Motor control device

Abstract

A motor control device includes an electric power converter for generating three-phase voltages based on voltage instruction-values, and for supplying them to a motor; an induced-voltage estimation unit for estimating an induced voltage of the motor, based on the voltage instruction-values and electric current feedback values; a dead-time disturbance-voltage compensation unit; and a dead-time disturbance compensation-voltage modification unit for compensating a dead-time disturbance voltage(s) generated in the electric power converter, based on a fundamental wave component of an estimated induced voltage.

Claims

1. A motor control device, comprising: a rotational position detector for detecting a rotational position of a motor; an electric current detection unit for detecting at least two phase currents among three phase currents applied to the motor; a coordinate conversion unit for mutually converting values of said three phase currents on predetermined coordinates, and values on d-q coordinates being coordinates on a rotating magnetic field of the motor; an electric current control unit for generating voltage instruction-values being control inputs of voltages, based on a difference between electric current instruction-values being control inputs for said three phase currents, and electric current feedback values being currents detected by the electric current detection unit and d-q coordinate-converted by the coordinate conversion unit; an electric power converter for generating, based on said voltage instruction-values, three-phase voltages being inputted through the electric current detection unit so that predetermined electric power is supplied to the motor; an induced-voltage estimation unit for estimating, based on said voltage instruction-values and said electric current feedback values, an induced voltage of the motor; a dead-time disturbance-voltage compensation unit for generating a dead-time disturbance compensation-voltage being a voltage to compensate disturbance, with respect to an output produced at the electric current detection unit, due to dead time which is a short-circuit prevention period of two switching devices provided in the electric power converter; and a dead-time disturbance compensation-voltage modification unit for modifying said dead-time disturbance compensation-voltage in accordance with said electric current instruction-values, and also having the induced-voltage estimation unit, and, based on a fundamental wave component of an induced voltage being estimated, a dead-time disturbance compensation-voltage memory unit for calculating an error of said dead-time disturbance compensation-voltage and for storing the error inside, wherein the motor control device comprises two operation modes of a dead-time disturbance compensation-voltage adjustment mode for setting said electric current instruction-values in accordance with an operation pattern of the motor in order to acquire said dead-time disturbance compensation-voltage, and a torque-ripple suppression control mode for estimating torque of the motor, and for curbing a torque ripple thereof based on the torque being estimated; and, after operations in the dead-time disturbance compensation-voltage adjustment mode are completed, the operations transfer to the torque-ripple suppression control mode.

2. The motor control device as set forth in claim 1, further comprising a torque-ripple compensation unit including the induced-voltage estimation unit, and a torque-ripple compensation instruction generation unit for curbing, in accordance with an induced voltage estimated in the induced-voltage estimation unit, a torque ripple in which the motor produces.

3. The motor control device as set forth in claim 2, wherein, in the dead-time disturbance compensation-voltage adjustment mode, the torque-ripple compensation instruction generation unit is in a resting state, and the dead-time disturbance compensation-voltage modification unit operates in a state in which an operation pattern of the motor is inputted so as to drive the motor in accordance with the operation pattern, so that, in the dead-time disturbance compensation-voltage memory unit, a dead-time disturbance compensation-voltage compensation map corresponding to said electric current instruction-values is made based on an error of a dead-time disturbance compensation-voltage being calculated, and stored; and, in the torque-ripple suppression control mode, a dead-time disturbance compensation-voltage amplitude error corresponding to said electric current instruction-values stored in the dead-time disturbance compensation-voltage adjustment mode is outputted from the dead-time disturbance compensation-voltage memory unit, and inputted into the dead-time disturbance-voltage compensation unit, so that, in a state in which a dead-time disturbance voltage is compensated in accordance with a dead-time disturbance compensation-voltage being modified, the torque-ripple compensation instruction generation unit operates, and performs a torque-ripple suppression control.

