Power converter

Abstract

To output, in a modulation method for suppressing excitation of common-mode noise for which a region capable of modulating is limited, an identical voltage vector by using only a region capable of modulating a voltage vector outside the region capable of modulating, thereby suppressing the common-mode noise without changing the modulation method.SOLUTION: A power conversion device 1 includes an inverter circuit 27 and a control device 21 for controlling switching of a switching element. The control device 21 includes a modulation unit 50 which sets a partial region in a base voltage space, which is a voltage vector region that can be output in the inverter circuit 27, as a region capable of modulating, and a phase voltage command correction unit 40 for calculating a corrected voltage vector obtained by correcting the voltage vector to fall within the region capable of modulating when the voltage vector falls within the base voltage space and outside the region capable of modulating. As a result, the modulation unit 50 performs output by using the corrected voltage vector.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This invention relates to a power conversion device that converts a DC voltage to an AC voltage. [Background technology]

[0002] Conventionally, various pulse width modulation (PWM) methods have been proposed to suppress conducted noise propagating through power supplies, but these methods can be broadly divided into two categories. One is a method that completely suppresses fluctuations in common-mode voltage, which are a source of common-mode noise, and the other is a method that partially suppresses the noise while tolerating fluctuations in common-mode voltage.

[0003] One of the former methods is pulse width modulation, which outputs only odd voltage vectors or only even voltage vectors. This method makes it possible to completely suppress fluctuations in common-mode voltage within the carrier period. There is also pulse width modulation that switches between outputting only odd voltage vectors or only even voltage vectors depending on the electrical angular phase. This method can also significantly suppress fluctuations in common-mode voltage (see, for example, Patent Document 1).

[0004] One of the latter methods is pulse width modulation, in which the rising and falling edges of the phase voltages of other phases are synchronized with the rising and falling edges of the phase voltages of a specific phase in a PWM pattern (see, for example, Patent Document 2). Furthermore, common-mode voltage fluctuations can also be suppressed by pulse width modulation of two phases, in which the switching of one phase is fixed and the other two phases are switched (see, for example, Patent Document 3). [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Patent No. 5397448 [Patent Document 2] WO2019 / 180763 [Patent Document 3] Patent No. 5298003 [Overview of the project] [Problems that the invention aims to solve]

[0006] While the former method (Patent Document 1) is the most effective method for suppressing common-mode voltage fluctuations, it has the drawback of limiting the linear output region (the maximum amplitude at which the voltage vector can complete one rotation with a constant radius) and thus limiting the possible modulation ratio due to restrictions on the voltage vector used. Therefore, it is difficult to apply when driving a compressor motor, or it is necessary to switch the modulation method as in Patent Document 3 when the rotational speed and modulation ratio are high. However, switching the modulation method causes fluctuations in the common-mode voltage during the switch, and the common-mode voltage fluctuation suppression effect deteriorates in modulation methods with a wider linear output region.

[0007] In contrast, the latter method (Patent Documents 2 and 3) allows the linear output region to be used up to its normal maximum, and can achieve a high modulation rate, but its effect in suppressing common-mode voltage fluctuations is still inferior to that of the former method.

[0008] In Patent Document 3, two-phase modulation and three-phase modulation are switched depending on the operating region. It is conceivable to switch between the former and latter methods described above in a similar manner, but this would result in the problem of a switching shock occurring in the pulse width modulation method.

[0009] The present invention was made to solve the aforementioned conventional technical problems, and aims to provide a power conversion device that enables high modulation rate operation in pulse width modulation, which has a high noise suppression effect but a limited modulation rate. [Means for solving the problem]

[0010] The present invention relates to a power conversion device that converts a DC voltage to an AC voltage, comprising: an inverter circuit that applies phase voltages generated at the connection points of switching elements of each phase to a load; and a control device that controls the switching of the switching elements of the inverter circuit, wherein the control device comprises: a modulation unit that defines a moduloable region as a part of the basic voltage space, which is the outputtable voltage vector region of the inverter circuit; and a phase voltage command correction unit that calculates a corrected voltage vector by correcting the command voltage vector to be within the moduloable region when the command voltage vector is within the basic voltage space but outside the moduloable region, and the modulation unit outputs using the corrected voltage vector.

[0011] In relation to the power conversion device described above, the modulation unit may include a first modulation processing unit having a first moduloable region that is the moduloable region, and a second modulation processing unit having a second moduloable region that is the moduloable region and is a region different from the first moduloable region, and the phase voltage command correction unit may be characterized by correcting the correction voltage vector within either the first moduloable region or the second moduloable region.

[0012] In relation to the power conversion device described above, the phase voltage command correction unit may have a modulation region selection unit that selects either the first moduloable region or the second moduloable region as the target for correction of the command voltage vector, and the modulation region selection unit may be characterized by switching between the first moduloable region and the second moduloable region based on the phase of the command voltage vector.

[0013] In relation to the above-described power conversion device, the first modulation processing unit may be characterized by performing pulse width modulation that outputs only odd voltage vectors, and the second modulation processing unit may be characterized by performing pulse width modulation that outputs only even voltage vectors.

[0014] In relation to the power converter described above, the phase voltage command correction unit may have a modulation region selection unit that selects either the first moduloable region or the second moduloable region as the target for correction of the command voltage vector, and the modulation region selection unit may be characterized in that it does not perform switching between the first moduloable region and the second moduloable region immediately after performing switching between the first moduloable region and the second moduloable region.

[0015] In relation to the above-described power converter, when defining the phase that serves as the boundary for switching between the first moduloable region and the second moduloable region in the modulation region selection unit as the switching boundary phase, the modulation region selection unit may be characterized in that it does not perform the switching between the first moduloable region and the second moduloable region if the voltage vector passes through the switching boundary phase in the reverse rotation direction.

[0016] In relation to the above-described power conversion device, the modulation region selection unit may be characterized by predicting the future command voltage vector to estimate a continuous switching between the first moduloable region and the second moduloable region, and if, as a result of the estimation, it is estimated that a continuous switching will occur, the switching between the first moduloable region and the second moduloable region will not be performed when the future command voltage vector arrives.

[0017] In relation to the above-described power conversion device, the device may also include a command calculation unit that generates the command voltage vector, wherein the command calculation unit compensates for the error between the corrected voltage vector calculated by the phase voltage command correction unit and the command voltage vector before correction in subsequent calculations of the command voltage vector.

[0018] In relation to the power conversion device described above, the phase voltage command correction unit may be characterized by setting the correction voltage vector on the boundary line defining the moduloable region.

[0019] In relation to the power conversion device described above, the phase voltage command correction unit may be characterized by setting the relationship between the voltage vector and the corrected voltage vector such that their lengths are the same and their phases are different.

[0020] In relation to the power conversion device described above, the phase voltage command correction unit may be characterized by setting the relationship between the voltage vector and the corrected voltage vector such that their phases are the same and their lengths are different. [Effects of the Invention]

[0021] According to the present invention, since the modulation unit can perform modulation control after correcting a voltage vector that belongs outside the operable region to bring it within the operable region, a modulation method that always suppresses the excitation of common-mode noise becomes available, and an excellent effect of suppressing common-mode noise overall can be achieved. Furthermore, according to another invention related to the present invention, since the next output voltage vector is calculated taking into account the error between the corrected voltage vector and the voltage vector before correction, it is possible to output a voltage vector equivalent to the previous command voltage vector even when the command voltage vector is within the range that the inverter circuit can output but outside the operable region of the modulation unit. [Brief explanation of the drawing]

[0022] [Figure 1] This is an electrical circuit diagram of a power conversion device according to an embodiment of the present invention. [Figure 2] This diagram shows the command values ​​for three-phase AC voltage. [Figure 3] This diagram represents the fundamental voltage space used to explain the linear output region. [Figure 4] This diagram shows the relationship between voltage vectors and phase voltages. [Figure 5] This is a diagram showing voltage vectors (basic voltage vectors). [Figure 6] This is a flowchart illustrating the operation of the control device for the power converter. [Figure 7]This is a diagram illustrating the voltage space used to explain the output region of an odd-voltage RSPWM. [Figure 8] This figure shows the correlation between the output region and the function value. [Figure 9] This diagram illustrates the linear output region of odd-voltage RSPWM. [Figure 10] This figure shows the output vector and output time of an odd-voltage RSPWM. [Figure 11] This figure shows an example of the output vector of an odd-voltage RSPWM. [Figure 12] This figure shows the PWM pattern for odd voltage RSPWM. [Figure 13] This figure shows the modulation waveform at a low modulation rate for odd-voltage RSPWM. [Figure 14] This is a diagram illustrating the voltage space used to explain the output region of an even-voltage RSPWM. [Figure 15] This figure shows the correlation between the output region and the function value. [Figure 16] This diagram illustrates the linear output region of even-voltage RSPWM. [Figure 17] This figure shows the output vector and output time of an even-voltage RSPWM. [Figure 18] This figure shows an example of the output vector of an even-voltage RSPWM. [Figure 19] This figure shows an example of the output vector of an even-voltage RSPWM. [Figure 20] This figure shows the PWM pattern for even voltage RSPWM. [Figure 21] This figure shows the modulation waveform at a low modulation rate for even-voltage RSPWM. [Figure 22] This figure shows the operating range of the full-voltage RSPWM. [Figure 23] This diagram shows the applicable ranges of odd-voltage and even-voltage RSPWM in the phase-determined full-voltage RSPWM in voltage space. [Figure 24] This figure shows the phase ranges to which odd-voltage and even-voltage RSPWM apply in the phase-determining full-voltage RSPWM. [Figure 25]This figure shows the modulation waveform at a low modulation rate for full-voltage RSPWM. [Figure 26] This diagram shows a voltage vector correction method represented in voltage space. [Figure 27] This diagram shows a voltage vector correction method represented in voltage space. [Figure 28] This diagram shows a voltage vector correction method represented in voltage space. [Figure 29] This diagram shows a voltage vector correction method represented in voltage space. [Figure 30] This diagram shows a voltage vector correction method represented in voltage space. [Figure 31] This figure shows the correspondence between each phase range of the voltage vector, the operable region where the voltage vector is modified, and the boundary line segment selected at the modification point in full-voltage RSPWM. [Figure 32] This is a block diagram showing the functional configuration of the phase voltage command correction unit in the control device. [Figure 33] This is a flowchart explaining the operation of the common-mode voltage command correction unit. [Figure 34] This diagram shows the voltage vector correction target in the voltage space for the non-operational region in full-voltage RSPWM. [Figure 35] This diagram shows a voltage vector correction method represented in voltage space. [Figure 36] This diagram shows a voltage vector correction method represented in voltage space. [Figure 37] This is a block diagram showing a modified example of the functional configuration of the phase voltage command correction unit in the control device. [Figure 38] This flowchart illustrates a modified operation of the common-mode voltage command correction unit. [Figure 39] This diagram shows a voltage vector correction method represented in voltage space. [Figure 40] This diagram shows a voltage vector correction method represented in voltage space. [Figure 41] This diagram shows a voltage vector correction method represented in voltage space. [Figure 42] This diagram shows a voltage vector correction method represented in voltage space. [Figure 43] This figure shows the modulation waveform of this power converter at a high modulation rate. [Figure 44] (A) is a conceptual diagram showing the average movement trajectory of the three-phase voltage command values ​​generated by this power converter, and (B) is a conceptual diagram showing an enlarged portion of the same movement trajectory. [Figure 45] This figure shows the modulation waveform of this power converter at a high modulation rate. [Figure 46] This diagram shows variations of the voltage vector correction method in voltage space. [Modes for carrying out the invention]

