FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
[See section 10, Rule 13]
MOTOR DRIVE APPARATUS
MITSUBISHI ELECTRIC CORPORATION, A CORPORATION ORGANISED
AND EXISTING UNDER THE LAWS OF JAPAN, WHOSE ADDRESS IS 7-3,
MARUNOUCHI 2-CHOME, CHIYODA-KU, TOKYO 1008310, JAPAN
THE FOLLOWING SPECIFICATION PARTICULARLY DESCRIBES THE
INVENTION AND THE MANNER IN WHICH IT IS TO BE PERFORMED
2
DESCRIPTION
Field
[0001] The present invention relates to a motor drive
5 apparatus that drives a synchronous motor having saliency
without a sensor.
Background
[0002] A rotating magnetic field is produced when a
10 multi-phase alternating current voltage is applied to a
stator of a synchronous motor. The synchronous motor
produces torque by a magnetic interaction between the
rotating magnetic field and a rotor. The multi-phase
alternating current voltage is an alternating current
15 voltage of three phases or four phases or more. When the
synchronous motor rotates, the phase and frequency need to
be synchronized between the rotating magnetic field and the
rotor. Thus, in order to drive the synchronous motor,
information on a rotational position or rotational
20 frequency of the rotor is required. There is a method of
using a position sensor or a speed sensor to obtain the
information on the rotational position or rotational
frequency of the rotor. On the other hand, the application
of a drive system that does not use such sensors is also
25 spreading in order to reduce the number of parts and the
number of wirings. The state and the drive system without
a sensor for acquiring the information on the rotational
position or rotational frequency of the rotor are called
“sensorless”.
30 [0003] When the synchronous motor is started without a
sensor, it is necessary to start switching control for
semiconductor elements of an inverter while synchronizing
the phase and frequency of an output voltage with a
3
rotating state of the rotor such that an overcurrent or the
like does not occur. Therefore, the sensorless drive
system of the synchronous motor usually distinguishes the
case where the synchronous motor is started from a case
5 other than the case where the synchronous motor is started.
Specifically, the sensorless drive system of the
synchronous motor is roughly divided into two algorithms
called “steady state estimation” and “initial estimation”.
[0004] The steady state estimation algorithm is applied
10 when the semiconductor elements of the inverter that drives
the motor perform a switching operation, and the torque or
rotating state of the motor is continuously controlled. On
the other hand, the initial estimation algorithm is applied
when the semiconductor elements of the inverter start the
15 switching operation from a state in which the switching
operation is stopped. That is, the steady state estimation
is started by using information on the rotational position
and rotational frequency of the rotor obtained by the
initial estimation.
20 [0005] As an example of the initial estimation, Patent
Literature 1 discloses a method of estimating the
rotational frequency of the rotor on the basis of a period
with which the current polarity of a specific phase is
reversed, and estimating the rotational position of the
25 rotor on the basis of a timing at which the current
polarity of a specific phase is reversed. Note that for
convenience, reversing the polarity of the phase current,
that is, the sign of the phase current, is called “zero
crossing”.
30
Citation List
Patent Literature
[0006] Patent Literature 1: Japanese Patent Application
4
Laid-open No. 2004-336866
Summary
Technical Problem
5 [0007] In an inverter whose output voltage is controlled
by pulse width modulation (PWM), it is known that the
actual output voltage may deviate from a voltage command
value intended by a controller due to the influence of an
on-voltage of a switching element included in the inverter,
10 the dead time in controlling the inverter, and the like. A
voltage error that is a difference between the voltage
command value and the actual output voltage becomes more
remarkable as the current flowing through the switching
element is smaller, that is, as a motor current is smaller.
15 Note that the motor current is a current flowing through
each phase of the motor, that is, a phase current of the
motor.
[0008] In the initial estimation described in Patent
