CN116458050A - Power conversion device - Google Patents

Abstract

The power conversion device (100) is provided with: a power conversion unit (1) that converts DC power into AC power and supplies the AC power to a load (3) in accordance with an on/off state amount determined by a combination of on/off parameters of a plurality of switching elements; a voltage output calculation unit (11) that calculates a voltage output value from the power conversion unit (1) on the basis of the on-off state quantity; an integrated value calculation unit (12) that integrates the voltage command value and the voltage output value calculated by the voltage output calculation unit (11) to obtain a voltage command integrated value and a voltage output integrated value; an on-off update determination unit (13) that outputs an update signal of the on-off state quantity based on the voltage command integrated value, the allowable value of the voltage command integrated value, and the voltage output integrated value; and an on-off decision table (14) for deciding on-off state amounts of the plurality of switching elements of the power conversion unit (1) based on the voltage instruction integrated value and the voltage output integrated value obtained by the integrated value calculation unit (12), and the update signal of the on-off update determination unit (13).

Description

Power conversion device

Technical Field

The present invention relates to a power conversion device.

Background

In a power conversion device using a PWM (pulse width modulation: pulse Width Modulation) system, a triangular wave carrier comparison PWM system is generally used in which a triangular wave carrier and a voltage command value are compared to determine the on/off state of a switching element, but since an output voltage is a rectangular pulse simulating a sine wave, a harmonic wave is generated in addition to the sine wave as a fundamental wave, and a ripple (also referred to as ripple) is generated in a current flowing through a rotating electric machine or a generated torque.

In order to solve this problem, a method (direct on-off control method) of directly determining the on-off states of a plurality of switching elements of a power conversion unit has been proposed. As one of the direct on-off control methods, there is known a direct torque control method in which an allowable value is set for a command value of torque and magnetic flux of a rotating electrical machine, and an on-off state is switched when the allowable value is exceeded. Since the on-off state is determined so as to suppress magnetic flux and torque ripple, current flowing through the rotating electrical machine or torque ripple generated can be reduced as compared to the PWM method. Further, by setting the allowable value to be large, the number of times of on-off state transition of the plurality of switching elements of the power conversion section becomes small, and therefore, on-off loss generated at the time of on-off state transition can also be reduced.

Further, as a direct on-off control system in which the direct torque control is improved, model predictive control is known. The model predictive control is a method of calculating the current flowing through the rotating electric machine and the torque or magnetic flux generated in all candidate on-off states of the power conversion unit based on the state equation of the rotating electric machine, and determining the on-off state based on the calculated values. By controlling the on-off state based on the predicted value of the driving state of the rotating electrical machine in this way, the time constant of the current, torque, and magnetic flux in the transient state can be increased, the current or torque ripple in the steady state can be reduced, the number of times of switching the on-off state can be reduced, and the on-off loss can be reduced, as compared with the PWM system and the direct torque control.

However, in order to calculate the values of the rotating electrical machines in all the candidate on-off states, it is necessary to calculate the state equation of the rotating electrical machines, and the calculated amount becomes very large. Further, there is a disadvantage that the rotating electrical machine used has a large number of parameters and is easily affected by parameter errors.

For example, in the prior art described in patent document 1 below, a direct on-off control method based on a value obtained by integrating a voltage value is studied. In this control method, the error between the voltage command value vector and the voltage output vector is integrated, and when the integrated value exceeds a boundary circle set for the voltage command value vector, the voltage output vector in the direction closest to the center of the boundary circle is output. Therefore, in order to determine the on-off state so that the voltage integration error of the voltage command vector and the voltage output vector stays in the boundary circle for a long period of time, the number of on-off transitions of the switching element in the power conversion unit is suppressed to a necessary minimum, and the output voltage is controlled to have a sinusoidal waveform.

In addition, in the prior art described in patent document 2 below, a direct on-off control system of magnetic flux by a rotating electrical machine is studied. The control method sets a reference rotating magnetic flux of a rotating electric machine in advance, and outputs a zero voltage vector and a non-zero voltage vector appropriately so that the calculated deviation of the magnetic flux falls within an allowable range set for the reference rotating magnetic flux. Therefore, since the magnetic flux of the rotating electrical machine is controlled so as to follow the reference rotating magnetic flux set by the circumference, the iron loss and copper loss and torque ripple caused by the harmonic current flowing through the rotating electrical machine can be suppressed.

In the prior art described in patent document 3 below, a direct on-off control scheme is studied by direct torque control based on model predictive control. The control method predicts the torque and the stator magnetic flux of the rotating electrical machine in a predetermined section based on a state equation of the rotating electrical machine, and searches for an on-off pattern (a combination of a plurality of on-off states) in which the number of on-off transitions of each phase switching element is minimized while the torque and the stator magnetic flux satisfy a desired allowable value in the predicted section. Therefore, in a steady state, the number of on-off transitions is minimized under the conditions of desired torque and pulsation of the stator magnetic flux. In addition, in a transient state such as a step command related to torque, the on-off state most following the torque command is selected from among predicted values of torque and stator magnetic flux of the rotating electrical machine, so that a high-speed torque response time can be realized.

Patent document 1: japanese patent laid-open No. 11-89244

Patent document 2: japanese patent laid-open No. 59-025592

Patent document 3: japanese patent laid-open publication No. 2011-152038

Disclosure of Invention

In the direct on-off control system based on the value obtained by integrating the voltage value described in patent document 1, when the integrated value of the error between the voltage command vector and the voltage output vector exceeds the boundary circle set for the voltage command value vector, the on-off state is switched so that the number of on-off transitions becomes small. However, since the voltage output vector in the direction closest to the center of the boundary circle is always selected, the zero voltage vector cannot be positively selected, and the effect of reducing the number of on-off state transitions is limited.

The direct on-off control method based on the magnetic flux of the rotating electric machine described in patent document 2 appropriately selects the zero voltage vector and the non-zero voltage vector so that the magnetic flux of the rotating electric machine falls within a circular allowable range set for the reference rotating magnetic flux. Therefore, the number of on-off transitions is easily reduced by selecting the zero voltage vector as compared with patent document 1, but the voltage vector to be selected needs to be switched for each region including the magnetic flux of the rotating electrical machine, and thus the table for determining the on-off state becomes complicated. Further, since whether or not to switch the voltage vector is determined based on the circular allowable range set for the vector of the reference rotating magnetic flux, the calculated amount at the time of switching determination becomes large.

Further, in the direct torque control by the model predictive control described in patent document 3, the on-off state is determined by calculating the predicted values of the stator magnetic flux and the torque of the rotating electric machine to be controlled, compared with patent documents 1 and 2, so that the on-off loss in the steady state can be reduced while maintaining the high-speed torque response time in the transient state.

The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a power conversion device that reduces the on-off loss of a power conversion unit that can be mounted in an inexpensive microcomputer by determining an on-off state amount based on a value obtained by integrating voltage values of respective phases of a multi-phase alternating current obtained from the on-off state amount of the power conversion unit and a value obtained by integrating a voltage command value.

The disclosed power conversion device is provided with:

a power conversion unit having a plurality of switching elements, which converts dc power into ac power and supplies the ac power to a load in accordance with the on-off state amounts of the switching elements;

a voltage output calculation unit that calculates a voltage output value of the multi-phase ac supplied from the power conversion unit based on the on-off state quantity;

an integrated value calculation unit that calculates a voltage command integrated value and a voltage output integrated value by integrating the voltage command value of the multi-phase ac and the voltage output value of the multi-phase ac calculated by the voltage output calculation unit, respectively; and

and an on-off determining unit that determines and outputs an on-off state quantity of the power converting unit using the voltage command integrated value and the voltage output integrated value.

ADVANTAGEOUS EFFECTS OF INVENTION

According to the power conversion device disclosed in the present invention, it is possible to provide a power conversion device that determines the on-off state quantity based on a value obtained by integrating the voltage values of each phase of the multi-phase ac obtained from the on-off state quantity of the power conversion unit and a value obtained by integrating the voltage command value, thereby reducing the on-off loss of the power conversion unit that can be mounted in an inexpensive microcomputer.

Drawings

Fig. 1 is a block diagram showing a configuration of a power conversion device according to embodiment 1.

Fig. 2 is a hardware configuration diagram of the power conversion device according to embodiment 1.

Fig. 3 is a diagram showing all candidate on-off states of the power conversion unit according to embodiment 1.

Fig. 4 is a diagram showing a relationship between an on-off state index of the power conversion unit according to embodiment 1 and the multiphase voltage output value Vout.

Fig. 5 is a diagram for explaining a method of determining update of the on-off state quantity of the power conversion unit according to embodiment 1 and a method of determining the same.

Fig. 6 is a flowchart showing an example of the operation of the power conversion device according to embodiment 1.

Fig. 7 is a block diagram showing the structure of a power conversion device according to embodiment 2.

Fig. 8 is a diagram for explaining a method of determining update of the on-off state quantity of the power conversion unit and a method of determining the same according to embodiment 2.

Fig. 9 is a flowchart showing an example of the operation of the power conversion device according to embodiment 2.

Fig. 10 is a block diagram showing the structure of a power conversion device according to embodiment 3.

Fig. 11 is a hardware configuration diagram of a power conversion device according to embodiment 3.

Fig. 12 is a diagram for explaining a method of determining update of the on-off state quantity of the power conversion unit and a method of determining the same according to embodiment 3.

Fig. 13 is a flowchart showing an example of the operation of the power conversion device according to embodiment 3.

Fig. 14 is a block diagram showing a configuration of a power conversion device according to embodiment 4.

Fig. 15 is a block diagram showing the configuration of a speed estimation calculation unit according to embodiment 4.

Fig. 16 is a hardware configuration diagram of a power conversion device according to embodiment 4.

Fig. 17 is a diagram for explaining a method of determining update of the on-off state quantity of the power conversion unit and a method of determining the same according to embodiment 4.

Fig. 18 is a flowchart showing an example of the operation of the power conversion device according to embodiment 4.

Fig. 19 is a block diagram showing a configuration of a power conversion device according to embodiment 5.

Fig. 20 is a block diagram illustrating machine learning based on trained models and teacher data according to embodiment 5.

Fig. 21 is a hardware configuration diagram for generating a trained model according to embodiment 5.

Fig. 22 is a flowchart showing an example of the operation of the power conversion device according to embodiment 5.

Fig. 23 is a block diagram showing a configuration of a power conversion device according to embodiment 6.

Fig. 24 is a diagram for explaining a method of calculating the on-off state amount and the duration of the on-off state amount of the power conversion device according to embodiment 6.

Fig. 25 is a flowchart showing an example of the operation of the power conversion device according to embodiment 6.

Fig. 26 is a block diagram showing a configuration of a power conversion device according to embodiment 7.

Fig. 27 is a diagram for explaining a method of calculating current harmonic data from a detected current value according to embodiment 7.

Fig. 28 is a diagram for explaining a method of calculating current low harmonic data from a detected current value according to embodiment 7.

Fig. 29 is a flowchart showing an example of the operation of the power conversion device according to embodiment 7.

Fig. 30 is a block diagram showing the structure of a power conversion device according to embodiment 8.

Fig. 31 is a hardware configuration diagram of a power conversion device according to embodiment 8.

Fig. 32 is a diagram showing all candidate on-off state amounts of the power conversion unit according to embodiment 8.

Fig. 33 is a diagram illustrating an on-off pattern according to embodiment 8.

Fig. 34 is a diagram illustrating the integrated values of all candidate voltage prediction values using the electric power output integrated value as an initial value according to embodiment 8.

Fig. 35 is a diagram illustrating a trace of the voltage command integrated value and the integrated value of the voltage predicted value of the 60-degree phase amount across the power command value according to embodiment 8.

Fig. 36 is a flowchart showing an example of the operation of the power conversion device according to embodiment 8.

Fig. 37 is a block diagram showing the structure of a power conversion device according to embodiment 9.

Fig. 38 is a diagram illustrating a trace of the integrated value of the voltage predicted value and a value obtained by setting an allowable value to the voltage command integrated value of the phase amount equal to or greater than 60 degrees across the power command value of the power conversion device according to embodiment 9.

Fig. 39 is a flowchart showing an example of the operation of the power conversion device according to embodiment 9.

Fig. 40 is a block diagram showing the structure of a power conversion device according to embodiment 10.

Fig. 41 is a hardware configuration diagram of a power conversion device according to embodiment 10.

Fig. 42 is a diagram illustrating a trace of the integrated value of the voltage prediction value and the value obtained by setting the allowable value to the voltage command integrated value of the phase amount equal to or greater than 60 degrees across the power command value according to embodiment 10.

Fig. 43 is a flowchart showing an example of the operation of the power conversion device according to embodiment 10.

Fig. 44 is a block diagram showing a configuration of a power conversion device according to embodiment 11.

Fig. 45 is a block diagram illustrating machine learning based on trained models and teacher data used in the power conversion device according to embodiment 11.

Fig. 46 is a hardware configuration diagram for generating a trained model used in the power conversion device according to embodiment 11.

Fig. 47 is a flowchart showing an example of the operation of the power conversion device according to embodiment 11.

Fig. 48 is a block diagram showing the structure of a power conversion device according to embodiment 12.

Fig. 49 is a block diagram showing a configuration of a speed estimation calculation unit of the power conversion device according to embodiment 12.

Fig. 50 is a flowchart showing an example of the operation of the power conversion device according to embodiment 12.

Detailed Description

Embodiment 1

The present invention relates to a power conversion device that converts dc power into ac power, and more particularly to a power conversion device that controls on/off states of a plurality of switching elements in a power conversion unit for supplying power to a rotating electrical machine. Next, a power conversion device according to embodiment 1 of the present invention will be described with reference to the drawings.

Fig. 1 is a block diagram showing a configuration of a power conversion device 100 according to embodiment 1.

As shown in fig. 1, a power conversion device 100 includes a power conversion unit 1 as a main circuit, and a control device 10 that controls output of the power conversion unit 1, and the power conversion device 100 is connected between a dc power supply 2 and a load 3.

The power conversion unit 1 converts dc power from the dc power supply 2 into ac power, supplies the ac power to the load 3, and drives the load 3. The load 3 is driven by ac power supplied from the power conversion unit 1. As the load 3, various motors such as a transformer, a reactor, an induction motor, and a synchronous motor can be used.

The control device 10 includes: a voltage output calculation unit 11 that calculates a multiphase voltage output value Vout output from the power conversion unit 1 to the load 3 based on the on-off state amounts SWS of the plurality of switching elements of the power conversion unit 1; an integrated value calculation unit 12 that integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an on-off update determination unit 13 that calculates an update signal Snew that determines whether or not to update the on-off state amounts SWS of the plurality of switching elements of the power conversion unit 1, based on the voltage command integrated value Pref, the voltage output integrated value Pout, and an allowable range Δpref set for the voltage command integrated value Pref (here, the allowable range is not limited to 1 value, and may be determined by a region having a width, hereinafter, the allowable range is referred to as an allowable range Δpref because both meanings are included); and an on-off decision table 14 for deciding on-off state amounts SWS of the plurality of switching elements of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew. The on/off state amount SWS of the power conversion unit 1 is determined by the on/off update determination unit 13 and the on/off determination table 14, and corresponds to an on/off determination unit 300 described later.

The on-off update determination unit 13 determines whether to update the on-off state amounts SWS of the plurality of switching elements of the power conversion unit 1 based on the voltage output integrated value Pout and the voltage permission value pdelt in which the permission range Δpref is set for the voltage instruction integrated value Pref. For example, when the on-off state quantity SWS of the power conversion unit 1 is updated, 1 is output as the update signal Snew, and when the on-off state quantity SWS is not updated, 0 is output as the update signal Snew. Details of the method of calculating the update signal Snew will be described later.

The on-off determination table 14 determines the on-off state quantity SWS of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew. Hereinafter, this determined on-off state amount will be referred to as a determined SWS. For example, when the update signal Snew is 1, the on-off state amount SWS of the power conversion unit 1 is updated, and when the update signal Snew is 0, the on-off state amount SWS of the power conversion unit 1 is not updated, that is, the on-off state amount SWS is maintained. Details of the on-off decision table 14 will be described later.

Fig. 2 is a hardware configuration diagram for realizing the power conversion device 100.

The power conversion unit 1 is configured by a three-phase inverter circuit that converts direct-current power of the direct-current power supply 2 into three-phase alternating-current power, and the power conversion unit 1 drives the load 3. The power conversion unit 1 includes a plurality of switching elements Q1 to Q6 each of which is connected in antiparallel with a diode D. The power conversion unit 1 is connected to input terminals of the respective phases of the load 3 via bus bars from connection points of the upper arm and the lower arm of the respective phases. In this case, the u-phase has switching elements Q1 and Q2, the v-phase has switching elements Q3 and Q4, and the w-phase has switching elements Q5 and Q6.

The control device 10 is composed of a processor 40 and a storage device 41.

The storage device 41 includes volatile storage devices (not shown) such as RAM (Random Access Memory) and nonvolatile auxiliary storage devices (not shown) such as HDD (Hard Disk Drive) and SSD (Solid State Drive). Further, as the nonvolatile auxiliary storage device, a flash memory may be used instead of the HDD.

The processor 40 executes a control program input from the storage device 41.

The storage device 41 has a secondary storage device and a volatile storage device. The control program 42 is input to the processor 40 from the auxiliary storage via the volatile storage.

The processor 40 outputs the processing data 43 such as the operation result to the volatile memory device of the storage device 41, and stores the processing data in the auxiliary storage device via the volatile memory device as necessary.

As described above, the control device 10 outputs the on-off state amounts SWS of the plurality of switching elements Q1 to Q6 of the power conversion unit 1 to control the power conversion unit 1.

Fig. 3 is a diagram showing an example of the on-off state amounts of the plurality of switching elements in the case where two levels of the power conversion unit 1 are targeted. The on-off state quantity SWS is determined by a combination of signals of on (: 1) and off (: 0) of the respective switching elements Q1 to Q6. In this case, the combination is uniquely and correspondingly determined by the on/off parameter (the value of the level indicating the on/off state) indicated by the level 1 as the on/off state corresponding to the on/off state and the level 0 as the off/off state corresponding to the off/off state, and thus can be defined as an index indicating the on/off state.

In fig. 3, the combinations of the values of the on-off state levels of the switching elements Q1 and Q2, Q3 and Q4, and Q5 and Q6 defining the on-off state SWu of u-phase, the on-off state SWv of v-phase, and the on-off state SWw of w-phase are 9, and therefore, the distinction is made by representing them by the 9 on-off state indices SW0 to SW 8.

Specifically, the switching elements Q1 to Q6 of the upper arm and the lower arm have 9 kinds of on-off state amounts SWS (on-off state amounts corresponding to the on-off state indexes SW0, SW1, SW2, SW3, SW4, SW5, SW6, SW 7) of which one is turned on and the other is turned off, and on-off state amounts SWS (on-off state amounts corresponding to the on-off state index SW 8) of which all the switching elements Q1 to Q6 are turned off when the operation of the power conversion section 1 is stopped.

When the multiphase voltage is a three-phase voltage, the voltage output calculation unit 11 outputs a value V to the three-phase voltage based on the on-off state quantity SWS of the power conversion unit 1 shown in fig. 3, as shown in fig. 4 ₃ The values of the respective phase voltage outputs of out, i.e., u-phase voltage Vu, v-phase voltage Vv, w-phase voltage Vw are calculated. As shown in fig. 4, the values of the u-phase voltage Vu, v-phase voltage Vv, and w-phase voltage Vw are shown in correspondence with the on-off state indexes SW0 to SW 8. Here, vdc represents a bus voltage Vdc of the dc power supply 2.

Fig. 5 is a diagram for explaining a method of determining the on-off state amount SWS of the power conversion unit 1 by the on-off update determination unit 13 and the on-off determination table 14. In detail, fig. 5 is composed of two views, i.e., fig. 5A and fig. 5B, which is a partial enlarged view of fig. 5A. In fig. 5A and 5B, a voltage command integrated value Pref obtained by integrating the multiphase voltage command value Vref and a voltage output integrated value Pout obtained by integrating the multiphase voltage output value Vout are shown as typical examples of the case where the on-off state amount SWS of the power conversion unit 1 is determined in uvw coordinate system which is a three-phase stationary coordinate system.

