Control device for an AC rotating machine and an electric power steering device

Abstract

This control device for an alternating-current rotating machine comprises: an inverter that applies a voltage to the alternating-current rotating machine; a direct-current power source that supplies direct-current power to the inverter; a current detector that detects a rotating machine current flowing to the alternating-current rotating machine; a first-axis voltage command value computer that computes a voltage command value for a first axis of two axes of rotation of the alternating-current rotating machine; and a second-axis voltage command value computer that computes a voltage command value for a second axis of the two axes of rotation thereof. The first-axis voltage command value computer computes the voltage command value for the first axis on the basis of: a corrected first deviation determined by multiplying a corrected gain by a first deviation; a second deviation; and a rotational angular velocity. The value to which the corrected gain is set is increased if a physical quantity in the alternating-current rotating machine has changed suddenly or if an output voltage of the direct-current power source has risen.

Description

Technical Field

[0001] The present disclosure relates to a control device for an AC rotating machine and an electric power steering device.

Background Art

[0002] Patent Document 1 discloses a technique related to so-called non-interference control that compensates for interference terms in the d-axis and q-axis (for example, d-axis and q-axis) in vector control of an AC rotating machine. Specifically, the control device of Patent Document 1 includes a non-interference controller that calculates interference components of the d-axis and q-axis, a non-interference error corrector that corrects an error of the interference component, a current PI controller that performs a PI (proportional integral) operation according to current deviations of the d-axis and q-axis, and the like.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the current PI controller of Patent Document 1, if noise is included in the detected rotating machine current value for calculating the current deviation, it may cause an increase in the driving noise or torque ripple in the AC rotating machine. In order to suppress such a phenomenon, it is possible to reduce the responsiveness of the current PI controller, but in that case, the following problems occur. In an AC rotating machine, for example, when the rotational angular velocity suddenly decreases, the induced voltage proportional to the rotational angular velocity also suddenly decreases, and as a result, the output voltage of the inverter may become excessive compared to the voltage required to make the rotating machine current the target value. In addition, such a situation may occur due to reasons other than a sudden decrease in the rotational angular velocity. If the responsiveness of the current PI controller in Patent Document 1 is reduced, in the above situation, the time during which the output voltage of the inverter becomes excessive becomes longer, which becomes a factor causing an overcurrent state in the AC rotating machine.

[0005] The present disclosure has been made in view of the above circumstances, and aims to provide a control device for an AC rotating machine and an electric power steering device capable of suppressing the occurrence of an overcurrent state while suppressing an increase in driving noise and the like caused by noise being included in the detected rotating machine current value.

Means for Solving the Problems

[0006] In order to solve the above problems, a control device for an AC rotating machine according to an aspect of the present disclosure includes an inverter that applies a voltage to the AC rotating machine, a DC power supply that supplies DC power to the inverter, a current detector that detects a rotating machine current flowing through the AC rotating machine, a first-axis voltage command value calculator that calculates a voltage command value for a first axis in a rotating two-axis of the AC rotating machine, and a second-axis voltage command value calculator that calculates a voltage command value for a second axis in the rotating two-axis. The first-axis voltage command value calculator multiplies a correction gain by a correction first deviation that is a deviation between a current command value on the first axis and a detected current value on the first axis of the rotating machine current, a second deviation that is a deviation between a current command value on the second axis and a detected current value on the second axis of the rotating machine current, and calculates the voltage command value for the first axis based on the rotational angular velocity of the AC rotating machine. The set value of the correction gain is increased when a physical quantity in the AC rotating machine changes rapidly or when the output voltage of the DC power supply increases.

[0007] An electric power steering device according to an aspect of the present disclosure includes the control device according to the above aspect, the AC rotating machine, and a driving force transmission mechanism that transmits the driving force of the AC rotating machine to a steering system of a vehicle.

Advantages of the Invention

[0008] According to the present disclosure, it is possible to provide a control device for an AC rotating machine and an electric power steering device that can suppress the occurrence of an overcurrent state while suppressing an increase in driving noise or the like caused by noise being included in the detected rotating electrical current value.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Embodiments for Carrying out the Invention

[0010] Hereinafter, with reference to the drawings, the control device for a rotating machine and the electric power steering apparatus according to the present disclosure will be described in detail.

[0011] Embodiment 1. FIG. 1 is a block diagram showing the schematic configuration of the control device for a rotating machine according to Embodiment 1. As shown in FIG. 1, the control device 1 includes a rotor position detector 11, an inverter 12, and a controller 13. The control device 1 controls the rotating machine 10 based on a torque command T_ref as a control command input from the outside of the control device 1.

[0012] The rotating electrical machine 10 is a three-phase AC rotating electrical machine having three-phase windings U, V, and W. Further, the rotating electrical machine 10 is an AC rotating electrical machine that can be controlled based on two rotating axes. In this specification, the "two rotating axes" means two axes that rotate in synchronization with the rotor of the rotating electrical machine 10 and are orthogonal to each other in the cross-sectional plane. The "cross-sectional plane" is a cross-section perpendicular to the central axis of the rotor. For example, the two rotating axes may be the d-q axes. The d-axis is an axis connecting the central axis of the rotor and the magnetic pole. The q-axis is an axis orthogonal to both the d-axis and the central axis. Further, the two rotating axes may be the γ-δ axes. The γ-axis is an axis shifted in the rotation direction with respect to the d-axis. The δ-axis is an axis orthogonal to both the γ-axis and the central axis. Of the two rotating axes, one is referred to as the first axis and the other is referred to as the second axis. For example, when the d-axis is the first axis, the q-axis is the second axis. Note that the q-axis may be the first axis and the d-axis may be the second axis. Similarly, when the γ-axis is the first axis, the δ-axis is the second axis.

[0013] Hereinafter, a case where the rotating electrical machine 10 is a permanent magnet synchronous rotating electrical machine and the two rotating axes are the d-q axes will be described. However, the rotating electrical machine 10 may be, for example, a wound field synchronous rotating electrical machine, an induction rotating electrical machine, a synchronous reluctance motor, or the like. Further, the d-axis and the q-axis in the following disclosure content may be replaced with the δ-axis and the γ-axis.

[0014] The rotor position detector 11 includes a resolver, an encoder, an MR (magnetoresistive) sensor, etc., and detects the rotor position θ using these. The rotor position θ is the position of the rotor of the rotating electrical machine 10 in the rotation direction. In the present embodiment, the rotor position θ of the rotating electrical machine 10 is detected using the rotor position detector 11. However, a configuration in which the rotor position θ of the rotating electrical machine 10 is estimated without using the rotor position detector 11 can also be adopted. That is, in the present disclosure, the control device 1 may not include the rotor position detector 11.

[0015] The inverter 12 is a power converter that applies a voltage to the rotating machine 10. Specifically, under the control of the controller 13, the inverter 12 converts the DC power supplied from the DC power source BT into AC power and supplies the converted AC power to the rotating machine 10. The DC power source BT includes devices for supplying DC power, such as a DC-DC converter, a diode rectifier, and a PWM rectifier, in addition to the battery. In this specification, the output voltage (DC bus voltage) of the DC power source BT is represented as Vdc.

[0016] The inverter 12 includes upper arm switching elements Sup, Svp, Swp, lower arm switching elements Sun, Svn, Swn, and shunt resistors Ru, Rv, Rw. The upper arm switching elements Sup, Svp, Swp are connected to the positive electrode of the DC power source BT. The lower arm switching elements Sun, Svn, Swn are respectively connected to the upper arm switching elements Sup, Svp, Swp and are connected to the negative electrode of the DC power source BT via the shunt resistors Ru, Rv, Rw.

[0017] Here, a U-phase series circuit is formed by the upper arm switching element Sup, the lower arm switching element Sun, and the shunt resistor Ru. In this U-phase series circuit, the connection point between the upper arm switching element Sup and the lower arm switching element Sun is connected to the winding U of the rotating machine 10.

[0018] Also, a V-phase series circuit is formed by the upper arm switching element Svp, the lower arm switching element Svn, and the shunt resistor Rv. In this V-phase series circuit, the connection point between the upper arm switching element Svp and the lower arm switching element Svn is connected to the winding V of the rotating machine 10.

[0019] Also, a W-phase series circuit is formed by the upper arm switching element Swp, the lower arm switching element Swn, and the shunt resistor Rw. In this W-phase series circuit, the connection point between the upper arm switching element Swp and the lower arm switching element Swn is connected to the winding W of the rotating machine 10.

[0020] As the upper arm switching elements Sup, Svp, Swp and the lower arm switching elements Sun, Svn, Swn, for example, semiconductor switches such as IGBT (Insulated Gate Bipolar Transistor), bipolar transistor, MOSFET (Metal - Oxide - Semiconductor Field Effect Transistor) can be used.

[0021] The switching signals Gup, Gvp, Gwp output from the controller 13 are respectively input to the upper arm switching elements Sup, Svp, Swp. The switching signals Gun, Gvn, Gwn output from the controller 13 are respectively input to the lower arm switching elements Sun, Svn, Swn. The upper arm switching elements Sup, Svp, Swp and the lower arm switching elements Sun, Svn, Swn are turned on or off by the switching signals Gup, Gvp, Gwp, Gun, Gvn, Gwn output from the controller 13. In this specification and the drawings, the switching signals Gup, Gvp, Gwp, Gun, Gvn, Gwn may be collectively referred to as "switching signals Gup~Gwn".

[0022] For example, when the switching signal Gup is "on command (=1)", the upper arm switching element Sup is turned on, and when the switching signal Gup is "off command (=0)", the upper arm switching element Sup is turned off. The same applies to the other switching elements (the upper arm switching elements Svp, Swp and the lower arm switching elements Sun, Svn, Swn). In this way, the inverter 12 generates AC power supplied to the rotary machine 10 from the DC power supplied by the DC power supply BT.

[0023] The shunt resistors Ru, Rv, Rw are resistance elements for current detection. The shunt resistor Ru outputs the terminal voltage VRu (=-Ru×iu) proportional to the current (rotating machine current) iu flowing through the winding U of the rotating machine 10 to the controller 13. The shunt resistor Rv outputs the terminal voltage VRv (=-Rv×iv) proportional to the current (rotating machine current) iv flowing through the winding V of the rotating machine 10 to the controller 13. The shunt resistor Rw outputs the terminal voltage VRw (=-Rw×iw) proportional to the current (rotating machine current) iw flowing through the winding W of the rotating machine 10 to the controller 13. In this specification and the drawings, the terminal voltages VRu, VRv, VRw may be collectively referred to as "terminal voltages VRu~VRw".

[0024] Here, the terminal voltages VRu, VRv, VRw are values obtained by multiplying the rotating machine currents iu, iv, iw and the resistance values of the shunt resistors Ru, Rv, Rw, and are amounts proportional to the currents iu, iv, iw. Therefore, it can be said that the terminal voltages VRu, VRv, VRw are values for detecting the current (detected values of the rotating machine current). Note that the inverter 12 may be integrated with the rotating machine 10. The integrated inverter 12 and rotating machine 10 are called a power pack.

[0025] The controller 13 uses the torque command T_ref, the terminal voltages VRu, VRv, VRw, and the rotor position θ as input values, and based on these, generates switching signals Gup~Gwn for driving the inverter 12. The controller 13 is, for example, a PWM controller realized by a discrete-time arithmetic unit such as a microcomputer or a DSP (Digital Signal Processor). The controller 13 includes a current command value calculator 21, a current detector 22, a coordinate converter 23 (detection coordinate converter), a voltage command value calculation unit 24, a coordinate converter 25 (control coordinate converter), a correction voltage generator 26, a PWM signal generator 27, a speed calculator 28, a change amount calculator 30, and an absolute value calculator 31.

[0026] The speed calculator 28 calculates the rotational angular velocity ω of the rotating machine 10 by performing a differential operation or a difference operation on the rotor position θ. The speed calculator 28 inputs the obtained rotational angular velocity ω to the current command value calculator 21, the d-axis voltage command value calculator 24d, and the q-axis voltage command value calculator 24q of the voltage command value calculation unit 24. Based on the torque command T_ref, the DC bus voltage Vdc, and the rotational angular velocity ω, the current command value calculator 21 calculates the current command values id_ref and iq_ref. The current command values id_ref and iq_ref are the command values (target values) of the current supplied to the rotating machine 10. id_ref is also called the "field weakening current command value", and iq_ref is also called the "torque current command value". As the calculation method executed by the current command value calculator 21, known MTPA (Maximum Torque Per Ampere) control, MTPV (Maximum Torque Per Voltage) control, and field weakening control may be appropriately combined and used for each operation range (the range of speed-torque characteristics). Note that the rotational angular velocity ω may be converted into the rotational speed of the rotor, and each control may be performed using the rotational speed.

