JP5928216B2 - Inverter control device - Google Patents

Description

  The present invention relates to a technique for controlling an inverter that drives a rotating machine. In particular, the present invention relates to a technique for improving the followability to a torque command when the DC voltage input to the inverter fluctuates greatly.

  Conventionally, once the input AC power is rectified to obtain DC power, the DC power is converted into variable voltage, variable frequency multi-phase (for example, three-phase) AC power by switching, and the AC power is used for multi-phase AC. Power converters that perform variable speed control of these AC rotating machines are widely used.

  The power converter includes a rectifier circuit, a DC link capacitor, and an inverter. The rectifier circuit performs full-wave rectification on the input AC voltage supplied from the commercial power supply and converts it to a DC voltage. The inverter converts the DC voltage into a three-phase AC voltage by switching and supplies it to a three-phase AC rotating machine that is a load.

  In this power converter, a capacitor input type rectifier circuit is most often used as the rectifier circuit. When the DC link capacitor connected to the output side of the capacitor input type rectifier circuit has a function of smoothing the output voltage of the rectifier circuit, the capacitance is set to be large by using an electrolytic capacitor.

  However, the use of such a large-capacity capacitor has problems that the power conversion device becomes large and the manufacturing cost increases.

  As an example of measures against the problem, an electrolytic capacitor-less power converter using a single-phase AC voltage as a power source has been developed (for example, Patent Document 1, Non-Patent Documents 1 and 2). For example, in an electrolytic capacitor-less power converter, a small-capacity film capacitor is used as a DC link capacitor.

Japanese Patent No. 4192979 Japanese Patent Application Laid-Open No. 2005-223991

Utsugi, Oishi, Haga, "A Study on Harmonic Suppression Control of Electrolytic Capacitor-less Inverter for IPMSM Drive", 2011 IEEJ Industrial Application Conference, No.1-43, ppI-265-I-268 Sekimoto and four others, “Development of global power supply harmonic regulation air conditioner by electrolytic capacitor-less inverter”, 2011 IEEJ Motor Drive Study Group, No.MD-11, pp51-56 Yo, Nakano, Wang, “Suppression of disturbance in repetitive control systems”, Transactions of the Society of Instrument and Control Engineers, Vol. 37, No. 5, pp. 397-402, 2001. Lightning and three others "A study on harmonic suppression control of electrolytic capacitorless inverter for IPMSM drive High harmonic suppression method of input current of IPM motor controlled by high power factor", IEEJ Technical Report, Joint Research on Consumer Electronics Meeting, March 4, 2011, HCA-11-11, pp57-62

  In the method described in Patent Document 1 or Non-Patent Documents 1 and 2, since the capacity of the DC link capacitor is small, the DC link voltage fluctuates at a frequency twice the power supply frequency. That is, the timing at which the DC link voltage becomes a small value close to zero is repeated at a cycle that is half the cycle of the power supply frequency.

  The torque of the rotating machine depends on the electric power applied to the rotating machine. Therefore, if the required torque does not change, it is necessary to flow a larger current as the DC link voltage takes a smaller value. However, as the current increases, the copper loss of the rotating machine increases.

  Therefore, Non-Patent Documents 1 and 4 disclose techniques for suppressing copper loss and harmonics by giving small electric power to a rotating machine when the DC link voltage takes a small value.

  However, in the techniques shown in Non-Patent Documents 1 and 4, the power output from the inverter (and thus given to the rotating machine) fluctuates at a frequency twice the power supply frequency, approximating the waveform of the DC link voltage. Such large fluctuations in power complicate torque control.

  In the control of Non-Patent Documents 1 and 4, a repetitive control system as exemplified in Non-Patent Document 3 is adopted. However, a current command is generated by using a bandpass filter.

  However, when iterative control is used, as shown in FIG. 2 of Non-Patent Document 3, the response is speeded up, but the gain of a plurality of frequencies is increased in a high frequency range. In Patent Document 2 and Non-Patent Document 2, it is possible to reduce or suppress a specific frequency in current control using repetitive control, but the frequency characteristics tend to increase in other high frequency bands. There are challenges.