Here, if it is assumed that the multiphase voltage command value Vref is in a steady state in the stationary coordinate system of the uvw three phases, it varies in accordance with the frequency of the voltage command value of the uvw three phases. For example, when the frequency is positive, the frequency changes circularly in the counterclockwise direction. Therefore, for the three-phase voltage command value V ₃ Three-phase voltage command integrated value P obtained by integrating ref ₃ ref also varies circularly in the counterclockwise direction.

Three-phase voltage output value V ₃ out is a value determined based on the on-off state quantity SWS of the power conversion unit 1 as shown in fig. 4, and is plotted on uvw coordinates in accordance with two values Vdc/2 and-Vdc/2 corresponding to the on-off state indexes SW0 to SW 7. The resultant vector of voltages of all the u-, v-, and w-phases becomes the voltage output value that is actually output. At this time, the on-off state meansThe resultant vector formed by the on-off state amounts SWs of the upper arms of the respective phases is zero for the respective on-off state amounts corresponding to the marks SW0, SW7, SW8, and therefore, the voltage output values of zero are 3 kinds, and the voltage output values of non-zero are 6 kinds.

As shown in fig. 5A and 5B, the three-phase voltage command value V is calculated by ₃ Three-phase voltage command integrated value P obtained by ref integration ₃ Allowable range Δp of ref setting ₃ ref, drawing the allowable range delta of the hexagon ₆ Pref. Since the integral value P is instructed for three-phase voltage ₃ The allowable range Δp is set for each phase of ref ₃ ref (in the figure, Δu represents the allowable range of u phase, Δv represents the allowable range of v phase, and Δw represents the allowable range of w phase), and thus the integrated value P is instructed for three-phase voltage ₃ ref sets the allowable range Δp ₃ Three-phase voltage allowable value P of ref ₃ delta hexagonal allowable range delta ₆ Pref. The on-off update determination unit 13 is based on the allowable range delta of the hexagon ₆ Pref, three-phase voltage output integral value P ₃ out, it is determined whether or not the on-off state amount SWS of the power conversion unit 1 needs to be updated. The method is described in further detail below.

As shown in fig. 5A and 5B, the three-phase voltage output integrated value P is set as a start point ₃ out and allowable range delta of hexagon ₆ The point at which the lower limit of the v phase of Pref crosses.

First, since the three-phase voltage outputs the integrated value P ₃ out and allowable range delta of hexagon ₆ The lower limit of the V phase of Pref crosses (the crossing point is point V _S ) Therefore, the on-off update determination unit 13 outputs 1 for updating the on-off state quantity SWS as the update signal Snew. Since the update signal Snew is 1, the on-off decision table 14 updates the on-off state quantity SWS.

At this time, the three-phase voltage output integral value P ₃ out reaches the lower limit value of v phase, which is not included in the three-phase voltage command integrated value P, and thus a non-zero voltage vector is output as the on-off state quantity SWS ₃ The allowable range Δ of a hexagon in which the traveling directions of the vectors of ref cross (the intersection point of both is the point V1i of fig. 5B) ₆ Pref's side (this side is side V1d of fig. 5B) and the adjacent sides (these sides are side V1c and side V1d of fig. 5B) are sides of these 3 sides (specifically, this side is side V1a of fig. 5B).

Here, the non-zero voltage vector is output as the three-phase voltage command integrated value P ₃ Allowable range delta of hexagon with crossed traveling direction of ref ₆ Pref (this side is side V1d of fig. 5B), and therefore, as the on-off state amount SWS1, the integrated value P is output to the three-phase voltage ₃ out is selected as the on-off state index SW3 that changes in the direction of the upper limit value of the V-phase (corresponding to the side V1d of fig. 5B).

Next, the three-phase voltage output integrated value P ₃ out reaches the upper limit value of u phase, which is not included in the three-phase voltage command integrated value P, and thus outputs a non-zero voltage vector as the on-off state quantity SWS ₃ The allowable range Δ of a hexagon in which the traveling directions of the vectors of ref cross (the intersection point of both is the point V2i of fig. 5B) ₆ Pref's side (this side is side V2c of fig. 5B) and the adjacent sides (these sides are side V2B and side V2d of fig. 5B) are sides of these 3 sides (specifically, this side is side V2f of fig. 5B).

Here, the non-zero voltage vector is output as the three-phase voltage command integrated value P ₃ Allowable range delta of hexagon with crossed traveling direction of ref ₆ Pref's side (specifically, this side is side V2c of fig. 5B) is the on-off state quantity SWS that varies, and therefore, as for the on-off state quantity SWS2, the integrated value P is output for the three-phase voltage ₃ out is selected by the on-off state index SW4 changing in the direction of the lower limit value of the u-phase (corresponding to the side V2c of fig. 5B).

Thereafter, the three-phase voltage output integral value P ₃ out arrival contains and three-phase voltage command integral value P ₃ The allowable range Δ of a hexagon in which the traveling directions of the vectors of ref cross (the intersection point of both is the point V3i of fig. 5B) ₆ Since Pref has a lower limit value of u-phase (corresponding to side V3c in fig. 5B) among 3 sides including the side and the adjacent side, the on-off state based on the on-off state index SW0 or SW7 that is the zero voltage vector is output as the on-off state amount SWS3 State quantity.

In summary, as shown in fig. 5A and 5B, the allowable range Δ of the hexagon as the starting time point ₆ When the lower limit value of the Pref v-phase crosses, the output value is outputted as an on-off state quantity SWS1 to the three-phase voltage command integrated value P ₃ Allowable range delta of hexagon with crossed traveling direction of ref ₆ The on-off state quantity corresponding to the on-off state index SW3 which changes in the direction of the upper limit value of the v-phase of Pref.

Then, when crossing the upper limit value of u phase, the three-phase voltage command integrated value P is outputted as the on-off state quantity SWS2 ₃ Allowable range delta of hexagon with crossed traveling direction of ref ₆ The on-off state quantity corresponding to the on-off state index SW4 which changes in the direction of the lower limit value of the u-phase of Pref.

When crossing the lower limit value of the u-phase, the on-off state quantity corresponding to the on-off state index SW0 or SW7, which is the zero-voltage vector, is output as the on-off state quantity SWS 3. That is, it can be said that in any case, the on-off state quantity is determined by a vector representing a change in the voltage command integrated value.

In this way, based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref of each phase, the on-off update determination unit 13 calculates the update signal Snew for determining whether to update the on-off state quantity SWS, and based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew, the on-off state quantity SWS of the power conversion unit 1 is determined in accordance with the on-off state index of the on-off determination table 14, whereby the on-off loss SWloss generated when the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 is transitioned can be reduced with a small calculation amount.

Next, a control operation in the power conversion device 100 according to embodiment 1 will be described in detail with reference to the drawings.

Fig. 6 is a flowchart illustrating a control operation in the power conversion device 100.

First, the voltage output calculation unit 11 calculates the multiphase voltage output value Vout based on the on-off state quantity SWS output from the on-off determination table 14 (step S1).

Next, the integrated value calculation unit 12 integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates the voltage command integrated value Pref and the voltage output integrated value Pout (step S2).

The on-off update determination unit 13 calculates an update signal Snew for updating the on-off state quantity SWS of the power conversion unit 1 as shown in fig. 5 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref set for the voltage command integrated value Pref (step S3).

The on-off determination table 14 determines the on-off state amount SWS of the power conversion unit 1 in accordance with the change in the on-off state amount as shown in fig. 5 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew (step S4).

The power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off determination table 14, and outputs the ac power to the load 3 (step S5).

The load 3 is driven and controlled by the ac power output from the power conversion unit 1 (step S6).

As described above, the power conversion device 100 according to embodiment 1 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a voltage output calculation unit 11 that calculates a multiphase voltage output value Vout based on the on-off state quantity SWS output from the on-off determination table 14; an integrated value calculation unit 12 that integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an on-off update determination unit 13 that determines whether or not to update the on-off state quantity SWS of the power conversion unit 1, based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref set for the voltage command integrated value Pref; and an on-off decision table 14 for deciding the on-off state quantity SWS of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, the traveling direction of the vector of the voltage command integrated value Pref calculated from the allowable range Δpref, and the update signal SnewThe on-off update determination unit 13 outputs an integrated value P at the three-phase voltage ₃ out arbitrary phase arrival is obtained by commanding the integrated value P to three-phase voltage ₃ ref sets a hexagonal allowable range delta composed of voltage allowable values Pdelta of the allowable range delta Pref ₆ In the case of any side of Pref, 1 is output as the update signal Snew, and in the other cases, 0 is output as the update signal Snew, and the on-off determination table 14 updates the on-off state amount SWS of the power conversion unit 1 when the update signal Snew is 1.

The on-off determination table 14 sets the allowable range Δ of the hexagon intersecting the traveling direction of the vector of the voltage command integrated value Pref in the voltage output integrated value Pout ₆ When 3 sides including Pref and adjacent sides are reached, a zero voltage vector is outputted as the on-off state quantity SWS of the power conversion unit 1, and when the remaining 3 sides are reached, a non-zero voltage vector in which the voltage output integrated value Pout changes in the traveling direction of the vector of the voltage command integrated value Pref is outputted as the on-off state quantity SWS of the power conversion unit 1.

Therefore, the power conversion device 100 according to embodiment 1 causes the voltage output integrated value Pout obtained by integrating the multiphase voltage output value Vout to follow the voltage command integrated value Pref obtained by integrating the multiphase voltage command value Vref, calculates the update determination of the on-off state quantity SWS of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the value of each phase of the allowable range Δpref, and determines the on-off state quantity SWS of the power conversion unit 1 based on the traveling direction of the vector of the voltage command integrated value Pref calculated based on the voltage command integrated value Pref, the update signal Snew, and thus can control the driving of the load 3 so as to reduce the on-off loss SWloss of the power conversion unit 1 by a calculated amount that can be provided in an inexpensive microcomputer.

Embodiment 2

Next, a power conversion device according to embodiment 2 will be described with reference to fig. 7. Here, fig. 7 is a block diagram showing the configuration of a power conversion device 100A according to embodiment 2.

As shown in fig. 7, the on-off determination table 14A of the power conversion device 100A differs from the on-off determination table 14 of the power conversion device 100 according to the above embodiment 1 in the selection method when a non-zero voltage vector is output as the on-off state quantity SWS of the power conversion unit 1. The same reference numerals are given to constituent elements having the same functions as those of embodiment 1, and description thereof will be omitted, focusing on differences from embodiment 1.

As shown in fig. 7, in the power conversion device 100A according to embodiment 2, an on-off decision table 14A included in the control device 10A is provided instead of the on-off decision table 14 included in the control device 10 according to embodiment 1. The on/off state amount SWS of the power conversion unit 1 is determined by the on/off update determination unit 13 and the on/off determination table 14A, and thus corresponds to the on/off determination unit 300 described later.

The on-off determination table 14A determines the on-off state amount SWS of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew. Details of the method for determining the on-off state amount SWS will be described later.

Fig. 8 illustrates a method of determining the on-off state amount SWS of the power conversion unit 1 by the on-off update determination unit 13 and the on-off determination table 14A. In fig. 8, the voltage command integrated value Pref and the voltage output integrated value Pout are set to determine the on-off state quantity SWS of the power conversion unit 1 in the uvw coordinates which are the stationary coordinates of three phases. Since the method of determining whether or not to update the on-off state amount SWS in the on-off update determination unit 13 is the same as that of embodiment 1, a description thereof will be omitted here.

The on-off determination table 14A in embodiment 2 differs from the on-off determination table 14 in embodiment 1 in the selection method of the non-zero voltage vector output as the on-off state quantity SWS.

In embodiment 1, the integrated value P is outputted at the three-phase voltage ₃ out arrival does not include and include the three-phase voltage command integral value P ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ In the case of 3 edges including Pref and adjacent edges, a non-zero output is obtainedThe voltage vector is outputted as the on-off state quantity SWS of the power conversion section 1 as a non-zero voltage vector varying in the traveling direction of the vector of the voltage command integrated value Pref.

In contrast, the non-zero voltage vector output of embodiment 2 is in the phase direction in which the voltage output integrated value Pout reaches and within the hexagonal allowable range Δ ₆ A non-zero voltage vector that varies within Pref. In fig. 8, the three-phase voltage output integrated value P is set as the start time point ₃ out and allowable range delta of hexagon ₆ The time point at which the upper limit value of the u-phase of Pref crosses.

First, since the three-phase voltage outputs the integrated value P ₃ out and allowable range delta of hexagon ₆ Pref crosses, so the on-off update determination section 13 outputs 1 as the update signal Snew. Since the update signal Snew is 1, the on-off decision table 14A updates the on-off state quantity SWS.

At this time, an integrated value P is output due to the three-phase voltage ₃ out reaches a value other than the integrated value P of the three-phase voltage command ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ The upper limit value of the u-phase of the 3 sides including the side of Pref and the adjacent side, therefore, a non-zero voltage vector is output as the on-off state quantity SWS.

For non-zero voltage vectors, the output is in the direction of the arriving phase and within the hexagonal tolerance range delta ₆ Non-zero voltage vector varying in Pref, and thus the on-off state quantity SWS1 output and three-phase voltage output integrated value P ₃ out is within the hexagonal allowable range delta ₆ The on-off state quantity corresponding to the on-off state index SW4 changing in the direction of the lower limit value of the u-phase of Pref.

Next, an integrated value P is output due to the three-phase voltage ₃ out reaches a value other than the integrated value P of the three-phase voltage command ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ The lower limit value of v phase of the side of Pref and the sides of 3 sides including the adjacent side, therefore, a non-zero voltage vector is output as the on-off state quantity SWS.

For non-zero voltage vectors, due to outputIn the direction of the arriving phases and within the hexagonal tolerance range delta ₆ Non-zero voltage vector varying in Pref, and thus on-off state quantity SWS2 output and three-phase voltage output integrated value P ₃ out is within the hexagonal allowable range delta ₆ The on-off state quantity corresponding to the on-off state index SW3 that varies in the upper limit value direction of the v-phase of Pref.

Next, an integrated value P is output due to the three-phase voltage ₃ out reaches a value other than the integrated value P of the three-phase voltage command ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ The upper limit value of the u-phase of the 3 sides including the side of Pref and the adjacent side, therefore, a non-zero voltage vector is output as the on-off state quantity SWS.

For non-zero voltage vectors, the output is in the direction of the arriving phase and within the hexagonal tolerance range delta ₆ Non-zero voltage vector varying in Pref, and thus on-off state quantity SWS3 output and three-phase voltage output integrated value P ₃ out is within the hexagonal allowable range delta ₆ The on-off state quantity corresponding to the on-off state index SW4 changing in the direction of the lower limit value of the u-phase of Pref.

Thereafter, the three-phase voltage output integral value P ₃ out arrival contains and three-phase voltage command integral value P ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ The lower limit value of u-phase in 3 sides including the side of Pref and the adjacent side is outputted as the on-off state quantity SWS4 as the on-off state quantity corresponding to the on-off state index SW0 or SW7 which becomes the zero voltage vector.

Therefore, in embodiment 2, as compared with the on-off determination method of embodiment 1, as in the transition from the on-off state quantity SWS1 to SWS2 in fig. 8, since the non-zero voltage vector is output so that the error of the voltage output integrated value Pout becomes smaller than the voltage command integrated value Pref, the harmonic voltage Vthd (also referred to as the harmonic component Vthd of the voltage) and the harmonic current Ithd caused by the harmonic voltage Vthd are suppressed without increasing the calculation consumption.

Next, a control operation in the power conversion device 100A according to embodiment 2 will be described in detail with reference to fig. 9. Fig. 9 is a flowchart illustrating a control operation in the power conversion device 100A.

First, the on-off update determination unit 13 outputs an update signal Snew for updating the on-off state quantity SWS of the power conversion unit 1 based on the allowable range Δpref, the voltage command integrated value Pref calculated by the voltage output calculation unit 11 and the integrated value calculation unit 12, and the voltage output integrated value Pout in the same flow as in embodiment 1 (steps S1 to S3).

Next, the on-off determination table 14A determines the on-off state quantity SWS of the power conversion unit 1 in accordance with the change of the on-off state quantity as shown in fig. 8 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew (step S7).

Then, the power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off determination table 14A, and outputs the ac power to the load 3, thereby controlling the driving of the load 3 (step S8 and step S9). Here, the operation of step S8 is the same as the operation of step S5 shown in fig. 6, and the operation of step S9 is the same as the operation of step S6 shown in fig. 6.

The power conversion device 100A according to embodiment 2 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a voltage output calculation unit 11 that calculates a multiphase voltage output value Vout based on the on-off state quantity SWS output from the on-off determination table 14A; an integrated value calculation unit 12 that integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an on-off update determination unit 13 that determines whether or not to update the on-off state quantity SWS of the power conversion unit 1, based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref set for the voltage command integrated value Pref; and an on-off determination table 14A for determining the on-off state quantity SWS of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, the traveling direction of the vector of the voltage command integrated value Pref calculated from the voltage command integrated value Pref, and the update signal Snew, and on-off update determination Section 13 outputs integrated value P at three-phase voltage ₃ out arbitrary phase arrival is obtained by commanding the integrated value P to three-phase voltage ₃ ref sets a hexagonal allowable range delta composed of voltage allowable values Pdelta of the allowable range delta Pref ₆ In the case of any side of Pref, 1 is output as the update signal Snew, and in the other cases, 0 is output as the update signal Snew, and the on-off determination table 14A updates the on-off state amount SWS of the power conversion unit 1 when the update signal Snew is 1.

The on-off determination table 14A sets the allowable range Δ of the hexagon intersecting the traveling direction of the vector of the voltage command integrated value Pref when the voltage output integrated value Pout reaches the allowable range Δ ₆ In the case of 3 sides including Pref and adjacent sides, a zero voltage vector is outputted as the on-off state quantity SWS of the power conversion unit 1, and in the case of reaching the remaining 3 sides, a voltage output integrated value Pout is outputted in the phase direction in which the voltage output integrated value Pout reaches and within the hexagonal allowable range Δ ₆ The non-zero voltage vector varying in Pref is used as the on-off state quantity SWS of the power conversion section 1.

Therefore, in the power conversion device 100A according to embodiment 2, compared with embodiment 1, an error between the voltage command integrated value Pref and the voltage output integrated value Pout can be reduced, and therefore the harmonic voltage Vthd and the harmonic current Ithd due to the harmonic voltage Vthd can be suppressed without increasing calculation consumption.

Embodiment 3

Next, a power conversion device according to embodiment 3 will be described with reference to fig. 10. Here, fig. 10 is a block diagram showing the structure of a power conversion device 100B according to embodiment 3.

As shown in fig. 10, the power conversion device 100B further includes: a bus voltage detection unit 15 for detecting a bus voltage Vdc of the power conversion unit 1; and an offset adjustment unit 16 for adjusting the offset values of the voltage command integrated value Pref and the voltage output integrated value Pout, wherein the power conversion device 100B differs from the power conversion device 100A according to embodiment 2 in that the voltage output calculation unit 11A calculates the multiphase voltage output value Vout based on the bus voltage Vdc detected by the bus voltage detection unit 15 and the on-off state quantity SWS output by the on-off decision table 14B, and the on-off update determination unit 13A determines whether to update the on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp calculated by the offset adjustment unit 16, and the on-off decision table 14B outputs a non-zero voltage vector as a selection method when the on-off state quantity SWS of the power conversion unit 1.

The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 and 2, and description thereof will be omitted, focusing on differences from embodiments 1 and 2.

As shown in fig. 10, the power conversion device 100B according to embodiment 3 further includes a bus voltage detection unit 15, as compared with the power conversion device 100A according to embodiment 2, instead of the voltage output calculation unit 11, the on-off update determination unit 13, and the on-off determination table 14A included in the control device 10A according to embodiment 2, the power conversion device 100B according to embodiment 3 includes the voltage output calculation unit 11A, the on-off update determination unit 13A, and the on-off determination table 14B included in the control device 10B, and the control device 10B includes the offset adjustment unit 16. The on/off state amount SWS of the power conversion unit 1 is determined by the on/off update determination unit 13A and the on/off determination table 14B, and corresponds to the on/off determination unit 300 described later.

First, the functions of the voltage output calculating unit 11A, the on-off update determining unit 13A, the on-off determination table 14B, the bus voltage detecting unit 15, and the offset adjusting unit 16 in the power converting apparatus 100B of embodiment 3, which are differences from the power converting apparatus 100A of embodiment 2, will be described.

The bus voltage detection unit 15 detects the bus voltage Vdc in the power conversion unit 1.