4. The motor control device as set forth in claim 1, wherein, in an operation pattern of the motor, speed of not at zero, but of a constant value or more is given to the motor, while a q-axis electric current instruction-value thereof being set at zero to undergo non-energization, and a plurality of electric current instruction-values is set as a d-axis electric current instruction-value.

5. The motor control device as set forth in claim 1, wherein, in an operation pattern of the motor, speed of not at zero, but of a constant value or more is given to the motor, while a d-axis electric current instruction-value thereof being set at zero to undergo non-energization, and a plurality of electric current instruction-values is set as a q-axis electric current instruction-value.

6. The motor control device as set forth in claim 1, wherein a calculator of said dead-time disturbance compensation-voltage calculates a dead-time disturbance-voltage compensation value as an amplitude value for three-phase voltages.

7. The motor control device as set forth in claim 1, wherein, in the dead-time disturbance compensation-voltage adjustment mode, a dead-time disturbance compensation-voltage amplitude error ΔV.sub.Td is calculated by a first equation (5) $\begin{array}{cc}\Delta V_{\mathrm{Td}}=\frac{e_q\mathrm{\_bar}}{2\sqrt{\frac{2}{3}}\sin \left(\frac{5\pi }{12}\right)}& \left(5\right)\end{array}$ when only a d-axis current undergoes energization, where e.sub.d.sub._bar designates a fundamental wave component of a d-axis induced-voltage estimate value e.sub.d of the motor, and, by a second equation (13) $\begin{array}{cc}\Delta V_{\mathrm{Td}}=\frac{e_q\mathrm{\_bar}-{\omega }_{\mathrm{re}}{\varphi }_f}{2\sqrt{\frac{2}{3}}\sin \left(\frac{5\pi }{12}\right)}& \left(13\right)\end{array}$ when only a q-axis current undergoes energization, where e.sub.q.sub._bar designates a fundamental wave component of a q-axis induced-voltage estimate value e.sub.q of the motor; φ.sub.f, magnet's magnetic flux; and ω.sub.re, electrical angular velocity.

8. The motor control device as set forth in claim 1, wherein the induced-voltage estimation unit changes inductance for the motor or a induced voltage constant, a torque constant, therefor depending on said electric current instruction-values.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) The foregoing and other objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.

(2) FIG. 1 is a diagram for explaining a general outline of dead-time disturbance voltages;

(3) FIG. 2 is a block diagram illustrating an example of a configuration of a first operation mode according to a motor control device of the present invention;

(4) FIG. 3 is a block diagram illustrating another example of a configuration of a first operation mode according to a motor control device of the present invention;

(5) FIG. 4 is a block diagram illustrating an example of a configuration of a second operation mode according to a motor control device of the present invention;

(6) FIG. 5 is a diagram showing an example of a map of dead-time disturbance compensation-voltage amplitude errors stored in a memory means in a first operation mode according to a motor control device of the present invention;

(7) FIG. 6 is a diagram for explaining an operation principle of the memory means in a first operation mode according to a motor control device of the present invention;

(8) FIG. 7 is a diagram showing an example of characteristics of an inductance L that is a motor constant, with respect to an instruction-current amplitude value; and

(9) FIG. 8 is a diagram showing an example of characteristics of a torque constant K.sub.t that is a motor constant, with respect to an instruction-current amplitude value.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

(10) Hereunder, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

Embodiment 1

(11) FIG. 2, FIG. 3 and FIG. 4 are block diagrams each illustrating an example of Embodiment 1 according to the present invention. This embodiment includes two operation modes: a dead-time disturbance compensation-voltage adjustment mode illustrated in FIG. 2 and FIG. 3 as a first operation mode, and a mechanical state-value estimation mode illustrated in FIG. 4 as a second operation mode. And then, a sequence is included in which, first the operations are performed as a dead-time disturbance compensation-voltage adjustment mode, and subsequently, the operations transfer to the mechanical state-value estimation mode.

(12) As the mechanical state-value estimation mode, the exemplary explanation will be made for a device which includes a torque-ripple suppression control mode in which motor's torque is estimated, and a torque ripple is curbed based on the estimated torque.

