CN1929290A - Motor control device - Google Patents

Abstract

A motor control device includes an estimator for estimating a rotor position of a motor having a salient pole by using a value corresponding to a q-axis inductance of the motor as an operation parameter where an estimated axes for the control corresponding to d-q axes are gamma-delta axes, and a controller for controlling the motor based on the estimated rotor position. The estimator generates a deviation between a d-axis and a q-axis by performing the estimation of the rotor position based on a value between a real q-axis inductance and a real d-axis inductance of the motor adopted as the operation parameter. The controller controls the motor so that a delta-axis component of a motor current supplied to the motor is maintained to be a predetermined value of zero or close to zero regardless of a value of a gamma-axis component of the motor current.

Description

Control device of electric motor

Technical field

The present invention relates to a kind of control device of electric motor that is used for controlling the work of motor.Also relate to a kind of electric motor drive system with this control device of electric motor.

Background technology

Developed in the past and do not used rotor-position sensor, but inferred the rotor-position of motor, according to this rotor-position of inferring, the control device of electric motor (position-sensorless control device) of control motor.One of the block diagram example of this control device of electric motor 103 has been shown among Figure 21.In the formation shown in Figure 21, the axle of inferring in the control of the corresponding d axle in the vector control of motor is made as the γ axle, the axle of inferring in the control of corresponding q axle is made as the δ axle.The relation of d axle, q axle, γ axle and δ axle has been shown among Figure 23.E among Figure 23 _Ex, generally be to be called expansion induced voltage (Expansion Zhang Lure to play Electricity and press) voltage vector.

Current detector 11 detects from the U phase current i of the motor current of the motor 1 of PWM inverter (inverter) 2 supply salient pole machines _uAnd V phase current i _v Coordinate converter 12 is with U phase current i _uAnd V phase current i _vBe transformed into γ shaft current i _γAnd δ shaft current i _δLocation/velocity estimator 120 (below be called " estimator 120 ") is inferred out and is inferred rotor position _eAnd infer motor speed omega _eAnd output.

Subtracter 19 is from electromotor velocity command value ω ^*Deduct from what estimator 120 was exported and infer motor speed omega _e, and export this subtraction result.Speed controlling portion 17 is according to the subtraction result (ω of subtracter 19 ^*-ω _e), generate δ shaft current i _δThe δ shaft current command value i that should follow _δ ^*Flux regulator portion 116 is according to δ shaft current command value i _δ ^*Deng, output γ shaft current i _γThe γ shaft current command value i that should follow _γ ^* Current control division 15 is according to the current error (i that is given through subtracter 13 and 14 _γ ^*-i _γ) and current error (i _δ ^*-i _δ) both sides converge to zero mode, output γ shaft voltage command value V _γ ^*With δ shaft voltage command value V _δ ^*

Coordinate converter 18 is inferred rotor position according to what estimator 120 exported _e, carry out γ shaft voltage command value V _γ ^*With δ shaft voltage command value V _δ ^*Inverse transformation, generate by U phase voltage command value Vu ^*, V phase voltage command value Vv ^*, and W phase voltage command value Vw ^*The three-phase voltage command value that is constituted, and output it to PWM inverter 2.PWM inverter 2 is according to this three-phase voltage command value (Vu ^*, Vv ^*, Vw ^*) signal that the production burst width modulated is crossed, will supply to motor 1 corresponding to the motor current of this voltage in three phases command value, drive motor 1.

The internal structure of estimator 120 has been shown among Figure 22.Estimator 120 has axis error and infers portion 130, proportional integral arithmetic unit 131 and integrator 132.Axis error is inferred portion 130 and is inferred axis error Δ θ between d axle and the γ axle.Axis error is inferred portion 130, for example uses following formula (1) to calculate axis error Δ θ.Here L _dAnd L _q, be the d axle inductance and the q axle inductance of each motor 1, Ra is the motor resistance of motor 1.In addition, s is the Laplace's operation symbol.Though proposed to be used for inferring the whole bag of tricks of rotor-position, as following formula (1), inferring with in the calculating formula, the value of the q axle inductance of motor is a lot of with the situation of parameter as computing.

$\mathrm{\Δ\θ}={\tan }^{-1}\left(\frac{-E_{\mathrm{ex\γ}}}{E_{\mathrm{ex\δ}}}\right)={\tan }^{-1}\left(\frac{-({v_{\mathrm{\γ}}}^{*}-(R_a+L_{d}s)i_{\mathrm{\γ}}+{\mathrm{\ω}}_e{L}_q{i}_{\mathrm{\δ}})}{{v_{\mathrm{\δ}}}^{*}-(R_a+L_{d}s)i_{\mathrm{\δ}}-{\mathrm{\ω}}_e{L}_q{i}_{\mathrm{\γ}}}\right)---\left(1\right)$

Above-mentioned formula (1) is the arithmetic expression of the axis error Δ θ shown in No. 3411878 communique of Japan special permission (below be called patent documentation 1).In addition, be to be that the d axle of benchmark and the difference between the γ axle (dc axle) are as Δ θ in the patent documentation 1 with the d axle, but will be that the d axle of benchmark and the difference between the γ axle (dc axle) are handled as Δ θ with the γ axle in this specification, so arithmetic expression and the formula (1) of the axis error Δ θ in the patent documentation 1, opposite in sign.In addition, in the formula (1), E _{Ex γ}With E _{Ex δ}Induced voltage E is expanded in expression respectively _Exγ axle composition and δ axle composition.

Proportional integral arithmetic unit 131, moving in order to realize PLL (Phase Locked Loop) with each position association that constitutes control device of electric motor 103, carry out proportional plus integral control, calculate and infer motor speed omega _e, make axis error infer the axis error Δ θ that portion 130 calculated and converge to zero.Integrator 132 Comparative Examples integrator computing units 131 are exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eProportional integral arithmetic unit 131 is exported infers motor speed omega _eInferred rotor position with integrator 132 exports _e,, export to each position of the control device of electric motor 103 that needs this value together as the output valve of estimator 120.

By constituting control device of electric motor 103 like this, make that the axis error Δ θ between d axle and the γ axle converges to zero, thereby can carry out stable Motor Control.In addition, keep under the zero situation d shaft current i at axis error Δ θ _dFollow the trail of γ shaft current command value i _γ ^*, q shaft current i _qFollow the trail of δ shaft current command value i _δ ^*

But, be used for utilizing the d shaft current i of the breakdown torque control of reluctance torque _dCalculating formula be widely known by the people, in the control device of electric motor 103 that is as above constituted, carry out under the situation of breakdown torque control, flux regulator portion 116 calculates γ shaft current command value i according to following formula (2) _γ ^*Here, Φ _αBe based on the armature magnetic linkage (flux linkage) of permanent magnet.

${i_{\mathrm{\γ}}}^{*}=\frac{{\mathrm{\Φ}}_a}{2(L_q-L_d)}-\sqrt{\frac{{{\mathrm{\Φ}}_a}^2}{4{(L_q-L_d)}^2}+{i_{\mathrm{\δ}}}^{*2}}---\left(2\right)$

In addition, in the TOHKEMY 2003-309992 communique (below be called patent documentation 2), the method for controlling position-less sensor of the phase place of the minimum adjustment motor current of the size that is used for making motor current is disclosed.

In addition, paper " Position and Speed Sensorless forIPMSM Based on Estimation of Position Error " (the T.IEE Japan that gloomy luxuriant hero etc. put down in writing, Vol.122-D, No.7,2002, the 722nd page～729 pages, below be called non-patent literature 1) in, employed computing is disclosed in the inferring of rotor-position with the error of parameter and the relation between the position deduction error (axis error).In addition, in No. 3312472 communique of Japan special permission (below be called patent documentation 3), TOHKEMY 2003-219682 communique (below be called patent documentation 4), TOHKEMY 2002-51597 communique (below be called patent documentation 5) and the TOHKEMY 2003-153582 communique (below be called patent documentation 6), the Motor Control technology of the injection that utilizes high frequency voltage and high-frequency current is disclosed.In addition, in the Japanese kokai publication hei 10-94298 communique, a kind of technology of using the switching of no transducer control about low speed with no transducer control and high speed is disclosed.

In order to use above-mentioned formula (2) to realize breakdown torque control, prerequisite is axis error Δ θ to be maintained zero.In addition, use in the calculating of axis error Δ θ of above-mentioned formula (1), need q axle inductance L _qValue as computing with parameter (motor parameter).Therefore, in order to carry out breakdown torque control, the q axle inductance L of the reality of motor 1 was searched in requirement in the past _qValue, directly use this actual q axle inductance L _qValue, obtain axis error Δ θ and (and even infer rotor position _e).

In addition,, can learn, need in motor, circulate corresponding to q shaft current i from above-mentioned formula (2) in order to carry out high-efficient operation according to the breakdown torque control of using reluctance torque etc. _qD shaft current i _dTherefore, in order to carry out such high-efficient operation, need calculate γ shaft current command value i one by one _γ ^*

In addition, be used for carrying out the γ shaft current command value i of breakdown torque control etc. _γ ^*Arithmetic expression in, have the not clear a plurality of motor parameters of a plurality of true value, if γ shaft current command value i _γ ^*Calculating in employed this motor parameter (calculate use parameter) and really between the motor parameter error is arranged, just can't carry out desired Motor Control.Therefore, need be used for dwindling the adjustment of this error, but for the adjustment of a plurality of motor parameters and be not easy, this is adjusted needs a lot of times as far as possible.

As mentioned above, in the former control device of electric motor, when carrying out breakdown torque control,

The 1st, need be used for axis error Δ θ is maintained the adjustment of (be used for inferring rotor-position) parameter of 0.

The 2nd, γ shaft current command value i _γ ^*Arithmetic expression (2) in employed parameter also need to adjust.

The 3rd, need need the γ shaft current command value i of complicated calculating one by one _γ ^*Calculating.

Rotor position presuming parameter adjustment and γ shaft current command value i _γ ^*Calculate the parameter adjustment of usefulness and carry out respectively, need the corresponding adjustment time.In addition, error and the γ shaft current command value i of rotor position presuming in the parameter adjustment _γ ^*Calculate and interact the feasible difficulty more that becomes of adjusting with the error in the parameter adjustment.In addition, because of the optimization of adjusting difficult caused parameter is difficult to realize, consequently, be difficult to realize the optimal drive of motor.

In addition, above-mentioned patent documentation 1 and 2 and the technology put down in writing of above-mentioned non-patent literature 1 in, can't address the above problem.In addition, in the patent documentation 2, use the calculating of Δ θ ≈ 0, so Δ θ is big more, it is just low more to infer precision.

Summary of the invention

Therefore, the objective of the invention is to, a kind of facilitation that helps to be used for to obtain the computing of breakdown torque control etc. with the adjustment of parameter is provided, and/or the control device of electric motor of the reduction of amount of calculation.In addition, also be to provide a kind of electric motor drive system with such control device of electric motor.

To achieve these goals, relevant the 1st control device of electric motor of the present invention, the axle parallel with the magnetic flux that permanent magnet produced that constitutes rotor is being made as the d axle, to be made as the γ axle corresponding to the axle of inferring in the control of d axle, to be made as under the situation of q axle than the axle of the leading 90 degree electric angles (electrical angle) of d axle, possess: will as the computing parameter, infer the estimator of the rotor-position of above-mentioned motor corresponding to the value of the q axle inductance of the motor that salient pole is arranged; And according to the above-mentioned rotor-position of being inferred, control the control part of above-mentioned motor, above-mentioned estimator, with the value between the q axle inductance of the reality of above-mentioned motor and the actual d axle inductance, be adopted as above-mentioned computing with after the value of parameter, carry out inferring of above-mentioned rotor-position, by between d axle and γ axle, producing deviation like this.

Concrete example as, above-mentioned control part is controlled above-mentioned motor, γ axle composition that make to supply with the motor current of above-mentioned motor remains near the set-point zero or zero.

As mentioned above, during the estimating device position, the computing parameter of corresponding q axle inductance is not used actual q axle inductance, and the value between the d axle inductance of the q axle inductance of use reality and reality by like this, is intended to produce deviation between d axle and the γ axle.Though in order to carry out breakdown torque control etc., need and to supply with motor corresponding to the q shaft current of d shaft current, but owing to produced aforesaid deviation, therefore, also can circulate corresponding to the d shaft current of the value of the q shaft current of reality even the γ axle composition of motor current remains near the set-point zero or zero.

Promptly,, do not need to calculate one by one the value of the γ axle composition of necessary motor current yet, only, just can realize the necessary breakdown torque control of d shaft current etc. by this γ axle being become to be divided near the set-point zero or zero if adopt above-mentioned formation.

In addition, in Figure 21 and the example in the past shown in Figure 22, need rotor position presuming to use the adjustment of parameter with computing, and be used for carrying out the adjustment of the computing of breakdown torque control with parameter, if but adopt above-mentioned formation, just can will be used for carrying out the adjustment of these computings of breakdown torque control etc., and unite with the adjustment of parameter corresponding to the computing of q axle inductance with parameters.Also promptly, can expect to be used for obtaining the facilitation of the computing of breakdown torque control, realize the reduction of adjustment time with the adjustment of parameter.In addition, owing to do not need one by one the value of the γ axle composition of calculating motor electric current, therefore also realized being used for the reduction of the amount of calculation of breakdown torque control etc.

Concrete example as, will than above-mentioned γ axle leading 90 the degree electric angles the axle be made as under the situation of δ axle, above-mentioned estimator is inferred corresponding to above-mentioned rotor-position, also infers the rotating speed of above-mentioned rotor; Above-mentioned control part, has the current-order calculating part, it generates the γ axle composition of above-mentioned motor current and γ shaft current command value and the δ shaft current command value that δ axle composition should be followed the trail of, and makes the above-mentioned rotating speed of being inferred follow the trail of outside electromotor velocity command value of being given; Above-mentioned current-order calculating part, no matter the value of above-mentioned δ shaft current command value how, all remains above-mentioned set-point with above-mentioned γ shaft current command value, by like this, irrelevant with the value of the δ axle composition of above-mentioned motor current, allow the γ axle composition of above-mentioned motor current keep above-mentioned set-point.

In addition, for example the γ axle of above-mentioned motor current is being become to be divided into above-mentioned set-point, and giving under the state of the given load torque of above-mentioned motor effect,, be made as and allow the size of above-mentioned motor current be the value of minimum value the value of above-mentioned computing with parameter.

By like this, can access breakdown torque control or be similar to the control that breakdown torque is controlled.

In addition, for example the γ axle of above-mentioned motor current is being become to be divided into above-mentioned set-point, and giving under the state of the given loading condition of above-mentioned motor effect, with the value of above-mentioned computing with parameter, being made as the loss that allows in the above-mentioned motor is the value of minimum value.

By like this, can access maximal efficiency control or be similar to the control that maximal efficiency is controlled.

In addition, for example, be respectively L at the q of the reality of establishing above-mentioned motor axle inductance and actual d axle inductance _qWith L _d, be under the situation of L as above-mentioned computing with the q axle inductance of parameter, above-mentioned estimator, use and satisfy:

L _d≤L＜(L _d+L _q)/2

L, carry out inferring of above-mentioned rotor-position.

In addition, for example above-mentioned computing can be fixed value with the value of parameter.

By like this, computing is more prone to the adjustment of parameter.

In addition, to achieve these goals, relevant the 2nd control device of electric motor of the present invention, in the control device of electric motor of this control of carrying out motor, it is characterized in that: the corresponding to rotating shaft of direction or the phase place rotating shaft more leading than this rotating shaft of the current phasor when direction is controlled with the realization breakdown torque are made as the qm axle, to be made as with the rotating shaft of this qm axle quadrature under the situation of dm axle, with the motor current that is circulated in the above-mentioned motor, be decomposed into qm axle composition that is parallel to above-mentioned qm axle and the dm axle composition that is parallel to above-mentioned dm axle, carry out the control of above-mentioned motor.

Adopt above-mentioned formation, can expect that also computing is with the facilitation of the adjustment of parameter etc.

Concrete example as, in above-mentioned the 2nd control device of electric motor, have the estimator of the rotor-position of inferring above-mentioned motor, and a control part of controlling above-mentioned motor according to the above-mentioned rotor-position of being inferred; The axle parallel with the magnetic flux that permanent magnet produced that constitutes rotor is being made as the d axle, to be made as the γ axle corresponding to the axle of inferring in the control of d axle, to be made as under the situation of δ axle than the axle of the leading 90 degree electric angles of γ axle, above-mentioned control part carries out the control of above-mentioned motor, make above-mentioned γ axle and above-mentioned δ axle, follow the trail of above-mentioned dm axle and above-mentioned qm axle respectively.

In addition, in for example above-mentioned the 2nd control device of electric motor, above-mentioned control part is controlled above-mentioned motor, makes the γ axle composition of above-mentioned motor current remain near zero or zero set-point.

By like this, owing to do not need one by one the value of the γ axle composition of calculating motor electric current, therefore can cut down the operand that is used for breakdown torque control etc.

In addition, in for example above-mentioned the 2nd control device of electric motor, above-mentioned estimator uses the axis error between above-mentioned qm axle and the above-mentioned δ axle, infers above-mentioned rotor-position.

In addition, in for example above-mentioned the 2nd control device of electric motor, to be made as under the situation of q axle than the axle of the leading 90 degree electric angles of above-mentioned d axle, above-mentioned estimator, the resolution of vectors of using the induced voltage on the q axle that is produced in above-mentioned motor is inferred above-mentioned rotor-position as the induced voltage vector on the qm axle under the situation of induced voltage vector on the qm axle and the induced voltage vector on the dm axle.

Like this, in for example above-mentioned the 2nd control device of electric motor, above-mentioned estimator uses the γ axle composition and the δ axle composition of the induced voltage vector on the above-mentioned qm axle, or uses the γ axle composition of the induced voltage vector on the above-mentioned qm axle, infers above-mentioned rotor-position.

In addition, in for example above-mentioned the 2nd control device of electric motor, above-mentioned estimator, use with the resolution of vectors of the magnetic linkage on the d axle of above-mentioned motor as the flux linkage vector on the dm axle under the situation of the flux linkage vector on flux linkage vector on the qm axle and the dm axle, infer above-mentioned rotor-position.

Like this, in for example above-mentioned the 2nd control device of electric motor, above-mentioned estimator uses the γ axle composition and the δ axle composition of the flux linkage vector on the above-mentioned dm axle, or uses the δ axle composition of the flux linkage vector on the above-mentioned dm axle, infers above-mentioned rotor-position.

In addition, in for example above-mentioned the 2nd control device of electric motor, above-mentioned control part has the above-mentioned rotor-position that uses above-mentioned estimator to infer, and the given fixed axis composition of above-mentioned motor current is transformed into the coordinate converter of γ axle composition and δ axle composition; Above-mentioned estimator according to γ axle composition and the δ axle composition from the resulting above-mentioned motor current of above-mentioned coordinate converter, is inferred the qm axle composition and the dm axle composition of above-mentioned motor current; Qm axle composition and the dm axle composition of use by inferring resulting above-mentioned motor current, and, infer above-mentioned rotor-position from the γ axle composition of the resulting above-mentioned motor current of above-mentioned coordinate converter and the error current between the δ axle composition.

In addition, in for example above-mentioned the 2nd control device of electric motor, also has the stack portion of the superimposed voltage different that superpose with this driving voltage frequency to the driving voltage that is used for driving above-mentioned motor; Above-mentioned estimator can be carried out according to corresponding to the superimposed current that circulates in the above-mentioned motor of being superimposed upon of above-mentioned superimposed voltage, infers the 1st of above-mentioned rotor-position and infers processing.

If superimposed voltage such as overlapped high-frequency rotational voltage are inferred rotor-position according to the superimposed current that circulates because of this stack, particularly when low speed rotation or rotation when stopping, can realizing good no transducer control.

In addition, in for example above-mentioned the 2nd control device of electric motor, above-mentioned estimator can also be carried out according to the drive current corresponding to above-mentioned driving voltage that contains in the above-mentioned motor current, infers the 2nd of above-mentioned rotor-position and infers processing; Corresponding to the velocity information of rotating speed of the above-mentioned rotor of expression, infer the above-mentioned the 1st and to handle and the above-mentioned the 2nd to infer and switch the actual performed processing of inferring in the processing.

By like this, can in big velocity interval, realize good no transducer control.

Concrete example as, above-mentioned estimator, have according to above-mentioned superimposed current, calculate axis error between above-mentioned qm axle and the above-mentioned δ axle as the 1st candidate axis error calculating part of the 1st candidate axis error, with according to above-mentioned drive current, calculate axis error between above-mentioned qm axle and the above-mentioned δ axle as the 2nd candidate axis error calculating part of the 2nd candidate axis error; Employed information in the inferring of above-mentioned rotor-position corresponding to above-mentioned velocity information, is switched between above-mentioned the 1st candidate axis error and above-mentioned the 2nd candidate axis error, infers and handles the switching of inferring processing with the above-mentioned the 2nd by carrying out the above-mentioned the 1st like this.

