JP5401537B2 - AC motor drive device - Google Patents

Description

  The present invention relates to a drive device for an AC motor that drives an AC motor such as an induction motor or a permanent magnet synchronous motor by pulse width modulation (PWM).

  When controlling an AC motor with high response and high accuracy, it is necessary to manage the torque of the motor. For this purpose, it is necessary to detect the phase current of the motor and control the torque current component and the excitation component current independently in the controller. This technique is widely used as vector control.

  DCCT (DC Current Transformer), which directly detects phase current as current detection means, is a current sensor that can detect a wide range from AC to DC and has excellent characteristics, but it is expensive and large in size. It is an obstacle to downsizing. In a small general-purpose inverter, a shunt resistor that is inexpensive and easy to mount is used instead of DCCT.

  As a current detection method using a shunt resistor, JP-A-63-80774 (Patent Document 1), JP-A-6-153526 (Patent Document 2), JP-A-2008-131770 (Patent Document 3). And JP-A-6-284747 (Patent Document 4).

  Japanese Patent Laid-Open No. 63-80774 (Patent Document 1) is a current detection method in which a shunt resistor is attached to the lower side of each arm of an inverter as shown in FIG. 21, and the shunt resistor is turned on when the lower arm is turned on. By sampling the voltage, the phase current can be detected.

  Japanese Patent Laid-Open No. 6-153526 (Patent Document 2) is a current detection method in which a shunt resistor is attached to a DC bus current as shown in FIG. The installation of a shunt resistor at this position is generally known as overcurrent detection of an inverter. However, in Japanese Patent Laid-Open No. 6-153526 (Patent Document 2), an electric motor is obtained from the value of the current flowing through the shunt resistor. The phase current is reproduced. That is, the phase of the current flowing through the shunt resistor is determined from the switching state of the inverter main circuit.

  Japanese Unexamined Patent Application Publication No. 2008-131770 (Patent Document 3) discloses an improved version of Patent Document 2. In Patent Document 2, for example, at low speed driving or in a region where the carrier frequency is high, the pulse width is short and it is difficult to detect current. Therefore, Patent Document 3 corrects the pulse width in consideration of the symmetry of PWM. Over time, detection accuracy is ensured.

  Japanese Patent Application Laid-Open No. 6-284747 (Patent Document 4) discloses a current detection method in which two shunt resistors for a DC bus current are prepared for the purpose of improving the accuracy of overcurrent detection. One shunt resistor is used to detect the combined current of the main circuit switching element of the lower arm, and the other shunt resistor is used to detect the combined current of the diode connected in parallel to the switching element.

JP-A-63-80774 JP-A-6-153526 JP 2008-131770 A JP-A-6-284747

  However, the current detection method described in Patent Document 1 has a problem that shunt resistors are required for each phase (three). Furthermore, the shunt resistance varies in each phase, which causes a three-phase imbalance. Since the sensor unbalance cannot be distinguished from the motor or load unbalance, correction on the control side is impossible, and the desired torque accuracy cannot be obtained.

  In the current detection methods disclosed in Patent Documents 2 and 3, since only one shunt resistor is required, there is no fear of variation between phases. However, under conditions where the pulse width is short, such as during low-speed rotation with a low modulation rate or at a high carrier frequency, the current flowing through the shunt resistor becomes a pulse-shaped current with a narrow width, which is affected by ringing and the like. It becomes difficult to detect. Although this is alleviated in Patent Document 3, there is still a limit. For example, if the pulse width is corrected to the detectable condition, the current can be detected, but the harmonics increase due to the correction of the pulses, and the resulting waveform distortion and heat generation due to the harmonic loss of the motor. , The decrease in efficiency becomes a problem.

  In the current detection method disclosed in Patent Document 4, two shunt resistors are required. However, since the object of the waveform is maintained, there is no concern about three-phase imbalance. However, the phase of the current flowing through the shunt resistor cannot be specified. For example, when the lower arm of a phase is turned on, it cannot be distinguished whether the switching element is turned on and a positive current flows, or whether the diode is turned on and a negative current flows. There's a problem.

  In Patent Document 4, accurate current detection like vector control is not aimed, but current detection at the time of regeneration and high modulation rate is reliably performed and used as a means for circuit protection. It is impossible to perform a simple motor control.

  In view of these problems, an object of the present invention is to provide a drive device for an AC motor that realizes phase current detection of an AC motor with high accuracy and high quality with a small number of parts and a current waveform with little distortion. It is in.

  Of the inventions disclosed in the present application, the outline of typical ones will be briefly described as follows.

  An AC motor drive device according to a representative embodiment of the present invention includes an inverter that outputs a sinusoidal AC current, a three-phase AC motor connected to the inverter, and a pulse width modulation signal for the inverter. And a controller for controlling the inverter to detect the combined current of the three-phase switching elements of the upper or lower arm of the inverter main circuit as a DC bus current detecting means. 1 current detection means and at least one of second current detection means for detecting a combined current of diodes connected in parallel to the three-phase switching elements, and using the current detection value, the three-phase Driving the three-phase AC motor by estimating and calculating the amplitude and phase of the AC current flowing in the AC motor, or the corresponding state quantities And features.

  The effects obtained by the representative ones of the inventions disclosed in the present application will be briefly described as follows.

  According to the three-phase AC motor driving apparatus according to the representative embodiment of the present invention, it is possible to provide a motor driving apparatus with high accuracy, high efficiency, low noise, and high response as well as downsizing of the apparatus.

  In addition, since current detection accuracy can be ensured, it is possible to realize a highly responsive motor drive using vector control. In particular, even in sensorless drive near zero speed, which has been considered difficult so far, harmonics can be increased. And an ideal motor drive device can be realized.

