JP5120621B2 - Control device for permanent magnet type synchronous motor - Google Patents

Description

  In particular, the present invention relates to a control device in which, in a permanent magnet type synchronous motor in which the rotor does not have saliency, the magnetic pole position of the rotor can be accurately calculated at the time of start-up and can be smoothly operated.

In order to reduce the cost of a control device for a permanent magnet type synchronous motor (hereinafter also referred to as PMSM), a so-called sensorless control technique that operates without using a magnetic pole position detector has been put into practical use.
Many of this kind of sensorless control techniques calculate the magnetic pole position using an induced voltage induced between the terminals of the motor by the permanent magnet of the rotor, and are applied to the operation in the medium to high speed region. However, in this method, it is difficult to obtain a sufficient induced voltage when the motor is stopped or when the motor is operated at a low speed. In this case, another magnetic pole position calculation technique is used.

Usually, the magnetic pole position calculation technique applied during low-speed operation differs depending on the electrical characteristics of the PMSM.
Here, the PMSM is divided into two types according to the structure of the rotor: a surface magnet structure permanent magnet synchronous motor (hereinafter also referred to as SPMSM) and an embedded magnet structure permanent magnet synchronous motor (hereinafter also referred to as IPMSM). Can be separated. Among these, as a magnetic pole position calculation technique of IPMSM, a method using saliency based on unbalance of the magnetic circuit of the rotor has been put into practical use.
On the other hand, as a magnetic pole position calculation technique for SPMSM and IPMSM having a small rotor saliency, a method using the magnetic saturation characteristic of the rotor is known.

  As a prior art using this magnetic saturation characteristic, for example, in Patent Document 1, a pulse voltage is applied in a plurality of directions to a PMSM during stoppage or extremely low speed operation, and a magnetic pole is determined from a difference in current before and after the pulse voltage is applied. In the method of estimating the magnetic pole position by measuring the change in inductance depending on the position, a method of calculating the magnetic pole position using the Fourier coefficient obtained from the current difference vector is described.

  In Patent Document 2, the electrical angle is changed at a constant angle in a stationary state of PMSM, the response time of the step response of the current is measured at each electrical angle, and the electrical angle-response time characteristic is approximated. Permanent magnet synchronous motor for calculating the coefficient of the Fourier series fundamental cosine wave term and the fundamental wave sine wave term and estimating the magnetic pole position by adding or subtracting 180 degrees to the electrical angle obtained from these coefficient values A control device is described.

Further, Patent Document 3 gives a minute change to the voltage command values of the dc axis and qc axis, which are estimated magnetic flux axes of PMSM, and the current change rate of the dc axis and qc axis at this time, or the inductance that is the reciprocal thereof. A method for calculating the magnetic pole position from the change is described.
According to this method, the magnetic pole position can be calculated in a short time by a relatively simple process. Furthermore, it is also possible to calculate the frequency command value so that the position calculation error obtained from the current change rates of the dc axis and the qc axis becomes zero, and to calculate the magnetic pole position by integrating the frequency command value.

Japanese Patent No. 3687603 (Claims 1 and 2, paragraphs [0011] to [0032], [0056] to [0078], FIG. 1, etc.) JP-A-11-262286 (Claim 4, paragraphs [0039] to [0051], FIG. 4 etc.) JP 2002-78392 A (paragraphs [0079] to [0082], [0100] to [0112], [0120] to [0121], FIGS. 16 to 19, FIGS. 23 to 25, FIG. 28, etc.)

In the prior art described in Patent Document 1, the rotor needs to be stationary while a plurality of pulse voltages are applied for magnetic pole position calculation. Therefore, it cannot be applied when the rotor is rotating.
Similarly, the prior art described in Patent Document 2 is also based on the principle of measuring the current step response time while the rotor is stationary, and is therefore applicable when the rotor is rotating. Impossible.

  On the other hand, according to the prior art described in Patent Document 3, the magnetic pole position can be calculated even when the rotor is rotating, but the inductance that can be measured using the current change rate is limited to four points. Therefore, the change in inductance depending on the magnetic pole position cannot be detected accurately, and there is a problem in the calculation accuracy of the magnetic pole position.