In addition, concrete example is as, above-mentioned estimator, have according to above-mentioned superimposed current, calculate the 1st candidate speed calculation portion of the rotating speed of above-mentioned rotor, and according to above-mentioned drive current, calculate the 2nd candidate speed calculation portion of the rotating speed of above-mentioned rotor as the 2nd candidate speed as the 1st candidate speed; Employed information in the inferring of above-mentioned rotor-position corresponding to above-mentioned velocity information, is switched between above-mentioned the 1st candidate speed and above-mentioned the 2nd candidate speed, infers and handles the switching of inferring processing with the above-mentioned the 2nd by carrying out the above-mentioned the 1st like this.

In addition, concrete example as, above-mentioned estimator, have according to above-mentioned superimposed current, calculate the 1st candidate position calculating part as the 1st candidate position of the candidate of the above-mentioned rotor-position that should infer, with according to above-mentioned drive current, calculate the 2nd candidate position calculating part as the 2nd candidate position of the candidate of the above-mentioned rotor-position that should infer; Employed information in the inferring of above-mentioned rotor-position corresponding to above-mentioned velocity information, is switched between above-mentioned the 1st candidate position and above-mentioned the 2nd candidate position, infers and handles the switching of inferring processing with the above-mentioned the 2nd by carrying out the above-mentioned the 1st like this.

In addition, for example, above-mentioned estimator, the above-mentioned the 1st infer handle with the above-mentioned the 2nd infer switch between the processing actual performed when inferring processing, corresponding to above-mentioned velocity information, or,, allow the actual performed processing of inferring handle transfer to inferring of the opposing party from a side the processing of inferring by having increased both sides' the processing of inferring of inferring the result of inferring processing corresponding to from switching the elapsed time of beginning.

By like this, can realize that level and smooth inferring handled switches.

In addition, concrete example is as, the voltage vector track on the rotatable coordinate axis of above-mentioned superimposed voltage, forms that to have with the d axle be the symmetric figure of benchmark.

In addition, concrete example is handled when inferring above-mentioned rotor-position inferring by the above-mentioned the 1st as, above-mentioned estimator, uses at least 1 composition in 2 compositions of quadrature of the vector that forms above-mentioned superimposed current, infers above-mentioned rotor-position.

In addition, concrete example as, above-mentioned estimator has the rotation of coordinate portion of the phasor coordinate of above-mentioned superimposed current being rotated the phase difference between above-mentioned dm axle and the above-mentioned d axle, handle when inferring above-mentioned rotor-position inferring by the above-mentioned the 1st, use at least 1 composition that forms by in 2 compositions of quadrature of the resulting current phasor of this rotation of coordinate, infer the axis error between above-mentioned qm axle and the above-mentioned δ axle, use this axis error to infer above-mentioned rotor-position.

In addition, to achieve these goals, associated motor drive system of the present invention is characterized in that having: motor; Drive the inverter of above-mentioned motor; And control aforesaid any control device of electric motor of above-mentioned motor by controlling above-mentioned inverter.

As mentioned above, associated motor control device of the present invention and electric motor drive system can realize being used for obtaining the facilitation of the computing of breakdown torque control etc. with the adjustment of parameter.In addition, can also realize the reduction of operand.

Description of drawings

Fig. 1 is the block diagram of the summary formation of the associated motor drive system of expression the 1st execution mode of the present invention.

Fig. 2 is the analytic modell analytical model figure of the associated motor of the 1st execution mode of the present invention.

Fig. 3 is the block diagram of the electric motor drive system of Fig. 1.

Fig. 4 is the interior block diagram of the location/velocity estimator of Fig. 3.

Fig. 5 for expression γ shaft current be under zero the condition and breakdown torque control corresponding to q shaft current with as calculating the figure that uses the relation between the q axle inductance of parameter.

Fig. 6 is used for the figure of the control in the electric motor drive system of the control of more satisfactory breakdown torque and Fig. 1.

Fig. 7 for expression γ shaft current be under zero the condition as computing with the q axle inductance of parameter and the figure of the relation between the motor current.

Fig. 8 is the polar plot of action of the motor of key diagram 1.

Fig. 9 is the figure of variation of the location/velocity estimator of presentation graphs 3.

Figure 10 is the block diagram of the summary formation of the associated motor drive system of expression the 2nd execution mode of the present invention.

Figure 11 is the analytic modell analytical model figure of the associated motor of the 2nd execution mode of the present invention.

Figure 12 is the analytic modell analytical model figure of the associated motor of the 2nd execution mode of the present invention.

The figure of one of the current locus of the motor current that circulates in the motor of Figure 13 for expression Figure 10 example.

Figure 14 is the block diagram of the electric motor drive system of Figure 10.

Figure 15 is the interior block diagram of the location/velocity estimator of expression Figure 14.

Figure 16 is the dependent curve chart of qm shaft current of relevant each inductance of expression the 2nd execution mode of the present invention.

Figure 17 is used for the figure of the control in the electric motor drive system of the control of more satisfactory breakdown torque and Figure 10.

Figure 18 infers the figure of the inside configuration example of portion for the axis error of expression Figure 15.

Figure 19 is the block diagram of the associated motor drive system of expression the 3rd execution mode of the present invention.

Figure 20 is the block diagram of the associated motor drive system of expression the 4th execution mode of the present invention.

Figure 21 is the block diagram of former control device of electric motor.

Figure 22 is the interior block diagram of the location/velocity estimator of expression Figure 21.

Figure 23 is the polar plot of action that is used for illustrating the motor of Figure 21.

Figure 24 is the block diagram of the associated motor drive system of expression the of the present invention the 5th and the 6th execution mode.

The figure of the voltage vector track of the superimposed voltage that Figure 25 is generated by the superimposed voltage generating unit of Figure 24 for illustration.

The figure of the current phasor track of the superimposed current that Figure 26 circulates in motor because of the stack of superimposed voltage shown in Figure 25 for expression.

The figure of the γ axle composition of the superimposed current that Figure 27 circulates in motor because of the stack of superimposed voltage shown in Figure 25 for expression and the long-pending of δ axle composition and the flip-flop that should amass.

The figure of the γ axle composition of the superimposed current that Figure 28 circulates in motor because of the stack of superimposed voltage shown in Figure 25 for expression and the long-pending of δ axle composition and the flip-flop that should amass.

Figure 29 is can be as the interior block diagram of the estimator of the location/velocity estimator of Figure 24.

Figure 30 infers the interior block diagram of portion for the axis error of Figure 29.

Figure 31 is the interior block diagram of the axis error calculating part of Figure 30.

Figure 32 is the figure (when circular rotational voltage superpose) of expression according to the current phasor track example of the front and back of the rotation of coordinate of the rotation of coordinate portion of Figure 30.

Figure 33 is the figure (when oval rotational voltage superpose) of expression according to the current phasor track example of the front and back of the rotation of coordinate of the rotation of coordinate portion of Figure 30.

Figure 34 is the figure (during alternating voltage stack) of expression according to the current phasor track example of the front and back of the rotation of coordinate of the rotation of coordinate portion of Figure 30.

Figure 35 is the interior block diagram (the 1st estimator example) of the relevant position/speed estimating device of the 6th execution mode of the present invention.

Figure 36 is the figure of function that is used for illustrating the hand-off process portion of Figure 35.

The figure that Figure 37 handles for the weighted average that the hand-off process portion of explanation by Figure 35 carried out.

The figure that Figure 38 handles for the weighted average that the hand-off process portion of explanation by Figure 35 carried out.

Figure 39 is the interior block diagram (the 2nd estimator example) of the relevant position/speed estimating device of the 6th execution mode of the present invention.

Figure 40 is the interior block diagram (the 3rd estimator example) of the relevant position/speed estimating device of the 6th execution mode of the present invention.

Embodiment

" the 1st execution mode "

Below embodiments of the present invention are elaborated.At first, the 1st execution mode of the present invention is described.Fig. 1 is the frame assumption diagram of the associated motor drive system of the 1st execution mode.1 for to be set at rotor (not shown) with permanent magnet, armature coil is set at the three-phase permanent-magnetic synchronous motors 1 (hereinafter to be referred as making " motor 1 ") of stator (not shown).Motor 1 is to imbed the salient pole machine that the magnet type synchronous motor is representative (motor with salient pole).

2 is PWM (Pulse Width modulation) inverter, supplies with the three-phase alternating voltage that U phase, V phase and W phase are constituted for motor 1 corresponding to the rotor-position of motor 1.If supply with the voltage of this motor 1 is motor voltage (armature voltage) Va, and the electric current of supplying with motor 1 from inverter 2 is motor current (armature supply) Ia.

3 is control device of electric motor (position-sensorless control device), and use motor current Ia infers the rotor-position of motor 1 etc., allows motor 1 export to PWM inverter 2 with the signal of desired rotating speed rotation with being used for.The rotating speed that this is desired is as electromotor velocity command value ω ^*Central Processing Unit) etc., never illustrated CPU (central processing unit: supply to control device of electric motor 3.

Fig. 2 is the analytic modell analytical model figure of motor 1.In the following description, armature coil is meant the coil that is arranged in the motor 1.The armature coil fixed axis of U phase, V phase, W phase has been shown among Fig. 2.1a is the permanent magnet of the rotor of formation motor 1.In the rotating coordinate system of the speed rotation identical with the magnetic flux that produced with permanent magnet 1a, the direction of establishing the magnetic flux that permanent magnet 1a produced is the d axle, and the axle of inferring in the control of corresponding d axle is the γ axle.In addition, though not shown, the phase place of spending electric angles from d axle leading 90 is the q axle, obtains inferring the δ axle of axle from the phase place of the leading 90 degree electric angles of γ axle.The rotating coordinate system of corresponding real axis is to select d axle and the q axle coordinate system as reference axis, and its reference axis is called the d-q axle.Rotating coordinate system in the control (inferring rotating coordinate system) is for being chosen as γ axle and δ axle the coordinate system of reference axis, and its reference axis is called γ-δ axle.

The d-q axle is rotated, and its rotating speed is called real motor speed omega.γ-δ axle also rotates, and its rotating speed is called infers motor speed omega _eIn addition, in certain moment rotating d-q axle, the phase place of d axle is a benchmark with the armature coil fixed axis of U phase, represents by θ (actual rotor position θ).Equally, in rotating γ of certain moment-δ axle, the phase place of γ axle is a benchmark with the armature coil fixed axis of U phase, passes through θ _e(infer rotor position _e) represent.Like this, the axis error Δ θ between d axle and the γ axle (the axis error Δ θ of d-q axle and γ-δ axle) is by Δ θ=θ-θ _eRepresent.

In the following explanation, the γ axle composition of motor voltage Va, δ axle composition, d axle composition and q axle composition are respectively by γ shaft voltage V _γ, δ shaft voltage V _δ, d shaft voltage V _dAnd q shaft voltage Vq represents that the γ axle composition of motor current Ia, δ axle composition, d axle composition and q axle composition are respectively by γ shaft current i _γ, δ shaft current i _δ, d shaft current i _dAnd q shaft current iq represents.

In addition, in the following description, Ra is motor resistance (resistance value of the armature coil of motor 1), L _d, L _qBe respectively d axle inductance (the d axle composition of the inductance of the armature coil of motor 1), q axle inductance (the q axle composition of the inductance of the armature coil of motor 1), Φ a is the caused armature magnetic linkage of permanent magnet 1a.In addition, L _d, L _q, determined value when Ra and Φ a are the manufacturing of electric motor drive system, these values are used in the computing of control device of electric motor.In addition, in various shown in afterwards, s is expressed as the Laplace's operation symbol.

Fig. 3 is the block diagram of the electric motor drive system of the inside formation of the control device of electric motor 3 of detailed description Fig. 1.Control device of electric motor 3 has current detector 1, coordinate converter 12, subtracter 13, subtracter 14, current control division 15, flux regulator portion 16, speed controlling portion 17, coordinate converter 18, subtracter 19 and location/velocity estimator 20 (hereinafter to be referred as making " estimator 20 ").Constitute each position of control device of electric motor 3, as required can the free application controls device 3 interior all values that generated.

Current detector 11 for example is made of Hall element etc., detects the U phase current i as the fixed axis composition of the motor current Ia that supplies with motor 1 from PWM inverter 2 _uAnd V phase current i _vThe U phase current i that coordinate converter 12 receives from current detector 11 _uAnd V phase current i _vTesting result, use the rotor position of inferring that estimator 20 supplied with _e, it is transformed into γ shaft current i _γAnd δ shaft current r _δFollowing formula (3) is used in this conversion.

$\left[\begin{array}{c}{i}_{\mathrm{\γ}}\\ i_{\mathrm{\δ}}\end{array}\right]=\sqrt{2}\left[\begin{array}{cc}\sin ({\mathrm{\θ}}_e+\mathrm{\π}/3)& \sin {\mathrm{\θ}}_e\\ \cos ({\mathrm{\θ}}_e+\mathrm{\π}/3)& \cos {\mathrm{\θ}}_e\end{array}\right]\left[\begin{array}{c}{i}_u\\ i_v\end{array}\right]---\left(3\right)$

Estimator 20 is inferred out and is inferred rotor position _eAnd infer motor speed omega _eAnd output.About inferring rotor position _eAnd infer motor speed omega _ePresuming method, will describe in detail in the back.

Subtracter 19 is from electromotor velocity command value ω ^*In deduct the motor speed omega of inferring from estimator 20 _e, export this subtraction result (velocity error).Speed controlling portion 17 is according to the subtraction result (ω of subtracter 19 ^*-ω _e), generate δ shaft current command value i _δ ^*This δ shaft current command value i _δ ^*, expression is as the δ shaft current i of the δ axle composition of motor current Ia _δThe value of the electric current that should follow the trail of.Flux regulator portion 16, output γ shaft current command value i _γ ^*This γ shaft current command value i _γ ^*, expression is as the γ shaft current i of the γ axle composition of motor current Ia _γThe value of the electric current that should follow the trail of.And the relation between the location/velocity estimator 20 will describe in detail in the back, but this γ shaft current command value i _γ ^*Keep " 0 " in the present embodiment.

The γ shaft current command value i that subtracter 13 is exported from flux regulator portion 16 _γ ^*In deduct the γ shaft current i that coordinate converter 12 is exported _γ, calculate current error (i _γ ^*-i _γ).The δ shaft current command value i that subtracter 14 is exported from speed controlling portion 17 _δ ^*In deduct the δ shaft current i that coordinate converter 12 is exported _δ, calculate current error (i _δ ^*-i _δ).

Current control division 15 receives each current error that subtracters 13,14 are calculated, from the γ shaft current i of coordinate converter 12 _γWith δ shaft current i _δ, and from the motor speed omega of inferring of estimator 20 _e, output γ shaft voltage command value v _γ ^*With δ shaft current command value v _δ ^*, make γ shaft current i _γFollow the trail of γ shaft current command value i _γ ^*, and δ shaft current i _δFollow the trail of δ shaft current command value i _δ ^*

Coordinate converter 18 is inferred rotor position according to what estimator 20 exported _e, carry out γ shaft voltage command value v _γ ^*With δ shaft voltage command value v _δ ^*Inverse transformation, generate the U phase voltage command value Vu of U phase constituent, V phase constituent and the W phase constituent of expression motor voltage Va ^*, V phase voltage command value V _v ^*, W phase voltage command value V _w ^*The three-phase voltage command value that is constituted outputs it to PWM inverter 2.In this inverse transformation, use following two formulas that equation constituted (4)

$\left[\begin{array}{c}{v_u}^{*}\\ {v_v}^{*}\end{array}\right]=\sqrt{\frac{2}{3}}\left[\begin{array}{cc}\cos {\mathrm{\θ}}_e& -\sin {\mathrm{\θ}}_e\\ \cos ({\mathrm{\θ}}_e-2\mathrm{\π}/3)& -\sin ({\mathrm{\θ}}_e-2\mathrm{\π}/3)\end{array}\right]\left[\begin{array}{c}{v_{\mathrm{\γ}}}^{*}\\ {v_{\mathrm{\δ}}}^{*}\end{array}\right]---\left(4\right)$

v _w ^*＝-(v _u ^*+v _v ^*)

PWM inverter 2 is according to the three-phase voltage command value (Vu that represents to load to the voltage of motor 1 ^*, Vv ^*, Vw ^*), the signal after the production burst width modulated will be supplied with motor 1, drive motor 1 corresponding to the motor current Ia of this three-phase voltage command value.

The inside that estimator 20 has been shown among Fig. 4 one of constitutes example.The estimator 20 of Fig. 4 has axis error and infers portion 30, proportional integral calculator 31 and integrator 32.

Axis error is inferred portion 30 and is calculated axis error Δ θ '.This axis error Δ θ ' can learn that according to explanation described later θ is different with the axis error Δ.The axis error of Figure 22 is inferred portion 130, use above-mentioned formula (1) to calculate axis error Δ θ, but the axis error of Fig. 4 is inferred portion 30, uses following formula (5) to calculate Δ θ '.

${\mathrm{\Δ\θ}}^{\′}={\tan }^{-1}\left(\frac{-({v_{\mathrm{\γ}}}^{*}-(R_a+L_{d}s)i_{\mathrm{\γ}}+{\mathrm{\ω}}_{e}L{i}_{\mathrm{\δ}})}{{v_{\mathrm{\δ}}}^{*}-(R_a+L_{d}s)i_{\mathrm{\δ}}-{\mathrm{\ω}}_{e}L{i}_{\mathrm{\γ}}}\right)---\left(5\right)$

In the formula (5), with Δ θ and the L in the above-mentioned formula (1) _q, replaced to Δ θ ' and L respectively.Therefore, axis error is inferred portion 30, and the computing of the corresponding q axle inductance of L when inferring rotor-position is handled with parameter, infers axis error Δ θ '.About this computing with the establishing method of parameter L value and with the relation of this establishing method in the meaning of axis error Δ θ ', will describe in detail in the back.

Proportional integral arithmetic unit 31 is used for realizing PLL (Phases Locked Loop), and is moving with each position association that constitutes control device of electric motor 3, carries out proportional plus integral control, calculates and infers motor speed omega _e, make axis error infer the axis error Δ θ ' that portion 30 calculated and converge to zero.Integrator 32 Comparative Examples integrator computing units 31 are exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eProportional integral arithmetic unit 31 is exported infers motor speed omega _eInferred rotor position with integrator 32 exports _e,, send to each position of the control device of electric motor 3 that needs this value all as the output valve of estimator 20.

Use under the situation of true value (actual value) as the L in the formula (5) of q axle inductance in hypothesis, also, at L=L _qSituation under, Δ θ '=Δ θ by based on control such as the PLL of proportional integral arithmetic unit 31 grades, makes axis error Δ θ ' (=Δ θ) converge to zero (also promptly identical with the formation of Figure 21 control).But the characteristic point of present embodiment is that computing is set as with parameter L satisfies following formula (6).Also promptly, the q axle inductance with the reality of motor 1 (also is L _q) with actual d axle inductance (also be L _d) between value, the computing that is adopted as corresponding q axle inductance is carried out the calculating of axis error with on the basis of parameter.In addition, satisfy L certainly _d＜L _q

L _d≤L＜L _q …(6)

In addition, preferably setting computing makes it satisfy following formula (7) with parameter L.

L _d≤L＜(L _d+L _q)/2 …(7)

By the L that will as above set as the computing of corresponding q axle inductance with the resulting axis error Δ of parameter θ ', θ is different certainly with the axis error Δ.Therefore,, make axis error Δ θ ' converge to zero, also can produce deviation (axis error of non-zero) between d axle and the γ axle even carry out PLL control.

In the present embodiment, have a mind to produce this deviation, actively utilize this deviation, make the γ shaft current command value i that flux regulator portion 16 is exported _γ ^*Be zero,, be similar to the control of breakdown torque control by like this.Below this control is studied.

At first, as shown in above-mentioned non-patent literature 1, inferring of rotor-position (also promptly inferred rotor position _eCalculating) in employed calculating with the error of parameter and the relation of position deduction error (axis error), described as shown in the formula (8).Here, Ra ' is employed as the value of computing with the motor resistance of parameter for being used for the arithmetic expression of inferring of rotor-position, and (Ra-Ra ') represents that this computing is with the error between parameter and the genuine motor resistance Ra.L _q' to be used for the arithmetic expression of inferring of rotor-position employed as the value of computing with the q axle inductance of parameter, (L in expression _q-L _q') represent that this computing is with parameter and the genuine error between the q axle inductance.

$\frac{R_a-{R_a}^{\′}}{\mathrm{\ω}}{i}_{\mathrm{\γ}}-(L_q-{L_q}^{\′})i_{\mathrm{\δ}}---\left(8\right)$

$=\sin \mathrm{\Δ\θ}\{{\mathrm{\Φ}}_a+(L_d-L_q)(i_{\mathrm{\γ}}\cos \mathrm{\Δ\θ}+i_{\mathrm{\δ}}\sin \mathrm{\Δ\θ})\}$

Now, establish L _q'=L.Also promptly, when inferring rotor-position, provide energetically and be equivalent to (L _q-L) error.Use above-mentioned formula (5) to infer axis error Δ θ ', be equivalent to actively provide and be equivalent to (L _q-L) error and infer axis error.Suppose in addition, (Ra-Ra ') be zero.In addition, as mentioned above, also consider γ shaft current i _γThe γ shaft current command value i that should follow the trail of _γ ^*Be made as zero situation.Also promptly, in the formula (8), be made as i _γ=0.Like this, formula (8) is out of shape as shown in the formula (9).