It is a block diagram showing the structure of the AC motor drive device in connection with the 1st Embodiment of this invention. It is a figure which shows the waveform of the principal part of the alternating current motor drive device in connection with the 1st Embodiment of this invention. It is the figure which showed the detail of the waveform of FIG. It is a figure which shows the motion of the detection electric current I0 concerning the 1st Embodiment of this invention. It is a block diagram showing the structure of the current estimator in connection with the 1st Embodiment of this invention. It is a figure showing the voltage command and electric current detection block of the AC motor drive device in connection with the 2nd Embodiment of this invention. It is a figure which shows the waveform of the principal part of the alternating current motor drive device in connection with the 2nd Embodiment of this invention. It is a figure which shows the operation state of the inverter main circuit in connection with the 2nd Embodiment of this invention. It is a figure showing the voltage command and electric current detection block of the alternating current motor drive device in connection with the 3rd Embodiment of this invention. It is a figure which shows the waveform of the shunt electric current of the alternating current motor drive device in connection with the 3rd Embodiment of this invention. It is a figure showing the voltage command and electric current detection block of the AC motor drive device in connection with the 4th Embodiment of this invention. It is a figure which shows the waveform of the principal part in the case of voltage correction "none" of the AC motor drive device in connection with the 4th Embodiment of this invention. It is a figure which shows the waveform of the principal part in case there is no voltage correction "with" of the AC motor drive device in connection with the 4th Embodiment of this invention. It is a figure of the waveform which shows the voltage correction state of the AC motor drive device in connection with the 4th Embodiment of this invention. It is a figure of the waveform which shows the voltage correction by the other method of the alternating current motor drive device in connection with the 4th Embodiment of this invention. It is a figure showing the voltage command and electric current detection block of the alternating current motor drive unit in connection with the 5th Embodiment of this invention. It is a figure showing the other example of the voltage command and electric current detection block of the alternating current motor drive device in connection with the 5th Embodiment of this invention. It is a figure showing the electric current estimator of the alternating current motor drive device in connection with the 6th Embodiment of this invention. It is a figure showing the voltage command and electric current detection block of the alternating current motor drive device in connection with the 7th Embodiment of this invention, and a motor constant measuring device. It is a figure which shows the waveform of the principal part of the alternating current motor drive device in connection with the 7th Embodiment of this invention. It is a block diagram of the prior art example of the current detection circuit which uses three shunt resistors. It is a block diagram of the prior art example of the current detection circuit which uses one shunt resistor.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Speed command generator, 2 ... Controller, 3 ... Inverter, 4 ... DC power supply, 5 ... Three-phase induction motor, 6a ... 1st shunt current detector, 6b ... 2nd shunt current detector, 7 ... Overcurrent detector, 8a to 8d ... adder / subtractor, 9 ... speed controller, 10 ... Id * generator, 11 ... d-axis current controller, 12 ... q-axis current controller, 13 ... speed estimator, 14 ... slip Calculation unit, 15 ... Electrical angle phase calculation unit, 16 ... Voltage command / current detection block, 17 ... dq coordinate converter, 18 ... dq inverse converter, 19 ... PWM generator, 20 ... Current estimator, 31 ... Main inverter Circuit 32 ... Output pre-driver 41 ... Commercial AC power supply 42 ... Rectifier 43 ... Smoothing capacitor.

  Embodiments of the present invention will be described below with reference to the drawings.

(First embodiment)
FIG. 1 is a block diagram showing the configuration of an AC motor driving apparatus according to the first embodiment of the present invention.

  This device is intended for driving an AC motor such as a three-phase induction motor. Broadly speaking, this apparatus includes a speed command generator 1, a controller 2, an inverter 3, an inverter power supply 4, and a three-phase induction motor 5 (hereinafter abbreviated as an induction machine) to be driven. The The inverter 3 includes an inverter main circuit 31, an output pre-driver 32, and first and second shunt current detectors 6a and 6b.

  The speed command generator 1 is a means for giving a rotational speed command ωr * of the induction machine 5 and is positioned above the controller 2.

  The controller 2 calculates the applied voltage to the induction machine 5 so that the actual rotational speed of the induction machine 5 matches the rotational speed command ωr *, and further performs pulse width modulation to perform main control of the inverter. A gate signal for switching the circuit is output. The controller 2 includes a speed controller (ASR) 9, a d-axis current command (Id *) generator 10, a d-axis current controller (IdACR) 11, a q-axis current controller (IqACR) 12, and a speed estimator 13. , A slip calculator 14, an electrical angle phase calculator 15, and a voltage command / current detection block 16. The voltage command / current detection block 16 is a feature of the present invention, and details will be described later. Reference numerals 8a to 8d denote adder / subtracters.

  The power source 4 includes a rectifier 42 that rectifies the commercial AC power source 41 and a smoothing capacitor 43 that smoothes voltage ripples of the rectified waveform. Further, an overcurrent detector 7 is provided for observing the values of the first and second shunt current detectors 6a and 6b, and immediately shutting off the gate signal and protecting the inverter when an excessive current is generated.

  Next, the operation of each component will be described.

  The speed command ωr * given from the speed command generator 1 is calculated by the adder / subtractor 8a to calculate the deviation from the speed estimated value ωrc. Based on this, the speed controller (ASR) 9 uses the torque current command Iq *. Is calculated. Usually, the necessary current command Iq * is generated while observing the actual speed ωr so that the rotational speed ωr of the induction machine 5 coincides with the command value ωr *. This embodiment is speed sensorless, and the speed estimator 13 estimates and calculates the rotational speed of the induction machine 5. The output ωrc is treated as equivalent to the actual speed and the torque current command Iq * is calculated. ing. Of course, it is needless to say that the present invention can also be applied to the case where the result of measuring the rotational speed using a speed detector is used instead of ωrc.

