CN104836501A - Method for permanent-magnet synchronous motor parameter on-line identification - Google Patents

Abstract

The invention relates to a method for permanent-magnet synchronous motor (PMSM) parameter on-line identification. The method is based on a speed sensorless vector control system of a virtual dc-qc coordinate system. The method comprises: 1) for a non-salient PMSM, establishing a qc-axis steady-state equation considering motor parameter error influence, the equation being u<qc> is approximately equal to R<s>i<qc>+ [omega]L<d>i<dc>+[omega][phi], wherein R<s> is the resistor of a stator winding, L<d> is an direct axis inductor, the [phi] is a magnetic linkage, [omega] is rotor angular velocity, u<qc> is qc-axis voltage, i<qc> is qc-axis current, i<dc> is dc-axis current; 2) and properly adjusting motor operating states repeatedly, and recording the steady-state value of each variable in the steady-state value, substituting into the equation to obtain a plurality of groups of steady-state relations, and obtaining the identification values of motor parameters through calculation of state variable increment. The method just duly adds short-time small disturbance of i<dc> and [omega] instructions in an original control strategy, having not influence on original control flow and system normal operation. The method is simple and fast in algorithm, and can realize non-salient or salient-pole PMSM parameter on-line identification in low cost.

Description

A kind of method of permanent magnet synchronous motor on-line parameter identification

Technical field

The invention belongs to electric drive and electric and electronic technical field, be specifically related to a kind of method being applied to permanent magnet synchronous motor on-line parameter identification.

Background technology

The advantages such as permanent magnet synchronous motor (PMSM:Permanent Magnet Synchronous Motor) has that specific power is high, efficient energy-saving, control are accurate, obtain in every field and apply more and more widely ^[]].The high performance control method of PMSM mainly contains vector control and direct torque control etc.Wherein, the senseless control system that system cost is low, environmental suitability is strong becomes study hotspot.

The performance of PMSM senseless control system largely depends on the estimation accuracy of rotor-position and rotating speed, and detection method can be divided into two classes.The first kind is by injecting signal specific, estimating Position And Velocity, have insensitive to the parameter of electric machine, be suitable for the advantages such as velocity interval is wide, but desirable hypothesis has been done to motor characteristic, as high-frequency signal injection ^[2]need the saliency utilizing motor, INFORM method ^[3]suppose that back-emf is completely sinusoidal, and calculation of complex, higher to the requirement of control chip.Equations of The Second Kind is the evaluation method based on first-harmonic excitation and back-emf, as virtual coordinate system method ^[4], sliding mode observer method ^[5].These class methods calculate simple, easily realize, but stronger to the dependence of the parameter of electric machine.

PMSM parameter has stator winding resistance R _s, d-axis inductance L _d, quadrature axis inductance L _qand magnetic linkage the different installation forms of permanent magnet, determine the difference of d-axis inductance and quadrature axis inductance.Permanent magnet submerged installation be salient-pole machine, L _d≠ L _q, and surface-pasted be non-salient pole machine, L _d=L _q.Experimental result shows, motor temperature can cause the parameter of electric machine to change, but limited extent; When frequency control runs, R _smainly there is change by a relatively large margin by the impact of skin effect; L _dcomparatively stable, L _qobvious change is there is along with load; mainly stablize the impact with demagnetization characteristic by magnet steel.For ensureing the effect of Equations of The Second Kind method, need the on-line identification parameter of electric machine.Traditional parameter identification method, as model reference adaptive method ^[6], extended Kalman filter method ^[7], least square method ^[8]deng, all suppose that rotor-position and tach signal are error free, there is no the impact that consideration brings identification result in the error of Speedless sensor situation upper/lower positions and tach signal.

Different control method to the demand of parameter of electric machine information and susceptibility different, be also the basis of on-line identification, based on the speed-less sensor vector control system of virtual dc-qc coordinate system structure as shown in Figure 1.State observation module exports according to voltage and current feedback estimates rotor speed and position, needs to use motor model and parameter; Phase-locked control module regulates the rotating speed of dc-qc coordinate system according to the position angular difference of dc-qc coordinate system and d-q coordinate system (rotor synchronous frame); Rotational speed control module according to the bias adjustment voltage vector of rotating speed, and adds to motor stator winding by space voltage vector modulation (SVPWM); Because simple possible, easily low cost realize, obtain applying comparatively widely.In Fig. 1, dotted portion is parameter identification module of the present invention, based on motor model and state variable feedback, and the on-line identification parameter of electric machine.

