The present invention relates to a control of an induction motor.
As a control of an induction motor, there is known a vector control for an induction motor, in which a three-phase AC current flowing through a stator is transformed into a two-axis orthogonal coordinate system in synchronization with a power supply angular frequency corresponding to a sum of a motor electric angular frequency and a slip frequency to generate an excitation current and a torque current, and the motor torque is controlled by adjusting the excitation current and the torque current.
When the slip angular frequency is controlled to be proportional to a ratio between the torque current and the rotor magnetic flux, the induction motor torque is proportional to a product between a rotor magnetic flux generated with a delay from the excitation current and a torque current orthogonal to the rotor magnetic flux. In addition, there is known a technique of providing a non-interference controller that cancels an interference term in advance in order to perform a control independently because the axes interfere with each other.
For example, in JP09-047097A, there is disclosed a technique of computing an interference voltage using a mathematical model by receiving an electric current command value as an input. In addition, in JP01-020688A, there is disclosed a technique of computing a non-interference voltage using a mathematical model by receiving a real electric current as an input.
However, for example, parameters used in the control described above, such as a self-inductance or a mutual inductance, change depending on a driving condition such as a torque or a rotation speed. Such a parameter change has non-linearity. However, the technique disclosed in JP09-047097A or JP01-020688A does not take account of an error caused by a parameter change. Therefore, responses of the torque and the electric current may become irregular.
In the control of the induction motor, the rotor magnetic flux generated by the excitation current inevitably has a delay. However, the technique disclosed in JP09-047097A or JP01-020688A does not take account of such a delay. That is, since the control is performed based on an input obtained by neglecting a transient response period until the rotor magnetic flux is generated, the control response may become instable in practice.
It is therefore an object of this disclosure to stably control an induction motor by avoiding an influence of a parameter error caused by a driving condition and an influence caused by a response delay of a rotor magnetic flux.
According to an aspect of this disclosure, there is provided a three-phase AC induction motor control device configured to control a torque based on a two-axis orthogonal coordinate system in synchronization with a power supply angular frequency. The three-phase AC induction motor control device includes: a non-interference controller configured to receive a motor rotation speed, a torque command value, and a power supply voltage as an input and compute a torque axis non-interference compensation voltage and a magnetic flux axis non-interference compensation voltage by referencing a map stored in advance; and a non-interference magnetic flux response filter configured to perform filtering, including a direct transfer term and a rotor magnetic flux response delay, for the torque axis non-interference compensation voltage.
The foregoing and additional features and characteristics of this disclosure will become more apparent from the following detailed description considered with the reference to the accompanying drawings.
Hereinafter, embodiments of this disclosure will be described with reference to the accompanying drawings.
The DC power supply 2 is a power supply capable of supplying a high voltage, such as a stack type lithium ion battery.
A pulse width modulation (PWM) converter 6 receives three-phase voltage command values vu*, vv*, and vw* computed by a coordinate transformer 12 described below and generates PWM_Duty driving signals Duu*, Dul*, Dvu*, Dvl*, Dwu*, and Dwl* of a switching element (such as IGBT) of the inverter 3 based on such command values.
The inverter 3 is a three-phase voltage type inverter, which switches a DC voltage of the DC power supply 2 to an AC voltage depending on the PWM_Duty driving signal Duu*, Dul*, Dvu*, Dvl*, Dwu*, and Dwl* and supplies the resulting voltage to the induction motor 1.
An electric current sensor 4 detects an electric current having at least two phases (for example, U-phase and V-phase) iu and iv out of the three-phase current supplied from the inverter 3 to the induction motor 1 and inputs it to the A/D converter 7. The A/D converter 7 inputs the current values ius and ivs obtained by performing A/D conversion for the current values iu and iv into the coordinate transformer 11. It is noted that, when the electric current sensor 4 is provided for two phases such as a U-phase and a V-phase as illustrated in
[Formula 1]
iws=−iws−ivs (1)
A magnetic pole position detector 5 inputs an A-phase pulse, a B-phase pulse, or a Z-phase pulse to a pulse counter 8 depending on a rotor angle of the induction motor 1. The pulse counter 8 computes a mechanical angle θrm of the rotor based on the input pulse and outputs the resulting angle to the angular velocity calculator 9.
The angular velocity calculator 9 computes a mechanical angular velocity ωrm of the rotor and an electrical angular velocity ωre of the rotor by multiplying the mechanical angular velocity ωrm of the rotor by the number p of pole pairs of the motor based on a temporal change rate of the input mechanical angle θrm.