  The adder / subtractor 8c calculates the deviation of the torque current command Iq * from the actual induction motor torque current component Iqc, and the d current controller (IqACR) 12 calculates the q axis. Voltage command Vqc * is calculated.

  The d-axis current command Id * generator 10 generates a d-axis current command Id * corresponding to the excitation current of the induction machine 5. The current command Id * is calculated by the adder / subtractor 8b as a deviation from the actual exciting current component Idc, and from the calculated value, the d-axis current controller (IdACR) 11 calculates the d-axis voltage command Vdc of the induction machine. * Is calculated. These d-axis current controller 11 and q-axis current controller 12 are configured by general proportional integral control.

  The speed estimator 13 estimates the rotational speed of the induction machine 5. In recent years, “speed sensorless vector control” that performs vector control of an induction machine without using a speed sensor such as a pulse encoder has been widely used, and this embodiment also assumes this. The speed estimation method will not be described in detail because any known method is adopted. In the case of induction machines, it is necessary to control the slip frequency appropriately according to the load condition. In the present embodiment, the slip calculator 14 calculates the slip frequency ωs according to Id * and Iq *, adds it to the estimated rotational speed ωrc, and calculates ω1 that is the driving frequency of the induction machine 5. doing.

  In the vector control, control is performed by observing the voltage and current of the induction machine 5 on the rotation coordinate axis with reference to the secondary magnetic flux of the induction machine 5. The phase angle θdc of the rotational coordinate axis can be obtained by integrating the drive frequency ω1 with an integrator. This calculation is performed by the electrical angle phase calculator 15.

  The voltage command / current detection block 16, which is a characteristic part of the present invention, includes a dq coordinate converter 17, a dq inverse converter 18, a PWM generator 19, and a current estimator 20. The dq coordinate converter 17 converts the current (AC amount) of the induction machine 5 into an excitation current component (d-axis current) Idc and a torque current component (q-axis current) Iqc on the dq coordinate axis. The dq inverse converter 18 converts the voltage commands Vdc *, Vqc * into voltage commands Vu *, Vv *, Vw * on three-phase AC coordinates. The PWM generator 19 generates a pulse width modulation signal for switching the inverter 3 based on the Vu *, Vv *, and Vw *. Finally, the current estimator 20 reads the values of the shunt currents IDCP and IDCN, and calculates “amplitude” and “phase”, which are the current feature amounts of the induction machine 5.

  The PWM generator 19 outputs a PWM (Pulse Width Modulation) signal for controlling the output pre-driver 32 and turns the inverter main circuit 31 on and off. A trigger signal CP for sampling the shunt current is sent from the PWM generator to the current estimator 20.

  The DC power supply 4 is a DC power supply that supplies energy to the inverter 3 as described above.

  The inverter main circuit unit 31 is composed of six switching elements Sup to Swn. The current detector includes a first shunt resistor 6a that detects the sum of all currents flowing through the lower switching elements of each phase, and a first detector that similarly detects the sum of currents flowing through the freewheeling diodes of each phase. 2 shunt resistors 6b. These detection signals are input to the controller 2 as analog signals.

  Next, the basic operation of this motor drive system will be described. The present embodiment is based on induction machine vector control technology. The principle of the vector control technique is to independently control the current Iq contributing to the torque and the current Id contributing to the magnetic flux on the rotation coordinate axis (dq coordinate axis) based on the secondary magnetic flux linked to the rotor. It is a technique. The d-axis current controller 11, the q-axis current controller 12, the dq converter 17, the dq inverse converter 18 and the like in FIG. 1 are main parts for realizing this vector control technique.

  Next, the voltage command / current detection block 16 which is a characteristic part of the present invention will be described with reference to FIGS.

  FIG. 2 is a diagram showing waveforms of main parts of the AC motor driving device according to the first embodiment of the present invention. 2, (a) is a phase current waveform flowing through the induction machine 5, (b) is a phase voltage command waveform to the induction machine 5, and (c) is used for pulse width modulation by the PWM generator 19. It is a triangular wave carrier waveform. Regarding the current to be detected, (d) is a DC bus current IDC0, and is a combined waveform of the first shunt current IDCP in (e) and the second shunt current IDCN in (f).

  In IDCP and IDCN, the current waveform of the phase having the largest absolute value appears as a dotted line in an arc shape. These arc-shaped waveforms change with a period of 60 degrees in electrical angle. Further, IDCP is a waveform only on the positive side, and IDCN is a waveform only on the negative side. When both are added, IDC0 is obtained. These enlarged waveforms are shown in FIG.

  In FIG. 3, the waveform of FIG. 2 is enlarged and, at the same time, pulse width modulated phase voltage waveforms Vu, Vv, Vw and line voltage waveforms Vuv, Vvw, Vwu are added.

  As shown in FIG. 3B, the pulse width modulation is generated by comparing the triangular wave carrier with the phase voltage commands Vu *, Vv *, Vw *, and the phase voltages Vu, Vv, Vw shown in FIG. Output from the inverter 3.

  IDCP and IDCN can detect any phase current of Iu, Iv, and Iw under the condition that the three-phase voltages are all -VDC / 2 (period indicated by [A] in the figure). In the example of FIG. 3, the v-phase current Iv can be detected. Therefore, if IDCP or IDCN is sampled at this timing, one phase can be reliably detected. As a method of creating the timing, for example, as shown in FIG. 3, a triangular wave carrier peak signal (CP) may be used as a trigger.