[0023] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. [Examples]

[0024] The power conversion device 1 in the embodiment to which the present invention is applied drives the motor 8 (load) of a so-called inverter-integrated electric compressor that constitutes the refrigerant circuit of a vehicle air conditioning system mounted on a vehicle such as an electric vehicle.

[0025] (1) Circuit configuration of power converter 1 In Figure 1, the power conversion device 1 of the embodiment includes a three-phase inverter circuit 27 and a control device 21. The inverter circuit 27 is a circuit that converts the DC voltage Vdc of a DC power source (vehicle battery: for example, 350V) 29 into a three-phase AC voltage and applies it to the motor 8. In this case, the motor 8 of the embodiment is an IPMSM (Interior Permanent Magnet Synchronous Motor).

[0026] The inverter circuit 27 has a U-phase half-bridge circuit 19U, a V-phase half-bridge circuit 19V, and a W-phase half-bridge circuit 19W. Each phase half-bridge circuit 19U to 19W has upper arm switching elements 18A to 18C and lower arm switching elements 18D to 18F, respectively. Furthermore, each switching element 18A to 18F has a freewheeling diode 31 connected in antiparallel. In this embodiment, each upper and lower arm switching element 18A to 18F is composed of an insulated gate bipolar transistor (IGBT) with a MOS structure incorporated into its gate portion.

[0027] The collectors of the upper arm switching elements 18A to 18C of the inverter circuit 27 are connected to the upper arm power line (positive bus) 10 of the DC power supply 29 and the smoothing capacitor 32. On the other hand, the emitters of the lower arm switching elements 18D to 18F of the inverter circuit 27 are connected to the lower arm power line (negative bus) 15 of the DC power supply 29 and the smoothing capacitor 32.

[0028] In this case, the emitter of the upper arm switching element 18A of the U-phase half-bridge circuit 19U and the collector of the lower arm switching element 18D are connected in series, the emitter of the upper arm switching element 18B of the V-phase half-bridge circuit 19V and the collector of the lower arm switching element 18E are connected in series, and the emitter of the upper arm switching element 18C of the W-phase half-bridge circuit 19W and the collector of the lower arm switching element 18F are connected in series.

[0029] Furthermore, the connection point (U-phase voltage Vu) between the upper arm switching element 18A and the lower arm switching element 18D of the U-phase half-bridge circuit 19U is connected to the U-phase armature coil of the motor 8, the connection point (V-phase voltage Vv) between the upper arm switching element 18B and the lower arm switching element 18E of the V-phase half-bridge circuit 19V is connected to the V-phase armature coil of the motor 8, and the connection point (W-phase voltage Vw) between the upper arm switching element 18C and the lower arm switching element 18F of the W-phase half-bridge circuit 19W is connected to the W-phase armature coil of the motor 8.

[0030] (Basic Configuration of the Control Device) Next, the control device 21 is composed of a microcomputer having a processor. In the embodiment, a rotational speed command value is input from the vehicle's ECU, and the motor current (phase current) is acquired from the motor 8. Based on these, the ON / OFF states (switching) of the switching elements 18A to 18F of the inverter circuit 27 are controlled. Specifically, the gate voltage applied to the gates of the switching elements 18A to 18F is controlled.

[0031] The control device 21 of the embodiment includes a dq-axis current command calculation unit 28, a phase voltage command value calculation unit 33, a phase voltage command correction unit 40, a modulation unit 50, a PWM signal generation unit 36, a gate driver 37, and current sensors 26A, 26B, and 26C composed of current transformers for measuring the motor currents (phase currents) of each phase flowing through the motor 8, namely, the U-phase current iu, the V-phase current iv, and the W-phase current iw. Each of the current sensors 26A to 26C is connected to the phase voltage command calculation unit 33.

[0032] In the embodiment, the current sensor 26A measures the U-phase current iu, the current sensor 26B measures the V-phase current iv, and the current sensor 26C measures the W-phase current iw. However, the U-phase current iu may be measured by the current sensor 26A, the V-phase current iv may be measured by the current sensor 26B, and the W-phase current iw may be obtained by calculation from these. Also, regarding the method of detecting the motor current of each phase, in addition to measuring with the current sensors 26A to 26C as in the embodiment, there are methods such as detecting the current value of the lower arm power supply line 15 with a shunt resistor and the phase voltage command calculation unit 33 estimating from the current value and the operating state of the motor 8. Therefore, the method of detecting and estimating each phase current is not particularly limited.

[0033] (dq-axis Current Command Calculation Unit) The dq-axis current command calculation unit 28 outputs a d-axis current command value and a q-axis current command value as target values for controlling the motor 8.

[0034] (Phase Voltage Command Calculation Unit) The phase voltage command calculation unit 33 calculates the electrical angle of the motor 8, the d-axis current command value Id ref , ref , ref , ref , ref , ref and the q-axis current command value Iq ref Based on the vector control using the d-axis current Id and q-axis current Iq obtained by converting the three-phase current detected by the current sensor 26B into the dq-axis, the phase voltage command values Vu, Vv, and Vw to be applied to the armature coils of each phase of the motor 8 are generated. ref (Hereinafter, the U-phase voltage command value Vu ref ), Vv ref (Hereinafter, the V-phase voltage command value Vv ref ), Vw ref (Hereinafter, the W-phase voltage command value Vw ref ) are calculated and output. Specifically, the phase voltage command calculation unit 33 includes a dq-axis current controller 34 and a coordinate conversion unit 35. Here, an example is shown where the phase voltage command calculation unit 33 outputs the phase voltage command value (command voltage vector) of a three-phase alternating current, but the present invention is not limited to this. As long as the command calculation unit can calculate any command voltage vector, other forms of output may be used.

[0035] The dq-axis current controller 34 compares the d-axis current Id and q-axis current Iq obtained by converting the three-phase current flowing through the motor 8 into the dq-axis with the d-axis current command value Id ref and the q-axis current command value Iq ref output from the dq-axis current command calculation unit 28, and performs feedback control (for example, PI control) on the current so that the two match. Further, the correction amount (Vm'-Vm) of the correction voltage vector Vm' output from the phase voltage command correction unit 40 described later is fed back to the dq-axis current controller 34 in the form of the d-axis correction amount Vd err and the q-axis correction amount Vq err , and these values are also reflected in the above PI control. As a result, the d-axis voltage command value Vd ref and the q-axis voltage command value Vq ref are output from the dq-axis current controller 34. As a specific example, the dq-axis current controller 34 may execute PI control according to the following mathematical formula (I).

Equation

[0036] The coordinate transformation unit 35 receives the d-axis voltage command value Vd from the dq-axis current controller 34. ref and q-axis voltage command value Vq ref Therefore, the α-axis voltage command value Vα is obtained using the following formula (II). ref and β-axis voltage command value Vβ ref Calculate these α-axis voltage command values ​​Vα ref and β-axis voltage command value Vβ ref Using equation (III), the voltage command value Vu for each phase of UVW ref , Vv ref VW ref Calculate the (phase voltage command value).

number

number

[0037] The above formula (III) is used with respect to the phase θm with respect to the α axis and the α-axis voltage command value Vα ref and β-axis voltage command value Vβ ref Rewriting this using the voltage vector (command voltage vector) Vm, which is composed of the following, we get the following equation (IV).

number

[0038] Figure 2 shows the voltage command values ​​Vu for each phase calculated using formula (IV). ref , Vv ref VW ref The waveform is shown, and Figure 3 shows the linear output region kH. The linear output region is the maximum amplitude at which the voltage vector can complete a clean rotation (draw a circle) in the voltage space diagram shown in Figure 3. Since the voltage space of a three-phase inverter is hexagonal as shown in Figure 3, theoretically the linear output region is the inscribed circle of the hexagon. In this application, the linear output region kH in general three-phase modulation is represented as 1, and the linear output regions of other modulation methods are normalized and discussed.