Literature 1, the current of the specific phase is
20 controlled such that it pulsates near zero, so that the
voltage error mentioned above is remarkable. When the
voltage error is large, the phase current of another phase
becomes too small, and the accuracy of estimating the
frequency is reduced. Alternatively, the phase current of
25 another phase becomes excessive to cause magnetic
saturation, and the accuracy of estimating the rotational
position is reduced. Note that the “another phase” here
means a phase whose phase current is too small or too large
with respect to the specific phase. As described above,
30 the voltage error being large causes a problem in that the
accuracy of estimating the frequency or the rotational
position is reduced.
[0009] Moreover, signals of a voltage sensor and a
5
current sensor include noise that enters an analog circuit.
Therefore, the voltage sensor or the current sensor cannot
always detect the accurate zero crossing timing.
[0010] Moreover, the number of times a detected value of
5 the current can be sampled by a digital circuit is limited.
In particular, during high-speed rotation of the motor, the
motor current pulsates with a short period, so that the
detected value of the current is often not obtained, and an
error of the zero crossing timing becomes more remarkable.
10 [0011] The present invention has been made in view of
the above, and an object of the present invention is to
provide a motor drive apparatus that can estimate, with
high accuracy, information on a rotational position or
rotational frequency of a synchronous motor that is
15 rotating.
Solution to Problem
[0012] In order to solve the above problem and achieve
the object, a motor drive apparatus according to the
20 present invention includes an inverter that drives a
synchronous motor with saliency, a controller that controls
an operating state of the inverter, and a detector that
detects phase currents of the synchronous motor. The
controller includes a voltage control unit that determines
25 an output voltage of the inverter, a coordinate
transformation unit that transforms the phase currents into
a two-phase current in a stationary coordinate system, a
pulsation extraction unit that extracts a pulsating current
from the two-phase current, and a phase synchronization
30 calculation unit that estimates and calculates a frequency
and a phase of the pulsating current. The voltage control
unit outputs a voltage command value at which none of the
phase currents equals zero.
6
Advantageous Effects of Invention
[0013] The motor drive apparatus according to the
present invention can estimate, with high accuracy, the
5 information on the rotational position or rotational
frequency of the synchronous motor that is rotating.
Brief Description of Drawings
[0014] FIG. 1 is a diagram of a configuration of a motor
10 drive apparatus according to a first embodiment.
FIG. 2 is a circuit diagram illustrating a
configuration of an inverter illustrated in FIG. 1.
FIG. 3 is a schematic diagram for explaining a voltage
command value output by a voltage control unit according to
15 the first embodiment.
FIG. 4 is a graph illustrating a relationship between
a phase of a voltage command value vector and magnitude of
a phase current in the first embodiment.
FIG. 5 is a block diagram illustrating a configuration
20 of a pulsation extraction unit according to a second
embodiment.
FIG. 6 is a block diagram illustrating a configuration
of a phase synchronization calculation unit according to
the second embodiment.
25 FIG. 7 is a block diagram illustrating a configuration
of a phase synchronization calculation unit according to a
third embodiment.
FIG. 8 is a graph for explaining an operation of
switching the gain of an amplifier according to the third
30 embodiment.
FIG. 9 is a diagram of a configuration of a motor
drive apparatus according to a fourth embodiment.
FIG. 10 is a block diagram illustrating a
7
configuration of a correction calculation unit according to
the fourth embodiment.
FIG. 11 is a block diagram illustrating an example of
a hardware configuration that implements arithmetic
5 functions of a controller of the fourth embodiment.
FIG. 12 is a block diagram illustrating another
example of a hardware configuration that implements the
arithmetic functions of the controller of the fourth
embodiment.
10