  In particular, as described in Non-Patent Documents 1 and 2, in an electrolytic capacitor-less power converter, it is necessary to reduce the power supply harmonics of the input current flowing on the power supply side. When the current control of the repetitive control is used, the gain of a plurality of frequencies is increased in a high frequency range, so that the characteristics of the power supply harmonics of the input current flowing on the power supply side may be deteriorated.

  Regarding this phenomenon, Non-Patent Document 4 introduces a technique that employs a band-pass filter together with repetitive control, and suppresses gains at frequencies other than twice the power supply frequency in repetitive control. However, this also increases the followability of control with respect to the torque command only at a frequency twice the power supply voltage, and does not suppress fluctuations in the torque command itself. Moreover, the bandpass filter described in Non-Patent Document 4 has a second-order configuration in both the denominator and numerator as shown in Equation (9) of Non-Patent Document 4, and control is necessary to realize this. There is a problem that the calculation load for realizing the system is high. Such an increase in calculation load requires a high-performance microcomputer and increases the cost.

  In view of the above-mentioned problems, the main object of the present invention is to realize control that suppresses copper loss of a rotating machine by reducing the current when the DC link voltage is small and follows fluctuations in the current command with a low calculation load. There is to do. Furthermore, a secondary purpose is to suppress fluctuations in the current command.

  The inverter control device (8) according to the present invention receives a DC voltage (Vdc) that fluctuates periodically, and outputs an inverter (2) that outputs a multiphase AC voltage (Vu, Vv, Vw) to the rotating machine (1). ) To control.

  The first aspect includes a current command generator (12) that generates a pair of current commands (iqr, idr) in a rotating coordinate system that rotates at a rotation angular frequency of the rotating machine, and a current ( Voltage command generation for generating a voltage command (Vqr, Vdr) in the rotating coordinate system from a deviation between a pair of currents (iq, id) representing iu, iv) in the rotating coordinate system and the pair of current commands (24a, 24b) and a coordinate converter (17) for obtaining command values (Vu *, Vv *, Vw *) of the AC voltage from the voltage command.

  The voltage command generator includes an amplifier (13a, 13b) that amplifies the deviation, and a proportional-integral controller (16a, 16b) that performs proportional-integral control on the sum of the output of the amplifier and the deviation. . The absolute value of the current command increases only during a period when the DC voltage increases, and decreases only during a period when the DC voltage decreases. The amplifier resonates at least at a frequency (ω1 / 2π) at which the DC voltage varies.

  A second aspect of the inverter control device according to the present invention is the first aspect, wherein a torque or speed command value (T *) and a waveform of the DC voltage (Vdc) are input, A command value is modulated at a frequency (2f) at which the DC voltage fluctuates, and a corrected command value (T **) that increases only when the DC voltage increases and decreases only when the DC voltage decreases. A command value corrector (11) to be generated is further provided.

  The current command generator (12) receives the correction command value and the current phase (β) and generates the current command (iqr, idr).

  A third aspect of the inverter control device according to the present invention is the first aspect or the second aspect, and the amplifier resonates at other frequencies.

  A fourth aspect of the inverter control device according to the present invention is the second aspect thereof, wherein the other frequency is an integral multiple of the frequency at which the DC voltage fluctuates.

  According to the first aspect of the inverter control device of the present invention, the current flowing through the rotating machine is suppressed when the DC voltage is small. Therefore, the copper loss of the rotating machine can be suppressed. Further, since the amplifier in the voltage command generator resonates at a frequency at which the DC voltage fluctuates, the amplification at that frequency becomes very large, and control that follows the fluctuation of the current command is performed. Compared with the case where a band-pass filter is used, the amplifier can realize a steep frequency characteristic with a low-order transfer function, thereby reducing a calculation load.

  According to the 2nd aspect of the inverter control apparatus concerning this invention, a torque command or a speed command is modulated and the electric current command said to a 1st aspect is produced | generated.

  According to the third aspect or the fourth aspect of the inverter control device according to the present invention, even if a trapezoidal wave is adopted as a torque or speed command value or a current command, the degree of follow-up to the fluctuation is good. It is.