The voltage output calculation unit 11A calculates the multiphase voltage output value Vout based on the on-off state quantity SWS output from the on-off determination table 14B and the busbar voltage Vdc detected by the busbar voltage detection unit 15.

Offset adjustment unit 16 adjusts the offset value of each of voltage command integrated value Pref and voltage output integrated value Pout to 0, and outputs corrected voltage command integrated value Prefcomp and corrected voltage output integrated value Poutcomp (step S13).

The on-off update determination unit 13A outputs an update signal Snew for determining whether to update the on-off state quantity SWS of the power conversion unit 1, based on the corrected voltage command integral value Prefcomp, the corrected voltage output integral value Poutcomp, and the allowable range Δpref set for the corrected voltage command integral value Prefcomp. For example, when the on-off state quantity SWS of the power conversion unit 1 is updated, 1 is output as the update signal Snew, and when the on-off state quantity SWS is not updated, 0 is output as the update signal Snew.

The on-off determination table 14B determines the on-off state amount SWS of the power conversion unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the update signal Snew. Details of the method for determining the on-off state amount SWS will be described later.

Fig. 11 is a diagram showing a hardware configuration for implementing the power conversion device 100B.

The difference from the power conversion device 100 according to embodiment 1 is that the hardware configuration of the power conversion device 100B according to embodiment 3 further includes a bus voltage detection unit 15. The bus voltage detection unit 15 is a means for detecting the bus voltage Vdc by measuring the voltage difference between the positive side (+) and the negative side (-) of the dc power supply 2.

Fig. 12 is a diagram for explaining a method of determining the on-off state amount SWS of the power conversion unit 1 by the on-off update determination unit 13A and the on-off determination table 14B. Fig. 12 shows a case where corrected voltage command integrated value Prefcomp and corrected voltage output integrated value Poutcomp determine on-off state quantity SWS of power conversion unit 1 in a uvw coordinate system which is a three-phase stationary coordinate system.

The method of determining whether or not to update the on-off state amount SWS in the on-off update determination unit 13A is merely to change the input value to the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp, and is the same as in embodiments 1 and 2, and therefore description thereof is omitted.

Next, differences from the on-off determination table 14B in embodiment 3 and the on-off determination table 14A in embodiment 2 will be described. The difference is a method of selecting a non-zero voltage vector to be output as the on-off state quantity SWS.

In embodiment 2, the integrated value P is outputted at the three-phase voltage ₃ out arrival does not include and include the three-phase voltage command integral value P ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ In the case of the side of Pref and the sides of 3 sides including the adjacent side, a non-zero voltage vector is output as the on-off state quantity SWS of the power conversion section 1, the non-zero voltage vector being output in the phase direction in which the voltage output integrated value Pout reaches and within the hexagonal allowable range Δ ₆ A non-zero voltage vector that varies within Pref.

In contrast, the non-zero voltage vector according to embodiment 3 outputs a non-zero voltage vector that changes toward any one of two sides including a side that is close to a side that intersects the traveling direction of the vector of the corrected voltage command integrated value Prefcomp.

In fig. 12, as a start time point, the corrected three-phase voltage output integrated value P is set ₃ Allowable range delta of outcomp and hexagon ₆ The time point at which the upper limit value of the u-phase of Pref crosses.

First, an integrated value P is output due to the corrected three-phase voltage ₃ Allowable range delta of outcomp and hexagon ₆ Pref crosses, so the on-off update determination section 13A outputs 1 as the update signal Snew. Since the update signal Snew is 1, the on-off decision table 14B updates the on-off state quantity SWS.

At this time, an integrated value P is output due to the corrected three-phase voltage ₃ outcomp reaches a value P which is not the three-phase voltage command integral value after the correction ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ The upper limit value of the u-phase of the 3 sides including the side of Pref and the adjacent side, therefore, a non-zero voltage vector is output as the on-off state quantity SWS.

For non-zero voltage vectors, the output is directed to the direction of travel of the vector including the sum-corrected voltage command integral value PrefcompSince the non-zero voltage vector that varies in either one of the two sides including the side where the side of the fork approaches, the on-off state amount SWS1 is output in the hexagonal allowable range Δ ₆ The on-off state quantity corresponding to the on-off state index SW4 changing in the direction of the lower limit value of the u-phase of Pref.

Next, an integrated value P is output due to the corrected three-phase voltage ₃ outcomp reaches a value P which is not the three-phase voltage command integral value after the correction ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ The upper limit value of w-phase of the side of 3 sides including the side of Pref and the adjacent side, therefore, a non-zero voltage vector is output as the on-off state quantity SWS.

Since the non-zero voltage vector is output to be changed toward either one of the two sides including the side that is close to the side intersecting the traveling direction of the vector of the corrected voltage command integrated value Prefcomp, the on-off state quantity SWS2 is output to be equal to the allowable range Δ in the hexagonal shape ₆ The on-off state quantity corresponding to the on-off state index SW3 that varies in the upper limit value direction of the v-phase of Pref.

Next, an integrated value P is output due to the corrected three-phase voltage ₃ outcomp reaches a value P which is not the three-phase voltage command integral value after the correction ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ The upper limit value of the u-phase of the 3 sides including the side of Pref and the adjacent side, therefore, a non-zero voltage vector is output as the on-off state quantity SWS.

As for the non-zero voltage vector, since the non-zero voltage vector that varies toward either one of the two sides including the side that is close to the side intersecting the traveling direction of the vector of the corrected voltage command integrated value Prefcomp is output, the allowable range Δ of the hexagon is output as the on-off state quantity SWS3 ₆ The on-off state quantity corresponding to the on-off state index SW4 changing in the direction of the lower limit value of the u-phase of Pref.

Thereafter, the corrected three-phase voltage output integrated value P ₃ outcomp arrival of the contained and corrected three-phase powerIntegral value P of pressure command ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ Since Pref has a lower limit value of u-phase in 3 sides including the adjacent side, an on-off state quantity corresponding to the on-off state index SW0 or SW7 which is a zero voltage vector is output as the on-off state quantity SWS 4.

Therefore, in embodiment 3, a non-zero voltage vector that varies in the phase direction (here, the direction parallel to a specific coordinate axis of uvw coordinate axes) is output at the time of the on-off state amount SWS1 in fig. 12, compared with the on-off determination method in embodiment 1, and therefore, the harmonic voltage Vthd and the harmonic current Ithd caused by the harmonic voltage Vthd are suppressed compared with embodiment 1.

In embodiment 3, a non-zero voltage vector that varies in the direction of the upper limit value of the v-phase is output to transition from the on-off state quantity SWS1 to the on-off state quantity SWS2 in fig. 12, compared with the on-off determining method of embodiment 2, the harmonic voltage Vthd and the harmonic current Ithd caused by the harmonic voltage Vthd are further suppressed compared with the method of embodiment 2 in which the output of the non-zero voltage vector varies in the direction of the phase in which the voltage output integrated value Pout reaches. The calculation consumption in the on-off determination method is also substantially the same as that in embodiments 1 and 2.

Next, a control operation in the power conversion device 100B according to embodiment 3 will be described in detail with reference to fig. 13. Fig. 13 is a flowchart illustrating a control operation in the power conversion device 100B (see fig. 10).

In fig. 13, first, the bus voltage detection unit 15 detects the bus voltage Vdc of the power conversion unit 1 (step S10).

Next, the voltage output calculation unit 11A calculates the multiphase voltage output value Vout based on the bus voltage Vdc detected by the bus voltage detection unit 15 and the on-off state quantity SWS output by the on-off determination table 14B (step S11).

Next, the processing is performed in the same flow as in embodiments 1 and 2, and the integrated value calculation unit 12 integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates the voltage command integrated value Pref and the voltage output integrated value Pout (step S12).

Offset adjustment unit 16 outputs voltage command integrated value Pref and voltage output integrated value Pout, each of which has an offset value adjusted to zero (0), as corrected voltage command integrated value Prefcomp and corrected voltage output integrated value Poutcomp.

The on-off update determination unit 13A outputs an update signal Snew for updating the on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the allowable range Δpref (step S14).

The on-off determination table 14B determines the on-off state quantity SWS of the power conversion unit 1 in accordance with the change in the on-off state quantity as shown in fig. 12 based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the update signal Snew (step S15).

Then, the power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off determination table 14B, and outputs the ac power to the load 3, thereby controlling the driving of the load 3 (step S16 and step S17).

The power conversion device 100B of embodiment 3 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a bus voltage detection unit 15 that detects a bus voltage Vdc of the power conversion unit 1; a voltage output calculation unit 11A that calculates a multiphase voltage output value Vout based on the busbar voltage Vdc detected by the busbar voltage detection unit 15 and the on-off state quantity SWS output by the on-off determination table 14B; an integrated value calculation unit 12 that integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an offset adjustment unit 16 that adjusts the offset values of the voltage command integrated value Pref and the voltage output integrated value Pout to 0, and outputs the voltage command integrated value Prefcomp and the voltage output integrated value Poutcomp after correction; an on-off update determination unit 13A based on the corrected data Voltage command integrated value Prefcomp, corrected voltage output integrated value Poutcomp, and allowable range Δpref set for corrected voltage command integrated value Prefcomp, and determining whether to update on-off state quantity SWS of power conversion unit 1; and an on-off decision table 14B for deciding the on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage command integral value Prefcomp, the corrected voltage output integral value Poutcomp, the traveling direction of the vector of the corrected voltage command integral value Prefcomp calculated from the corrected voltage command integral value Prefcomp, and the update signal Snew, the on-off update decision unit 13A determining the on-off state quantity SWS of the power conversion unit 1 based on the three-phase voltage output integral value P ₃ out arbitrary phase arrival is obtained by commanding the integrated value P to three-phase voltage ₃ ref sets a hexagonal allowable range delta composed of voltage allowable values Pdelta of the allowable range delta Pref ₆ In the case of any side of Pref, 1 is output as the update signal Snew, and in the other cases, 0 is output as the update signal Snew, and in the case where the update signal Snew is 1, the on-off determination table 14B updates the on-off state amount SWS of the power conversion unit 1.

The on-off decision table 14B sets the corrected voltage output integrated value Poutcomp to the allowable range Δ including a hexagon intersecting the traveling direction of the vector of the corrected voltage command integrated value Prefcomp ₆ When 3 sides including Pref and adjacent sides output zero voltage vectors as on-off state amounts SWS of the power conversion unit 1, and when the remaining 3 sides are reached, non-zero voltage vectors that change toward any one of two sides including a side that is close to a side intersecting a traveling direction of a vector of the corrected voltage command integrated value Prefcomp are output.

Therefore, the power conversion device 100B according to embodiment 3 can reduce the error between the voltage command integrated value Pref and the voltage output integrated value Pout, compared with embodiments 1 and 2, and therefore can suppress the harmonic voltage Vthd and the harmonic current Ithd due to the harmonic voltage Vthd without increasing the calculation consumption, and the value obtained by integrating the multiphase voltage values by the offset adjustment section 16 is always in multiphase balance, so that the performance of the update determination of the on-off state quantity SWS and the determination of the on-off state quantity SWS is not deteriorated.

Embodiment 4

Next, a power conversion device according to embodiment 3 will be described with reference to fig. 14. Here, fig. 14 is a block diagram showing the structure of a power conversion device 100C according to embodiment 3.

As shown in fig. 14, the power conversion device 100C differs from the power conversion devices 100, 100A, 100B of embodiments 1 to 3 in that the load 3 is replaced with the rotating electric machine 4, a current detection unit 17 for detecting a current flowing between the power conversion unit 1 and the rotating electric machine 4 is provided, as an angular velocity estimation means of the rotating electric machine 4, a speed estimation operation unit 21 which is an adaptive magnetic flux observer, a speed controller 19 and a current controller 20 for controlling an angular velocity and a current of the rotating electric machine 4 are provided, and an on-off state quantity is calculated based on a voltage command integrated value Prefcomp corrected for the allowable range Δpref and a corrected voltage output integrated value Poutcomp used in the on-off update determination unit 13B.

The following description will be given mainly on differences from embodiments 1 to 3, with the same reference numerals and omitted descriptions for the constituent elements having the same functions as those of embodiments 1 to 3.

Fig. 14 is a block diagram showing a configuration of a power conversion device 100C according to embodiment 4. As shown in fig. 14, a power conversion device 100C according to embodiment 4 differs from embodiment 3 in that a load 3 is replaced with a rotating electrical machine 4, a current detection unit 17 is provided between a power conversion unit 1 and the rotating electrical machine 4, and a control device 10C includes an on-off update determination unit 13B instead of the on-off update determination unit 13A, and further includes: a speed estimation calculation unit 21 for estimating the angular speed and phase of the rotating electrical machine 4; a speed controller 19 for controlling the angular speed ω of the rotating electric machine 4 _rm Performing control; a current controller 20 for controlling a current Ir of the rotating electric machine 4; and an allowable range calculation section 18 for calculating a newly set allowable range Δpref from the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp.

Next, the functions of the on-off update determination unit 13B, the current detection unit 17, the allowable range calculation unit 18, the speed controller 19, the current controller 20, and the speed estimation calculation unit 21 according to embodiment 4, which are differences from embodiment 3, will be described.

The current detection unit 17 detects a current flowing between the power conversion unit 1 and the rotating electrical machine 4.

The allowable range calculation unit 18 calculates the newly set allowable range Δpref based on the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp calculated by the offset adjustment unit 16.

Here, when calculating the allowable range Δpref, phase information calculated from the corrected voltage command integrated value Prefcomp may be used. The allowable range Δpref may be calculated as a function of an arbitrary value depending on the input value.

The on-off update determination unit 13B outputs an update signal Snew of whether or not to update the on-off state quantity SWS of the power conversion unit 1, based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the allowable range Δpref calculated by the allowable range calculation unit 18. For example, when the on-off state quantity SWS of the power conversion unit 1 is updated, 1 is output as the update signal Snew, and when the on-off state quantity SWS is not updated, 0 is output as the update signal Snew.

The speed controller 19 calculates an angular velocity estimated value hω by the velocity estimation calculation unit 21 based on the angular velocity command value ωrmref _rm Calculating angular velocity ω for rotating electrical machine _rm A current command value Iref to be controlled. As a method for calculating the current command value Iref, any method is used as long as the angular velocity ω of the rotating electrical machine is set _rm A method for matching the angular velocity command value ωrmref may be used, and a proportional-integral controller (PI controller) or a proportional-integral-derivative controller (PID controller) may be used.

The current controller 20 calculates a multiphase voltage command value Vref for controlling the current Ir flowing through the rotating electric machine based on the current command value Iref calculated by the speed controller 19, the detected current value Iuvw detected by the current detecting unit 17, and the phase estimated value hθ calculated by the speed estimating and calculating unit 21. As a method of calculating the multiphase voltage command value Vref, any method may be used as long as the current Ir flowing through the rotating electric machine matches the current command value Iref in the rotating coordinate system, and the multiphase voltage command value Vref may be calculated by coordinate conversion by calculating the voltage command value in the rotating coordinate system using a PI controller or a PID controller after coordinate conversion of the current Ir flowing through the rotating electric machine into the value in the rotating coordinate system.

The speed estimation calculation unit 21 calculates the angular speed ω of the rotating electrical machine based on the multiphase voltage command value Vref and the detected current value Iuvw of the current detection unit 17 _rm And the phase θ.

Fig. 15 is a block diagram showing the configuration of the speed estimation calculation unit 21. The speed estimation calculation unit 21 is constituted by an adaptive observer, and is configured to calculate the phase θ and the angular velocity ω of the rotating electrical machine 4 _rm An estimation operation is performed. The adaptive observer is also called an adaptive flux observer because it is defined by a state equation in which the stator flux Φs and the rotor flux Φr of the rotating electrical machine 4 are used as state variables. The adaptive observer may be configured using an extended induced voltage, a current, or the like as a state variable.

The speed estimation calculation unit 21 shown in fig. 15 uses the multiphase voltage command value Vref and the detected current value Iuvw, and estimates the angular speed value hω _rm And phase estimated value hθ, and outputting calculated angular velocity estimated value hω _rm And a phase estimation value hθ. The multiphase voltage command value Vref is a value calculated by the current controller 20, and the detected current value Iuvw is a value detected by the current detecting unit 17. Here, the multiphase voltage command value Vref is input to the speed estimation computing unit 21, but the voltage value output from the power conversion unit 1 may be detected, and the detected voltage output value may be set as the input value of the speed estimation computing unit 21.

The speed estimation computing unit 21 includes a model deviation computing unit 22, an angular speed estimator 23, a first-order angular frequency computing unit 24, and an integrator 25. The model deviation calculation unit 22 is based on multiple phasesVoltage command value Vref, detected current value Iuvw, first-order angular frequency ω ₁ Estimated value hω of angular velocity _rm The model deviation epsilon is calculated. The angular velocity estimator 23 estimates an angular velocity estimate hω based on the model deviation ε _rm And performing operation. The first-order angular frequency calculator 24 calculates a first-order angular frequency ω based on the magnetic flux estimated value hΦ and the current estimated value hi and the angular velocity estimated value ₁ And performing operation. The integrator 25 is for the first order angular frequency ω ₁ The integration is performed to output a phase estimation value hθ.

The model deviation calculation unit 22 includes a current estimator 221, a subtractor 222, and a deviation calculation unit 223. The current estimator 221 detects a current value Iuvw and a first-order angular frequency ω based on the multiphase voltage command value Vref ₁ Estimated value hω of angular velocity _rm The magnetic flux estimated value hΦ and the current estimated value hi are calculated, and the calculated magnetic flux estimated value hΦ and current estimated value hi are outputted. The subtractor 222 calculates a current deviation Ierr by subtracting the detected current value Iuvw from the current estimated value hi, and outputs the calculated current deviation Ierr.

The deviation calculator 223 calculates the model deviation epsilon based on the current deviation Ierr and the magnetic flux estimated value hΦ calculated by the subtractor 222. Here, when the current deviation Ierr is a vector and the magnetic flux estimated value hΦ is a vector, the vector of the current deviation Ierr is input, the orthogonal component of the vector of the magnetic flux estimated value hΦ is extracted as a scalar, and the extracted scalar is output as the model deviation epsilon. As the estimated value of the magnetic flux A method of extracting the orthogonal component of the vector of (a) as a scalar quantity, a method of transforming the vector of the current deviation Ierr onto a rotating orthogonal coordinate system, and a method of calculating the magnitude of the outer product value between the vector of the current deviation Ierr and the vector of the magnetic flux estimated value hΦ are known.

The current estimator 221 calculates a current estimated value hi and a magnetic flux estimated value hΦ from the state equation of the rotating electrical machine 4. Here, the rotating electric machine 4 is assumed to be a normal permanent magnet embedded synchronous motor, but any type of motor may be used as long as the state equation of the induction motor, the surface permanent magnet synchronous motor, the winding excitation synchronous motor, the reluctance synchronous motor, or the like can be satisfied. That is, the current estimator 221 can estimate the current of the rotating electrical machine other than the permanent magnet embedded synchronous motor in the same manner.

When the rotating electric machine 4 is a permanent magnet embedded synchronous motor, the state equations are expressed by the following equations (1) and (2). Here, "L _d "represents d-axis inductance," L _q "denotes the inductance of q-axis," i _d "means d-axis current," i _q "means q-axis current," Φ _ds "means d-axis stator flux," Φ _qs "means q-axis stator flux," Φ _dr The symbol "a" represents the d-axis rotor magnetic flux and the symbol "a" is the estimated value (for example, "hΦ", which is the estimated value of Φ, is the same as the other estimated values). In addition, "R _a "represents armature resistance," omega ₁ "means first order angular frequency," v _d "represents d-axis voltage," v _q "represents q-axis voltage," h ₁₁ "to" h ₃₂ "means observer gain.

In addition, the first-order angular frequency omega ₁ Given by the following formula (3). In formula (3), "h ₄₁ ”、“h ₄₂ "means observer gain.

[ mathematics 1]

[ math figure 2]

[ math 3]

/>

The above-described formulas (1) and (2) are formulas based on a normal induced voltage, but the same calculation can be performed even if the formulas (1) and (2) are deformed to express an extended induced voltage. The expression (1) and the expression (2) are mathematical expressions in dq coordinates of a rotating coordinate system, but the expression (1) and the expression (2) are transformed to be represented by another coordinate system such as an αβ coordinate system of two-phase communication or a uvw coordinate system of three-phase communication in a stationary coordinate system, and the same calculation can be performed. Since the above formula (1) contains the estimated angular velocity value hω _rm Therefore, at the angular velocity estimated value hω _rm Angular velocity ω from actual rotation _rm If the current estimated value hi does not match, an error occurs.