  When the rotor is rotating even a little, an induced voltage is induced between the terminals of the electric motor by the permanent magnet of the rotor. However, in the prior arts of Patent Document 1 and Patent Document 3, the influence of the current that flows due to the induced voltage is not considered in the magnetic pole position calculation, and this current is not controlled. For this reason, when the rotor is rotating, the magnetic saturation characteristics change, or the inductance change cannot be measured accurately from the current difference or current change rate, and an error occurs in the magnetic pole position calculation. Furthermore, when the rotor rotates at a high speed, the current due to the induced voltage becomes excessive, and the electric motor and the power converter may be damaged.

  Therefore, a problem to be solved by the present invention is to provide a control device that can calculate the magnetic pole position with high accuracy even when the rotor is rotating and that does not possibly damage the device.

In order to solve the above-mentioned problem, the invention according to claim 1 applies a rectangular wave alternating high-frequency voltage composed of two pulse voltages having the same amplitude and pulse width and different polarities to the PMSM in a plurality of vector directions. Then, a high-frequency current in a direction parallel to the alternating high-frequency voltage is detected. Then, a Fourier series primary component that is a component that changes in one cycle depending on the vector direction of the alternating high-frequency voltage is extracted from this high-frequency current, and a magnetic pole position or magnetic pole position calculation error is calculated using this primary component. .
According to the present invention, the magnetic pole position can be accurately calculated even when the fundamental current flows in the armature winding by the induced voltage induced between the terminals of the motor by the permanent magnet of the rotating rotor. it can. In addition, since magnetic pole position information is extracted from the high-frequency current when alternating high-frequency voltages are applied in a plurality of vector directions using the Fourier series, it is possible to obtain the magnetic pole position and the magnetic pole position calculation error with high accuracy. it can.

The invention according to claim 2 detects a high-frequency current on arbitrarily defined orthogonal coordinates, and extracts a Fourier series zero-order component that is a component that does not change depending on the vector direction of the alternating high-frequency voltage from the detected high-frequency current, A magnetic pole position or magnetic pole position calculation error is calculated from the Fourier series zero-order component.
According to the present invention, the calculation of the magnetic pole position and the like equivalent to claim 1 can be easily realized.

According to a third aspect of the present invention, the alternating high-frequency voltage is applied in a direction orthogonal to the estimated magnetic pole direction, a high-frequency current in the direction orthogonal to the estimated magnetic pole direction flowing at this time is detected, and a Fourier class is detected from the high-frequency current. The zero-order component is extracted to calculate the magnetic pole position or the magnetic pole position calculation error.
According to the present invention, it is possible to calculate the magnetic pole position and the like more easily than the first and second aspects.

Invention, in the control apparatus according to any one of claims 1 to 3, a speed calculating means for calculating the speed of the rotor in proportion-integral operation using the magnetic pole position calculation error, the calculation according to claim 4 Magnetic pole position calculation means for calculating the magnetic pole position by integrating the speed calculation value obtained by the means, and can calculate the magnetic pole position with high accuracy even when the rotor is rotating at high speed. it can.

The invention according to claim 5, in the control apparatus according to any one of claims 1 to 3, a sample point to start the application of the alternating high frequency voltage, or a sample point to terminate the application of the alternating high frequency voltage Means for detecting a fundamental wave current from a current detection value of the motor, a current regulator for calculating a fundamental wave voltage command value so as to eliminate a deviation between the current command value of the motor and the fundamental wave current detection value, and the fundamental wave voltage Means for calculating a voltage command value of the electric motor by adding a command value and the alternating high-frequency voltage; and means for controlling a terminal voltage of the electric motor to the voltage command value . , during one period of the alternating high frequency voltage, a shall Gyosu control constant.
According to the present invention, the fundamental wave current can be controlled to zero by controlling the current command value to zero, and the magnetic pole position and the like are accurately calculated even when the rotor speed is high and the induced voltage is large. be able to.

  The invention according to claim 6 is the one in which the invention described in any one of claims 1 to 5 is applied to the correction of the error of the magnetic pole position detector, and can improve the torque control accuracy and increase the efficiency of the electric motor. .

According to the present invention, the magnetic pole position of a permanent magnet type synchronous motor having no saliency in the rotor can be calculated with high accuracy using the magnetic saturation characteristics of the motor iron core. It is possible to calculate the magnetic pole position accurately even when the Further, in a PMSM drive system that does not have a magnetic pole position detector, the PMSM can be smoothly started by starting operation using the magnetic pole position calculation value obtained by the present invention.
Furthermore, in the case of a drive system equipped with a magnetic pole position detector, more accurate control can be achieved by correcting the error of the magnetic pole position detection value using the magnetic pole position calculation value obtained by the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, the principle of magnetic pole position calculation will be described.
FIG. 6 is a diagram showing the definition of coordinate axes used for magnetic pole position calculation.