It is a circuit diagram which shows the power converter device in 1st Embodiment, and its periphery. It is a graph which shows the waveform of various quantities in a 1st embodiment. It is a block diagram which illustrates the composition of the output operation means in a 1st embodiment. It is a block diagram which shows the structure of the amplifier in 1st Embodiment. It is a Bode diagram which shows the transfer function in a 1st embodiment. It is a Bode diagram which shows the transfer function in a 1st embodiment. It is a Bode diagram which shows the transfer function in a 1st embodiment. It is a graph which shows the waveform of various quantities in 2nd Embodiment. It is a block diagram which shows the structure of the amplifier in 2nd Embodiment. It is a Bode diagram which shows the transfer function in a 2nd embodiment.

First embodiment.
FIG. 1 is a circuit diagram showing a power converter and its surroundings in the first embodiment of the present invention.

  In order to drive an AC rotating machine (hereinafter, simply referred to as “rotating machine”) 1, the power conversion device includes a rectifier 5, a DC link capacitor 3, an inverter 2 as a power converter, and an output as an inverter control device. Voltage calculation means 8. In this embodiment, the capacity of the DC link capacitor 3 is reduced and the above-described electrolytic capacitor-less power converter is employed. The rotating machine 1 functions as a three-phase load of the inverter 2, and for example, a synchronous machine or an induction machine can be adopted.

  The rectifying means 5 converts the AC voltage of the single-phase AC power source 6 into a DC voltage, and, for example, a full-wave rectification type diode bridge is employed. The capacity of the DC link capacitor 3 is small, and the DC link voltage Vdc supported by it is periodically changed. Here, since the rectification means 5 performs full-wave rectification, the DC link voltage Vdc fluctuates at a period half that of the AC voltage of the single-phase AC power supply 6. The DC link voltage Vdc is detected by the DC voltage detection means 4.

  The inverter 2 receives the DC link voltage Vdc and outputs multiphase AC voltages Vu, Vv, Vw corresponding to the torque command T * to the rotating machine 1.

  The output voltage calculation means 8 should be given to the rotating machine 1 by the inverter 2 based on the DC voltage Vdc detected by the DC voltage detecting means 4, the currents iu and iv flowing through the rotating machine 1, and the phase angle θ of the rotating machine 1. Voltage commands Vu *, Vv *, Vw *, which are command values of the voltages Vu, Vv, Vw, are output.

  The currents iu and iv are detected by current detectors 7a and 7b, respectively, and the phase angle θ of the rotating machine 1 is detected by the phase detector 9.

  The current detectors 7 a and 7 b are provided in a connection for connecting the inverter 2 and the rotating machine 1. For example, a current transformer can be used. In addition, the phase current may be detected using a current flowing in the inverter 2 such as a current flowing through the DC link (bus current) using other known methods.

  Since the sum of the three-phase currents flowing through the rotating machine 1 is zero, the remaining one-phase current can be obtained from the detected currents (currents iu and iv in this case) for two phases. Of course, if the currents iu and iv are grasped as the U-phase current and the V-phase current, respectively, the current iu can be obtained by detecting the W-phase current and the V-phase current.

  FIG. 2 is a graph showing waveforms of various quantities in the present embodiment. FIG. 5A shows the waveform of the DC link voltage Vdc, FIG. 5B shows the waveform of the corrected torque command T ** together with the torque command T *, and FIG. 4C shows the q-axis current command iqr. And the waveform of the d-axis current command idr. FIG. 2 shows only various waveforms in one cycle of the DC link voltage Vdc, that is, in the half cycle T / 2 of the AC voltage of the single-phase AC power supply 6. Since the waveforms in the other periods are the same as those in the half cycle T / 2 minutes, they are omitted in FIG.

  During steady state operation, the torque command T * does not vary greatly as indicated by the broken line in FIG. 2B, and takes a substantially constant value, for example, in a half cycle T / 2. Torque output from the rotating machine 1 is covered by electric power input to the rotating machine 1. Therefore, if an almost constant torque is to be output in the half cycle T / 2, a large current needs to flow through the rotating machine 1 when the DC link voltage Vdc is small.

  However, increasing the current flowing through the rotating machine 1 increases the copper loss of the rotating machine 1. From this point of view, it is desirable to reduce the torque when the DC link voltage Vdc is low. Therefore, the corrected torque command T ** is preferably set to be small when the DC link voltage Vdc is small and large when the DC link voltage Vdc is large to some extent.