[0027] Next, the PWM signal generator 27 will be described. Based on the corrected voltage command values vu′, vv′, vw′ output from the corrected voltage generator 26, the PWM (Pulse Width Modulation) signal generator 27 outputs the switched signals Gup to Gwn. The corrected voltage command values vu′, vv′, vw′ will be described later.

[0028] Figure 2 is a diagram for explaining the generation principle of the switching signal in Embodiment 1. The PWM signal generator 27 generates the switching signals Gup to Gwn by comparing the corrected voltage command values vu′, vv′, vw′ with a carrier triangular wave (carrier wave) C having a period Tc (frequency fc). The corrected voltage command values vu′, vv′, vw′ respectively correspond to the U phase, V phase, and W phase.

[0029] Specifically, if the corrected voltage command value vu′ is greater than the carrier triangular wave C, the PWM signal generator 27 turns on the switching signal Gup (“1”) and turns off the switching signal Gun (“0”). Conversely, if the corrected voltage command value vu′ is less than the carrier triangular wave C, the PWM signal generator 27 turns off the switching signal Gup (“0”) and turns on the switching signal Gun (“1”).

[0030] Also, if the corrected voltage command value vv′ is greater than the carrier triangular wave C, the PWM signal generator 27 turns on the switching signal Gvp (“1”) and turns off the switching signal Gvn (“0”). Conversely, if the corrected voltage command value vv′ is less than the carrier triangular wave C, the PWM signal generator 27 turns off the switching signal Gvp (“0”) and turns on the switching signal Gvn (“1”).

[0031] Also, if the corrected voltage command value vw′ is greater than the carrier triangular wave C, the PWM signal generator 27 turns on the switching signal Gwp (“1”) and turns off the switching signal Gwn (“0”). Conversely, if the corrected voltage command value vw′ is less than the carrier triangular wave C, the PWM signal generator 27 turns off the switching signal Gwp (“0”) and turns on the switching signal Gwn (“1”).

[0032] Note that a short - circuit prevention time (dead time) may be provided for the switching signals Gup to Gwn so that the upper - arm switching elements Sup, Svp, Swp and the lower - arm switching elements Sun, Svn, Swn of the inverter 12 do not simultaneously turn on.

[0033] The switching signals Gup to Gwn include a pattern in which all of the lower - arm switching elements Sun, Svn, Swn are turned on during one electrical angle cycle of the rotating machine 10. Specifically, as in the section D in FIG. 2, a pattern in which all of the switching signals Gun, Gvn, Gwn are turned on (1) is included.

[0034] Here, the voltage after PWM modulation applied from the inverter 12 to the rotating machine 10 includes components of the corrected voltage command values vu′, vv′, vw′, as well as components that are integer multiples of the period Tc of the carrier triangular wave C. Due to this, currents with components that are integer multiples of the period Tc are supplied to the rotating machine 10, and depending on the value of the period Tc, the rotating machine 10 may generate abnormal noise.

[0035] To prevent the generation of such abnormal noise, for example, when the rotating machine 10 is used as a motor for performing steering assist of an electric power steering, the period Tc of the carrier triangular wave C may be set to 60 [μs] or less. By setting Tc = 60 [μs], the frequency fc (= 1 / Tc) of the abnormal noise becomes 16.6 kHz, which is less likely to be noise that humans feel uncomfortable with. More preferably, the period Tc of the carrier triangular wave C may be set to about 50 [μs]. By setting Tc = 50 [μs], the frequency fc (= 1 / Tc) of the abnormal noise becomes about 20 kHz, which is almost inaudible to humans. The frequency band that humans can hear is about 20 Hz to 20 kHz. Hereinafter, Tc = 50 [μs] will be described.

[0036] Subsequently, the current detector 22 shown in FIG. 1 will be described. The current detector 22 outputs the pre-conversion detected currents ius, ivs, iws using the voltages VRu, VRv, VRw across the shunt resistors Ru, Rv, Rw and the switching signals Gup to Gwn output from the PWM signal generator 27. Specifically, the current detector 22 acquires the voltages VRu, VRv, VRw across the shunt resistors Ru, Rv, Rw at the timing "X" shown in FIG. 2. This timing "X" is the timing when the carrier triangular wave C reaches its maximum value (DC bus voltage Vdc).

[0037] At the timing "X", as shown in FIG. 2, all the switching signals Gun, Gvn, Gwn input to the lower arm switching elements Sun, Svn, Swn are on ("1"). Therefore, the current detector 22 obtains the values of the pre-conversion detected currents ius, ivs, iws by dividing the voltages VRu, VRv, VRw across the shunt resistors Ru, Rv, Rw by -Ru, -Rv, -Rw, respectively.

[0038] The coordinate converter 23 performs coordinate conversion based on the pre-conversion detected currents ius, ivs, iws detected by the current detector 22 and the rotor position θ detected by the rotor position detector 11. As a result, the coordinate converter 23 calculates the detected current values id, iq on the rotating two axes (d, q axes). Further, the coordinate converter 23 inputs the calculation results (the detected current values id, iq after coordinate conversion) to the voltage command value calculation unit 24.

[0039] The voltage command value calculation unit 24 calculates the voltage command values vd, vq on the rotating two axes (d, q axes) based on the current command values id_ref, iq_ref calculated by the current command value calculator 21, the detected current values id, iq, the rotational angular velocity ω, the DC bus voltage Vdc, and the change amount Δωabs of the absolute value of the angular velocity calculated by the change amount calculator 30. The change amount calculator 30 will be described later. Hereinafter, the details of the voltage command value calculation unit 24 will be described.

[0040] The voltage command value calculation unit 24 includes a first deviation calculator 24a, a second deviation calculator 24b, a d-axis voltage command value calculator 24d, and a q-axis voltage command value calculator 24q. The first deviation calculator 24a calculates a d-axis current deviation ed, which is the deviation between the d-axis current command value id_ref and the d-axis detected current value id. The second deviation calculator 24b calculates a q-axis current deviation eq, which is the deviation between the q-axis current command value iq_ref and the q-axis detected current value iq. The value of the d-axis current deviation ed calculated by the first deviation calculator 24a is input to the d-axis voltage command value calculator 24d and the q-axis voltage command value calculator 24q. The value of the q-axis current deviation eq calculated by the second deviation calculator 24b is input to the d-axis voltage command value calculator 24d and the q-axis voltage command value calculator 24q.

[0041] The d-axis voltage command value calculator 24d calculates the d-axis voltage command value vd using the d-axis current deviation ed, the q-axis current deviation eq, the DC bus voltage Vdc, the rotational angular velocity ω, and the change amount Δωabs. The q-axis voltage command value calculator 24q calculates the q-axis voltage command value vq using the q-axis current deviation eq, the d-axis current deviation ed, the DC bus voltage Vdc, the rotational angular velocity ω, and the change amount Δωabs.

[0042] In this specification, the deviation between the current command value and the detected current value on the first axis of the rotating two axes may be referred to as the first deviation. Similarly, the deviation between the current command value and the detected current value on the second axis of the rotating two axes may be referred to as the second deviation. For example, when the d-axis is the first axis, the d-axis current deviation ed, which is the deviation between the d-axis current command value id_ref and the d-axis detected current value id, is the "first deviation", and the q-axis current deviation eq is the "second deviation". Similarly, when the q-axis is the first axis, the q-axis current deviation eq is the "first deviation", and the d-axis current deviation ed is the "second deviation". Furthermore, the corrected first deviation may be referred to as the "corrected first deviation", and the corrected second deviation may be referred to as the "corrected second deviation".

[0043] The details of the d-axis voltage command value calculator 24d are shown in FIG. 3, and the details of the q-axis voltage command value calculator 24q are shown in FIG. 4. Hereinafter, with reference to FIG. 3, the d-axis voltage command value calculator 24d will be described. The d-axis voltage command value calculator 24d includes a d-axis transient amplifier 120d, a d-axis proportional amplifier 101d, a limiter 102d, a d-axis integral amplifier 103d, a high-pass filter (HPF) 104d, a response angular frequency amplifier 105d, an inductance amplifier 106d, a multiplier 107d, an integrator 109d (self-axis integrator), an other-axis integral input amplifier 121d, an integrator 122d (other-axis integrator), an other-axis integral output amplifier 123d, a subtractor 124d, a limiter 110d, and an adder 111d.

[0044] The d-axis transition amplifier 120d multiplies the d-axis current deviation ed by Ktrd to calculate the corrected d-axis current deviation ed’. That is, ed’ = ed × Ktrd. “Ktrd” is a correction gain whose value changes based on the change amount Δωabs, and details will be described later. The d-axis proportional amplifier 101d calculates the d-axis proportional output Vdp by multiplying the corrected d-axis current deviation ed’ by Kpd. That is, Vdp = ed’ × Kpd. “Kpd” is the d-axis proportional gain that is multiplied to obtain a preferable response of the rotational electrical machine current actually flowing with respect to the current command value id_ref. For example, Kpd = ωcc × Ld. Here, ωcc is the response angular frequency (more specifically, the reciprocal of the time constant of the feedback control system) for adjusting so that the frequency response of the rotational electrical machine current with respect to the current command value is in a preferable range, and Ld is the d-axis inductance of the rotational electrical machine 10. However, the value of Kpd is not limited to ωcc × Ld, and may be appropriately adjusted by actually measuring the responsiveness of the rotational electrical machine current actually flowing with respect to the current command value id_ref, etc. The d-axis proportional output Vdp calculated by the d-axis proportional amplifier 101d is input to the limiter 102d.

[0045] The limiter 102d compares the d-axis proportional output Vdp with the upper limit value (Vlimit) and the lower limit value (-Vlimit), and outputs the limited d-axis proportional output Vdp’ based on the comparison result. However, Vlimit = Kmax·Vdc / 2 0.5 is set. Kmax is the maximum voltage utilization rate of the inverter 12, and is appropriately set according to the desired output. For example, Kmax may be 1.

[0046] The specific operation performed by the limiter 102d is as follows. (A) When Vdp < -Vlimit, Vdp’ = -Vlimit (B) When -Vlimit ≤ Vdp ≤ Vlimit, Vdp’ = Vdp (C) When Vlimit < Vdp, Vdp’ = Vlimit

[0047] That is, when the d-axis proportional output Vdp exceeds the upper limit value (Vlimit) or falls below the lower limit value (-Vlimit), the limiter 102d limits the value of Vdp and outputs it as the limited d-axis proportional output Vdp'. Also, when the d-axis proportional output Vdp is equal to or greater than the lower limit value and equal to or less than the upper limit value, the limiter 102d outputs the value of Vdp as the limited d-axis proportional output Vdp' without change. The value of the limited d-axis proportional output Vdp' output by the limiter 102d is input to the limiter 110d and the adder 111d.

[0048] The d-axis integral amplifier 103d calculates the self-axis integral input Cid by multiplying the d-axis current deviation ed by Kid. That is, Cid = ed × Kid. "Kid" is the integral gain multiplied to make the steady-state value of the d-axis current deviation ed zero. For example, Kid = ωcc × R. Here, R is the winding resistance value of the rotating machine 10. However, the value of Kid is not limited to ωcc × R and may be adjusted as appropriate based on measurement results and the like. The self-axis integral input Cid calculated by the d-axis integral amplifier 103d is input to the integrator 109d.

[0049] The high-pass filter 104d reduces the low-frequency component of the q-axis current deviation eq and outputs it to the response angular frequency amplifier 105d. The response angular frequency amplifier 105d multiplies the q-axis current deviation eq with its low-frequency component reduced by the high-pass filter 104d by ωcc and outputs it to the multiplier 107d.

[0050] The inductance amplifier 106d multiplies the rotational angular velocity ω by Lq and outputs it to the multiplier 107d. Note that "Lq" is the value of the q-axis inductance of the rotating machine 10. The multiplier 107d obtains the cross-axis integral input Bid by multiplying the output of the response angular frequency amplifier 105d and the output of the inductance amplifier 106d. The multiplier 107d outputs the cross-axis integral input Bid to the cross-axis integral input amplifier 121d. The cross-axis integration input amplifier 121d multiplies the cross-axis integration input Bid by Kind to calculate the corrected cross-axis integration input Bid'. "Kind" is an integration input switching gain that changes based on the d-axis current deviation ed, and its value will be described later. The integrator 122d inputs the corrected cross-axis integration input Bid' and performs an integration operation, and outputs the result as the cross-axis integration output Vdi2. The cross-axis integration output amplifier 123d multiplies the cross-axis integration output Vdi2 by Koutd to calculate the corrected cross-axis integration output Vdi2'. "Koutd" is an integration output switching gain that changes based on the change amount Δωabs, and its value will be described later.