[0039] Furthermore, by combining the High and Low states of the U-phase voltage Vu, V-phase voltage Vv, and W-phase voltage Vw, they can be represented as eight voltage vectors (fundamental voltage vectors) V0 to V7, as shown in Figure 4. Of these, V1, V3, and V5 are odd voltage vectors, V2, V4, and V6 are even voltage vectors, and V0 and V7 are zero voltage vectors. When each voltage vector is shown in voltage space, it looks like Figure 5.

[0040] The coordinate transformation unit 35 of the embodiment further uses the following equations (V) and (VI) to obtain the voltage command value Vu for the two-phase modulation of each phase. ref2 , Vv ref2 VW ref2 It is calculating this.

number

number

[0041] Note that Vmod in formulas (V) and (VI) represents the voltage command value Vu for three-phase modulation. ref , Vv ref VW ref From the voltage command value Vu of two-phase modulation ref2 , Vv ref2 VW ref2 This is a correction value used to calculate the three-phase voltage command value Vu in formula (V). ref , Vv ref VW ref The smallest value (min) among them is added to Vdc / 2, and in formula (VI), the three-phase voltage command value Vu ref , Vv ref VW ref The maximum value (max) among them is subtracted from Vdc / 2.

[0042] And the voltage command value Vu for three-phase modulation ref , Vv ref VW refIf the sign of the phase with the maximum amplitude among the (phase voltage command values) is negative, two-phase modulation is performed by fixing the upper arm switching element of the phase with the maximum amplitude ON using formula (VI), and the voltage command value Vu ref , Vv ref VW ref If the sign of the phase with the maximum amplitude is positive, two-phase modulation is performed by fixing the lower arm switching element of the phase with the maximum amplitude ON using formula (V). Alternatively, if the sign of the phase with the maximum amplitude is positive, two-phase modulation may be performed by fixing the upper arm switching element of the phase with the maximum amplitude ON using formula (VI), and if the sign of the phase with the maximum amplitude is negative, two-phase modulation may be performed by fixing the lower arm switching element of the phase with the maximum amplitude ON using formula (V).

[0043] (Phase voltage command correction unit) The phase voltage command correction unit 40 processes the α-axis voltage command value Vα, which is temporarily converted by the coordinate transformation unit 35. ref and β-axis voltage command value Vβ ref The voltage vector Vm, which is composed of these elements, is modified to correct the α-axis voltage command value Vα'. ref and the corrected β-axis voltage command value Vβ' ref A corrected voltage vector Vm' is generated and passed to the coordinate transformation unit 35. The phase voltage command correction unit 40 uses equation (III) to calculate the corrected α-axis voltage command value Vα'. ref and the corrected β-axis voltage command value Vβ' ref From the corrected phase voltage command value Vu' of three-phase modulation ref , Vv' ref , VW' ref Calculate the corrected voltage command value Vu' for the two-phase modulation of each phase using formulas (V) and (VI). ref2 , Vv' ref2 , VW' ref2 The following is calculated and passed to the coordinate transformation unit 35. Note that the specific correction method in the phase voltage command correction unit 40 is closely related to the modulation method of the modulation unit 50, so the modulation unit 50 will be explained first, and the details of the phase voltage command correction unit 40 will be described later.

[0044] (Modulation section) The modulation unit 50 includes an odd-voltage modulation processing unit 52 and an even-voltage modulation processing unit 54. The odd-voltage modulation processing unit 52 uses the aforementioned corrected voltage vector Vm' to perform odd-side pulse width modulation, outputting only the odd-voltage vectors V1, V3, and V5 from the basic voltage vector during one control cycle. The even-voltage modulation processing unit 54 performs even-side pulse width modulation, outputting only the even-voltage vectors V2, V4, and V6 from the basic voltage vector during one control cycle. In odd-side pulse width modulation, the corrected α-axis voltage command value Vα' ref and the corrected β-axis voltage command value Vβ' ref From there, the ON times tu, tv, and tw of the upper arm switching elements of each phase are directly generated, and the voltage vectors (V1, V3, V5) and their output times are output. In even-side pulse width modulation, the corrected α-axis voltage command value Vα' is generated. ref and the corrected β-axis voltage command value Vβ' ref From this, the ON times tuv, tvw, and twu of the two-phase upper arm switching elements are directly generated, and the voltage vectors (V2, V4, V6) and their output times are output. In this application, this odd-side pulse width modulation will be hereinafter referred to as odd-voltage RSPWM (Remote State PWM), and the even-side pulse width modulation will be hereinafter referred to as even-voltage RSPWM (Remote State PWM). This pulse width modulation is based on the concept of instantaneous spatial vector modulation, which performs spatial vector modulation without waiting for the next sampling point.

[0045] The modulation unit 50 switches between odd-voltage RSPWM and even-voltage RSPWM based on the determination of the positive or negative sign of the maximum phase. In this application, this pulse width modulation will be hereafter referred to as the maximum phase determination type full-voltage RSPWM.

[0046] (PWM signal generation section) The PWM signal generation unit 36 ​​receives the voltage vector and output time output by the modulation unit 50, compares its magnitude with the carrier signal, and generates and outputs a PWM signal that serves as a drive command signal for the U-phase inverter 19U, V-phase inverter 19V, and W-phase inverter 19W of the inverter circuit 27.

[0047] (Gate driver) The gate driver 37 generates gate voltages for the switching elements 18A and 18D of the U-phase inverter 19U, the gate voltages for the switching elements 18B and 18E of the V-phase inverter 19V, and the gate voltages for the switching elements 18C and 18F of the W-phase inverter 19W, based on the PWM signal output from the PWM signal generation unit 36.

[0048] Then, each switching element 18A to 18F of the inverter circuit 27 is driven ON / OFF based on the gate voltage output from the gate driver 37. That is, when the gate voltage is ON (a predetermined voltage value), the switching element operates ON, and when the gate voltage is OFF (zero), the switching element operates OFF. When the switching elements 18A to 18F are IGBTs as described above, the gate driver 37 is a circuit for applying the gate voltage to the IGBTs based on the PWM signal, and is composed of photocouplers, logic ICs, transistors, etc.

[0049] Then, the voltage at the connection point between the upper arm switching element 18A and the lower arm switching element 18D of the U-phase half-bridge circuit 19U is applied (output) to the U-phase armature coil of the motor 8 as the U-phase voltage Vu (phase voltage), the voltage at the connection point between the upper arm switching element 18B and the lower arm switching element 18E of the V-phase half-bridge circuit 19V is applied (output) to the V-phase armature coil of the motor 8 as the V-phase voltage Vv (phase voltage), and the voltage at the connection point between the upper arm switching element 18C and the lower arm switching element 18F of the W-phase half-bridge circuit 19W is applied (output) to the W-phase armature coil of the motor 8 as the W-phase voltage Vw (phase voltage).

[0050] (Detailed explanation of the modulation section's operation) Next, the operation of the modulation unit 50 in this embodiment will be described with reference to Figure 6 and subsequent figures. Figure 6 is a flowchart illustrating the overall flow of pulse width modulation performed by the modulation unit 50. In step S1, the modified α-axis voltage command value Vα' is modified by the phase voltage command modification unit 40, which will be described in detail later. ref and the corrected β-axis voltage command value Vβ' ref, and the corrected phase voltage command value Vu' for three-phase modulation. ref , Vv' ref , VW' ref The modulation unit 50 receives this signal.

[0051] Then, in step S2, the corrected phase voltage command value Vu' for the three-phase modulation described above is used. ref , Vv' ref , VW' ref It is determined whether the sign of the phase with the maximum amplitude is positive or negative. Then, the corrected phase voltage command value Vu' for three-phase modulation is determined. ref , Vv' ref , VW' ref If the phase with the maximum amplitude has a positive sign, proceed to step S3 and perform odd voltage RSPWM.

[0052] (Odd voltage RSPWM) In step S3 of the odd-voltage RSPWM, the modulation unit 50 uses the following formulas (VII) and (VIII) to set the modified α-axis voltage command value Vα'. ref and the corrected β-axis voltage command value Vβ' ref From this, the ON times tu, tv, and tw of the upper arm switching elements 18A, 18B, and 18C for each phase are calculated. In formula (VII), V1, V3, and V5 are odd voltage vectors, and Ts is one control period. This control period Ts may be one carrier period. However, this control period Ts should be a period sufficiently shorter than one electrical angle period. Also, Su, Sv, and Sw are functions corresponding to the output regions (Sectors) A to C in the voltage space shown in Figure 7, and the correspondence between each output region and the functions Su, Sv, and Sw is shown in Figure 8. These functions Su, Sv, and Sw select the voltage vector in spatial vector modulation. Note that the calculation result of formula (VIII) is the same as that of formula (V), so either can be used in step S3.

number

number

[0053] Next, in step S4, the zero voltage output time t0, which is the time during which the zero voltage vector V0 is output, is calculated using formula (IX). Here, the corrected α-axis voltage command value Vα' is corrected in advance via the modulation range determination unit 42, which will be described later. ref and the corrected β-axis voltage command value Vβ' ref Because this method is employed, the zero-voltage output time t0 will always be a value greater than or equal to zero.

number

[0054] In step S5, the ON time of each phase's upper arm switching element is corrected using the following formula (X) based on the calculated zero-voltage output time t0. This is done by adding t0 / 3 to all of the ON times tu, tv, and tw. As a result, the zero-voltage output time is eliminated, and fluctuations in the common-mode voltage Vc of motor 8 are resolved.

number

[0055] Then, in step S6, as shown in Figure 10, the modulation unit 50 determines the odd voltage vectors (V1, V3, V5) and their output times based on each output region (Sector). These values ​​are then finally output to the PWM signal generation unit 36 ​​in step S14. Figure 10 shows the relationship between the output region (Sector), voltage vectors, and output times of the odd voltage RSPWM.