  That is, it is desirable that the corrected torque command T ** also fluctuates with the frequency (2 / T), like the DC link voltage Vdc. More preferably, from the viewpoint of suppressing torque fluctuation, it is desirable that the period during which the corrected torque command T ** fluctuates is short.

  Therefore, in the present embodiment, the corrected torque command T ** is used instead of the torque command T *, and a trapezoidal wave is used for the waveform. Referring to FIG. 2, the trapezoidal wave increases in the period tr only during the period when the DC link voltage Vdc increases, and decreases during the period td only during the period when the DC link voltage Vdc decreases. The average value of the corrected torque command T ** in the half cycle (T / 2) is preferably equal to the torque command T *. As a result, control corresponding to the torque command T * can be performed.

  Reflecting such a waveform of the corrected torque command T **, the waveforms of the q-axis current command iqr and the d-axis current command idr also become trapezoidal waves. Here, assuming a so-called field weakening, the d-axis current command idr is negative.

  As described above, when the inverter 2 is controlled based on the corrected torque command T ** that fluctuates at the frequency (2 / T), the output voltage calculation means 8 needs to have good followability at the frequency (2 / T). . For example, when the power supply frequency is 50 Hz, the frequency at which the DC link voltage Vdc fluctuates is 100 Hz. Therefore, it is necessary to set the current control response in the output voltage calculation means 8 to 100 Hz or more.

  For such high-performance and high-speed response of current control, Non-Patent Document 4 employs a band-pass filter together with repetitive control, and the gain at frequencies other than twice the power supply frequency is suppressed in repetitive control. . However, as described above, such a configuration increases the calculation load. Therefore, in the output voltage calculation means 8 of the present embodiment, the corrected torque command T ** is suppressed while suppressing the copper loss of the rotating machine 1 by reducing the torque when the DC link voltage Vdc is small as follows. The control that follows the fluctuation of the is realized with a low calculation load.

  As shown in FIG. 1, the output voltage calculation means 8 includes a DC link voltage Vdc (to be precise, its value and / or waveform) detected by the DC voltage detection means 4, and current detectors 7a and 7b. The currents iu and iv (to be precise, the values thereof) and the phase θ detected by the phase detector 9 are input.

  Further, a torque command T * for the AC rotating machine 1 to output a desired torque is input. The torque command T * may be a value calculated by multiplying the deviation between the speed command in the case of speed control and the detected actual speed of the AC rotating machine 1 by gain multiplication or the like, that is, an output value of speed control. That is, the control amount corresponds to the torque output by the motor.

  FIG. 3 is a block diagram illustrating the configuration of the output voltage calculation means 8. The output voltage calculation means 8 includes a torque command corrector 11, a dq-axis current command generator 12, voltage command generators 24a and 24b, coordinate converters 10 and 17, and subtractors 14a and 14b.

  The coordinate converter 10 is a three-phase / dq axis converter, and converts the currents iu and iv into a q axis current iq and a d axis current id in the dq coordinate system. Here, the dq coordinate system is a rotating coordinate system that has a d-axis and a q-axis that advances by π / 2 relative to the d-axis and rotates at the rotation angular frequency of the rotating machine 1. Originally, the three-phase current is converted into the current on the dq axis, but the q-axis current iq and the d-axis current id are actually obtained from the feature that the sum of the three-phase currents is zero as described above. The phase current required for this is sufficient for two phases.

  As is well known, when coordinate conversion of a three-phase voltage or a three-phase current into rotationally orthogonal two axes is required, a control coordinate axis is required, and the phase of this control coordinate axis is assumed to be θ. In the present embodiment, the phase θ is detected by the phase detector 9 shown in FIG.

  The torque command corrector 11 receives the DC link voltage Vdc and the torque command T *, and outputs the above-described corrected torque command T **. Specifically, a waveform of the DC link voltage Vdc is input, and the torque command T ** is modulated at a frequency (2f = 2 / T) at which the DC link voltage Vdc varies. As a result, a corrected torque command T ** that increases only during a period in which the DC link voltage Vdc increases and decreases only in a period in which the DC link voltage Vdc decreases is obtained.