(13) In FIG. 2 and FIG. 3, the motor control device has an electric current control unit 1, a d-q/three-phase conversion unit 2, a three-phase/d-q conversion unit 3, an electric power converter 4, an electric current detection unit 5, subtracters 6 and 7, a rotational position detector 8, and a dead-time disturbance compensation-voltage modification unit 100. And then, the motor control device controls a PM motor (hereinafter simply referred to as a “motor”) 9 as an alternating-current rotating machine by means of the electric power converter 4.

(14) In addition, into a dead-time disturbance-voltage compensation unit 11, a three-phase output current [i] detected by means of the electric current detection unit 5 is inputted, so that, based on the three-phase output current [i], a dead-time disturbance compensation-voltage [v*.sub.Td] of a three-phase rectangular wave expressed by Equation (2) is generated and outputted, being inputted into an adder 12.

(15) In this embodiment, a motor's torque control device (which can be, in other words, referred to as an electric current control device in this embodiment) is constituted of the electric current control unit 1, the d-q/three-phase conversion unit 2, the three-phase/d-q conversion unit 3, the electric power converter 4, the electric current detection unit 5, the subtracter 6 and 7, the rotational position detector 8, the motor 9, the dead-time disturbance-voltage compensation unit 11 and the adder 12; from these constituent items, a torque control (which can be, in other words, referred to as an electric current control in this embodiment) is performed based on respective difference between a q-axis electric current instruction-value i.sub.q* and a q-axis electric current detected-value i.sub.q, and that between a d-axis electric current instruction-value i.sub.d* and a d-axis electric current detected-value i.sub.d; and thus, the motor 9 is driven so that the electric currents equal to the i.sub.q* and i.sub.d* flow through the motor, which produces its torque equal to a torque instruction-value t* (for the details of the operations of torque control, or those of electric current control, refer to Sugimoto, Koyama, and Tamai [Authors], “Actualities of Theory and Design of AC Servo Systems,” 1990, pp. 54-57 and pp. 72-85, [then] Sogo Denshi Shuppan-sha, Tokyo, for example).

(16) Moreover, the relationship between the torque instruction-value τ* and the q-axis electric current instruction-value i.sub.q* is given as the following Equation (3).

(17) $\begin{array}{cc}{\tau }^{*}=\frac{i_q^{*}}{K_t}=\frac{i_q^{*}}{P_m{\varphi }_f}& \left(3\right)\end{array}$
where K.sub.t designates a torque constant of the motor; P.sub.m, the number of pole pairs; and φ.sub.f, magnet's magnetic flux.

(18) First, the explanation will be made for the operations of the dead-time disturbance compensation-voltage adjustment mode that is a first operation mode. In the motor control device of the present invention, an operation pattern for acquiring dead-time disturbance compensation-voltage is set to have three features (a), (b) and (c), as follows:

(19) (a) To the motor 9, its speed is given not at 0, but at a predetermined rotational velocity ω.sub.td or more, so that it is rotating;

(20) (b) the q-axis electric current instruction-value i.sub.q* is set at 0; and

(21) (c) a plurality of values are set as d-axis electric current instruction-values i.sub.d*, and, in each of i.sub.d*, the operations of this operation mode are performed through repetition thereof. In this operation mode, each of the instruction values is set in accordance with the operation pattern for acquiring dead-time disturbance compensation-voltage, so that the motor 9 is subjected to be driven. In these cases, the explanation will be made for the operations of the dead-time disturbance compensation-voltage modification unit 100.

(22) In an induced-voltage estimation unit 10, an induced-voltage estimate value vector [e.sub.dq] as a motor's estimated induced voltage is estimated according to the calculation of the following Equation (4), based on a motor constant, a vector [i.sub.dq] made of a d-axis electric current detected-value i.sub.d and a q-axis electric current detected-value i.sub.q, a vector [v*.sub.dq] made of voltage instruction-values v.sub.d* and v.sub.q* to the motor 9, and a motor's electrical angle θ.sub.re detected by the rotational position detector 8 such as an encoder or the like.