In FIG. 6, d and q axes are rotational coordinates that rotate in synchronization with the rotor, and the magnetic pole direction of the rotor is defined as the d axis, and the advance direction of 90 [deg] from the d axis is defined as the q axis. The γ and δ axes are the estimated axes of the d and q axes, and the current control and voltage control of the PMSM are performed by the control device using various quantities on the γ and δ axes. The rotational speeds of the d and q axes and the rotational speeds of the γ and δ axes are ω _r and ω ₁ , respectively.
Here, an angle difference θ _err between the γ-axis (γ, δ-axis) and the d-axis (d, q-axis), that is, a position calculation error is defined by Equation 1.

On the other hand, for the x and y axes in FIG. 6, the vector direction of the alternating high-frequency voltage applied to the stator winding of the permanent magnet synchronous motor to measure the inductance when calculating the magnetic pole position is 90 [ deg] The advance direction is defined as the y-axis. When an angle difference (also referred to as an angle) with respect to the x axis with respect to the γ axis is defined as δ _x , an angle difference θ _xerr between the x axis (x, y axis) and the d axis (d, q axis) is , Expressed by Equation 2.

Next, FIG. 7 shows the relationship between the d-axis current and the flux linkage, and the relationship between the d-axis current and the d-axis inductance.
The linkage flux when the d-axis current I _d is zero is equal to the permanent magnet flux Ψ _m . When the d-axis current I _d is controlled to the plus side, the permanent magnet magnetic flux and the magnetic flux generated by the d-axis current I _d are combined to increase the magnetic flux, the motor iron core is magnetically saturated, and the d-axis inductance decreases. To do.
On the other hand, if the d-axis current I _d is controlled to the minus side, the permanent magnet magnetic flux and the magnetic flux generated by the d-axis current I _d are canceled out, so that the magnetic saturation of the iron core of the motor is reduced and the d-axis inductance increases. To do. That is, the inductance is changed by the value of d-axis current I _d.
From the above, the magnetic saturation characteristics change depending on the angle difference between the d-axis and the current vector, and the inductance changes accordingly.

FIG. 8 shows the relationship between the angle difference θ _xerr between the x axis and the d axis and the x axis inductance when a constant current is passed in the positive direction of the x axis. In FIG. 8, in order to explain the principle in an easy-to-understand manner, the case of SPMSM in which the rotor has no saliency (d-axis inductance and q-axis inductance are equal) is shown.
As shown in FIG. 8, the x-axis inductance changes in one electrical angle cycle with respect to the angle difference θ _xerr . By utilizing this fact, the angle δ _x that minimizes the x-axis inductance with the γ-axis angle θ ₁ being constant is obtained from the relational expressions of Formula 1 and Formula 2 to obtain the d-axis angle (magnetic pole position). Can be calculated.

Here, the x-axis inductance can be measured from an x-axis alternating high-frequency current flowing at this time by applying an alternating high-frequency voltage in the x-axis direction.
FIG. 9 shows the angle δ _x , the x-axis alternating high-frequency voltage v _xh , the high-frequency current i _xh , the fundamental wave voltage v _xf , the fundamental wave current i _xf , the x-axis voltage v _x , the x-axis current i _x , and the x-axis high-frequency current. The waveform of amplitude I _xh is shown. In order to understand the basic principle, first, a case where the current flowing by the induced voltage or the fundamental wave voltage can be ignored will be described.

The alternating high-frequency voltage v _xh applied to the x-axis is a rectangular wave obtained by synthesizing two pulse voltages having the same amplitude and pulse width but different polarities. Note that one cycle of the alternating high frequency voltage _{v xh} and _{T h.}
When measuring the inductance at the time was passed a constant current to the x-axis positive direction, as shown, in the period T _h, control positively in the first half cycle amplitude V _h of x-axis high frequency voltage v _xh In the next half cycle, it is controlled to be negative (hereinafter, the application of the high-frequency voltage is defined as “application of a high-frequency voltage in the x-axis positive direction”). At this time, the x-axis current i _x first changes in the positive direction of the x-axis as shown in the figure, and then returns to the value immediately before the start of the application of the alternating high-frequency voltage v _xh . As is clear from FIG. 9, a high-frequency current i _xh having an amplitude I _xh having a DC bias component in the positive direction of the x-axis flows on the x-axis.