  The dq axis current command generator 12 receives the corrected torque command T ** and the current phase β, and generates current commands iqr and idr in the dq coordinate system. The current phase β is grasped as, for example, a phase advance angle with respect to the q axis of the current flowing through the rotating machine 1.

  The subtractor 14a outputs a deviation Δd of the d-axis current id with respect to the current command idr, and the subtractor 14b outputs a deviation Δq of the q-axis current iq with respect to the current command iqr.

  The coordinate converter 17 is a dq / three-phase axis converter, and converts the d-axis voltage command Vdr and the q-axis voltage command Vqr into three-phase voltage commands Vu *, Vv *, and Vw *. The voltage commands Vu *, Vv *, Vw * are command values for the AC voltages Vu, Vv, Vw applied to the rotating machine 1. The inverter 2 controls the rotating machine 1 to perform an operation corresponding to the torque command T * based on the voltage commands Vu *, Vv *, Vw *.

  Since the coordinate converters 10 and 17, the torque command corrector 11, the dq axis current command generator 12, and the subtractors 14 a and 14 b are well known, description of their specific configuration is omitted in the present embodiment.

  As is well known, the operation of the coordinate converters 10 and 17 requires a control coordinate axis, and the phase of the control coordinate axis is θ. In the present embodiment, the phase θ is detected by the phase detector 9 shown in FIG.

  The voltage command generator 24a generates a voltage command Vdr from the deviation Δd, and the voltage command generator 24b generates a voltage command Vqr from the deviation Δq. Since the voltage command generators 24a and 24b generate a voltage command value from the current deviation, the voltage command generators 24a and 24b are grasped together with the subtracters 14a and 14b as a current control unit.

  The voltage command generator 24a includes an amplifier 13a, an adder 15a, and a PI controller 16a. The voltage command generator 24b includes an amplifier 13b, an adder 15b, and a PI controller 16b.

  FIG. 4 is a block diagram showing the configuration of the amplifiers 13a and 13b, which is employed in both the amplifiers 13a and 13b. The configuration includes a constant multiplier 18, a subtracter 19, a multiplier 20, and integrators 21 and 22.

  In the present embodiment, it is desirable that the amplifiers 13a and 13b have a high gain only at the frequency 2f. For this purpose, the current control unit constitutes a filter for increasing the gain only at the specific angular frequency ω1. For example, the frequency is 2f = 100 Hz.

  First, the filter transfer function F0 for increasing the gain with respect to the pole jω1 is expressed by the equation (1). The symbol s indicates a differential element.

  The pole of the transfer function F0 shown in the equation (1) is jω1, and becomes zero in a steady state (s = 0) and a sufficiently high band (s = ∞), so that only a gain is obtained for the pole jω1. Is possible.

  FIG. 5 is a Bode diagram, and graphs G1 and P1 show the gain and phase of equation (1) of the transfer function F0 when the specific angular frequency ω1 is 2π × 100 [rad / sec].

  As can be seen from FIG. 5, the transfer function F0 has a higher gain with respect to the pole. However, since the transfer function F0 has a characteristic that the gain is large even in the vicinity of the transfer function F0, the aim is to reduce the gain in the vicinity of the pole and to reduce the gain in the high frequency range. Specifically, the transfer function F1 of the filter obtained by dividing the transfer function F0 by the specific frequency ω1 is obtained so that the gain of the filter is inversely proportional to the frequency.

  The graph G2 in FIG. 5 shows the gain of the transfer function F1 when the specific angular frequency ω1 is 2π × 100 [rad / sec]. The phase of the transfer function F1 is shown by the graph P1 as with the phase of the transfer function F0.

From the graph G2, it can be seen that the gain characteristic of the transfer function F1 has a gain of 1 (= 0 dB) or more only at a specific frequency (ω1 / 2π) = 100 Hz (10 ² Hz). Further, from the comparison between the graphs G1 and G2, it can be confirmed that the gain in other frequency bands is lower in the transfer function F1 than in the transfer function F0. Therefore, according to the expression (2), it is possible to design a filter that suppresses the influence on other frequency bands while increasing the gain only at a specific frequency.