(23) $\begin{array}{cc}\begin{array}{c}\frac{d{\theta }_{\mathrm{re}}}{dt}={\omega }_{\mathrm{re}}=P_m{\omega }_{\mathrm{rm}}\\ \left[e_{\mathrm{dq}}\right]=\left[v_{\mathrm{dq}}^{*}\right]-\left\{\left(R+\mathrm{sL}\right)I+{\omega }_{\mathrm{re}}\mathrm{LJ}\right\}\left[i_{\mathrm{dq}}\right]\end{array}\}& \left(4\right)\end{array}$

(24) Here, symbol R designates resistance of motor's winding; L, self inductance thereof; P.sub.m, the number of pole pairs; s, differential operator; I, unit matrix; J, skew symmetric matrix; ω.sub.rm, mechanical angular velocity; and ω.sub.re, electrical angular velocity.

(25) And then, the induced-voltage estimate value vector [e.sub.dq] is inputted into a dead-time disturbance compensation-voltage memory unit 101. Next, the explanation will be made for the operations of the dead-time disturbance compensation-voltage memory unit 101. As for a fundamental wave component e.sub.d bar of a d-axis induced-voltage estimate value e.sub.d in the induced-voltage estimate value vector [e.sub.dq] a dead-time disturbance compensation-voltage amplitude error ΔV.sub.Td is given according to the calculation of the following Equation (5).

(26) $\begin{array}{cc}\Delta V_{\mathrm{Td}}=\frac{e_d\mathrm{\_bar}}{2\sqrt{\frac{2}{3}}\sin \left(\frac{5\pi }{12}\right)}& \left(5\right)\end{array}$

(27) The operations described above are repeated in each of the cases for a plurality of i.sub.d*. And then, by using an arbitrary approximation method of a linear interpolation, for example, or the like, a map of ΔV.sub.Td is made with respect to a motor's electric current amplitude instruction-value i* made of i.sub.d* and i.sub.q* (refer to Equation (6) below) as shown in FIG. 5, and is stored in the dead-time disturbance compensation-voltage memory unit 101.
i*=√{square root over ((id*).sup.2+(iq*).sup.2)}  (6)
In the second operation mode as will be described later, the aforementioned i* is calculated by Equation (6), in a case in which values of i.sub.d* and i.sub.q* are individually set and the motor control is performed, and a value corresponding to the calculated one is called from the map. And then, as illustrated in FIG. 3, the map's value stored in this memory unit is inputted into the dead-time disturbance-voltage compensation unit 11, which is so arranged that a new voltage instruction-value [v*.sub.Td] with respect to the i* can be determined, so that this operation mode ends its operations, and the operations transfer to those in the second operation mode.

(28) Hereinafter, the explanation will be made for a principle of Equation (5).

(29) Now, in the dead-time disturbance-voltage compensation unit 11, the dead-time disturbance-voltage compensation based on Equation (2) is performed; however, as already described above, a dead-time disturbance compensation-voltage amplitude V.sub.Td used in Equation (2) has an error ΔV.sub.Td originated from a delay(s) of switching devices or the like of the electric power converter, and thus, it is necessary to modify the V.sub.Td by the quantity of the error.

(30) Here, because the term [i.sub.dq] in Equation (4) contains a drive current [i*.sub.M] based on a voltage instruction-value, and a disturbance current [i*.sub.Td] generated due to the dead-time disturbance voltage, it can be said that an induced-voltage estimate value vector [e.sub.dq] contains an error voltage(s) due to the [i*.sub.Td].

(31) By the way, when attention is only paid to a fundamental wave component of induced voltage, an induced voltage on the d-axis becomes “0,” so that, by subtracting a fundamental wave component of induced voltage on the d-axis, “0,” from a fundamental wave component e.sub.d bar of a d-axis induced-voltage estimate value e.sub.d, the estimation error of a fundamental wave component of the induced-voltage estimation value due to the [i*.sub.Td] can be obtained.