[0051] The integrator 109d performs an integration operation on the d-axis integration input Cid, and outputs the result as the d-axis integration output Vdi1 to the subtracter 124d. The subtracter 124d subtracts the d-axis integration output Vdi1 from the corrected cross-axis integration output Vdi2', and outputs the result as the d-axis integration output Vdi to the limiter 110d.

[0052] The limiter 110d compares the d-axis integration output Vdi with the upper limit value (Vlimit - Vdp') and the lower limit value (-Vlimit - Vdp'), and outputs the limited d-axis integration output Vdi' based on the comparison result. The specific operation performed by the limiter 110d is as follows. (D) When Vdi < -Vlimit - Vdp', Vdi' = -Vlimit - Vdp' (E) When -Vlimit - Vdp' ≤ Vdi ≤ Vlimit - Vdp', Vdi' = Vdi (F) When Vlimit - Vdp' < Vdi, Vdi' = Vlimit - Vdp'

[0053] That is, when the d-axis integrated output Vdi exceeds the upper limit value (Vlimit - Vdp’) or is less than the lower limit value (-Vlimit - Vdp’), the limiter 110d limits the value of Vdi and outputs it as the limited d-axis integrated output Vdi’. Also, when the d-axis integrated output Vdi is greater than or equal to the lower limit value and less than or equal to the upper limit value, the limiter 110d outputs the value of Vdi as the limited d-axis integrated output Vdi’ as it is. The value of the limited d-axis integrated output Vdi’ output by the limiter 110d is input to the adder 111d.

[0054] Here, the reason why the limiter 110d performs the limiting determination using the limited d-axis proportional output Vdp’ is to prevent the excessive accumulation of the limited d-axis integrated output Vdi’ and obtain the anti-windup effect. The adder 111d adds the limited d-axis proportional output Vdp’ and the limited d-axis integrated output Vdi’ to obtain the voltage command value vd for the d-axis. The obtained voltage command value vd is input to the coordinate converter 25 as shown in FIG. 1.

[0055] Next, the q-axis voltage command value calculator 24q will be described with reference to FIG. 4. The q-axis voltage command value calculator 24q includes a q-axis transient amplifier 120q, a q-axis proportional amplifier 101q, a limiter 102q, a q-axis integral amplifier 103q, a high-pass filter (HPF) 104q, a response angular frequency amplifier 105q, an inductance amplifier 106q, a multiplier 107q, an integrator 109q (self-axis integrator), an integrator 122q (other-axis integrator), an other-axis integrated output amplifier 123q, an adder 124q, a limiter 110q, and an adder 111q.

[0056] The q-axis transient amplifier 120q multiplies the q-axis current deviation eq by Ktrq to calculate the corrected q-axis current deviation eq’. That is, eq’ = eq × Ktrq. “Ktrq” is a correction gain whose value changes based on the change amount Δωabs, and the details will be described later. The q-axis proportional amplifier 101q calculates the q-axis proportional output Vqp by multiplying the corrected q-axis current deviation eq’ by Kpq. That is, Vqp = eq’ × Kpq. "Kpq" is the q-axis proportional gain multiplied to obtain a preferable response of the rotational machine current actually flowing with respect to the current command value iq_ref. For example, Kpq = ωcc × Lq. Here, as described above, ωcc is the response angular frequency and Lq is the q-axis inductance. However, the value of Kpq is not limited to ωcc × Lq and may be appropriately adjusted by actually measuring the responsiveness of the rotational machine current actually flowing with respect to the current command value iq_ref. The q-axis proportional output Vqp calculated by the q-axis proportional amplifier 101q is input to the limiter 102q.

[0057] The limiter 102q compares the q-axis proportional output Vqp with the upper limit value (Vlimit) and the lower limit value (-Vlimit), and outputs the limited q-axis proportional output Vqp’ based on the comparison result. The specific calculation performed by the limiter 102q is as follows. (G) When Vqp < -Vlimit, Vqp’ = -Vlimit (H) When -Vlimit ≤ Vqp ≤ Vlimit, Vqp’ = Vqp (I) When Vlimit < Vqp, Vqp’ = Vlimit

[0058] That is, when the q-axis proportional output Vqp exceeds the upper limit value (Vlimit) or is less than the lower limit value (-Vlimit), the limiter 102q limits the value of Vqp and outputs it as the limited q-axis proportional output Vqp’. Also, when the q-axis proportional output Vqp is greater than or equal to the lower limit value and less than or equal to the upper limit value, the limiter 102q outputs the value of Vqp as the limited q-axis proportional output Vqp’ as it is. The value of the limited q-axis proportional output Vqp’ output by the limiter 102q is input to the limiter 110q and the adder 111q.

[0059] The q-axis integral amplifier 103q calculates the self-axis integral input Ciq by multiplying the q-axis current deviation eq by Kiq. That is, Ciq = eq × Kiq. Here, Kiq is the integral gain multiplied to make the steady-state value of the q-axis current deviation eq zero. For example, Kiq = ωcc × R. However, the value of Kiq is not limited to ωcc × R and may be adjusted as appropriate based on measurement results and the like. The self-axis integral input Ciq calculated by the q-axis integral amplifier 103q is input to the integrator 109q.

[0060] The high-pass filter 104q reduces the low-frequency component of the d-axis current deviation ed and outputs it to the response angular frequency amplifier 105q. The response angular frequency amplifier 105q multiplies the d-axis current deviation ed with its low-frequency component reduced by the high-pass filter 104q by ωcc and outputs it to the multiplier 107q.

[0061] The inductance amplifier 106q multiplies the rotational angular velocity ω by Ld and outputs it to the multiplier 107q. The multiplier 107q obtains the cross-axis integral input Biq by multiplying the output of the response angular frequency amplifier 105q and the output of the inductance amplifier 106q. The multiplier 107q outputs the cross-axis integral input Biq to the integrator 122q. The integrator 122q inputs the cross-axis integral input Biq and performs an integration operation, and outputs the result as the cross-axis integral output Vqi2. In the cross-axis integral output amplifier 123q, the cross-axis integral output Vqi2 is multiplied by Koutq to calculate the corrected cross-axis integral output Vqi2’. "Koutq" is an integral output switching gain whose value changes based on the change amount Δωabs, and the details will be described later.

[0062] The integrator 109q performs an integration operation on the q-axis integral input Ciq and outputs the result as the q-axis integral output Vqi1 to the adder 124q. The adder 124q adds the q-axis integral output Vqi1 and the corrected cross-axis integral output Vqi2’ and outputs the result as the q-axis integral output Vqi.

[0063] The limiter 110q compares the q-axis integrated output Vqi with the upper limit value (Vlimit - Vqp') and the lower limit value (-Vlimit - Vqp'), and outputs the limited q-axis integrated output Vqi' based on the comparison result. The specific operations performed by the limiter 110q are as follows. (J) When Vqi < -Vlimit - Vqp', Vqi' = -Vlimit - Vqp' (K) When -Vlimit - Vqp' ≤ Vdi ≤ Vlimit - Vqp', Vqi' = Vqi (L) When Vlimit - Vqp' < Vqi, Vqi' = Vlimit - Vqp'

[0064] That is, when the q-axis integrated output Vqi exceeds the upper limit value (Vlimit - Vqp') or is less than the lower limit value (-Vlimit - Vqp'), the limiter 110q limits the value of Vqi and outputs it as the limited q-axis integrated output Vqi'. Also, when the q-axis integrated output Vqi is greater than or equal to the lower limit value and less than or equal to the upper limit value, the limiter 110q outputs the value of Vqi as the limited q-axis integrated output Vqi' as it is. The value of the limited q-axis integrated output Vqi' output by the limiter 110q is input to the adder 111q.

[0065] Here, the reason why the limiter 110q performs the limit determination using the limited q-axis proportional output Vqp' is to prevent the excessive accumulation of the limited q-axis integrated output Vqi' and to obtain the effect of anti-windup. The adder 111q adds the limited q-axis proportional output Vqp' and the limited q-axis integrated output Vqi' to obtain the voltage command value vq for the q-axis. The obtained voltage command value vq is input to the coordinate converter 25.

[0066] As shown in FIG. 1, the coordinate converter 25 receives the voltage command values vd, vq on the rotating two axes (d, q axes) and the rotor position θ. The coordinate converter 25 performs coordinate conversion on the voltage command values vd, vq based on the rotor position θ and calculates the voltage command values vu, vv, vw on the three-phase coordinates.

[0067] The correction voltage generator 26 generates correction voltage command values vu′, vv′, vw′ based on the voltage command values vu, vv, vw output from the coordinate converter 25 and the offset voltage voffset. FIG. 5 is a flowchart showing the processing performed by the correction voltage generator 26 in the first embodiment.

[0068] When the processing of the flowchart shown in FIG. 5 starts, first, step S11 is executed. In step S11, the correction voltage generator 26 selects the smallest value among the voltage command values vu, vv, vw and sets that value as Vmin. Next, in step S12, the correction voltage generator 26 sets the value of the offset voltage voffset to 0.5Vdc - Vmin. Next, in step S13, the correction voltage generator 26 obtains the correction voltage command values vu′, vv′, vw′ by subtracting the offset voltage voffset from the voltage command values vu, vv, vw respectively.

[0069] FIG. 6 is a diagram showing an example of the waveforms of the voltage command values vu, vv, vw and the correction voltage command values vu′, vv′, vw′ in the first embodiment. In FIG. 6, the waveform graphs of the voltage command values vu, vv, vw are shown on the upper side, and the waveform graphs of the correction voltage command values vu′, vv′, vw′ are shown on the lower side. However, in each graph shown in FIG. 6, the DC bus voltage Vdc = 10V is assumed. Here, the voltage range that the inverter 12 can output is from 0 (the minimum value of the carrier triangular wave C) to Vdc (the maximum value of the carrier triangular wave C). Therefore, when the DC bus voltage Vdc = 10V, the voltage range that the inverter 12 can output is 0V to 10V as shown in FIG. 6.

[0070] By performing the processing as in steps S11 to S13 shown in FIG. 5, as shown in the lower graph of FIG. 6, the minimum value among the corrected voltage command values vu′, vv′, vw′ always coincides with the inverter output lower limit value. This means that the corrected voltage command values vu′, vv′, vw′ are offset downward as equally as possible among the three phases within the range where voltage saturation does not occur, and the on-time of the lower-arm switching elements Sun, Svn, Swn is made as large as possible. This point is advantageous from the viewpoint of current detection in an inverter provided with shunt resistors Ru, Rv, Rw in series with the lower-arm switching elements Sun, Svn, Swn as in this embodiment.

[0071] Note that the calculation method of the offset voltage voffset is not limited to the method described above. For example, in step S11 of FIG. 5, in addition to the process of selecting the smallest value from among the voltage command values vu, vv, vw and setting it to Vmin, a process of selecting the largest value from among the voltage command values vu, vv, vw and setting it to Vmax may be performed. Then, in step S12, a process of setting voffset to "(Vmax + Vmin) / 2" may be performed. In this way, the method of setting voffset = (Vmax + Vmin) / 2 is called third-harmonic addition or HIP modulation. Also, the upper-arm switching element corresponding to the phase with the largest voltage command value among the three phases may always be kept in the on state, and so-called two-phase modulation may be employed.

[0072] The corrected voltage command values vu′, vv′, vw′ generated by the corrected voltage generator 26 through the above calculations are input to the PWM signal generator 27. Then, the PWM signal generator 27 performs PWM modulation on the corrected voltage command values vu′, vv′, vw′, and outputs the result as switching signals Gup to Gwn to the inverter 12.

[0073] FIG. 7 is a Bode diagram showing the transfer characteristics from the q-axis current command value iq_ref to the q-axis detected current value iq in the control device 1 according to Embodiment 1 when the rotational angular velocity ω is high. In FIG. 7, the gain diagram is shown on the upper side and the phase diagram is shown on the lower side. In both the gain diagram and the phase diagram, "the present disclosure" shows the transfer characteristics according to Embodiment 1, and "other-axis integral input (Bid, Biq) = 0" is the characteristic when 0 (zero) is forcibly input as the values of Bid and Biq (see FIGS. 3 and 4).