  The configuration of the amplifiers 13a and 13b that realize the characteristic of the equation (2) is shown in FIG. The constant multiplier 18 multiplies the deviation Δd (or deviation Δq) by a constant with the gain K1. The subtracter 19 subtracts the output of the multiplication unit 20 from the output of the constant multiplication unit 18. The integrator 21 performs an integration operation on the output of the subtracter 19 and outputs a signal Hd (or signal Hq). The integration unit 22 performs an integration operation on the signal Hd (or the signal Hq), and the multiplication unit 20 multiplies the square of the specific angular frequency ω1 on the result of the integration operation. It is obvious that the amplifiers 13a and 13b can realize the transfer function F1 by such a configuration.

  The transfer function F1 has a characteristic corresponding to the admittance of a series resonance circuit that resonates at a specific angular frequency ω1. In other words, it can be said that the amplifiers 13a and 13b have a characteristic of resonating at a specific frequency (ω1 / 2π) that is a frequency at which the DC link voltage Vdc varies. By utilizing such resonance characteristics, as is clear from the equation (2), the transfer function is expressed by a polynomial of a relatively low order in which the denominator is a quadratic expression and the numerator is a linear expression.

  By employing the amplifiers 13a and 13b that resonate at a specific frequency as described above, the technique of this embodiment has an advantage that the calculation load is lower than the technique exemplified in Non-Patent Document 4.

  In the voltage command generator 24a, the signal Hd output from the amplifier 13a and the deviation Δd are added by the adder 15a. The PI controller 16a generates a voltage command Vdr by performing proportional integration control on the addition result. Similarly, in the voltage command generator 24b, the signal Hq output from the amplifier 13b and the deviation Δq are added by the adder 15b. The PI controller 16b generates a voltage command Vqr by performing proportional integration control on the addition result.

  In the PI controllers 16a and 16b, any transfer function is expressed by Expression (3). However, gain Kp and integration time constant Ti were introduced. The first term on the right side of Equation (3) indicates proportional control, and the second term on the right side indicates integral control.

  FIG. 6 is a Bode diagram, and graphs G3 and P3 represent the gain and phase of the transfer function represented by Expression (3), respectively. Graphs G21 and P1 are also shown. The graph G21 shows the gains of the amplifiers 13a and 13b, but the value of the gain K1 is different from the graph G2 shown in FIG. As described above, the gain K1 of the constant multiplier 18 is also an effective parameter for adjusting the gain characteristic at the specific frequency (ω1 / 2π).

  Since the phase curve of the Bode diagram does not depend on the value of the gain K1, the graph P1 is common in FIGS.

  FIG. 7 is a Bode diagram, and graphs G4 and P4 represent the gain and phase of the transfer function of the voltage command generator 24a (or voltage command generator 24b), respectively.

  As described above, the PI controllers 16a and 16b, the amplifiers 13a and 13b, and the adders 15a and 15b included in the voltage command generators 24a and 24b do not increase the gain at the peripheral frequencies, and a specific frequency (here, 100 Hz). Only gain characteristics can be improved.

  As described above, in the present embodiment, in the electrolytic capacitorless power conversion device in which the capacity of the DC link capacitor 3 is reduced, the torque command T * is changed at the same frequency as the DC link voltage Vdc. Modify to *. Thereby, when the DC link voltage Vdc is low, the current flowing to the rotating machine 1 can be reduced, and the copper loss of the rotating machine 1 is reduced. This contributes to reducing the size and cost of the rotating machine 1.

  The current control for responding to the corrected torque command T ** has good followability because the gain of the transfer function is increased at the frequency at which the corrected torque command T ** fluctuates. Such a good followability contributes to a quick start-up of torque.

  In addition, since the gain of current control cannot be increased at other frequencies, the harmonic component of the input current flowing on the power supply side is not increased.

  In addition, the amplifiers 13a and 13b used to obtain such frequency characteristics resonate at a frequency at which the DC link voltage Vdc fluctuates. Therefore, compared with the case where a band-pass filter is used, the steep frequency characteristics are lower. It can be realized with a transfer function. This is desirable from the viewpoint of suppressing calculation load.

  Further, by adopting a trapezoidal wave for the corrected torque command T **, it is possible to suppress torque command fluctuations, and hence torque fluctuations.