[0056] Circle Q1 in Figure 9 shows the linear output region when odd voltage RSPWM (odd RSPWM) is used alone. On the other hand, in this embodiment, output is performed using the entire range of the triangular region Z1 (hereinafter referred to as the odd-side operational region) formed by connecting the vertices of the odd voltage vectors (V1, V3, V5). The odd-side operational region Z1 has three line segments Z1a, Z1b, and Z1c that define the region. Figure 11 shows, for example, the output times of the odd voltage vectors V1, V3, and V5 calculated based on the corrected voltage vector Vm' in output region A. Furthermore, Figure 12 shows the output patterns of the odd voltage vectors V1, V3, and V5 in output region A.

[0057] Furthermore, if the corrected voltage vector Vm' falls within the range of circle Q1 in Figure 9, it means that a low modulation rate control state is continuing. For example, assuming that modulation is performed using "only" odd voltage RSPWM at a low modulation rate, the modulation waveforms of each phase of UVW will be as shown in Figure 13, and there will be no fluctuation in the common-mode voltage Vc (modulation waveform of only odd voltage RSPWM at a low modulation rate).

[0058] On the other hand, if the corrected voltage vector Vm' does not fall within the range of circle Q1, it means that the control state has a high modulation rate. Details of the modulated waveform in this case will be described later.

[0059] (Even voltage RSPWM)

[0060] Meanwhile, in step S2, the corrected phase voltage command value Vu' for three-phase modulation is used. ref , Vv' ref , VW' ref If the sign of the phase with the maximum amplitude is negative, the process proceeds to step S8 and executes even voltage RSPWM. In step S8, the modulation unit 50 uses the following equations (XI) and (XII) to calculate the modified α-axis voltage command value Vα'. ref and the corrected β-axis voltage command value Vβ' ref From this, calculate the ON times tuv, tvw, and twu of the two-phase upper arm switching element.

number

number

[0061] Furthermore, tuv is the ON time of the upper arm switching elements 18A and 18B of the U and V phases, tvw is the ON time of the upper arm switching elements 18B and 18C of the V and W phases, and twu is the ON time of the upper arm switching elements 18C and 18A of the W and U phases. In addition, V2, V4, and V6 in equation (XI) are even voltage vectors, and Suv, Svw, and Swu are functions corresponding to the output regions (Sectors) A to C of the voltage space shown in Figure 14. The correspondence between each output region and the functions Suv, Svw, and Swu is shown in Figure 15. These functions Suv, Svw, and Swu select the voltage vector in spatial vector modulation.

[0062] Furthermore, the modulation unit 50 calculates the OFF times tu (upper bar), tv (upper bar), and tw (upper bar) of the upper arm switching elements 18A, 18B, and 18C of each phase using the following formula (XIII). Note that the calculation result of formula (XIII) is the same as that of formula (XII), so either one may be used in step S8.

number

[0063] Next, in step S9, the zero voltage output time t7, which is the time for outputting the zero voltage vector V7, is calculated using formula (XIV). Here, the corrected α-axis voltage command value Vα' is corrected in advance via the modulation range determination unit 42, which will be described later. ref and the corrected β-axis voltage command value Vβ' ref Because this method is employed, the zero-voltage output time t7 will always be a value greater than or equal to zero.

number

[0064] In step S10, the OFF time of each phase's upper arm switching element is corrected using the following formulas (XV) and (XVI) based on the calculated zero-voltage output time t7. This is done by adding t7 / 3, which is the OFF time, to all of the OFF times tu (upper bar), tv (upper bar), and tw (upper bar). As a result, the zero-voltage output time is eliminated, and fluctuations in the common-mode voltage Vc of the motor 8 are resolved.

number

number

[0065] Then, in step S11, as shown in Figure 17, the modulation unit 50 determines the even voltage vectors (V2, V4, V6) and their output times based on each output region (Sector). These values ​​are then finally output to the PWM signal generation unit 36 ​​in step S14. Figure 17 shows the relationship between the output region (Sector), voltage vectors, and output time of the even voltage RSPWM.

[0066] The circle Q2 in Figure 16 shows the linear output region when this even voltage RSPWM (even RSPWM) is used alone. On the other hand, in this embodiment, output is performed using the entire range of the triangular region Z2 (hereinafter referred to as the even-side operational region) formed by connecting the vertices of the even voltage vectors (V2, V4, V6). The even-side operational region Z2 has three line segments Z2a, Z2b, and Z2c that define the region. Figure 18 shows, for example, the output times of the even voltage vectors V2, V4, and V6 based on the corrected voltage vector Vm' in output region A, and Figure 19 shows, for example, the output times of the even voltage vectors V2, V4, and V6 based on the corrected voltage vector Vm' in output region C. Furthermore, Figure 20 shows the output patterns of each voltage vector V4, V2, and V6 in output region C.

[0067] In addition, when the corrected voltage vector Vm' is within the range of the circle Q2, it means that a control state with a low modulation ratio continues. For example, assuming that in the low modulation ratio state, modulation is performed only with the even voltage RSPWM, the modulation waveforms of each phase of U, V, and W are as shown in FIG. 21, and there is no fluctuation in the common mode voltage Vc (modulation waveform with only the even voltage RSPWM at low modulation ratio).

[0068] On the other hand, when the corrected voltage vector Vm' is not within the range of the circle Q2, it means that a control state with a high modulation ratio is in effect. Details of the modulation waveform in this case will be described later.

[0069] (Full voltage RSPWM of maximum phase determination formula) In this embodiment, modulation control is performed while switching between the odd voltage RSPWM and the even voltage RSPWM. As a result, it becomes RSPWM using all voltage vectors. This is referred to as full voltage RSPWM. Among the full voltage RSPWM, in this embodiment, the corrected phase voltage command values Vu' ref , Vv' ref , Vw' ref Among them, depending on whether the sign of the phase with the maximum amplitude is positive or negative, the odd voltage RSPWM and the even voltage RSPWM are switched, so it becomes the full voltage RSPWM of the maximum phase determination formula.

[0070] FIG. 22 shows the operable region Z3 of the full voltage RSPWM of the maximum phase determination formula. The operable region of the full voltage RSPWM (hereinafter, the full voltage operable region) Z3 is a region where the odd-side operable region Z1 and the even-side operable region Z2 are superimposed. Also, a total of six regions (hereinafter, the inoperable region) X that are outside the range of the full voltage operable region Z3 exist so as to surround the periphery of the full voltage operable region Z3. The inoperable region X is an isosceles triangle.

[0071] Note that circle Q3 in Figure 22 shows the linear output region when the maximum phase determination type full-voltage RSPWM is used alone. Circle Q3, which is the linear output region for full-voltage RSPWM, is larger than circles Q1 and Q2, which are the linear output regions when odd-voltage RSPWM or even-voltage RSPWM are performed alone, respectively. On the other hand, in this embodiment, output is performed using the entire range of the full-voltage operation region Z3, beyond this linear output region (circle Q3).

[0072] Furthermore, if the corrected voltage vector Vm' falls within the range of circle Q3, it means that a low modulation rate control state is continuing. On the other hand, if the corrected voltage vector Vm' does not fall within the range of circle Q3, it means that a high modulation rate control state is in effect.

[0073] (Phase-based total voltage RSPWM) In the full-voltage RSPWM of the maximum phase determination formula described above, the modulation unit 50 controls the corrected phase voltage command value Vu' for three-phase modulation. ref , Vv' ref , VW' ref The system determines whether the sign of the phase with the maximum amplitude is positive or negative, and switches between odd-voltage RSPWM and even-voltage RSPWM accordingly. However, the present invention is not limited to this, and for example, the system may switch between odd-voltage RSPWM and even-voltage RSPWM based on the phase θm' of the corrected voltage vector with respect to the α-axis of the corrected voltage vector Vm'. In this application, this pulse width modulation will be hereinafter referred to as phase-determined full-voltage RSPWM.

[0074] Figure 23 shows the phase region odd to which odd-voltage RSPWM is applied and the phase region even to which even-voltage RSPWM is applied within the operating range of the phase-determined full-voltage RSPWM. Figure 24 shows the correspondence between each phase range and the odd-voltage and even-voltage RSPWM applied to them. As shown in Figures 23 and 24, one electrical angle period is divided into six regions (330°<θm'≦30°, 30°<θm'≦90°, 90°<θm'≦150°, 150°<θm'≦210°, 210°<θm'≦270°, 270°<θm'≦330°), and odd-voltage RSPWM and even-voltage RSPWM are switched alternately.

[0075] Note that the operational output region Z3 of the phase-determining full-voltage RSPWM is the same as that of the maximum phase-determining full-voltage RSPWM in Figure 22. In other words, the operational region Z3 of the full-voltage RSPWM (hereinafter referred to as the full-voltage operational region) is the region where the odd-side operational region Z1 and the even-side operational region Z2 are superimposed. In addition, there are a total of six regions X that are outside the range of the full-voltage operational region Z3 (hereinafter referred to as the non-operational region), surrounding the full-voltage operational region Z3.

[0076] Furthermore, if the corrected voltage vector Vm' falls within the range of circle Q3, it means that a low modulation rate control state is continuing. When the corrected voltage vector Vm' falls within the range of circle Q3, the modulation waveforms of each phase of UVW by full-voltage RSPWM are as shown in Figure 25. In this case, the common-mode voltage Vc fluctuates when switching between odd-voltage and even-voltage RSPWM (modulation waveform by full-voltage RSPWM at low modulation rate).

[0077] On the other hand, if the corrected voltage vector Vm' does not fall within the range of circle Q3, it means that the control state has a high modulation rate. Details of the modulated waveform in this case will be described later.