[0074] As shown in the gain diagram of FIG. 7, "other-axis integral input (Bid, Biq) = 0" has a gain drop near 100 Hz. In contrast, in "the present disclosure", there is no drop near 100 Hz, and the transfer characteristics of an ideal first-order lag system are obtained. That is, according to the control device 1 according to the present disclosure, it is possible to obtain a suitable control result even when the rotational angular velocity ω is high.

[0075] As shown in FIG. 1, the rotational angular velocity ω is input to the absolute value calculator 31. The absolute value calculator 31 calculates the absolute value of the rotational angular velocity ωabs based on the rotational angular velocity ω, and outputs the result to the change amount calculator 30. The change amount calculator 30 calculates the change amount Δωabs of the absolute value of the angular velocity based on the absolute value of the rotational angular velocity ωabs. As a method for calculating the change amount Δωabs, for example, a high-pass filter may be applied to the absolute value of the rotational angular velocity ωabs. Alternatively, the change amount Δωabs may be calculated by taking the difference between the current value and the previous value of the absolute value of the rotational angular velocity ωabs.

[0076] The d-axis transient amplifier 120d and the q-axis transient amplifier 120q of the voltage command value calculation unit 24 vary the values of the correction gains Ktrd and Ktrq as shown in FIG. 8 according to the change amount Δωabs of the absolute value of the angular velocity. The correction gains Ktrd and Ktrq are switched between the first correction value Ktr_H and the second correction value Ktr_L. The second correction value Ktr_L is smaller than the first correction value Ktr_H. In FIG. 8, Δω1 is the first threshold value of the angular velocity change amount, and Δω2 is the second threshold value of the angular velocity change amount. The first threshold value Δω1 of the angular velocity change amount is smaller than 0. The second threshold value Δω2 of the angular velocity change amount is larger than the first threshold value Δω1 of the angular velocity change amount.

[0077] When Δωabs is greater than or equal to Δω2, the correction gains Ktrd and Ktrq are set to the second correction value Ktr_L. When the correction gains Ktrd and Ktrq are set to a low value (Ktr_L), the influence of the noise included in the detected current values id and iq is reduced. When Δωabs is less than or equal to Δω1, it means that the rotational angular velocity ω has decreased rapidly, so the correction gains Ktrd and Ktrq are set to a high value (Ktr_H). Also, when Δωabs is between Δω1 and Δω2, the values of the correction gains Ktrd and Ktrq are continuously switched along the slope in the range of Ktr_H to Ktr_L. Thereby, the shock when the set values of the correction gains Ktrd and Ktrq are switched can be reduced. However, the above slope may not be provided, or Δω1 = Δω2 may be set.

[0078] Making the values of the correction gains Ktrd and Ktrq greater than 1 is equivalent to making the values of the proportional gains Kpd and Kpq greater than 1. For example, Ktr_L may be set to 1 and Ktr_H may be set to a value greater than 1. In the example of FIG. 8, the d-axis correction gain Ktrd and the q-axis correction gain Ktrq have the same set value. However, they may have different set values. Specifically, one or more of Ktr_H, Ktr_L, Δω1, and Δω2 may be set to different values for the d-axis and the q-axis.

[0079] Thus, when the rotational angular velocity ω rapidly decreases, the voltage command value calculation unit 24 in the first embodiment can quickly decrease the output voltage of the inverter 12 because it sets the correction gains Ktrd and Ktrq to high values (Ktr_H).

[0080] In addition, the cross-axis integral output amplifiers 123d and 123q of the voltage command value calculation unit 24 vary the values of the integral output switching gains Koutd and Koutq as shown in FIG. 9 according to the value of Δωabs calculated by the change amount calculator 30. When Δωabs is greater than or equal to Δω2, the values of Koutd and Koutq are set to 1 to enable interference compensation. When Δωabs is less than or equal to Δω1, it means that the rotational angular velocity ω has rapidly decreased, so the values of Koutd and Koutq are set to 0 to disable interference compensation. When Δωabs is between Δω1 and Δω2, the values of Koutd and Koutq are continuously switched along the slope within the range of 0 to 1. This can reduce the shock when the set values of Koutd and Koutq are switched. However, the slope may not be necessary, or Δω1 = Δω2 may be used.

[0081] Thus, when the rotational angular velocity ω rapidly decreases, the voltage command value calculation unit 24 in the first embodiment can quickly decrease the output voltage of the inverter 12 because it makes the interference compensation term zero (i.e., disables interference compensation).

[0082] In the voltage command value calculation unit 24 of the first embodiment, the other-axis integration input amplifier 121d of the d-axis varies the value of the integration input switching gain Kind according to the d-axis current deviation ed, as shown in FIG. 10. When the d-axis current deviation ed is less than or equal to the threshold value Ith, the value of Kind is set to 1, and the integration operation for interference compensation is performed by the integrator 122d. When the d-axis current deviation ed is greater than the threshold value Ith, it means that a sufficient negative d-axis current is flowing. Therefore, the value of Kind is set to 0, and the integration for interference compensation in the integrator 122d is substantially stopped. Since Kind is multiplied by the value input to the integrator 122d, shock is less likely to occur even without providing a slope for reducing shock associated with the switching of the set value of Kind. For this reason, no slope is provided in this embodiment. However, a slope may be provided so that the value of Kind gradually changes between 0 and 1. The threshold value Ith is a value of 0 or more. As a specific value of the threshold value Ith, for example, when the q-axis current command value iq_ref is changed in a sinusoidal shape with the assumed maximum amplitude, it is a value larger than the amplitude of the oscillation of the generated d-axis current. The value of the threshold value Ith may be 0, and in this case, the effect of suppressing excessive integration is most obtained.

[0083] In this way, the voltage command value calculation unit 24 in the first embodiment stops the integration of interference compensation when the d-axis current deviation ed is greater than the threshold value Ith. With this configuration, it is possible to suppress the output voltage of the inverter 12 from becoming excessively large when the rotational angular velocity ω suddenly decreases.

[0084] First modification example of the first embodiment. In the first embodiment, consideration is given to suppressing an overcurrent state caused by a sudden decrease in the rotational angular velocity ω of the AC rotating machine 10. Here, even when the armature currents iu, iv, iw of the AC rotating machine 10 suddenly increase, or when the current command values id_ref, iq_ref suddenly decrease, a similar overcurrent state may be caused. Therefore, in the first modification example, the control when the armature currents iu, iv, iw suddenly increase, or when the current command values id_ref, iq_ref suddenly decrease will be described. Hereinafter, the description of the parts overlapping with those in Embodiment 1 will be omitted, and the description will focus on the different points.

[0085] As shown in FIG. 11, the control device 1 according to the first modification includes a controller 13A. The configuration of the controller 13A is different from that of the controller 13 according to Embodiment 1 in that it includes a change amount calculator 30A instead of the change amount calculator 30, and includes an absolute value calculator 32, an absolute value calculator 33, a subtractor 34, and a limiter 35 instead of the absolute value calculator 31. Further, as shown in FIGS. 12 and 13, the processing in the d-axis voltage command value calculator 24d and the q-axis voltage command value calculator 24q is different from the processing (FIGS. 3 and 4) according to Embodiment 1.

[0086] In FIG. 11, the absolute value calculator 32 calculates a current command absolute value Irefabs based on the d-axis and q-axis current command values id_ref and iq_ref. The specific calculation performed by the absolute value calculator 32 is as follows. Irefabs = √(id_ref 2 + iq_ref 2 )

[0087] The absolute value calculator 33 calculates a current absolute value Iabs based on the d-axis and q-axis detected current values id and iq. The specific calculation performed by the absolute value calculator 33 is as follows. Iabs = √(id 2 +iq 2 ) Note that the current absolute value Iabs may be calculated using the pre-conversion detected currents ius, ivs, and iws instead of the detected current values id and iq.

[0088] The subtractor 34 subtracts the current command absolute value Irefabs from the current absolute value Iabs to calculate a current excess amount Iov0. That is, Iov0 = Iabs - Irefabs. The limiter 35 limits the lower limit of the overcurrent amount Iov0 to 0 (zero) and calculates the overcurrent amount Iov after the limitation. That is, when the value of the overcurrent amount Iov0 is zero or more, the limiter 35 sets Iov = Iov0, and when the value of the overcurrent amount Iov0 is less than zero, the limiter 35 sets Iov = 0. According to such processing of the limiter 35, the overcurrent amount Iov after the limitation changes only when the absolute value of the current Iabs is greater than the absolute value of the current command Irefabs.

[0089] The change amount calculator 30A calculates the change amount ΔIov of the overcurrent amount based on the overcurrent amount Iov after the limitation. The change amount ΔIov of the overcurrent amount can be calculated, for example, by applying a high-pass filter to the overcurrent amount Iov after the limitation. Alternatively, ΔIov may be calculated by taking the difference between the current value and the previous value of the overcurrent amount Iov after the limitation. The change amount ΔIov of the overcurrent amount is calculated based on the result of subtracting the absolute value of the current command Irefabs from the absolute value of the current Iabs. Therefore, when the absolute value of the current Iabs rapidly increases or the absolute value of the current command Irefabs rapidly decreases, the change amount ΔIov of the overcurrent amount becomes large.

[0090] As shown in FIG. 12, in the d-axis voltage command value calculator 24d according to the first modification example, the d-axis transient amplifier 120d multiplies the d-axis current deviation ed by Ktrd to calculate the corrected d-axis current deviation ed'. The correction gain Ktrd in the first modification example is a value that changes based on the change amount ΔIov of the overcurrent amount, and the details will be described later.

[0091] The other-axis integral output amplifier 123d according to the first modification example multiplies the other-axis integral output Vdi2 by Koutd to calculate the corrected other-axis integral output Vdi2'. The value of the integral output switching gain Koutd in the first modification example changes based on the change amount ΔIov of the overcurrent amount. Other operations in the d-axis voltage command value calculator 24d are the same as those in the first embodiment.

[0092] As shown in Fig. 13, in the q-axis voltage command value calculator 24q according to the first modification example, the q-axis transient amplifier 120q multiplies the q-axis current deviation eq by Ktrq to calculate a corrected q-axis current deviation eq'. The value of the correction gain Ktrq in the first modification example changes based on the change amount ΔIov of the current excess amount.

[0093] The other-axis integral output amplifier 123q according to the first modification example multiplies the other-axis integral output Vqi2 by Koutq to calculate a corrected other-axis integral output Vqi2'. The value of the integral output switching gain Koutq in the first modification example changes based on the change amount ΔIov of the current excess amount. Other calculations in the q-axis voltage command value calculator 24q are the same as those in the first embodiment.

[0094] In the first modification example, the d-axis transient amplifier 120d and the q-axis transient amplifier 120q vary the d-axis and q-axis correction gains Ktrd and Ktrq as shown in Fig. 14 in response to the change amount ΔIov of the current excess amount calculated by the change amount calculator 30A. In Fig. 14, "ΔI1" is the first threshold value of the excess current change amount, and "ΔI2" is the second threshold value of the excess current change amount. The first threshold value ΔI1 of the excess current change amount is smaller than the second threshold value ΔI2 of the excess current change amount.

[0095] As shown in Fig. 14, when ΔIov is less than or equal to ΔI1, by setting the correction gains Ktrd and Ktrq to a low value Ktr_L, the influence of noise included in the detected current values id and iq is reduced. When ΔIov is greater than or equal to ΔI2, it means that the rotating machine currents iu, iv, and iw have increased rapidly, or the current command values id_ref and iq_ref have decreased rapidly. Therefore, in this case, the correction gains Ktrd and Ktrq are increased to a high value Ktr_H. Also, when ΔIov is between ΔI1 and ΔI2, the values of the correction gains Ktrd and Ktrq are continuously switched along the slope in the range of Ktr_H to Ktr_L. Thereby, the shock when the set values of the correction gains Ktrd and Ktrq are switched can be reduced. However, the above slope may not be present, or ΔI1 = ΔI2 may be set.

[0096] In the example of FIG. 14, the correction gain Ktrd for the d-axis and the correction gain Ktrq for the q-axis have the same set value. However, they may have different set values. Specifically, one or more of Ktr_H, Ktr_L, ΔI1, and ΔI2 may be set to different values for the d-axis and the q-axis.