  The present embodiment can be applied even when the single-phase AC power supply 6 is replaced with a three-phase AC power supply. In this case, the rectifier 5 is a three-phase full-wave rectifier diode bridge. In this case, since one cycle of the DC link voltage Vdc is 1/6 of the cycle T of the three-phase AC power supply, the amplifiers 13a and 13b employ a configuration that resonates at a specific angular frequency ω1 = 12π / T. .

Second embodiment.
FIG. 8 is a graph showing waveforms of various quantities in the present embodiment, and corresponds to FIG. 2 of the first embodiment.

  As in the first embodiment, the waveform indicated by the corrected torque command T ** increases only during the period tr when the DC link voltage Vdc increases, and decreases during the period td only when the DC link voltage Vdc decreases. To do. However, unlike the first embodiment, the correction torque command T ** has a different increase rate (waveform slope) between the first half tr1 and the second half tr2 of the period tr, and a decrease rate between the first half td2 and the second half td1 of the period td. (Wave slope) is different. Along with this, the current commands iqr and idr also exhibit trapezoidal waves whose inclined portions are bent. In the present embodiment, a technique corresponding to such a corrected torque command T ** that exhibits a trapezoidal wave in which an inclined portion is bent will be described.

  In such a case, the corrected torque command T ** also fluctuates at other frequencies that are twice the power supply frequency. Also in this embodiment, the configuration shown in FIG. 3 is adopted for the output voltage calculation means 8. However, in the present embodiment, the amplifiers 13a and 13b increase the gain only at the specific angular frequencies ω1 and ω2. Expressions (4) and (5) are obtained in the same manner as Expressions (1) and (2).

  FIG. 9 is a block diagram showing the configuration of the amplifiers 13a and 13b, which is employed in both the amplifiers 13a and 13b. The configuration includes constant multipliers 18a and 18b, subtractors 19a and 19b, multipliers 20a and 20b, integrators 21a, 21b, 22a and 22b, and an adder 23.

  The constant multiplication unit 18a corresponds to the constant multiplication unit 18 of the first embodiment, and multiplies the deviation Δd (or deviation Δq) by a gain K1. The constant multiplier 18b corresponds to the constant multiplier 18 of the first embodiment, and multiplies the deviation Δd (or deviation Δq) by a gain K2.

  The subtractors 19a and 19b are both equivalent to the subtractor 19 of the first embodiment, the integration units 21a and 21b are both equivalent to the integration unit 21 of the first embodiment, and the integration units 22a and 22b are both This corresponds to the integration unit 22 of the first embodiment.

  The multiplication unit 20a corresponds to the multiplication unit 20 of the first embodiment, and multiplies the square of the specific angular frequency ω1. The multiplication unit 20b corresponds to the multiplication unit 20 of the first embodiment, and multiplies the square of the specific angular frequency ω2. The adder 23 adds the outputs of the integrators 21a and 22b and outputs a signal Hd (or signal Hq).

  It is apparent that the amplifiers 13a and 13b that realize the transfer function F3 of the above equation (5) can be obtained by the configuration of FIG.

  FIG. 10 is a Bode diagram, and graphs G5 and P5 represent the gain and phase of the transfer function represented by Expression (5), respectively. Graphs G6 and P6 are also shown. Graphs G6 and P6 represent the gain and phase of the transfer function of the voltage command generator 24a (or voltage command generator 24b), respectively.

  Thus, it is desirable that the amplifiers 13a and 13b resonate in addition to a frequency that is twice the frequency at which the DC link voltage Vdc fluctuates in accordance with the frequency component of the waveform indicated by the corrected torque command T **.

  Although the case where two specific angular frequencies ω1, ω2 are employed has been described in the present embodiment, three or more specific angular frequencies ω1, ω2,... Ωn (n ≧ 3) may be employed. In this case, a plurality of configurations shown in FIG. 4 are provided in parallel with the deviation Δd (or deviation Δq) and the adder 23.