With the physical location of rotor magnetic pole for benchmark, set up synchronous d-q coordinate system, then the model of PMSM is as follows:

In formula, p is differential operator, u _dand u _qbe respectively d-axis and quadrature-axis voltage, ω is rotor velocity (=p θ), and θ is rotor position angle.

As shown in Figure 2, vector control is with rotor position angle estimated value θ _cfor launching in the dc-qc coordinate system of benchmark, it differs Δ θ=θ with d-q co-ordinate system location angle _c-θ, alpha-beta is rest frame.The transformational relation of state variable is

$\left[\begin{array}{ccc}{u}_{\mathrm{dc}}& \mathrm{or}& i_{\mathrm{dc}}\\ u_{\mathrm{qc}}& \mathrm{or}& i_{\mathrm{qc}}\end{array}\right]=\left[\begin{array}{cc}\cos \mathrm{\&Delta;\&theta;}& \sin \mathrm{\&Delta;\&theta;}\\ -\sin \mathrm{\&Delta;\&theta;}& \cos \mathrm{\&Delta;\&theta;}\end{array}\right]\left[\begin{array}{ccc}{u}_d& \mathrm{or}& i_d\\ u_q& \mathrm{or}& i_q\end{array}\right]---\left(2\right)$

Then motor model becomes:

u _dc＝R _si _dc+pL _di _dc-ω _cL _qi _qc-(L _d-L _q)i _qcpΔθ+E _0xsinΔθ (3)

u _qc＝R _si _qc+pL _di _qc+ω _cL _qi _dc+(L _d-L _q)i _dcpΔθ+E _0xcosΔθ (4)

In formula, ω _cfor the dc-qc coordinate system angular speed that phase-locked control obtains, θ _c=∫ ω _cdt.

During stable state, can ignore differential term, voltage equation and position angle deviation are respectively

u _dc＝R _si _dc-ω _cL _qi _qc+E _0xsinΔθ (6)

u _qc＝R _si _qc+ω _cL _qi _dc+E _0xcosΔθ (7)

$\tan \mathrm{\&Delta;\&theta;}=\frac{u_{\mathrm{dc}}-R_s{i}_{\mathrm{dc}}+{\mathrm{\&omega;}}_c{L}_q{i}_{\mathrm{qc}}}{u_{\mathrm{qc}}-R_s{i}_{\mathrm{qc}}-{\mathrm{\&omega;}}_c{L}_q{i}_{\mathrm{dc}}}---\left(8\right)$

The phase-locked link of Systematical control is by regulating ω _c, make Δ θ asymptotic convergence in 0.If the parameter of electric machine used in formula (8) is accurate, during stable state, dc-qc coordinate system will overlap with d-q coordinate system, and namely rotating speed is all consistent with true value with the estimation result of position; But if the parameter of electric machine has error, then two coordinate systems will have deviation all the time.

If the steady state deviation angle of two coordinate systems is Δ θ ', definition parameter of electric machine error is as follows:

ΔR _s＝R′ _s-R _s,ΔL _d＝L′ _d-L _d,ΔL _q＝L′ _q-L _q

Wherein, R _s, L _dand L _qfor parameter of electric machine true value, R ' _s, L ' _dwith L ' _qfor the set point of the parameter of electric machine in control program.Due to R _s, i _dcand Δ θ ' is all less, formula (8) is done with lower aprons,

Then the impact of parameter of electric machine error on rotor position estimation is as follows:

It can thus be appreciated that, although increase along with frequency, Δ R _sobviously increase because of skin effect, but Δ R _s/ ω reduces on the contrary, and i _dcusually very little, therefore Δ R _slimited to estimation Influence on test result; Δ L _qon impact and the load current i of estimation result _qcrelevant, affect little during underloading, during heavy duty, impact is large; Δ L _destimation result is not affected substantially.

The error of the parameter of electric machine not only affects the dynamic characteristic of control system, also can cause the deviation of vector control during stable state, and motor operational efficiency reduces.Therefore, on-line parameter identification is significant for the performance of guarantee and raising senseless control system.