[0097] As described above, when the rotational machine currents iu, iv, iw rapidly increase, the voltage command value calculation unit 24 in the first modification example sets the correction gains Ktrd and Ktrq to high values (Ktr_H), so that the output voltage of the inverter 12 can be rapidly decreased.

[0098] Also, the other-axis integral output amplifiers 123d and 123q for the d-axis and the q-axis in the first modification example vary the values of the integral output switching gains Koutd and Koutq as shown in FIG. 15 according to the value of ΔIov calculated by the change amount calculator 30A. When ΔIov is less than or equal to ΔI1, the values of Koutd and Koutq are set to 1 to enable interference compensation. When ΔIov is greater than or equal to ΔI2, it means that the rotational machine currents iu, iv, iw have rapidly increased with respect to the target value, or the current command values id_ref and iq_ref have rapidly decreased. Therefore, in this case, the values of Koutd and Koutq are set to 0 to disable interference compensation. When ΔIov is between ΔI1 and ΔI2, the values of Koutd and Koutq are continuously switched along the slope in the range of 0 to 1. Thereby, the influence of the shock when the set values of Koutd and Koutq are switched can be reduced. However, the above slope may not be provided, or ΔI1 = ΔI2 may be used.

[0099] As described above, when the rotational machine currents iu, iv, iw rapidly increase with respect to the target value, the voltage command value calculation unit 24 in the first modification example sets the interference compensation term to zero (that is, disables interference compensation), so that the output voltage of the inverter 12 can be rapidly decreased.

[0100] Second modification example of Embodiment 1. In Embodiment 1 and the first modification example, the case where internal factors (such as the rotational angular velocity ω, the rotational machine currents iu, iv, iw, etc.) of the rotating machine 10 change rapidly was considered. Here, even when external factors of the rotating machine 10 change rapidly, there is a possibility of causing a similar overcurrent state. Therefore, in the second modification example, as an example of external factors of an AC rotating machine, control when the DC bus voltage Vdc changes rapidly will be described. For example, when the DC bus voltage Vdc suddenly increases, even if there is no change in the corrected voltage commands vu', vv', vw', the output voltage of the inverter 12 may become excessive, so there is a possibility of an overcurrent state. Hereinafter, descriptions of parts overlapping with Embodiment 1 will be omitted, and the different points will be mainly described.

[0101] As shown in FIG. 16, the control device 1 according to the second modification example includes a controller 13B. The configuration of the controller 13B is different from that of the controller 13 according to Embodiment 1 in that it includes a change amount calculator 36 instead of an absolute value calculator 31 and a change amount calculator 30. Also, as shown in FIGS. 17 and 18, the processing in the d-axis voltage command value calculator 24d and the q-axis voltage command value calculator 24q is different from the processing according to Embodiment 1 (FIGS. 3 and 4).

[0102] In FIG. 16, the change amount calculator 36 calculates the change amount ΔVdc of the DC bus voltage based on the DC bus voltage Vdc. The change amount ΔVdc of the DC bus voltage can be calculated, for example, by applying a high-pass filter to the DC bus voltage Vdc. Also, ΔVdc may be calculated by taking the difference between the current value and the previous value of the DC bus voltage Vdc.

[0103] As shown in FIG. 17, in the d-axis voltage command value calculator 24d according to the second modification example, the d-axis transient amplifier 120d multiplies the d-axis current deviation ed by Ktrd to calculate the corrected d-axis current deviation ed'. The correction gain Ktrd in the second modification example is a value that changes based on the change amount ΔVdc of the DC bus voltage, and the details will be described later.

[0104] In the other-axis integral output amplifier 123d according to the second modification example, the other-axis integral output Vdi2 is multiplied by Koutd to calculate the corrected other-axis integral output Vdi2'. The value of the integral output switching gain Koutd according to the second modification example changes based on the change amount ΔVdc of the DC bus voltage. Other calculations in the d-axis voltage command value calculator 24d are the same as those in the first embodiment.

[0105] As shown in FIG. 18, in the q-axis voltage command value calculator 24q according to the second modification example, the q-axis transient amplifier 120q multiplies the q-axis current deviation eq by Ktrq to calculate the corrected q-axis current deviation eq'. The correction gain Ktrq in the second modification example is a value that changes based on the change amount ΔVdc of the DC bus voltage, and the details will be described later.

[0106] The other-axis integral output amplifier 123q according to the second modification example multiplies the other-axis integral output Vqi2 by Koutq to calculate the corrected other-axis integral output Vqi2'. The value of the integral output switching gain Koutq in the second modification example changes based on the change amount ΔVdc of the DC bus voltage. Other calculations in the q-axis voltage command value calculator 24q are the same as those in the first embodiment.

[0107] In the second modification example, the d-axis transient amplifier 120d and the q-axis transient amplifier 120q vary the d-axis and q-axis correction gains Ktrd and Ktrq as shown in FIG. 19 according to the change amount ΔVdc of the DC bus voltage calculated by the change amount calculator 36. In FIG. 19, "ΔV1" is the first voltage change amount threshold value, and "ΔV2" is the second voltage change amount threshold value. The first voltage change amount threshold value ΔV1 is smaller than the second voltage change amount threshold value ΔV2.

[0108] As shown in Fig. 19, when ΔVdc is less than or equal to ΔV1, by setting the correction gains Ktrd and Ktrq to a low value Ktr_L, the influence of the noise included in the detected current values id and iq can be reduced. When ΔVdc is greater than or equal to ΔV2, it means that the DC bus voltage Vdc has increased sharply, so Ktrd and Ktrq are increased to a high value (Ktr_H). Also, when ΔVdc is between ΔV1 and ΔV2, the values of the correction gains Ktrd and Ktrq are continuously switched along the slope within the range of Ktr_H to Ktr_L. Thereby, the shock when the set values of the correction gains Ktrd and Ktrq are switched can be reduced. However, the above slope may not be necessary, or ΔV1 = ΔV2 may be set.

[0109] Note that in the example of Fig. 19, the same set value is used for the d-axis correction gain Ktrd and the q-axis correction gain Ktrq. However, different set values may be used for the d-axis correction gain Ktrd and the q-axis correction gain Ktrq. Specifically, one or more of Ktr_H, Ktr_L, ΔV1, and ΔV2 may be set to different values for the d-axis and the q-axis.

[0110] Thus, since the voltage command value calculation unit 24 in the second modification example sets the correction gains Ktrd and Ktrq to a high value (Ktr_H) when the DC bus voltage Vdc increases sharply, it is possible to quickly reduce the output voltage of the inverter 12.

[0111] Also, for the cross-axis integration output amplifiers 123d and 123q in the second modification example, the values of the integration output switching gains Koutd and Koutq are varied as shown in Fig. 20 according to the value of ΔVdc calculated by the change amount calculator 36. When ΔVdc is less than or equal to ΔV1, the values of Koutd and Koutq are set to 1 to enable interference compensation. When ΔVdc is greater than or equal to ΔV2, it means that the DC bus voltage Vdc has increased sharply, so the values of Koutd and Koutq are set to 0 to disable interference compensation. When ΔVdc is between ΔV1 and ΔV2, the values of Koutd and Koutq are continuously switched along the slope in the range of 0 to 1. Thereby, the influence of the shock when the set values of Koutd and Koutq are switched can be reduced. However, the above slope may not be necessary, or ΔV1 = ΔV2 may be used.

[0112] Thus, since the voltage command value calculation unit 24 in the second modification example makes the interference compensation term zero when the DC bus voltage Vdc increases sharply (that is, disables interference compensation by setting the set values of the integration output switching gains Koutd and Koutq to 0), it is possible to quickly reduce the output voltage of the inverter 12.

[0113] In the second modification example, the variation of the DC bus voltage Vdc has been described as an example of an external factor of the AC rotating machine 10. However, the above description may be applied to other external factors (for example, variations in the load torque of the rotating machine 10, etc.) for control.

[0114] As described above, the control device 1 according to the first embodiment, the first modification, and the second modification includes an inverter 12 that applies a voltage to the AC rotating machine 10, a DC power supply BT that supplies DC power to the inverter 12, a current detector 22 that detects the rotating machine currents iu, iv, and iw flowing through the rotating machine 10, a d-axis voltage command value calculator 24d (first-axis voltage command value calculator) that calculates a voltage command value vd for the d-axis (first axis) on the two rotating axes of the rotating machine 10, and a q-axis voltage command value calculator 24q (second-axis voltage command value calculator) that calculates a voltage command value vq for the q-axis (second axis) on the two rotating axes. The d-axis voltage command value calculator 24d multiplies a corrected d-axis current deviation ed' (corrected first deviation), which is the deviation between the current command value id_ref on the d-axis and the detected current value id on the d-axis multiplied by a correction gain Ktrd, the current command value iq_ref on the q-axis, the q-axis current deviation eq (second deviation), which is the deviation between the q-axis detected current value iq on the q-axis, and the rotational angular velocity ω of the AC rotating machine 10 to calculate the voltage command value vd for the d-axis. The set value of the correction gain Ktrd is increased to a first correction value Ktr_H when a physical quantity in the AC rotating machine 10 changes rapidly or when the DC bus voltage Vdc (output voltage) in the DC power supply BT increases. Also, the correction gain Ktrq for the q-axis may be set in the same manner.

[0115] By setting the values of the correction gains Ktrd and Ktrq in this way, in a transient state (for example, when the rotational angular velocity ω rapidly decreases, when the rotating machine currents iu, iv, iw rapidly increase, when the DC bus voltage Vdc rapidly increases, etc.), the output voltage of the inverter 12 can be rapidly decreased. Therefore, the overcurrent state of the rotating machine 10 can be appropriately suppressed. Also, by increasing the correction gain Ktrd only in the transient state, that is, by setting the correction gains Ktrd and Ktrq to low values (second correction value Ktr_L) during normal operation, the deterioration of the motor driving noise and torque ripple due to the influence of noise included in the detection results of the rotating machine currents iu, iv, iw can be suppressed.

[0116] Further, when the physical quantity of the AC rotating machine 10 changes suddenly or the change amount ΔVdc of the output voltage increases, the voltage command value calculators 24d and 24q may set the correction gains Ktrd and Ktrq to the first correction value Ktr_H and set the interference compensation term to zero. More specifically, the values of Koutd and Koutq may be set to 0 to disable interference compensation. With this configuration, it is possible to make the excess amount of current supplied to the AC rotating machine 10 smaller in the transient state.

[0117] Further, the d-axis voltage command value calculator 24d may perform an integration operation using a value obtained by multiplying the q-axis current deviation eq by the rotational angular velocity ω of the AC rotating machine, the q-axis inductance Lq, and the response angular frequency ωcc for adjusting the frequency response of the rotating machine current with respect to the current command value, and calculate an interference compensation term based on the result of the integration operation. Thereby, a desired frequency response can be obtained without depending on the rotational angular velocity ω of the rotating machine. Similarly, the q-axis voltage command value calculator 24q may perform an integration operation using a value obtained by multiplying the d-axis current deviation ed by the rotational angular velocity ω of the AC rotating machine, the inductance Ld on the d-axis, and the response angular frequency ωcc for adjusting the frequency response of the rotating machine current with respect to the current command value, and calculate an interference compensation term based on the result.

[0118] Further, when the d-axis current deviation ed is greater than the threshold value Ith, the d-axis voltage command value calculator 24d may set the input value to the integration operation for calculating the interference compensation term to zero. More specifically, as shown in FIG. 10, the value of the integration input switching gain Kind may be set to 0. With this configuration, it is possible to suppress the output voltage of the inverter 12 from becoming excessively large when the rotational angular velocity ω suddenly decreases.

[0119] Further, the physical quantity may be the rotational angular velocity ω of the AC rotating machine 10, and when the rotational angular velocity ω suddenly decreases, the voltage command value calculators 24d and 24q may increase the set values of the correction gains Ktrd and Ktrq to the first correction value Ktr_H. With this configuration, when the rotational angular velocity ω suddenly decreases, the output voltage of the inverter 12 can be quickly decreased to appropriately suppress the overcurrent state of the rotating machine 10.

[0120] Also, the physical quantity may be the rotational machine currents iu, iv, iw, and when the rotational machine currents iu, iv, iw rapidly increase with respect to the target values, the voltage command value calculators 24d and 24q may increase the set values of the correction gains Ktrd and Ktrq to obtain a first correction value Ktr_H. With this configuration, when the rotational machine currents iu, iv, iw rapidly increase, the output voltage of the inverter 12 can be rapidly decreased, and the overcurrent state of the rotational machine 10 can be appropriately suppressed.