Similarly, the inductance when a current through the constant current to the x-axis negative direction, in the period T _h, the amplitude V _h of x-axis high frequency voltage v _xh controlled to negative in the first half cycle, the next half The period is controlled to be positive (hereinafter, application of this high-frequency voltage is defined as “application of a high-frequency voltage in the negative direction of the x-axis”), and measurement can be made from the high-frequency current i _xh flowing at this time.
When the fundamental wave current (i _xf in FIG. 9) that flows due to the induced voltage or the fundamental wave voltage (v _{xf in} FIG. 9) can be ignored, the high frequency current amplitude I _xh is expressed by the x and y axis high frequency currents i _xh and i _yh. Since it is equal to the x and y-axis current detection values i _x and i _y , it can be calculated by Equation 3.

When the fundamental wave current that flows due to the induced voltage or fundamental wave voltage cannot be ignored, it is necessary to extract a high-frequency current from the detected current value. If the induced voltage and the fundamental wave voltage is constant during one period T _h of the alternating high frequency voltage, a high-frequency current amplitude calculation value when the high frequency voltage applied to the x-axis positive direction, a high frequency voltage to the x-axis negative direction By calculating the average of the calculation value of the high-frequency current amplitude at the time of application, the influence of these fundamental wave voltage components can be offset and the high-frequency current amplitude I _xh can be calculated correctly. Specifically, the high-frequency current amplitude I _xh is calculated by the following formula 4.

Next, the relationship between the high frequency current and the magnetic pole position is derived. First, the x and y-axis state equations for high frequency components are derived under the following conditions.
The angle difference θ _err approximates a constant value.
The angle δ _x is controlled to a constant value.
It is assumed that the frequency of the alternating high frequency voltage to be applied is sufficiently higher than the electrical angular frequency ω ₁ and that the high frequency voltage is dominated by the transient voltage term. With this assumption, the voltage drop due to the armature resistance and the voltage drop due to the armature reaction approximate to zero.
Approximate the inductance change due to magnetic saturation by a sinusoidal function of the angle difference θ _xerr .

  Under the above conditions, the x- and y-axis state equations for the high-frequency components are as follows:

The amplitude vector I _xyh of the high-frequency current i _xh that flows by applying the alternating high-frequency voltage v _xh shown in FIG.

FIG. 10 shows the relationship between the angle difference θ _xerr and the x-axis high-frequency current amplitude I _xh, and here, the case of SPMSM is illustrated for easy understanding of the principle.
From the relationship between the angle difference θ _xerr and the angle δ _x shown in FIG. 10, Equation 6 and Equation 2, the x-axis high-frequency current i _xh is detected by applying the x-axis alternating high-frequency voltage v _xh while changing the angle δ _x. Then, the magnetic pole position can be calculated from the direction in which the component that changes in one cycle of the electrical angle of the angle δ _{x in} the x-axis high-frequency current i _xh is maximized. Here, the component that changes in one cycle of the electrical angle of the angle δ _x can be calculated by Equation 7 as a primary component of the Fourier series.

Here, as shown in FIG. 9, when the high-frequency voltage v _xh is sequentially applied at the angle δ _x to the x-axis plus direction and the x-axis minus direction to detect the x-axis high-frequency current i _xh , the angle δ _x is set. From the x-axis high-frequency current i _xh when changing from zero to 180 [deg], the calculation is performed according to Expression 8 instead of Expression 7.

The angle difference (magnetic pole position calculation error) θ _err is obtained by Equation 9 from the Fourier series obtained by Equation 7 or Equation 8.

When γ-axis angle θ ₁ = 0, from Equation 1, θ _r = −θ _err , so that magnetic pole position θ _r can be calculated by Equation 10.

Hereinafter, embodiments of the present invention based on the above principle will be described.
FIG. 1 is a block diagram showing a first embodiment of the present invention corresponding to claim 1. In FIG. 1, first, as described above, the γ-axis angle (magnetic pole position calculation value) θ ₁ is controlled to zero and input to the current coordinate converter 14 and the voltage coordinate converter 15.

Angle computing unit 20, is increased in steps a predetermined angle [delta] _x from zero every period T _h of the alternating high frequency voltage v _x to be applied to the permanent magnet synchronous motor 80 to 180 [deg], below the high frequency voltage coordinate Input to the converter 22.
x-axis frequency voltage calculator 21 applies an alternating high-frequency voltage every cycle T _h in the x-axis positive direction and the x-axis negative direction. Specifically, the x-axis high-frequency voltage command value v _xh ^* is calculated as v _xh in FIG. 9 and the y-axis high-frequency voltage v _yh ^* is controlled to zero.