(32) Meanwhile, in Embodiment 1, the current control is performed in such a manner that d-axis electric current instruction-values i.sub.d* are set at arbitrary constant values, and the q-axis electric current instruction-value i.sub.q* is set at “0.” In these cases, consideration will be given to an influence of a dead-time disturbance voltage appearing on the d-q axes. This is because, due to a three-phase/d-q conversion of [v.sub.Td] in Equation (1), it can be understood that, as shown in FIG. 6, disturbance in the shape of half-waves (refer to the curve indicated by the symbol “S” in the figure) appears on the energization axis, whereas disturbance in the shape of sawtooth waveforms (refer to the line indicated by the symbol “T” in the figure) appears on the non-energization axis.

(33) Here, FIG. 6 is a graph showing an influence of ΔV.sub.Td viewed on the d-q axes in a case in which ΔV.sub.Td=V.sub.Td; the vertical axis indicates a value of ΔV.sub.Td/V.sub.Td shown by the unit of p.u. which is normalized power, and the horizontal axis, an electrical angle-phase shown by the unit of radian (rad). The curve indicated by the symbol “S” in the figure shows an error voltage (in the shape of half-waves) on the energization axis, and the broken line indicated by the symbol “Sa” shows a mean value (unit: p.u.) of this error voltage. In addition, the line indicated by the symbol “T” shows an error voltage at the time of non-energization, and is the shape of sawtooth waveforms.

(34) Accordingly, from FIG. 6, it can be understood that, when only an error of fundamental wave component (refer to the arrow “A” in the figure) is considered, the influence appears only on the axis undergoing energization. When direct-current values of the error voltage in the shape of half-waves in FIG. 6 is calculated in a simple manner, the value of half-wave bosses in FIG. 6 is given by Equation (7), and the value of recesses therein is given by Equation (8) as shown below.
2√{square root over (⅔)} sin(π/2)  (7)
2√{square root over (⅔)} sin(π/3)  (8)
Accordingly, by taking on a value between π/2 and π/3 in the electrical angle-phases (taking on the arithmetic mean), (π/2+π/3)/2=5π/12; and therefore, an acquired current value is given as follows:
2√{square root over (⅔)} sin(5π/12)  (9)
(refer to the straight line “Sa” in the figure that is a mean value of the error voltage).

(35) Therefore, the following Equation (10) is held between the error voltage paid attention only to a fundamental wave component of an estimated induced voltage and a dead-time disturbance compensation-voltage amplitude error ΔV.sub.Td as a three-phase rectangular wave.

(36) $\begin{array}{cc}2\sqrt{\frac{2}{3}}\sin \left(\frac{5\pi }{12}\right)\times \Delta V_{\mathrm{Td}}=\left(e_d\mathrm{\_bar}-0\right)& \left(10\right)\end{array}$
From this Equation (10), Equation (5) is derived.

(37) Next, the explanation will be made for the operations of the torque-ripple suppression control mode that is the second operation mode.

(38) This mode is one of application examples on technologies for estimating a mechanical state-value; the mode allows curbing the torque ripple based on an estimation value of motor's torque.

(39) In FIG. 4, the motor control device newly includes an adder 13 and a torque-ripple compensation unit 200 in comparison with the first operation mode, and the torque-ripple compensation unit 200 includes the induced-voltage estimation unit 10 and a torque-ripple compensation instruction generation unit 201.

(40) In this operation mode, d-axis and q-axis electric current instruction-values i.sub.d* and i.sub.q* are inputted into the dead-time disturbance compensation-voltage memory unit 101, and ΔV.sub.Td corresponding to those is inputted into the dead-time disturbance-voltage compensation unit 11. And then, in the dead-time disturbance-voltage compensation unit 11, a dead-time disturbance compensation-voltage v.sup.*.sub.Td in which a dead-time disturbance compensation-voltage amplitude V.sub.Td is modified using the following Equation (11) in place of Equation (2) is generated and outputted into the adder 12, so that the dead-time disturbance compensation is performed.
[v.sub.Td*]=sgn[i]×(V.sub.Td+ΔV.sub.Td)  (11)

(41) Next, the explanation will be made for the operations of a torque-ripple suppression control in this operation mode, namely, the operations of the torque-ripple compensation unit 200.