[0121] Also, the physical quantity may be the current command values id_ref and iq_ref, and when the current command values id_ref and iq_ref rapidly decrease, the voltage command value calculators 24d and 24q may increase the set values of the correction gains Ktrd and Ktrq to obtain a first correction value Ktr_H. With this configuration, when the current command values id_ref and iq_ref rapidly decrease, the output voltage of the inverter 12 can be rapidly decreased, and the overcurrent state of the rotational machine 10 can be appropriately suppressed.

[0122] Also, the physical quantity may be an external factor (e.g., load torque) of the AC rotational machine 10 excluding the DC bus voltage Vdc, and when the external factor rapidly changes, the voltage command value calculators 24d and 24q may increase the set values of the correction gains Ktrd and Ktrq to obtain a first correction value Ktr_H. With this configuration, when the external factor rapidly changes, the output voltage of the inverter 12 can be rapidly decreased, and the overcurrent state of the rotational machine 10 can be appropriately suppressed.

[0123] Embodiment 2. Next, Embodiment 2 of the present disclosure will be described. The basic configuration is the same as that of Embodiment 1. Therefore, the description of the parts overlapping with Embodiment 1 will be omitted, and the description will focus on the differences. FIG. 21 is a block diagram showing a schematic configuration of a control device 2 for a rotating machine according to Embodiment 2. As shown in FIG. 21, the control device 2 is different from the control device 1 according to Embodiment 1 in that it includes an amplitude calculator 29. FIGS. 22 and 23 are block diagrams showing the configurations of a d-axis voltage command value calculator 24d and a q-axis voltage command value calculator 24q according to Embodiment 2, respectively. The configurations of the d-axis voltage command value calculator 24d and the q-axis voltage command value calculator 24q also differ between Embodiment 1 and Embodiment 2.

[0124] As shown in FIG. 21, the amplitude calculator 29 receives a d-axis voltage command value vd, a q-axis voltage command value vq, and a DC bus voltage Vdc. The amplitude calculator 29 calculates a voltage amplitude m based on the following equation (2-1). The voltage amplitude m is a value obtained by normalizing the magnitudes of the d-axis voltage command value vd and the q-axis voltage command value vq using the DC bus voltage Vdc of the inverter 12. m = {(vd 2 + vq 2 ) / (Vdc / 2)} 0.5 …(2-1)

[0125] As shown in FIG. 22, the d-axis voltage command value calculator 24d includes a limiting amplifier 112d. The limiting amplifier 112d inputs, as a secondary corrected cross-axis integral input Bid”, the result obtained by multiplying the corrected cross-axis integral input Bid’ output from the cross-axis integral input amplifier 121d by K to the integrator 122d. “K” is a limiting gain that is multiplied when integrating the current deviation eq on the q-axis, which is not the control target, as seen from the d-axis voltage command value calculator 24d. The voltage amplitude m is input to the limiting amplifier 112d, and the value of the limiting gain K is determined based on the value of the voltage amplitude m. The integrator 122d inputs the secondary corrected cross-axis integral input Bid” and performs an integration operation, and outputs the result as a cross-axis integral output Vdi2. In the d-axis voltage command value calculator 24d according to the present embodiment, the voltage amplitude m is also input to the d-axis proportional amplifier 101d. The present embodiment is different from Embodiment 1 in that the proportional gain Kpd and the limiting gain K on the d-axis are varied according to the voltage amplitude m.

[0126] As shown in Fig. 23, the q-axis voltage command value calculator 24q of Embodiment 2 includes a limiting amplifier 112q. The limiting amplifier 112q inputs the result of multiplying the cross-axis integral input Biq output from the multiplier 107q by K as the corrected cross-axis integral input Biq” to the integrator 122q. “K” is a limiting gain that is multiplied when integrating the current deviation ed on the d-axis, which is not the control target, as seen from the q-axis voltage command value calculator 24q. A voltage amplitude m is input to the limiting amplifier 112q, and the value of the limiting gain K is determined based on the value of the voltage amplitude m. The integrator 122q inputs the corrected cross-axis integral input Biq” and performs an integration operation, and outputs the result as the cross-axis integral output Vqi2. In the q-axis voltage command value calculator 24q according to this embodiment, the voltage amplitude m is also input to the q-axis proportional amplifier 101q. This embodiment is different from Embodiment 1 in that the proportional gain Kpq and the limiting gain K of the q-axis are varied according to the voltage amplitude m.

[0127] Fig. 24 shows the relationship between the voltage amplitude m and the proportional gains Kpd, Kpq, and the limiting gain K. The vertical axis of the upper graph in Fig. 24 is the proportional gains Kpd, Kpq, and the vertical axis of the lower graph is the limiting gain K. In both the upper and lower graphs, the horizontal axis is the voltage amplitude m. In Fig. 24, m1 is the first voltage threshold and m2 is the second voltage threshold. The proportional gains Kpd, Kpq are switched between a first proportional value Kp_H and a second proportional value Kp_L. The second proportional value Kp_L is smaller than the first proportional value Kp_H. The limiting gain K is switched between a first limiting value K_H and a second limiting value K_L. The second limiting value K_L is smaller than the first limiting value K_H.

[0128] As shown in Fig. 24, when the voltage amplitude m is higher than the first voltage threshold m1, the values of the proportional gains Kpd, Kpq are set to the second proportional value Kp_L. Also, when the voltage amplitude m is lower than the first voltage threshold m1, the proportional gains Kpd, Kpq are set to the first proportional value Kp_H. When the voltage amplitude m is higher than the second voltage threshold value m2, the limiting gain K is set to the first limiting value K_H. Also, when the voltage amplitude m is lower than the second voltage threshold value m2, the limiting gain K is set to the second limiting value K_L. Here, the first voltage threshold value m1 is higher than the second voltage threshold value m2. Therefore, when the values of the proportional gains Kpd and Kpq are Kp_L, the value of the limiting gain K is K_H.

[0129] In this way, only when the limiting gain K is at the high value K_H (i.e., in a state where the non-interference control between the d-q axes is sufficiently effective), by processing to lower the proportional gains Kpd and Kpq, the stability of the control can be ensured even when the proportional gains Kpd and Kpq are low. Also, when the limiting gain K is at the low value K_L, the proportional gains Kpd and Kpq are set to a high value Kp_H that can ensure stability so that the stability of the control can be ensured even in a state where the non-interference control between the d-q axes is not sufficient. Specifically, the values of Kp_H and Kp_L are set so that the responses from the current command values id_ref and iq_ref to the rotating machine currents iu, iv, and iw are good within the ranges of 300 Hz to 1000 Hz and 100 Hz to 300 Hz, respectively. For example, the value of K_H is set to 0.7 or more, and the value of K_L is set to 0.3 or less.

[0130] Next, the preferred setting of the first voltage threshold m1 will be described. Gxn shown in FIG. 25 shows an example of the waveform of any one of the switching signals Gun, Gvn, and Gwn in the lower arm switching elements Sun, Svn, and Swn. Further, VRx shows an example of the waveform of any one of the voltages VRu, VRv, and VRw across the shunt resistors Ru, Rv, and Rw. In the example shown in FIG. 25, after the signal Gxn changes from 0 to 1, ringing occurs in the voltage VRx across the shunt resistor for several μs. Ringing is a phenomenon in which the voltage across the shunt resistor fluctuates for a certain period of time when switching occurs in the inverter 12. If the current detector 22 acquires the pre-conversion detected currents ius, ivs, and iws based on the voltages VRu, VRv, and VRw across the shunt resistors including this ringing, the detection result will include an error. If the pre-conversion detected currents ius, ivs, and iws include an error, the detected current values id and iq after coordinate conversion will also include an error.

[0131] In order to accurately obtain the pre-conversion detected currents ius, ivs, and iws, it is preferable that the on-time of the corresponding lower arm switching elements Sun, Svn, and Swn is longer than the time threshold Tmin set according to the convergence time of the ringing. In order to make the on-time of the lower arm switching elements Sun, Svn, and Swn longer than the time threshold Tmin, the voltage input to the PWM signal generator 27 may be equal to or lower than Vdc×(Tc - Tmin) / Tc shown by the dashed-dotted line in FIG. 2.

[0132] In FIG. 2, the corrected voltage commands vu’, vv’, and vw’ (that is, the voltages input to the PWM signal generator 27) are all equal to or lower than Vdc×(Tc - Tmin) / Tc. Therefore, it is possible to eliminate the error due to ringing and accurately acquire the pre-conversion detected currents ius, ivs, and iws based on the voltages VRu, VRv, and VRw across the shunt resistors. In this way, the detection method for obtaining the pre-conversion detected currents ius, ivs, and iws based on the voltages VRu, VRv, and VRw across the shunt resistors corresponding to the three phases respectively is called "three-phase detection".

[0133] Incidentally, when the voltage amplitude m increases due to an increase in the rotational speed of the rotating machine 10 or the like, a part of the corrected voltage commands vu’, vv’, vw’ may take a large value close to the maximum value Vdc of the carrier triangular wave C. In the example of FIG. 26, the value of the corrected voltage command vu’ is larger than the value of Vdc×(Tc - Tmin) / Tc (hereinafter also referred to as the “upper limit value”). In this case, the difference between the time when Gun switches from 0 to 1 and the time at timing X becomes small. Therefore, the ringing effect is included in the voltage VRu across the shunt resistor of the u-phase obtained at timing X.

[0134] Therefore, the pre-conversion detected current for the phase in which the on-times of the switching signals Gun, Gvn, Gwn of the lower-arm switching elements are shorter than the time threshold Tmin may be generated from the other two phases. In this way, the detection method of obtaining one of the pre-conversion detected currents of the three phases based on the pre-conversion detected currents of the remaining two phases is called “two-phase detection”. For example, when the on-time of the switching signal Gun is shorter than the time threshold Tmin, the pre-conversion detected current ius of the U-phase may be calculated by ius = -ivs - iws. Similarly, when the on-time of the switching signal Gvn is shorter than the time threshold Tmin, it may be calculated as ivs = -ius - iws, and when the on-time of the switching signal Gwn is shorter than the time threshold Tmin, it may be calculated as iws = -ius - ivs.

[0135] The relationship between the number of phases used for current detection and the voltage amplitude m is as follows. That is, when the voltage amplitude m is low and the instantaneous values of the three-phase modified voltage commands vu’, vv’, and vw’ are all below the upper limit value (Vdc×(Tc - Tmin) / Tc), "three-phase detection" is used. Alternatively, when the voltage amplitude m is high and the instantaneous value of any one of the three-phase modified voltage commands vu’, vv’, and vw’ is above the upper limit value (Vdc×(Tc - Tmin) / Tc), "two-phase detection" is used. Here, when comparing "three-phase detection" and "two-phase detection", "three-phase detection" has better accuracy. Therefore, in order to improve the control accuracy of the rotating machine 10, it is preferable to use "three-phase detection" as much as possible. However, from the perspective of increasing the output of the rotating machine 10, it may be preferable to increase the voltage amplitude m and use "two-phase detection".

[0136] Therefore, in this embodiment, the value of the first voltage threshold m1 is set to "(Tc - Tmin) / Tc". It can be said that "(Tc - Tmin) / Tc" is a numerical value obtained by normalizing the above-mentioned upper limit value "Vdc×(Tc - Tmin) / Tc" by dividing it by Vdc. By setting the first voltage threshold m1 in this way, the following effects can be obtained. When the voltage amplitude m is greater than the first voltage threshold m1, "two-phase detection" is used, but at this time, the proportional gains Kpd and Kpq are set to a low value Kp_L. This can reduce the influence of the decrease in current detection accuracy and prevent the vibration and noise generated by the rotating machine 10 from increasing. At the same time, since the limit gain K is set to a high value K_H, non-interference control is effective and the control stability is maintained. Also, when the voltage amplitude m is less than the first voltage threshold m1, "three-phase detection" is used, but at this time, the proportional gains Kpd and Kpq are set to a high value Kp_H (see Fig. 24). Since the current detection accuracy is good in three-phase detection, the dependence on non-interference control is reduced. Thereby, the vibration and noise generated in the rotating machine 10 due to the noise included in the rotational angular velocity ω can be reduced.