Description of Embodiments
[0015] A motor drive apparatus according to embodiments
of the present invention will now be described in detail
with reference to the drawings. Note that the present
15 invention is not limited to the following embodiments.
[0016] First Embodiment.
FIG. 1 is a diagram of a configuration of a motor
drive apparatus according to a first embodiment. FIG. 2 is
a circuit diagram illustrating a configuration of an
20 inverter illustrated in FIG. 1. In FIG. 1, a motor drive
apparatus 100 includes a motor 2 including a rotor 2a, an
inverter 1 that drives the motor 2, a controller 3 that
controls an operating state of the inverter 1, and a
detector 4 that detects a phase current of the motor 2.
25 [0017] The inverter 1 receives direct current power from
a power source 110 and applies a voltage with variable
amplitude and variable frequency to the motor 2. The
inverter 1 adjusts the voltage applied to the motor 2 by
performing pulse width modulation (PWM) control on a
30 plurality of semiconductor elements not shown in FIG. 1.
[0018] The motor 2 is a synchronous motor with saliency.
An example of the synchronous motor with saliency is a
synchronous reluctance motor (hereinafter denoted as
8
"SynRM"). A rotor of the SynRM has a characteristic that
magnetic resistance in a radial direction changes in
accordance with an angle of rotation with respect to a
cylindrical axis. Such a characteristic is called
5 “saliency”. When a current flows through a stator of the
SynRM by applying a voltage thereto, a magnetic field is
produced that crosses the circumference of the rotor in the
radial direction. At this time, torque for rotating the
rotor is generated in a direction in which the magnetic
10 flux increases, that is, in a direction in which the
magnetic resistance of a magnetic path decreases. The
torque thus generated due to the saliency of the rotor is
called reluctance torque.
[0019] FIG. 2 illustrates a circuit configuration when
15 the inverter 1 is a three-phase inverter. The inverter 1
illustrated in FIG. 2 includes a leg 10A in which a
semiconductor element UP of an upper arm and a
semiconductor element UN of a lower arm are connected in
series, a leg 10B in which a semiconductor element VP of an
20 upper arm and a semiconductor element VN of a lower arm are
connected in series, and a leg 10C in which a semiconductor
element WP of an upper arm and a semiconductor element WN
of a lower arm are connected in series. The legs 10A, 10B,
and 10C are connected in parallel to one another.
25 [0020] A bus voltage is applied to the inverter 1
through direct current buses 15a and 15b. The inverter 1
drives the motor 2 by converting the direct current power
of the power source 110 supplied through the direct current
buses 15a and 15b into alternating current power, and
30 supplying the alternating current power obtained by the
conversion to the motor 2.
[0021] FIG. 2 illustrates, as an example, a case where
the semiconductor elements UP, UN, VP, VN, WP, and WN are
9
metal-oxide-semiconductor field-effect transistors
(MOSFETs). The semiconductor element UP includes a
transistor 10a and a diode 10b connected in anti-parallel
to the transistor 10a. The other semiconductor elements UN,
5 VP, VN, WP, and WN each have a similar configuration. The
anti-parallel means that an anode side of the diode is
connected to a first terminal corresponding to a source of
the MOSFET, and a cathode side of the diode is connected to
a second terminal corresponding to a drain of the MOSFET.
10 [0022] Note that instead of the MOSFETs, the
semiconductor elements UP, UN, VP, VN, WP, and WN may be,
for example, insulated gate bipolar transistors (IGBTs).
[0023] Moreover, FIG. 2 illustrates the configuration
including the three legs in each of which the semiconductor
15 element of the upper arm and the semiconductor element of
the lower arm are connected in series, but it is not
limited to this configuration. The number of legs may be
four or more. Furthermore, in FIG. 2, one leg includes one
pair of the semiconductor elements of the upper and lower
20 arms, but one leg may include a plurality of pairs of the
semiconductor elements of the upper and lower arms.
[0024] Also, when the transistor 10a of each of the
semiconductor elements UP, UN, VP, VN, WP, and WN is a
MOSFET, at least one of the semiconductor elements UP, UN,