Here, the model deviation epsilon is defined by the following expression (4), and the angular velocity estimator 23 is used by the velocity estimation calculation unit 21 to estimate the angular velocity value hω _rm Adjustment is made so that the model deviation epsilon becomes zero. The angular velocity estimator 23 is configured by directly connecting an integrator to a PI controller, for example.

[ mathematics 4]

The first-order angular frequency calculator 24 is based on the above equation (3) and on the magnetic flux estimated value hΦ, the current estimated value hi, and the angular velocity estimated value hω _rm For first order angular frequency omega ₁ And performing operation. The integrator 25 is configured to integrate the first-order angular frequency omega ₁ The phase estimation value hθ is calculated by integrating. The adaptive observer has an advantage of robustness against a fluctuation in the number of flux links and does not generate a speed estimation error in a steady state. Therefore, the adaptive observer can efficiently observe the angular velocity ω of the rotating electrical machine 4 _rm Estimation is performed.

Fig. 16 is a hardware configuration diagram for realizing the power conversion device 100C.

In fig. 16, the hardware configuration of power conversion device 100C differs from power conversion device 100B of embodiment 3 in that current detection unit 17 is newly added between power conversion unit 1 and rotating electrical machine 4.

The current detection unit 17 outputs the three-phase current value I to the rotating electrical machine 4 to the power conversion unit 1 _3uvw And (5) detecting. Here, any current detector such as CT (Current Transformer) detector or shunt resistor may be used for the current detecting unit 17. It is also possible to use a method of detecting the current of two phases among the currents of three phases and calculating the current of the remaining one phase. In addition, a single-shunt current detection method in which the three-phase ac current value is restored by one current detector may be used.

Fig. 17 is a diagram for explaining a method of determining the on-off state amount SWS of the power conversion unit 1 by the on-off update determination unit 13B and the on-off determination table 14B. Fig. 17 shows a case where corrected voltage command integrated value Prefcomp and corrected voltage output integrated value Poutcomp determine on-off state quantity SWS of power conversion unit 1 in a uvw coordinate system which is a three-phase stationary coordinate system.

The on-off update determination unit 13B in embodiment 4 differs from the on-off update determination unit 13A in embodiment 3 in that the on-off update determination unit 13B is calculated based on the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp as an allowable range Δpref used in calculating the update signal Snew of whether to update the on-off state quantity SWS.

In embodiment 3, the corrected three-phase voltage command value V ₃ The allowable range Δpref of a hexagon drawn by ref setting the allowable range Δpref ₆ The size of Pref is always constant, but in embodiment 4, the allowable range Δ of the hexagon ₆ The size of Pref varies.

As shown in fig. 17, as a starting point, the corrected three-phase voltage output integrated value P is set ₃ Allowable range delta of outcomp and hexagon ₆ The point at which the upper limit of Pref's u-phase crosses.

First, an integrated value P is output due to the corrected three-phase voltage ₃ Allowable range delta of outcomp and hexagon ₆ Pref crosses, so the on-off update determination unit 13B outputsAnd out of 1 as an update signal Snew. The on-off determination table 14B performs processing in the same flow as in embodiment 3, and outputs the on-off state quantity SWS1 within the hexagonal allowable range Δ ₆ The on-off state quantity corresponding to the on-off state index SW4 changing in the direction of the lower limit value of the u-phase of Pref.

Here, if the allowable range Δ of the hexagon ₆ The magnitude of Pref is changed based on the phase of the corrected voltage command integrated value Prefcomp as shown in fig. 17, and after the on-off state quantity SWS1 is used, the corrected three-phase voltage output integrated value P ₃ outcomp arrival does not include the corrected three-phase voltage command integral value P ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ The lower limit value of v phase of 3 sides including the side of Pref and the adjacent side, therefore, the allowable range delta of the hexagon is outputted and calculated in terms of the on-off state quantity SWS2 ₆ The on-off state quantity corresponding to the on-off state index SW3 that varies in the upper limit value direction of the v-phase of Pref.

Next, the corrected three-phase voltage output integrated value P ₃ outcomp arrival does not include the corrected three-phase voltage command integral value P ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ The upper limit value of u-phase of 3 sides including the side of Pref and the adjacent side, therefore, the on-off state quantity SWS3 is outputted and outputted within the allowable range delta of the hexagon ₆ The on-off state quantity corresponding to the on-off state index SW4 changing in the direction of the lower limit value of the u-phase of Pref.

Thereafter, the corrected three-phase voltage output integrated value P ₃ outcomp reaches the three-phase voltage command integral value P after inclusion and correction ₃ Allowed range delta of hexagons where travel directions of vectors of refcomp intersect ₆ Since Pref has a lower limit value of u-phase in 3 sides including the adjacent side, an on-off state quantity corresponding to the on-off state index SW0 or SW7 which is a zero voltage vector is output as the on-off state quantity SWS 4.

Therefore, in embodiment 4, as compared with the on-off determination method of embodiment 3, the on-off state amount SWS1 is shifted to SWS2 as shown in fig. 17As in the case of the above, the three-phase voltage command integrated value P is set in the direction of the non-zero voltage vector change ₃ When the traveling direction of the vector of ref approaches, the allowable range Δpref becomes small, and the three-phase voltage command integrated value P can be reduced ₃ ref and three-phase voltage output integral value P ₃ out.

Therefore, with the on-off update determination method in embodiment 4, the harmonic voltage Vthd and the harmonic current Ithd due to the harmonic voltage Vthd are further suppressed as compared with embodiment 3.

Next, a control operation in the power conversion device 100C according to embodiment 4 will be described in detail with reference to fig. 18. Here, fig. 18 is a flowchart illustrating a control operation in the power conversion device 100C (see fig. 16).

First, in the same flow as in embodiment 3, the voltage output calculation unit 11A calculates the multiphase voltage output value Vout based on the bus voltage Vdc detected by the bus voltage detection unit 15 and the on-off state quantity SWS output by the on-off determination table 14B (step S20, step S21).

Next, the current detection unit 17 detects a current flowing between the power conversion unit 1 and the rotating electrical machine 4 (step S22).

The speed estimation calculation unit 21 calculates the angular speed ω of the rotating electric machine 4 based on the multiphase voltage command value Vref and the detection current value Iuvw detected by the current detection unit 17 _rm The estimated value of (a) and the estimated value of the phase θ are calculated (step S23).

The speed controller 19 and the current controller 20 calculate an angular velocity estimated value hω based on the angular velocity command value ωrmref and the velocity estimated calculation unit 21 _rm The phase estimated value hθ and the detected current value Iuvw are calculated to obtain the multiphase voltage command value Vref (step S24).

In the same flow as in embodiment 3, the integrated value calculation unit 12 and the offset adjustment unit 16 execute processing to calculate the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp (step S25, step S26).

The allowable range calculation unit 18 calculates the allowable range Δpref based on the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp (step S27).

The on-off update determination unit 13B outputs the update signal Snew based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the allowable range Δpref calculated by the allowable range calculation unit 18 (step S28).

The on-off determination table 14B determines the on-off state quantity SWS of the power conversion unit 1 in accordance with the change in the on-off state quantity as shown in fig. 17 based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the update signal Snew (step S29).

Then, the power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off determination table 14B, and outputs the ac power to the rotating electrical machine 4, thereby performing drive control of the rotating electrical machine 4 (steps S30 and S31).

The power conversion device 100C according to embodiment 4 includes: a power conversion unit 1 that converts direct-current power from a direct-current power supply 2 into alternating-current power and supplies the alternating-current power to a rotating electrical machine 4; a bus voltage detection unit 15 that detects a bus voltage Vdc of the power conversion unit 1; a voltage output calculation unit 11A that calculates a multiphase voltage output value Vout based on the busbar voltage Vdc detected by the busbar voltage detection unit 15 and the on-off state quantity SWS output by the on-off determination table 14B; a current detection unit 17 that detects a current flowing between the power conversion unit 1 and the rotating electrical machine 4; a speed estimation calculation unit 21 that performs estimation calculation of the angular speed and phase of the rotating electrical machine 4 based on the multiphase voltage command value Vref and the detected current value Iuvw; a speed controller 19 for estimating an angular speed h omega based on the angular speed command value ωrmref _rm Calculating a current command value Iref; a current controller 20 that calculates a multiphase voltage command value Vref based on the current command value Iref, the detection current value Iuvw, and the phase estimation value hθ; an integrated value calculating unit 12 for integrating the multiphase voltage command value Vref calculated by the current controller 20 and the multiphase voltage output value Vout calculated by the voltage output calculating unit 11ACalculating a voltage command integral Pref and a voltage output integral Pout; an offset adjustment unit 16 that adjusts the offset values of the voltage command integrated value Pref and the voltage output integrated value Pout to 0, and outputs the voltage command integrated value Prefcomp and the voltage output integrated value Poutcomp after correction; an allowable range calculation unit 18 that calculates an allowable range Δpref based on the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp; an on-off update determination unit 13B that determines whether to update the on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the allowable range Δpref calculated by the allowable range calculation unit 18; and an on-off decision table 14B for deciding the on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage command integral Prefcomp, the corrected voltage output integral Poutcomp, the traveling direction of the vector of the corrected voltage command integral Prefcomp calculated from the corrected voltage command integral Prefcomp, the update signal Snew, and the allowable range calculation unit 18 based on the corrected voltage command integral Prefcomp, the corrected voltage output integral Poutcomp, and if the corrected three-phase voltage command integral P ₃ When the traveling direction of the refcomp vector approaches the traveling direction of the non-zero voltage vector, the allowable range Δpref is calculated to be small, and the on-off update determination unit 13B determines whether to update the on-off state quantity SWS based on the allowable range Δpref calculated by the allowable range calculation unit 18, so that the three-phase voltage command integrated value P is reduced ₃ ref and three-phase voltage output integral value P ₃ The on-off state quantity SWS is updated by the way of the error between out.

Therefore, in the power conversion device 100C according to embodiment 4, since the update of the on-off state quantity SWS is determined so as to reduce the error between the voltage command integrated value Pref and the voltage output integrated value Pout, the harmonic voltage Vthd and the harmonic current Ithd caused by the harmonic voltage Vthd can be further suppressed as compared with embodiments 1 to 3.

Embodiment 5

Next, a power conversion device according to embodiment 5 will be described with reference to fig. 19. Here, fig. 19 is a block diagram showing a configuration of a power conversion device 100D according to embodiment 5.

As shown in fig. 19, the power conversion device 100C according to embodiment 4 is different from the control device 10C having the allowable range calculation unit 18 in that the control device 10D is provided with a control device 10D, and the control device 10D has a trained model 26, and performs calculation using the trained model 26 (data obtained by performing machine learning is output) in calculating the allowable range Δpref. The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 to 4, and description thereof will be omitted, focusing on differences from embodiments 1 to 4.

Fig. 19 is a block diagram showing a configuration of a power conversion device 100D according to embodiment 5. As shown in fig. 19, the power conversion device 100D according to embodiment 5 has a trained model 26 instead of the allowable range calculation unit 18, as compared with embodiment 4. Therefore, the following describes the difference from embodiment 4, that is, the function of the trained model 26 of embodiment 5.

The trained model 26 calculates the allowable range Δpref based on the information obtained by machine learning based on the teacher data, and based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, the angular velocity estimated value, and the phase estimated value.

Fig. 20 is a block diagram illustrating a method of creating a trained model and machine learning based on teacher data. As shown in fig. 20, the learning unit 50 performs machine learning based on teacher data 57 obtained from the learning data 51 prepared in advance to generate a trained model.

The learning data 51 includes a corrected voltage command integrated value Prefcomp, a corrected voltage output integrated value Poutcomp, and an angular velocity estimated value hω _rm A phase estimation value hθ, and an allowable range Δpref. The learning data 51 stores a value calculated to drive the rotating electric machine 4, and for example, a control method for reducing the on-off loss SWloss of the power conversion unit 1, that is, a control method for reducing the on-off loss SWloss, may be used as compared with the PWM method The learning data 51 is generated using a control scheme of model predictive control, selective harmonic cancellation, low order harmonic cancellation, or a control scheme using an optimal pulse mode.

As shown in fig. 20, learning data 51 is input to the teacher data acquisition unit 52. The teacher data acquisition unit 52 includes an input data acquisition unit 53 and a tag data acquisition unit 54.

The input data acquisition unit 53 acquires the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the angular velocity estimated value hω from the learning data 51 _rm The phase estimation value hθ is output to the learning unit 50 as teacher input data 55.

The tag data obtaining unit 54 obtains the allowable range Δpref from the learning data 51 as teacher tag data 56, and outputs the teacher tag data 56 to the learning unit 50.

The teacher data 57 is composed of teacher input data 55 and teacher label data 56, and the learning unit 50 performs machine learning based on the teacher data 57, which is a combination of the teacher input data 55 and the teacher label data 56.

The learning with teacher data of the machine learning in embodiment 5 is performed by a neural network configured by combining perceptrons. The teacher data 57 thus obtained is supplied to the neural network, and learning is repeated while changing the weights for the respective sensors so that the output of the neural network is the same as the teacher label data 56.

In the learning process, the weights are adjusted in such a manner that the output errors of the respective sensors are reduced by repeatedly executing the processing realized by the error back-propagation method. That is, learning with teacher data eliminates errors between the teacher tag data 56 and output data of the neural network while adjusting the weight value.

Thus, the characteristics of the teacher data 57 are learned, and reasoning is performed based on the input, so that a trained model for deriving the result is obtained.

Thus, the trained model generated by machine learning has the characteristics of the teacher data 57. For example, if the learning data 51 that becomes the teacher data 57 is used for model predictive control, the trained model calculates the allowable range Δpref as a value equivalent to that when the rotating electric machine 4 is controlled by the model predictive control, and therefore by using this allowable range Δpref, the on-off loss SWloss of the power conversion unit 1 can be reduced as compared with the PWM method.

The neural network used in the learning by the learning unit 50 may be three layers, may be more layers, or may be a neural network that performs machine learning by deep learning.

FIG. 21 is a hardware architecture diagram for generating a trained model. The machine learner 60 is implemented by a hardware structure shown in fig. 21 by performing machine learning for generating a trained model by the machine learner 60 functioning as a neural network.

The machine learner 60 is composed of a processor 61 and a storage 62.

The storage device 62 has, for example, a RAM 63 as a volatile storage device, and an HDD 64 as a nonvolatile auxiliary storage device. Further, as the nonvolatile auxiliary storage device, an SSD or a flash memory may be used instead of the HDD.

The HDD 64 stores the learning program 65 and the teacher data 66, and also stores the generated learning result 67.

Various learning programs 65 are input to the processor 61 from the HDD 64 via the RAM 63, and the processor 61 executes the various learning programs 65 input thereto. The learning program 65 causes the processor 61 to perform learning with teacher data. That is, the teacher data 66 is also input from the HDD 64 to the processor 61 via the RAM 63, and is learned in accordance with the learning program 65.

The processor 61 outputs the data of the learning result 67 to the RAM 63 of the storage device 62, and if necessary, stores the data in the HDD 64 via the RAM 63.

The learning program 65 is a program including a command for causing the processor 61 to execute learning with teacher data, and generates data of a result of machine learning (learning result 67).

The machine learner 60 described above may be implemented by PC (Personal Computer), a server device, or the like. However, since the amount of computation is large, GPU (Graphics Processing Units) may be mounted on a PC, for example, and the GPU may be used for computation processing for learning with teacher data by a technique called GPGPU (General-Purpose computing on Graphics Processing Units) to perform processing at high speed.

Next, a control operation in the power conversion device 100D according to embodiment 5 will be described with reference to fig. 22. Here, fig. 22 is a flowchart illustrating a control operation in the power conversion device 100D.

First, in the same flow as in embodiment 4, the voltage output calculation unit 11A calculates the multiphase voltage output value Vout based on the bus voltage Vdc detected by the bus voltage detection unit 15 and the on-off state quantity SWS output by the on-off determination table 14B (step S40, step S41).

Next, in the same flow as in embodiment 4, the speed estimation computing unit 21 calculates the angular speed ω of the rotating electric machine 4 based on the multiphase voltage command value Vref and the detected current value Iuvw detected by the current detecting unit 17 _rm And the estimated value of the phase θ (step S42, step S43).

In addition, in the same flow as in embodiment 4, the speed controller 19 and the current controller 20 calculate an angular velocity estimated value hω based on the angular velocity command value ωrmref and the velocity estimation calculation unit 21 _rm The phase estimated value hθ and the detection current value Iuvw calculate the multiphase voltage command value Vref (step S44).

In the same flow as in embodiment 4, the integrated value calculation unit 12 and the offset adjustment unit 16 execute processing to calculate the corrected voltage command integrated value Prefcomp and the corrected voltage output integrated value Poutcomp (step S45, step S46).

Trained model 26 is based on corrected voltage command integral value Prefcomp, corrected voltage output integral value Poutcomp, angular velocity estimate value hω _rm The phase estimation value hθ calculates the allowable range Δpref (step S47).

In the same flow as in embodiment 4, the on-off update determination unit 13B outputs the update signal Snew based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the allowable range Δpref calculated by the allowable range calculation unit 18, and the on-off determination table 14B determines the on-off state quantity SWS of the power conversion unit by the method shown in fig. 17 based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the update signal Snew (steps S48 and S49).

Then, the power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off determination table 14B, and outputs the ac power to the rotating electrical machine 4, thereby performing drive control of the rotating electrical machine 4 (steps S50 and S51).

The power conversion device 100D according to embodiment 5 includes: a power conversion unit 1 that converts direct-current power from a direct-current power supply 2 into alternating-current power and supplies the alternating-current power to a rotating electrical machine 4; a bus voltage detection unit 15 that detects a bus voltage Vdc of the power conversion unit 1; a voltage output calculation unit 11A that calculates a multiphase voltage output value Vout based on the busbar voltage Vdc detected by the busbar voltage detection unit 15 and the on-off state quantity SWS output by the on-off determination table 14B; a current detection unit 17 that detects a current flowing between the power conversion unit 1 and the rotating electrical machine 4; a speed estimation calculation unit 21 for estimating the angular speed ω of the rotating electric machine 4 based on the multiphase voltage command value Vref and the detected current value Iuvw _rm And phase θ; a speed controller 19 for estimating an angular speed h omega based on the angular speed command value ωrmref _rm Calculating a current command value Iref; a current controller 20 that calculates a multiphase voltage command value Vref based on the current command value Iref, the detection current value Iuvw, and the phase estimation value hθ; an integrated value calculation unit 12 that integrates the multiphase voltage command value Vref calculated by the current controller 20 and the multiphase voltage output value Vout calculated by the voltage output calculation unit 11A, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an offset adjustment unit 16 that adjusts the offset values of the voltage command integrated value Pref and the voltage output integrated value Pout to 0, and outputs the voltage command integrated value Prefcomp and the voltage output integrated value Poutcomp after correction; training wellBased on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, the angular velocity estimated value hω _rm Reasoning the phase estimation value hθ, and calculating the allowable range delta Pref; an on-off update determination unit 13B that determines whether to update the on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage command integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, and the allowable range Δpref calculated from the trained model 26; and an on-off decision table 14B for deciding an on-off state quantity SWS of the power conversion unit 1 based on the corrected voltage instruction integrated value Prefcomp, the corrected voltage output integrated value Poutcomp, the traveling direction of the vector of the corrected voltage instruction integrated value Prefcomp calculated from the corrected voltage instruction integrated value Prefcomp, and the update signal Snew,

Therefore, since power conversion device 100D according to embodiment 5 calculates based on voltage command integrated value Prefcomp, corrected voltage output integrated value Poutcomp, angular velocity estimated value, phase estimated value, and trained model 26, which are corrected for allowable range Δpref, allowable range Δpref can be changed so that performance such as learning data used for generating trained model 26 can be obtained as compared with embodiment 4. Therefore, the power conversion device 100D can obtain performance as in the manner of controlling the rotating electrical machine 4 used in creating the learning data.

Embodiment 6

Next, a power conversion device according to embodiment 6 will be described with reference to fig. 23. Here, fig. 23 is a block diagram showing the configuration of a power conversion device 100E according to embodiment 6.