Third embodiment.
As described in the second embodiment, the amplifiers 13a and 13b vary in the DC link voltage Vdc even when the inclined portion of the waveform of the corrected torque command T ** is bent or not. It is desirable to resonate at frequencies other than twice the frequency to be transmitted. This is because the frequency spectrum of a trapezoidal wave usually includes a frequency component that is an integral multiple of the fundamental period of the trapezoidal wave.

  Therefore, the amplifiers 13a and 13b shown in the second embodiment are employed, and the frequency at which they resonate is desirably an integer multiple of the frequency (4π / T) at which the DC link voltage fluctuates. This is because even if the trapezoidal wave is adopted as the corrected torque command T **, the degree of follow-up to the fluctuation is good.

<Deformation>
The torque command T * is a torque command value, but a corrected command value can be adopted for the speed command value as well as the corrected torque command T **. In this case, the torque command corrector 11 is replaced with a command value corrector that generates a corrected speed command from the speed command. Specifically, the command value corrector modulates the speed command with a frequency 2f at which the DC link voltage Vdc fluctuates, and increases only in a period in which the DC link voltage Vdc increases, and in a period in which the DC link voltage Vdc decreases. Generates a corrected speed command that only decreases.

  In addition, as described above, the waveforms of the q-axis current command iqr and the d-axis current command idr are also trapezoidal waves, reflecting the waveform of the corrected torque command T **. That is, in each of the above-described embodiments, the absolute values of the waveforms of the q-axis current command iqr and the d-axis current command idr increase only during the period in which the DC link voltage dc increases, and in the period in which the DC link voltage Vdc decreases. It can be grasped that only decreases.

  Eventually, if the q-axis current command iqr and the d-axis current command idr have waveforms that exhibit the above increase / decrease, even if the factor that caused the waveforms is a torque command, it is a speed command. However, the above embodiment can be applied to other factors. If the amplifiers 13a and 13b that resonate at least at the frequency (ω1 / 2π) with respect to the deviation Δd of the d-axis current id with respect to the current command idr and the deviation Δq of the q-axis current iq with respect to the current command iqr are used, Control following changes in the q-axis current command iqr and the d-axis current command idr is performed, and the effects of the above-described embodiment can be obtained.

DESCRIPTION OF SYMBOLS 1 AC rotary machine 2 Inverter 8 Output voltage calculating means 11 Torque command modifier 12 dq axis current command generator 13a, 13b Amplifier 16a, 16b PI controller 17 Coordinate converter 24a, 24b Voltage command generator

Claims (4)

An inverter control device (8) for controlling an inverter (2) that inputs a periodically changing DC voltage (Vdc) and outputs a multiphase AC voltage (Vu, Vv, Vw) to a rotating machine (1). And
A current command generator (12) that generates a pair of current commands (iqr, idr) in a rotating coordinate system that rotates at a rotational angular frequency of the rotating machine;
A voltage command (Vqr, Vdr) in the rotating coordinate system is calculated from a deviation between a pair of currents (iq, id) representing the current (iu, iv) flowing in the rotating machine in the rotating coordinate system and the pair of current commands. ) To generate voltage command generators (24a, 24b);
A coordinate converter (17) for obtaining a command value (Vu *, Vv *, Vw *) of the AC voltage from the voltage command;
The voltage command generator is
An amplifier (13a, 13b) for amplifying the deviation;
A proportional-integral controller (16a, 16b) that performs proportional-integral control on the sum of the output of the amplifier and the deviation;
The absolute value of the current command increases only in a period in which the DC voltage increases, decreases only in a period in which the DC voltage decreases,
The amplifier is an inverter control device that resonates at least at a frequency (ω1 / 2π) at which the DC voltage varies. Input the torque or speed command value (T *) and the waveform of the DC voltage (Vdc), and modulate the command value with the frequency (2f) at which the DC voltage fluctuates to increase the DC voltage. A command value corrector (11) that generates a corrected command value (T **) that increases only during a period when the DC voltage decreases and decreases only during a period when the DC voltage decreases.
Further comprising
The inverter control device according to claim 1, wherein the current command generator (12) receives the correction command value and a current phase (β) and generates the current command (iqr, idr).   The inverter control device according to claim 1, wherein the amplifier resonates at other frequencies.   The inverter control device according to claim 3, wherein the other frequency is an integral multiple of the frequency at which the DC voltage varies.

Images