$\frac{i_d}{\sqrt{{i_d}^2+{i_q}^2}}\{{\mathrm{\&Phi;}}_a+(L_d-L_q)i_d\}+(L_q-L)\sqrt{{i_{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">d</figure-callout>}}}^2+{i_{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">q</figure-callout>}}}^2}=0---\left(9\right)$

Afterwards, will control corresponding to d shaft current i with breakdown torque _dFormula (10) be updated in the formula (9), find the solution L and just obtain following formula (11).In addition, formula (10) is general known formula, if will satisfy the d shaft current i of formula (10) corresponding to q shaft current iq _dSupply with motor 1, just obtain breakdown torque control.

$i_d=\frac{{\mathrm{\&Phi;}}_a}{2(L_q-L_d)}-\sqrt{\frac{{{\mathrm{\&Phi;}}_a}^2}{4{(L_q-L_d)}^2}+{i_q}^2}---\left(10\right)$

$L=L_q+\frac{-2{(L_q-L_d)}^3{i_{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">q</figure-callout>}}}^2}{{{\mathrm{\&Phi;}}_a}^2-{\mathrm{\&Phi;}}_a\sqrt{{{\mathrm{\&Phi;}}_a}^2+4{(L_d-L_q)}^2{i_{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">q</figure-callout>}}}^2}+4{(L_d-L_q)}^2{i_q}^2}---\left(11\right)$

Can learn that from the deriving method of formula (11) the represented L of formula (11) is illustrated in γ shaft current command value i _γ ^*Be made as under the zero situation, in order to obtain desirable breakdown torque control, axis error is inferred the q axle inductance value as computing usefulness parameter that 30 in portion should adopt.

L shown in the formula (11) is q shaft current i _qFunction.Below, for specializing of illustrating, list Φ a=0.2411[Vs/rad], L _d=0.003[H], L _q=0.008[H] numerical example, describe.I in this case _qWith the relation of L, shown in the curve 60 of Fig. 5.At γ shaft current command value i _γ ^*Be set as under zero the situation, control the value of corresponding to L with breakdown torque, at 1[A]≤iq≤40[A] in, probably be in 0.003[H] to 0.0042[H] scope in.Also promptly, as can be known with γ shaft current command value i _γ ^*Be made as under zero the situation, control the value of corresponding to L, L relatively with breakdown torque _q(be 0.008[H in this case]) L that are present in more _d(be 0.003[H in this case]) side.

Be conceived to this point, in the present embodiment, adopt and satisfy the computing parameter L of above-mentioned formula (6) or (7), and with γ shaft current command value i _γ ^*Be made as zero, by realizing control like this near breakdown torque control.For example, under the above-mentioned numerical example, will be fixed as L=0.0039[H with parameter L with the irrelevant computing of iq] situation under, the d shaft current i that is circulated in the motor 1 _dWith q shaft current i _qBetween relation, dotted line 62 expressions by Fig. 6.Solid line 61 is that the d shaft current i under the situation of desirable breakdown torque control is carried out in expression _dWith q shaft current i _qBetween the curve of relation, but can learn that dotted line 62 and solid line 61 are very similar curves from Fig. 6.

Although i _γ ^*=0, but still circulation is corresponding to q shaft current i _qD shaft current i _d, be because adopt and to satisfy the computing parameter L of above-mentioned formula (6) or (7), as computing parameter, causedly between d axle and γ axle, produce deviation corresponding to q axle inductance.In addition, in the curve 60 of Fig. 5, at i _q=30[A] time, L=0.0039[H], therefore certain solid line 61 and dotted line 62 are at iq=30[A] locate to intersect.

In addition, for specializing of illustrating, and to γ shaft current command value i _γ ^*Be made as zero example and be illustrated, but γ shaft current command value i _γ ^*Value not need strictly be zero, so long as near the value zero just can (also promptly need only i _γ ^*≈ 0 just can).In other words, γ shaft current command value i is being discussed _γ ^*The situation of value under " zero ", should be interpreted as having " essence zero " of the amplitude of certain degree.Even this is because i _γ ^*And be zero imprecisely, but so long as can regard zero degree in fact as, just can access the control of controlling near breakdown torque.

Computing should realize the control of above-mentioned approximate breakdown torque control with the value of parameter L, selects in the scope that satisfies above-mentioned formula (6) or formula (7).Specifically, by with γ shaft current command value i _γ ^*Be made as near the set-point zero or zero, with γ shaft current i _γBe made as this set-point and give given load torque for motor 1.Afterwards, under this state, select to make the value of the minimum computing of the size of motor current Ia in the scope that satisfies above-mentioned formula (6) or formula (7) with parameter L.At i _γ ^*Under the condition of ≈ 0, make that the size of motor current Ia is the value of the L of minimum value, as shown in Figure 7, be present in L _dWith L _qBetween, even Φ is a, L _d, L _qValue adopt various values, such L also satisfies above-mentioned formula (7).

At i _γ ^*Under the condition of ≈ 0, when the size of selecting to make motor current Ia was the value of L of minimum value, in this given load torque, its L became the desirable computing parameter that realizes breakdown torque control.In addition, such computing is just investigated, is set in the design phase with the value of parameter L.

Like this, by setting in advance suitably, do not need to calculate one by one i as the value of employed computing in the inferring of rotor-position with the q axle inductance of parameter _γ ^*, by allowing i _γ ^*≈ 0 just can realize the control near breakdown torque control.Therefore, at first obtained being used for the reduction effect of the amount of calculation of breakdown torque control.In addition, in Figure 21 and the example in the past shown in Figure 22, need rotor position presuming to use the adjustment of parameter with computing, and be used for carrying out the adjustment of the computing of breakdown torque control with parameter, but in the present embodiment, only, just can access control near breakdown torque control by adjusting rotor position presuming with the computing parameter L.By like this, greatly cut down and adjusted the needed time, improved the efficient of time.

In addition, more than to computing with parameter L and q shaft current i _qValue irrelevant, the example that is made as fixed value (in the above-mentioned example, L=0.0039[H]) is illustrated, but computing also can be corresponding to q shaft current i with parameter L _qValue (corresponding to δ shaft current command value i _δ ^*Value) change.For example, shown on the curve 60 of Fig. 5, if computing uses parameter L corresponding to q shaft current i _qValue (corresponding to δ shaft current command value i _δ ^*Value) change, even i then _γ ^*≈ 0, also can access desirable breakdown torque control (in this case, the solid line among Fig. 6 61 is overlapping fully with dotted line 62).In addition, how corresponding to q shaft current i _qValue (corresponding to δ shaft current command value i _δ ^*Value) set the computing parameter L, can study in advance in the design phase.

In addition, obtain breakdown torque control or be similar to breakdown torque control control method as mentioned above, but, can also obtain having utilized other controls of reluctance torque according to the establishing method of computing with parameter L.

For example, by with γ shaft current command value i _γ ^*Be made as near the set-point zero or zero, can be with γ shaft current i _γBe made as this set-point and given loading condition is sent to motor 1.Afterwards, under this state, in the scope that satisfies above-mentioned formula (6) or formula (7), select to make the value of the minimum computing of loss (copper loss and iron loss) the motor 1 with parameter L.At i _γ ^*Allow the loss be the value of the L of minimum value under the condition of ≈ 0,, be present in L with identical under the situation in the breakdown torque control _dWith L _qBetween, even Φ is a, L _d, L _qValue adopt various values, such L also satisfies above-mentioned formula (7).

i _γ ^*Under the condition of ≈ 0, when selecting to make loss to be the value of L of minimum value, under this given loading condition, this L becomes the computing parameter that realizes maximal efficiency control.In addition, such computing is studied and is set in the design phase with the value of parameter L.In addition, above-mentioned " given loading condition " is meant the condition that motor 1 for example is rotated with given rotating speed, or loads the condition of given load torque for motor 1.

In addition, two formulas that equation constituted (12a) and the computing of (12b) carrying out necessity below current control division 15 uses.In addition, speed controlling portion 17 and proportional integral arithmetic unit 31 use following formula (13) and (14) respectively, carry out necessary computing.

${v_{\mathrm{\&gamma;}}}^{*}=(K_{\mathrm{cp}}+\frac{K_{\mathrm{ci}}}{s})({i_{\mathrm{\&gamma;}}}^{*}-i_{\mathrm{\&gamma;}})-{\mathrm{\&omega;}}_e{L}_q{i}_{\mathrm{\&delta;}}---\left(12a\right)$

${v_{\mathrm{\&delta;}}}^{*}=(K_{\mathrm{cp}}+\frac{K_{\mathrm{ci}}}{s})({i_{\mathrm{\&delta;}}}^{*}-i_{\mathrm{\&delta;}})+{\mathrm{\&omega;}}_e(L_d{i}_{\mathrm{\&gamma;}}+{\mathrm{\&Phi;}}_a)---\left(12b\right)$

i _δ ^*＝(K _sp+K _si/s)·(ω ^*-ω _e) …(13)

ω _e＝(K _p+K _i/s)·Δθ’ …(14)

Here, Kcp, Ksp and Kp are proportionality coefficients, and Kci, Ksi and Ki are integral coefficients, predefined value when they are the design of electric motor drive system.

[about estimator]

According to one of the presuming method of the rotor-position of above-mentioned estimator 20 example, can adopt various presuming methods.(also promptly infer rotor position carrying out inferring of rotor-position _eCalculating) time, can adopt presuming method arbitrarily, just can with the presuming method of parameter so long as use corresponding to the computing of the q axle inductance of motor 1.

For example, can use the method for being put down in writing in the above-mentioned non-patent literature 1 to infer rotor-position.In the above-mentioned non-patent literature 1, use following formula (15) to calculate axis error Δ θ.Under the symbol in using present embodiment and the situation of mark, e _γAnd e _δ, represent the γ axle composition and the δ axle composition of the induced voltage that produced because of the rotation of motor 1 and armature magnetic linkage Φ a respectively according to permanent magnet 1a.In addition, s is the Laplace's operation symbol, and g is the gain of interference observer (outer random オ Block ザ one バ (observer)).

$\mathrm{\&Delta;\&theta;}={\tan }^{-1}\left(\frac{-e_{\mathrm{\&gamma;}}}{e_{\mathrm{\&delta;}}}\right)$

$={\tan }^{-1}\frac{-\frac{g}{s+g}(v_{\mathrm{\&gamma;}}+{\mathrm{\&omega;}}_e{L}_q{i}_{\mathrm{\&delta;}}-(L_{d}s+R_a)i_{\mathrm{\&gamma;}})}{\frac{g}{s+g}(v_{\mathrm{\&delta;}}-{\mathrm{\&omega;}}_e{L}_q{i}_{\mathrm{\&gamma;}}-(L_{d}s+R_a)i_{\mathrm{\&delta;}})}---\left(15\right)$

Under the axis error that the method for inferring axis error according to induced voltage shown in the formula (15) is applied to Fig. 4 was inferred situation in the portion 30, axis error was inferred portion 30 and can be used following formula (16) to calculate axis error Δ θ '.In the formula (16), with Δ θ and the L in the above-mentioned formula (15) _q, replaced to Δ θ ' and L respectively.Like this, identical with the formation of Fig. 4, if in order to make this axis error Δ θ ' converge to zero, infer motor speed omega and calculate by proportional integral arithmetic unit 31 _eAnd calculate by integrator 32 and to infer rotor position _e, just between d axle and γ axle, produce deviation.

${\mathrm{\&Delta;\&theta;}}^{\&prime;}={\tan }^{-1}\left(\frac{-e_{\mathrm{\&gamma;}}}{e_{\mathrm{\&delta;}}}\right)$

$={\tan }^{-1}\frac{-\frac{g}{s+g}(v_{\mathrm{\&gamma;}}+{\mathrm{\&omega;}}_{e}L{i}_{\mathrm{\&delta;}}-(L_{d}s+R_a)i_{\mathrm{\&gamma;}})}{\frac{g}{s+g}(v_{\mathrm{\&delta;}}-{\mathrm{\&omega;}}_{e}L{i}_{\mathrm{\&gamma;}}-(L_{d}s+R_a)i_{\mathrm{\&delta;}})}---\left(16\right)$

In addition, can use the spy to open the method described in the 2004-96979 communique etc. in addition, infer rotor-position.

In addition, can also adopt the formation of inferring axis error (rotor-position) according to the magnetic linkage that becomes the root of induced voltage, replace the formation of Fig. 4.This method is illustrated.At first, the expansion induced voltage equation on the real axis is generally represented by following formula (17).E in the formula (17) _Ex, through type (18) is represented, is called the expansion induced voltage.In addition, the p in the following formula is the symbol of differentiating.

$\left[\begin{array}{c}{v}_d\\ v_q\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_d\\ i_q\end{array}\right]+\left[\begin{array}{c}0\\ E_{\mathrm{ex}}\end{array}\right]---\left(17\right)$

E _ex＝ω((L _d-L _q)i _d+Φ _a)-(L _d-L _q)(pi _q) …(18)

With the formula on the real axis (17), coordinate transform just obtains formula (19) to Control Shaft.

$\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}\\ v_{\mathrm{\&delta;}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+E_{\mathrm{ex}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;\&theta;}\\ \cos \mathrm{\&Delta;\&theta;}\end{array}\right]+({\mathrm{\&omega;}}_e-\mathrm{\&omega;})L_d\left[\begin{array}{c}-i_{\mathrm{\&delta;}}\\ i_{\mathrm{\&gamma;}}\end{array}\right]---\left(19\right)$

In addition, ignoring expression expansion induced voltage E _ExThe situation of transition item (the 2nd on the right) of formula (18) under magnetic flux, as shown in the formula being defined as expansion magnetic flux Φ shown in (20) _Ex

Φ _ex＝(L _d-L _q)i _d+Φ _a …(20)

But under the certain state of electromotor velocity or load, the size of motor current and the variation of phase place are very little, therefore as the 2nd on the right of the formula (18) of the differential term of q shaft current, than ω Φ _ExEnough little, can regard zero as.In addition, under the situation of ground drive motor 1 of not lacking of proper care since actual motor speed ω with infer motor speed omega _eBe very approaching value, so the 3rd on the right of formula (19) is also than ω Φ _ExEnough little, can regard zero as.Therefore, if ignore the 3rd on the right of the 2nd on the right of formula (18) and formula (19) and take in, then formula (19) becomes following formula (21).

$\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}\\ v_{\mathrm{\&delta;}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+E_{\mathrm{ex}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;\&theta;}\\ \cos \mathrm{\&Delta;\&theta;}\end{array}\right]$

$\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+{\mathrm{\&omega;\&Phi;}}_{\mathrm{ex}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;\&theta;}\\ \cos \mathrm{\&Delta;\&theta;}\end{array}\right]---\left(21\right)$

The polar plot of relation etc. of the voltage of the various piece in the expression motor 1 has been shown among Fig. 8 here.Motor on-load voltage Va is by expansion induced voltage E _Ex=ω Φ _Ex, the voltage drop vector RaIa among the motor resistance Ra and armature coil inductance in the voltage drop vector V _LAnd represent.Because expansion magnetic flux Φ _ExBe the magnetic flux (L that the magnetic flux Φ a that produces of permanent magnet and d shaft current are produced _d-L _q) i _dAnd, so the direction of vector is consistent with the d axle.Pass through L _qThe vector that Ia is represented is the vector of the magnetic flux that produced of q axle inductance and motor current Ia, symbol 70 expression Φ _ExWith L _qThe resultant flux vector of Ia.

In addition, Φ _δBe expansion magnetic flux Φ _Exδ axle composition.Therefore, Φ _δ=Φ _ExSin Δ θ sets up.In addition, by the 1st in the matrix that launches above-mentioned formula (21) row and put in order, derive following formula (22).

${\mathrm{\&Phi;}}_{\mathrm{ex}}\sin \mathrm{\&Delta;\&theta;}=-(\frac{v_{\mathrm{\&gamma;}}-(L_{d}s+R_a)i_{\mathrm{\&gamma;}}}{\mathrm{\&omega;}}+L_q{i}_{\mathrm{\&delta;}})---\left(22\right)$

Usually, the magnetic flux that permanent magnet produced, enough bigger than the magnetic flux that d shaft current produced, for Φ a＞＞(L _d-L _q) i _dTherefore, can think Φ _ExNecessarily, also be Φ _Ex≈ Φ a.Like this,, can be similar to,, set up following formula (23) then with reference to formula (22) by sin Δ θ ≈ θ if axis error Δ θ is less.

${\mathrm{\&Phi;}}_{\mathrm{\&delta;}}={\mathrm{\&Phi;}}_{\mathrm{ex}}\sin \mathrm{\&Delta;\&theta;}=-(\frac{v_{\mathrm{\&gamma;}}-(L_{d}s+R_a)i_{\mathrm{\&gamma;}}}{\mathrm{\&omega;}}+L_q{i}_{\mathrm{\&delta;}})---\left(23\right)$

Φ _asinΔθ

Φ _aΔθ

Can learn Φ from above-mentioned formula (23) _δBe approximately equal to the δ axle composition (as the δ axle magnetic flux of the magnetic flux composition of the δ axle of the permanent magnet 1a (Fig. 2) that is parallel to motor 1) of armature magnetic linkage Φ a.Also promptly, Φ _δ≈ (certain value) * Δ θ.Therefore, allow this Φ by controlling _δConverge to zero, make axis error Δ θ also converge to zero.Also promptly, can be according to Φ _δInfer rotor-position or electromotor velocity.

Therefore, the estimator 20 among Fig. 3 and Fig. 4 can replace with estimator 20a shown in Figure 9.Estimator 20a infers portion 33, proportional integral arithmetic unit 31a and integrator 32a by δ axle magnetic flux and constitutes.Because as long as axis error Δ θ converges to zero, δ axle magnetic flux is inferred portion 33 just can infer δ axle magnetic flux Φ _δ, therefore the same with said method, should between d axle and γ axle, have a mind to produce deviation, δ axle magnetic flux is inferred portion 33 and is inferred out δ axle magnetic flux Φ according to following formula (24) _δ'.Promptly, as computing parameter, do not use actual L corresponding to q axle inductance yet _q, use the L that satisfies above-mentioned formula (6) or formula (7), calculate δ axle magnetic flux Φ _δ'.

${{\mathrm{\&Phi;}}_{\mathrm{\&delta;}}}^{\&prime;}=-(\frac{{v_{\mathrm{\&gamma;}}}^{*}-(L_{d}s+R_a)i_{\mathrm{\&gamma;}}}{{\mathrm{\&omega;}}_e}+L{i}_{\mathrm{\&delta;}})---\left(24\right)$

Proportional integral arithmetic unit 31a, identical with the proportional integral arithmetic unit 31 of Fig. 4, moving with each position association that constitutes control device of electric motor 3, carry out proportional plus integral control, calculate and infer motor speed omega _e, make δ axle magnetic flux infer the δ axle magnetic flux Φ that portion 33 is calculated _δ' converge to zero.Integrator 32a Comparative Examples integrator computing unit 31a is exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eProportional integral arithmetic unit 31a is exported infers motor speed omega _eInferred rotor position with integrator 32a exports _e,, export to each position of the control device of electric motor 3 that needs this value together as the output valve of estimator 20a.

In addition, can learn, owing to include L from formula (24) _dThe item and i _γMultiply each other, therefore the value of this item is less.Also promptly, when inferring rotor-position, d axle inductance L _dInfluence less (reason is i _γValue compare r _δValue little a lot).Consider this point, in inferring, in the employed formula (24), can use L as L _dValue.In this case, in the control identical with employed control in the non-salient pole machine that does not use reluctance torque (surface magnet formula synchronous motor etc.), because salient pole machine can high-efficient operation, therefore do not need with magnet imbed structure do not distinguish and change control on an equal basis, versatility is higher.The range of such versatility is also set up under the situation of use formula (5) and formula (16) etc.

In addition, use Φ in the formula (23) _Ex≈ Φ a's is approximate, but also can not use this approximate, infers δ axle magnetic flux.In this case, can infer δ axle magnetic flux Φ according to following formula (25) _δ'.In this case, the computing parameter of corresponding q axle inductance is not used actual L _q, use the L that satisfies above-mentioned formula (6) or formula (7).

${{\mathrm{\&Phi;}}_{\mathrm{\&delta;}}}^{\&prime;}=-\frac{{\mathrm{\&Phi;}}_a}{(L_d-L)i_{\mathrm{\&gamma;}}+{\mathrm{\&Phi;}}_a}(\frac{{v_{\mathrm{\&gamma;}}}^{*}-(L_{d}s+R_a)i_{\mathrm{\&gamma;}}}{{\mathrm{\&omega;}}_e}+L{i}_{\mathrm{\&delta;}})---\left(25\right)$

" the 2nd execution mode "

Next, the 2nd execution mode of the present invention is described.In the explanation of above-mentioned the 1st execution mode, present embodiment and other execution modes described later, unless otherwise specified, mark has identical symbol person all identical, and simultaneously, mark has identical mark (θ or ω etc.) person all identical.Therefore, omit sometimes about having marked prosign or mark person's repeat specification.