[0078] (Detailed explanation of the operation of the phase voltage command correction unit) Returning to Figure 1, the phase voltage command correction unit 40 sets the α-axis voltage command value Vα (before correction). ref and β-axis voltage command value Vβref It is determined whether the voltage vector Vm, which is composed of the above, is within the range of the full voltage operating region Z3. If the voltage vector Vm is within the range of the full voltage operating region Z3, no correction of the voltage vector Vm is necessary, and the voltage vector Vm is set to corrected voltage vector Vm'. On the other hand, as shown in Figure 26, if the voltage vector Vm is outside the range of the full voltage operating region Z3, i.e., in the non-operational region X, a voltage vector that is close to this voltage vector Vm (approximate range) and within the range of the full voltage operating region Z3 is calculated and set as the corrected voltage vector Vm'. The non-operational region X refers to the area within the hexagonal basic voltage space B enclosed by the basic voltage vectors V1 to V6 in vector control, and outside the range of the full voltage operating region Z3. The approximate ranges include, for example, the range where the vector length N' of the corrected voltage vector Vm' is 0.3N ≤ N' ≤ 1.7N, 0.5N ≤ N' ≤ 1.5N, or 0.7N ≤ N' ≤ 1.3N, relative to the vector length N of the voltage vector Vm. The approximate ranges also include, for example, the range where the voltage phase θm' of the corrected voltage vector Vm' is θm-120° ≤ θm' ≤ θm+120°, θm-90° ≤ θm' ≤ θm+90°, θm-60° ≤ θm' ≤ θm+60°, or θm-45° ≤ θm' ≤ θm+45°, relative to the voltage phase θm of the voltage vector Vm.

[0079] For example, there are several methods for calculating the corrected voltage vector Vm' within this approximate range K, such as (Method A) to (Method E).

[0080] (Method A: Correction to the boundary between the non-operational region and the voltage-operable region) For example, as shown in Figure 27, the corrected voltage vector Vm' is defined as any coordinates on the boundary lines Y1 and Y2 between the non-operational region X to which the voltage vector Vm belongs and the full-voltage operating region Z3. While this example illustrates the correction on the boundary lines Y1 and Y2 of the full-voltage operating region Z3, the present invention is not limited to this. The correction may also be made on the boundary line between the non-operational region X and the odd-numbered operating region Z1, or within the odd-numbered operating region Z1. Furthermore, the correction may also be made on the boundary line between the non-operational region X and the even-numbered operating region Z2, or within the even-numbered operating region Z2.

[0081] (Method B: Voltage Phase Correction) As shown in Figure 28, the length of the voltage vector Vm is not changed, and only the phase θm of the voltage vector Vm is modified to obtain the modified voltage vector Vm' at the point where it intersects with the boundary lines Y1 and Y2 of the non-operational region X and the fully-operable region Z3. When there are two candidate modified voltage vectors, one for forward rotation (positive direction modification) and one for reverse rotation (negative direction modification), it is preferable to select the one with a smaller absolute value of the voltage phase θm modification amount. Here, the case of modification on the boundary lines Y1 and Y2 of the fully-operable region Z3 is illustrated, but the present invention is not limited to this, and the modification may also be made on the boundary line between the non-operational region X and the odd-side operating region Z1, or within the odd-side operating region Z1, or on the boundary line between the non-operational region X and the even-side operating region Z2, or within the even-side operating region Z2.

[0082] (Method C: Correction of voltage vector length) As shown in Figure 29, the voltage vector Vm' is modified by changing only the length of the voltage vector Vm without changing its phase θm, and the point where it intersects with the boundary lines Y1 and Y2 of the non-operational region X and the fully-operable region Z3 is defined as the modified voltage vector Vm'. Here, the case where the modification is made on the boundary lines Y1 and Y2 of the fully-operable region Z3 is illustrated, but the present invention is not limited to this, and the modification may also be made on the boundary line between the non-operational region X and the odd-side operating region Z1 or within the odd-side operating region Z1, or on the boundary line between the non-operational region X and the even-side operating region Z2 or within the even-side operating region Z2.

[0083] (Method D: Parallel translation correction of the coordinates of the voltage vector to the outermost line of the inoperable region X) As shown in FIG. 30, the α-axis coordinate Vα and β-axis coordinate Vβ of the voltage vector Vm are translated in parallel with the outermost line Y3 (the contour line of the basic voltage space B) of the inoperable region X, and the place where they intersect the boundary lines Y1 and Y2 of the all-voltage operable region Z3 is set as the corrected voltage vector Vm'. When there are two corrected voltage vector candidates, one for positive rotation (positive direction correction) and one for reverse rotation (negative direction correction) of the phase θm, it is preferable to select the one with the smaller absolute value of the correction amount of the voltage phase θm. Here, an example is shown where the correction is made on the boundary lines Y1 and Y2 of the all-voltage operable region Z3, but the present invention is not limited to this, and the correction may be made on the boundary line between the inoperable region X and the odd-side operable region Z1 or within the odd-side operable region Z1, or on the boundary line between the inoperable region X and the even-side operable region Z2 or within the even-side operable region Z2.

[0084] (Method E: Combined correction of the above (Method A) to (Method D)) The corrected voltage vector Vm' may be calculated by appropriately combining the above (Method A) to (Method D).

[0085] In the above (Method A) to (Method E), an example was shown where the voltage vector Vm is corrected on the boundary of the all-voltage operable region Z3. However, in advance, either the odd-side operable region Z1 or the even-side operable region Z2 may be selected as the correction destination region, and the correction may be narrowed down and made on the boundary of either the odd-side operable region Z1 or the even-side operable region Z2. For example, when applying the all-voltage RSPWM of the maximum phase determination formula, among the phase voltage command values Vu ref , Vv ref , Vw ref if the sign of the phase with the maximum amplitude is negative, the correction to the even-side operable region Z2 may be selected, and if the sign of the phase with the maximum amplitude is positive, the correction to the odd-side operable region Z1 may be selected.

[0086] Furthermore, for example, when applying a phase-determined full-voltage RSPWM, as shown in Figure 31, the modification to the even-side operable region Z2 and the modification to the odd-side operable region Z1 may be switched according to the range of phase θm of the voltage vector Vm with respect to the α axis. Note that Figure 31 shows the boundary line segment (see Figure 31) that is selected when modifying the voltage vector Vm on the neighbor boundary line of the even-side operable region Z2 or the odd-side operable region Z1 based on the phase θm.

[0087] (Method B: Detailed explanation of voltage phase correction operation)

[0088] Next, assuming the application of the phase-determining full-voltage RSPWM, a detailed example of the operation of the phase voltage command correction unit 40 when the above (Method B) is adopted will be described. Figure 33 is a flowchart illustrating the operation flow of the phase voltage command correction unit 40. As shown in Figure 32, the phase voltage command correction unit 40 includes a modulation range determination unit 42, a modulation range selection unit 44, a correction command calculation unit 46, a correction command selection unit 48, and a feedback processing unit 49.

[0089] The modulation range determination unit 42 determines whether the coordinates of the voltage vector Vm are within the range of the full voltage operating range Z3 (step S41). If the voltage vector Vm is within the range of the full voltage operating range Z3, no correction of the voltage vector Vm is necessary, so the process proceeds to step S42, where the correction command calculation unit 46 sets the voltage vector Vm itself to the corrected voltage vector Vm'. Next, the process proceeds to step S80, where this corrected voltage vector Vm' is output to the modulation unit 50.

[0090] On the other hand, if in step S41 the coordinates of the voltage vector Vm are outside the range of the full voltage operating region Z3, i.e., located in the non-operating region X, then the voltage vector Vm needs to be corrected, and the process proceeds to step S44, where the modulation region selection unit 44 is activated.

[0091] In step S44, the modulation region selection unit 44 refers to the phase θm of the voltage vector Vm with respect to the α axis and selects either to modify it to the even-side operable region Z2 or to modify it to the odd-side operable region Z1, according to the relationship in Figure 31. When the relationship in Figure 31 is illustrated in Figure 34, the non-operable region X has an odd-number modification region X1 that modifies the voltage vector to the odd-side operable region Z1 and an even-number modification region X2 that modifies the voltage vector to the even-side operable region Z2. If the voltage vector Vm belongs to the odd-number modification region X1, the modulation region selection unit 44 modifies the voltage vector so that it is within the range of the odd-side operable region Z1. If the voltage vector Vm belongs to the even-number modification region X2, the modulation region selection unit 44 modifies the voltage vector so that it is within the range of the even-side operable region Z2. In this case, it is preferable that the voltage vector Vm of the odd-numbered correction region X1 be corrected to an odd-numbered operable region Z1 adjacent to the odd-numbered correction region X1, and to a triangular independent region Z1t that does not overlap with the even-numbered operable region Z2. Similarly, it is preferable that the voltage vector Vm of the even-numbered correction region X2 be corrected to an even-numbered operable region Z2 adjacent to the even-numbered correction region X2, and to a triangular independent region Z2t that does not overlap with the odd-numbered operable region Z1.

[0092] Furthermore, the three line segments Z1a, Z1b, and Z1c that demarcate the odd-numbered operating region Z1 can be expressed as functions of the α and β axes, resulting in the following equation (XVII). This function is used when calculating the corrected coordinates of the voltage vector Vm.

number

[0093] Similarly, the three line segments Z2a, Z2b, and Z2c that demarcate the even-numbered operating region Z2 can be expressed as functions of the α and β axes, resulting in the following equation (XVIII). This function is used when calculating the corrected coordinates of the voltage vector Vm.

number

[0094] For the sake of explanation, formulas (XVII) and (XVIII) above will be generalized to formula (XIX) below.

number

[0095] In step S44, if the voltage vector Vm is in the state shown in Figure 35, for example, that is, if the phase θm is within the range of 330° < θm ≤ 30°, the odd-side operable region Z1 is selected based on the correlation table in Figure 31. Then, the process proceeds to step S50, where the correction command calculation unit 46 corrects the voltage vector Vm to fall within the range of the odd-side operable region Z1.