25 VP, VN, WP, and WN may be formed of a wide bandgap
semiconductor such as silicon carbide, gallium nitride
material, or diamond.
[0025] Wide bandgap semiconductors generally have higher
withstand voltage characteristic and heat resistance than
30 silicon semiconductors. Therefore, effects of the
withstand voltage characteristic and heat resistance can be
obtained when the MOSFET formed of the wide bandgap
semiconductor is used for at least one of the semiconductor
10
elements UP, UN, VP, VN, WP, and WN.
[0026] A connection point 12 between the semiconductor
element UP of the upper arm and the semiconductor element
UN of the lower arm is connected to a first phase (for
5 example, a U phase) of the motor 2, a connection point 13
between the semiconductor element VP of the upper arm and
the semiconductor element VN of the lower arm is connected
to a second phase (for example, a V phase) of the motor 2,
and a connection point 14 between the semiconductor element
10 WP of the upper arm and the semiconductor element WN of the
lower arm is connected to a third phase (for example, a W
phase) of the motor 2. In the inverter 1, the connection
points 12, 13, and 14 each form an alternating current
terminal.
15 [0027] Returning to FIG. 1, the description of the motor
drive apparatus 100 will be continued. The controller 3
includes a voltage control unit 50 that determines the
output voltage of the inverter 1, a coordinate
transformation unit 20 that transforms phase currents into
20 a two-phase current in a stationary coordinate system, a
pulsation extraction unit 30 that extracts a pulsating
current from the two-phase current, and a phase
synchronization calculation unit 40 that estimates and
calculates a frequency and a phase of the pulsating current.
25 [0028] The voltage control unit 50 calculates a voltage
command value that is a command value of the voltage to be
output by the inverter 1. Switching states of the
semiconductor elements UP, UN, VP, VN, WP, and WN of the
inverter 1 are determined on the basis of the voltage
30 command value. Here, the voltage command value for setting
the torque of the motor 2 to a desired value needs to be
calculated on the basis of rotational position and
rotational frequency of the rotor 2a.
11
[0029] There is a method of using a position sensor or a
speed sensor to obtain information on the rotational
position or rotational frequency of the rotor 2a. However,
since such sensors are often installed coaxially with the
5 motor, the allowable shaft length of the motor is reduced
if the installation space of the motor is limited.
Therefore, the method of using the position sensor or the
speed sensor has a disadvantage in that the motor output
ends up being limited. Also, the position sensor or the
10 speed sensor requires wiring the signal line of the sensor
to hardware on which the controller 3 is mounted.
Therefore, the method of using the position sensor or the
speed sensor has a problem in that the component cost
increases and there is a risk of disconnection. For these
15 reasons, the application of the sensorless drive system,
which is a drive system not using the position sensor or
the speed sensor, is spreading. The motor drive apparatus
100 according to the first embodiment is also assumed to
drive the motor without a sensor.
20 [0030] When the motor 2 is started without a sensor, it
is necessary to start switching control of the inverter 1
with the phase and frequency of the output voltage being
synchronized with a rotating state of the rotor 2a such
that an overcurrent or the like does not occur. From this
25 point of view, the sensorless drive system of the
synchronous motor is divided into two algorithms, i.e.,
steady state estimation and initial estimation, as
described above.
[0031] As described above, the steady state estimation
30 is the algorithm applied when the inverter 1 starts
switching and the torque or rotational frequency of the
motor 2 is continuously controlled.
[0032] One typical method of the steady state estimation
12
is a method of using an induced voltage (also called a
"counter electromotive voltage") of the motor 2. This
method calculates the induced voltage on the basis of a
mathematical model of the motor, and defines a phase
5 difference between an axis of coordinates of a true dq
coordinate system corresponding to the position of the