  In Japanese Patent Laid-Open No. 6-153526 (Patent Document 2), which is a known example, the DC bus current IDC0 is sampled during the periods [B] and [C] in FIG. In this case, as is apparent from the figure, it is necessary to detect a current having a very narrow pulse width, and it is easily affected by ringing and noise. In particular, detection is difficult in the case of a thin pulse waveform as shown in FIG. However, in IDCP and IDCN, the pulse width can be sufficiently secured, so that detection can be performed reliably.

  Next, a method for obtaining the amplitude and phase of the current flowing through the induction machine 5 from IDCP and IDCN will be described with reference to FIG.

  FIG. 4 shows the phase voltage command Vu * and the waveform of the detection current I0 of the IDCP (or no problem even if IDCN is used) (waveform sampled at the carrier peak) under different conditions. I1 in the figure is the amplitude (peak value) of the current, and δ in the figure is the phase angle of the current. Under the condition of power factor 1 ((b) in the figure), it can be seen that the waveform near the peak of the current detection waveform I0 completely overlaps with the peak of the phase voltage command. That is, it can be seen that the current amplitude and phase information can be obtained by searching for the peak of the current detection value I0. In the case of an induction machine, since the power factor rarely becomes 1, it is assumed that it is a delayed phase.

  In the induction machine, in the no-load state ((c) in the figure), the excitation current is dominant, so the current is delayed by approximately 90 degrees with respect to the voltage. For this reason, the peak of the current I0 is in the vicinity of the second half of the period every 60 degrees. As the load increases, the power factor gradually increases and the peak phase advances. From the peak phase angle, the current phase is calculated by calculating backward.

  FIG. 5 shows a configuration diagram of the current estimator 20 that performs these operations. As shown in the figure, the shunt current IDCP and IDCN are sampled by the trigger CP from the PWM generator 19 in the current sample holders 202a and 202b. Next, the current detection value I0 is selected in the switch 201 for selecting either IDCP or IDCN. Both show the same value, but the pulse width varies depending on the electrical characteristics and operating state of the induction machine. Which one to select can be set according to the operating conditions of the induction machine, etc. with the wider pulse width. Or you may make it switch according to conditions.

  I0 is searched by the peak searcher 203, and the amplitude I1 and the phase δ are calculated. Since I1 and δ are polar coordinate systems, they are converted by the dq coordinate converter 17 into Idc and Iqc which are coordinates on the dq axis.

  As described above, according to the present embodiment, it is possible to calculate the amplitude and phase of the phase current of the induction machine from the detection current of the shunt current, thereby enabling the induction machine to be driven with high-performance vector control. It becomes.

(Second Embodiment)
Next, a second embodiment of the present invention will be described.

  In the first embodiment, the shunt current is sampled under the condition that the output phase voltage of the inverter is all negative (−VDC / 2), and the peak value of the waveform is searched for, so that the phase current of the induction machine Amplitude and phase information were obtained. In the first embodiment, vector control with the simplest configuration is possible, but since only current information with an electrical angle of 60 degrees can be obtained, there may be a problem in responsiveness.

  Since vector control is one of the merits of being able to increase the response of torque control, there remains a problem in that respect. In the second embodiment, this problem is solved.

  FIG. 6 shows a voltage command / current detection block 16B in the second embodiment. The second embodiment can be realized by using this voltage command / current detection block 16B instead of the voltage command / current detection block 16 in FIG.

  In FIG. 6, parts having the same numbers as those in FIGS. 1 and 5 represent the same parts. The newly added component includes a current estimator 20B, a current phase calculator 21 that calculates a current phase θi from current values Idc and Iqc on the dq axis, and a current phase Io from the current phase θi. There is an I0 phase selector 22 for discrimination. The current estimator 20B has an adder / subtractor 8e for calculating the sum of IDCP and IDCN, sample holders 202c and 202d for sample-holding the added value, and timing setting for giving hold timings T1 and T2 to these sample holders. A container 204 is provided. Furthermore, a phase current reproducer 205 that reproduces the phase currents Iu, Iv, and Iw from the outputs Ia1 and Ia2 of the sample holders 202c and 202d and the current value I0 output from the switch 201 is provided. The dq coordinate converter 17B is a dq coordinate converter that converts a three-phase AC signal (here, Iu, Iv, Iw) into a quantity on the dq coordinate axis.

  In FIG. 6, the current can be reproduced by a conventional single shunt resistor by using a newly added component (for example, JP-A-6-153526 (Patent Document 2)). In the timing setting unit 204, the timing at which IDC0 is sampled is set based on the magnitude relationship between the voltage commands Vu *, Vv *, and Vw *. In the normal PWM method, a two-phase current can be detected among the three-phase currents.

  FIG. 7 shows the timing T1, T2 and the state of the shunt current. Under the conditions in the figure, there are four types of switching modes 1 to 4, and a current appears in IDC0 in modes 2 and 3 and becomes zero in modes 1 and 4. IDCP and IDCN are zero in mode 1, and in mode 4, the same Iv is observed for both. These states are shown in FIG.

  As shown in FIG. 8A, in mode 1, all the upper switches of each phase are turned on, and the terminal voltage of the induction machine 5 becomes zero. At this time, the inverter 3 and the induction machine 5 are in a disconnected state. The current flowing through the induction machine is short-circuited via the upper switching element and diode, and no current flows through the shunt resistor. In mode 2 shown in FIG. 5B, only the V-phase is turned on in the lower arm, so that the V-phase current flows as the shunt current. If the V-phase current is “positive”, a current flows in the IDCN, and if it is “negative”, a current is generated in the IDCP. In mode 3 ((c) in the figure), the U-phase current flows through the shunt current in the same manner as in mode 2.

  In the state of mode 4 in FIG. 4D, the phase current of the induction machine 5 is short-circuited through the lower arm, and the same current (reverse direction) flows through the IDCP and IDCN at the same time.