[0096] Prior to the correction, the correction command calculation unit 46 defines a perfect circle P with the coordinates of the voltage vector Vm as its radius as a function of the α and β axes, as shown in Figure 35. As a result, the perfect circle P is given by the following formula (XX).

number

[0097] Next, since the phase θm of this voltage vector Vm belongs to the range 0° < θm ≤ 30°, we select the function representing the line segment Z1a from equation (XVIII) based on the correlation table in Figure 31, and calculate the two intersection points (VAα', VAβ') and (VBα', VBβ') of this function and the function representing the perfect circle P in equation (XX). The reason for selecting the line segment Z1a is that when this voltage vector Vm is rotated in both forward and reverse directions, it intersects this line segment Z1a first.

[0098] Using the function generalized by equation (XIX), the coordinates of the two intersection points are given by equation (XXI).

number

[0099] As a result of the above, in step S50, the correction command calculation unit 46 calculates first and second voltage vectors (hereinafter referred to as first and second correction candidate voltage vectors) VAm' and VBm' which are candidates for correction to the odd-side operable region Z1.

[0100] Next, the process proceeds to step S52, where the correction command selection unit 48 selects one voltage vector from the first and second candidate correction voltage vectors VAm' and VBm', and designates this as the corrected voltage vector Vm'. Prior to this selection, the correction command selection unit 48 calculates the first phase correction amount θAmod of the first candidate correction voltage vector VAm' with respect to voltage vector Vm, and the second phase correction amount θBmod of the second candidate correction voltage vector VBm' with respect to voltage vector Vm, using the following formula (XXII).

number

[0101] The correction command selection unit 48 compares the first phase correction amount θAmod and the second phase correction amount θBmod, and determines the smaller of the two correction candidate voltage vectors VAm' and VBm' (in this case, the first correction candidate voltage vector VAm') as the correction voltage vector Vm'. Then, proceeding to step S80, this correction voltage vector Vm' is output to the modulation unit 50. Furthermore, proceeding to step S82, the feedback processing unit 49 converts the correction amount (Vm'-Vm) of the correction voltage vector Vm' into a dq-axis voltage to obtain the d-axis correction amount Vd err and q-axis correction amount Vq err This generates a current and feeds it back to the dq-axis current controller 34.

[0102] Returning to step S44, if the voltage vector Vm is in the state shown in Figure 36, for example, i.e., the phase θm is within the range of 30° < θm ≤ 90°, then the even-side operable region Z2 is selected based on the correlation table in Figure 31. Then, proceeding to step S60, the correction command calculation unit 46 corrects the voltage vector Vm to fall within the range of the even-side operable region Z2.

[0103] Prior to the correction, the correction command calculation unit 46 defines a perfect circle P with the coordinates of the voltage vector Vm as its radius using the above formula (XX), as shown in Figure 36. Furthermore, since the phase θm of this voltage vector Vm belongs to 30° < θm ≤ 60°, based on the correlation table in Figure 31, the unit selects a function representing the line segment Z2a from formula (XVIII) and calculates the two intersection points (VAα', VAβ') and (VBα', VBβ') of this function and the function representing the perfect circle P in formula (XX). The reason for selecting the line segment Z2a is that when this voltage vector Vm is rotated in both forward and reverse directions, it intersects with this line segment Z2a first.

[0104] As a result of the above, in step S60, the correction command calculation unit 46 calculates first and second voltage vectors (hereinafter referred to as first and second correction candidate voltage vectors) VAm' and VBm' which are candidates for correction to the even-side operable region Z2.

[0105] Next, the process proceeds to step S62, where the correction command selection unit 48 selects one voltage vector from the first and second candidate correction voltage vectors VAm' and VBm', and designates this as the corrected voltage vector Vm'. Prior to this selection, the correction command selection unit 48 calculates the first phase correction amount θAmod of the first candidate correction voltage vector VAm' with respect to voltage vector Vm, and the second phase correction amount θBmod of the second candidate correction voltage vector VBm' with respect to voltage vector Vm, using the above formula (XXII).

[0106] The correction command selection unit 48 compares the first phase correction amount θAmod and the second phase correction amount θBmod, and determines the smaller of the two correction candidate voltage vectors VAm',VBm' (in this case, the second correction candidate voltage vector VBm') as the correction voltage vector Vm'. Then, proceeding to step S80, this correction voltage vector Vm' is output to the modulation unit 50. Furthermore, proceeding to step S82, the feedback processing unit 49 converts the correction amount (Vm'-Vm) of the correction voltage vector Vm' into a dq axis voltage to obtain the d axis correction amount Vd err and q-axis correction amount Vq err This generates a current and feeds it back to the dq-axis current controller 34.

[0107] (Method C: Detailed explanation of the operation of voltage vector length correction)

[0108] Next, assuming the application of the phase-determining full-voltage RSPWM, a detailed example of the operation of the phase voltage command correction unit 40 when the above (Method C) is adopted will be described. Figure 38 is a flowchart illustrating the operation flow of the phase voltage command correction unit 40. As shown in Figure 37, the phase voltage command correction unit 40 includes a modulation range determination unit 42, a modulation range selection unit 44, a correction command calculation unit 46, and a feedback processing unit 49.

[0109] The modulation range determination unit 42 determines whether the coordinates of the voltage vector Vm are within the range of the full voltage operating range Z3 (step S71). If the voltage vector Vm is within the range of the full voltage operating range Z3, no correction of the voltage vector Vm is necessary, so the process proceeds to step S72, where the correction command calculation unit 46 sets the voltage vector Vm itself as the corrected voltage vector Vm'. Next, the process proceeds to step S100, where this corrected voltage vector Vm' is output to the modulation unit 50.

[0110] On the other hand, if in step S71 the coordinates of the voltage vector Vm are outside the range of the full voltage operating region Z3, i.e., located in the non-operating region X, then the voltage vector Vm needs to be corrected, and the process proceeds to step S74, where the modulation region selection unit 44 is activated.

[0111] In step S74, the modulation region selection unit 44 refers to the phase θm of the voltage vector Vm with respect to the α axis and selects either to modify it to the even-side operable region Z2 or to modify it to the odd-side operable region Z1, according to the correlation table in Figure 31. For example, if the voltage vector Vm is in the state shown in Figure 39, i.e., the phase θm is within the range of 330° < θm ≤ 30°, then the odd-side operable region Z1 is selected based on the correlation table in Figure 31. Then, the process proceeds to step S80, where the modification command calculation unit 46 modifies the voltage vector Vm to within the odd-side operable region Z1.

[0112] Prior to the correction, the correction command calculation unit 46 defines the radial line segment E, which is in phase with the voltage vector Vm, as a function of the α and β axes, as shown in Figure 39. As a result, the radial line segment E is given by the following formula (XXIII).

number

[0113] Next, since the phase θm of this voltage vector Vm belongs to the 0° < θm ≤ 30° range, we select the function representing the line segment Z1a from equation (XVII) based on the correlation table in Figure 31, and calculate the intersection point (Vα', Vβ') of this function with the function representing the radial line segment E in equation (XXIII). The reason for selecting the line segment Z1a is that it is the closest boundary line to this voltage vector Vm.

[0114] For the sake of explanation, formulas (XVII) and (XVIII) above are generalized by formula (XXIV) below.

number

[0115] The coordinates of the intersection point of this generalized formula (XXIV) and formula (XXIII) are given by the following formula (XXV).

number

[0116] As a result of the above, in step S80, the correction command calculation unit 46 calculates a corrected voltage vector Vm' which is the result of the correction to the odd-side operable region Z1. Then, proceeding to step S100, this corrected voltage vector Vm' is output to the modulation unit 50. Furthermore, proceeding to step S102, the feedback processing unit 49 converts the correction amount (Vm'-Vm) of the corrected voltage vector Vm' into a dq-axis voltage to obtain the d-axis correction amount Vd err and q-axis correction amount Vq err This generates a current and feeds it back to the dq-axis current controller 34.

[0117] Returning to step S74, if the voltage vector Vm is in the state shown in Figure 40, i.e., the phase θm is within the range of 30° < θm ≤ 90°, then the even-side operable region Z2 is selected based on the correlation table in Figure 31. Then, proceeding to step S90, the correction command calculation unit 46 corrects the voltage vector Vm to fall within the range of the even-side operable region Z2.

[0118] Prior to the correction, the correction command calculation unit 46 defines the radial line segment E, which is in phase with the voltage vector Vm, as a function of the α and β axes, as shown in Figure 40. As a result, the radial line segment E is given by the above formula (XXIII).

[0119] Next, since the phase θm of this voltage vector Vm belongs to the 30° < θm ≤ 60° range, we select the function representing the line segment Z2a from equation (XVIII) based on the correlation table in Figure 31, and calculate the intersection point (Vα', Vβ') of this function and the function representing the radial line segment E in equation (XXIII). The reason for selecting the line segment Z2a is that it is the closest boundary line to this voltage vector Vm.

[0120] As a result of the above, in step S90, the correction command calculation unit 46 calculates a corrected voltage vector Vm' which is the result of the correction to the even-side operable region Z2. Then, proceeding to step S100, this corrected voltage vector Vm' is output to the modulation unit 50. Furthermore, proceeding to step S102, the feedback processing unit 49 converts the correction amount (Vm'-Vm) of the corrected voltage vector Vm' into a dq-axis voltage to obtain the d-axis correction amount Vd err and q-axis correction amount Vq err This generates a current and feeds it back to the dq-axis current controller 34.