The coordinate transformer 12 transforms a two-axis orthogonal DC coordinate system (γ-δ axes) rotating at a power supply angular velocity ω described below into a three-phase AC coordinate system (uvw axes). The coordinate transformer 12 receives a γ-axis voltage command value (excitation voltage command value) vγs*, a δ-axis voltage command value (torque voltage command value) vδs*, and a power supply angle θ obtained by integrating the power supply angular velocity ω using the angle transformer 10, computes the UVW-phase voltage command values vu*, vv*, and vw* through a coordinate transformation process of Formula (2), and outputs the UVW-phase voltage command values vu*, vv*, and vw* to the PWM converter 6.
The coordinate transformer 11 transforms the three-phase AC coordinate system (uvw axes) to the two-axis orthogonal DC coordinate system (γ-δ axes) described above. The coordinate transformer 11 receives a U-phase current ius, a V-phase current ivs, a W-phase current iws obtained through Formula (1), and the angle θ obtained by integrating the power supply angular velocity ω using the angle transformer 10 and computes the γ-axis current (excitation current) iγs and the δ-axis current (torque current) iδs based on Formula (3).
The electric current command value calculator 13 receives a target motor torque, a motor rotation speed (mechanical angular velocity ωrm), and a DC power supply voltage Vdc and computes a γ-axis current command value (excitation current command value) iγs* and a δ-axis current command value (torque current command value) iδs*. It is noted that the DC power supply voltage Vdc is directly detected using a voltage sensor. The target motor torque is set using a torque setting unit (not illustrated). For example, if this system is applied to an electric vehicle, the target torque is set based on a depression level of an accelerator pedal from a driver and the like.
A non-interference controller 17 receives the target motor torque, the motor rotation speed (mechanical angular velocity ωrm), and the DC power supply voltage Vdc, reads a non-interference compensation voltage (a torque axis non-interference compensation voltage or a magnetic flux axis non-interference compensation voltage) necessary to compensate for an interference voltage between the γ-δ orthogonal coordinate axes from a map stored in a memory in advance, and outputs the non-interference compensation voltage. For the magnetic flux axis, the value of the map is directly added to the voltage command value vγs* as a magnetic flux axis non-interference compensation voltage vγs_dcpl. For the torque axis, a value obtained by filtering the value of the map using a non-interference magnetic flux response filter 18 is added to the voltage command value vδs* as a torque axis non-interference compensation voltage vδs_dcpl. The non-interference controller 17 and the non-interference magnetic flux response filter 18 will be described in detail below.
A γ-axis current feedback controller (excitation current feedback (FB) controller) 15 performs a process of causing the measured γ-axis current value (excitation current) iγs to follow the γ-axis current command value (excitation current command value) iγs* with desired responsiveness without a steady-state offset. Similarly, a δ-axis current feedback controller (torque current FB controller) 16 also performs a process of causing the δ-axis current (torque current) iδs to follow the δ-axis current command value (torque current command value) iδs*.
Typically, if the γ-δ axes non-interference voltage correction is operated ideally, it provides a simple control target characteristic of one-input and one-output. Therefore, the γ-axis current feedback controller 15 and the δ-axis current feedback controller 16 can be implemented using a simple PI feedback compensator.
A value obtained by correcting the voltage command value output from the γ-axis current feedback controller 15 with the non-interference voltage vγs_dcpl* described above is the γ-axis voltage command value (excitation voltage command value) vγs* input to the coordinate transformer 12. Similarly, a value obtained by correcting a voltage command value output from the δ-axis current feedback controller 16 using the non-interference voltage vδs_dcpl* described above is the δ-axis voltage command value (torque voltage command value) vδs* input to the coordinate transformer 12.
The slip frequency controller 14 receives the γ-axis current (excitation current) iγs and the δ-axis current (torque current) iδs and computes the slip angular velocity ωse based on Formula (4). It is noted that Formula (4) describes only steady-state values. A value obtained by adding the computed slip angular velocity ωse to the rotor electric angular velocity ωre is output as the power supply angular velocity ω. By performing the slip frequency control, the torque of the induction motor is proportional to a product between the γ-axis current (excitation current) iγs and the δ-axis current (torque current) iδs.
Here, the non-interference controller 17 and the non-interference magnetic flux response filter 18 will be described in detail.
As described above, the non-interference controller 17 receives the target motor torque, the motor rotation speed (mechanical angular velocity ωrm), and the DC power supply voltage Vdc as input values and stores a map for extracting the non-interference compensation voltage necessary to compensate for the interference voltage between the γ-δ orthogonal axis coordinates in a memory in advance. It is difficult to store, in this map, the non-interference compensation voltage in consideration of a transient response. In this regard, the steady-state non-interference compensation voltages appropriate to each input value are obtained in advance through experiments and are stored.