Figure 10 is the frame assumption diagram of the associated motor drive system of the 2nd execution mode.The associated motor drive system of the 2nd execution mode has motor 1, inverter 2 and control device of electric motor 3a.

Control device of electric motor 3a, use motor current Ia infers the rotor-position of motor 1 etc., allows motor 1 export to PWM inverter 2 with the signal of desired rotating speed rotation with being used for.The rotating speed that this is desired is as electromotor velocity command value ω ^*Central Processing Unit) etc., never illustrated CPU (central processing unit: supply with control device of electric motor 3a.

Figure 11 and Figure 12 are the analytic modell analytical model figure of the motor 1 that uses present embodiment.Among Figure 11, show the armature coil fixed axis of U phase, V phase, W phase.In the present embodiment, d axle, q axle, γ axle and δ axle, actual rotor position θ, infer rotor position _eWith axis error Δ θ, and actual motor speed ω with infer motor speed omega _e, with the same definition of the 1st execution mode (with reference to Fig. 2).

And, when being controlled with the realization breakdown torque, direction should supply with the corresponding to rotating shaft of direction of the current phasor of motor 1, be defined as the qm axle.And, will postpone 90 axles of spending electric angles than qm axle and be defined as the dm axle.The reference axis that dm axle and qm axle are constituted is called the dm-qm axle.

The solid line 61 of Fig. 6 of current locus during from the control of expression realization breakdown torque can be learnt, realizes that the motor current of breakdown torque control has positive q axle composition and the d axle composition of bearing.Therefore, the qm axle is the phase place axle more leading than q axle.Among Figure 11 and Figure 12, the direction that is rotated counterclockwise is the direction of advance of phase place.

The phase place (angle) of the q axle of being seen from the qm axle is expressed as θ _m, the phase place (angle) of the qm axle of being seen from the δ axle is expressed as Δ θ _mIn this case, certainly, the phase place of the d axle of being seen from the dm axle also is expressed as θ _m, the phase place of the dm axle of being seen from the γ axle also is expressed as Δ θ _mθ _mIt is the angle of advancing of the qm axle (dm axle) seen from q axle (d axle).Δ θ _mAxis error (axis error between dm-qm axle and γ-δ axle) between expression qm axle and the δ axle.Axis error Δ θ between d axle and the γ axle is by Δ θ=Δ θ _m+ θ _mRepresent.

As mentioned above, dm axial ratio d axle phase place is leading, at this moment, and θ _mGet negative value.Equally, under the leading situation of γ axial ratio dm axle phase place, Δ θ _mGet negative value.About vector (E shown in Figure 12 _mDeng), will be explained below.

In addition, the dm axle composition of motor current Ia and qm axle composition are respectively by dm shaft current i _DmAnd qm shaft current i _QmExpression.The dm axle composition of motor voltage Va and qm axle composition are respectively by dm shaft voltage V _DmAnd qm shaft voltage V _QmExpression.

In the present embodiment, infer the axis error Δ θ between qm axle (dm axle) and the δ axle (γ axle) _m, and allow and converge to the dm axle as the γ axle of inferring axle and (also promptly, allow axis error Δ θ _mConverge to zero).Like this, by motor current Ia being decomposed into the qm shaft current i that is parallel to the qm axle _QmWith the dm shaft current i that is parallel to the dm axle _Dm, motor 1 is carried out vector control.

In this case too, need be used for inferring axis error Δ θ _m(be used for allowing Δ θ _mConverge to zero) infer adjustment with parameter, but by carrying out this adjustment, finish breakdown torque control simultaneously and realize using parameter adjustment.Also promptly, because axis error is inferred the adjustment that realizes using parameter with the control of parameter adjustment double as breakdown torque, therefore adjust and become very easy.

In addition, can learn that the current locus of the motor current Ia when carrying out breakdown torque control shown in the solid line 82 of Figure 13, is positioned on the qm axle from the definition of qm axle.Therefore, when carrying out breakdown torque control, do not need the γ shaft current command value i of the complexity shown in the above-mentioned formula (2) _γ ^*Calculating, alleviated computing load.At this moment, γ shaft current command value i _γ ^*Set equally with the 1st execution mode.Also promptly, γ shaft current command value i for example _γ ^*With i _δValue irrelevant, get near the set-point zero or zero.

Next, the working voltage equation describes meaning of the present invention and concrete control method.At first, the expansion induced voltage equation on the real axis, through type (26) expression, expansion induced voltage E _ExThrough type (27) expression.Formula (26) is identical with above-mentioned formula (17), and formula (27) is identical with above-mentioned formula (18).In addition, the p in the following formula is the symbol of differentiating.

$\left[\begin{array}{c}{v}_d\\ v_q\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_d\\ i_q\end{array}\right]+\left[\begin{array}{c}0\\ E_{\mathrm{ex}}\end{array}\right]---\left(26\right)$

E _ex＝ω((L _d-L _q)i _d+Φ _a)-(L _d-L _q)(pi _q) …(27)

If with the formula on the real axis (26), coordinate transform becomes as the γ-δ axle of inferring axle in the control, then obtains formula (28), if ignore the 3rd on the right of formula (28) for simplification, then obtains formula (29).

$\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}\\ v_{\mathrm{\&delta;}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+E_{\mathrm{ex}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;\&theta;}\\ \cos \mathrm{\&Delta;\&theta;}\end{array}\right]-{\left(\mathrm{p\&Delta;\&theta;}\right)L}_d\left[\begin{array}{c}-i_{\mathrm{\&delta;}}\\ i_{\mathrm{\&gamma;}}\end{array}\right]---\left(28\right)$

$\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}\\ v_{\mathrm{\&delta;}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+E_{\mathrm{ex}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;\&theta;}\\ \cos \mathrm{\&Delta;\&theta;}\end{array}\right]---\left(29\right)$

Be conceived to the dm-qm axle, rewriting formula (29) just obtains formula (30).

$\left[\begin{array}{c}{v}_{\mathrm{dm}}\\ v_{\mathrm{qm}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_q\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]+E_{\mathrm{ex}}\left[\begin{array}{c}-\sin {\mathrm{\&theta;}}_m\\ \cos {\mathrm{\&theta;}}_m\end{array}\right]---\left(30\right)$

Here, definition (31) is set up.And if consider i _d=i _QmSin θ _m, then formula (32) is set up.

L _q1i _qm＝sinθ _m{Φ _a+(L _d-L _q)i _d} …(31)

L _q1i _qm＝sinθ _m{Φ _a+(L _d-L _q)i _d}＝sinθ _m{Φ _a+(L _d-L _q)i _qmsinθ _m}

…(32)

If use formula (32) is out of shape formula (30), then obtain formula (33).And E _mThrough type (34) expression.L _q1, be to depend on θ _mImaginary inductance.L _Q1Be for the E in the 2nd on the right of formula (30) _ExSin θ _m, as according to imaginary inductance caused voltage drop handle, and for convenience defined.In addition, L _Q1Get negative value.

$\left[\begin{array}{c}{v}_{\mathrm{dm}}\\ v_{\mathrm{qm}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}(L_q+L_{q1})\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]+{\mathrm{<figure-callout\; id="0"\; label="E\; m"\; filenames="A20061012184200632.png,A20061012184200641.png"\; state="\{\{state\}\}">E</figure-callout>}}_m\left[\begin{array}{c}0\\ 1\end{array}\right]---\left(33\right)$

E _m＝(ω((L _d-L _q)i _d+Φ _a)-(L _d-L _q)(pi _q))cosθ _m＝E _excosθ _m …(34)

Here, equation: L _m=L _q+ L _q1 approximate (because the θ that sets up _mDepend on i _qAnd i _Qm, so L _Q1Depend on i _qWith i _QmIn addition, L _qAlso owing to magnetically saturated influence depends on i _qWith i _QmL _Q1I _qDependence and L _qI _qDependence focuses on L _m, also consider i when inferring _qWith i _QmInfluence).So, formula (33) is deformed into following formula (35).In addition, though the back also will illustrate this L _mBe equivalent to the computing parameter L in the 1st execution mode.

$\left[\begin{array}{c}{v}_{\mathrm{dm}}\\ v_{\mathrm{qm}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_q& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]+{\mathrm{<figure-callout\; id="0"\; label="E\; m"\; filenames="A20061012184200632.png,A20061012184200641.png"\; state="\{\{state\}\}">E</figure-callout>}}_m\left[\begin{array}{c}0\\ 1\end{array}\right]---\left(35\right)$

And then, if formula (35) is out of shape, then obtain following formula (36).Here, E _ExmRepresent by following formula (37).

$\left[\begin{array}{c}{v}_{\mathrm{dm}}\\ v_{\mathrm{qm}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_m& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]+\left[\begin{array}{c}0\\ E_m\end{array}\right]+\mathrm{\&omega;}(L_q-L_m)\left[\begin{array}{cc}0& 0\\ 1& 0\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]$

$=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_m& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]+\left[\begin{array}{c}0\\ E_m+\mathrm{\&omega;}(L_q-L_m)i_{\mathrm{dm}}\end{array}\right]$

$=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_m& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{dm}}\\ i_{\mathrm{qm}}\end{array}\right]+\left[\begin{array}{c}0\\ E_{\mathrm{exm}}\end{array}\right]---\left(36\right)$

E _exm＝(ω((L _d-L _q)i _d+Φ _a)-(L _d-L _q)(pi _q))cosθ _m+ω(L _q-L _m)i _dm

＝E _m+ω(L _q-L _m)i _dm

…(37)

If there is axis error Δ θ between γ-δ axle and the dm-qm axle _m, then formula (36) is out of shape as shown in the formula (38).Also promptly, to be deformed into formula (28) the same with formula (26), if with the formula on the dm-qm axle (36) coordinate transform to γ-δ axle, then obtain formula (38).

$\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}\\ v_{\mathrm{\&delta;}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_m& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+E_{\mathrm{exm}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;}{\mathrm{\&theta;}}_m\\ \cos \mathrm{\&Delta;}{\mathrm{\&theta;}}_m\end{array}\right]-{\left(\mathrm{p\&Delta;}{\mathrm{\&theta;}}_m\right)L}_d\left[\begin{array}{c}-i_{\mathrm{\&delta;}}\\ i_{\mathrm{\&gamma;}}\end{array}\right]---\left(38\right)$

In addition, if be approximately p Δ θ _m≈ 0, i _Dm≈ 0, (L _d-L _q) (pi _q) ≈ 0, the then represented E of through type (37) _Exm, be approximately as shown in the formula shown in (39).

E _exm＝(ω((L _d-L _q)i _d+Φ _a)-(L _d-L _q)(pi _q))cosθ _m+ω(L _q-L _m)i _dm

≈(ω((L _d-L _q)i _δsinθ _m+Φ _a)-(L _d-L _q)(pi _q))cosθ _m+ω(L _q-L _m)i _dm

≈ω((L _d-L _q)i _δsinθ _m+Φ _a)cosθ _m

…(39)

In addition, to θ _mFind the solution, with " L _m=L _q+ L _q1 " be updated to resulting formula in the above-mentioned formula (32), and supposition i _δ≈ i _Qm, then obtain following formula (40).Shown in (40), because θ _mBe i _δFunction, E then _ExmAlso be i _δFunction.

${\mathrm{\&theta;}}_m={\sin }^{-1}\left(\frac{{\mathrm{\&Phi;}}_a-\sqrt{{{\mathrm{\&Phi;}}_a}^2+4(L_q-L_m)(L_q-L_d){i_{\mathrm{\&delta;}}}^2}}{2i_{\mathrm{\&delta;}}(L_q-L_d)}\right)---\left(40\right)$

Contrast Figure 12 is to E _ExWith E _mAnd E _ExmBetween relation be illustrated.With E _Ex, E _mAnd E _ExmBe thought of as the voltage vector in the rotating coordinate system.In this case, E _ExCan be called expansion induced voltage vector.Expansion induced voltage vector E _ExIt is the induced voltage vector on the q axle.Consideration will be expanded induced voltage vector E _Ex, be decomposed into induced voltage vector on the qm axle and the induced voltage vector on the dm axle.Can learn from above-mentioned formula (34), be E by this induced voltage vector that decomposes on the resulting qm axle _mDecompose induced voltage vector (E on the represented dm axle of resulting symbol 80 by Figure 12 by this in addition _ExSin θ _m), be based on imaginary inductance L _q1 voltage drop vector.

Comparison expression (34) can be learnt E with formula (37) _ExmBe to E _mAdd ω (L _q-L _m) i _DmTherefore, in the rotating coordinate system, E _ExmAlso and E _mThe same, become the induced voltage vector on the qm axle.When carrying out breakdown torque control, as mentioned above, because i _Dm≈ 0, so E _ExmWith E _m(substantially) unanimity.

Next, with reference to Figure 12, to corresponding E _Ex, E _mAnd E _ExmMagnetic flux describe.E _ExBe the magnetic linkage Φ of motor 1 _ExThe induced voltage that rotation produced (with reference to above-mentioned formula (20)) with motor 1.In other words, Φ _ExPass through E _ExCalculate divided by ω and (but to ignore the represented E of formula (27) _ExTransition item (the 2nd on the right)).

If think Φ _ExBe the flux linkage vector in the rotating coordinate system, flux linkage vector Φ then _ExIt is the flux linkage vector on the d axle.Consideration is with flux linkage vector Φ _Ex, be decomposed into flux linkage vector on the qm axle and the flux linkage vector on the dm axle.If will be defined as Φ by this flux linkage vector that decomposes on the resulting dm axle _m, Φ then _m=E _m/ ω.In addition, decompose flux linkage vector (Φ on the represented qm axle of resulting symbol 81 by Figure 12 by this _ExSin θ _m), be based on imaginary inductance L _Q1Magnetic flux vector.

If " Φ _Exm=E _Exm/ ω ", Φ then _ExmBecome Φ _mAdd (L _q-L _m) i _DmObtain.Therefore, in the rotating coordinate system, Φ _ExmAlso and Φ _mThe same, become the flux linkage vector on the dm axle.When carrying out breakdown torque control, as mentioned above, because i _Dm≈ 0, so Φ _ExmWith Φ _m(substantially) unanimity.

Next, the example of the concrete electric motor drive system that utilizes above-mentioned the separate equations is shown.Figure 14 is the block diagram of in-built electric motor drive system that shows in detail the control device of electric motor 3a of Figure 10.Control device of electric motor 3a has current detector 11, coordinate converter 12, subtracter 13, subtracter 14, current control division 15, flux regulator portion 16, speed controlling portion 17, coordinate converter 18, subtracter 19 and location/velocity estimator 40 (hereinafter to be referred as making " estimator 40 ").Also promptly, the control device of electric motor 3a of Figure 14 has replaced to estimator 40 with the estimator in the control device of electric motor of Fig. 3 20.Constitute each position of control device of electric motor 3a, as required can the interior all values that is generated of free application of electric motors control device 3a.

Current detector 11 detects the U phase current i as the fixed axis composition of motor current Ia _uAnd V phase current i _vThe U phase current i that coordinate converter 12 receives from current detector 11 _uAnd V phase current i _vTesting result, use the rotor position of inferring that estimator 40 supplied with _e, it is transformed into γ shaft current i _γAnd δ shaft current i _δThis conversion is the same with the 1st execution mode, uses above-mentioned formula (3).

Estimator 40 is inferred out and is inferred rotor position _eAnd infer motor speed omega _eAnd output.Concrete presuming method about estimator 40 will be explained below.

Subtracter 19 is from electromotor velocity command value ω ^*In deduct the motor speed omega of inferring from estimator 40 _e, export this subtraction result (velocity error).Speed controlling portion 17 is according to the subtraction result (ω of subtracter 19 ^*-ω _e), generate δ shaft current command value i _δ ^* Flux regulator portion 16, output γ shaft current command value i _γ ^*This γ shaft current command value i _γ ^*, as mentioned above, set the same with the 1st execution mode.For example, i _γ ^*Get near the set-point zero or zero.

The γ shaft current command value i that subtracter 13 is exported from flux regulator portion 16 _γ ^*In deduct the γ shaft current i that coordinate converter 12 is exported _γ, calculate current error (i _γ ^*-i _γ).The δ shaft current command value i that subtracter 14 is exported from speed controlling portion 17 _δ ^*In deduct the δ shaft current i that coordinate converter 12 is exported _δ, calculate current error (i _δ ^*-i _δ).

Current control division 15 receives each current error that subtracters 13,14 are calculated, from the γ shaft current i of coordinate converter 12 _γWith δ shaft current i _δ, and from the motor speed omega of inferring of estimator 40 _e, output γ shaft voltage command value v _γ ^*With δ shaft voltage command value v _δ ^*, make γ shaft current i _γFollow the trail of γ shaft current command value i _γ ^*, and δ shaft current i _δFollow the trail of δ shaft current command value i _δ ^*

Coordinate converter 18 is inferred rotor position according to what estimator 40 exported _e, carry out γ shaft voltage command value v _γ ^*With δ shaft voltage command value v _δ ^*Inverse transformation, generate Vu ^*, V _v ^*, V _w ^*The three-phase voltage command value that is constituted outputs it to PWM inverter 2.In this inverse transformation, the same with the 1st execution mode, use above-mentioned formula (4).PWM inverter 2 will be supplied with motor 1, drive motor 1 corresponding to the motor current Ia of this three-phase voltage command value.

The inside that estimator 40 has been shown among Figure 15 one of constitutes example.The estimator 40 of Figure 15 has axis error and infers portion 41, proportional integral calculator 42 and integrator 43.Proportional integral arithmetic unit 42 and integrator 43, proportional integral arithmetic unit 31 and the integrator 32 with Fig. 4 is identical respectively.

Axis error is inferred portion 41 and is used V _γ ^*, V _δ ^*, i _γAnd i _δAll or part of of value, calculate axis error Δ θ _mProportional integral arithmetic unit 42 is used for realizing PLL (Phase Locked Loop), and is moving with each position association that constitutes control device of electric motor 3a, carries out proportional plus integral control, calculates and infers motor speed omega _e, make axis error infer the axis error Δ θ that portion 41 is calculated _mConverge to zero.Integrator 43 Comparative Examples integrator computing units 42 are exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eProportional integral arithmetic unit 42 is exported infers motor speed omega _eInferred rotor position with integrator 43 exports _e,, send to each position that needs the control device of electric motor of this value 3a all as the output valve of estimator 40.

Axis error is inferred the axis error Δ θ of portion 41 _mComputational methods, can adopt various computational methods.Below, infer the axis error Δ θ of portion 41 as axis error _mThe computational methods (θ of estimator 40 in other words, _eComputational methods), illustration the 1st, the 2nd, the 3rd, the 4th and the 5th computational methods.

In addition, infer portion 41 at axis error and utilize under the situation of each formula described in this specification the V in the separate equations _γ, V _δAnd the value of ω, use V respectively _γ ^*, V _δ ^*And ω _eValue.In addition, the content (L that has illustrated in each computational methods _mThe determining method etc. of value), other computational methods and other execution modes described later in all can be suitable for.

[the 1st computational methods]

At first, to axis error Δ θ _mThe 1st computational methods describe.In the 1st computational methods, with the induced voltage E that is produced in the motor 1 _ExBe decomposed into induced voltage vector on the qm axle and the induced voltage vector on the dm axle.Afterwards, use is as the induced voltage vector E of the induced voltage vector on the qm axle _Exm(≈ E _m, with reference to Figure 12), calculate axis error Δ θ _m,, calculate phase place (θ as the γ axle of inferring axle in the control by like this _e) (also promptly inferring rotor-position).

As establish induced voltage vector E _Exmγ axle composition and δ axle composition, be respectively E _{Exm γ}With E _{Exm δ}, then can learn Δ θ from Figure 12 _m=tan ^-1(E _{Exm γ}/ E _{Exm δ}) set up.Like this, if use the 1st of above-mentioned determinant (38) is gone and the 2nd result who goes after being out of shape, then Δ θ _mBe expressed as following formula (41) (ignore determinant (38) the right the 3rd).In addition, in the formula (41), suppose final Δ θ _mLess, use tan ^-1(E _{Exm γ}/ E _{Exm δ}) ≈ (E _{Exm γ}/ E _{Exm δ}) approximate.