The high-frequency voltage coordinate converter 22 converts the x and y-axis high-frequency voltage command values v _xh ^* and v _yh ^* into the γ and δ-axes for each angle δ _x according to Equation 11, and converts the coordinates into the γ and δ-axis high-frequency voltage command values. v _γh ^* and v _δh ^* are output.

The high frequency voltage command values v _γh ^* and v _δh ^* are added to the fundamental voltage command values v _γf ^* and v _δf ^* for each axis by the adders 23a and 23b, respectively, and the γ and δ-axis voltage command values v _γ ^{* are obtained.} , V _δ ^* is obtained. However, in this embodiment, the fundamental wave voltage command values v _γf ^* and v _δf ^* are both controlled to zero.
The γ and δ-axis voltage command values v _γ ^* and v _δ ^* are converted into voltage command values v _u ^* , v _v ^* , and v _w ^* of the three phases based on the magnetic pole position calculation value θ ₁ by the voltage coordinate converter 15. Converted.

On the other hand, the rectifier circuit 60 rectifies the AC voltage of the three-phase AC power supply 50, converts it into a DC voltage, and supplies it to a power converter 70 such as an inverter.
The PWM circuit 13 uses the voltage command values v _u ^* , v _v ^* , v _w ^* and the input voltage E _dc detected by the voltage detection circuit 12 to output the output voltage of the power converter 70 for each phase. A gate signal for controlling to voltage command values v _u ^* , v _v ^* , v _w ^* is generated. The power converter 70 controls the internal semiconductor switching element based on the gate signal, so that the phase terminal voltages of the permanent magnet type synchronous motor 80 are converted into voltage command values v _u ^* , v _v ^* , v _w ^* . Control.

The current coordinate converter 14 detects the phase current detection values i _u and i _w detected by the u-phase current detector 11u and the w-phase current detector 11w, respectively, based on the magnetic pole position calculation value θ ₁ and detects γ and δ-axis currents. Coordinates are converted to values i _γ and i _δ , respectively.
The bandpass filter 30 calculates γ and δ-axis high-frequency current amplitudes I _γh and I _δh from the γ and δ-axis current detection values i _γ and i _{δ according} to Equation 12.

The high-frequency current coordinate converter 31 uses the angle δ _x to coordinate-transform the γ and δ-axis high-frequency current amplitudes I _γh and I _δh using Equation 13 to obtain the x and y-axis high-frequency current amplitudes.
Note that C _xy in Equation 13 is the same as that described in Equation 11.

The Fourier series calculator 32 calculates components (first-order components of the Fourier series) I _xha1 and I _xhb1 that change in one electrical angle cycle of the angle δ _x from the x-axis high-frequency current amplitude I _xh by the above-described Expression 8.
Initial magnetic pole position calculation unit 33, using the above components _{I xha1, I xhb1,} is intended for calculating the magnetic pole position theta _r using Equation 10, may be started motor 80 by using the magnetic pole position theta _r.

Next, FIG. 2 is a block diagram showing a second embodiment of the present invention corresponding to claims 1 and 4.
In the present embodiment, the magnetic pole position calculation error θ _err is calculated from the high frequency current detected by applying the alternating high frequency voltage, and the magnetic pole position and speed are calculated so that the calculation error θ _err becomes zero. The magnetic pole position can be accurately calculated even when rotating. Below, it demonstrates centering on a different point from 1st Embodiment, attaches | subjects the same referential mark about the same location, and abbreviate | omits description.

In FIG. 2, the position error calculator 121 _obtains the magnetic pole position calculation error θ _err from the Fourier series I _xha1 and I _xhb1 output from the Fourier series calculator 32 by the above-described Expression 9.
The speed calculator 122 performs a proportional integral calculation of the magnetic pole position calculation error θ _{err according} to Equation 14 to obtain a speed calculation value ω ₁ .

The magnetic pole position calculator 123 integrates the speed calculation value ω ₁ as shown in Equation 15 to obtain the magnetic pole position calculation value θ ₁ .

Further, in order to speed the convergence of the magnetic pole position calculation value theta _1, at the time of operation start, to preset the initial value of the magnetic pole position in the magnetic pole position calculation value theta _1. The initial value of the magnetic pole position may be calculated from the Fourier series I _xha1 and I _xhb1 calculated by setting θ ₁ to zero _according to the equation 10.