(42) In the induced-voltage estimation unit 10, similarly to the first operation mode, an induced-voltage estimate value vector [e.sub.dq] is calculated, and is inputted into the torque-ripple compensation instruction generation unit 201. In the torque-ripple compensation instruction generation unit 201, motor's torque τ is estimated according to the following Equation (12), based on the induced-voltage estimate value vector [e.sub.dq] and a d-axis and q-axis detection-current vector [i.sub.dq].

(43) $\begin{array}{cc}\tau =\frac{\left[i_{\mathrm{dq}}^T\right]\left[e_{\mathrm{dq}}\right]}{{\omega }_{\mathrm{rm}}}& \left(12\right)\end{array}$

(44) Here, [i.sup.T.sub.dq] designates transpose of the vector [i.sub.dq].

(45) And then, a vibration component contained in a torque estimation value τ is extracted; a torque-ripple compensation signal i.sub.q*.sub.rip is calculated so as to cancel out the vibration, and is inputted into the adder 13. Note that, there are many publicly known technologies in relation to a generation method of the torque-ripple compensation signal i.sub.q*.sub.rip based on this torque estimation value τ; however, in the present invention, it is possible to use an arbitrary technology publicly known.

(46) The adder 13 adds the torque-ripple compensation signal i.sub.q*.sub.rip to a q-axis electric current instruction-value i.sub.q*, and the value is inputted into the subtracter 6 as a new q-axis electric current instruction-value. According to this arrangement, it can be said that, in the motor 9, a pulsation torque based on the torque-ripple compensation signal i.sub.q*.sub.rip is produced, in addition to the torque due to the q-axis electric current instruction-value i.sub.q* (namely, torque instruction τ*), in a reverse phase to a torque ripple, and it can be said that the torque ripple is curbed because the pulsation torque and the torque ripple are canceled out with each other.

(47) The induced-voltage estimate value vector [e.sub.dq] and estimation torque calculated from Equation (4) based on the induced-voltage estimate value vector contain vibration components synchronized with a torque ripple; however, a 6f-component being a principal of them results in the superposition of a component originated from a dead-time disturbance voltage and a component where the motor itself generates, so that it is difficult to extract only the component(s) that the motor itself generates.

(48) However, in this embodiment, the dead-time disturbance compensation is put to work using a dead-time disturbance-voltage compensation value stored at the time of the operations in the first operation mode, so that the 6f-component originated from the dead-time disturbance voltage is canceled out, and thus, it becomes possible to only extract the 6f-component where the motor itself generates.

(49) In addition, when a dead-time disturbance compensation-voltage amplitude error ΔV.sub.Td is to be acquired in the first operation mode, the operations can be performed without producing torque by the motor 9, because the d-axis only undergoes energization.

(50) In addition, in the second operation mode, the induced-voltage estimation unit 10 is not engaged in the dead-time disturbance-voltage compensation operations, and thus, the dead-time disturbance compensation operations do not necessarily interfere with the torque-ripple suppression control.

(51) Moreover, as shown for example in FIG. 7, when dependency characteristics of inductance “L” being a motor constant with respect to an electric current (refer to the solid line indicated by the symbol “Lv” in the figure; the broken line “Lc” indicates the characteristics where inductance “L” has no dependence on the electric current) are known, it is also possible to achieve that, in conjunction with setting i.sub.d* and i.sub.q* by an operation pattern for acquiring dead-time disturbance compensation-voltage, and also setting a value of inductance “L” with respect to an electric current amplitude instruction-value i* made of i.sub.d* and i.sub.q* to that of inductance “Lv” in the induced-voltage estimation unit 10, the performance is enhanced without allowing the influence of disturbance originated from motor's parameter variation.