  Modes 1 and 4 are states in which the output voltage is zero, and the period becomes longer in a region where the output voltage of the inverter 3 is small (condition that the drive frequency of the induction machine is low). Conversely, the periods of modes 2 and 3 are shortened. As described in the first embodiment, in order to reproduce the phase current from the waveform of IDC0, it is necessary to sample a very narrow pulse current, which is easily affected by ringing and noise. For example, the period of mode 2 in FIG. 7 is very short, and the setting of T1 is at the limit. It can be seen that the V-phase current values can be detected as I0 and Ia1 under the conditions shown in the figure. However, the detected value of Ia1 is obtained by sampling a minute pulse, and is a condition that is difficult to obtain accuracy. It can be seen that the V-phase current Iv can be detected more stably by using the value from I0 than from the IDC0. Therefore, the phase current reproducer 205 preferentially uses the current detected as I0, and selects the remaining phases from Ia1 and Ia2. In the case of FIG. 7, Ia2 is selected and detected as Iu.

  Note that the phase current reproducer 205 must know which phase current I0 is. For this purpose, a current phase calculator 21 and an I0 phase selector 22 are used. The current phase calculator 21 performs the following calculation using the past current values Idc and Iqc.

θi = θdc + atan (Iqc / Idc) (1)
Here, θdc is a reference phase inside the controller, and atan is an arctangent calculation.

  From the above equation, the instantaneous phase θi of the current of the induction machine 5 is obtained. If θi is obtained, it can be determined which of the three-phase currents is generated in IDCP and IDCN. The conditional expression is as follows.

(I) −30 ≦ θi ≦ 30 [deg] → Iu
(Ii) 30 ≦ θi ≦ 90 [deg] → −Iw
(Iii) 90 ≦ θi ≦ 150 [deg] → Iv
(Iv) 150 ≦ θi ≦ 210 [deg] → −Iu
(V) 210 ≦ θi ≦ 270 [deg] → Iw
(Vi) 270 ≦ θi ≦ 330 [deg] → −Iv (2)
These are stored in the I0 phase selector 22 as table data. According to the above, it is possible to instantaneously determine which phase current is the current sampled as I0. This information is given as a signal Picp to the phase current reproducer 205, and is used preferentially when performing phase current reproduction.

  According to the present embodiment, more accurate current detection can be realized, and it is possible to contribute to high response of vector control while being a simple circuit.

  Note that, in FIG. 6 and the embodiments after this embodiment, when obtaining IDC0, IDCP and IDCN are combined as analog values, and then sampled and held. However, it is also possible to add both after sample-holding and taking them into the controller as digital quantities. When actually using a microcomputer, such processing is performed. However, for convenience of description of the embodiment, the processing order is simply reversed. In operation, both are equivalent.

(Third embodiment)
Next, a third embodiment of the present invention will be described.

  The second embodiment combines both the function of estimating which phase the detected current I0 is, and the function of reproducing the phase current from the IDC0 by the conventional method. As described above, it is sometimes difficult to reproduce the phase current with the conventional technology. In the present embodiment, an example in which current detection is possible with a simpler configuration by considering the conditions that are difficult to reproduce in advance will be described.

  FIG. 9 shows a voltage command / current detection block 16C in the third embodiment. The third embodiment can be realized by using the voltage command / current detection block 16C in place of the voltage command / current detection block 16 in FIG.

  9, parts having the same numbers as those in FIGS. 1, 5, and 6 represent the same parts. The newly added parts are a timing setter 204C and a phase current reproducer 205C, and the sample holder 202D in FIG. 6 is deleted, and IDC0 is sampled at only one timing.

  FIG. 10 shows a shunt current waveform that simulates an actual waveform. When the induction machine 5 is actually driven by the inverter 3, the “ringing” as shown in the figure always occurs every time the switching element performs the switching operation. The magnitude (amplitude), frequency, duration, etc. of this ringing depends on the characteristics of the inverter and induction machine, installation conditions, cable length, etc., and it is difficult to take a general measure. Generally, the longer the cable connecting the inverter and the induction machine, the longer the ringing duration.

  When ringing occurs as shown in FIG. 10, for example, it is difficult to sample the current in a narrow pulse width period ([A] in the figure). Under such conditions, “ringing” is detected, and the accuracy is significantly degraded.

  According to this embodiment, current information of at least one phase can be obtained by detecting IDCP or IDCN, and therefore it is not necessary to detect the phase current under conditions that are difficult to detect. It is only necessary to select the one with the better condition, that is, the one with the wider pulse width and read the current. In FIG. 10, only a current having a wide pulse width (in this case, Iu) is sampled at the timing of T1, and used for phase current reproduction.

  In these operations, the signal Picp indicating which phase is I0 is given to the timing setting unit 204C, and the timing T1 is set so as to sample other phases. The timing may be set once in the carrier up or down cycle. Therefore, it is possible to detect the current of the induction machine 5 with one sample holder and a simple algorithm.

(Fourth embodiment)
Next, a fourth embodiment of the present invention will be described.

  In the third embodiment, an example is shown in which current reading from IDC0 is sufficient only once in a half cycle of the carrier.

  However, there is a limit even when IDC0 is sampled only once. For example, when the cable extends from the inverter 3 to the induction machine 5 as much as 100 m, the duration time of the ringing reaches 10 μs. The triangular wave carrier may be increased to about 20 kHz, and the influence of the ringing of several tens of μs is large with respect to the carrier period (50 μs), and it is difficult to cope with the third embodiment.

  Therefore, this problem is solved by correcting the voltage command itself and adding a means for forcibly expanding the pulse width.