Next, FIG. 3 is a block diagram showing a third embodiment of the present invention corresponding to claims 1, 4 and 5.
In this embodiment, a function of controlling the fundamental current to zero is added to the second embodiment shown in FIG. 2, and the change in magnetic saturation characteristics due to the fundamental current flowing due to the induced voltage is reduced. This is a highly accurate calculation. Below, it demonstrates centering on a different point from 2nd Embodiment, attaches | subjects the same referential mark about the same location, and abbreviate | omits description.

In FIG. 3, the notch filter 131 removes the high-frequency current from the detected γ and δ-axis current values i _γ and i _δ , and detects the γ and δ-axis fundamental currents i _γf and i _δf . From FIG. 9, the x-axis high frequency current _{i xh} the sample points to end the sample points or applied to start the application of x-axis high frequency voltage _{v xh} in period _{T h} (n-2), becomes zero in (n) . Therefore, the notch filter 131 detects γ and δ-axis fundamental wave currents i _γf and i _δf from the detected γ and δ-axis current values i _γ and i _{δ at the} sample points (n−2) and (n).

In FIG. 3, the γ-axis fundamental wave voltage command value v _γf ^* is obtained by calculating a deviation between the γ-axis current command value i _γ ^* and the γ-axis fundamental wave current i _γf by the subtractor 132a. Amplification is performed by the regulator 133a. On the other hand, the δ-axis fundamental wave voltage command value v _δf ^* is obtained by calculating a deviation between the δ-axis current command value i _δ ^* and the δ-axis fundamental wave current i _δ by the subtractor 132b. It amplifies by 133b and calculates.
Here, the γ and δ axis current command values i _γ ^* and i _δ ^* are both controlled to zero. Further, gamma, [delta]-axis fundamental voltage command value _v γf ^_*, v δf _* is to prevent interference with the high-frequency current between the high-frequency voltage one period T _h, it shall be controlled to a constant value.
Note that the magnetic pole position calculation operation in the present embodiment is the same as that in the second embodiment.

Next, FIG. 4 is a block diagram showing a fourth embodiment of the present invention corresponding to claims 2 and 4.
In this embodiment, the calculation of the magnetic pole position calculation error θ _err in the third embodiment of FIG. 3 is simplified, the high-frequency current coordinate converter 31 in the third embodiment is omitted, and the output from the bandpass filter 30 is performed. The γ and δ-axis high frequency current amplitudes I _γh and I _δh are directly input to the Fourier series calculator 141.
First, the principle of this embodiment will be described.
There is a relationship of Equation 16 between the x and y axis high frequency voltage and the γ and δ axis high frequency current.

From Equation 16, when an alternating high-frequency voltage is applied to the x-axis, the γ and δ-axis high-frequency current amplitudes I _γh and I _δh are as shown in Equation 17.

From Equations 16 and 17, the γ and δ-axis high-frequency current amplitudes I _γh and I _δh do not depend on the value of the angle δ _x , in other words, the magnetic pole position calculation error θ _err on the arbitrarily defined orthogonal coordinates. Independent components are included. gamma, [delta]-axis high frequency current amplitude I _{y H,} I _.delta.h from the angle [delta] the zero-order component does not depend on the value of _x I _γha0, when the I _Derutaha0 determined by Fourier series, so Equation 18.

As shown in FIG. 9, at each angle δ _x , when alternating high-frequency voltage is applied in the x-axis plus direction (angle difference δ _x direction) and the x-axis minus direction (angle difference δ _x + π direction), the angle From the γ and δ-axis high-frequency current amplitudes I _γh and I _δh when δ _x is changed from zero to 180 [deg], the Fourier series is calculated by Equation 19.

The magnetic pole position calculation error θ _err can be calculated using Expression 20 from the zero-order component of the Fourier series shown in Expression 19.

When the angle γ-axis θ ₁ = 0, θ _r = −θ _err from Equation 1, so that the magnetic pole position θ _r can be calculated by Equation 21.

