrm{sgn}\left(e_1(k-1)\right)]\\ {\tilde{z}}_3\left(k\right)={\tilde{z}}_3(k-1)+T\·{\mathrm{\δ}}_1\·\mathrm{sgn}\left(e_1(k-1)\right)\\ {\hat{z}}_2\left(k\right)={\hat{z}}_2(k-1)+T\·[-k_1\·z_2(k-1)+k_2\·{\tilde{z}}_4(k-1)\\ +k_3\·u_{\mathrm{s\β}}(k-1)+{\mathrm{\λ}}_2\·{\left|\right|e_2(k-1)\left|\right|}^{0.5}\·\mathrm{sgn}\left(e_2(k-1)\right)]\\ {\tilde{z}}_4\left(k\right)={\tilde{z}}_4(k-1)+T\·{\mathrm{\δ}}_2\·\mathrm{sgn}\left(e_2(k-1)\right)\end{array}\right.$

Wherein T represents system treatment cycle, and k represents certain once-through operation.

Finally, above formula is followed the tracks of stator current component i _{s α}and i _{s β}, output observed result with they and rotor flux have following relation:

${\mathrm{\ψ}}_{\mathrm{r\α}}={\∫}_0^t(-{\tilde{z}}_3+\frac{L_m}{T_r}\·{\hat{z}}_1)\·\mathrm{dt}$

${\mathrm{\ψ}}_{\mathrm{r\β}}={\∫}_0^t(-{\tilde{z}}_4+\frac{L_m}{T_r}\·{\hat{z}}_2)\·\mathrm{dt}$

Wherein, L _mfor magnetizing inductance, T _rfor rotor time constant;

In the implementation procedure of above formula, there are two point values to illustrate.

The first, in above formula, comprise pure integral element.Pure integral element is affected by initial value for integral and drift, has direct current biasing and initial phase problem.For addressing this problem, with low-pass first order filter, replace pure integral element, and suitable amplitude and phase compensation in addition.In addition, according to ultimate attainment distortion (Super-Twisting), although the rotor flux observer that theoretical form builds also contains pure integral element, it has applied sliding formwork simultaneously, so the meeting that affects of direct current biasing is corrected by sliding formwork effect.

The second, in above formula, contain rotor time constant T _r.When motor moves, T _rvalue can be because temperature rise and slip change the kelvin effect unusual fluctuation causing, and change.But, in real system ratio large two orders of magnitude, so T _rvariation little on the observed result impact of rotor flux.

During concrete enforcement, use the low-pass first order filter with amplitude and phase compensation, replace pure integral element, its flow chart as shown in Figure 3.Suppose that current stator frequency is ω _e, the cut-off frequency ω of low-pass first order filter _c, compensating gain COEFFICIENT K should be K=ω _c/ ω _e.During actual use, can keep K is normal value 3, and cut-off frequency ω _cgiven according to motor running condition.When whole system has just brought into operation, by cut-off frequency ω _cbe made as a suitable fixed value (as 30rad/s); Treat following space bit angle setting θ ₁after stable, ω _cby θ ₁the stator angular frequency that converts and obtain through differential _edetermine.

The rotor flux being obtained by above-mentioned observer is expressed as [ψ under rest frame _{r α}ψ _{r β}] ^t, as the reference value of model reference adaptive method.

(b) by the rotor flux value [ψ of above-mentioned observation _{r α}ψ _{r β}] ^t, calculate its space bit angle setting θ ₁.

${\mathrm{\θ}}_1=a\tan \left(\frac{{\mathrm{\ψ}}_{\mathrm{r\α}}}{{\mathrm{\ψ}}_{\mathrm{r\β}}}\right)$

The foundation that this space bit angle setting is rotor flux linkage orientation, i.e. the space bit angle setting of synchronous speed coordinate system.

(c), using rotor flux current model as adjustable model, by adaptive law, come the process of identification Rotational Speed of Asynchronous Motor as follows.

First, build the observer based on rotor flux current model:

${\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}=\frac{1}{1+T_r\·p}\·(L_m\·i_{\mathrm{s\α}}-{\widehat{\mathrm{\ω}}}_r\·T_r\·{\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}})$

${\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}=\frac{1}{1+T_r\·p}\·(L_m\·i_{\mathrm{s\β}}+{\widehat{\mathrm{\ω}}}_r\·T_r\·{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}})$

In above formula, ${\left[\begin{array}{cc}{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}& {\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\end{array}\right]}^T$ Be the rotor flux value of estimating, p is differential operator, the Rotational Speed of Asynchronous Motor that identification obtains, L _mfor magnetizing inductance, T _rfor rotor time constant.