  FIG. 11 shows a voltage command / current detection block 16D in the fourth embodiment. The fourth embodiment can be realized by using this voltage command / current detection block 16D instead of the voltage command / current detection block 16 in FIG.

  In FIG. 11, parts having the same numbers as those in FIGS. 1, 5, 6 and 9 represent the same parts. The newly added parts are a voltage corrector 23 and adder / subtracters 8f to 8h. The voltage corrector 23 corrects the voltage command by adding correction voltages dVu, dVv, dVw to the three-phase voltage commands Vu *, Vv *, Vw *. This correction is performed with the minimum unit of the carrier cycle as a reference (in the case of triangular wave comparison, the monotonic increase or monotonic decrease of the carrier is the minimum unit), and the multiples are corrected as one voltage command change cycle. . Within this voltage command change period, the correction voltages of each phase are calculated so that the average becomes zero, thereby reducing the influence on the original voltage command.

  FIG. 12 shows a voltage command waveform, a phase voltage waveform, and an IDC0 waveform before correction, and FIG. 13 shows each waveform after correction. In these waveforms, the duration of ringing occurring in IDC0 is shown as being long.

  As shown in FIG. 12, before the voltage correction is performed, since the ringing is long, a part of the ringing is detected even at the timing of T1. In contrast, FIG. 13 shows the correction of the voltage command so as to widen the pulse width. In the figure, the correction amount of only Vu * is added to widen only the width of the detected pulse. The voltage is corrected in the monotonic increase period of the carrier, and in the next monotonic decrease period (denoted as TDOWN in the figure), the corrected amount is subtracted so that the average becomes zero. As a result, current detection at the timing when the ringing is settled can be realized. Further, when this method is used, the current detection is performed intermittently, and as shown in the figure, the current is detected at the TUP cycle, and is not detected at the TDOWN cycle.

  This operation is shown in FIG. A voltage applied for current detection is referred to as “detection correction voltage”, and a voltage for canceling the detection correction voltage is referred to as “correction voltage”. Current detection is performed based on a current detection flag. In this example, current detection is always performed only during the monotonically increasing period of carriers.

  Such correction of the pulse width modulation signal is also known in the prior art (for example, Japanese Patent Application Laid-Open No. 2008-131770 (Patent Document 3)). In the conventional method, when conditions become severe, it becomes necessary to correct two phases of the three-phase voltage. In this case, a large pulsation component is generated in the current, and a detection error may occur due to correction. Alternatively, losses such as electromagnetic noise and motor iron loss increase as the amount of harmonics increases.

  According to the present embodiment, only one phase of the three-phase voltage command needs to be corrected even under the worst conditions. In FIG. 11, correction amounts are added to all of Vu *, Vv *, and Vw *, but in practice, only one phase is corrected, and the other correction amounts are given as zero. Therefore, the amount of harmonics generated is less than before, and the increase in detection error and loss can be reliably improved.

  In Japanese Patent Laid-Open No. 2008-131770 (Patent Document 3), a monotonically increasing triangular wave carrier and an odd number of monotonically decreasing periods are used as voltage command change periods to improve detection accuracy and suppress harmonic generation. Is going. This method can also be introduced in this embodiment. This is shown in FIG.

  The operation principle is described in Japanese Patent Application Laid-Open No. 2008-131770 (Patent Document 3), which is a well-known example, and is omitted. This known example is characterized in that when a correction voltage is applied, the correction voltage is divided and added before and after the detection correction voltage. As a result, the current detection flag appears alternately when the carrier monotonously increases and when the carrier monotonously decreases. This maintains the symmetry of the current waveform, which suppresses the increase in harmonics and improves detection accuracy. In this correction method, if conditions become severe, it is necessary to correct two phases out of three-phase voltage commands. However, if it is introduced into the present invention, as described above, it is only necessary to perform correction for one phase, so that accuracy can be further improved and loss can be reduced as compared with known examples.

(Fifth embodiment)
Next, a fifth embodiment of the present invention will be described.

  In the fourth embodiment, when current detection of IDC0 is difficult, current detection is performed by correcting the voltage command and changing the pulse width. However, changing the voltage command will cause waveform distortion. For example, in a system that does not require much transient response characteristics but wants to minimize loss only, voltage correction is not desired. The fifth embodiment is an example for solving these problems.

  FIG. 16 shows a voltage command / current detection block 16E in the fifth embodiment. The fifth embodiment can be realized by using this voltage command / current detection block 16E instead of the voltage command / current detection block 16 in FIG.

  In FIG. 16, parts having the same numbers as those in FIGS. 1, 5, 6, 9, and 11 represent the same parts. The newly added parts to FIG. 11 are a dq inverse coordinate converter 18b and a current changeover switch 24.

  The dq inverse coordinate converter 18b receives Idc and Iqc as calculation results and converts them into three-phase currents Iuc, Ivc and Iwc. This process is performed based on the calculation result one time before the cycle in which the current detection / calculation is performed. The current changeover switch 24 is a switch that determines whether to select one of the three-phase current obtained by reproducing the phase current and the three-phase currents Iuc, Ivc, and Iwc. If the switch is “1”, the reproduction current is selected, and if it is “0”, the three-phase current (Iuc, Ivc, Iwc) calculated backward from the previous result is selected. This selection is made for each phase.

  In the case where the pulse width is short and detection from IDC0 is difficult, in this embodiment, reproduction processing from Ia is not performed. The voltage corrector 23 always outputs a zero correction voltage. Instead, I0 can be reliably detected and is used for current reproduction. However, since at least the amplitude and phase of the phase current of the induction machine 5 cannot be specified without the information on the other phase, any one of Iuc, Ivc, and Iwc, which are past values, is used here. Calculate the amount that cannot be reproduced by borrowing from past values. This method is a mere “delay” in the control operation and affects the control system. However, since I0 is used as a detection value, the control system is unlikely to become unstable. At least the state where the current is maximized by I0 is managed.