Pertinent literature:

[1] Wang Xiuhe. magneto [M]. the second edition. Beijing: China Electric Power Publishing House, 2010;

[2]Guo Qingding,Luo Ruifu,Wang Limei.Neural Network Adaptive Observer Based Positionand Velocity Sensorless Control of PMSM.AMC'96,1996：41-46；

[3]Schroedl M.Sensorless Control of AC Machines at Low Speed and Standstill Based onthe"INFORM″Method.IEEE Conference on Industrial Application,1996:270-277；

[4]Kiyoshi Sakamoto,Yoshitaka Iwaji,Tsunehiro Endo.Position and Speed SensorlessControl for PMSM Drive Using Direct Position Error Estimation.IECON’01:The 27 ^thAnnual Conference of the IEEE Industrial Electronics Society,2001:1680-1685；

[5]Song Chi,Longya Xu.Position Sensorless Control of PMSM Based on Novel SlidingMode Observer over Wide Speed Range.IEEE IPEMC 2006(3):1-7；

[6] Chen Zhenfeng, Zhong Yanru, Li Jie. the identification of embedded permagnetic synchronous motor self adaptation on-line parameter. Electric Machines and Control, 2010,14 (4): 9-13;

[7]X.Jiang,Z.Zhang,P.Sun and Z.Zhu.Estimation of Temperature Rise in Stator Windingand Rotor Magnet of PMSM Based on EKF.IEEE ICCEE 2010vol.8:24-27；

[8] Li Ping. the modeling of permagnetic synchronous motor and parameter identification. Computer Simulation, 2011,28 (8): 401-404.

Summary of the invention

The invention provides the method for a kind of permanent magnet synchronous motor (PMSM) on-line parameter identification, the method is realized by the speed-less sensor vector control system of virtual dc-qc coordinate system, effectively can improve the performance of this control system.

The inventive method is the on-line identification technology based on the systematic steady state characterisitic parameter under microvariations input and multiple running status.

Owing to using R ' in above-mentioned formula (8) _swith L ' _q, under phase-locked control action, Δ θ → 0, the steady state voltage equation (3) of dc axle is forced to be locked as

u _dc＝R′ _si _dc-ωL′ _qi _qc(10)

So cannot parameter identification be used for, belongs to typical and owe order problem.

And the steady-state equation of qc axle (4) is by the impact of motor parameter error, become

For non-salient pole PMSM, L _d=L _q, above formula (11) becomes formula (12)

Obviously, only under single running status, the on-line identification to multiple parameter of electric machine cannot be realized.

The inventive method takes following steps:

1), with reference to mentioned above, for non-salient pole PMSM (L _d=L _q), set up the qc axle steady-state equation (12) after considering parameter of electric machine error effect:

Wherein, R _sfor stator winding resistance, L _dfor d-axis inductance, for magnetic linkage, ω is rotor velocity, u _qcfor qc shaft voltage, i _qcfor qc shaft current, i _dcfor dc shaft current;

2), repeatedly motor operating state suitably to be regulated, and the steady-state value of each variable under recording lower state, substitute into qc axle steady-state equation (12) and obtain and organize steady state relation more, and calculate R by state variable increment _s, L _d, identifier

Specific implementation process is as shown in on-line identification flow chart 4:

21), keep rotating speed constant, setting i _dc ^*=0, enforcement controls and records the steady-state value ω of each variable ₀, u _qc0and i _qc0, substitute into formula (12) and obtain formula (13):

22), slightly i is changed _dc ^*, setting i _dc ^*=Δ i _dc≠ 0, implement to control and record corresponding steady-state value u _qc1and i _dc1, substitute into formula (12) and obtain formula (14):

Due to rotating speed and load constant, then i _qc1≈ i _qc0, formula (14) and formula (13) are subtracted each other, and arrangement can obtain L _didentifier

${\hat{L}}_d=\frac{u_{\mathrm{qc}1}-u_{\mathrm{qc}0}}{{\mathrm{\&omega;}}_0{i}_{\mathrm{dc}1}}---\left(15\right)$

23), slightly ω is regulated ^*, setting ω ^*=ω ₀+ Δ ω, and keep i _dc ^*=0, implement to control and record corresponding steady-state value ω ₁, u _qc2and i _qc2, substitute into formula (12) and formula (16) can be obtained:

In view of slightly regulating ω in short-term ^*time, load torque is constant, i _qc2≈ i _qc0, formula (13) × ω ₁with formula (16) × ω ₀subtract each other, the identifier of Rs can be obtained

${\hat{R}}_s=\frac{{\mathrm{\&omega;}}_1{u}_{\mathrm{qc}0}-{\mathrm{\&omega;}}_0{u}_{\mathrm{qc}2}}{({\mathrm{\&omega;}}_1-{\mathrm{\&omega;}}_0)i_{\mathrm{qc}0}}---\left(17\right)$

Formula (16) and formula (13) are subtracted each other, and can obtain identifier

Thus the on-line identification achieved 3 parameters of electric machine.