$\mathrm{\&Delta;}{\mathrm{\&theta;}}_m={\tan }^{-1}\frac{-E_{\mathrm{exm\&gamma;}}}{E_{\mathrm{exm\&delta;}}}={\tan }^{-1}\frac{-(v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}+\mathrm{\&omega;}{L}_m{i}_{\mathrm{\&delta;}})}{v_{\mathrm{\&delta;}}-(R_a+p{L}_d)i_{\mathrm{\&delta;}}-\mathrm{\&omega;}{L}_m{i}_{\mathrm{\&gamma;}}}$

$\&ap;-\frac{v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}+\mathrm{\&omega;}{L}_m{i}_{\mathrm{\&delta;}}}{v_{\mathrm{\&delta;}}-(R_a+p{L}_d)i_{\mathrm{\&delta;}}-\mathrm{\&omega;}{L}_m{i}_{\mathrm{\&gamma;}}}---\left(41\right)$

Axis error is inferred portion 41, calculates Δ θ in use formula (41) _mThe time, can ignore differential term pL _di _γAnd pL _di _δIn addition, Δ θ _mThe needed L of calculating _mThe calculating of value in, utilize following formula (42).To L _Q1Find the solution " i _Dm=0 with following formula (43) and (44) " be updated to resulting formula in the above-mentioned formula (32), and utilize its result, can access formula (42).

$L_m=L_q+{\mathrm{<figure-callout\; id="1"\; label="L\; q"\; filenames="A20061012184200631.png,A20061012184200632.png"\; state="\{\{state\}\}">L</figure-callout>}}_q$

$=L_q+\frac{i_d\{{\mathrm{\&Phi;}}_a+(L_d-L_q)i_d\}}{{{\mathrm{<figure-callout\; id="2"\; label="i\; d"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">i</figure-callout>}}_d}^2+{i_q}^2}---\left(42\right)$

$i_{\mathrm{qm}}=\sqrt{{i_{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">d</figure-callout>}}}^2+{i_q}^2}---\left(43\right)$

$\sin {\mathrm{\&theta;}}_m=\frac{i_d}{\sqrt{{i_d}^2+{i_q}^2}}---\left(44\right)$

And then, utilize with breakdown torque and control corresponding to d shaft current i _dFormula (45) with as i _dAnd i _qAnd i _QmThe formula (43) of relational expression (approximate expression), above-mentioned formula (42) is out of shape, then L _mBecome i _QmFunction (also promptly from L _mArithmetic expression in eliminate i _dWith i _q).Therefore, axis error is inferred portion 41 by supposition i _δ≈ i _Qm, can be according to i _δCalculate and pass through i _QmThe represented L of function _mValue.Afterwards, use the L that is calculated _mValue, calculate axis error Δ θ from formula (41) _m

$i_d=\frac{{\mathrm{\&Phi;}}_a}{2(L_q-L_d)}-\sqrt{\frac{{{\mathrm{\&Phi;}}_a}^2}{4{(L_q-L_d)}^2}+{i_q}^2}---\left(45\right)$

In addition, can also suppose i _δ≈ iq _m, utilize L _mAs i _δThe represented approximate expression of function, obtain L _mValue, perhaps in advance will be corresponding to i _δL _mValue prepare as list data, by obtaining L with reference to this list data _mValue.

Expression L has been shown among Figure 16 _d, L _qAnd L _mI _QmCurve chart under dependent certain numerical example (is established i _γ ^*≈ 0).As shown in figure 16, L _mValue depend on i _Qm, along with i _QmIncrease and increase.The L that sets in the present embodiment _m, be equivalent to the computing parameter L in the 1st execution mode, control corresponding to L with breakdown torque _mValue, the same with L, relative L _qThe L that are present in more _dSide (with reference to Fig. 5 and Fig. 7).

L _mValue, the result is the same with the 1st execution mode, is made as to satisfy following formula (46) or formula (47).By like this, the control device of electric motor 3a of present embodiment, the same with the 1st execution mode, between d axle and γ axle, have a mind to produce deviation, by establishing i _γ ^*≈ 0, realizes the control of approximate breakdown torque control.

L _d≤L _m＜L _q …(46)

L _d≤L _m＜(L _d+L _q)/2 …(47)

In addition, L _mAlso can adopt fixed value.Also promptly, also can adopt and i _δThe irrelevant fixed value of value as L _mValue.At L _mUnder the situation for given fixed value, d shaft current i _dWith q shaft current i _qBetween relation, solid line 83 expressions by Figure 17.Dotted line 84 is that the d shaft current i under the situation of desirable breakdown torque control is carried out in expression _dWith q shaft current i _qBetween the curve of relation, can learn that from Figure 17 solid line 83 and dotted line 84 are very similar curves.

[the 2nd computational methods]

Next, to axis error Δ θ _mThe 2nd computational methods describe.In the 2nd computational methods, also the same with above-mentioned the 1st computational methods, usability is answered voltage vector E _Exm, calculate axis error Δ θ _m,, calculate phase place (θ as the γ axle of inferring axle in the control by like this _e) (also promptly inferring rotor-position).But in the 2nd computational methods, usability is not answered voltage vector E _Exmδ axle composition E _{Exm δ}Specifically, use following formula (48) to calculate axis error Δ θ _mIn addition, in the formula (48), suppose final Δ θ _mLess, use sin ^-1(E _{Exm γ}/ E _Exm) ≈ (E _{Exm γ}/ E _Exm) approximate.

$\mathrm{\&Delta;}{\mathrm{\&theta;}}_m={\sin }^{-1}\left(\frac{-E_{\mathrm{exm\&gamma;}}}{E_{\mathrm{exm}}}\right){=\sin }^{-1}\frac{-(v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}+\mathrm{\&omega;}{L}_m{i}_{\mathrm{\&delta;}})}{E_{\mathrm{exm}}}$

$\&ap;-\frac{v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}+\mathrm{\&omega;}{L}_m{i}_{\mathrm{\&delta;}}}{E_{\mathrm{exm}}}---\left(48\right)$

Axis error is inferred portion 41, calculates Δ θ in use formula (48) _mThe time, can ignore differential term pL _di _γIn addition, L _mValue decide by the method identical with method in above-mentioned the 1st computational methods.

E in the formula (48) _ExmCalculating in, use above-mentioned formula (39).As E _ExmCalculate and use approximate expression, for example can use following formula (49), (50) or (51).Formula (49) is to have utilized " p Δ θ _m≈ 0, i _Dm≈ 0, (L _d-L _p) (pi _q) ≈ 0 " and the approximate expression of approximate formula (37), formula (50) is further to have utilized " cos θ _m≈ 1 " the approximate expression of approximate formula (49), formula (51) is further to have utilized " (L _d-L _p) i _δSin θ _m＜＜Φ a " the approximate expression of approximate formula (50).In addition, when utilizing formula (49), (50) or (51), use ω _eValue as ω.

E _exm≈ω((L _d-L _q)i _δsinθ _m+Φ _a)cosθ _m …(49)

E _exm≈ω((L _d-L _q)i _δsinθ _m+Φ _a) …(50)

E _exm≈ωΦ _a …(51)

In order to calculate the θ that contains in formula (49) etc. _m, and use following formula (40).Can learn θ from formula (40) _mBe i _δFunction, so E _ExmAlso be i _δFunction.Because E _ExmCalculating comparatively complicated, therefore preferably use the suitable approximate expression that is suitable for calculating.In addition, can also be in advance with corresponding to i _δE _ExmValue prepare as list data, by obtaining E with reference to this list data _ExmValue.

[the 3rd computational methods]

Next, to axis error Δ θ _mThe 3rd computational methods describe.In the 3rd computational methods, will with the magnetic linkage Φ of the armature coil interlinkage of motor 1 _Ex, be decomposed into flux linkage vector on the qm axle and the flux linkage vector on the dm axle.Afterwards, use is as the flux linkage vector Φ of the flux linkage vector on the dm axle _Exm(≈ Φ _m, with reference to Figure 12), calculate axis error Δ θ _m,, calculate phase place (θ as the γ axle of inferring axle in the control by like this _e) (also promptly inferring rotor-position).

As establish flux linkage vector Φ _Exmγ axle composition and δ axle composition, be respectively Φ _{Exm γ}With Φ _{Exm δ}, then can learn from Figure 12, set up Δ θ _m=tan ^-1(Φ _{Exm γ}/ Φ _{Exm δ}).Because Φ _ExmBe E _ExmDivided by ω, Δ θ then _mBe expressed as following formula (52).In addition, in the formula (52), suppose final Δ θ _mLess, use tan ^-1(Φ _{Exm δ}/ Φ _{Exm γ}) ≈ (Φ _{Exm δ}/ Φ _{Exm γ}) approximate.

$\mathrm{\&Delta;}{\mathrm{\&theta;}}_m={\tan }^{-1}\frac{-{\mathrm{\&Phi;}}_{\mathrm{exm\&delta;}}}{{\mathrm{\&Phi;}}_{\mathrm{exm\&gamma;}}}={\tan }^{-1}\frac{-(\frac{v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}}{\mathrm{\&omega;}}+L_m{i}_{\mathrm{\&delta;}})}{\frac{v_{\mathrm{\&delta;}}-(R_a+p{L}_d)i_{\mathrm{\&delta;}}}{\mathrm{\&omega;}}-L_m{i}_{\mathrm{\&gamma;}}}$

$\&ap;-\frac{\frac{v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}}{\mathrm{\&omega;}}+L_m{i}_{\mathrm{\&delta;}}}{\frac{v_{\mathrm{\&delta;}}-(R_a+p{L}_d)i_{\mathrm{\&delta;}}}{\mathrm{\&omega;}}-L_m{i}_{\mathrm{\&gamma;}}}---\left(52\right)$

Axis error is inferred portion 41, calculates Δ θ in use formula (52) _mThe time, can ignore differential term pL _di _γAnd pL _di _δIn addition, L _mValue decide by the method identical with method in above-mentioned the 1st computational methods.

[the 4th computational methods]

Next, to axis error Δ θ _mThe 4th computational methods describe.In the 4th computational methods, also the same with above-mentioned the 3rd computational methods, use flux linkage vector Φ _Exm, calculate axis error Δ θ _m,, calculate phase place (θ as the γ axle of inferring axle in the control by like this _e) (also promptly inferring rotor-position).But in the 4th computational methods, do not use flux linkage vector Φ _Exmγ axle composition Φ _{Exm γ}Specifically, use following formula (53) to calculate axis error Δ θ _mIn addition, in the formula (53), suppose final Δ θ _mLess, use sin ^-1(Φ _{Exm δ}/ Φ _Exm) ≈ (Φ _{Exm δ}/ Φ _Exm) approximate.

$\mathrm{\&Delta;}{\mathrm{\&theta;}}_m={\sin }^{-1}\left(\frac{-{\mathrm{\&Phi;}}_{\mathrm{exm\&delta;}}}{{\mathrm{\&Phi;}}_{\mathrm{exm}}}\right){=\sin }^{-1}\frac{-(\frac{v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}}{\mathrm{\&omega;}}+L_m{i}_{\mathrm{\&delta;}})}{{\mathrm{\&Phi;}}_{\mathrm{exm}}}$

$\&ap;\frac{-(\frac{v_{\mathrm{\&gamma;}}-(R_a+p{L}_d)i_{\mathrm{\&gamma;}}}{\mathrm{\&omega;}}+L_m{i}_{\mathrm{\&delta;}})}{{\mathrm{\&Phi;}}_{\mathrm{exm}}}---\left(53\right)$

Axis error is inferred portion 41, calculates Δ θ in use formula (53) _mThe time, can ignore differential term pL _di _γIn addition, L _mValue decide by the method identical with method in above-mentioned the 1st computational methods.

Φ in the formula (53) _ExmCalculating in, use both sides with above-mentioned formula (39) divided by the formula after the ω.As Φ _ExmCalculate and use approximate expression, for example can use following formula (54), (55) or (56).Following formula (54), (55) and (56), the both sides that are respectively formula (49), (50) and (51) are divided by resulting formula after the ω.In addition, when utilizing formula (54), (55) or (56), use ω _eValue as ω.

Φ _exm≈((L _d-L _q)i _δsinθ _m+Φ _a)cosθ _m …(54)

Φ _exm≈((L _d-L _q)i _δsinθ _m+Φ _a) …(55)

Φ _exm≈Φ _a …(56)

In order to calculate the θ that contains in formula (54) etc. _m, and use following formula (40).Can learn θ from formula (40) _mBe i _δFunction, so Φ _ExmAlso be i _δFunction.Because Φ _ExmCalculating comparatively complicated, therefore preferably use the approximate expression that is suitable for calculating.In addition, can also be in advance with corresponding to i _δΦ _ExmValue prepare as list data, by obtaining Φ with reference to this list data _ExmValue.

K (i _δ)=1/ Φ _Exm, obtain K (i _δ) as after the correction coefficient, the axis error of the 4th computational methods is inferred the internal structure of portion 41, as shown in figure 18.In addition, also can be according to i _δValue change the gain of being adopted in the proportional integral arithmetic unit 42 (proportionality coefficient and integral coefficient), replace adopting correction coefficient K (i _δ).

[the 5th computational methods]

Next to axis error Δ θ _mThe 5th computational methods describe.In the 5th computational methods, electric current (electric current of motor model (model)) on the use dm-qm axle and the error current between the electric current on γ-δ axle calculate axis error Δ θ _m,, calculate phase place (θ as the γ axle of inferring axle in the control by like this _e) (also promptly inferring rotor-position).

Use formula that this method is described.At first, if ignore the 3rd on the right of following formula (38), then obtain following formula (57).

$\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}\\ v_{\mathrm{\&delta;}}\end{array}\right]=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_m& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+E_{\mathrm{exm}}\left[\begin{array}{c}-\sin \mathrm{\&Delta;}{\mathrm{\&theta;}}_m\\ \cos \mathrm{\&Delta;}{\mathrm{\&theta;}}_m\end{array}\right]$

$=\left[\begin{array}{cc}{R}_a+p{L}_d& -\mathrm{\&omega;}{L}_m\\ \mathrm{\&omega;}{L}_m& R_a+p{L}_d\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\\ i_{\mathrm{\&delta;}}\end{array}\right]+\left[\begin{array}{c}{E}_{\mathrm{exm\&gamma;}}\\ E_{\mathrm{exm\&delta;}}\end{array}\right]---\left(57\right)$

If by T sampling period _δCarry out discretization, then formula (57) can be rewritten as following formula (58).

$\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\left(n\right)\\ i_{\mathrm{\&delta;}}\left(n\right)\end{array}\right]=\left[\begin{array}{cc}1-\frac{R_a}{L_d}{T}_s& \mathrm{\&omega;}(n-1)\frac{L_m}{L_d}{T}_s\\ -\mathrm{\&omega;}(n-1)\frac{L_m}{L_d}{T}_s& 1-\frac{R_a}{L_d}{T}_s\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}(n-1)\\ i_{\mathrm{\&delta;}}(n-1)\end{array}\right]$

$+\frac{T_s}{L_d}\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}(n-1)\\ v_{\mathrm{\&delta;}}(n-1)\end{array}\right]-\frac{T_s}{L_d}\left[\begin{array}{c}{E}_{\mathrm{exm\&gamma;}}(n-1)\\ E_{\mathrm{exm\&delta;}}(n-1)\end{array}\right]---\left(58\right)$

In addition, infer the resulting current i of inferring of calculating of portion 41 by axis error _{M γ}And i _{M δ}, using a model has calculated E _{Exm γ}And E _{Exm δ}Infer induced voltage E _{Mexm γ}And E _{Mexm δ}, represent by following formula (59).

$\left[\begin{array}{c}{i}_{\mathrm{M\&gamma;}}\left(n\right)\\ i_{\mathrm{M\&delta;}}\left(n\right)\end{array}\right]=\left[\begin{array}{cc}1-\frac{R_a}{L_d}{T}_s& \mathrm{\&omega;}(n-1)\frac{L_m}{L_d}{T}_s\\ -\mathrm{\&omega;}(n-1)\frac{L_m}{L_d}{T}_s& 1-\frac{R_a}{L_d}{T}_s\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}(n-1)\\ i_{\mathrm{\&delta;}}(n-1)\end{array}\right]$

$+\frac{T_s}{L_d}\left[\begin{array}{c}{v}_{\mathrm{\&gamma;}}(n-1)\\ v_{\mathrm{\&delta;}}(n-1)\end{array}\right]-\frac{T_s}{L_d}\left[\begin{array}{c}{E}_{\mathrm{Mexm\&gamma;}}(n-1)\\ E_{\mathrm{Mexm\&delta;}}(n-1)\end{array}\right]---\left(59\right)$

Axis error is inferred portion 41, calculates respectively to infer induced voltage E _{Mexm γ}And E _{Mexm δ}, as E _{Mexm γ}And E _{Exm δ}Presumed value.In addition owing to use L _mReplace L _q, the calculation current i _{M γ}And i _{M δ}, therefore infer current i _{M γ}And i _{M δ}, can be referred to as the dm axle composition of inferring motor current Ia and the electric current of qm axle composition respectively.

As the fixed axis composition (i of basis by current detector 11 detected motor current Ia _uWith i _v) current i _γWith i _δ, and by calculating the resulting current i of inferring _{M γ}And i _{M δ}Between the error current Δ i of difference _γWith Δ i _δ,, represent by following formula (60) according to formula (58) and (59).

$\left[\begin{array}{c}{\mathrm{\&Delta;i}}_{\mathrm{\&gamma;}}\left(n\right)\\ {\mathrm{\&Delta;i}}_{\mathrm{\&delta;}}\left(n\right)\end{array}\right]=\left[\begin{array}{c}{i}_{\mathrm{\&gamma;}}\left(n\right)\\ i_{\mathrm{\&delta;}}\left(n\right)\end{array}\right]-\left[\begin{array}{c}{i}_{\mathrm{M\&gamma;}}\left(n\right)\\ i_{\mathrm{M\&delta;}}\left(n\right)\end{array}\right]=-\frac{T_s}{L_d}\left[\begin{array}{c}{E}_{\mathrm{exm\&gamma;}}(n-1)-E_{\mathrm{Mexm\&gamma;}}(n-1)\\ E_{\mathrm{exm\&delta;}}(n-1)-E_{\mathrm{Mexm\&delta;}}(n-1)\end{array}\right]$

$=-\frac{T_s}{L_d}\left[\begin{array}{c}\mathrm{\&Delta;}{E}_{\mathrm{exm\&gamma;}}(n-1)\\ \mathrm{\&Delta;}{E}_{\mathrm{exm\&delta;}}(n-1)\end{array}\right]---\left(60\right)$

Here, Δ E _{Exm γ}Be induced voltage E _{Exm γ}With as induced voltage E _{Exm γ}Presumed value infer induced voltage E _{Exm γ}Between error, Δ E _{Exm δ}Be induced voltage E _{Exm δ}With as induced voltage E _{Exm δ}Presumed value infer induced voltage E _{Mexm δ}Between error.

Can learn the error of the presumed value of induced voltage (Δ E from formula (60) _{Exm γ}Deng) and error current (Δ i _γDeng) proportional.Therefore, can the use error electric current by the error convergence of the presumed value of induced voltage.Also promptly, infer induced voltage E _{Exm γ}With E _{Mexm δ}, can be as having inferred induced voltage E exactly _{Exm γ}And E _{Exm δ}Person's (can infer induced voltage exactly).

Specifically, this infer induced voltage, use the last induced voltage of inferring to calculate with last estimation error.More specifically say,, calculate one by one and infer induced voltage E by following formula (61) _{Exm γ}With E _{Mexm δ}Here, g is used for allowing the feedback oscillator of error convergence of same definite value of induced voltage.

$\left[\begin{array}{c}{E}_{\mathrm{Mexm\&gamma;}}\left(n\right)\\ E_{\mathrm{Mexm\&delta;}}\left(n\right)\end{array}\right]=\left[\begin{array}{c}{E}_{\mathrm{Mexm\&gamma;}}(n-1)\\ E_{\mathrm{Mexm\&delta;}}(n-1)\end{array}\right]+g\left[\begin{array}{c}{\mathrm{\&Delta;E}}_{\mathrm{exm\&gamma;}}(n-1)\\ {\mathrm{\&Delta;E}}_{\mathrm{exm\&delta;}}(n-1)\end{array}\right]$

$=\left[\begin{array}{c}{E}_{\mathrm{Mexm\&gamma;}}(n-1)\\ E_{\mathrm{Mexm\&delta;}}(n-1)\end{array}\right]-\frac{g\&CenterDot;L_d}{T_s}\left[\begin{array}{c}{\mathrm{\&Delta;i}}_{\mathrm{\&gamma;}}\left(n\right)\\ {\mathrm{\&Delta;i}}_{\mathrm{\&delta;}}\left(n\right)\end{array}\right]---\left(61\right)$ …(61)

Afterwards, as described in the above-mentioned the 1st or the 2nd computational methods, axis error is inferred portion 41 and is used following formula (62) or (63), calculates axis error Δ θ _m

${\mathrm{\&Delta;\&theta;}}_m={\tan }^{-1}\frac{-E_{\mathrm{Mexm\&gamma;}}\left(n\right)}{E_{\mathrm{Mexm\&delta;}}\left(n\right)}\&ap;\frac{-E_{\mathrm{Mexm\&gamma;}}\left(n\right)}{E_{\mathrm{Mexm\&delta;}}\left(n\right)}---\left(62\right)$

${\mathrm{\&Delta;\&theta;}}_m={\sin }^{-1}\frac{-E_{\mathrm{Mexm\&gamma;}}\left(n\right)}{E_{\mathrm{exm}}}\&ap;\frac{-E_{\mathrm{Mexm\&gamma;}}\left(n\right)}{E_{\mathrm{exm}}}---\left(63\right)$

In addition, in formula (58)～formula (63), be recorded in the mark (n or n-1) in the bracket " () ", expression is by T sampling period _δDiscretization situation under sampling time.N is a natural number, and n represents to follow the moment that (n-1) arrived.Constitute each position of control device of electric motor 3a, per sampling period T _sCalculate and export each value one by one.Specifically, i for example _γ(n) and i _δ(n), be n the i in the sampling time _γWith i _δ, i _γ(n-1) and r _δ(n-1), be i in (n-1) individual sampling time _γWith i _δi _γWith i _δIn addition too.