  When IDC0 cannot be detected, FIG. 16 shows a method of substituting from past detection values. However, as shown in FIG. 17, current commands Idc * and Iqc * are used instead of Idc and Iqc. Also good. In this case, since it is not a delay factor, the stability is rather improved.

(Sixth embodiment)
Next, a sixth embodiment of the present invention will be described.

  In a preferred embodiment of the present invention, two shunt resistors are used as a basic configuration, like the shunt resistors 6a and 6b in FIG. The shunt resistor is usually a low resistance of several tens of mΩ to several hundreds of mΩ, and the error is large. Furthermore, since it is a resistor, the value changes depending on the temperature.

  In the case of using a plurality of shunt resistors, it is their variation that is particularly problematic. In the conventional example (Japanese Patent Laid-Open No. 63-80774 (Patent Document 1)) in which each arm is provided with a shunt resistor, there is a high possibility that torque pulsation will occur due to three-phase imbalance.

  In the embodiment of the present invention, there is a possibility that the influence of the variation appears depending on the direction of the current. The present embodiment relates to a method for correcting the variation.

  FIG. 18 shows a current estimator 20G in the sixth embodiment. The sixth embodiment can be realized by using the current estimator 20G in the voltage command / current detection block in the previous embodiments.

  In FIG. 18, a gain setting unit 206 and variable gain amplifiers 207a and 207b are added. The gain setting unit 206 inputs the outputs of the sample holders 202a and 202b and calculates the deviation between them. This calculation is performed during the period when the output voltage of the inverter 3 is negative (−VDC / 2) in all three phases. During this period, the absolute values of the currents flowing through the two should be the same, except for the signs. If there is a difference between the two, it can be considered that this is due to variations in the shunt resistors 6a and 6b.

  Either the variable gain amplifier 207a or 207b is corrected so as to eliminate this variation. As a result, variations in shunt resistance are corrected, and the occurrence of torque pulsation is suppressed. In FIG. 18, a variable gain amplifier is introduced as an analog amplifier. However, in actuality, it is more difficult to input a digital quantity with an AD converter and then perform gain correction as software processing. Is advantageous.

(Seventh embodiment)
Next, a seventh embodiment of the present invention will be described.

  In a drive device for an AC motor, “vector control” that can handle an AC motor in the same way as a DC motor is often used. In this vector control, it is particularly necessary to match the electric constant of the motor. Already, high-end models of general-purpose inverters are equipped with an “auto-tuning function” that automatically adjusts this electrical constant. This auto-tuning function uses the inverter 3 to apply a single-phase voltage to the induction machine 5 and measures the inductance and resistance of the induction machine from the alternating current that flows. The single phase is used to fix the rotor of the induction machine 5. With such single-phase excitation, the leakage inductance and winding resistance of the motor can be calculated. Usually, since current detection accuracy is required, it is generally applied to an inverter equipped with a current sensor (DCCT).

  In the present embodiment, an auto-tuning function by the shunt current detection means that is a feature of the present invention is provided.

  FIG. 19 shows a voltage command / current detection block 16H and a motor constant measuring instrument 25 in the seventh embodiment. For example, if these are introduced into the third embodiment shown in FIG. 9, the seventh embodiment can be realized. Of course, the present invention can be applied to other embodiments.

  In FIG. 19, a motor constant measuring device 25 functions when the auto-tuning function is operated, and creates a voltage command for measuring the electrical constant of the induction machine 5. The voltage changeover switch 26 can be switched between normal operation and auto-tuning operation by each switch. Normally, all switches are set to the “1” side, but are switched to “0” during auto-tuning, and the output of the motor constant measuring device 25 is applied to the induction machine 5 via the inverter 3. As a result, a current flows through the induction machine 5, the shunt current values IDCP and IDCN described so far are sampled at the triangular wave carrier peak, and one of the values I0 is read. The motor constant measuring device 25 measures the motor constant from the relationship with the applied voltage using this value I0.

  The operation of FIG. 19 will be described using the waveform of FIG.

  During auto-tuning operation, a single-phase voltage is applied to generate an alternating magnetic field instead of the usual three-phase balanced voltage. As shown in FIG. 20A, a voltage is applied from Vu to Vv, Vw. Vv and Vw have the same waveform, both of which are in phase with Vu and have half the amplitude.

  As a result, an alternating current as shown in FIG. The phase difference from the voltage command is caused by the leakage inductance of the induction machine 5. The winding resistance and inductance of the induction machine 5 can be calculated from the phase and amplitude of the current.

  At this time, the shunt currents are as shown in FIGS. Since the AC motor reaches the rated current with a slight applied voltage in the stopped state, IDC0 is a set of pulse currents with a very narrow width as shown in the figure. On the contrary, IDCP and IDCN have a wide energization width, making it easier to read current. When conventional one-shunt current detection is used, the waveform shown in FIG. 20D is obtained, and voltage correction as shown in the fourth embodiment is required. However, voltage correction may increase the detection error and is not suitable for the auto-tuning function, and the tuning accuracy is likely to deteriorate.

  In the present embodiment, IDCP or IDCN in FIGS. 5E and 5F is introduced without using IDC0 for auto-tuning. When IDCP or IDCN is sampled at the peak of the triangular wave carrier, an arc-shaped current waveform obtained by full-wave rectification of the phase current Iu is obtained as shown by the dotted lines in FIGS. From this waveform, the motor constant can be calculated with high accuracy.

  Specifically, it can be obtained by searching for the peak of the arc-shaped waveform.