rotor 2a and an axis of coordinates of an estimated dq
coordinate system in which the voltage command value is
calculated. Then, the estimated dq coordinate system is
10 corrected to eliminate the phase difference, and as a
result, estimated values of the position and rotational
frequency of the rotor 2a are obtained. Note that the dq
coordinate system represents a rotating coordinate system
when the motor 2 is vector-controlled, and is a widely
15 known concept.
[0033] Another method of the steady state estimation is
a method of applying a high-frequency voltage to the motor
2 and using a current response at that time. This method
obtains estimated values of the rotational position and
20 rotational frequency of the rotor 2a by using the fact that
the current on the dq coordinates has an elliptical path
when the high-frequency voltage is applied to the motor 2
with saliency. This method is often used under low-speed
operating conditions where the induced voltage is small.
25 [0034] Meanwhile, as described above, the initial
estimation is the algorithm applied when the inverter 1
starts switching from a state in which the switching is
stopped. As described above, the steady state estimation
algorithm requires an initial value of either one or both
30 of the information on the rotational position and
rotational frequency of the rotor 2a when starting the
calculation. Energization started while a difference
between the initial value and a true value is large causes
13
an inconvenience such as overcurrent. For this reason, the
steady state estimation is started by using the information
on the rotational position and rotational frequency of the
rotor 2a obtained by the initial estimation. The initial
5 estimation algorithm is thus executed only for a short time
when the inverter 1 is started.
[0035] The operation of the voltage control unit 50
illustrated in FIG. 1 will now be described. In the first
embodiment, a method of the initial estimation will be
10 described in detail. Note that a method of the steady
state estimation is not particularly limited.
[0036] When the voltage control unit 50 receives a
command to start the inverter 1 from a higher-level control
system (not shown), the voltage control unit 50 calculates
15 a voltage command value such that no motor current of the
phases equals zero. Furthermore, the voltage control unit
50 generates the voltage command value that is a direct
current voltage and has a voltage vector in the same
direction as or in the opposite direction to any of the
20 phases of the motor 2. Note that a reason for generating
such a voltage command value will be described later.
[0037] FIG. 3 is a schematic diagram for explaining the
voltage command value output by the voltage control unit 50
according to the first embodiment. Here, the motor 2 is
25 assumed to be a three-phase motor. Phases of the threephase
motor are referred to as a u-phase, a v-phase, and a
w-phase. The u-phase, v-phase, and w-phase form a threephase
coordinate system. The uvw three-phase coordinate
system is a stationary coordinate system. Note that
30 although a voltage command value vector is in the same
direction as the u-phase as an example in FIG. 3, the
direction is not limited thereto. The voltage command
value vector may be in the same direction as the v-phase or
14
the w-phase.
[0038] First, an average value of a u-phase current is
represented by “iu0”, an average value of a v-phase current
is represented by “iv0”, and an average value of a w-phase
5 current is represented by “iw0”. The inverter 1 then
outputs a voltage according to the voltage command value
illustrated in FIG. 3. As a result, the average value “iv0”
of the v-phase current and the average value “iw0” of the
w-phase current have the sign opposite to that of the
10 average value “iu0” of the u-phase current and the
magnitude half that thereof.
[0039] Moreover, arrows “α” and “β” in FIG. 3 indicate
coordinate axes when the voltage and current are subjected
to three-phase to two-phase transformation. That is, “α”
15 and “β” form a two-phase coordinate system. The αβ twophase
coordinate system is a stationary coordinate system
as with the uvw three-phase coordinate system. A
transformation matrix from the uvw three-phase coordinate
system to the αβ two-phase coordinate system is given by
20 the following formula.
[0040] [Formula 1]
  