[0121] (Method D: Detailed explanation of operation when the coordinates of the voltage vector are translated parallel to the outermost line of the unoperable region X) Next, assuming the application of the phase-determining full-voltage RSPWM, a detailed example of the operation of the phase voltage command correction unit 40 when the above (Method D) is adopted will be described. Note that the internal configuration and operation flow flowchart of the phase voltage command correction unit 40 are the same as those in Figures 37 and 38, so these will be used as reference.

[0122] The modulation range determination unit 42 determines whether the coordinates of the voltage vector Vm are within the range of the full voltage operating range Z3 (step S71). If the voltage vector Vm is within the range of the full voltage operating range Z3, no correction of the voltage vector Vm is necessary, so the process proceeds to step S72, where the correction command calculation unit 46 sets the voltage vector Vm itself as the corrected voltage vector Vm'. Next, the process proceeds to step S100, where this corrected voltage vector Vm' is output to the modulation unit 50.

[0123] On the other hand, if in step S71 the coordinates of the voltage vector Vm are outside the range of the full voltage operating region Z3, i.e., located in the non-operating region X, then the voltage vector Vm needs to be corrected, and the process proceeds to step S74, where the modulation region selection unit 44 is activated.

[0124] In step S74, the modulation region selection unit 44 refers to the phase θm of the voltage vector Vm with respect to the α axis and selects either the modification to the even-side operable region Z2 or the modification to the odd-side operable region Z1, according to the correlation table in Figure 31.

[0125] In step S74, if the voltage vector Vm is in the state shown in Figure 41, for example, that is, if the phase θm is within the range of 330° < θm ≤ 30°, the odd-side operable region Z1 is selected based on the correlation table in Figure 31. Then, the process proceeds to step S80, where the correction command calculation unit 46 corrects the voltage vector Vm to fall within the range of the odd-side operable region Z1.

[0126] Prior to the correction, the correction command calculation unit 46 applies formula (VIII) to the voltage vector Vm, as shown in Figure 41, to calculate the ON time command values ​​tu, tv, and tw for each phase's upper arm switching elements 18A, 18B, and 18C. Note that no adjustments using formulas (VIX) and (X) are made here. As a result, one of tu, tv, or tw is always zero. In the case of the voltage vector Vm in Figure 41, tw is always zero.

[0127] Since the phase θm of the voltage vector Vm belongs to the range 0° < θm ≤ 30°, the relationship tu > tv always holds. Therefore, while maintaining the larger value tu, the tv side is corrected using the relationship tv' = Ts - tu. In other words, tv is corrected to tv' using the relationship Ts = tu + tv'. As a result, as shown in Figure 41, a corrected voltage vector Vm' that falls within the odd-side operating region Z1 is calculated based on the corrected ON time command value (tu, tv').

[0128] Furthermore, the coordinates (Vα',Vβ') of this corrected voltage vector Vm' on the α and β axes are the intersection points of the line segment F that passes through the coordinates of the voltage vector Vm before correction and is parallel to the outermost line Y3 of the non-operational region X, and the function that represents the line segment Z1a selected from equation (XVII) based on the correlation table in Figure 31.

[0129] To generalize the above explanation, we define t1 and t2 as the two non-zero command values ​​among the ON times tu, tv, and tw calculated by applying formula (VIII) to the voltage vector Vm. In this case, the coordinates (t1', t2') that fall within the odd-side operational region Z1 of the modified voltage vector Vm' maintain the larger of t1 and t2 as is, and modify the smaller of the two command values, resulting in the following formula (XXVI).

number

[0130] As a result of the above, in step S80, the correction command calculation unit 46 calculates a corrected voltage vector Vm' which is the result of the correction to the odd-side operable region Z1. Then, proceeding to step S100, this corrected voltage vector Vm' is output to the modulation unit 50. Furthermore, proceeding to step S102, the feedback processing unit 49 converts the correction amount (Vm'-Vm) of the corrected voltage vector Vm' into a dq-axis voltage to obtain the d-axis correction amount Vd err and q-axis correction amount Vq err This generates a current and feeds it back to the dq-axis current controller 34.

[0131] Returning to step S74, if the voltage vector Vm is in the state shown in Figure 42, for example, i.e., the phase θm is within the range of 30° < θm ≤ 90°, then the even-side operable region Z2 is selected based on the correlation table in Figure 31. Then, proceeding to step S90, the correction command calculation unit 46 corrects the voltage vector Vm to fall within the range of the even-side operable region Z2.

[0132] Prior to the correction, the correction command calculation unit 46 applies formula (XIII) to the voltage vector Vm, as shown in Figure 42, to calculate the OFF time command values ​​tu (upper bar), tv (upper bar), and tw (upper bar) for each phase's upper arm switching elements 18A, 18B, and 18C. Note that no adjustment using formula (XIV) is performed here. As a result, one of tu (upper bar), tv (upper bar), or tw (upper bar) is always zero. Note that in the case of the voltage vector Vm in Figure 42, tu (upper bar) is always zero.

[0133] Since the phase θm of the voltage vector Vm belongs to the range 30° < θm ≤ 60°, the relationship tw (upper bar) > tv (upper bar) always holds true. Therefore, while maintaining the larger value, tw (upper bar), we modify the tv (upper bar) side using the relationship tv' (upper bar) = Ts - tw (upper bar). In other words, we modify tv (upper bar) to tv' (upper bar) using the relationship Ts = tw (upper bar) + tv' (upper bar). As a result, as shown in Figure 42, the corrected voltage vector Vm' within the even-side operational region Z2 is calculated based on the corrected OFF time command values ​​(tv' (upper bar), tw (upper bar). The coordinates (Vα', Vβ') of this corrected voltage vector Vm' on the α and β axes are the intersection points of a line segment F that passes through the coordinates of the voltage vector Vm before correction and is parallel to the outermost line Y3 of the non-operational region X, and a function that represents the line segment Z2a selected from formula (XVIII) based on the correlation table in Figure 31.

[0134] As a result of the above, in step S90, the correction command calculation unit 46 calculates a corrected voltage vector Vm' which is the result of the correction to the even-side operable region Z2. Then, proceeding to step S100, this corrected voltage vector Vm' is output to the modulation unit 50. Furthermore, proceeding to step S102, the feedback processing unit 49 converts the correction amount (Vm'-Vm) of the corrected voltage vector Vm' into a dq-axis voltage to obtain the d-axis correction amount Vd err and q-axis correction amount Vq err This generates a current and feeds it back to the dq-axis current controller 34.

[0135] To generalize the above explanation, we define the remaining two non-zero command values ​​as t1 (upper bar) and t2 (upper bar) among the OFF times tu (upper bar), tv (upper bar), and tw (upper bar) calculated by applying formula (XIII) to the voltage vector Vm. In this case, the coordinates (t1' (upper bar), t2' (upper bar)) within the even-side operational region Z2 of the corrected voltage vector Vm' are given by formula (XXVII) below, since the larger of the two command values ​​t1 (upper bar) and t2 (upper bar) is retained and the smaller command value is corrected.

number

[0136] As described above, with the power converter 1 of this embodiment, the phase voltage command correction unit 40 converts the voltage vector Vm belonging to the non-operational region X into a corrected voltage vector Vm' within the full-voltage operation region Z3, and then the modulation unit 50 performs modulation control. This makes it possible to always use the same modulation scheme for modulation. As a result, even if the voltage vector Vm partially falls outside the full-voltage operation region Z3 when controlling at a high modulation rate, it is all corrected to fall within the full-voltage operation region Z3. Therefore, it is always driven with a modulation scheme that does not excite common-mode noise, and thus common-mode noise is significantly suppressed overall.

[0137] Furthermore, in the power converter 1, when the phase voltage command correction unit 40 converts the voltage vector Vm into a corrected voltage vector Vm' within the full-voltage operation region Z3, it selects a range close to the non-operational region X to which the voltage vector Vm belongs. In addition, since the amount of correction of the corrected voltage vector Vm' is fed back and controlled, the amount of correction of the corrected voltage vector Vm' can be suppressed, making it possible to stabilize the three-phase voltage commands Vu, Vv, and Vw output from the modulation unit 50.

[0138] Figure 43 shows the waveforms B (reference) of the three-phase voltage commands Vu, Vv, and Vw output from the modulation unit 50 by the power converter 1 of this embodiment, and the common-mode voltage waveform Vc. It can be seen that the fluctuation of the common-mode voltage waveform Vc is suppressed significantly more than in the conventional power converter 1.

[0139] (Modification control of the selection method for the operable region in the modulation region selection unit) As already mentioned, in the modulation region selection unit 44, the phase θm of the voltage vector Vm with respect to the α axis is referenced, and in the method of selecting either correction to the even-side operable region Z2 or correction to the odd-side operable region Z1 according to the correlation table in Figure 31, as long as the rotation speed of the phase θm is always positive, the modulation region selection unit 44 alternately switches between the even-side operable region Z2 and the odd-side operable region Z1 at phase intervals of 60° for the phase θm.

[0140] By the way, in this embodiment, in order to stabilize the output control of the motor 8, current control is performed by feeding back the amount of correction of the corrected voltage vector Vm' to the dq-axis current controller 34, as shown in equation (I). The corrected voltage vector Vm' has a phase θm that moves forward or backward and a vector length that increases or decreases relative to the original voltage vector Vm, so when current control is performed by feeding back information about these correction amounts to the dq-axis current controller 34, the output d-axis voltage command value Vd ref and q-axis voltage command value Vq ref That influence will be reflected there.