As mentioned above, in the present embodiment, allow axis error Δ θ _mConverge to zero, make the γ axle follow the trail of the dm axle.Consequently, i _γWith i _δFollow the trail of i respectively _DmWith i _QmAlso promptly, we can say control device of electric motor 3a, be decomposed into qm axle composition and dm axle composition, carry out the drive controlling of motor 1 electric current that is circulated in the motor 1.Decompose resulting effect as mentioned above by this.

" the 3rd execution mode "

In addition, the formation of the control device of electric motor 3a shown in Figure 14 can also be out of shape shown in the control device of electric motor 3b of Figure 19.Implement the execution mode of this distortion, be made as the 3rd execution mode of the present invention.Control device of electric motor 3b with the estimator 40 of the control device of electric motor 3a among Figure 14, replaces to location/velocity estimator 45 (hereinafter to be referred as making estimator 45), θ _mCalculating part 46 and arithmetic unit 47.Point except this displacement, control device of electric motor 3a and the electric motor drive system of Figure 14 are identical with control device of electric motor 3b and the electric motor drive system of Figure 19.Therefore omit the formation of same part and the explanation of action.

Estimator 45 uses i _γ, i _δ, v _γ ^*And v _δ ^*, infer the phase place of the d axle of seeing mutually from U, with this presumed value as θ _DqeIn addition, estimator 45 is the same with estimator 40 in the 2nd execution mode, also calculates and infers motor speed omega _eIn addition, passing through θ _DqeCarry out differential and obtain the motor speed omega of inferring that estimator 45 exported _eSituation under, the resulting motor speed omega of inferring _eThough, should be called the presumed value of the rotating speed of d axle exactly, under normal state, this presumed value is regarded the rotational speed omega with the γ axle as _eIdentical.

θ _mCalculating part 46 will be from the i of speed controlling portion 17 _δ ^*, as the i in the above-mentioned formula (40) _δ, use above-mentioned formula (40) to calculate θ _mAt this moment, can also be with corresponding to i _δ ^*θ _mValue prepare as list data in advance, by with reference to this list data, obtain θ _mValue.

The θ that arithmetic unit 47 uses estimator 45 to be exported _DqeWith θ _mThe θ that calculating part 46 is exported _m, calculate θ _e, with the θ that is calculated _eExport to coordinate converter 12 and 18.

Like this, estimator 45, θ in the 3rd execution mode _mThe position that calculating part 46 and arithmetic unit 47 are constituted calculates the phase place (θ as the γ axle of inferring axle in the control _e).The formation of the 3rd execution mode also can access the action effect identical with the 2nd execution mode.

" the 4th execution mode "

In addition, the 1st～the 3rd execution mode adopts the mode that estimator is inferred rotor-position is set, but also can detect actual rotor-position.Also promptly, the control device of electric motor 3c of Figure 20 be can use, Fig. 3, Figure 14 or control device of electric motor shown in Figure 19 replaced.

With the electric motor drive system that includes control device of electric motor 3c shown in Figure 20, describe as the 4th execution mode of the present invention.Figure 20 is the block diagram of the associated motor drive system of the 4th execution mode.Electric motor drive system has motor 1, inverter 2 and control device of electric motor 3c.

Control device of electric motor 3c has current detector 11, coordinate converter 12, subtracter 13, subtracter 14, current control division 15, flux regulator portion 16, speed controlling portion 17, coordinate converter 18, subtracter 19, position detector 50, differentiator 51, θ _mCalculating part 52 and arithmetic unit 53.Also promptly, control device of electric motor 3c has replaced to " position detector 50, differentiator 51, θ with the estimator 40 of Figure 14 _m Calculating part 52 and arithmetic unit 53 ".Point except this displacement, all with control device of electric motor 3a and the electric motor drive system of Figure 14, and control device of electric motor 3c and the electric motor drive system of Figure 20 are the same.Constitute each position of control device of electric motor 3c, as required can the interior all values that is generated of free application controls device 3c.

Because the various piece in the control device of electric motor 3c, carry out work according to the actual rotor position of being detected rather than the rotor-position of being inferred, so in the present embodiment, " γ and the δ " with in the 2nd execution mode replaces to " dm and qm ".

Position detector 50 detects the actual rotor position θ of motor 1 by rotary encoder formations such as (rotary encoder), and this value is sent to differentiator 51 and arithmetic unit 53.51 pairs of actual rotor positions of differentiator θ carries out differential, calculates actual motor speed omega, and this value is exported to subtracter 19, flux regulator portion 16 and current control division 15.

In addition, under normal condition, actual motor speed omega is regarded as identical with the rotating speed of dm-qm axle.Therefore, though, also can use the output valve θ of arithmetic unit 53 with the input value of θ as differentiator 51 _DmReplace θ, as the input value of differentiator 51.

Speed controlling portion 17 is according to the subtraction result (ω of subtracter 19 ^*-ω), generate qm shaft current i _QmThe qm shaft current command value i that should follow the trail of _Qm ^* Flux regulator portion 16, output dm shaft current i _DmThe dm shaft current command value i that should follow the trail of _Dm ^*This dm shaft current command value i _Dm ^*Set the same with the 2nd execution mode.Also promptly, for example with i _Dm ^*Be made as near the set-point zero or zero.

The i that subtracter 13 is exported from flux regulator portion 16 _Dm ^*In, deduct the i that coordinate converter 12 is exported _Dm, calculate current error (i _Dm ^*-i _Dm).The i that subtracter 14 is exported from speed controlling portion 17 _Qm ^*In, deduct the i that coordinate converter 12 is exported _Qm, calculate current error (i _Qm ^*-i _Qm).

Current control division 15 receives each current error that subtracters 13 and 14 are calculated, from the i of coordinate converter 12 _DmWith i _Qm, and from the actual motor speed ω of differentiator 51, output V _DmThe dm shaft voltage command value V that should follow the trail of _Dm ^*And V _QmThe qm shaft voltage command value V that should follow the trail of _Qm ^*, make i _DmFollow the trail of i _Dm ^*, and i _QmFollow the trail of i _Qm ^*

θ _mCalculating part 52 will be from the i of speed controlling portion 17 _Qm ^*As the i in the above-mentioned formula (40) _δUse, use following formula (40) to calculate θ _mAt this moment, can also be with corresponding to i _Qm ^*(i _δ ^*) θ _mValue, prepare as list data in advance, by with reference to this form, obtain θ _mValue.

θ and θ that arithmetic unit 53 use location detectors 50 are detected _mThe θ that calculating part 52 is calculated _m, calculate the phase theta of the dm axle of being seen from the U phase _Dm, with the θ that is calculated _DmExport to coordinate converter 12 and 18.

Coordinate converter 18 is according to resulting θ _Dm, with V _Dm ^*And V _Qm ^*, be transformed into V _u ^*, V _v ^*, and V _w ^*The voltage in three phases command value that is constituted will be exported to the PWM inverter by the resulting value of conversion.PWM inverter 2 will be corresponding to the motor current I of this voltage in three phases command value _e, supply with motor 1, drive motor 1.

By the formation of execution mode 4, also can access effect and the effect identical with the 2nd execution mode.

" the 5th execution mode "

But the no transducer control that the 2nd and the 3rd execution mode (Figure 14 and Figure 19) is illustrated is the control according to the induced voltage that is produced etc., and is therefore particularly useful when the high speed rotating of motor 1.But when low speed rotation, it infers precision not necessarily enough, in addition, can't use when rotation stops.In the 5th execution mode, during to low speed rotation or rotation when stopping especially effectively the no transducer control based on the dm-qm axle describe.

It among Figure 24 the frame assumption diagram of the associated motor control device 3d of the 5th execution mode.Control device of electric motor 3d has current detector 11, coordinate converter 12, subtracter 13, subtracter 14, current control division 15, flux regulator portion 16, speed controlling portion 17, coordinate converter 18, subtracter 19, location/velocity estimator 200 (hereinafter to be referred as making " estimator 200 "), superimposed voltage generating unit 201, adder 202 and adder 203.Constitute each position of control device of electric motor 3d, as required can the interior all values that is generated of free application controls device 3d.

Control device of electric motor 3d, be with the difference of the control device of electric motor 3a of Figure 14, superimposed voltage generating unit 201 and adder 202 and 203 have been increased newly, and the estimator 40 among the control device of electric motor 3a of Figure 14 replaced to location/velocity estimator 200 (hereinafter to be referred as making estimator 200), other are all the same with control device of electric motor 3d and 3a.In addition, the item of being put down in writing in the 2nd execution mode, short of contradiction all can be applicable to present embodiment.

Estimator 200 is inferred out and is inferred rotor position _eAnd infer motor speed omega _eAnd output.Concrete presuming method about estimator 200 will be explained below.Coordinate converter 12 uses the rotor position of inferring that estimator 200 given _e, the U phase current i that current detector 11 is detected _uAnd V phase current i _vBe transformed into γ shaft current i _γAnd δ shaft current i _δ

Subtracter 19 is from electromotor velocity command value ω ^*In deduct the motor speed omega of inferring from estimator 200 _e, output subtraction result (velocity error).Speed controlling portion 17 is according to the subtraction result (ω of subtracter 19 ^*-ω _e), generate δ shaft current command value i _δ ^* Flux regulator portion 16, output γ shaft current command value i _γ ^*This γ shaft current command value i _γ ^*, with the same settings such as the 1st execution modes.For example, i _γ ^*Get near the set-point zero or zero.

The γ shaft current command value i that subtracter 13 is exported from flux regulator portion 16 _γ ^*In deduct the γ shaft current i that coordinate converter 12 is exported _γ, calculate current error (i _γ ^*-i _γ).The δ shaft current command value i that subtracter 14 is exported from speed controlling portion 17 _δ ^*In deduct the δ shaft current i that coordinate converter 12 is exported _δ, calculate current error (i _δ ^*-i _δ).

Current control division 15 receives each current error that subtracters 13,14 are calculated, from the γ shaft current i of coordinate converter 12 _γWith δ shaft current i _δ, and from the motor speed omega of inferring of estimator 200 _e, output γ shaft voltage command value v _γ ^*With δ shaft voltage command value v _δ ^*, make γ shaft current i _γFollow the trail of γ shaft current command value i _γ ^*, and δ shaft current i _δFollow the trail of δ shaft current command value i _δ ^*

Superimposed voltage generating unit 201 generates and is used for and γ shaft voltage command value v _γ ^*With δ shaft voltage command value v _δ ^*The superimposed voltage of stack and output.This superimposed voltage is by relative v _γ ^*γ axle superimposed voltage v _{H γ} ^*(the γ axle composition of superimposed voltage) is with relative v _δ ^*δ axle superimposed voltage v _{H δ} ^*(the δ axle composition of superimposed voltage) constitutes.Below, always be called γ axle superimposed voltage v _{H γ} ^*With δ axle superimposed voltage v _{H δ} ^*, be also referred to as superimposed voltage v sometimes _{H γ} ^*With superimposed voltage v _{H δ} ^*

The γ shaft voltage command value v that adder 202 is exported for current control division 15 _γ ^*Add γ axle superimposed voltage v _{H γ} ^*, with its addition result (v _γ ^*+ v _{H γ} ^*) export to coordinate converter 18.The δ shaft voltage command value v that adder 203 is exported for current control division 15 _δ ^*Add δ axle superimposed voltage v _{H δ} ^*, with its addition result (v _δ ^*+ v _{H δ} ^*) export to coordinate converter 18.

Coordinate converter 18 is inferred rotor position according to what estimator 200 exported _e, be superimposed with v _{H γ} ^*γ shaft voltage command value (also be (v _γ ^*+ v _{H γ} ^*)) and be superimposed with v _{H δ} ^*δ shaft voltage command value (also be (v _δ ^*+ v _{H δ} ^*)) inverse transformation, generate three-phase voltage command value (Vu ^*, V _V ^*, V _W ^*), and output it to PWM inverter 2.In this inverse transformation, use the v in the above-mentioned formula (4) _γ ^*With v _δ ^*Replaced to (v respectively _γ ^*+ v _{H γ} ^*) and (v _δ ^*+ v _{H δ} ^*) formula.PWM inverter 2 will be supplied with motor 1, drive motor 1 corresponding to the motor current Ia of this three-phase voltage command value.

Like this, pass through v _γ ^*With v _δ ^*In the represented driving voltage that is used for drive motor 1, superimposed voltage has superposeed.By the stack of this superimposed voltage, passing through γ shaft current command value i _γ ^*With δ shaft current command value i _δ ^*In the represented drive current that is used for drive motor 1, stack is corresponding to the superimposed current of above-mentioned superimposed voltage.

By the superimposed voltage that superimposed voltage generating unit 201 is generated, for example be the rotational voltage of high frequency.Here, " high frequency " is meant that for example the frequency of this superimposed voltage is compared enough big with the frequency of driving voltage.Therefore, according to the frequency of the above-mentioned superimposed current that this superimposed voltage superposeed, compare also enough big with the frequency of above-mentioned drive current.In addition, " rotational voltage " is meant and allows the voltage vector of superimposed voltage form circular voltage on the fixed coordinates axle.

Too, the voltage vector track of the superimposed voltage that superimposed voltage generating unit 201 is generated forms the such circle of voltage vector track 210 of Figure 25 for example under situation about considering on d-q axle or the γ-rotatable coordinate axis such as δ axle.In superimposed voltage is under the three-phase equilibrium voltage condition, its voltage vector track, and shown in voltage vector track 210, being formed on the rotatable coordinate axis with the initial point is the just round of center.This rotational voltage (superimposed voltage) be and motor 1 nonsynchronous high frequency voltage, so the loading of this rotational voltage can not cause motor 1 rotation.

In addition, be to imbed magnet type synchronous motor etc., L at motor 1 _d＜L _qDuring establishment, because of the superimposed voltage that forms voltage vector track 210 makes the current phasor track of the superimposed current that circulated in the motor 1, shown in the current phasor track 211 of Figure 26, being formed on γ-δ axle with the initial point is the center, the γ direction of principal axis is a long axis direction, and the δ direction of principal axis is the ellipse of short-axis direction.But current phasor track 211 is that the axis error Δ θ between d axle and the γ axle is the current phasor track under zero the situation.The current phasor track of the superimposed current under the non-vanishing situation of axis error Δ θ is the represented ellipse of current phasor track 212, and its long axis direction and γ direction of principal axis are inconsistent.Also promptly under the non-vanishing situation of axis error Δ θ, on γ-δ axle, be the center with the initial point, current phasor track 211 tilts, and depicts current phasor track 212.

If establish the γ axle composition and the δ axle composition of superimposed current, be respectively γ axle superimposed current i _{H γ}And δ axle superimposed current i _{H δ}, its product (i then _{H γ}* i _{H δ}) in, there is the flip-flop depend on by the inclination of the represented ellipse of current phasor track 212.Product (i _{H γ}* i _{H δ}), in the 1st and the 3rd quadrant of current phasor track, get on the occasion of, in addition, in the 2nd and the 4th quadrant, get negative value, therefore when ellipse does not tilt (under the situation of current phasor track 211), do not contain flip-flop, if but oval inclination (under the situation of current phasor track 212), just contain flip-flop.In addition, the I among Figure 26, II, III, IV, the 1st, the 2nd, the 3rd, the 4th quadrant on expression γ-δ axle.

Among Figure 27, the time is transverse axis, and axis error Δ θ is the product (i under zero the situation _{H γ}* i _{H δ}) represent by curve 220 and 221 respectively with the flip-flop of this product.Among Figure 28, the time is transverse axis, the product (i under the non-vanishing situation of axis error Δ θ _{H γ}* i _{H δ}) represent by curve 222 and 223 respectively with the flip-flop of this product.Can learn product (i from Figure 27 and Figure 28 _{H γ}* i _{H δ}) flip-flop, be zero under the situation of Δ θ=0 °, non-vanishing under the situation of Δ θ ≠ 0 °.In addition, this flip-flop increases along with the size increase of axis error Δ θ (being directly proportional with axis error Δ θ substantially).Therefore, make this flip-flop converge to zero if control for the time being, then axis error Δ θ also converges to zero.

The estimator 200 of Figure 24 is conceived to this point and infers action.But in order to infer the dm-qm axle, infer action, make axis error Δ θ between dm axle and the γ axle _mRather than the axis error Δ θ of d axle and γ axle converges to zero.

Interior block diagram as estimator (speed/positional estimator) 200a of one of estimator 200 example has been shown among Figure 29.Estimator 200a has axis error and infers portion 231, proportional integral calculator 232 and integrator 233.Proportional integral arithmetic unit 232 and integrator 233, proportional integral arithmetic unit 31 and the integrator 32 with Fig. 4 is identical respectively.

Axis error is inferred portion 231 and is used i _γAnd i _δ, calculate axis error Δ θ _mProportional integral arithmetic unit 232 is used for realizing PLL (Phase Locked Loop), and is moving with each position association that constitutes control device of electric motor 3d, carries out proportional plus integral control, calculates and infers motor speed omega _e, make axis error infer the axis error Δ θ that portion 231 is calculated _mConverge to zero.Integrator 233 Comparative Examples integrator computing units 232 are exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eProportional integral arithmetic unit 232 is exported infers motor speed omega _eInferred rotor position with integrator 233 exports _e,, send to each position that needs the control device of electric motor of this value 3d all as the output valve of estimator 200a.In addition, infer motor speed omega _eAlso export to axis error and infer portion 231.

The axis error that Figure 29 has been shown among Figure 30 is inferred the inside configuration example of portion 231.As shown in figure 30, axis error is inferred portion 231 and is had BPF (band pass filter) 241, LPF (low pass filter) 242, θ _mCalculating part 243, rotation of coordinate portion 244 and axis error calculating part 245.In addition, as shown in figure 31, axis error calculating part 245 has multiplier 246, LPF247, coefficient multiplier 248.Now, establish the superimposed voltage v that superimposed voltage generating unit 201 is generated _{H γ} ^*With v _{H δ} ^*Frequency (the electric angle speed on γ-δ reference axis) be ω _h

BFP241, the γ shaft current i that is exported from the coordinate converter 12 of Figure 24 _γWith δ shaft current i _δIn, extract ω out _hFrequency content, output γ axle superimposed current i _{H γ}And δ axle superimposed current i _{H δ}BPF241 is to receive i _γWith i _δAs input signal, by including ω in the passband _hThe band pass filter of frequency, for example is that its centre frequency by passband is ω typically _hIn addition, remove the frequency content of drive current by BPF241.

LFP242, the γ shaft current i that is exported from the coordinate converter 12 of Figure 24 _γWith δ shaft current i _δIn, remove ω as radio-frequency component _hFrequency content after, export to θ _mCalculating part 243.Also promptly, by LPF242, from γ shaft current i _γWith δ shaft current i _δMiddle superimposed current (the i that removes _{H γ}With i _{H δ}) composition.

θ _mCalculating part 243 is according to having removed ω _hFrequency content after γ shaft current i _γWith δ shaft current i _δValue, calculate phase theta _m(with reference to Figure 11).Specifically, will remove ω _hFrequency content after δ shaft current i _δValue, as the i in the above-mentioned formula (40) _δ, use following formula (40) to calculate θ _mAt this moment, also can be with corresponding to i _δθ _mValue prepare as list data in advance, by with reference to this list data, obtain θ _mValue.

Rotation of coordinate portion 244 uses following formula (64), allows superimposed current i _{H γ}And i _{H δ}Formed current phasor i _h, rotation of coordinate is passed through θ _mRepresented phase bit position calculates current phasor i _HmAt this moment, use θ _mThe θ that calculating part 243 is calculated _mValue.Current phasor i _hAnd i _HmBe expressed as following formula (65a) and (65b).i _{H γ}And i _{H δ}Be to form current phasor i _h2 compositions of quadrature, it is respectively current phasor i _hγ axle composition and δ axle composition.i _{Hm γ}And i _{Hm δ}Be to form current phasor i _Hm2 compositions of quadrature.The i that is calculated by rotation of coordinate portion 244 _{Hm γ}And i _{Hm δ}, send to axis error calculating part 245.