(52) As described above, in this embodiment, by executing each of the sequences reaching from the first operation mode to the second operation mode one after another, an effect can be achieved that, based on observations of electrical state-values, the dead-time disturbance compensation can be easily performed, and there exists an effect in addition that, based on the observations of electrical state-values, the dead-time disturbance compensation and estimation of a mechanical state-value can be executed without interfering with each other.

Embodiment 2

(53) This embodiment includes two operation modes, similarly to Embodiment 1: the dead-time disturbance compensation-voltage adjustment mode illustrated in FIG. 2 and FIG. 3 as a first operation mode, and a mechanical state-value estimation mode illustrated in FIG. 4 as a second operation mode. And then, a sequence is included in which, first the operations are performed as a dead-time disturbance compensation-quantity memory mode, and subsequently, the operations transfer to the mechanical state-value estimation mode.

(54) As the mechanical state-value estimation mode, the exemplary explanation will be made for a device which includes a torque-ripple suppression control mode in which motor's torque is estimated, and a torque ripple is curbed based on the estimated torque.

(55) In Embodiment 2, in the dead-time disturbance compensation-voltage adjustment mode, an operation pattern for acquiring dead-time disturbance compensation-voltage is set to have three features (d), (e) and (f) different from those in Embodiment 1, as follows:

(56) (d) To the motor 9, its speed is given not at 0, but at a rotational velocity ω.sub.td or more, so that it is rotating;

(57) (e) the d-axis electric current instruction-value i.sub.d* is set at 0; and

(58) (f) a plurality of values are set as q-axis electric current instruction-values i.sub.q*, and, in each of i.sub.q*, the operations of this operation mode are performed through repetition thereof.

(59) As described above, it can be said that the operations are performed mainly based on the energization on the q-axis. In the induced-voltage estimation unit 10, an induced-voltage estimate value vector [e.sub.dq] is calculated in accordance with Equation (4), and is inputted into the dead-time disturbance compensation-voltage memory unit 101.

(60) Here, in the dead-time disturbance compensation-voltage memory unit 101, the calculation of the following Equation (13) is performed, in place of Equation (5), for a fundamental wave component e.sub.q.sub._bar of a q-axis induced-voltage estimate value e.sub.q in the [e.sub.dq].

(61) $\begin{array}{cc}\Delta V_{\mathrm{Td}}=\frac{e_q\mathrm{\_bar}-{\omega }_{\mathrm{re}}{\varphi }_f}{2\sqrt{\frac{2}{3}}\sin \left(\frac{5\pi }{12}\right)}& \left(13\right)\end{array}$

(62) The operations described above are repeated in each of the cases for a plurality of i.sub.q*, and a dead-time disturbance compensation-voltage amplitude error ΔV.sub.Td is calculated by Equation (13) in all the cases. And then, by using an arbitrary approximation method of a linear interpolation or the like, a map of ΔV.sub.Td is made with respect to an electric current amplitude instruction-value i* made of i.sub.d* and i.sub.q*, and is stored in the memory unit, so that this operation mode ends its operations, and the operations transfer to those in the second operation mode.

(63) Here, as for a principle of Equation (13), it is basically similar to the case of Equation (5). Namely, because the current control is performed in such a manner that q-axis electric current instruction-values i.sub.q* are set at arbitrary constant values, and the d-axis electric current instruction-value i.sub.d* is set at “0,” it can be said that an estimation error of a fundamental wave component of an induced-voltage estimation value only appears on a q-axis component in the [e.sub.dq].

(64) Meanwhile, when attention is only paid to a fundamental wave component of induced voltage, an induced voltage on the q-axis is given as ω.sub.reφ.sub.f from magnet's magnetic flux φ.sub.f and an electrical angular velocity ω.sub.re, so that, by subtracting a fundamental wave component (ω.sub.reφ.sub.f of the induced voltage on the q-axis from a fundamental wave component e.sub.q.sub._bar of a q-axis induced-voltage estimate value e.sub.q, the estimation error of a fundamental wave component of the induced-voltage estimation value can be obtained.