The non-interference magnetic flux response filter 18 performs filtering for the torque axis non-interference compensation voltage output from the non-interference controller 17, including a direct transfer term expressed in a transfer function of Formula (5) and a rotor magnetic flux response delay.
It is noted that a factor σ of Formula (5) denotes a leakage coefficient set to “σ=1−M2/(Ls·Lr).” “τf” denotes a time constant of the rotor magnetic flux response delay and is typically defined as a ratio Lr/Rr between the inductance Lr of the rotor side (second order side) and the resistance Rr. “s” denotes a Laplacian operator.
As described above, since the non-interference controller 17 computes the non-interference compensation voltage by referencing the map, it is possible to accurately perform non-interference compensation by avoiding an influence from a steady-state parameter error caused by a driving condition such as the torque, the rotation number, and the power supply voltage, and an influence from the response delay of the magnetic flux of the rotor. As a result, it is possible to remarkably improve responsiveness of an electric current control system. In addition, the number of operations such as multiplication, addition, or subtraction is relatively small, and an operation period can be set to be relatively long. This also contributes to reduction of a computation load.
Comparing
As described above, according to the first embodiment, the non-interference controller 17 outputs the non-interference compensation voltage by referencing a map based on a driving condition. Therefore, it is possible to reduce a computation load and avoid an influence from a change of the parameter caused by a driving condition. In addition, the non-interference magnetic flux response filter 18 performs filtering for the torque axis interference compensation voltage, which is one of the output values of the non-interference controller 17, including the direct transfer term and the rotor magnetic flux response delay. Therefore, it is possible to accurately perform non-interference compensation by avoiding an influence from the response delay of the rotor magnetic flux. As a result, it is possible to remarkably improve responsiveness of the electric current control system.
Unlike the first embodiment, electric current delay filters 20 and 21 are provided to perform filtering for the γ-axis and δ-axis non-interference compensation voltages. Hereinafter, a description will focus on a difference between the first and second embodiments.
As the electric current delay filters 20 and 21, a digital filter obtained by discretizing a transfer characteristic G2(s) of Formula (6) using Tustin's approximation or the like is employed. A time constant τc is determined depending on a response delay of the electric current feedback control system.
Since the electric current delay filters 20 and 21 are provided to perform a delay process including simulation of a delay in the electric current control as described above, it is possible to accurately perform non-interference compensation at a high frequency and improve torque responsiveness even at a high frequency range.
Although the transient responsiveness of the γ-axis current (excitation current) iγs, the δ-axis current (torque current) iδs, or the torque is improved in the first embodiment, it is recognized that an electric current waveform and a torque waveform vibrate immediately after the step as seen from the enlarged view of the portion immediately after the step as illustrated in
As described above, according to the second embodiment, the electric current delay filters 20 and 21 are provided to simulate the electric current control delay. Therefore, in addition to the effects of the first embodiment, it is possible to further improve the transient responsiveness of the γ-axis current (excitation current) iγs, the δ-axis current (torque current) iδs, or the torque.
Unlike the first embodiment, the parameter used in the non-interference magnetic flux response filter 18 is a variable value obtained by inputting an estimated or measured value of the rotor characteristic. That is, the characteristic of the non-interference magnetic flux response filter 18 is variable depending on the rotor characteristic. The rotor characteristic includes, for example, a temperature, a resistance, and an inductance of the rotor. In addition, the characteristic of the non-interference magnetic flux response filter includes, for example, a time constant, an occupation ratio between a direct transfer term and a transient term, and the like. Hereinafter, a description will focus on a difference between the first and third embodiments.
In the first and second embodiments, it is possible to improve robustness of the control system for a parameter error caused by a difference of the driving condition. However, the parameters used in the non-interference magnetic flux response filter 18 are also affected by an external factor other than the driving condition, such as high sensitivity against a temperature change. Therefore, a parameter error may be generated by an external factor. In this regard, the parameters used in the non-interference magnetic flux response filter 18 are set to be variable depending on the rotor characteristic in order to match a temperature change. For example, if a temperature of the rotor is used as an input, a resistance Rr of the rotor side (second order side) can be corrected using a temperature coefficient α that changes depending on a material of the rotor as expressed in Formula (7). As a result, a time constant τf of the rotor magnetic flux response delay of Formula (5) changes.