Secondly, by the output valve [ψ based on the theoretical sliding mode observer of ultimate attainment distortion (Super-Twisting) _{r α}ψ _{r β}] ^tas reference value, ${\left[\begin{array}{cc}{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}& {\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\end{array}\right]}^T$ As adjustable value, according to model reference adaptive (MRAS) principle, under rest frame, adaptive error value is tried to achieve by following formula:

$\mathrm{\ϵ}={\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\·{\mathrm{\ψ}}_{\mathrm{r\α}}$

Finally, according to the adaptive law of stability principle design speed, with the equivalence of PI controller, therefore the output valve of " rotating speed adaptation module " PI controller is Identification of Rotational Speed of Asynchronous Motor result

3. by the amount under rest frame, and the rotor flux space bit angle setting θ being obtained by observer ₁, input " 2 mutually static/2 same pace coordinate transformation modules ", obtains the stator current i under synchronous speed coordinate system _scomponent i _sdand i _sq.

4. rotor flux amplitude is controlled: input speed set-point by " rotor flux amplitude judge module ", judge whether motor operates under weak magnetic state.If motor is in permanent magnetic flux running status, output stator electric current d axle component set-point for normal value.If motor weak magnetic field operation, obtains according to look-up table will as set-point, i _sdas value of feedback, by " electric current loop PI controller module ", the set-point of output stator voltage d axle component

5. speed and torque are controlled: outside appointment as rotary speed setting value, by identification result as value of feedback, by " speed ring PI controller module ", output stator electric current q axle component set-point will as set-point, i _sqas value of feedback, by " electric current loop PI controller module ", the set-point of output stator voltage q axle component

6. inverter control signal produces: by above-mentioned stator voltage set-point of trying to achieve with and rotor flux space bit angle setting θ ₁, input " 2 same pace/2 phase static coordinate conversion module ", obtains the component of stator voltage under rest frame with above-mentioned component calculates inverter trigger impulse via Realization of pulse width modulation based on voltage space vector module, is defeated by inverter, can control each phase voltage of supplying with threephase asynchronous machine.

Embodiment 1

Provide a specific embodiment below, the parameter of dispatching from the factory of a threephase asynchronous machine is as following table:

Parameter of electric machine table

Apply the method for asynchronous machine rotor flux observation of the present invention and Speed Identification, controlling motor running condition stops for startup-40r/min-500r/min-1000r/min-500r/min-40r/min-, move altogether approximately 40 seconds, the duration of each operation conditions and result as shown in Figure 4, Figure 5, two different typesettings that figure is same result.

In Fig. 4: A is that rotating speed, the B of algorithm identification of the present invention is that rotating speed, C that photoelectric code disk reads are that poor, the D that two kinds of methods obtain rotating speeds is as the flux observation result (being close to coincidences) with reference to model and adjustable model, and E is the waveform of black surround place amplification in figure.

In Fig. 5: A is that poor, C that Speed Identification is the flux observation of reference model and adjustable model with code-disc rotating speed (be close to and overlap), B are to be as with reference to model flux observation result as adjustable model flux observation result, D, and E is the waveform that in figure, amplify at black surround place.

Above-mentioned Fig. 4, Fig. 5 are exported by control system digital-to-analogue conversion, with the DL750 of YOKOGAWA company oscillograph recording, obtain.

Under 40r/min operation conditions, the rotor flux observer based on ultimate attainment distortion theory, its observed result is more accurate than conventional art, can be as with reference to model, so the present invention's Speed Identification knot is under the low speed comparatively accurate.

Claims (2)

1. a method for asynchronous machine rotor flux observation and Speed Identification, is characterized in that, according to the stator voltage u of the asynchronous machine recording in real time _s, stator current i _s, by the sliding mode observer based on ultimate attainment distortion theory, observation obtains rotor flux ψ _rand as reference value, the rotor flux of being tried to achieve by rotor flux current model as adjustable value, form the model reference adaptive system based on rotor flux, by adaptive law, carry out identification Rotational Speed of Asynchronous Motor

By the described sliding mode observer based on ultimate attainment distortion theory, observation obtains rotor flux ψ _rprocess as follows:

First, threephase asynchronous machine is write to following form in the Mathematical Modeling under rest frame:

$\left\{\begin{array}{c}{\stackrel{\·}{i}}_{\mathrm{s\α}}=-k_1\·i_{\mathrm{s\α}}+k_2\·(b\·{\mathrm{\ψ}}_{\mathrm{r\α}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}})+k_3\·u_{\mathrm{s\α}}\\ {\stackrel{\·}{i}}_{\mathrm{s\β}}=-k_1\·i_{\mathrm{s\β}}+k_2\·(b\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}})+k_3\·u_{\mathrm{s\β}}\\ {\stackrel{\·}{\mathrm{\ψ}}}_{\mathrm{r\α}}=a\·i_{\mathrm{s\α}}-b\·{\mathrm{\ψ}}_{\mathrm{r\α}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}}\\ {\stackrel{\·}{\mathrm{\ψ}}}_{\mathrm{r\β}}=a\·i_{\mathrm{s\β}}-b\·{\mathrm{\ψ}}_{\mathrm{r\β}}+{\mathrm{\ω}}_r\·{\mathrm{\Ψ}}_{\mathrm{r\α}}\end{array}\right.$

In above formula, i _{s α}and i _{s β}for stator current i _scomponent, u _{s α}and u _{s β}for stator voltage u _scomponent, ψ _{r α}and ψ _{r β}for rotor flux ψ _rcomponent;