$\left[\begin{array}{c}{i}_{\mathrm{hm\&gamma;}}\\ i_{\mathrm{hm\&delta;}}\end{array}\right]=\left[\begin{array}{cc}\cos {\mathrm{\&theta;}}_m& \sin {\mathrm{\&theta;}}_m\\ -\sin {\mathrm{\&theta;}}_m& \cos {\mathrm{\&theta;}}_m\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{h\&gamma;}}\\ i_{\mathrm{h\&delta;}}\end{array}\right]---\left(64\right)$

$i_h=\left[\begin{array}{c}{i}_{\mathrm{h\&gamma;}}\\ i_{\mathrm{h\&delta;}}\end{array}\right]---\left(65a\right)$

$i_{\mathrm{hm}}=\left[\begin{array}{c}{i}_{\mathrm{hm\&gamma;}}\\ i_{\mathrm{hm\&delta;}}\end{array}\right]---\left(65b\right)$

With reference to Figure 32 of the current phasor track example before and after this rotation of coordinate of expression, the meaning of rotation of coordinate is remarked additionally.The just situation of round rotational voltage of having considered to have superposeed has also promptly loaded the situation of the superimposed voltage of the voltage vector track 210 of depicting Figure 25.In this case, because of the magnetic salient pole of motor 1, the current phasor i on the rotatable coordinate axis _hTrack, shown in current phasor track 251, form the axisymmetric ellipse of relative d axle (also promptly forming the corresponding to ellipse of d direction of principal axis and long axis direction).Rotation of coordinate portion 244 is for making the relative dm axle of this ellipse axial symmetry, and gives current phasor i _hThe effect spin matrix calculates current phasor i _HmBy like this, current phasor i _HmTrack shown in current phasor track 252.

On rotatable coordinate axis, current phasor track 252 forms oval, and its long axis direction is at Δ θ _mConsistent in the time of=0 ° with the dm direction of principal axis, but at Δ θ _mIn the time of ≠ 0 ° and the dm direction of principal axis inconsistent.Therefore, if with current phasor i _HmThe product (i of 2 compositions of quadrature _{Hm γ}* i _{Hm δ}) flip-flop, be expressed as (i _{Hm γ}* i _{Hm δ}) _DC, then, with product (i _{H γ}* i _{H δ}) flip-flop the same with relation between the axis error Δ θ, flip-flop (i _{Hm γ}* i _{Hm δ}) _DCAt axis error Δ θ _mBe to be zero under zero the situation, substantially and axis error Δ θ _mBe directly proportional.Therefore, be K if establish proportionality coefficient, axis error Δ θ then _mCan represent by following formula (66).

Δθ _m＝K·(i _hmγ×i _hmδ) _DC …(66)

For the represented calculating of realization formula (66), axis error calculating part 245 constitutes as shown in figure 31.Also promptly, multiplier 246 calculates the i that is calculated by rotation of coordinate portion 244 _{Hm γ}With i _{Hm δ}Product, LPF247 extracts this product (i _{Hm γ}* i _{Hm δ}) flip-flop, obtain (i _{Hm γ}* i _{Hm δ}) _DCCoefficient multiplier 248 is given the flip-flop (i that LPF247 exported _{Hm γ}* i _{Hm δ}) _DCMultiply by proportionality coefficient K, calculate the represented axis error Δ θ of formula (66) _mThe axis error Δ θ that coefficient multiplier 248 is exported _m, infer the axis error Δ θ that portion 231 is inferred as the axis error of Figure 29 _m, send to proportional integral arithmetic unit 232, infer motor speed omega as mentioned above _eAnd infer rotor position _eCalculating, make axis error Δ θ _mConverge to zero.Also promptly, make γ-δ axle follow the trail of dm-qm axle (inferring the dm-qm axle).

As mentioned above, if the superimposed voltage of overlapped high-frequency according to the superimposed current composition that correspondingly circulates, is inferred rotor-position, particularly under the halted state or low-speed running state of motor 1, can infer rotor-position well.

[about the variation of superimposed voltage]

The most typical example listed the situation of the superimposed voltage that superimposed voltage generating unit 201 generated for just round rotational voltage, but the superimposed voltage that superimposed voltage generating unit 201 is generated can also adopt various superimposed voltage.And the voltage vector track of (the d-q axle is first-class) on rotatable coordinate axis of superimposed voltage needs not be the axisymmetric track of relative d axle.Say in more detail, superimposed voltage voltage vector track of (the d-q axle is first-class) on rotatable coordinate axis, bag initial point in need depicting, and to have with the d axle be the symmetric figure of benchmark.Current phasor track at the caused superimposed current of loading of superimposed voltage is under the precondition of the axisymmetric track of relative d axle, the axis error that constitutes Figure 29 is inferred portion 231, therefore by allowing the relative d axle of voltage vector track axial symmetry, just satisfy this precondition.

Here, " interior bag initial point " is meant, the initial point of (the d-q axle is first-class) on the rotatable coordinate axis is present in above-mentioned " having symmetric figure " inside.In addition, " is that benchmark has symmetry with the d axle " is meant the 1st quadrant part and the 2nd quadrant figure partly of the track of the voltage vector on the d-q axle, and between the figure of the 3rd quadrant and the 4th quadrant part, satisfying with the d axle is the line symmetric relation of axle.

For example, the voltage vector track of the superimposed voltage on the rotatable coordinate axis in (the d-q axle is first-class), can be to be the ellipse of short-axis direction or long axis direction with the d direction of principal axis, also can be the line (also being that superimposed voltage can be an alternating voltage) on d axle or the q axle, can also be to be the quadrangle at center with the initial point.

Among Figure 33, show oval-shaped rotational voltage as the current phasor i under the situation of superimposed voltage loading _hAnd i _HmTrack.Among Figure 34, show and only to have the alternating voltage of d axle composition as the current phasor i under the situation of superimposed voltage loading _hAnd i _HmTrack.

But, not under the situation of just justifying at the voltage vector track that loads superimposed voltage, need be with the θ of Figure 30 _mThe calculated value phase theta of calculating part 243 _m, send to the superimposed voltage generating unit 201 of Figure 24.This be because, under the voltage vector track of superimposed voltage is just round situation, the voltage vector track of superimposed voltage and the phase-independent of superimposed voltage, relative d axle axial symmetry, but,, need phase theta in order to allow the relative d axle of this voltage vector track axial symmetry not being under the situation of just justifying _mInformation.

For example, in superimposed voltage is under the situation of just round rotational voltage, superimposed voltage generating unit 201 generates by the represented superimposed voltage of following formula (67), in superimposed voltage is that superimposed voltage generating unit 201 generates by the represented superimposed voltage of following formula (68) under the situation of oval rotational voltage or alternating voltage.Here, V _{H γ}And V _{H δ}, be respectively the axial amplitude of γ of superimposed voltage and the axial amplitude of δ of superimposed voltage.The t express time.

$\left[\begin{array}{c}{v_{\mathrm{h\&gamma;}}}^{*}\\ {v_{\mathrm{h\&delta;}}}^{*}\end{array}\right]=\left[\begin{array}{c}{V}_{\mathrm{h\&gamma;}}\cos \left({\mathrm{\&omega;}}_{h}t\right)\\ V_{\mathrm{h\&delta;}}\sin \left({\mathrm{\&omega;}}_{h}t\right)\end{array}\right]---\left(67\right)$

$\left[\begin{array}{c}{v_{\mathrm{h\&gamma;}}}^{*}\\ {v_{\mathrm{h\&delta;}}}^{*}\end{array}\right]=\left[\begin{array}{c}{V}_{\mathrm{h\&gamma;}}\cos ({\mathrm{\&omega;}}_{h}t+{\mathrm{\&theta;}}_m)\\ V_{\mathrm{h\&delta;}}\sin ({\mathrm{\&omega;}}_{h}t+{\mathrm{\&theta;}}_m)\end{array}\right]---\left(68\right)$

[derivation of the theoretical formula of axis error]

More than to utilizing axis error Δ θ _mWith flip-flop (i _{Hm γ}* i _{Hm δ}) _DCBe directly proportional, the method for inferring of carrying out the dm-qm axle is illustrated, and here the theoretical formula of the principle of inferring about this is studied.And, for convenience of explanation, carry out about the investigation under the situation of inferring of d-q axle.Also promptly, calculate the derivation of the theoretical formula under the situation of the axis error Δ θ between d axle and the γ axle.

At first, represent by following formula (69) about the equation of stack composition.Here, following formula (70a), (70b), (70c), (70d) and (70e) establishment.P is the symbol of differentiating in addition.

$p\left[\begin{array}{c}{i}_{\mathrm{h\&gamma;}}\\ i_{\mathrm{h\&delta;}}\end{array}\right]=\frac{1}{L_d{L}_q}\left[\begin{array}{cc}{L}_{\mathrm{\&delta;}}& -L_{\mathrm{\&gamma;\&delta;}}\\ -L_{\mathrm{\&gamma;\&delta;}}& L_{\mathrm{\&gamma;}}\end{array}\right]\left[\begin{array}{c}{v_{\mathrm{h\&gamma;}}}^{*}\\ {v_{\mathrm{h\&delta;}}}^{*}\end{array}\right]---\left(69\right)$

L _γ＝L ₀+L ₁cos2Δθ …(70a)

L _δ＝L ₀-L ₁cos2Δθ …(70b)

L _γδ＝L ₁sin2Δθ …(70c)

${\mathrm{<figure-callout\; id="0"\; label="L"\; filenames="A20061012184200632.png,A20061012184200641.png"\; state="\{\{state\}\}">L</figure-callout>}}_0=\frac{L_d+L_q}{2}---\left(70d\right)$

${\mathrm{<figure-callout\; id="1"\; label="L"\; filenames="A20061012184200631.png,A20061012184200632.png"\; state="\{\{state\}\}">L</figure-callout>}}_1=\frac{L_d-L_q}{2}---\left(70e\right)$

If the superimposed voltage that is loaded is by following formula (67) expression, 2 component i of quadrature of the superimposed current that circulates corresponding to the loading of this superimposed voltage then _{H γ}With i _{H δ}, represent by following formula (71).S in the formula (71) is the Laplace's operation symbol, θ _h=ω _hT.

$\left[\begin{array}{c}{i}_{\mathrm{h\&gamma;}}\\ i_{\mathrm{h\&delta;}}\end{array}\right]=\frac{1}{L_d{L}_q}\left[\begin{array}{cc}{L}_{\mathrm{\&delta;}}& -L_{\mathrm{\&gamma;\&delta;}}\\ -L_{\mathrm{\&gamma;\&delta;}}& L_{\mathrm{\&gamma;}}\end{array}\right]\frac{1}{s}\left[\begin{array}{c}{v_{\mathrm{h\&gamma;}}}^{*}\\ {v_{\mathrm{h\&delta;}}}^{*}\end{array}\right]$

$=\frac{1}{{\mathrm{\&omega;}}_h{L}_d{L}_q}\left[\begin{array}{cc}{L}_0-{\mathrm{<figure-callout\; id="1"\; label="L"\; filenames="A20061012184200631.png,A20061012184200632.png"\; state="\{\{state\}\}">L</figure-callout>}}_1\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">cos</figure-callout>}2\mathrm{\&Delta;\&theta;}& -{\mathrm{<figure-callout\; id="1"\; label="L"\; filenames="A20061012184200631.png,A20061012184200632.png"\; state="\{\{state\}\}">L</figure-callout>}}_1\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">sin</figure-callout>}2\mathrm{\&Delta;\&theta;}\\ -{\mathrm{<figure-callout\; id="1"\; label="L"\; filenames="A20061012184200631.png,A20061012184200632.png"\; state="\{\{state\}\}">L</figure-callout>}}_1\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">sin</figure-callout>}2\mathrm{\&Delta;\&theta;}& {\mathrm{<figure-callout\; id="0"\; label="L"\; filenames="A20061012184200632.png,A20061012184200641.png"\; state="\{\{state\}\}">L</figure-callout>}}_0+{\mathrm{<figure-callout\; id="1"\; label="L"\; filenames="A20061012184200631.png,A20061012184200632.png"\; state="\{\{state\}\}">L</figure-callout>}}_1\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="A20061012184200681.png,A20061012184200692.png"\; state="\{\{state\}\}">cos</figure-callout>}2\mathrm{\&Delta;\&theta;}\end{array}\right]\left[\begin{array}{c}{V}_{\mathrm{h\&gamma;}}\sin {\mathrm{\&theta;}}_h\\ -V_{\mathrm{h\&delta;}}\cos {\mathrm{\&theta;}}_h\end{array}\right]---\left(71\right)$

If according to above-mentioned formula (71), 2 compositions of quadrature of superimposed current are put in order, then obtain following formula (72).Here, K ₁～K ₇Be if determined L _d, L _q, V _{H γ}And V _{H δ}Just the coefficient of Que Dinging.

i _hγ×i _hδ＝K ₁sin(2θ _h)+K ₂sin(2Δθ)+K ₃sin(4Δθ)+K ₄sin(2Δθ+2θ _h)

+K ₅sin(2Δθ-2θ _h)+K ₆sin(4Δθ+2θ _h)+K ₇sin(4Δθ-2θ _h)

…(72)

Current phasor i _hThe product (i of 2 compositions of quadrature _{H γ}* i _{H δ}) flip-flop be expressed as (i _{H γ}* i _{H δ}) _DCBecause flip-flop does not comprise because of θ _hAnd therefore the item of change is expressed as shown in the formula (73).

(i _hγ×i _hδ) _DC＝K ₂sin(2Δθ)+K ₃sin(4Δθ) …(73)

Under the situation of Δ θ ≈ 0, owing to can be approximately sin (2 Δ θ) ≈ 2 Δ θ, sin (4 Δ θ) ≈ 4 Δ θ, so axis error Δ θ can represent by following formula (74).K in the formula (74) passes through COEFFICIENT K ₂With K ₃The coefficient that sets.In addition, be under the situation of just round rotational voltage in superimposed voltage, COEFFICIENT K ₃Vanishing, according to formula (73), 4 times the sine of Δ θ does not have mutually.

Δθ＝K·(i _hγ×i _hδ) _DC …(74)

By abduction, be applicable to axis error Δ θ with above-mentioned formula (74) _m, just can access above-mentioned formula (66).

[variation of axis error calculating part]

Above illustration use the resulting current phasor i of rotation of coordinate by the rotation of coordinate portion 244 of Figure 30 _Hm2 compositions of quadrature (also be i _{Hm γ}With i _{Hm δ}) both sides, calculate axis error Δ θ _mSituation, (also be i but also can only use 1 composition in 2 compositions of this quadrature _{Hm γ}Or i _{Hm δ}), calculate axis error Δ θ _mBut only using 1 composition to calculate axis error Δ θ _mSituation under, the superimposed voltage that the superimposed voltage generating unit 201 of Figure 24 is generated must be the alternating voltage that only has d axle composition or q axle composition.

The stack alternating voltage (also is i according to 1 composition of the vector of the superimposed current that circulates corresponding to the stack of this alternating voltage _{H γ}Or i _{H δ}), the method that calculates axis error Δ θ just known very early (for example with reference to above-mentioned patent documentation 3).Therefore the detailed description of omitting this method.By the d-q axle is replaced to the dm-qm axle, and use this method, can only use current phasor i _Hm1 composition (also be i _{Hm γ}Or i _{Hm δ}), calculate axis error Δ θ _m

In addition, overlapped high-frequency voltage, the method for inferring rotor-position and electromotor velocity according to the electric current that is correspondingly circulated has a lot (reference example such as above-mentioned patent documentation 4～6).These methods also can be diverted to the dm-qm axle, infer processing.

" the 6th execution mode "

By controlling corresponding to the no transducer that is particularly suitable for low speed rotation state (and halted state) of the 5th execution mode, with the no transducer control combination that is particularly suitable for the high speed rotating state corresponding to the 2nd or the 3rd execution mode, can in comprising the big velocity interval of halted state, realize good no transducer control.The 6th execution mode is described as the execution mode corresponding to this combination.2nd, the item that has illustrated in the 3rd and the 5th execution mode, short of contradiction just can be applicable to the 6th execution mode.

Because all formations of associated motor control device of the 6th execution mode, therefore identical with Figure 24 omitted diagram separately.But in the associated motor control device of the 6th execution mode, the inside of estimator 200 constitutes different with the estimator 200a of Figure 29.Therefore, inside formation and the action of conduct with the estimator 200 of the difference of the 5th execution mode described.

Below with the 1st, the 2nd and the 3rd estimator example, describe as the example of the estimator 200 that can be applicable to present embodiment.

[the 1st estimator example]

At first, the 1st estimator example is described.Figure 35 is the inside configuration example of the relevant position/speed estimating device 200b (hereinafter to be referred as making " estimator 200b ") of the 1st estimator.Estimator 200b can use the estimator 200 among Figure 24.

Estimator 200b has the 1st axis error and infers portion's the 261, the 2nd axis error and infer portion 262, hand-off process portion 263, proportional integral arithmetic unit 264 and integrator 265.

The 1st axis error infer portion 261 be with the 5th execution mode in the axis error of illustrated Figure 29 infer the identical position of portion 231, according to i _γAnd i _δ, calculate the axis error of qm axle and δ axle.But, in the 5th execution mode, the axis error Δ θ that the axis error that this calculated should be calculated as estimator 200 (or 200a) _mHandle, and in the present embodiment, with its axis error Δ θ that should calculate as estimator 200b _mCandidate.Also promptly, the 1st axis error is inferred portion 261, and the qm axle that will be calculated via axis error calculating part 245 grades (with reference to Figure 30) of the inscape that becomes self and the axis error of δ axle are as axis error Δ θ _mCandidate calculate and export.The 1st axis error is inferred the output valve of portion 261, is called the 1st candidate axis error Δ θ _M1In addition, the 1st axis error is inferred portion 261, is calculating the 1st candidate axis error Δ θ _m1 o'clock, the output valve of application percentage integrator computing unit 264 (was inferred motor speed omega as required _e).

The 2nd axis error infer portion 262 be with the 2nd execution mode in the axis error of illustrated Figure 15 infer the identical position of portion 41, use V _γ ^*, V _δ ^*, i _γAnd i _δAll or part of of value, according to the 1st～the 5th illustrated in the 2nd execution mode computational methods etc., calculate the axis error of qm axle and δ axle, as axis error Δ θ _mCandidate and output.The 2nd axis error is inferred the output valve of portion 262, is called the 2nd candidate axis error Δ θ _M2In addition, the back also will illustrate, the 2nd candidate axis error Δ θ _M2Calculating the time employed i _γAnd i _δValue in, should not contain the caused superimposed current (i of superimposed voltage _{H γ}And i _{H δ}) composition.In addition, the 2nd axis error is inferred portion 262, is calculating the 2nd candidate axis error Δ θ _M2The time, the output valve of application percentage integrator computing unit 264 (is inferred motor speed omega as required _e).

Hand-off process portion 263 is with the 1st candidate axis error Δ θ _M1Receive as the 1st input value, simultaneously with the 2nd candidate axis error Δ θ _M2Receive as the 2nd input value,, calculate output valve and export according to the 1st input value and the 2nd input value corresponding to the velocity information of the rotating speed of the rotor of expression motor 1.Among the estimator 200b, usage ratio integrator computing unit 264 is calculated infers motor speed omega _eAs velocity information.But also can use electromotor velocity command value ω ^*As velocity information.

Hand-off process portion 263 is for example shown in Figure 36, corresponding to velocity information the either party in the 1st input value and the 2nd input value is directly exported as output valve.In this case, hand-off process portion 263 is passing through the represented rotating speed of velocity information less than given threshold velocity VT _HThe time, the 1st input value is exported as output valve, greater than threshold velocity V _THThe time, the 2nd input value is exported as output valve.

In addition, can also allow hand-off process portion 263 be weighted average treatment.In this case, hand-off process portion 263 is passing through the represented rotating speed of velocity information less than the 1st given threshold velocity V _TH1The time, the 1st input value is exported as output valve, greater than the 2nd given threshold velocity V _TH2The time, the 2nd input value is exported as output valve.In addition, be in the 1st threshold velocity V by the represented rotating speed of velocity information _TH1To the 2nd threshold velocity V _TH2Scope in the time, the weighted average of the 1st input value and the 2nd input value is calculated and exports as output valve.Here, set up V _TH1＜V _HT2

Weighted average is handled, and for example carries out corresponding to the represented rotating speed of velocity information.Also promptly, shown in the ideograph of Figure 37, be in the 1st threshold velocity V by the represented rotating speed of velocity information _TH1To the 2nd threshold velocity V _TH2Scope in the time, carry out the weighted average of the 1st input value and the 2nd input value,, increase the contribution rate (contributionratio) of the 2nd input value output valve along with the increase of this rotating speed, along with the minimizing of this rotating speed, increase the contribution rate of the 1st input value to output valve.

In addition, for example weighted average is handled, and the elapsed time after beginning corresponding to switching carries out.Also promptly, for example shown in the ideograph of Figure 38, with the represented rotating speed of velocity information from less than the 1st threshold velocity V _TH1State begin to enter into greater than the 1st threshold velocity V _TH1The moment t1 of state be benchmark, begin output valve is switched to the 2nd input value from the 1st input value.When moment t1, output valve for example is the 1st input value.Afterwards, carry out the weighted average of the 1st input value and the 2nd input value,, increase the contribution rate of the 2nd input value output valve along with the elapsed time since moment t1 increases.Having passed through the moment after the preset time from the moment t1 that switches beginning, make output valve consistent with the 2nd input value, finish to switch.At the represented rotating speed of velocity information from greater than the 2nd threshold velocity V _TH2State to less than the 2nd threshold velocity V _TH2The situation of state transition under too.In addition, be weighted under the situation of average treatment the 1st threshold velocity V in the elapsed time of switching beginning corresponding to distance _TH1With the 2nd threshold velocity V _TH2Also can be identical.