As shown in fig. 23, the on-off determining unit 300 is provided, and the on-off determining unit 300 determines the on-off state quantity SWS of the power converting unit 1 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref, and the on-off determining unit 300 is configured by the on-off calculating unit 27 and the on-off output unit 28, and the on-off calculating unit 27 calculates the on-off state quantity SWS of the power converting unit 1 and the duration Tsw of the on-off state quantity SWS as the setting signal SetSW based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref, and the on-off output unit 28 outputs the on-off state quantity SWS of the power converting unit 1 based on the setting signal SetSW calculated by the on-off calculating unit 27.

The power conversion device according to embodiment 6 differs from the configuration of the power conversion devices according to embodiments 1 to 5 in that the duration Tsw of the on-off state quantity SWS is calculated. The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 to 5, and description thereof will be omitted, focusing on differences from embodiments 1 to 5.

First, the functions of the on-off calculating unit 27 and the on-off output unit 28, which are differences from embodiments 1 to 5, will be described below.

The on-off calculation unit 27 calculates the on-off state quantity SWS of the power conversion unit 1 and the duration Tsw of the on-off state quantity SWS as the setting signal SetSW based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref. Details of the on-off state quantity SWS and the calculation method of the duration Tsw of the on-off state quantity SWS will be described later.

The on-off output unit 28 determines the on-off state amount SWS of the power conversion unit 1 based on the setting signal SetSW calculated by the on-off calculation unit 27.

Fig. 24 is a diagram for explaining a method of calculating the on-off state quantity SWS and the duration Tsw of the on-off state quantity SWS in the on-off calculation section 27. Fig. 24 shows a case where the voltage command integrated value Pref and the voltage output integrated value Pout calculate the on-off state quantity SWS of the power conversion unit 1 and the duration Tsw of the on-off state quantity SWS in a uvw coordinate system which is a three-phase stationary coordinate system. The determination of the on-off state amount SWS is the same as that of embodiment 2.

In fig. 24, the three-phase voltage output integrated value P is set as the start time point ₃ out and allowable range delta of hexagon ₆ The time point at which the upper limit value of the u-phase of Pref crosses. First, three-phase voltage output integral value P ₃ out and allowable range delta of hexagon ₆ Pref crosses, thus arrive at a packet that is notIncluding and three-phase voltage command integral value P ₃ Allowable range delta of hexagon with crossed traveling directions of ref vector ₆ Since Pref has an upper limit value of u-phase of 3 sides including the adjacent side, a non-zero voltage vector is outputted as the on-off state quantity SWS1. For non-zero voltage vectors, the sum-and-three-phase voltage command integral value P is included due to the output orientation ₃ Since the non-zero voltage vector varies on either one of the two sides including the side where the traveling direction of the vector ref crosses is close, the on-off state amount SWS1 is selected to correspond to the on-off state index SW4 that varies in the direction of the lower limit value of the u-phase of the hexagonal allowable range.

Then, an integrated value P is outputted to the three-phase voltage when the on-off state index SW4 is outputted ₃ out and allowable range delta of hexagon ₆ The duration T1sw until the upper or lower limit of Pref crosses is calculated. Here, in the hexagonal allowable range Δ ₆ The time until reaching the upper limit value of w phase in the side of Pref is the duration T1sw.

Thereafter, the on-off state quantity SWS is also calculated in the same flow as in embodiment 2. In fig. 24, the duration Tsw of each on-off state quantity SWS is calculated, and the calculated on-off state quantity SWS and duration Tsw are output as the setting signal SetSW.

Therefore, in embodiment 6 as well, the duration Tsw of each on-off state quantity SWS is calculated as compared with the on-off determination method of embodiment 2, so that it is not necessary to determine whether or not the on-off state quantity SWS needs to be updated successively as compared with embodiment 2, and the calculation period can be set long, so that it is an easy-to-install method in an inexpensive microcomputer.

Next, a control operation in the power conversion device 100E according to embodiment 6 will be described in detail with reference to fig. 25. Fig. 25 is a flowchart illustrating a control operation in the power conversion device 100E.

First, the processing is performed in the same flow as in embodiment 2, and the voltage output calculation unit 11 calculates the multiphase voltage output value Vout, and the integrated value calculation unit 12 integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates the voltage command integrated value Pref and the voltage output integrated value Pout (step S52 and step S53).

Next, the on-off calculation unit 27 calculates the on-off state quantity SWS and the duration Tsw of the on-off state quantity SWS as shown in fig. 24 based on the voltage command integrated value Pref, the voltage output integrated value Pout, and the allowable range Δpref (step S54).

Then, the on-off output section 28 sequentially outputs the on-off state quantity SWS from SWS1 to SWS3 based on the on-off state quantity SWS calculated by the on-off calculating section 27 and the duration Tsw (step S55).

The power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off output unit 28, outputs the ac power to the load 3, and controls the driving of the load 3 (step S56).

The power conversion device 100E according to embodiment 6 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a voltage output calculation unit 11 that calculates a multiphase voltage output value Vout based on the on-off state quantity SWS output by the on-off output unit 28; an integrated value calculation unit 12 that integrates the multiphase voltage command value Vref and the multiphase voltage output value Vout, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an on-off calculation unit 27 that calculates an on-off state quantity SWS and a duration Tsw of the power conversion unit 1 as a setting signal SetSW based on the voltage command integrated value Pref, the voltage output integrated value Pout, and an allowable range Δpref set for the voltage command integrated value Pref; and an on-off output unit 28 that outputs the on-off state quantity SWS of the power conversion unit 1 based on the setting signal SetSW calculated by the on-off calculation unit 27.

Therefore, the power conversion device 100E according to embodiment 6 is a system that is easy to install in an inexpensive microcomputer because it is possible to make a long calculation cycle without determining whether or not the on-off state amount SWS needs to be updated successively, as compared with embodiments 2 to 5.

Embodiment 7

Next, a power conversion device according to embodiment 7 will be described with reference to fig. 26. Here, fig. 26 is a block diagram showing the structure of a power conversion device 100F according to embodiment 7.

As shown in fig. 26, a power conversion device 100F according to embodiment 7 differs from embodiment 6 in that a current detection unit 17 is provided between a power conversion unit 1 and a load 3, and further includes: a harmonic processing unit 29 that calculates a harmonic current Ithd based on the detection current value Iuvw detected by the current detection unit 17; a harmonic current controller 30 that calculates an allowable range Δpref based on the harmonic current command value Ithdref and the harmonic current Ithd; a low-frequency extraction unit 31 that calculates a current low-frequency value Ifund based on the detected current value Iuvw; and a current controller 20 that calculates a multiphase voltage command value Vref based on the current command value Iref and the current low frequency value ifind.

Next, the functions of the current detection unit 17, the harmonic processing unit 29, the harmonic current controller 30, the low frequency extraction unit 31, and the current controller 20 according to embodiment 7, which are differences from embodiment 6, will be described.

The current detection unit 17 detects a current flowing between the power conversion unit 1 and the load 3.

The harmonic processing unit 29 calculates the harmonic current Ithd based on the 2 or more detected current values Iuvw. Here, the harmonic current Ithd is calculated, for example, from data obtained by digitizing harmonic components included in the current or from current spectrum data expressed on a frequency axis. The harmonic current controller 30 calculates the allowable range Δpref based on the harmonic current command value Ithdref and the harmonic current Ithd. Details of the method of calculating the harmonic current Ithd and the allowable range Δpref will be described later.

The low-frequency extraction unit 31 calculates a current low-frequency value ifind based on 2 or more detection current values Iuvw. Here, the current low-frequency value Ifund is a fundamental wave of the current. The current controller 20 calculates the multiphase voltage command value Vref based on the current command value Iref and the current low frequency value ifind. Details of the method of calculating the low-frequency current value Ifund will be described later.

Fig. 27 shows a case where the harmonic current Ithd is calculated by the harmonic processing section 29 based on a current waveform of 1 cycle amount from a current drawn with 2 or more detection current values Iuvw. If a fast Fourier transform (FFT: fast Fourier Transform) is performed on the current waveform of the time axis of FIG. 27, the current spectrum of the frequency axis is calculated. The current spectrum of the frequency axis is calculated from the fundamental and harmonics of the current waveform. The harmonic current Ithd corresponds to the harmonic of fig. 27.

The harmonic current command value Ithdref changes whether the harmonic of fig. 27 is a component of each order, a total value of each order of the harmonic, or a ratio of the harmonic to the fundamental wave. For example, in the case where the harmonic current command value Ithdref is supplied as the ratio of the harmonic wave to the fundamental wave, the harmonic current Ithd is also calculated from the ratio of the harmonic wave to the fundamental wave of the current spectrum of the frequency axis.

The fundamental wave of the frequency axis obtained in fig. 27 may be used as the current low-frequency value Ifund. The allowable range Δpref is adjusted so as to eliminate the difference between the harmonic current command value Ithdref and the harmonic current Ithd.

Fig. 28 shows a case where the current low-frequency value Ifund is calculated by the low-frequency extracting section 31 by current oversampling based on the above 2 detected current values Iuvw. Here, the current oversampling is to detect a current at a period shorter than a normal operation period.

As can be seen from fig. 28, in the case where no current is over-sampled, it is difficult to calculate the current fundamental wave as the current low-frequency value Ifund from the detected current value Iuvw. In contrast, in the case of the current-carrying sampling of fig. 28, the current is detected at a short period, and the current low-frequency value Ifund is calculated by the average processing at regular intervals, whereby a value close to the current fundamental wave is obtained from the detected current value Iuvw as the current low-frequency wave.

Next, a control operation in the power conversion device 100F according to embodiment 7 will be described in detail with reference to fig. 29. Fig. 29 is a flowchart illustrating a control operation in the power conversion device 100F.

First, the current flowing through the load 3 is measured as a detection current value Iuvw by the current detection unit 17 (step S57). At this time, the detection current value Iuvw is measured a predetermined number of times (step S58).

Next, based on 2 or more detected current values Iuvw, the harmonic processing section 29 and the low-frequency extracting section 31 calculate a harmonic current Ithd and a current low-frequency value Ifund, respectively (step S59, refer to fig. 27 and 28).

Then, the allowable range Δpref is calculated based on the harmonic current command value Ithdref and the harmonic current Ithd (step S60), and the multiphase voltage command value Vref is calculated based on the current command value Iref and the current low frequency value ifind (step S61).

Thereafter, the processing is performed in the same flow as in embodiment 6, and the power conversion unit 1 converts the dc power of the dc power supply 2 into ac power based on the on-off state amount SWS determined by the on-off output unit 28, and outputs the ac power to the load 3, thereby performing drive control of the load 3 (steps S52 to S56).

The power conversion device 100F according to embodiment 7 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a harmonic processing unit 29 that has a current detection unit 17 between the power conversion unit 1 and the load 3, and calculates a harmonic current Ithd based on a detection current value Iuvw detected by the current detection unit 17; a harmonic current controller 30 that calculates an allowable range Δpref based on the harmonic current command value Ithdref and the harmonic current Ithd; a low-frequency extraction unit 31 that calculates a current low-frequency value Ifund based on the detected current value Iuvw; a current controller 20 that calculates a multiphase voltage command value Vref based on the current command value Iref and the current low frequency value ifind; a voltage output calculation unit 11 that calculates a multiphase voltage output value Vout based on the on-off state quantity SWS output by the on-off output unit 28; an integrated value calculation unit 12 that integrates a multiphase voltage command value Vref (hereinafter, also simply referred to as a voltage command value Vref) and a multiphase voltage output value Vout, respectively, and calculates a voltage command integrated value Pref and a voltage output integrated value Pout; an on-off calculation unit 27 that calculates an on-off state quantity SWS and a duration Tsw of the power conversion unit 1 as setting signals based on the voltage command integrated value Pref, the voltage output integrated value Pout, and an allowable range Δpref set for the voltage command integrated value Pref; and an on-off output unit 28 that outputs the on-off state quantity SWS of the power conversion unit 1 based on the setting signal SetSW calculated by the on-off calculation unit 27.

In the power conversion device 100F according to embodiment 7, compared with embodiments 1 to 6, since the harmonic current Ithd of the load 3 can be controlled according to the command value, the desired harmonic current Ithd can be obtained under various operating conditions.

Embodiment 8

Next, a power conversion device 100G according to embodiment 8 will be described with reference to the drawings.

Fig. 30 is a block diagram showing the structure of a power conversion device according to embodiment 8.

As shown in fig. 30, a power conversion device 100G according to embodiment 8 includes: a power conversion unit 1 which is a main circuit; a bus voltage detection unit 15 that detects a bus voltage Vdc of the power conversion unit 1; and a control device 10E that controls the output of the power conversion unit 1, wherein the power conversion device 100G is connected between the dc power supply 2 and the load 3.

The power conversion unit 1 converts dc power from the dc power supply 2 into ac power, supplies the ac power to the load 3, and drives the load 3. The load 3 is driven by ac power supplied from the power conversion unit 1. For example, various motors such as a transformer, a reactor, an induction motor, and a synchronous motor can be used as the load 3.

The control device 10E includes: a voltage prediction unit 32 that calculates a voltage prediction value Vpred based on the bus voltage Vdc detected by the bus voltage detection unit 15 and all the candidate on-off state amounts SWSall in the power conversion unit 1; an on-off predicting unit 33 that predicts, based on the voltage command value Vref and all the candidate voltage prediction values vpreadall calculated by the voltage predicting unit 32, on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and a duration Tsw (hereinafter, simply referred to as duration Tsw) of the on-off state amounts that the on-off state amounts are continued; and an on-off output unit 28 that outputs a signal "determination SWS" for determining the on-off state amounts of the plurality of switching elements of the power conversion unit 1 based on the on-off state amounts SWS and the duration Tsw calculated by the on-off prediction unit 33. The on/off state amount SWS of the power conversion unit 1 is determined by the on/off prediction unit 33 and the on/off output unit 28, and thus corresponds to the on/off determination unit 300 described above.

The voltage predicting unit 32 calculates the multiphase voltage output value Vout at all candidate on-off state amounts SWSall that can be obtained by the plurality of switching elements of the power converting unit 1 as the voltage predicted value Vpred. Details of all the candidate on-off state amounts SWSall of the power conversion unit 1 will be described later.

The on-off predicting unit 33 integrates the voltage command value Vref and the voltage predicted value Vpred calculated by the voltage predicting unit 32, respectively, and expands the voltage command integrated value Pref and the integrated value Ppred of the voltage predicted value, which are the values obtained by integrating the respective values, to a desired section. At this time, the initial value C of the integrated value Ppred of the voltage predicted value is calculated from the on-off state quantity SWS and the duration Tsw calculated by the on-off predicting unit 33 and the bus voltage Vdc detected by the bus voltage detecting unit 15, and the multiphase voltage output value Vout is integrated as the voltage output integrated value Pout. The evaluation value J is calculated from the predicted value extending to the desired section, and the on-off state amounts SWS of the plurality of switching elements of the power conversion unit 1 and the duration Tsw thereof are calculated from the calculated evaluation value J. The details of the prediction value expansion method and the evaluation value will be described later.

Fig. 31 is a hardware configuration diagram for realizing power conversion device 100G.

The power conversion unit 1 is configured by a three-phase inverter circuit that converts direct-current power of the direct-current power supply 2 into three-phase alternating-current power, and the power conversion unit 1 drives the load 3. The power conversion unit 1 includes a plurality of switching elements Q1 to Q6 each of which is connected in antiparallel with a diode D. The power conversion unit 1 is connected to input terminals of the respective phases of the load 3 via bus bars from connection points of the upper arm and the lower arm of the respective phases. In this case, the u-phase has switching elements Q1 and Q2, the v-phase has switching elements Q3 and Q4, and the w-phase has switching elements Q5 and Q6.

The bus voltage detection unit 15 is a mechanism for detecting the bus voltage Vdc by measuring the voltage difference between the positive side (+) and the negative side (-) of the dc power supply 2.

The control device 10E has hardware including a processor 40 and a storage device 41.

The storage device 41 includes volatile storage devices (not shown) such as RAM (Random Access Memory) and nonvolatile auxiliary storage devices (not shown) such as HDD (Hard Disk Drive) and SSD (Solid State Drive). Further, as the nonvolatile auxiliary storage device, a flash memory may be used instead of the HDD.

The processor 40 executes a control program input from the storage device 41.

The storage device 41 has a secondary storage device and a volatile storage device. The control program 42 is input to the processor 40 from the auxiliary storage via the volatile storage.

The processor 40 outputs the processing data 43 such as the operation result to the volatile memory device of the storage device 41, and stores the processing data 43 in the auxiliary storage device via the volatile memory device as necessary.

As described above, the control device 10E outputs the on-off state amounts SWS of the plurality of switching elements Q1 to Q6 of the power conversion unit 1 to control the power conversion unit 1.

Fig. 32 is a diagram showing an example of the case where all candidate on-off state amounts SWSall of the power conversion unit 1 are 2-level. The on-off state quantity SWS is a combination of signals of on (: 1) and off (: 0) of the respective switching elements Q1 to Q6. The switching elements Q1 to Q6 of the upper arm and the lower arm have 9 on-off state indexes of 8 on-off state indexes SWn (n is an integer of 0 to 7) in which one is turned on and the other is turned off, and on-off state indexes (on-off state indexes SW 8) in which all the switching elements (all the switching elements Q1 to Q6) are turned off when the operation of the power conversion section 1 is stopped.

Fig. 33 is a diagram illustrating the on-off pattern SWP of the power conversion unit 1. Fig. 33 illustrates, as an example of the on-off pattern SWP, an on-off pattern SWP in which data sets of the on-off state amounts SWS and the durations Tsw of the 3 power conversion units 1 are combined.

In the present invention, the on-off pattern SWP is set by combining data sets of the on-off state amounts SWS and the durations Tsw of the two power conversion units 1. In the case of fig. 33, the on-off pattern SWP is a combination set such that the 1 st on-off state amount SWS1 (here, corresponding to the on-off state index SW1 of fig. 32) is continued for a time period T1SW, the 2 nd on-off state amount SWS2 (here, corresponding to the on-off state index SW2 of fig. 32) is continued for a time period T2SW, and the 3 rd on-off state amount SWS3 (here, corresponding to the on-off state index SW7 of fig. 32) is continued for a time period T3SW.

Fig. 34 is a diagram illustrating all candidate voltage predicted values vpreadall of the voltage output integrated value Pout from the on-off predicting unit 33. The integrated value pprendll of all candidate voltage predictors vpreadall (hereinafter, for simplicity, the integrated value pprendll of all candidate voltage predictors) is plotted based on the on-off state index SWn (n is an integer of 0 to 8) of fig. 32 from the point where the integrated value Pout of the voltage output calculated by the on-off predicting unit 33 is set as the initial value C.

In fig. 34, the on-off state quantity SWs corresponding to the on-off state indexes SW0, SW7, and SW8 is zero in the multiphase voltage output value Vout (hereinafter, also simply referred to as the voltage output value Vout), and therefore the integrated value Ppred of the voltage predicted value matches the voltage output integrated value Pout. Therefore, the integrated value Ppred of the total 6 voltage predicted values indicated by the dotted lines (corresponding to the on-off state indexes SW1 to SW 6) is depicted.

Fig. 35 is a diagram illustrating a method of expanding a prediction section of the voltage command integrated value Pref and the integrated value Ppred of the voltage predicted value in the on-off prediction unit 33. Fig. 35 shows, as an example, a case where the voltage command integrated value Pref and the integrated value Ppred of the voltage predicted value are extended to a 60-degree phase amount of the voltage command value Vref from the prediction start time point in the αβ coordinate which is the 2-phase stationary coordinate system. In fig. 35, the voltage command integrated value Pref is indicated by solid arrows, and the integrated value Ppred of the voltage predicted value is indicated by 5 broken arrows (on-off state amounts SWS1 to SWS 5).

Since the voltage command value Vref changes according to the frequency of the voltage command value Vref if the voltage command value Vref is assumed to be in a steady state on a stationary coordinate, for example, if the frequency is positive, the voltage command value Vref is rotated circumferentially counterclockwise. Therefore, the voltage command integrated value Pref also performs a circumferential rotation in the counterclockwise direction. That is, in fig. 35, the amount of movement is 60 degrees from the prediction start time point to the prediction end time point.

In contrast, since the voltage predicted value Vpred is the voltage output value Vout at the on-off state quantity SWS of the power conversion unit 1, only the value 2 is set as the on-off state of the switching element. However, since the trajectory up to the voltage predicted value Vpred can be expressed if the integrated value Ppred of the voltage predicted value is the integrated value Ppred of the voltage predicted value, the integrated value Ppred of the voltage predicted value can calculate the trajectory of the 60-degree phase amount of the voltage command value as shown in fig. 35.