[Formula 7]
Rt=R20{1+α20(t−20)} (7)
It is noted that “Rt” of Formula (7) denotes a resistance at a temperature of t[° C.], “α20” denotes a temperature coefficient at a temperature of 20[° C.] as a reference, and “R20” denotes a resistance at a temperature of 20[° C.] as a reference.
As a result, it is possible to accurately perform non-interference compensation even when a parameter error is generated due to a change of the rotor characteristic. Accordingly, it is possible to remarkably improve responsiveness of an electric current control system.
It is noted that a means for estimating a resistance Rr of the rotor side (second order side) may be provided, and the parameters used in the non-interference magnetic flux response filter 18 may be corrected by using the estimated resistance Rr as an input. In addition, an inductance out of the parameters used in the non-interference magnetic flux response filter 18 depends on an electric current. Therefore, a map may be created in advance such that an electric current is used as an input, and an inductance is used as an output, and the inductance may be corrected using this map.
It is recognized that, by setting the parameters used in the non-interference magnetic flux response filter 18 to be variable in consideration of a temperature change as described above, it is possible to further improve an electric current response or a torque response, compared to the first embodiment.
Hereinbefore, according to the third embodiment, in addition to the effects of the first embodiment, it is possible to improve an electric current response or a torque response when there is a parameter error by setting the characteristic of the non-interference magnetic flux response filter 18 to be variable depending on the rotor characteristic.
As the electric current delay filters 40 and 41, a digital filter obtained by discretizing the transfer characteristic G2(s) expressed in Formula (6) described above using Tustin's approximation or the like is employed. The time constant τc is determined based on a response delay of the electric current feedback control system.
In the second embodiment described above, the non-interference controller 17 performs filtering for the non-interference compensation voltage computed based on a map using the electric current delay filters 20 and 21 that simulates an electric current control delay. As a result, it is possible to sufficiently improve torque responsiveness and the like. However, the map is referenced using an electric current command value without considering an electric current control delay. Therefore, an influence of the electric current control delay still remains in transient responsiveness, compared to an ideal non-interference compensation based on a mathematical model.
In contrast, according to the fourth embodiment, filtering, in which the electric current control delay is simulated, is performed for the electric current command values iγs* and iδs* computed using the electric current command value calculator 13, and the filtering result is used as an input of the non-interference controller 17. Therefore, it is possible to reflect an electric current control delay in a more appropriate position. Accordingly, it is possible to accurately perform non-interference compensation even when a parameter error is generated due to a difference of the driving condition. As a result, it is possible to remarkably improve responsiveness of an electric current control system.
As described above, by reflecting an electric current control delay in an appropriate position, it is possible to further improve responsiveness of the γ-axis current (excitation current) iγs and the δ-axis current (torque current) iδs at the time of the torque step, compared to the first embodiment, as indicated by the encircled portions of
As described above, according to the fourth embodiment, in addition to the effects of the first embodiment, it is possible to further improve responsiveness of an electric current control system by performing filtering, in which an electric current control delay is simulated, for the electric current command values iγs and iδs.
Although various embodiments of this disclosure have been described hereinbefore, they are just for illustrative purposes and are not intended to specifically limit the technical scope of the invention. Instead, it would be appreciated that various changes or modifications may be possible without departing from the spirit and scope of the invention.
This application claims priority based on Japanese Patent Application No. 2012-065887 filed with the Japan Patent Office on Mar. 22, 2012, the entire contents of which are expressly incorporated herein by reference.
Number | Date | Country | Kind |
---|---|---|---|
2012-065887 | Mar 2012 | JP | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/JP2013/056622 | 3/11/2013 | WO | 00 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2013/141059 | 9/26/2013 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
8344680 | Kitanaka | Jan 2013 | B2 |
8674647 | Iwaji et al. | Mar 2014 | B2 |
20100259207 | Kitanaka | Oct 2010 | A1 |
20110204831 | Iwaji et al. | Aug 2011 | A1 |
Number | Date | Country |
---|---|---|
2 211 457 | Jul 2010 | EP |
64-020688 | Jan 1989 | JP |
9-047097 | Feb 1997 | JP |
Entry |
---|
L. Umanand et al.,Adaptation of the Rotor Time Constant for Variations in the Rotor Resistance of an Induction Motor, Power Electronics Specialists Conference, PESC, IEEE, 1994, pp. 738-743. |
D. Fodor et al., Digitized Vector Control of Induction Motor with DSP, Industrial Electronics, Control and Instrumentation, vol. 3, Sep. 5, 1994, pp. 2057-2062. |
Number | Date | Country | |
---|---|---|---|
20150048774 A1 | Feb 2015 | US |