R _sfor stator resistance, L _mfor the steady-state value of magnetizing inductance, L _sand L _rbe respectively stator, inductor rotor, T _rfor rotor time constant, ω _rfor rotor speed, other scales that occur in above formula are shown:

$\mathrm{\σ}=1-L_m^2/L_s{L}_r$

$k_1=R_s/{\mathrm{\σL}}_s+L_m^2/{\mathrm{\σL}}_s{L}_r{T}_r$

k ₂＝L _m/σL _sL _r

k ₃＝1/σL _s

a＝L _m/T _r

b＝1/T _r

Secondly, the conversion of above-mentioned model being done based on ultimate attainment distortion theory is as follows:

$\left\{\begin{array}{c}{z}_1=i_{\mathrm{s\α}}\\ z_2=i_{\mathrm{s\β}}\\ z_3=b\·{\mathrm{\ψ}}_{\mathrm{r\α}}+{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\β}}\\ z_4=b\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\mathrm{\ω}}_r\·{\mathrm{\ψ}}_{\mathrm{r\α}}\end{array}\right.$

Z wherein ₁～z ₄for mathematics intermediate variable;

After conversion, the Mathematical Modeling under threephase asynchronous machine rest frame is written as:

$\left\{\begin{array}{c}{\stackrel{\·}{z}}_1=-k_1\·z_1+k_2\·z_3+k_3\·u_{\mathrm{s\α}}\\ {\stackrel{\·}{z}}_2=-k_1\·z_2+k_2\·z_4+k_3\·u_{\mathrm{s\β}}\end{array}\right.$

So, according to the form of ultimate attainment distortion theory, build rotor flux observer:

$\left\{\begin{array}{c}{\stackrel{\stackrel{\·}{^}}{z}}_1=-k_1\·z_1+k_2\·{\tilde{z}}_3+k_3\·u_{\mathrm{s\α}}+{\mathrm{\λ}}_1\·{\left|e_1\right|}^{0.5}\·\mathrm{sgn}\left(e_1\right)\\ {\stackrel{\stackrel{\·}{~}}{z}}_3={\mathrm{\δ}}_1\·\mathrm{sgn}\left(e_1\right)\\ {\stackrel{\stackrel{\·}{^}}{z}}_2=-k_1\·z_2+k_2\·{\tilde{z}}_4+k_3\·u_{\mathrm{s\β}}+{\mathrm{\λ}}_2\·{\left|e_2\right|}^{0.5}\·\mathrm{sgn}\left(e_2\right)\\ {\stackrel{\stackrel{\·}{~}}{z}}_4={\mathrm{\δ}}_2\·\mathrm{sgn}\left(e_2\right)\end{array}\right.$

Wherein, e ₁and e ₂for error amount, have sgn () is-symbol function, λ ₁, λ ₂and δ ₁, δ ₂for predefined sliding formwork gain;

Finally, the measured value obtaining from above formula with they and rotor flux have following relation:

${\mathrm{\ψ}}_{\mathrm{r\α}}={\∫}_0^t(-{\tilde{z}}_3+\frac{L_m}{T_r}\·{\hat{z}}_1)\·\mathrm{dt}$

${\mathrm{\ψ}}_{\mathrm{r\β}}={\∫}_0^t(-{\tilde{z}}_4+\frac{L_m}{T_r}\·{\hat{z}}_2)\·\mathrm{dt}$

In above formula, L _mfor magnetizing inductance, T _rfor rotor time constant;

The rotor flux being obtained by above-mentioned observer is expressed as [ψ under rest frame _{r α}ψ _{r β}] ^t, as the reference value of model reference adaptive method.

2. the method for rotor flux observation according to claim 1 and Speed Identification, is characterized in that, described rotor flux current model is as adjustable model, by adaptive law identification Rotational Speed of Asynchronous Motor process as follows:

First, build the observer based on rotor flux current model,

${\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}=\frac{1}{1+T_r\·p}\·(L_m\·i_{\mathrm{s\α}}-{\widehat{\mathrm{\ω}}}_r\·T_r\·{\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}})$

${\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}=\frac{1}{1+T_r\·p}\·(L_m\·i_{\mathrm{s\β}}+{\widehat{\mathrm{\ω}}}_r\·T_r\·{\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}})$

In above formula, be the rotor flux value of observation, p is differential operator, the Rotational Speed of Asynchronous Motor that identification obtains, T _rfor rotor time constant;

Secondly, by the output valve [ψ of the sliding mode observer based on ultimate attainment distortion theory _{r α}ψ _{r β}] ^tas reference value, as adjustable value, according to model reference adaptive method, adaptive error value is tried to achieve by following formula:

$\mathrm{\ϵ}={\widehat{\mathrm{\ψ}}}_{\mathrm{r\α}}\·{\mathrm{\ψ}}_{\mathrm{r\β}}-{\widehat{\mathrm{\ψ}}}_{\mathrm{r\β}}\·{\mathrm{\ψ}}_{\mathrm{r\α}}$

Finally, according to the adaptive law of stability principle design speed, with the equivalence of PI controller, so the output valve of PI controller is Identification of Rotational Speed of Asynchronous Motor result