In the case of fig. 35, the trajectory of the voltage command value Vref of an amount of 60 degrees is predicted in the order of the on-off state amounts SWS from the on-off state amounts SWS1 to SWS 5. That is, in the prediction calculation, the length of the track is changed only by changing the time for integrating the voltage prediction value Vpred at each on-off state quantity SWS. For this trajectory, when the stator magnetic flux Φ is intended to be calculated by model predictive control, which is a conventional method, for example, it is necessary to calculate the state equation of the load 3 successively. That is, the method of the present invention significantly reduces the amount of calculation compared to the conventional method.

The total value Perrsum of the integral deviation Perr (hereinafter, simply referred to as the total value Perrsum of the integral deviation), which is the difference between the voltage command integral Pref of the 60-degree phase amount of the voltage command value calculated as shown in fig. 35 and the integral value Ppred of the voltage prediction value, may be calculated, and the total value Perrsum of the integral deviation may be used as the evaluation value J to determine the on-off state amount SWS of the power conversion unit 1, or the on-off state amount SWS of the power conversion unit 1 may be determined based on the total value SWcountsum of the on-off switching times SWcount of the on-off state amount SWS during switching of the integral deviation Perrsum and the on-off state amount SWS (hereinafter, simply referred to as the total value SWcountsum of the on-off switching times) by swcountsum×perrsum. Note that fig. 35 does not explicitly show the total value Perrsum of the integral deviation.

Next, a control operation in the power conversion device 100G according to embodiment 8 will be described with reference to the drawings.

Fig. 36 is a flowchart illustrating a control operation in the power conversion device 100G.

First, the bus voltage detection unit 15 detects the bus voltage Vdc of the dc power source 2 (step S62).

Next, the voltage predicting unit 32 calculates the voltage predicted value vpreadall candidates based on the bus voltage Vdc acquired in step S62 and the on-off state amounts swsal of all candidates of the power converting unit 1 (step S63).

Next, the on-off predicting unit 33 obtains the voltage command value Vref (step S64).

Then, the on-off predicting unit 33 expands the predicted value of the 60-degree phase amount across the voltage command value as shown in fig. 35 based on the integrated value Ppredall and the voltage command integrated value Pref of the voltage predicted value of all candidates obtained by integrating the voltage predicted value vpendall of the candidates obtained in step S63 and the voltage command value Vref obtained in step S64, respectively (step S65).

Next, based on the predicted values expanded in step S65, the on-off predicting unit 33 calculates the evaluation values J for all the candidate on-off state amounts SWSall (step S66). Here, the evaluation value J is, for example, a total value Perrsum of an integral deviation of the voltage command integrated value Pref and the integrated value Ppred of the voltage predicted value after the 60-degree phase of the voltage command value is extended.

Next, based on the evaluation values J at all the candidate on-off state amounts swsal calculated in step S66, the on-off predicting section 33 searches for the on-off state amount SWS that minimizes the evaluation value J (step S67). For example, when the evaluation value J is the total value Perrsum of the integral deviation, the on-off state amount SWS in which the total value Perrsum of the integral deviation is minimum is selected.

Next, based on the on-off state quantity SWS selected by the search in step S67 and the duration Tsw of the on-off state quantity, the on-off output unit 28 determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 (step S68).

Finally, the power conversion unit 1 controls the on/off state quantity SWS and controls the load 3 in accordance with the on/off state quantity SWS determined in step S68 and the duration Tsw of the on/off state quantity (step S69).

The power conversion device 100G according to embodiment 8 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a bus voltage detection unit 15 for detecting a bus voltage Vdc of the dc power supply 2; a voltage prediction unit 32 that calculates a voltage prediction value Vpred based on the bus voltage Vdc and all the candidate on-off state amounts SWSall in the power conversion unit 1; an on-off predicting unit 33 that calculates on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and a duration Tsw for which the on-off state amounts are continued, based on the voltage command value Vref and all the candidate voltage predicted values vpreadall calculated by the voltage predicting unit 32; and an on-off output unit 28 for determining the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 based on the on-off state quantity SWS calculated by the on-off prediction unit 33 and the duration Tsw, the on-off prediction unit 33 outputting the on-off state quantity SWS and the duration Tsw for minimizing the integrated deviation based on the integrated value Perrsum of the integrated deviation calculated from the integrated value pprendall of the candidate voltage prediction values and the voltage command integrated value Pref of the 60-degree phase quantity across the voltage command value, and the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 being determined by the on-off output unit 28 accordingly.

Therefore, in the power conversion device 100G according to embodiment 8, the voltage output integrated value Pout of the power conversion unit 1 is made to follow the voltage command integrated value Pref and the total value Perrsum of the integrated deviations of the 60-degree phase amounts across the voltage command value is made to be minimum, as compared with embodiment 7, so that the load 3 can be driven so as to reduce the harmonic voltage Vthd and the harmonic current Ithd.

Here, the evaluation value J is described as the total value Perrsum of the integral deviation, but the evaluation value J can also take into consideration the total value SWcountsum of the on/off switching times of the on/off state quantity SWS, and by setting the evaluation value J as swcountsum×perrsum of the 60-degree phase quantity across the voltage command value, the load 3 can be driven so as to reduce the on/off switching times of the power conversion unit 1, that is, so as to reduce the on/off loss SWloss.

In addition, the 60-degree phase is set here, but even if the phase is not 60 degrees, but is a shorter section or a longer section, the on-off state quantity SWS and the duration Tsw of the power conversion unit 1 can be calculated in the same flow.

Embodiment 9

Next, a power conversion device 100H according to embodiment 9 will be described with reference to the drawings.

Fig. 37 is a block diagram showing a configuration of a power conversion device 100H according to embodiment 9.

As shown in fig. 37, in the power conversion device 100H according to embodiment 9, a control device 10F of the power conversion device 100H includes an on-off prediction unit 33A instead of the on-off prediction unit 33 included in the control device 10E according to embodiment 8. The on/off state amount SWS of the power conversion unit 1 is determined by the on/off prediction unit 33A and the on/off output unit 28, and thus corresponds to the on/off determination unit 300 described above.

The on-off predicting unit 33A of the power conversion device 100H according to embodiment 9 is different from the on-off predicting unit 33 of the power conversion device 100G according to embodiment 8 in that the on-off state quantity SWS and the duration Tsw of the power conversion unit 1 are calculated using the allowable range Δpref with respect to the voltage command integrated value Pref. The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 to 8, and description thereof will be omitted, focusing on differences from embodiments 1 to 8.

The on-off predicting unit 33A calculates the on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and the duration Tsw for which the on-off state amounts are continued, based on the voltage allowable value pdelt that sets the allowable range Δpref to the voltage command integrated value Pref obtained by integrating the voltage command value Vref and the integrated value Ppred of the voltage predicted value obtained by integrating the voltage predicted value Vpred calculated by the voltage predicting unit 32. At this time, as in embodiment 8, the initial value C of the integrated value Ppred of the voltage predicted value is calculated from the on-off state quantity SWS and the duration Tsw calculated by the on-off predicting unit 33A and the bus voltage Vdc detected by the bus voltage detecting unit 15, and the voltage output integrated value Pout is set to the initial value C. Details of the expansion methods and evaluation values of Pdelta and Ppred will be described later.

The allowable range Δpref set for the voltage command integrated value Pref obtained by integrating the voltage command value Vref in the on-off predicting section 33A indicates a range in which the voltage output integrated value Pout is allowed with respect to the voltage command integrated value Pref. Therefore, the magnitude of the allowable range Δpref becomes a value that determines a trade-off relationship between the harmonic voltage Vthd and the on-off switching number SWcount, that is, if the value is set large, the harmonic voltage Vthd becomes large, but the on-off switching number SWcount of the plurality of switching elements in the power conversion unit 1 becomes small, whereas if the allowable range Δpref is set small, the harmonic voltage Vthd becomes small, but the on-off switching number SWcount of the plurality of switching elements in the power conversion unit 1 becomes large.

Fig. 38 is a diagram illustrating an example of a method of expanding the voltage allowable value Pdelta and the integrated value Ppred of the voltage predicted value, which are set to the allowable range Δpref, with respect to the voltage command integrated value Pref obtained by integrating the voltage command value Vref in the on-off predicting unit 33A.

Fig. 38 shows, as an example, a case where the voltage command integrated value Pref and the integrated value Ppred of the voltage predicted value are extended to a phase of 60 degrees or more of the voltage command value from the prediction start time point in the αβ coordinate which is the 2-phase stationary coordinate system. In fig. 38, the voltage command integrated value Pref is indicated by solid arrows, and the integrated value Ppred of the voltage predicted value is indicated by 4 broken arrows (on-off state amounts SWS1 to SWS 4).

Since the voltage command value Vref changes according to the frequency of the voltage command value Vref if the voltage command value Vref is assumed to be in a steady state on a stationary coordinate, for example, if the frequency is positive, the voltage command value Vref is rotated circumferentially counterclockwise. Therefore, the voltage command integrated value Pref also performs a circumferential rotation in the counterclockwise direction.

Since the voltage predicted value Vpred is the voltage output value Vout formed by the on-off state quantity SWS of the power conversion unit 1 with respect to the voltage command value Vref, the locus of Ppred is plotted by the on-off state quantity SWS as shown in fig. 34.

The switching elements of each phase can take on the on-off state indexes SW0 to SW8, and the on-off state indexes SW0 to SW8 are determined by on-off parameters (values indicating the level of the on-off state) indicated by 1 or 0 corresponding to on/off corresponding to the above 2 levels of Q1 to Q6 shown in fig. 32, so that the total of 9 voltage output values Vout are obtained. However, the on-off state indexes SW0, SW7, and SW8 have zero resultant vectors in the on-off state amounts SWs of the upper arms of the respective phases, and therefore, the non-zero voltage output values Vout are 6 kinds, and the zero voltage output values Vout are 3 kinds. Here, the integrated value Ppred of the voltage predicted values obtained by integrating the total of 9 voltage output values Vout is calculated.

Fig. 38 depicts a voltage allowable value Pdelta in which an allowable range Δpref is set for the voltage command integrated value Pref on the αβ coordinate. That is, in order to set the allowable range Δpref for the voltage command integrated value Pref shown on the αβ coordinate, an upper limit value Pupper of the voltage allowable value and a lower limit value Plower of the voltage allowable value are plotted. The upper limit value Pupper of the voltage allowable value and the lower limit value Plower of the voltage allowable value are circumferentially rotated with time in a steady state, similarly to the voltage command integrated value Pref.

The integrated value Ppred of the voltage predicted value linearly changes with time according to the on-off state quantity SWS as shown in fig. 34. Therefore, the voltage allowable value Pdelta and the integrated value Ppred of the voltage predicted value, in which the allowable range Δpref is set for the voltage command integrated value Pref, are expressed by a time function, and the cross arrival time Tcross until the upper limit value Pupper of the voltage allowable value or the lower limit value Plower of the voltage allowable value crosses the integrated value Ppred of the voltage predicted value is calculated.

In fig. 38, the integral value Ppred of the voltage predicted value crosses the upper limit value Pupper of the voltage allowable value when the duration T1SW of the on-off state quantity SWS1 (corresponding to the on-off state index SW 3), the integral value Ppred of the voltage predicted value crosses the lower limit value Plower of the voltage allowable value when the duration T2SW of the on-off state quantity SWS2 (corresponding to the on-off state index SW 4), the integral value Ppred of the voltage predicted value crosses the upper limit value Pupper of the voltage allowable value when the duration T3SW of the on-off state quantity SWS3 (corresponding to the on-off state index SW 3), and the integral value Ppred of the voltage predicted value crosses the upper limit value Pupper of the voltage allowable value when the duration T4SW of the on-off state quantity SWS4 (corresponding to the on-off state index SW 4).

In embodiment 9, when the predicted value is extended by using the allowable range Δpref, the calculation amount is further reduced compared to embodiment 8 because only the time function can be solved in comparison with the method for extending the predicted value of the unused allowable range Δpref of embodiment 8. That is, as described above, since the integrated value Ppred of the voltage predicted value can be linearly changed with time with respect to the change of the on-off state quantity SWS, the calculation amount can be further reduced.

Here, the total value Tcrosssum of the cross arrival time tcoross until the intersection of Pdelta and the integrated value Ppred of the voltage predicted value is calculated based on the predicted value calculated as shown in fig. 38, and the reciprocal 1/Tcrosssum of the total value Tcrosssum of the time until the intersection is used as the evaluation value J to determine the on-off state quantity SWS of the power conversion unit 1, or the evaluation value J may be calculated based on the total value Tcrosssum of the time until the intersection and the total value SWcountsum of the on-off switching times based on SWcountsum/Tcrosssum to determine the on-off state quantity SWS of the power conversion unit 1.

Next, a control operation in the power conversion device 100H according to embodiment 9 will be described with reference to the drawings.

Fig. 39 is a flowchart illustrating a control operation in the power conversion device 100H.

First, the on-off predicting unit 33A obtains the voltage command value Vref and all the candidate voltage predicted values vpreadall calculated by the voltage predicting unit 32 (steps S62 to S64).

Next, the on-off predicting unit 33A obtains the allowable range Δpref for setting the voltage command integrated value Pref (step S9).

Then, the on-off predicting unit 33A calculates a total value tcosssum of time until the voltage allowable value Pdelta in which the allowable range Δpref is set crosses the integrated value Ppred of the voltage predicted value based on the voltage allowable value Pdelta in which the allowable range Δpref is set and the integrated values of all the candidate voltage predicted values for the voltage command integrated value Pref, based on the predicted values of the phases of 60 degrees or more extended to the voltage command integrated value (step S71).

The on-off predicting unit 33A calculates the evaluation value J among all the candidate on-off state amounts based on the reciprocal 1/Tcrosssum of the total value Tcrosssum of the time until the intersection calculated in step S71 (step S72).

The on-off predicting unit 33A searches for the on-off state quantity SWS of the power converting unit 1 and the duration Tsw for which the on-off state quantity is continued in the same flow as in embodiment 8 based on the evaluation value J calculated in step S72, and the power converting unit 1 controls the on-off state quantity SWS based on the searched on-off state quantity SWS and duration Tsw to control the load 3 (steps S73 to S75).

The power conversion device 100H according to embodiment 9 includes: a power conversion unit 1 that converts dc power from a dc power supply 2 into ac power and supplies the ac power to a load 3; a bus voltage detection unit 15 for detecting a bus voltage Vdc of the dc power supply 2; a voltage prediction unit 32 that calculates a voltage prediction value Vpred based on the bus voltage Vdc and all the candidate on-off state amounts SWSall in the power conversion unit 1; an on-off predicting unit 33A that calculates on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and a duration Tsw during which the on-off state amounts are continued, based on a voltage allowable value Ppredall in which an allowable range Δpref is set for the voltage command integrated value Pref and the integrated values Ppredall of all the candidate voltage predicted values calculated by the voltage predicting unit 32; and an on-off output unit 28 for determining the on-off state quantity SWS and the duration Tsw of the plurality of switching elements of the power conversion unit 1 based on the on-off state quantity SWS and the duration Tsw calculated by the on-off prediction unit 33A, the on-off prediction unit 33A calculating a total value tcoussum of a time until the voltage allowable value Pdelta of the allowable range Δpref and the integrated value Ppredall of all the candidate voltage predicted values intersect with each other for a voltage command integrated value Pref of the voltage command value equal to or greater than 60 degrees, and outputting the on-off state quantity SWS and the duration Tsw so that the inverse 1/tcousssum of the time until the intersection is minimized, and the on-off output unit 28 determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 accordingly.

Therefore, since the power conversion device 100H according to embodiment 9 uses the time function to express the voltage allowable value Pdelta and the integrated value Ppred of the voltage predicted value, in which the allowable range Δpref is set for the voltage command integrated value Pref, and expands the prediction interval to the 60 degree phase or more of the voltage command value, the amount of calculation for the prediction operation can be reduced as compared with embodiment 1, and the load 3 can be driven with the integrated deviation Perr of the voltage command value Vref and the voltage output value Vout being restricted to be constant by the allowable range Δpref. Therefore, the power conversion device 100H according to embodiment 9 can easily set the prediction section longer than that of embodiment 8, and thus can improve the suppression effect of the voltage harmonics.

Here, the evaluation value J is described as the reciprocal 1/tcrossu of the total time until the intersection, but the load 3 can be driven so that the on-off loss SWloss of the power conversion unit 1 is also reduced by taking the total value SWcountsum of the on-off switching times of the on-off state quantity SWS into consideration and setting the evaluation value J to SWcountsum/tcrossum of the phase quantity of 60 degrees or more across the voltage command value. Here, the 60-degree phase is assumed, but even if the phase is not 60 degrees, but a shorter section or a longer section, the on-off state quantity SWS and the duration Tsw of the power conversion unit 1 can be calculated in the same flow.

Embodiment 10

Next, a power conversion device 100I according to embodiment 10 will be described with reference to the drawings.

Fig. 40 is a block diagram showing the structure of a power conversion device 100I according to embodiment 10. As shown in fig. 40, in the power conversion device 100I according to embodiment 10, the load 3 is replaced with the rotating electrical machine 4, the current detection unit 17 is provided between the power conversion unit 1 and the rotating electrical machine 4, and the control device 10G of the power conversion device 100I includes the on-off prediction unit 33B instead of the on-off prediction unit 33A provided in the control device 10G. The control device 10G further includes a state observation unit 34, wherein the state observation unit 34 calculates a driving state quantity Mstate representing a driving state of the rotating electric machine 4 based on the detected current value Iuvw detected by the current detection unit 17, and further includes an allowable range calculation unit 18, wherein the allowable range calculation unit 18 calculates an allowable range Δpref with respect to the voltage command integrated value Pref based on the driving state quantity Mstate calculated by the state observation unit 34. The on/off state amount SWS of the power conversion unit 1 is determined by the on/off prediction unit 33B and the on/off output unit 28, and thus corresponds to the on/off determination unit 300 described above.

The on-off predicting unit 33B of the power conversion device 100I is different from the on-off predicting unit 33A of the power conversion device 100H according to embodiment 9 in that the allowable range Δpref with respect to the voltage command integrated value Pref is changed in accordance with the driving state of the rotating electrical machine 4. The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 to 9, and description thereof will be omitted, focusing on differences from embodiments 1 to 9.

As described above, the power conversion device 100I is different from embodiments 8 and 9 in having the current detection unit 17.

As shown in fig. 40, the power conversion device 100I includes a voltage predicting unit 32, an on/off predicting unit 33B, an on/off outputting unit 28, a state observing unit 34, and an allowable range calculating unit 18, which are components of the power conversion unit 1, the bus voltage detecting unit 15, the current detecting unit 17, and the control device 10G. Next, the functions of the current detecting unit 17, the state observing unit 34, and the allowable range calculating unit 18, which are different from those of embodiment 9, will be described.

The current detection unit 17 detects a current flowing between the power conversion unit 1 and the rotating electrical machine 4, and outputs the detected current to the state observation unit 34.

The state observation unit 34 calculates a driving state quantity Mstate representing the driving state of the rotating electric machine 4 based on the voltage command value Vref and the detected current value Iuvw detected by the current detection unit 17, and outputs the calculated driving state quantity Mstate. Here, the drive state quantity Mstate includes, for example, at least any one of a voltage command value Vref and a detection current value Iuvw detected by the current detection unit 17, which are converted to an αβ coordinate which is a stationary coordinate of the two phases, a magnetic flux Φαβ of the two phases calculated based on the voltage command value Vref and the current αβ of the two phases, a magnetic flux Φ of the rotating electric machine 4 calculated based on the magnetic flux Φαβ of the two phases, or a first-order angular frequency ω ₁ Angular velocity omega _rm Torque τ calculated based on the two-phase magnetic flux Φαβ and the two-phase current iαβ, and loss Mloss of the rotating electrical machine 4 calculated from the detected current value Iuvw.

The allowable range calculation unit 18 calculates an allowable range Δpref set for the voltage command integrated value Pref based on the driving state quantity Mstate of the rotating electrical machine 4 calculated by the state observation unit 34, and outputs the calculated allowable range Δpref to the on-off prediction unit 33B. The allowable range Δpref has a value that determines the trade-off relationship between the harmonic voltage Vthd and the on-off switching number SWcount as described above. Therefore, for example, when the torque of the rotating electrical machine 4 is reduced, the harmonic voltage Vthd is set so as to reduce the allowable range Δpref. The allowable range Δpref may be calculated as a function depending on an arbitrary value related to the rotating electrical machine 4.