[0137] The second voltage threshold value m2 may be set to a value lower than the first voltage threshold value m1. However, if the noise included in the detected value of the rotational angular velocity ω does not pose a problem, the switching of the value of the limit gain K at the second voltage threshold value m2 is not necessarily required. Specifically, if the voltage amplitude m is equal to or greater than the first voltage threshold value m1, the values of the proportional gains Kpd and Kpq are set to Kp_L, and if not, the values of the proportional gains Kpd and Kpq are set to Kp_H, and the value of the limit gain K may be set to be constant (K_H) regardless of the voltage amplitude m.

[0138] Also, generally including the AC rotating electrical machine 10 according to the present embodiment, the applied voltage to the rotating electrical machine and the rotational speed (or rotational angular velocity ω) of the rotating electrical machine are generally in a proportional relationship. Therefore, the switching of the proportional gains Kpd and Kpq and the limit gain K may be performed using thresholds related to the rotational speed (or rotational angular velocity ω) instead of the thresholds related to the voltage amplitude m described above. For example, the rotational speed when the voltage amplitude m matches the first voltage threshold value m1 may be set as the first rotational speed threshold value n1, and the rotational speed when the voltage amplitude m matches the second voltage threshold value m2 may be set as the second rotational speed threshold value n2. In this case, the switching of the proportional gains Kpd and Kpq or the limit gain K based on the comparison between the voltage amplitude m and the threshold values m1 and m2 described above can be replaced with switching based on the comparison between the rotational speed and the threshold values n1 and n2. Similarly, the rotational angular velocity ω when the voltage amplitude m matches the first voltage threshold value m1 may be set as the first rotational angular velocity threshold value ω1, and the rotational angular velocity ω when the voltage amplitude m matches the second voltage threshold value m2 may be set as the second rotational angular velocity threshold value ω2. In this case, the switching of the proportional gains Kpd and Kpq or the limit gain K based on the comparison between the voltage amplitude m and the threshold values m1 and m2 described above can be replaced with switching based on the comparison between the rotational angular velocity ω and ω1 and ω2.

[0139] Alternatively, switching according to the voltage amplitude m and switching according to the rotational speed (or rotational angular velocity ω) may be used in combination. As a specific example, the switching of the proportional gains Kpd and Kpq may be performed based on the comparison between the voltage amplitude m and the first voltage threshold value m1, and the switching of the limit gain K may be performed based on the comparison between the rotational speed (or rotational angular velocity ω) and the second rotational speed threshold value n2 (or the second rotational angular velocity threshold value ω2).

[0140] In the above description, it has been explained that the values of the proportional gains Kpd and Kpq are switched based on the comparison between the voltage amplitude m and the first voltage threshold value m1, or the comparison between the rotational speed and the first rotational speed threshold value n1, or the comparison between the rotational angular velocity ω and the first rotational angular velocity threshold value ω1. However, for example, when the rotational machine current becomes excessive and exceeds the threshold value, the values of the proportional gains Kpd and Kpq may be fixed at Kp_H regardless of the voltage amplitude m or the rotational speed. When the rotational machine current is excessive, by setting the proportional gains Kpd and Kpq to large values, the state in which the rotational machine current becomes excessive can be shortened. Thereby, a failure of the control device of the rotary machine 10 can be prevented.

[0141] Also, when a part of the control device of the rotary machine 10 fails and continuous operation is performed with the remaining non-failed parts, the proportional gains Kpd and Kpq may always be set to a large value (Kp_H) to enhance the safety of the system. Also, when switching the values of the proportional gains Kpd and Kpq between Kp_L and Kp_H, a slope (gradient) may be added and the switching may be gradually performed so that the control does not become discontinuous. Similarly, when switching the value of the limit gain K between K_L and K_H, a slope (gradient) may be added and the switching may be gradually performed.

[0142] As described above, in the second embodiment, the first-axis voltage command value calculator (for example, the d-axis voltage command value calculator 24d) calculates the voltage command value (for example, Vd) of the first axis using a value obtained by multiplying the corrected first deviation (for example, the corrected d-axis current deviation ed') by a proportional gain (for example, Kpd). The proportional gain is switched between a first proportional value Kp_H and a second proportional value Kp_L that is smaller than the first proportional value Kp_H. When the voltage amplitude m calculated based on the voltage command value is equal to or greater than the first voltage threshold m1, or when the rotational speed of the AC rotating machine 10 is equal to or greater than the first rotational speed threshold n1, or when the rotational angular velocity ω of the AC rotating machine 10 is equal to or greater than the first rotational angular velocity threshold ω1, the proportional gain is set to the second proportional value Kp_L. According to this configuration, when the rotational angular velocity ω is large, by setting the proportional gain to a small value (Kp_L), the high-frequency noise component included in the corrected first deviation can be reduced, and the control device 2 for the low-noise AC rotating machine 10 can be provided.

[0143] Also, in the second embodiment, the first-axis voltage command value calculator performs an integration operation using a value obtained by multiplying the second deviation by a limiting gain K, and the limiting gain K is switched between a first limiting value K_H and a second limiting value K_L that is smaller than the first limiting value K_H. When the voltage amplitude m is equal to or less than a second voltage threshold m2 that is smaller than the first voltage threshold m1, or when the rotational speed of the AC rotating machine 10 is equal to or less than a second rotational speed threshold n2 that is smaller than the first rotational speed threshold n1, or when the rotational angular velocity ω of the AC rotating machine 10 is equal to or less than a second rotational angular velocity threshold ω2 that is smaller than the first rotational angular velocity threshold ω1, the limiting gain K is set to the second limiting value K_L. According to this configuration, when the rotational angular velocity ω is small, by setting the limiting gain to a small value (K_L), the input of the second deviation to the integration operation can be limited to a small value. Therefore, the high-frequency noise component included in the second deviation can be reduced, and the control device 2 for the low-noise AC rotating machine 10 can be provided.

[0144] In addition, the inverter 12 includes three sets of upper-arm switching elements Sup, Svp, Swp, lower-arm switching elements Sun, Svn, Swn, and shunt resistors Ru, Rv, Rw, which respectively correspond to three phases (U, V, W). When the voltage amplitude m is smaller than the first voltage threshold value m1, the current detector 22 detects the rotational machine currents iu, iv, iw in the three phases based on the voltages VRu, VRv, VRw across the corresponding shunt resistors (that is, "three-phase detection" is used). According to this configuration, it is possible to suppress the detection results of the rotational machine currents iu, iv, iw from including errors due to the influence of ringing. Therefore, the AC rotating machine 10 can be accurately controlled.

[0145] Also, when the rotational machine current exceeds the threshold value (that is, when it is excessive), regardless of the voltage amplitude m, etc., the proportional gain Kpd, Kpq may be set to the first proportional value Kp_H. In this case, the time during which the state where the rotational machine current is excessive continues can be shortened. Therefore, a failure of the control device 2 can be prevented.

[0146] Also, in the control device 2, when a failure occurs at at least one location, regardless of the voltage amplitude m, etc., the proportional gain Kpd, Kpq may be set to the first proportional value Kp_H. In this case, the stability of the system at the time of failure occurrence can be improved, and the occurrence of a secondary failure can be suppressed.

[0147] Embodiment 3. Next, Embodiment 3 will be described. In this embodiment, the case where the technology described in Embodiment 1 or 2 is applied to the control of the rotating machine included in the electric power steering device will be described. As shown in Fig. 27, the electric power steering apparatus 100 according to this embodiment includes a control device 3, a steering wheel 101, a rotary machine 10, and a torque detector 103. The electric power steering apparatus 100 is mounted on a vehicle. The steering wheel 101 is operated by a driver. By operating the steering wheel 101, the front wheels 102 of the vehicle are driven. Since the basic configuration of the control device 3 is the same as that of the control device 1 in the first embodiment, detailed description thereof is omitted, and the different points will be mainly described.

[0148] The torque detector 103 detects the steering torque Ts of the steering wheel 101 by the driver and outputs the detection result to the control device 3. The driving force of the rotary machine 10 is transmitted to the steering system 100s of the vehicle via the driving force transmission mechanism 104. The steering system 100s includes the steering wheel 101, the front wheels 102, etc. The electric power steering apparatus 100 uses the driving force generated by the rotary machine 10 as an assist torque to assist the driver in steering the vehicle.

[0149] The current command value calculator 21 in the control device 3 differs from that in the first embodiment in the calculation of the torque current command value (q-axis current command value iq_ref), and thus this point will be described. The steering torque Ts and the vehicle running speed S are input to the current command value calculator 21, and based on these inputs, the q-axis current command value iq_ref is calculated.

[0150] Fig. 28 is a graph showing the set value of the q-axis current command value iq_ref according to the steering torque Ts and the vehicle running speed S. As shown in this graph, the larger the steering torque Ts, the larger the value of the current command value iq_ref and the steeper the gradient of its change. Also, the larger the vehicle running speed S, the smaller the value of the current command value iq_ref. Incidentally, the value of the current command value iq_ref may be determined by further considering the damping torque, etc. obtained based on the rotational angular velocity ω.

[0151] As shown in FIG. 28, the lower the driving speed S of the vehicle, the greater the gradient of the q-axis current command value iq_ref with respect to the steering torque Ts. As a result, when the detected noise is included in the steering torque Ts, the influence on the q-axis current command value iq_ref becomes large, leading to the generation of vibration and noise from the rotary machine 10 or an uncomfortable feeling of touching the steering wheel 101.

[0152] Therefore, in the voltage command value calculation unit 24 in the third embodiment, the proportional gain Kpd in the d-axis voltage command value calculator 24d and the proportional gain Kpq in the q-axis voltage command value calculator 24q are varied as shown in FIG. 29 according to the driving speed S of the vehicle. "Kp_H1" is referred to as the first speed reference value, and "Kp_L1" is referred to as the second speed reference value. The second speed reference value Kp_L1 is smaller than the first speed reference value Kp_H1.

[0153] When the driving speed S of the vehicle is s1 or less, the values of the proportional gains Kpd and Kpq are set to Kp_L1. In this way, by setting the values of the proportional gains Kpd and Kpq low, it is possible to suppress the q-axis current command value iq_ref from reacting sensitively to the pulsation of the steering torque Ts. Also, when the driving speed S of the vehicle is s2 or more, since the influence on the q-axis current command value iq_ref due to the pulsation of the steering torque Ts is small, the proportional gains Kpd and Kpq are set to a high value Kp_H1. In this way, by switching the proportional gains Kpd and Kpq according to the driving speed S of the vehicle, it is possible to provide an electric power steering device that suppresses the generation of vibration and abnormal noise of the rotary machine 10 caused by noise included in the steering torque detection value and the driver from feeling discomfort from the steering wheel 101.

[0154] As shown in FIG. 29, when the driving speed S of the vehicle is within the range of s1 to s2, the values of the proportional gains Kpd and Kpq are continuously changed within the range of Kp_H1 to Kp_L1. Thereby, it is possible to suppress a sense of discomfort felt by the driver due to the sudden switching of the set values of the proportional gains Kpd and Kpq. Note that the proportional gains Kpd and Kpq may be switched stepwise between Kp_H1 and Kp_L1.

[0155] As described above, the electric power steering apparatus 100 according to the present embodiment includes a control device 3, an AC rotary machine 10, and a driving force transmission mechanism 104 that transmits the driving force of the AC rotary machine 10 to the steering system 100s of the vehicle. With such a configuration, it is possible to provide an electric power steering apparatus 100 that achieves both quietness and steering stability.

[0156] Further, in the present embodiment, the proportional gains Kpd and Kpq are switched between a first speed reference value Kp_H1 and a second speed reference value Kp_L1 that is smaller than the first speed reference value Kp_H1. When the traveling speed S of the vehicle is smaller than the threshold value s2, the voltage command value calculation unit 24 sets the proportional gains Kpd and Kpq to values lower than the first speed reference value Kp_H1. According to this configuration, it is possible to enhance quietness in a situation where the vehicle speed is low and the steering sound is likely to be heard.

[0157] Embodiment 4 Next, Embodiment 4 will be described. As shown in FIG. 30, the electric power steering apparatus 100 according to the present embodiment includes the same control device 3 as that in Embodiment 3. Embodiment 4 is different from the technique described in Embodiment 3 in that the detected value input to the voltage command value calculation unit 24 is changed from the vehicle traveling speed S to the steering torque Ts, and a part of the calculation content in the voltage command value calculation unit 24 is different. Since other points are the same as those in Embodiment 3, the description thereof will be omitted.