Next, the operation of the present embodiment will be described with reference to FIG.
The Fourier series calculator 141 calculates components I _γha0 and I _δha0 that do not depend on the angle δ _x from the γ and δ-axis high-frequency current amplitudes I _δh and I _{δh using} a Fourier series.
The position error calculator 142 calculates the magnetic pole position calculation error θ _{err according} to Equation 20 described above.
Further, to speed the convergence of the magnetic pole position calculation value, preset the initial value of the magnetic pole position in the magnetic pole position calculation value theta ₁ at operation start. The initial value of the magnetic pole position is calculated by using Equation 21 using the Fourier series I _γha0 and I _δha0 by setting the θ ₁ to zero and calculating the Fourier series I _γha0 and I _δha0 of the γ and δ-axis high-frequency currents. .

Next, a fifth embodiment of the present invention corresponding to claim 3 will be described with reference to FIG. In this embodiment, the magnetic pole position calculation is further simplified in the fourth embodiment.
First, the principle of this embodiment will be described.
From Equation 18 described above, the Fourier series I _δha0 of the δ-axis high-frequency current is a sine wave function of the magnetic pole position calculation error θ _err . Therefore, when the magnetic pole position calculation error θ _err is near zero, it can be approximated that the Fourier series I _δha0 and the magnetic pole position calculation error θ _err are in a proportional relationship. Therefore, if the magnetic pole position and velocity are calculated so that the Fourier series I _δha0 becomes zero, the magnetic pole position can be calculated accurately.

In order to obtain I _δha0 from Equation 18, _only information on δ-axis high-frequency current amplitude I _δh , that is, high-frequency current amplitude information in a direction orthogonal to the estimated magnetic pole direction (γ-axis direction) (δ-axis direction) is necessary. It is. Therefore, the alternating high frequency voltage only needs to be applied in the δ-axis plus direction and the δ-axis minus direction. From Equation 19, I _{δha0 at} this time can be calculated from the δ-axis high-frequency current amplitude I _{δh according} to Equation 22 below.

Next, the operation of this embodiment will be described with reference to FIGS. 11 shows the δ-axis high-frequency voltage v _δh , the high-frequency current i _δh , the fundamental wave voltage v _δf , the fundamental wave current i _δf , the δ-axis voltage v _δ , the δ-axis current i _δ , and the δ-axis high frequency current in this embodiment. A waveform of amplitude I _δh is shown.

In FIG. 5, a δ-axis high-frequency voltage calculator 151 provided in place of the angle calculator 20, the x-axis high-frequency voltage calculator 21, and the high-frequency voltage coordinate converter 22 in the fourth embodiment is a δ-axis high-frequency voltage command value v. _δh ^* is controlled to the same rectangular wave alternating high-frequency voltage as v _δh shown in FIG. Here, the γ-axis high-frequency voltage command value v _γh ^* is controlled to zero.
The Fourier series calculator 152 calculates the Fourier series I _δha0 based on the δ-axis high-frequency current amplitude I _δh by the above-described equation 22. The speed calculator 153 amplifies the Fourier series I _δha0 and calculates the speed calculation value ω ₁ using Equation 23.

In the present embodiment, in order to speed up the convergence of the magnetic pole position calculation value, preset the initial value to the magnetic pole position calculation value theta ₁ at operation start. The initial value θ ₁₀ of the magnetic pole position is calculated by calculating the Fourier series I _δha0 of the high-frequency current with θ ₁ set to zero, and the initial magnetic pole position calculator 154 is calculated from the Fourier series I _δha0 by Equation 24.

  In addition, as described in claim 6, when the present invention is applied to a drive system of a permanent magnet type synchronous motor having a magnetic pole position detector, the error of the magnetic pole position detection value is the first to fifth described above. By correcting using the magnetic pole position calculation value obtained by any of the embodiments, the magnetic pole position can be obtained more accurately.

1 is a block diagram showing a first embodiment of the present invention. It is a block diagram which shows 2nd Embodiment of this invention. It is a block diagram which shows 3rd Embodiment of this invention. It is a block diagram which shows 4th Embodiment of this invention. It is a block diagram which shows 5th Embodiment of this invention. It is a figure which shows the definition of the coordinate axis used for a magnetic pole position calculation. It is a figure which shows the relationship between d-axis current and a linkage flux, and the relationship between d-axis current and d-axis inductance. It is a figure which shows the relationship between angle difference (theta) _xerr and x-axis inductance. It is a figure which shows the voltage waveform and current waveform in embodiment of this invention. It is a figure which shows the relationship between angle difference (theta) _xerr and x-axis high frequency current amplitude. It is a figure which shows the voltage waveform and current waveform in 5th Embodiment of this invention.