  It is also possible to calculate the current amplitude and phase by Fourier series expansion or Fourier transform. 20 (e) and 20 (f) contain the largest number of harmonics twice the fundamental frequency, and the components may be obtained by Fourier series expansion. FIG. 20 (g) shows a waveform of a reference double frequency function. These are multiplied by the arc-shaped component detected in (e) or (f) of the same figure, and the average value is calculated to obtain the content of the double component. If the relationship between the content rate and the original current waveform is obtained in advance, the phase and amplitude of the current can be calculated, and the inductance and resistance of the induction machine 5 can be derived from the result.

  The embodiment according to the present invention has been described above. For the sake of explanation, the induction motor is used as the AC motor. However, all embodiments can be applied to other AC motors such as a permanent magnet type synchronous motor, a wound type synchronous motor, and other AC motors. is there. The calculation of the current estimation part can be applied to any AC motor. Moreover, it can be applied not only to electric motors but also to power converters that handle three-phase AC.

  Although the shunt resistor is connected to the lower arm side (minus side) of the main circuit, there is no problem if a similar circuit is attached to the upper arm side. The present invention is equally applicable.

  The controller according to the present invention can be realized by a microcomputer, and each component used in the description can be realized as logic or software inside the microcomputer.

  As described above, the present invention is a technology for making an AC motor drive device a compact, high-precision, low-loss system. The range of application of this device can be used for rotational speed control of industrial general-purpose inverters, fans, pumps, compressors, elevators, conveyors, etc. using them.

Claims (10)

An AC motor drive device comprising: an inverter that outputs a sinusoidal AC current; a three-phase AC motor connected to the inverter; and a controller that outputs a pulse width modulation signal to the inverter and controls the inverter. There,
First current detecting means for detecting a combined current of each phase switching element of the upper arm or the lower arm of the inverter main circuit, and a first current detecting means for detecting a combined current of a diode connected in antiparallel to each of the phase switching elements. At least one of the two current detection means;
The first, using at least one of the current detection value of the second current detection means, the amplitude and phase of the alternating current flowing in the three-phase AC motor and a estimation calculation section for estimating,
The estimation calculation means is performed for each period of the pulse width modulation signal,
In the period in which the output line voltage of the inverter is zero, that is, in the period in which all the switching elements of the upper arm of each of the three phases are on, or in which the switching elements of the lower arm of each of the three phases are all on An AC motor driving apparatus characterized in that the amplitude and phase of an AC current flowing through the three-phase AC motor are estimated and calculated using a current value detected by a current detection means and / or a second current detection means . In claim 1 ,
In the period in which the output line voltage of the inverter is zero, that is, in the period in which all the switching elements of the upper arm of each of the three phases are turned on or all the switching elements of the lower arm of each of the three phases are turned on The detected current value I0 of the current detecting means and / or the second current detecting means is read into the controller,
Further, in the period when the output line voltage of the inverter is not zero, the current detection value Ia obtained by combining the detection current values of the first and second current detection means is read into the controller,
The estimating calculation means, based on said two current detection values I0 and Ia, three-phase alternating current of an amplitude and phase estimation calculation AC motor driving device which is characterized in that. In claim 2 , two or more types of voltages with non-zero output line voltage of the inverter are continuously output, and one type of voltage is selected from the two types of output line voltages. An AC motor drive device that detects the current detection value Ia during a voltage output period. In any one of claims 1-3, wherein the pulse width modulated signal, three-phase AC signal and, using those generated by comparing a triangular wave carrier signal, the triangular wave carrier is monotonically increasing or monotonically decreasing to become period Is regarded as a unit period, and a plurality of the unit periods are defined as voltage command change periods
A correction signal such that the average value of the correction amount in the voltage command change period becomes zero or substantially zero is added to at least one phase of the three-phase AC signal, and a correction result in the voltage command change period is at least 1 A drive device for an AC motor, wherein two line voltage pulse widths are secured to a predetermined amount or more. 5. The AC motor drive device according to claim 4 , wherein an odd number of 3 or more is used as the voltage command change period as the number of triangular wave carrier unit periods in the voltage command change period. In any one of claims 1 to 5, pouring the estimation calculation of the current amplitude and phase in the controller, the detected current detected value in the past of the three-phase AC motor, the current estimated value or the three-phase AC motor using either the current command value, a current amplitude of current, the AC motor and estimating the position phase drive. In any one of claims 1 to 6, when the output voltage of the inverter becomes zero, i.e., the switching elements of the upper arm of the three-phase, respectively are turned on, or, a three-phase switching elements of the respective lower arm all is on In a certain state, current detection gain correction means for reading the detection values of the first and second current detection means into the controller and adjusting the current detection gain so that the difference between the two current detection values becomes zero. A drive device for an AC motor, comprising: In claim 1,
First current detecting means for detecting a combined current of each phase switching element of the upper arm or the lower arm of the inverter main circuit, and a first current detecting means for detecting a combined current of a diode connected in antiparallel to each of the phase switching elements. At least one of the two current detection means;
The inverter outputs an alternating current to the three-phase AC motor, the alternating current is detected using a current detection value of at least one of the current detection means, and a resistance value and an inductance of the three-phase AC motor A drive apparatus for an AC motor, comprising constant measuring means for measuring an electrical constant including a value. The measurement of the electrical constant according to claim 8 , wherein an operation corresponding to Fourier series expansion or Fourier transform is used in the controller, and the operation has a period twice as long as the alternating current as a periodic function. AC motor drive device characterized by introducing a function. In claim 1,
A drive apparatus for an AC motor, wherein a carrier frequency for PWM control of the inverter is set to exceed 10 kHz, and vector control is applied to the three-phase AC motor.

Images