For salient pole PMSM (L _d≠ L _q), cannot directly identification quadrature axis inductance L in said method principle _q.But can utilize and be provided by electric machine manufacturer or quadrature axis inductance L that off-line identification obtains _qwith d-axis inductance L _d, precalculate its ratio K=L _q/ L _d, then by following formula approximate estimation quadrature axis inductance L _qidentifier

${\hat{L}}_q=K\&CenterDot;{\hat{L}}_d---\left(19\right)$

Wherein, for step 22) the d-axis inductance L that obtains _didentifier.

Above-mentioned steps 22) in, described Δ i _dcvalue be the 3-6% of Rated motor electric current, preferably 5%.

Above-mentioned steps 23) in, the value of described Δ ω is run the 1-3% of angular speed, preferably 2%.

Steady-state value based on low-pass filtering calculates and feedback, ensures stability and the accuracy of on-line identification.The impact of the controlled characteristic of real system, load characteristic and the effect of PWM voltage, can not be in completely desirable stable state, voltage and electric current all have certain pulsation.The change of each state variable steady-state value caused by above-mentioned microvariations input is very little, often lower than its ripple amplitude.Therefore, according to the instantaneous values feedback of voltage and electric current under different running status in on-line identification, the change of steady-state value may be pulsed flood, identification result accurately cannot be obtained.Matlab software platform establishes detailed motor and controller model to the PMSM of a 1.1kW/380V/3000rpm, switching frequency 8kHz, voltage vector control cycle 250 μ s, rotating speed control cycle 5ms.Under action of small disturbance, the simulation waveform of voltage and electric current as shown in Figure 3, and wherein (a) is original waveform, and (b) is the waveform obtained after carrying out root wood filtering process to continuous 1000 sampled datas.As can be seen here, the ripple amplitude of voltage is about 10V, and the ripple amplitude of electric current is greater than 1A.After low-pass filtering treatment, just can obtain the effective information of steady-state value change under action of small disturbance.Work as i _dc ^*when becoming 0.4A from 0, u _qcadd about 1.5V; Work as N _r ^*when becoming 1530rpm from 1500, u _qcadd about 2.9V.

Therefore, the step 21 in embodiment), 22), 23) all low-pass filtering treatment has been carried out to voltage and current signal in control procedure, thus improve stability and the accuracy of on-line identification further.

Permanent magnet synchronous motor on-line identification method of the present invention only in time adds i in original control strategy _dcand the microvariations in short-term of ω instruction, normally run not impact to original control flow and system, and do not increase any state variable, algorithm is simple and direct, can realize the on-line identification of non-salient pole or the inapparent PMSM parameter of electric machine of saliency at low cost.

Accompanying drawing explanation

Fig. 1 is the structure chart of the speed-less sensor vector control system based on virtual dc-qc coordinate system;

Fig. 2 is synchronous coordinate system graph of a relation, and wherein, alpha-beta is rest frame, and d-q is rotor synchronous frame, and dc-qc is hypothetical rotor synchronous coordinate system;

Fig. 3 is the simulation waveform of voltage and electric current under action of small disturbance, (a) original waveform, the waveform after (b) low-pass filtering;

Fig. 4 on-line identification flow chart of the present invention;

Fig. 5 is the actual measurement Parameters variation characteristic of PMSM, (a) temperature characterisitic, (b) frequency characteristic.

Embodiment

The inventive method is applicable to non-salient pole or the inapparent PMSM control system of saliency, can easily realize by wherein adding corresponding software program in variable frequency control.

Embodiment: the System's composition of embodiment is identical with Fig. 1, as previously mentioned, it adopts frequency-variable controller and the eddy-current brake load of applicant's independent development to system configuration.Tested motor nonideal non-salient pole PMSM, tested rating of electric machine parameter is as shown in table 1:

The tested parameter of electric machine of table 1

(DC va method measures R to use LCR electric bridge and off-line parameter identification method _s, direct current attenuation method measures L _dand L _q, idle end voltage method measures ) detect the parameter of electric machine, confirm its variation characteristic.

Fig. 5 (a) is observed temperature characteristic.Along with the rising of motor temperature T, R _s, L _dand L _qincrease, and slightly reduce.When temperature rises to 100 DEG C from 25 DEG C, R _s, L _dand L _qincrease 7.5%, 7.5% and 15% respectively, reduce 8.1%.

Fig. 5 (b) is practical frequency characteristic (using LCR electric bridge).Along with the rising of Injection Current frequency f, because winding conducting wire is comparatively thick, skin effect is obvious, R _senlarge markedly.When injecting 100Hz electric current, R _s129%, L is increased when injecting than direct current _dand L _qreduce 7.7% and 7.3% respectively.