Explanation of symbols

50 Three-phase AC power supply 60 Rectifier circuit 70 Power converter 80 Permanent magnet synchronous motor (PMSM)
11u u-phase current detection circuit 11w w-phase current detection circuit 12 voltage detection circuit 13 PWM circuit 14 current coordinate converter 15 voltage coordinate converter 20 angle calculator 21 x-axis high-frequency voltage calculator 22 high-frequency voltage coordinate converter 23a adder 23b Adder 30 Bandpass filter 31 High-frequency current coordinate converter 32 Fourier series calculator 33 Initial magnetic pole position calculator 121 Position error calculator 122 Speed calculator 123 Magnetic pole position calculator 131 Notch filter 132a Subtractor 132b Subtractor 133a γ-axis current Controller 133b δ-axis current controller 141 Fourier series calculator 142 Position error calculator 151 δ-axis high-frequency voltage calculator 152 Fourier series calculator 153 Speed calculator 154 Initial magnetic pole position calculator

Claims (6)

Taking the terminal voltage and current of a permanent magnet synchronous motor as vectors,
Means for applying a rectangular wave alternating high-frequency voltage composed of two pulse voltages having the same amplitude and pulse width and different polarities to the electric motor in a plurality of vector directions;
Means for detecting a high-frequency current flowing in the motor when the alternating high-frequency voltage is applied, in a direction parallel to the alternating high-frequency voltage;
Means for extracting a Fourier series primary component that is a component that changes in one cycle depending on the vector direction of the alternating high-frequency voltage from the high-frequency current;
Means for calculating a magnetic pole position or a magnetic pole position calculation error of the electric motor from the Fourier series primary component;
A control device for a permanent magnet type synchronous motor. Taking the terminal voltage and current of a permanent magnet synchronous motor as vectors,
Means for applying a rectangular wave alternating high-frequency voltage composed of two pulse voltages having the same amplitude and pulse width and different polarities to the electric motor in a plurality of vector directions;
Means for detecting a high-frequency current on orthogonal coordinates, which is a current that flows through the electric motor when the alternating high-frequency voltage is applied;
Means for extracting a Fourier series zero-order component that is a component that does not change depending on the vector direction of the alternating high-frequency voltage from the high-frequency current;
Means for calculating a magnetic pole position of the electric motor or a magnetic pole position calculation error from the Fourier series zero-order component;
A control device for a permanent magnet type synchronous motor. Taking the terminal voltage and current of a permanent magnet synchronous motor as vectors,
Means for applying an alternating high frequency voltage of a rectangular wave composed of two pulse voltages having the same amplitude and pulse width and different polarities to the motor in a direction perpendicular to the estimated magnetic pole direction;
Means for detecting a high-frequency current flowing in the motor when the alternating high-frequency voltage is applied, in a direction orthogonal to the estimated magnetic pole direction;
Means for extracting a Fourier series zero-order component that is a component that does not change depending on the vector direction of the alternating high-frequency voltage from the high-frequency current;
Means for calculating a magnetic pole position of the electric motor or a magnetic pole position calculation error from the Fourier series zero-order component;
A control device for a permanent magnet type synchronous motor. In the control device according to any one of claims 1 to 3,
Speed calculation means for calculating the rotor speed by proportional-integral calculation of the magnetic pole position calculation error;
Magnetic pole position calculation means for calculating the magnetic pole position by integrating the speed calculation value obtained by the calculation means;
A control device for a permanent magnet type synchronous motor. In the control device according to any one of claims 1 to 3,
Means for detecting a fundamental wave current from a current detection value of a sample point at which application of the alternating high-frequency voltage is started or a sample point at which application of the alternating high-frequency voltage is terminated;
A current regulator that calculates a fundamental voltage command value so as to eliminate a deviation between the current command value of the electric motor and a detected fundamental wave current value;
Means for calculating the voltage command value of the motor by adding the fundamental voltage command value and the alternating high-frequency voltage;
Means for controlling the terminal voltage of the motor to the voltage command value;
With
A control apparatus for a permanent magnet type synchronous motor, wherein the fundamental wave voltage command value is controlled to be constant during one cycle of the alternating high-frequency voltage. In the control device according to any one of claims 1 to 5,
Magnetic pole position detection means for detecting the magnetic pole position of the rotor,
A control apparatus for a permanent magnet type synchronous motor, wherein an error in a magnetic pole position detection value by the detection means is corrected using a magnetic pole position calculation value obtained by any one of claims 1 to 5.

Images