Fig. 41 is a hardware configuration diagram for realizing the power conversion device 100I.

The hardware configuration of the power conversion device 100I shown in fig. 41 differs from the power conversion device 100G of embodiment 8 in that a current detection unit 17 is newly added between the power conversion unit 1 and the rotating electrical machine 4.

The current detection unit 17 detects the current values of the three phases output from the power conversion unit 1 to the rotating electric machine 4. Here, any current detector such as CT (Current Transformer) detector or shunt resistor may be used for the current detecting unit 17. It is also possible to use a method of detecting the current of two phases among the currents of three phases and calculating the current of the remaining one phase. In addition, a single-shunt current detection method in which the three-phase ac current value is restored by one current detector may be used.

Fig. 42 is a diagram for explaining a method of expanding the voltage allowable value Pdelta and the integrated value Ppred of the voltage predicted value, in which the allowable range Δpref is set for the voltage command integrated value Pref obtained by integrating the voltage command value Vref, in the on-off predicting unit 33B according to embodiment 10.

Fig. 42 shows, as an example, a case where the voltage command integrated value Pref and the integrated value Ppred of the voltage predicted value are extended to a phase of 60 degrees or more of the voltage command value from the prediction start time point in the αβ coordinate which is the 2-phase stationary coordinate system. In fig. 42, the voltage command integrated value Pref is indicated by a solid arrow, and the integrated value Ppred of the voltage predicted value is indicated by 2 broken arrows (on-off state amounts SWS1 to SWS 2).

Therefore, next, similarly to embodiment 9, the voltage allowable value Pdelta and the integrated value Ppred of the voltage predicted value, in which the allowable range Δpref is set for the voltage command integrated value Pref, are expressed by a time function, and the cross arrival time Tcross until the upper limit value Pupper of the voltage allowable value or the lower limit value Plower of the voltage allowable value crosses the integrated value Ppred of the voltage predicted value is calculated.

In fig. 42, the magnitude of the allowable range Δpref changes in the middle of the prediction calculation based on the allowable range Δpref calculated by the allowable range calculating unit 18. Here, for example, a case is assumed in which, in a case where the allowable range Δpref is calculated as a function of the voltage output integrated value Pout, the magnitude of the allowable range Δpref varies depending on the voltage output integrated value Pout.

As shown in fig. 42, the integrated value Ppred of the voltage predicted value crosses the upper limit value Pupper of the voltage allowable value for which the allowable range Δpref is set when the duration T1SW of the on-off state quantity SWS1 (corresponding to the on-off state index SW 3) and then the integrated value Ppred of the voltage predicted value crosses the upper limit value Pupper of the voltage allowable value for which the allowable range Δpref is set when the duration T2SW of the on-off state quantity SWS2 (corresponding to the on-off state index SW 4).

In embodiment 10, since the allowable range Δpref is changed in accordance with the driving state of the rotating electrical machine 4, it is possible to consider not only the harmonic voltage Vthd and the number of on-off switching SWcount of the power conversion unit 1 but also the torque τ, the magnetic flux Φ, and the first-order angular frequency ω of the rotating electrical machine 4, as compared with the case where the allowable range Δpref of embodiment 9 is a constant value ₁ The on-off state quantity SWS of the power conversion unit 1 is determined in a state where the performance of Mloss is lost.

Next, a control operation in the power conversion device 100I according to embodiment 10 will be described with reference to the drawings.

Fig. 43 is a flowchart illustrating a control operation in the power conversion device 100I.

First, the on-off predicting unit 33B obtains the voltage command value Vref and all candidate voltage predicted values vpreadall calculated by the voltage predicting unit 32 (steps S62 to S64).

Next, the current detection unit 17 detects a current flowing through the rotating electric machine 4 (step S76).

Next, the state observation unit 34 calculates a driving state quantity Mstate of the rotating electric machine 4 based on the detected current value Iuvw and the voltage command value Vref detected in step S76 (step S77).

The allowable range calculation unit 18 calculates an allowable range Δpref set for the voltage command integrated value Pref based on the driving state quantity Mstate calculated in step S77 (step S78).

Then, the on-off predicting unit 33B executes processing in the same flow as in embodiment 9 based on the voltage allowable value Pdelta in which the allowable range Δpref is set and the integrated value pprendall of the candidate voltage predicted values for the voltage command integrated value Pref, calculates the total value tcoussum of the time until the allowable range Δpref is set and the integrated value pprendmost of the voltage predicted values are crossed based on Pdelta and pprendall of the phases equal to or greater than 60 degrees extended to the voltage command value, and takes the reciprocal 1/tcousssum of the calculated total value tcoussum of the time until the crossing as the evaluation value J (steps S79 to S81).

The on-off predicting unit 33B searches the on-off state quantity SWS of the power converting unit 1 and the duration Tsw for which the on-off state quantity is continued in the same flow as in the embodiments 8 and 9 based on the evaluation value J calculated in step S81, and determines the on-off state quantity SWS of the power converting unit 1 based on the searched on-off state quantity SWS and duration Tsw (steps S82 and S83).

The power conversion unit 1 converts the dc power into ac power by the on-off state amount SWS of the power conversion unit 1 determined in step S83, and controls the rotating electrical machine 4 (step S84).

The power conversion device 100I according to embodiment 10 includes: a power conversion unit 1 that converts direct-current power from a direct-current power supply 2 into alternating-current power and supplies the alternating-current power to a rotating electrical machine 4; a bus voltage detection unit 15 for detecting a bus voltage Vdc of the dc power supply 2; a voltage prediction unit 32 that calculates a voltage prediction value Vpred based on the bus voltage Vdc and all the candidate on-off state amounts SWSall in the power conversion unit 1; a current detection unit 17 that detects, as a detected current value Iuvw, a current flowing between the power conversion unit 1 and the rotating electrical machine 4; a state observation unit 34 that calculates a drive state quantity Mstate indicating the drive state of the rotating electrical machine 4, based on the voltage command value Vref and the detection current value Iuvw; an allowable range calculation unit 18 that calculates an allowable range Δpref set for the voltage command integrated value Pref based on the driving state quantity Mstate calculated by the state observation unit 34; an on-off predicting unit 33B that calculates on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and a duration Tsw during which the on-off state amounts are continued, based on a voltage allowable value Ppredall in which an allowable range Δpref is set for the voltage command integrated value Pref and the integrated values Ppredall of all the candidate voltage predicted values calculated by the voltage predicting unit 32; and an on-off output unit 28 that determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 based on the on-off state quantity SWS and the duration Tsw calculated by the on-off prediction unit 33B, wherein the on-off prediction unit 33B calculates a total value tcosssum of time until the voltage allowable value Pdelta of the allowable range Δpref and the integrated value Ppredall of all the candidate voltage predicted values intersect with each other, and outputs the on-off state quantity SWS and the duration Tsw that minimize the reciprocal 1/tcosssum of the time until the intersection, and the on-off output unit 28 determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 accordingly.

Therefore, since the power conversion device 100I according to embodiment 10 changes the allowable range Δpref in accordance with the driving state of the rotating electrical machine 4, it is possible to control the on-off state amount SWS of the power conversion unit 1 taking into consideration the torque ripple of the rotating electrical machine 4, the magnetic flux ripple, and the loss Mloss, as compared with embodiment 9.

Here, the evaluation value J is described as the reciprocal 1/tcrossu of the total value tcrossu of the time until the intersection, but the evaluation value J is also taken into consideration by the total value SWcountsum of the on and off switching times of the on-off state quantity SWS, and the evaluation value J is SWcountsum/tcrossum of the phase quantity equal to or greater than 60 degrees across the voltage command value, whereby the rotating electric machine 4 can be driven such that the on-off loss SWloss of the power conversion unit 1 is also reduced. Here, the 60-degree phase is assumed, but even if the phase is not 60 degrees, but a shorter section or a longer section, the on-off state quantity SWS and the duration Tsw of the power conversion unit 1 can be calculated in the same flow.

Embodiment 11

Next, a power conversion device 100J according to embodiment 11 will be described with reference to the drawings.

Fig. 44 is a block diagram showing a configuration of a power conversion device 100J according to embodiment 11. As shown in fig. 44, a power conversion device 100J according to embodiment 11 includes a control device 10H instead of the control device 10G, and the control device 10H includes a trained model 26 instead of the allowable range calculation unit 18 included in the control device 10G. The trained model 26 calculates an allowable range Δpref set for the voltage command integrated value Pref based on the information obtained by machine learning based on the teacher data and based on the driving state quantity Mstate input from the state observation unit 34.

Here, the allowable range Δpref of the power conversion device 100J according to embodiment 11 is calculated using the trained model 26 obtained by performing machine learning, which is different from the allowable range calculation unit 18 of the power conversion device 100I according to embodiment 10. The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 to 10, and description thereof will be omitted, focusing on differences from embodiments 1 to 10.

Fig. 45 is a block diagram illustrating a method of creating a trained model and machine learning based on teacher data.

As shown in fig. 45, the learning unit 50 performs machine learning based on teacher data 57 obtained from the learning data 51 prepared in advance to generate a trained model.

The learning data 51 includes the magnetic flux Φ, torque τ, loss Mloss, and allowable range Δpref set for the voltage command integrated value Pref of the rotating electrical machine 4. The learning data 51 stores the value calculated to drive the rotary electric machine 4, and may be generated by a control method for reducing the on-off loss SWloss of the power conversion unit 1 as compared with the PWM method, that is, a control method using model predictive control, selective harmonic cancellation, low-order harmonic cancellation, or a control method using an optimal pulse pattern.

The learning data 51 is input to the teacher data acquisition unit 52. The teacher data acquisition unit 52 includes an input data acquisition unit 53 and a tag data acquisition unit 54.

The input data acquisition unit 53 acquires the magnetic flux Φ, the torque τ, and the loss Mloss of the rotating electrical machine 4 from the learning data 51 as teacher input data 55, and outputs the teacher input data to the learning unit 50.

The tag data obtaining unit 54 obtains the allowable range Δpref set for the voltage command integrated value Pref from the learning data 51, and outputs the allowable range Δpref to the learning unit 50 as teacher tag data 56.

The teacher data 57 is composed of teacher input data 55 and teacher label data 56, and the learning unit 50 performs machine learning based on the teacher data 57, which is a combination of the teacher input data 55 and the teacher label data 56.

The learning with teacher data of the machine learning in embodiment 11 is performed by a neural network configured by combining perceptrons. Specifically, the magnetic flux Φ, the torque τ, and the loss Mloss of the rotating electric machine 4 are set as the teacher input data 55, and the allowable range Δpref set for the voltage command integrated value Pref is set as the teacher label data 56. The teacher data 57 thus obtained is supplied to the neural network, and learning is repeated while changing the weights for the respective sensors so that the output of the neural network is the same as the teacher label data 56.

In the learning process, the weights are adjusted in such a manner that the output errors of the respective sensors are reduced by repeatedly executing the processing realized by the error back-propagation method. That is, learning with teacher data eliminates errors between the teacher tag data 56 and output data of the neural network while adjusting the weight value.

Thus, the characteristics of the teacher data 57 are learned, and a trained model for deriving a result based on the input reasoning is obtained.

Thus, the trained model generated by machine learning has the characteristics of the teacher data 57. For example, if the learning data 51, which is the teacher data 57, is used for model predictive control, the trained model controls the voltage command integrated value Pref to a value equivalent to that when the rotating electric machine 4 is controlled by the model predictive control, and therefore, the on-off loss SWloss of the power conversion unit 1 can be reduced as compared with the PWM method.

The neural network used in the learning by the learning unit 50 may be three layers, may be more layers, or may be a neural network that performs machine learning by deep learning.

FIG. 46 is a hardware architecture diagram for generating a trained model. The machine learner 60 is implemented by a hardware structure shown in fig. 46 by performing machine learning for generating a trained model by the machine learner 60 functioning as a neural network.

The machine learner 60 is composed of a processor 61 and a storage 62.

The storage device 62 has, for example, a RAM 63 as a volatile storage device, and an HDD 64 as a nonvolatile auxiliary storage device. Further, as the nonvolatile auxiliary storage device, an SSD or a flash memory may be used instead of the HDD.

The HDD 64 stores the learning program 65 and the teacher data 66, and also stores the generated learning result 67.

Various learning programs 65 are input to the processor 61 from the HDD 64 via the RAM 63, and the processor 61 executes the various learning programs 65 input thereto. The learning program 65 causes the processor 61 to perform learning with teacher data. That is, the teacher data 66 is also input from the HDD 64 to the processor 61 via the RAM 63, and is learned in accordance with the learning program 65.

The processor 61 outputs the data of the learning result 67 to the RAM 63 of the storage device 62, and if necessary, stores the data in the HDD 64 via the RAM 63.

The learning program 65 is a program including a command for causing the processor 61 to execute learning with teacher data, and generates data of a result of machine learning (learning result 67).

The machine learner 60 described above may be implemented by PC (Personal Computer), a server device, or the like. However, since the amount of computation is large, GPU (Graphics Processing Units) may be mounted on a PC, for example, and the GPU may be used for computation processing for learning with teacher data by a technique called GPGPU (General-Purpose computing on Graphics Processing Units) to perform processing at high speed.

Next, a control operation in the power conversion device 100J according to embodiment 11 will be described.

Fig. 47 is a flowchart illustrating a control operation in the power conversion device 100J.

First, the on-off predicting unit 33B obtains the voltage command value Vref and all candidate voltage predicted values vpreadall calculated by the voltage predicting unit 32 (steps S62 to S64) by performing the process in the same flow as in embodiment 10.

Next, the state observation unit 34 calculates the magnetic flux Φ, the torque τ, and the loss Mloss of the rotating electric machine 4 as the driving state quantity Mstate based on the voltage command value Vref and the detected current value Iuvw, similarly to embodiment 10 (steps S76 and S77).

Next, based on the driving state quantity Mstate calculated in step S77, the trained model 26 calculates the allowable range Δpref set for the voltage command integrated value Pref (step S85).

Then, the on-off predicting unit 33B executes processing in the same flow as in embodiment 10 based on the voltage allowable value pprendall in which the allowable range Δpref is set and the integrated value pprendall of the candidate voltage predicted values for the voltage command integrated value Pref, calculates the total value tcoussum of the time until the allowable range Δpref is set and the integrated value pprendall of the voltage predicted values for the voltage command integrated value Pref crosses based on pda and pprendall of the phases equal to or greater than 60 degrees extended to the voltage command integrated value Pref, and takes the reciprocal 1/tcosssum of the calculated total value tcosssum of the time until the crossing as the evaluation value J (steps S86 to S88).

The on-off predicting unit 33B searches the on-off state quantity SWS of the power converting unit 1 and the duration Tsw for which the on-off state quantity is continued in the same flow as in embodiment 8 based on the evaluation value J calculated in step S88, and determines the on-off state quantity SWS of the power converting unit 1 based on the searched on-off state quantity SWS and duration Tsw (steps S89, S90).

The power conversion unit 1 converts dc power into ac power by the on-off state amount SWS of the power conversion unit 1 determined in step S90 in the same manner as in embodiment 10, and controls the rotating electrical machine 4 (step S91).

The power conversion device 100J of embodiment 11 includes: a power conversion unit 1 that converts direct-current power from a direct-current power supply 2 into alternating-current power and supplies the alternating-current power to a rotating electrical machine 4; a bus voltage detection unit 15 for detecting a bus voltage Vdc of the dc power supply 2; a voltage prediction unit 32 that calculates a voltage prediction value Vpred based on the bus voltage Vdc and all the candidate on-off state amounts SWSall in the power conversion unit 1; a current detection unit 17 that detects, as a detected current value Iuvw, a current flowing between the power conversion unit 1 and the rotating electrical machine 4; a state observation unit 34 that calculates a drive state quantity Mstate indicating the drive state of the rotating electrical machine 4, based on the voltage command value Vref and the detection current value Iuvw; a trained model 26 for calculating an allowable range Δpref set for the voltage command integrated value Pref based on information obtained by machine learning based on teacher data and based on the driving state quantity Mstate input from the state observation unit 34; an on-off predicting unit 33B that calculates on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and a duration Tsw during which the on-off state amounts are continued, based on a voltage allowable value Ppredall in which an allowable range Δpref is set for the voltage command integrated value Pref and the integrated values Ppredall of all the candidate voltage predicted values calculated by the voltage predicting unit 32; and an on-off output unit 28 that determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 based on the on-off state quantity SWS and the duration Tsw calculated by the on-off prediction unit 33B, wherein the on-off prediction unit 33B calculates a total value tcosssum of time until the voltage allowable value Pdelta of the allowable range Δpref and the integrated value Ppredall of all the candidate voltage predicted values intersect with each other, and outputs the on-off state quantity SWS and the duration Tsw that minimize the reciprocal 1/tcosssum of the time until the intersection, and the on-off output unit 28 determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 accordingly.

Therefore, since power conversion device 100J according to embodiment 11 generates allowable range Δpref set for voltage command integrated value Pref based on driving state quantity Mstate of rotating electric machine 4 and trained model 26, allowable range Δpref can be changed so that performance such as learning data for generating trained model 26 can be obtained as compared with embodiment 10. Therefore, the power conversion device 100J can obtain performance such as the control system of the rotating electric machine 4 used at the time of creation of the learning data.

Here, the evaluation value J is described as the reciprocal 1/tcrossu of the total value tcrossu of the time until the intersection, but the evaluation value J is also taken into consideration by the total value SWcountsum of the on and off switching times of the on-off state quantity SWS, and the evaluation value J is SWcountsum/tcrossum of the phase quantity equal to or greater than 60 degrees across the voltage command value, whereby the rotating electric machine 4 can be driven such that the on-off loss SWloss of the power conversion unit 1 is also reduced. Here, the 60-degree phase is assumed, but even if the phase is not 60 degrees, but a shorter section or a longer section, the on-off state quantity SWS and the duration Tsw of the power conversion unit 1 can be calculated in the same flow.

Embodiment 12

Next, a power conversion device 100K according to embodiment 12 will be described with reference to the drawings.

Fig. 48 is a block diagram showing the structure of power conversion device 100K according to embodiment 12. As shown in fig. 48, a power conversion device 100K according to embodiment 12 is more advantageous in that a control device 10I of the power conversion device 100K includes a speed controller 19, a current controller 20, and a speed estimation calculation unit 21, and the speed estimation calculation unit 21 includes an angular velocity ω to the rotating electric machine 4, as compared with embodiments 10 and 11 _rm And means for estimating. The speed estimation calculation unit 21 calculates the angular speed ω of the rotating electric machine 4 based on the voltage command value Vref and the detected current value Iuvw of the current detection unit 17 _rm And the phase θ.

The speed calculation of the rotating electrical machine 4 of the power conversion device 100K according to embodiment 12 is not performed by the state observation unit 34, but is performed using the newly provided speed estimation calculation unit 21. The power conversion devices 100I and 100J according to embodiments 10 and 11 are different from each other in that the speed estimation calculation unit 21 is provided, in other words, in that an adaptive flux observer is provided. The same reference numerals are given to the constituent elements having the same functions as those of embodiments 1 to 11, and description thereof will be omitted, focusing on differences from embodiments 1 to 11.

Fig. 49 is a block diagram showing the configuration of the speed estimation calculation unit 21. The speed estimation calculation unit 21 is constituted by an adaptive observer, and is configured to calculate the phase θ and the angular velocity ω of the rotating electrical machine 4 _rm An estimation operation is performed. The adaptive observer is also called an adaptive flux observer because it is defined by a state equation in which the stator flux Φs and the rotor flux Φr of the rotating electrical machine 4 are used as state variables. The adaptive observer may be configured using an extended induced voltage, a current, or the like as a state variable.

The speed estimation calculation unit 21 shown in fig. 49 uses the voltage command value Vref and the detected current value Iuvw to calculate the angular speed ω of the rotating electric machine 4 _rm Is described as an estimated value (hereinafter, referred to as an estimated angular velocity value hω _rm ) And an estimated value of the phase θ of the rotating electrical machine 4 (hereinafter, referred to as a phase estimated value hθ) are calculated, and the calculated angular velocity estimated value hω is outputted _rm And a phase estimation value hθ. The voltage command value Vref is a value calculated by the current controller 20 (not shown), and the detected current value Iuvw is a value detected by the current detecting unit 17. Here, the voltage command value Vref is input to the speed estimation computing unit 21, but the voltage output value Vout output from the power conversion unit 1 to the rotating electrical machine 4 may be detected and set as the input value of the speed estimation computing unit 21.