(65) Therefore, in this embodiment, the following equation is held between the error voltage paid attention only to a fundamental wave component of an estimated induced voltage and a dead-time disturbance compensation-voltage amplitude error ΔV.sub.Td as a three-phase rectangular wave.

(66) $\begin{array}{cc}2\sqrt{\frac{2}{3}}\sin \left(\frac{5\pi }{12}\right)\times \Delta V_{\mathrm{Td}}=e_q\mathrm{\_bar}-{\omega }_{\mathrm{re}}{\varphi }_f& \left(14\right)\end{array}$
From Equation (14), Equation (13) is derived. In addition, in regard to the torque-ripple suppression control mode that is the second operation mode, the operations equivalent or similar to those in Embodiment 1 are performed.

(67) Moreover, as shown for example in FIG. 7, when known are dependency characteristics of inductance “L” being a motor constant with respect to an electric current (refer to the solid line indicated by the symbol “Lv” in the figure), it is also possible to achieve that, in conjunction with setting i.sub.d* and i.sub.q* by the operation pattern for acquiring dead-time disturbance compensation-voltage, and also setting a value of inductance with respect to an electric current amplitude instruction-value i*, made of i.sub.d* and i.sub.q*, to that of inductance “Lv” in the induced-voltage estimation unit 10 and the dead-time disturbance compensation-voltage memory unit 101, the performance is enhanced without allowing the influence of disturbance originated from motor's parameter variation.

(68) While on the other hand, with respect to an electric current amplitude instruction-value i* made of i.sub.d* and i.sub.q*, and with respect to a motor having a property inwhich its torque becomes saturated, when known are saturation characteristics of the torque with respect to the electric current (refer to the solid line indicated by the symbol “Ktv” in FIG. 8; the broken line “Ktc” indicates the characteristics where a torque constant “K.sub.t” has no dependence on the electric current), it is possible to show the saturation characteristics, in the form shown in FIG. 8, as a saturation curve Ktv of the torque constant K.sub.t with respect to the electric current amplitude instruction-value i*. Because magnet's magnetic flux φ.sub.f=Kt/P.sub.m, as made known in Equation (3), it is also possible to achieve that, by using for the φ.sub.f in Equation (13) a φ.sub.f with respect to an individual electric current amplitude instruction-value i* obtained from the saturation curve Ktv of the torque constant K.sub.t shown in FIG. 8 as φ.sub.f=K.sub.t/P.sub.m, the performance is enhanced without allowing the influence of disturbance originated from motor's parameter variation.

(69) As described above, in this embodiment, by executing each of the sequences reaching from the first operation mode to the second operation mode one after another similarly to Embodiment 1, an effect can be achieved that, based on observations of electrical state-values, the dead-time disturbance compensation can be easily performed, and there exists an effect in addition that, based on the observations of electrical state-values, the dead-time disturbance compensation and estimation of a mechanical state-value can be executed without interfering with each other. Note that, in the present invention, each of the embodiments can be freely combined, and/or each of the embodiments can be appropriately modified or eliminated without departing from the scope of the invention.

EXPLANATION OF NUMERALS AND SYMBOLS

(70) Numeral “1” designates an electric current control unit; “2,” d-q/three-phase conversion unit; “3,” three-phase/d-q conversion unit; “4,” electric power converter; “5,” electric current detection unit; “6,” “7,” subtracter; “8,” rotational position detector; “9,” PM motor (motor); “10,” induced-voltage estimation unit; “11,” dead-time disturbance-voltage compensation unit; “12,” “13,” adder; “100,” dead-time disturbance compensation-voltage modification unit; “101,” dead-time disturbance compensation-voltage memory unit; “200,” torque-ripple compensation unit; and “201,” torque-ripple compensation instruction generation unit.