First carry out emulation to the inventive method to confirm.R _s, L _dwith true value be set to 2 Ω, 12mH and 0.46Wb respectively, and the initial set value in control program is respectively 0.5 Ω, 13.3mH and 0Wb.The simulation waveform of parameter identification process as shown in Figure 3, wherein, i _dc ^*: 0 → 0.4A; N _r ^*: 1500 → 1530rpm; A () illustrates original waveform, (b) illustrates the waveform after low-pass filtering treatment.R _s, L _dwith identification result be respectively 1.98 Ω, 12.1mH and 0.459Wb, error is respectively 0.8% ,-4% and-0.2%.

The control chip of frequency-variable controller adopts the SH71253 of Renesas, switching frequency 8kHz, voltage vector control cycle 250 μ s, rotating speed control cycle 5ms.On-line parameter identification subprogram shown in Fig. 4 is added, R in original control program _sand L _qinitial set value be respectively 1.7 Ω and 16.7mH, all have larger deviation with true value.Set different running status, carry out repeatedly identification.In identification process, current i _dcdisturbance input be 0.125A, rotating speed N _rdisturbance input be 24rpm.Table 2 is part of test results.

Table 2 on-line identification experimental result

From table 2 experimental result, along with rotating speed rising, load down, R _sincrease gradually, L _dslightly reduce, and identification result have faint reduction trend because of saliency approximate error, with offline inspection result coincide.

Claims (8)

1. a method for permanent magnet synchronous motor (PMSM) on-line parameter identification, based on the speed-less sensor vector control system of virtual dc-qc coordinate system, comprises the following steps:

1) for non-salient pole PMSM (L _d=L _q), set up the qc axle steady-state equation (12) after considering parameter of electric machine error effect:

Wherein, R _sfor stator winding resistance, L _dfor d-axis inductance, for magnetic linkage, ω are rotor velocity, u _qcfor qc shaft voltage, i _qcfor qc shaft current, i _dcfor dc shaft current;

2) repeatedly motor operating state suitably to be regulated, and the steady-state value of each variable under recording lower state, substitute into qc axle steady-state equation (12) and obtain and organize steady state relation more, and calculate R by state variable increment _s, L _d, identifier

2. the method for claim 1, is characterized in that comprising further: step 3) estimation quadrature axis inductance L _q, for salient pole PMSM (L _d≠ L _q), quadrature axis inductance L _qidentifier pass through estimation, wherein, for step 2) L that obtains _didentifier, K=L _q/ L _d, L _q, L _dfor electric machine manufacturer provide or off-line identification obtain quadrature axis inductance and d-axis inductance.

3. method as claimed in claim 1 or 2, is characterized in that: step 2) specific implementation process as follows:

21) keep rotating speed constant, setting i _dc ^*=0, enforcement controls and records the steady-state value ω of each variable ₀, u _qc0and i _qc0, substitute into formula (12) and obtain formula (13):

22) slightly i is changed _dc ^*, setting i _dc ^*=Δ i _dc≠ 0, implement to control and record corresponding steady-state value u _qc1and i _dc1, substitute into formula (12) and obtain formula (14):

Due to rotating speed and load constant, then i _qc1≈ i _qc0, formula (14) and formula (13) are subtracted each other, and arrangement can obtain L _didentifier

23) slightly ω is regulated ^*, setting ω ^*=ω ₀+ Δ ω, and keep i _dc ^*=0, implement to control and record corresponding steady-state value ω ₁, u _qc2and i _qc2, substitute into formula (12) and obtain formula (16):

In view of slightly regulating ω in short-term ^*time, load torque is constant, i _qc2≈ i _qc0, formula (13) × ω ₁with formula (16) × ω ₀subtract each other, the identifier of Rs can be obtained

Formula (16) and formula (13) are subtracted each other, and can obtain identifier

。

4. method as claimed in claim 3, is characterized in that: in step 21), 22), 23) all low-pass filtering treatment is carried out to voltage and current signal in control procedure.

5. method as claimed in claim 3, is characterized in that: step 22) in, described Δ i _dcvalue be the 3-6% of Rated motor electric current.

6. method as claimed in claim 3, is characterized in that: step 23) in, the value of described Δ ω is run the 1-3% of angular speed.

7. method as claimed in claim 3, is characterized in that: step 22) described in Δ i _dcvalue be the 3-6% of Rated motor electric current; Step 23) described in the value of Δ ω be run the 1-3% of angular speed.

8. method as claimed in claim 7, is characterized in that: step 22) described in Δ i _dcvalue be 5% of Rated motor electric current; Step 23) described in the value of Δ ω be run angular speed 2%.