The speed estimation computing unit 21 includes a model deviation computing unit 22, an angular speed estimator 23, a first-order angular frequency computing unit 24, and an integrator 25. The model deviation calculation unit 22 detects a current value Iuvw based on the voltage command value Vref and the first-order angular frequency ω ₁ Estimated value hω of angular velocity _rm The model deviation epsilon is calculated. The angular velocity estimator 23 estimates an angular velocity estimate hω based on the model deviation ε _rm And performing operation. The first-order angular frequency calculator 24 calculates a first-order angular frequency based on an estimated value of the magnetic flux Φ (hereinafter, referred to as a magnetic flux estimated value hΦ) and an estimated value of the current i (hereinafter, referred to as a magnetic flux estimated value h ΦThe surface is described as a current estimated value hi) and an angular velocity estimated value hω _rm For first order angular frequency omega ₁ And performing operation. The integrator 25 is for the first order angular frequency ω ₁ The integration is performed to output a phase estimation value hθ.

The model deviation calculation unit 22 includes a current estimator 221, a subtractor 222, and a deviation calculation unit 223. The current estimator 221 detects a current value Iuvw and a first-order angular frequency ω based on the voltage command value Vref ₁ Estimated value hω of angular velocity _rm The magnetic flux estimated value hΦ and the current estimated value hi are calculated, and the calculated magnetic flux estimated value hΦ and current estimated value hi are outputted. The subtractor 222 calculates a current deviation Ierr by subtracting the detected current value Iuvw from the current estimated value hi, and outputs the calculated current deviation Ierr.

The deviation calculator 223 calculates the model deviation epsilon based on the current deviation Ierr and the magnetic flux estimated value hΦ calculated by the subtractor 222. Here, when the current deviation Ierr is a vector and the magnetic flux estimated value hΦ is a vector, the vector of the current deviation Ierr is taken as an input, the orthogonal component of the vector of the magnetic flux estimated value hΦ is extracted as a scalar, and the extracted scalar is outputted as the model deviation epsilon. As a method of extracting the orthogonal component of the vector of the magnetic flux estimated value hΦ in a scalar quantity, a method of converting the vector of the current deviation Ierr into coordinates on a rotating orthogonal coordinate system and a method of calculating the magnitude of the outer product value between the vector of the current deviation Ierr and the vector of the magnetic flux estimated value hΦ are known.

The current estimator 221 calculates a current estimated value hi and a magnetic flux estimated value hΦ from the state equation of the rotating electrical machine 4. Here, the rotary electric machine 4 is assumed to be a normal embedded permanent magnet synchronous motor, but any other type of motor may be used as long as the state equation of the induction motor, the surface permanent magnet synchronous motor, the winding excitation synchronous motor, the reluctance synchronous motor, or the like is satisfied. That is, the current estimator 221 can estimate the current in the same manner for other types of rotating electrical machines.

In the case where the rotating electric machine 4 is a permanent magnet embedded synchronous motor, the state equation is expressed byThe expression of the above-mentioned expression (1) and expression (2). Here, symbol L _d Inductance representing d-axis, symbol L _q Inductance representing q-axis, symbol i _d Representing d-axis current, symbol i _q Representing q-axis current, symbol Φ _ds Representing d-axis stator flux, symbol Φ _qs Representing the q-axis stator flux, the sign phi _dr The symbol "" represents the d-axis rotor magnetic flux and "" represents the estimated value (for example, "hΦ", which is the estimated value of Φ, is the same as the other estimated values). In addition, the symbol R _a Representing armature resistance, symbol v _d Representing d-axis voltage, symbol v _q Represents q-axis voltage, symbol h ₁₁ 、h ₁₂ 、h ₂₁ 、h ₂₂ 、h ₃₁ 、h ₃₂ Representing the observer gain. Here, the first order angular frequency ω ₁ Given by way of equation (3) above. In formula (3), the symbol h ₄₁ Symbol h ₄₂ Representing the observer gain.

The above-described formulas (1) and (2) are formulas based on a normal induced voltage, but the same calculation can be performed even if the formulas (1) and (2) are deformed to express an extended induced voltage. The above equations (1) and (2) are mathematical expressions in dq coordinates on rotational coordinates, but the same calculation can be performed by performing coordinate transformation on the equations (1) and (2) and expressing them by another coordinate system such as two-phase alternating α β coordinates or three-phase alternating uvw coordinates on stationary coordinates.

Since the above formula (1) contains the estimated angular velocity value hω _rm Therefore, at the angular velocity estimated value hω _rm And the actual angular velocity omega _rm If the current estimated value hi does not match, an error occurs. Here, the model deviation epsilon is defined by the above equation (4), and the angular velocity estimator 23 is used by the velocity estimation calculation unit 21 to estimate the angular velocity value hω _rm Adjustment is made so that the model deviation epsilon becomes zero. The angular velocity estimator 23 is configured by directly connecting an integrator to a proportional-integral controller, for example.

The first-order angular frequency calculator 24 is based on the magnetic flux estimation value hΦ, the current estimation value hi, and the angle based on the above equation (3)Velocity estimation value hω _rm For first order angular frequency omega ₁ And performing operation. The integrator 25 is configured to integrate the first-order angular frequency omega ₁ The phase estimation value hθ is calculated by integrating. The adaptive observer has an advantage of robustness against a fluctuation in the number of flux links and does not generate a speed estimation error in a steady state. Therefore, the adaptive observer can efficiently observe the angular velocity ω of the rotating electrical machine 4 _rm Estimation is performed.

Next, a control operation in the power conversion device 100K according to embodiment 12 will be described with reference to the drawings.

Fig. 50 is a flowchart illustrating a control operation in the power conversion device 100K.

First, the on-off predicting unit 33B obtains all candidate voltage predicted values vpreadall calculated by the voltage predicting unit 32 in the same flow as in embodiment 1 (steps S62 and S63).

The current detection unit 17 detects the current flowing through the rotating electric machine 4 in the same manner as in embodiment 10 (step S92).

Through the flow of fig. 49, the speed estimation computing unit 21 calculates the angular speed ω of the rotating electric machine 4 based on the voltage command value Vref and the detected current value Iuvw _rm The estimated value of (h 2), h omega _rm The estimated value of the phase θ of the rotating electrical machine 4, that is, hθ is calculated (step S93).

Next, based on the angular velocity estimated value hω calculated by the velocity estimation calculation unit 21 _rm And the phase estimated value hθ, the detected current value Iuvw, the speed controller 19 and the current controller 20 execute arithmetic processing, and the on-off predicting unit 33B obtains the voltage command value Vref (step S94).

Next, the state observation unit 34 calculates an angular velocity estimation value hω based on the voltage command value Vref, the detection current value Iuvw, and the velocity estimation calculation unit 21 _rm The phase estimation value hθ calculates the magnetic flux Φ, torque τ, and loss Mloss of the rotating electrical machine 4 as the driving state quantity Mstate (step S95).

Next, the trained model 26 calculates the allowable range Δpref set for the voltage command integrated value Pref based on the driving state quantity Mstate calculated in step S95, as in embodiment 4 (step S96).

Then, the on-off predicting unit 33B executes processing in the same flow as in embodiment 10 based on the voltage allowable value pprendall in which the allowable range Δpref and the integrated value pprendall of the candidate voltage predicted values are set for the voltage command integrated value Pref, calculates the total value tcoussum of the time until the allowable range Δpref is set to intersect with the integrated value pprendall of the voltage predicted values in the 60 degree phase or more based on the Pdelta and pprendall extended to the voltage command value, and takes the calculated reciprocal 1/tcousssum of the time until the intersection as the evaluation value J (steps S97 to S99).

Next, the on-off predicting unit 33B searches the on-off state quantity SWS of the power converting unit 1 and the duration Tsw for which the on-off state quantity continues in the same flow as in embodiment 8 based on the evaluation value J calculated in step S99, and determines the on-off state quantity SWS of the power converting unit 1 based on the searched on-off state quantity SWS and duration Tsw (steps S100, S101).

The power conversion unit 1 converts dc power into ac power by the on-off state amount SWS of the power conversion unit 1 determined in step S101 in the same manner as in embodiment 10, and controls the rotating electrical machine 4 (step S102).

The power conversion device 100K according to embodiment 12 includes: a power conversion unit 1 that converts direct-current power from a direct-current power supply 2 into alternating-current power and supplies the alternating-current power to a rotating electrical machine 4; a bus voltage detection unit 15 for detecting a bus voltage Vdc of the dc power supply 2; a voltage prediction unit 32 that calculates a voltage prediction value Vpred based on the bus voltage Vdc and all the candidate on-off state amounts SWSall in the power conversion unit 1; a current detection unit 17 that detects, as a detected current value Iuvw, a current flowing between the power conversion unit 1 and the rotating electrical machine 4; a speed estimation calculation unit 21 for estimating the angular speed ω of the rotating electric machine 4 based on the voltage command value Vref and the detected current value Iuvw _rm And the estimated value of the phase theta are calculated; a speed controller 19 based on the angular speed command value ωrmref and estimated from the speedThe estimated angular velocity value hω calculated by the calculating unit 21 _rm Calculating a current command value Iref; a current controller 20 that calculates a voltage command value Vref based on the current command value Iref calculated by the speed controller 19, the phase estimation value hθ calculated by the speed estimation operation unit 21, and the detected current value Iuvw detected by the current detection unit 17; a state observation unit 34 for estimating an angular velocity estimated value hω calculated by the velocity estimation calculation unit 21 based on the voltage command value Vref calculated by the current controller 20, the detected current value Iuvw detected by the current detection unit 17 _rm The phase estimation value hθ calculates a driving state quantity Mstate indicating the driving state of the rotating electric machine 4; a trained model 26 for calculating an allowable range Δpref set for the voltage command integrated value Pref based on information obtained by machine learning based on teacher data and based on the driving state quantity Mstate input from the state observation unit 34; an on-off predicting unit 33B that calculates on-off state amounts SWS of the plurality of switching elements of the power converting unit 1 and a duration Tsw during which the on-off state amounts are continued, based on a voltage allowable value Ppredall in which an allowable range Δpref is set for the voltage command integrated value Pref and the integrated values Ppredall of all the candidate voltage predicted values calculated by the voltage predicting unit 32; and an on-off output unit 28 for determining the on-off state quantity SWS and the duration Tsw of the plurality of switching elements of the power conversion unit 1 based on the on-off state quantity SWS and the duration Tsw calculated by the on-off prediction unit 33B, the on-off prediction unit 33B calculating a total value tcoussum of a time until the voltage allowable value Pdelta of the allowable range Δpref and the integrated value Ppredall of all the candidate voltage predicted values intersect with respect to the voltage command integrated value Pref of the voltage command value equal to or greater than 60 degrees of the phase quantity, and outputting the on-off state quantity SWS and the duration Tsw so that the inverse 1/tcousssum of the time until the intersection is minimized, and the on-off output unit 28 determines the on-off state quantity SWS of the plurality of switching elements of the power conversion unit 1 accordingly.

Therefore, the power conversion device 100K according to embodiment 12 uses the speed estimation calculation unit 21 to accurately calculate the angular speed ω of the rotating electric machine 4 _rm Since estimation is performed, in comparison with embodiment 11,the speed control of the rotating electric machine 4 can be made high, and the angular velocity ω when the driving state quantity Mstate, which is the input of the trained model 26, is calculated can be improved _rm Since the accuracy of (a) is also improved, the trained model 26 can appropriately calculate the allowable range Δpref in accordance with the driving state of the rotating electrical machine 4.

Here, the evaluation value J is described as the total value tcousssum of the time until the intersection, but the evaluation value J is also taken into consideration as the total value SWcountsum of the on/off switching times of the on/off state quantity SWS, and the rotating electric machine 4 can be driven so that the on/off loss SWloss of the power conversion unit 1 is also reduced by setting the evaluation value J to SWcountsum/tcousssum of the phase quantity of 60 degrees or more across the voltage command value. Here, the 60-degree phase is assumed, but even if the phase is not 60 degrees, but a shorter section or a longer section, the on-off state quantity SWS and the duration Tsw of the power conversion unit 1 can be calculated in the same flow.

The configuration shown in the above embodiment is an example of the present invention, and other known techniques may be combined, and a part of the configuration may be omitted or changed without departing from the scope of the present invention.

Description of the reference numerals

A power conversion unit, a 2 dc power supply, a 3-load, a 4-rotation motor, 10A, 10B, 10C, 10D, 10E, 10F, 10G, 10H, 10I control devices, 11A voltage output calculation unit, 12 integration value calculation unit, 13A, 13B on-off update determination unit, 14A, 14B on-off determination table, 15 bus voltage detection unit, 16 offset adjustment unit, 17 current detection unit, 18 allowable range calculation unit, 19 speed controller, 20 current controller, 21 speed estimation calculation unit, 22 model deviation calculation unit, 23 angular velocity estimator, 24 first-order angular frequency arithmetic unit, 25 integrator, 26 trained model, 27 on-off calculation unit, 28 on-off output unit, 29 harmonic processing unit, 30 harmonic current controller, 31 frequency extraction unit, 32 voltage prediction unit, 33A, 33B prediction unit, 34 state observation unit, 40 processor, 41 storageA storage device, 42 control program, 43 processing data, 50 learning portion, 51 learning data, 52 teacher data acquisition portion, 53 input data acquisition portion, 54 tag data acquisition portion, 55 teacher input data, 56 teacher tag data, 57 teacher data, 60 machine learning device, 61 processor, 62 storage device, 63ram,64hdd,65 learning program, 66 teacher data, 67 learning result, 100A, 100B, 100C, 100D, 100E, 100F, 100G, 100H, 100I, 100J, 100K power conversion device, 221 current estimator, 222 subtractor, 223 offset arithmetic device, 300 on-off determination portion, C initial value, D diode, ifind current low frequency value, iref current command value, ithd harmonic current, ithdref harmonic current command value, iuvw detection current value, iαβ two-phase current, J evaluation value, mloss loss, mtrete driving state quantity, pdelta voltage allowable value, perr integral deviation (difference between Pref and pd), the total value of Perrsum integral deviation, the lower limit value of a permissible value of a Plower voltage, a Pref voltage instruction integral value, a Pout voltage output integral value, an integral value of a Ppred voltage predicted value, an integral value of a voltage predicted value of all candidates of Ppredall, an upper limit value of a purper voltage permissible value, Q1 to Q6 switching elements, a setSW setting signal, the number of switch on/off times, the total value of switch on/off times, switch loss of switch, switch on/off state quantity of all candidates of switch on/off, switch on/off state quantity of switch on (switch on) of switch on/off state quantity (n is a positive number), switch on/off mode of switch on/off, switch on/off time of switch on, switch on/off arrival time of Tcross, tsw, T1sw, T2sw duration of all candidates of switch on/off, vdc bus voltage, vout voltage output value (multiphase voltage output value), vpres voltage predicted value of all candidates of switch on/off, vref command value, harmonic components of Vthd voltage, hi current estimate, hθ phase estimate, hΦ magnetic flux estimate, hω _rm Angular velocity estimation value, Δpref allowable range, epsilon model deviation, θ phase (phase of rotating electric machine), τtorque, Φ magnetic flux, Φαβ -phase magnetic flux, ω ₁ First order angular frequency omega _rm Angular velocity, ωrmref angular velocity command value.

Claims (15)

1. An electric power conversion device, comprising:

a power conversion unit having a plurality of switching elements, which converts dc power into ac power and supplies the ac power to a load in accordance with the on-off state amounts of the switching elements;

a voltage output calculation unit that calculates a voltage output value of the multi-phase ac supplied from the power conversion unit based on the on-off state quantity;

an integrated value calculation unit that calculates a voltage command integrated value and a voltage output integrated value by integrating the voltage command value of the multi-phase ac and the voltage output value of the multi-phase ac calculated by the voltage output calculation unit, respectively; and

and an on-off determining unit that determines and outputs an on-off state quantity of the power converting unit using the voltage command integrated value and the voltage output integrated value.

2. The power conversion device according to claim 1, wherein,

the on-off determination unit includes:

an on-off calculation unit that outputs a setting signal for determining an on-off state amount of the power conversion unit based on the voltage command integrated value, the voltage output integrated value, and an allowable range input from the outside; and

And an on-off output unit that determines an on-off state amount of the power conversion unit based on the setting signal.

3. The power conversion apparatus according to claim 2, wherein,

the on-off calculation unit calculates an on-off state quantity from a relationship among the voltage instruction integrated value, the voltage output integrated value, and the allowable range.

4. A power conversion apparatus according to claim 2 or 3, wherein,

the on-off calculating unit outputs the on-off state quantity of the power converting unit as the setting signal.

5. The power conversion apparatus according to any one of claims 2 to 4, wherein,

the on-off calculation unit outputs the on-off state quantity of the power conversion unit and the duration of the on-off state quantity as the setting signal.

6. The power conversion apparatus according to any one of claims 2 to 5, wherein,

the on-off calculation unit calculates a duration until the voltage output integral value crosses the limit value of the allowable range set for the voltage command integral value when the current on-off state quantity is continuously used, calculates the on-off state quantity based on any one of the crossed phases, an on-off decision table for deciding the on-off state quantity, a phase of the voltage command integral value at the time of crossing, and a phase of the voltage output integral value, and outputs the duration and the on-off state quantity as the setting signal.

7. The power conversion apparatus according to any one of claims 2, 3, and 5,

the on-off calculation unit calculates a first duration until a limit value of the allowable range set for the voltage command integrated value crosses the voltage output integrated value when the current on-off state quantity is continuously used, calculates a second duration from a crossing position to a next crossing of the limit value of the allowable range and the voltage output integrated value based on a plurality of on-off state quantities, selects a first on-off state quantity having the longest second duration, and outputs the first duration and the first on-off state quantity as the setting signal.

8. The power conversion apparatus according to claim 7, wherein,

the on-off calculating unit selects the first on-off state quantity having the longest second duration, switches the first on-off state quantity after the voltage output integrated value crosses the limit value of the allowable range, calculates a third duration until the voltage output integrated value crosses the limit value of the allowable range next based on a plurality of on-off state quantities, selects the second on-off state quantity having the longest third duration, and outputs a combination of the first and second durations and the on-off state quantity as the setting signal.

9. The power conversion apparatus according to claim 7 or 8, wherein,

the on-off calculation part calculates the N-th on-off state quantity with the longest duration time to the N+1-th on-off state quantity, and outputs the combination of the N durations time and the on-off state quantity.

10. The power conversion apparatus according to any one of claims 7 to 9, characterized in that,

the on-off calculation unit selects the on-off state quantity having the longest duration when the on-off state quantity is selected from the plurality of on-off state quantities, or selects the on-off state quantity having the smallest evaluation value based on the on-off switching times and duration of each phase of the on-off state quantity, and based on the evaluation value obtained from the total value of the on-off switching times and duration of each phase.

11. The power conversion device according to any one of claims 2 to 10, characterized in that,

the load is a rotating electric machine, the power conversion device further includes a current detection unit that detects a current flowing through the rotating electric machine as a detected current value,

the power conversion device includes:

a harmonic processing unit that calculates current harmonic data, which is a harmonic component of the detected current value detected by the current detecting unit;

A harmonic current controller that calculates the allowable range based on the current harmonic data and the harmonic current command value calculated by the harmonic processing unit;

a low-frequency extraction unit that calculates a current low-frequency value that is a low-frequency component of the detected current value detected by the current detection unit; and

and a current controller for calculating a voltage command value of the multi-phase alternating current based on the current low frequency value and the current command value calculated by the low frequency extraction unit.

12. The power conversion device according to claim 11, wherein,

the low frequency extraction unit detects 2 or more of the detection current values, and calculates the current low frequency value based on a plurality of detection current values.

13. The power conversion device according to claim 11 or 12, characterized in that,

the harmonic processing unit detects 2 or more of the detected current values, and calculates the current harmonic data based on the detected current values.

14. The power conversion device according to any one of claims 1 to 13, characterized in that,

the on-off determination unit calculates the voltage command integrated value and the voltage output integrated value after the offset adjustment.

15. The power conversion device according to any one of claims 1 to 14, characterized in that,

the voltage output calculation unit calculates a voltage output value of the multi-phase ac supplied from the power conversion unit based on a dc power supply voltage of the dc power supply and the on-off state quantity.