[0158] In FIG. 29 described in Embodiment 3, whether the vehicle traveling speed S is low or high, the smaller the steering torque Ts, the smaller the slope of the q-axis current command value iq_ref with respect to the steering torque Ts. That is, the smaller the steering torque Ts, the more margin there is in the steering stability, so there is room to decrease the limit gain K described in Embodiment 2. Therefore, in the voltage command value calculation unit 24 in the present embodiment 4, as shown in FIG. 31, the limit gain K is varied.

[0159] In FIG. 31, “ΔTs” is the amount of change per unit time of the steering torque Ts, “K_H1” is the first torque reference value, and “K_L1” is the second torque reference value. The second torque reference value K_L1 is smaller than the first torque reference value K_H1. The limit gain K varies within the range of K_L1 to K_H1 according to the magnitude of the amount of change ΔTs. Specifically, when ΔTs is less than or equal to the first torque threshold value ΔTs1, the value of the limit gain K is set to K_L1. When ΔTs is greater than or equal to the second torque threshold value ΔTs2, the value of the limit gain K is set to K_H1. When ΔTs is within the range of ΔTs1 to ΔTs2, the value of the limit gain K continuously varies within the range of K_L1 to K_H1. As a result, it is possible to reduce the vibration and noise of the rotary machine 10 in a region where the variation of the steering torque Ts is small.

[0160] In FIG. 31, the horizontal axis is ΔTs. However, the same effect can be obtained by using the gradient (Δiq_ref / ΔTs) of the graph in FIG. 28 as the horizontal axis. In this specification, the above gradient (Δiq_ref / ΔTs) is referred to as the “current torque gradient”. That is, the current torque gradient is the ratio of the amount of change (Δiq_ref) of the q-axis current command value iq_ref to the amount of change (ΔTs) of the steering torque Ts. When the current torque gradient is smaller than the threshold value, the proportional gains Kpd and Kpq may be set to a small value Kp_L1. Also, when at least one of the current command values iq_ref and id_ref is smaller than the threshold value, or when the steering torque Ts is smaller than the threshold value, the proportional gains Kpd and Kpq may be set to a small value (for example, Kp_L1).

[0161] When the steering torque Ts or the current command values iq_ref and id_ref are small, the steering noise tends to be prominent. Therefore, in the fourth embodiment, when the current command values iq_ref and id_ref are below the threshold, or when the gradient (current torque gradient Δiq_ref / ΔTs) in the graph representing the relationship between the steering torque Ts and the torque current command value iq_ref is smaller than the threshold, or when the change amount ΔTs of the steering torque Ts per unit time is smaller than the first torque threshold ΔTs1, a configuration is proposed in which the proportional gains Kpd and Kpq are set to values lower than the first speed reference value Kp_H1. Thereby, in a situation where the steering noise is likely to be prominent, the non-interference control can be weakened to ensure quietness.

[0162] In the third and fourth embodiments, instead of inputting the rotational angular velocity ω to the inductance amplifiers 106d and 106q along with the non-interference control, the result of the rotational angular velocity ω passing through a low-pass filter may be input. The cut-off frequency of the low-pass filter is preferably higher than the upper limit value (for example, 5 Hz) of the steering frequency of the electric power steering apparatus 100. In this case, there is an effect of reducing the abnormal noise and vibration of the rotary machine 10 caused by the noise component included in the rotational angular velocity ω generated by introducing the non-interference control. Therefore, a quiet electric power steering apparatus 100 can be provided.

[0163] As described above, the first to fourth embodiments have been described. However, the present disclosure is not limited to the above-described embodiments, and can be freely changed without departing from the spirit of the present disclosure. Also, the above-described first to fourth embodiments can be combined as appropriate.

[0164] Furthermore, each component included in the rotation machine control devices 1 to 4 and the electric power steering device 100 described above has a computer system inside. Then, a program for realizing the functions of each component included in the rotation machine control devices 1 to 4 and the electric power steering device 100 described above is recorded on a computer-readable recording medium, and the program recorded on this recording medium is read into the computer system and executed, whereby the processing in each component included in the rotation machine control devices 1 to 4 and the electric power steering device 100 described above may be performed. Here, "reading the program recorded on the recording medium into the computer system and executing it" includes installing the program in the computer system. The "computer system" referred to here shall include hardware such as an OS and peripheral devices.

[0165] In addition, the "computer system" may include a plurality of computer devices connected via a network including a communication line such as the Internet, WAN, LAN, or dedicated line. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, magneto-optical disk, ROM, CD-ROM, or a storage device such as a hard disk built into the computer system. Thus, the recording medium storing the program may be a non-transitory recording medium such as a CD-ROM.

[0166] In addition, the recording medium also includes an internal or external recording medium provided so as to be accessible from a distribution server for distributing the program. Note that the program may be divided into a plurality of parts, downloaded at different timings, and then combined by each component included in the control devices 1 to 4 of the rotating machine and the electric power steering device 100. Also, the distribution servers for distributing each of the divided programs may be different. Further, the "computer-readable recording medium" includes those that hold a program for a certain period of time, such as a volatile memory (RAM) inside a computer system that becomes a server or a client when a program is transmitted via a network. Also, the above program may be for realizing a part of the functions described above. Furthermore, it may be a so-called difference file (difference program) that can realize the functions described above in combination with a program already recorded in the computer system.

Explanation of Signs

[0167] 1 to 4... Control devices 10... AC rotating machine 12... Inverter 22... Current detector 24d, 24q... (First-axis voltage command value calculator, second-axis voltage command value calculator) 100... Electric power steering device 104... Driving force transmission mechanism id, iq... Detected current value id_ref, iq_ref... Current command value iq_ref... Torque current command value iu, iv, iw... Rotating machine current K... Limiting gain K_H... First limit value K_L... Second limit value Kp_H... First proportional value Kp_H1... First speed reference value Kp_L... Second proportional value Kp_L1... Second speed reference value Kpd, Kpq... Proportional gain Ld…d-axis inductance m…voltage amplitude m1…first voltage threshold m2…second voltage threshold n1…first rotational speed threshold n2…second rotational speed threshold Ru, Rv, Rw…shunt resistance S…travel speed 100s…vehicle steering system Sun, Svn, Swn…lower arm switching element Sup, Svp, Swp…upper arm switching element Ts…steering torque vd, vq…(voltage command value of the first axis, voltage command value of the second axis) VRu, VRv, VRw…terminal voltage vu, vv, vw…voltage command value ΔTs…change amount of steering torque ω…rotational angular velocity ω1…first rotational angular velocity threshold ω2…second rotational angular velocity threshold ωcc…response angular frequency

Claims

1. A control device for an AC rotating machine, comprising: an inverter that applies a voltage to the AC rotating machine; a DC power supply that supplies DC power to the inverter; a current detector that detects a rotor current flowing through the AC rotating machine; a first-axis voltage command value calculator that calculates a voltage command value for a first axis in a two-axis rotation of the AC rotating machine; a second-axis voltage command value calculator that calculates a voltage command value for a second axis in the two-axis rotation; The first-axis voltage command value calculator multiplies a correction gain by a first deviation, which is a deviation between a current command value on the first axis and a detected current value on the first axis of the rotor current, to obtain a corrected first deviation; a second deviation, which is a deviation between a current command value on the second axis and a detected current value on the second axis of the rotor current; and calculates a voltage command value for the first axis based on the rotational angular velocity of the AC rotating machine, wherein a set value of the correction gain is increased when a physical quantity in the AC rotating machine changes suddenly or when an output voltage of the DC power supply increases. A control device for an AC rotating machine.

2. The control device for an AC rotating machine according to claim 1, wherein the first-axis voltage command value calculator sets an interference compensation term to zero when a physical quantity in the AC rotating machine changes suddenly or when an output voltage of the DC power supply increases.

3. The control device for an AC rotating machine according to claim 1 or 2, wherein the first-axis voltage command value calculator performs an integration operation using a value obtained by multiplying the second deviation by a product of the rotational angular velocity of the AC rotating machine, an inductance on the second axis, and a response angular frequency for adjusting a frequency response of the rotor current with respect to the current command value, and calculates an interference compensation term based on a result of the integration operation.

4. The control device for an AC rotating machine according to claim 2 or 3, wherein the first-axis voltage command value calculator sets an input value to an integration operation for calculating an interference compensation term to zero when the first deviation is greater than a threshold value.

5. The physical quantity is a rotational angular velocity of the AC rotating machine, and the first-axis voltage command value calculator increases a set value of the correction gain when the rotational angular velocity suddenly decreases. The control device for an AC rotating machine according to any one of claims 1 to 4.

6. The physical quantity is the rotor current, The first-axis voltage command value calculator of the AC rotating machine control device according to any one of claims 1 to 4 increases the set value of the correction gain when the rotating machine current rapidly increases with respect to the target value.

7. The physical quantity is the current command value on the first axis, The first-axis voltage command value calculator of the AC rotating machine control device according to any one of claims 1 to 4 increases the set value of the correction gain when the current command value on the first axis rapidly decreases.

8. The physical quantity is an external factor of the AC rotating machine excluding the output voltage of the DC power supply, The first-axis voltage command value calculator of the AC rotating machine control device according to any one of claims 1 to 4 increases the set value of the correction gain when the external factor rapidly changes.

9. The first-axis voltage command value calculator calculates the voltage command value of the first axis using a value obtained by multiplying the corrected first deviation by a proportional gain, The proportional gain is switched between a first proportional value and a second proportional value smaller than the first proportional value, When the voltage amplitude calculated based on the voltage command value is equal to or greater than a first voltage threshold, or when the rotational speed of the AC rotating machine is equal to or greater than a first rotational speed threshold, or when the rotational angular velocity of the AC rotating machine is equal to or greater than a first rotational angular velocity threshold, the proportional gain is set to the second proportional value. The control device for an AC rotating machine according to any one of claims 1 to 8.

10. The first-axis voltage command value calculator performs an integration operation using a value obtained by multiplying the second deviation by a limit gain, The limit gain is switched between a first limit value and a second limit value smaller than the first limit value, When the voltage amplitude is equal to or less than a second voltage threshold smaller than the first voltage threshold, or when the rotational speed of the AC rotating machine is equal to or less than a second rotational speed threshold smaller than the first rotational speed threshold, or when the rotational angular velocity of the AC rotating machine is equal to or less than a second rotational angular velocity threshold smaller than the first rotational angular velocity threshold, the limit gain is set to the second limit value. The control device for an AC rotating machine according to claim 9.

11. The inverter includes three sets of upper-arm switching elements, lower-arm switching elements, and shunt resistors respectively corresponding to three phases. The current detector according to claim 9 or 10, wherein when the voltage amplitude is smaller than the first voltage threshold, the rotational machine current in the three phases is detected based on the voltage across each corresponding shunt resistor, of the control device for an AC rotating electrical machine.

12. The first axis voltage command value calculator according to any one of claims 9 to 11, wherein when the rotational machine current exceeds a threshold value, the set value of the proportional gain is increased, of the control device for an AC rotating electrical machine.

13. The first axis voltage command value calculator according to any one of claims 9 to 11, wherein when a failure occurs at at least one location in the control device for the AC rotating electrical machine, the set value of the proportional gain is increased, of the control device for an AC rotating electrical machine.

14. A control device for an AC rotating electrical machine according to any one of claims 1 to 13, the AC rotating electrical machine, and a driving force transmission mechanism that transmits the driving force of the AC rotating electrical machine to a steering system of a vehicle, an electric power steering apparatus comprising the same.

15. The first axis voltage command value calculator calculates the voltage command value of the first axis using a value obtained by multiplying the corrected first deviation by a proportional gain, the proportional gain is switched between a first speed reference value and a second speed reference value smaller than the first speed reference value, The first axis voltage command value calculator according to claim 14, wherein when the traveling speed of the vehicle is smaller than a threshold value, the proportional gain is set to a value smaller than the first speed reference value, of the electric power steering apparatus.

16. The first axis voltage command value calculator, when the current command value is equal to or less than a threshold value, or when the steering torque is equal to or less than a threshold value, or when the gradient in a graph representing the relationship between the steering torque and the torque current command value is smaller than a threshold value, or when the amount of change in the steering torque per unit time is smaller than a threshold value, sets the proportional gain to a value smaller than the first speed reference value, of the electric power steering apparatus according to claim 15.

17. The first axis voltage command value calculator according to any one of claims 14 to 16, wherein an integration operation is performed using a result of passing the rotational angular velocity of the AC rotating electrical machine through a low-pass filter having a cut-off frequency higher than an upper limit value of the steering frequency of the electric power steering apparatus.

Images