[0141] Figure 44(A) shows the effect of modifying the voltage vector Vm on the phase voltage command value. Line D in Figure 44(A) represents the d-axis voltage command value Vd ref and q-axis voltage command value Vq ref The α-axis voltage command value Vα is obtained by transforming the coordinates to the αβ axis using equation (II). ref and β-axis voltage command value Vβ refThis is a conceptual diagram representing the averaged movement trajectory of the voltage vector Vm. Note that this "averaging of the movement trajectory" is a concept where, because the voltage vector Vm actually output from the dq-axis current controller 34 oscillates significantly in the rotational and radial directions along the αβ axis, making it difficult to illustrate, a virtual filter in the rotational and radial directions is applied to smooth out the movement trajectory. As shown in region D1 of line D, the voltage vector Vm rotates in the opposite direction near the switching boundary phases (30°, 90°, 150°, 210°, 270°, 330°) for switching between the even-side operable region Z2 and the odd-side operable region Z1 in the correlation table in Figure 31.

[0142] Figure 44(B) is a conceptual diagram showing a portion of the movement trajectory in this region D1, magnified without smoothing. Since the actual voltage vector Vm oscillates finely, the movement trajectory of the voltage vector Vm oscillates circumferentially around the switching boundary phase of 270°. If the movement trajectory of the voltage vector Vm is defined in the control timing time series k1 to k7 for each control period Ts, and the correlation table in Figure 31 is applied directly to k1 to k7, the modified regions become: k1: odd-side operable region Z1, k2: odd-side operable region Z1, k3: even-side operable region Z2, k4: odd-side operable region Z1, k5: even-side operable region Z2, k6: odd-side operable region Z1, k7: even-side operable region Z2, k8: even-side operable region Z2. As a result, during the short control period from k3 to k7 (5 control cycles), the odd-side operable region Z1 and the even-side operable region Z2 switch alternately, exciting unintended common-mode voltage fluctuations as shown by the dotted line C in Figure 43.

[0143] Therefore, the modulation region selection unit 44 modifies the correlation table in Figure 31. Specifically, when the voltage vector Vm rotates in the reverse direction and passes through the switching boundary phase (270° in Figure 44(B)) in the reverse rotation direction, and the operable region after the reverse rotation (odd-numbered operable region Z1 in Figure 44(B)) is about to be selected, it is forcibly replaced with the operable region before the reverse rotation (even-numbered operable region Z2 in Figure 44(B)). In other words, as in the control timing time series k3→k4, when the voltage vector Vm rotates in the reverse direction and it is possible to select the even-numbered operable region Z2 before the reverse rotation → odd-numbered operable region Z1 after the reverse rotation, the selection at k4 is forcibly replaced with the even-numbered operable region Z2 before the reverse rotation. Similarly, if the voltage vector Vm rotates in the reverse direction, such as from k5 to k6 in the control timing time series, and it is possible to select between the even-side operable region Z2 before the reverse rotation and the odd-side operable region Z1 after the reverse rotation, the selection at k6 is forcibly replaced with the even-side operable region Z2 before the reverse rotation. In other words, the switching at the switching boundary phase is made to have a unidirectional characteristic in the forward rotation direction, and switching in the reverse rotation direction is prohibited.

[0144] To perform this corrective control, the modulation region selection unit 44 selects the d-axis voltage command value Vd output from the dq-axis current controller 34 at the next (future) control timing k+1, from the current control timing k. ref and q-axis voltage command value Vq ref We can infer this using the following formula (XXVIII).

number

[0145] Estimated d-axis voltage command value Vd ref and q-axis voltage command value Vq ref By calculating the phase θm, it is possible to determine whether the voltage vector Vm rotates in the reverse direction at the next control timing k+1, and whether the switching boundary phase passes in the reverse rotation direction. Only if it passes in the reverse rotation direction should the operable region before the reverse rotation be selected.

[0146] Figure 45 shows the waveforms B (reference) of the three-phase voltage commands Vu, Vv, and Vw output from the modulation unit 50, and the common-mode voltage waveform Vc, when corrective control is applied in the modulation region selection unit 44. It can be seen that the fluctuation of the common-mode voltage waveform Vc is suppressed compared to the fluctuation in Figure 43.

[0147] In this embodiment, examples of modulation methods utilizing a portion of the fundamental vector region B include odd-voltage RSPWM, even-voltage RSPWM, and full-voltage RSPWM. However, the present invention is not limited to these, and other modulation methods can be employed. Furthermore, in this embodiment, the modulation unit 50 is shown as having two modulation methods: an odd-voltage modulation processing unit 52 that executes odd-voltage RSPWM and an even-voltage modulation processing unit 54 that executes even-voltage RSPWM. However, the present invention is not limited to this. The modulation unit 50 may have a single modulation method. Moreover, the modulation unit 50 may have a first modulation processing unit that executes another first modulation method and a second modulation processing unit that executes another second modulation method, and these may be switched as appropriate. Furthermore, in the present invention, the modulation unit 50 may employ three or more modulation methods and switch them as appropriate.

[0148] Furthermore, in this embodiment, as (Method A) in the phase voltage command correction unit 40, the case in which the voltage vector Vm is corrected to the boundary between the non-operational region and the voltage operationable region is illustrated, but the present invention is not limited to this, and the correction can also be made inside this boundary. Also, (Method A), (Method B), (Method C), and (Method D) as voltage vector correction methods are illustrative, and other correction methods can be adopted. For example, as an example of combining (Method A), (Method B), and (Method C), as shown in Figure 46, it is also preferable to define a perpendicular line G from the coordinates of the voltage vector Vm to the nearest boundary line Y1, and correct the voltage vector Vm' at the intersection of this perpendicular line G and the boundary line Y1.

[0149] Furthermore, although the above embodiments were described using the driving of the motor (load) of an electric compressor as an example, the present invention is not limited to this and is also effective when driving motors other than the motor of an electric compressor. Moreover, the present invention is applicable to various power conversion devices that convert DC voltage to AC voltage using an inverter and apply it to a load. [Explanation of Symbols]

[0150] 1. Power converter 8 motors 18A~18F Up / Down Arm Switching Element 19U U-phase inverter 19V V-phase inverter 19W W-phase inverter 21 Control device 27 Inverter Circuit 28 dq axis current command calculation section 33 Phase voltage command calculation unit 34 dq axis current controller 35 Coordinate Transformation Unit 36 PWM signal generation section 37 Gate Driver 40-phase voltage command correction unit 42 Modulation range determination unit 44 Modulation Region Selection Section 46 Modified command calculation section 48 Correction Command Selection Unit 49 Feedback Processing Unit 50 Modulation section

Claims

1. In a power conversion device that converts DC voltage to AC voltage, An inverter circuit that applies the phase voltage generated at the connection point of the switching elements of each phase to the load, The inverter circuit comprises a control device for controlling the switching of the switching elements, The control device is A modulation unit that modulates a portion of the fundamental voltage space, which is the outputtable voltage vector region of the inverter circuit, into a modulatorable region, The system includes a phase voltage command correction unit that calculates a corrected voltage vector by correcting the command voltage vector to fall within the moduloable region when the command voltage vector is within the basic voltage space but outside the moduloable region, The power conversion device is characterized in that the modulation unit outputs using the modified voltage vector.

2. The modulation unit is A first modulation processing unit having a first moduloable region which is the moduloable region, The system comprises a second modulation processing unit having a second moduloable region which is the moduloable region and is a region different from the first moduloable region, The aforementioned phase voltage command correction unit is The power conversion device according to claim 1, characterized in that the modified voltage vector is modified within either the first moduloable region or the second moduloable region.

3. The phase voltage command correction unit includes a modulation region selection unit that selects either the first moduloable region or the second moduloable region as the target for correction of the command voltage vector. The power conversion device according to claim 2, characterized in that the modulation region selection unit switches between the first moduloable region and the second moduloable region based on the phase of the command voltage vector.

4. The first modulation processing unit performs pulse width modulation that outputs only odd voltage vectors, The power conversion device according to claim 2, characterized in that the second modulation processing unit performs pulse width modulation that outputs only even voltage vectors.

5. The phase voltage command correction unit includes a modulation region selection unit that selects either the first moduloable region or the second moduloable region as the target for correction of the command voltage vector. The power converter according to claim 2, characterized in that the modulation region selection unit does not perform the switching of the first moduloable region and the second moduloable region immediately after performing the switching of the first moduloable region and the second moduloable region.

6. When defining the phase that serves as the boundary for switching between the first moduloable region and the second moduloable region in the modulation region selection unit as the switching boundary phase, The power conversion device according to claim 3, characterized in that the modulation region selection unit does not perform switching between the first moduloable region and the second moduloable region when the voltage vector passes the switching boundary phase in the reverse rotation direction.

7. The power conversion device according to claim 3, wherein the modulation region selection unit predicts the future command voltage vector to estimate a continuous switching between the first moduloable region and the second moduloable region, and if, as a result of the estimation, it is estimated that a continuous switching will occur, the switching between the first moduloable region and the second moduloable region is not performed when the future command voltage vector arrives.

8. The unit comprises a command calculation unit that generates the command voltage vector, The command calculation unit, The power conversion device according to claim 1, characterized in that the error between the corrected voltage vector calculated by the phase voltage command correction unit and the command voltage vector before correction is compensated for in subsequent calculations of the command voltage vector.

9. The power conversion device according to claims 1 to 7, characterized in that the phase voltage command correction unit sets the correction voltage vector on the boundary line defining the moduloable region.

10. The power conversion device according to claims 1 to 7, characterized in that the phase voltage command correction unit sets the relationship between the voltage vector and the corrected voltage vector such that their lengths are the same and their phases are different.

11. The power conversion device according to claims 1 to 7, characterized in that the phase voltage command correction unit sets the relationship between the voltage vector and the corrected voltage vector such that their phases are the same and their lengths are different.

Images