In addition, threshold velocity V _THFor example be the interior rotating speed of scope of 10rps (rotation per second)～30rps, the 1st threshold velocity V _TH1For example be the rotating speed in 10rps～20rps scope, the 2nd threshold velocity V _TH2It for example is the rotating speed in 20rps～30rps scope.

Among the estimator 200b of Figure 35,, send to the proportional integral arithmetic unit 264 of the effect of playing speed estimating portion with the output valve of hand-off process portion 263.Proportional integral arithmetic unit 264 is used for realizing PLL, and is moving with each position association that constitutes control device of electric motor 3d, carries out proportional plus integral control, calculates and infers motor speed omega _e, make the output valve of hand-off process portion 263 converge to zero.Integrator 265 Comparative Examples integrator computing units 264 are exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eProportional integral arithmetic unit 264 is exported infers motor speed omega _eInferred rotor position with integrator 265 exports _e,, send to each position that needs the control device of electric motor of this value 3d all as the output valve of estimator 200b.

[the 2nd estimator example]

Next, the 2nd estimator example is described.Figure 39 is the inside configuration example of the relevant position/speed estimating device 200c (hereinafter to be referred as making " estimator 200c ") of the 2nd estimator.Estimator 200c can use the estimator 200 among Figure 24.

Estimator 200c has the 1st axis error and infers portion's the 261, the 2nd axis error and infer portion 262, hand-off process portion 263, proportional integral arithmetic unit 266 and 267 and integrator 268.Among the estimator 200c, carry out hand-off process in the stage of the speed of inferring.

The 1st axis error among the estimator 200c infers portion 261 and the 2nd error is inferred portion 262, identical with among the estimator 200b of Figure 35.But the 1st error is inferred portion 261, is calculating the 1st candidate axis error Δ θ _M1The time, as required with output valve (the 1st candidate speed omega described later of proportional integral arithmetic unit 266 _E1) as inferring motor speed omega _eUtilize.Equally, the 2nd error is inferred portion 262, is calculating the 2nd candidate axis error Δ θ _M2The time, as required with output valve (the 2nd candidate speed omega described later of proportional integral arithmetic unit 267 _E2) as inferring motor speed omega _eCarry out processing and utilizing.

Proportional integral arithmetic unit 266 is used for realizing PLL, and is moving with each position association that constitutes control device of electric motor 3d, carries out proportional plus integral control, calculates and infers electromotor velocity, makes the 1st candidate axis error Δ θ _M1Converge to zero.Proportional integral arithmetic unit 266 is calculated infers electromotor velocity, as the 1st candidate speed omega _E1Output.

Proportional integral arithmetic unit 267 is used for realizing PLL, and is moving with each position association that constitutes control device of electric motor 3d, carries out proportional plus integral control, calculates and infers electromotor velocity, makes the 2nd candidate axis error Δ θ _M2Converge to zero.Proportional integral arithmetic unit 267 is calculated infers electromotor velocity, as the 2nd candidate speed omega _E2Output.

Hand-off process portion 263 among the estimator 200c, identical with among the estimator 200b of Figure 35.But among the estimator 200c, the 1st input value of hand-off process portion 263 and the 2nd input value become the 1st candidate speed omega respectively _E1With the 2nd candidate speed omega _E2Therefore, the hand-off process portion 263 among the estimator 200c corresponding to the velocity information of rotating speed of the rotor of expression motor 1, exports the 1st candidate speed omega _E1, the 2nd candidate speed omega _E2Or its weighted average.In addition, velocity information can also be used electromotor velocity command value ω ^*The output valve of hand-off process portion 263 becomes the motor speed omega of inferring that is suitable for rotating speed _e

268 pairs of hand-off process portions 263 of integrator are exported infers motor speed omega _eCarry out integration, calculate and infer rotor position _eHand-off process portion 263 is exported infers motor speed omega _eInferred rotor position with integrator 268 exports _e,, send to each position that needs the control device of electric motor of this value 3d all as the output valve of estimator 200c.

[the 3rd estimator example]

Next, the 3rd estimator example is described.Figure 40 is the inside configuration example of the relevant position/speed estimating device 200d (hereinafter to be referred as making " estimator 200d ") of the 3rd estimator.Estimator 200d can use the estimator 200 among Figure 24.

Estimator 200d has the 1st axis error and infers portion's the 261, the 2nd axis error and infer portion 262, hand-off process portion 263, proportional integral arithmetic unit 266 and 267 and integrator 269 and 270.Among the estimator 200d, carry out hand-off process in the stage of estimated position.

The 1st axis error among the estimator 200d infers portion 261 and the 2nd error is inferred portion 262, and proportional integral arithmetic unit 266 and 267, identical with among the estimator 200c of Figure 39.The 1st candidate speed omega that integrator 269 Comparative Examples integrator computing units 266 are exported _E1Carry out integration, calculate the 1st candidate position θ _E1The 2nd candidate speed omega that integrator 270 Comparative Examples integrator computing units 267 are exported _E2Carry out integration, calculate the 2nd candidate position θ _E2

Hand-off process portion 263 among the estimator 200d, identical with among the estimator 200b of Figure 35.But among the estimator 200d, the 1st input value of hand-off process portion 263 and the 2nd input value become the 1st candidate position θ respectively _E1With the 2nd candidate position θ _E2Therefore, the hand-off process portion 263 among the estimator 200d corresponding to the velocity information of rotating speed of the rotor of expression motor 1, exports the 1st candidate position θ _E1, the 2nd candidate position θ _E2Or its weighted average.In addition, velocity information can also be used electromotor velocity command value ω ^*The output valve of hand-off process portion 263 becomes the rotor position of inferring that is suitable for rotating speed _e

Hand-off process portion 263 is exported infers rotor position _e,, send to each position that needs the control device of electric motor of this value 3d as the output valve of estimator 200d.When needing, can also infer rotor position to what hand-off process portion 263 was exported _eCarry out differential, calculate and infer motor speed omega _e

In addition, the item of being put down in writing in the explanation of estimator 200b, short of contradiction just can be applicable to estimator 200c and 200d.

In addition, when low speed rotation etc., need the 1st axis error to infer the 1st candidate axis error Δ θ that portion 261 is calculated _M1The moment, need be based on the generation of the superimposed voltage of superimposed voltage generating unit 201, but when high speed rotating etc., do not needing the 1st axis error to infer the 1st candidate axis error Δ θ that portion 261 is calculated _M1The moment, can stop generation based on the superimposed voltage of superimposed voltage generating unit 201.For the 2nd candidate axis error Δ θ _M2Calculating necessary be corresponding to driving voltage (v _γ ^*And v _δ ^*) the drive current composition that circulated, this be because, superimposed current becomes to be divided into the 2nd candidate axis error Δ θ _M2The noise of calculating.

But, being weighted under the situations such as average treatment, both be superimposed with superimposed voltage, need to calculate the 2nd candidate axis error Δ θ again _M2In this case, can be to γ shaft current i from coordinate converter 12 _γAnd δ shaft current i _δImplement high territory by handling, will remove superimposed current (i _{H γ}And i _{H δ}) the γ shaft current i of composition _γAnd δ shaft current i _δValue be used for the 2nd candidate axis error Δ θ _M2Calculating.

By forming estimator (200b, 200c or 200d) like this, under the rotation halted state or low speed rotation state of motor 1, carry out and use value (the Δ θ that becomes branch to calculate according to superimposed current _M1, ω _E1Or θ _E1) low speed with inferring processing, under the high speed rotating state of motor 1, carry out to use value (the Δ θ that becomes branch to calculate according to drive current _M2, ω _E2Or θ _E2) high speed with inferring processing.Therefore, can in big velocity interval, realize good no transducer control.

Suppose corresponding to rotating speed, the low speed of inferring the d-q axle before switching is controlled with no transducer, control with no transducer with the high speed of inferring the dm-qm axle, then need the switching between different coordinates, therefore in the switching that realizes smoothness, go wrong corresponding to the 2nd or the 3rd execution mode.In addition, under the situation that realizes breakdown torque control, need be under based on the control of d-q axle corresponding to the d shaft current (γ shaft current) of q shaft current (δ shaft current), relative therewith, under control based on the dm-qm axle, make dm shaft current (γ shaft current) be zero or be roughly zero, so γ shaft current command value i _γ ^*Be accompanied by to switch and become discontinuous.On the other hand, if as present embodiment, under control, switch, just solved the problems referred to above based on the dm-qm axle.

In addition, handle, slowly carry out low speed with inferring processing and, having guaranteed the continuity of the presumed value of rotor-position and electromotor velocity, thereby can infer the switching of processing glibly at a high speed with the switching of inferring processing by using weighted average.But, at the estimator 200b of image pattern 35 like this, under the situation about switching in the stage that error is inferred, because only in the answer speed that the characteristic by PLL sets, the presumed value of rotor-position and electromotor velocity changes, even therefore be weighted average treatment, also guaranteed the continuity of its presumed value.

" distortion etc. "

Illustrated item in each execution mode, short of contradiction just can be applied to other execution modes.For example, illustrated item (formula etc.) in the 2nd execution mode all can be applicable to the 3rd～the 6th execution mode.In addition, the concrete numerical value shown in the above-mentioned explanation only is example, certainly it is changed to various numerical value.

In above-mentioned the 1st～the 6th execution mode, to flux regulator portion 16 output zero or be roughly zero i _γ ^*Or i _Dm ^*Be illustrated, but carry out to export the i that has corresponding to the value of this rotating speed in the weak flux controlled rotating speed at needs _γ ^*Or i _Dm ^*

In addition, current detector 11 can adopt the formation that directly detects motor current as shown in Fig. 3 waits, and also can replace this method, reproduces motor current from the transient current of the DC electric current of mains side, by detecting motor current like this.

In addition, part or all of the function of the control device of electric motor in each execution mode for example uses the software (program) be installed in the general microcomputer etc. to realize.Use software to realize under the situation of control device of electric motor, showing the block diagram presentation function block diagram of each formation of control device of electric motor.Certainly, also can not pass through software (program), only constitute control device of electric motor by hardware.

[selecting of qm axle]

In addition,, the 2nd～the 6th execution mode is illustrated, but, can also obtains the desired torque control different with breakdown torque control by the utilization foregoing to have realized that breakdown torque control (or approximate control) is prerequisite.Certainly, also can access the effects such as facilitation of above-mentioned parameter adjustment this moment.

For example, in the 2nd～the 6th execution mode, should supply with the corresponding to rotating shaft of direction of the current phasor of motor 1 in the time of will controlling with the realization breakdown torque with direction and compare, the rotating shaft that phase place is more leading is adopted as the qm axle.By like this, can reduce iron loss, improve the efficient of motor.If suitably improved the phase place of qm axle, then can realize maximal efficiency control.

Under the situation that realizes breakdown torque control, L _mValue calculate by following formula (42), but by adopting the little value of the value calculated than following formula (42) as L _mValue, can improve the efficient of motor.

In the 1st execution mode (Fig. 3), from control device of electric motor 3, remove the part of estimator 20, constituted control part.In the 2nd execution mode (Figure 14), from control device of electric motor 3a, remove the part of estimator 40, constituted control part.In the 3rd execution mode (Figure 19), estimator 45, θ from control device of electric motor 3b, have been removed _mThe part of calculating part 46 and arithmetic unit 47 constitutes control part.

In the 4th execution mode (Figure 20), position detector 50, θ _mCalculating part 52 and arithmetic unit 53 constitute θ _DmCalculating part has been removed θ from control device of electric motor 3c _DmThe part of calculating part constitutes control part.

In the 5th and the 6th execution mode (Figure 24), from control device of electric motor 3d, removed the part of estimator 200, constituted control part with superimposed voltage generating unit 201.

Among the estimator 200b of relevant Figure 35 of the 6th execution mode, the 1st axis error infers portion 261 and the 2nd axis error is inferred portion 262, plays the function of the 1st candidate axis error calculating part and the 2nd candidate axis error calculating part respectively.Among the estimator 200c of relevant Figure 39 of the 6th execution mode, the 1st axis error is inferred the position that portion 261 and proportional integral arithmetic unit 266 are constituted, play the function of the 1st candidate speed calculation portion, the 2nd axis error is inferred portion 262 and the position that proportional integral arithmetic unit 267 is constituted, and plays the function of the 2nd candidate speed calculation portion.Among the estimator 200d of relevant Figure 40 of the 6th execution mode, the 1st axis error is inferred the position that portion 261, proportional integral arithmetic unit 266 and integrator 269 are constituted, play the function of the 1st candidate position calculating part, the 2nd axis error is inferred the position that portion 262, proportional integral arithmetic unit 267 and integrator 270 are constituted, and plays the function of the 2nd candidate position calculating part.

In each execution mode, coordinate converter 12 and 18, subtracter 13 and 14 and current control division 15 constitute the voltage instruction calculating part.Flux regulator portion 16, speed controlling portion 17 and subtracter 19 constitute the current-order calculating part.

In addition, in the present embodiment, simplify narration sometimes, only by mark (i _γDeng) record, show quantity of state corresponding to this mark etc.Also promptly, in this specification, " i for example _γ" and " γ shaft current i _γ" referred to identical.

The present invention is applicable to all electrical equipment that use motor.For example go for the electric automobile that the rotation by motor drives, and the compressor that uses in the air-conditioning etc. etc.

Claims (17)

1. a control device of electric motor is being made as the d axle with the axle parallel with the magnetic flux that permanent magnet produced that constitutes rotor, will be made as the γ axle corresponding to the axle of inferring in the control of d axle, will be made as under the situation of q axle than the axle of the leading 90 degree electric angles of d axle,

Possess: will as the computing parameter, infer the estimator of the rotor-position of above-mentioned motor corresponding to the value of the q axle inductance of the motor that salient pole is arranged; And, control the control part of above-mentioned motor according to the above-mentioned rotor-position of being inferred, it is characterized in that:

Above-mentioned estimator, the value with between the q axle inductance of the reality of above-mentioned motor and the actual d axle inductance is adopted as above-mentioned computing with after the value of parameter, carries out inferring of above-mentioned rotor-position, by producing deviation like this between d axle and γ axle.

2. control device of electric motor as claimed in claim 1 is characterized in that:

Above-mentioned control part is controlled above-mentioned motor, and the feasible γ axle composition of supplying with the motor current of above-mentioned motor remains near the set-point zero or zero.

3. control device of electric motor as claimed in claim 2 is characterized in that:

To be made as under the situation of δ axle than the axle of the leading 90 degree electric angles of above-mentioned γ axle,

Above-mentioned estimator is inferred corresponding to above-mentioned rotor-position, also infers the rotating speed of above-mentioned rotor;

Above-mentioned control part, has the current-order calculating part, it generates the γ axle composition of above-mentioned motor current and γ shaft current command value and the δ shaft current command value that δ axle composition should be followed the trail of, and makes the above-mentioned rotating speed of being inferred follow the trail of outside electromotor velocity command value of being given;

Above-mentioned current-order calculating part, no matter the value of above-mentioned δ shaft current command value how, all remains above-mentioned set-point with above-mentioned γ shaft current command value, by like this, irrelevant with the value of the δ axle composition of above-mentioned motor current, allow the γ axle composition of above-mentioned motor current keep above-mentioned set-point.

4. control device of electric motor as claimed in claim 1 is characterized in that:

At the q of the reality of establishing above-mentioned motor axle inductance and actual d axle inductance, be respectively L _qWith L _d, be under the situation of L as above-mentioned computing with the q axle inductance of parameter,

Above-mentioned estimator, use and satisfy:

L _d≤L＜(L _d+L _q)/2

L, carry out inferring of above-mentioned rotor-position.

5. a control device of electric motor carries out the control of motor, it is characterized in that:

The corresponding to rotating shaft of direction or the phase place rotating shaft more leading than this rotating shaft of the current phasor when direction is controlled with the realization breakdown torque are made as the qm axle, will be made as with the rotating shaft of this qm axle quadrature under the situation of dm axle,

With the motor current that is circulated in the above-mentioned motor, be decomposed into qm axle composition that is parallel to above-mentioned qm axle and the dm axle composition that is parallel to above-mentioned dm axle, carry out the control of above-mentioned motor.

6. control device of electric motor as claimed in claim 5 is characterized in that:

Estimator with the rotor-position of inferring above-mentioned motor, and the control part of controlling above-mentioned motor according to the above-mentioned rotor-position of being inferred;

The axle parallel with the magnetic flux that permanent magnet produced that constitutes rotor is being made as the d axle, will be made as the γ axle, will be made as under the situation of δ axle than the axle of the leading 90 degree electric angles of γ axle corresponding to the axle of inferring in the control of d axle,

Above-mentioned control part carries out the control of above-mentioned motor, makes above-mentioned γ axle and above-mentioned δ axle, follows the trail of above-mentioned dm axle and above-mentioned qm axle respectively.

7. control device of electric motor as claimed in claim 6 is characterized in that:

Above-mentioned control part is controlled above-mentioned motor, makes the γ axle composition of above-mentioned motor current remain near zero or zero set-point.

8. control device of electric motor as claimed in claim 6 is characterized in that:

Above-mentioned estimator uses the axis error between above-mentioned qm axle and the above-mentioned δ axle, infers above-mentioned rotor-position.

9. control device of electric motor as claimed in claim 6 is characterized in that:

To be made as under the situation of q axle than the axle of the leading 90 degree electric angles of above-mentioned d axle,

Above-mentioned estimator, the resolution of vectors of using the induced voltage on the q axle that is produced in above-mentioned motor is inferred above-mentioned rotor-position as the induced voltage vector on the qm axle under the situation of induced voltage vector on the qm axle and the induced voltage vector on the dm axle.

10. control device of electric motor as claimed in claim 6 is characterized in that:

Above-mentioned estimator, use with the resolution of vectors of the magnetic linkage on the d axle of above-mentioned motor as the flux linkage vector on the dm axle under the situation of the flux linkage vector on flux linkage vector on the qm axle and the dm axle, infer above-mentioned rotor-position.

11. control device of electric motor as claimed in claim 6 is characterized in that:

Above-mentioned control part has the above-mentioned rotor-position that uses above-mentioned estimator to infer, and the given fixed axis composition of above-mentioned motor current is transformed into the coordinate converter of γ axle composition and δ axle composition;

Above-mentioned estimator according to γ axle composition and the δ axle composition from the resulting above-mentioned motor current of above-mentioned coordinate converter, is inferred the qm axle composition and the dm axle composition of above-mentioned motor current;

Qm axle composition and the dm axle composition of use by inferring resulting above-mentioned motor current, and, infer above-mentioned rotor-position from the γ axle composition of the resulting above-mentioned motor current of above-mentioned coordinate converter and the error current between the δ axle composition.

12. control device of electric motor as claimed in claim 6 is characterized in that:

Also has the stack portion of the superimposed voltage different that superpose to the driving voltage that is used for driving above-mentioned motor with this driving voltage frequency;

Above-mentioned estimator can be carried out the superimposed current that circulates based on corresponding to the stack of above-mentioned superimposed voltage in above-mentioned motor, infer the 1st of above-mentioned rotor-position and infer processing.

13. control device of electric motor as claimed in claim 12 is characterized in that:

Above-mentioned estimator,

Can also carry out according to the drive current that contains in the above-mentioned motor current, infer the 2nd of above-mentioned rotor-position and infer processing corresponding to above-mentioned driving voltage;

Corresponding to the velocity information of rotating speed of the above-mentioned rotor of expression, infer the above-mentioned the 1st and to handle and the above-mentioned the 2nd to infer and switch the actual performed processing of inferring in the processing.

14. control device of electric motor as claimed in claim 13 is characterized in that:

Above-mentioned estimator, the above-mentioned the 1st infer handle with the above-mentioned the 2nd infer switch between the processing actual performed when inferring processing,

Corresponding to above-mentioned velocity information, or,, allow the actual performed processing of inferring handle transfer to inferring of the opposing party from a side the processing of inferring by having added both sides' the processing of inferring of inferring the result of inferring processing corresponding to from switching the elapsed time of beginning.

15. control device of electric motor as claimed in claim 12 is characterized in that:

Voltage vector track on the rotatable coordinate axis of above-mentioned superimposed voltage, it is the symmetric figure of benchmark that formation has with the d axle.

16. an electric motor drive system is characterized in that having:

Motor;

Drive the inverter of above-mentioned motor; And

Control the control device of electric motor as claimed in claim 1 of above-mentioned motor by controlling above-mentioned inverter.

17. an electric motor drive system is characterized in that having:

Motor;

Drive the inverter of above-mentioned motor; And

Control the control device of electric motor as claimed in claim 5 of above-mentioned motor by controlling above-mentioned inverter.

Images