  


 

 


 
  


 



w vu
0 3 2 3 2
2 3 1 1 2 1 2
(1)
[0041] Note that the transformation matrix differs from
the one represented by the above formula (1) depending on
25 how the coordinate axes are defined, but it is common to
define the coordinate axes such that the α axis coincides
with any of the u, v, and w axes. When the α and β axes
are defined as in FIG. 3 and the above formula (1), the
voltage command value need only be set such that “vα” is
30 non-zero and “vβ” is zero. Note that the d and q axes
15
illustrated in FIG. 3 are ones obtained by a rotating
coordinate transformation of the αβ two-phase coordinate
system by an angle of rotation θ of the rotor 2a.
[0042] Now, when a direct current voltage is applied to
5 the stator of the motor that is rotating, pulsation occurs
in the phase current due to saliency. Properties of the
pulsating current will be described in detail below.
[0043] First, a voltage equation of the SynRM in the αβ
coordinate system is represented by the following formula.
10 [0044] [Formula 2]
 

 

 

 



  


 



 
 

 i
i
PL R PL
R PL PL
v
v
s
s
(2)
[0045] In the above formula (2), “vα” and “vβ” represent
the voltages obtained by two-phase transformation, and “iα”
and “iβ” represent the currents obtained by two-phase
15 transformation. Moreover, “P” represents a differential
operator, and “Rs” represents a coil resistance.
Furthermore, “Lα”, “Lβ”, and “Lαβ” are defined by the
following formula.
[0046] [Formula 3]
 
L L L  2
L L L 2
L L sin2
L L L cos2
L L L cos2
1 d q
0 d q
1
0 1
0 1
 
 
 
  
  



20 (3)
[0047] In the above formula (3), “Lα” represents an α-
axis inductance, “Lβ” represents a β-axis inductance, and
“Lαβ” represents an αβ-axes mutual inductance. Furthermore,
“θ” represents an angle of rotation of the rotor 2a, “L0”
25 represents an average inductance, “L1” represents a
differential inductance, “Ld” represents a d-axis
16
inductance, and “Lq” represents a q-axis inductance.
[0048] In the motor with saliency, the d-axis inductance
Ld and the q-axis inductance Lq are different, so that the
differential inductance L1 is non-zero according to the
5 fifth equation of the above formula (3). Therefore, as
represented by first and second equations of the above
formula (3), the α-axis inductance Lα and the β-axis
inductance Lβ change according to the angle of rotation θ
of the rotor 2a.
10 [0049] Moreover, the differential operator P in the
above formula (2) acts on all of the α-axis inductance Lα,
the β-axis inductance Lβ, and the αβ-axes mutual inductance
Lαβ as well as the α-axis current iα and the β-axis current
iβ. Thus, the following formula is obtained by expanding
15 the terms of the differential operator in the above formula
(2).
[0050] [Formula 4]
 

 

  
   

 

  
 
   


 

  
  
 

 

 
  
    


 






 Pi
Pi
sin2 cos2
L L cos2 sin2 i
i
cos2 sin2
R 2 L sin2 cos2 v
v
s 1 0 1
(4)
20 [0051] In the above formula (4), “ω” represents the
rotational frequency of the rotor 2a, and ω=Pθ. The
energized state in FIG. 3 is equivalent to applying a
clockwise rotating magnetic field to the motor that is at
rest when viewed from the rotor 2a. Therefore, the
25 pulsation phase of the α-axis current iα and the β-axis
current iβ is such that the α-axis current iα is 90 degrees
ahead of the β-axis current iβ.
[0052] Moreover, when viewed from the stator, the
saliency of the inductance equally affects both of the α
30 and β axes, so that the amplitude of the pulsation of the
α-axis current iα and the amplitude of the pulsation of the
β-axis current iβ are equal. On the basis of these, the α-
17
axis current iα and the β-axis current iβ are divided into
an average value component (iα0, iβ0) and a pulsating
current component (iα1, iβ1) and are then represented by the
following formula.
5 [0053] [Formula 5]
 
        
       
   
    i : i i i icos 2
i : i i i isin 2
0 1 0
0 1 0
(5)
[0054] In the above formula (5), “φ” represents an
unknown phase angle, and “Δi” represents the amplitude of
the pulsating current. Here, the average values iα0 and iβ0
10 of the two-phase current are obtained by solving the
equation obtained by ignoring all the terms related to the
inductance component in the above formula (2). Since the
β-axis voltage vβ is zero as described above, the average
value iα0 of the α-axis current and the average value iβ0 of
15 the β-axis current are iα0=vα/Rs and iβ0=0, respectively.
[0055] Also, the following formula is obtained by
substituting the above formula (5) into the above formula
(4).
[0056] [Formula 6]
 
 
 
  


 

  
  
    


 

  
   
   


 


 
   sin 2
2 L i cos 2 cos 2
R i sin 2 cos2
0 2 L i sin2 1 0 s 0 20 (6)
[0057] The following formula is further obtained by
rearranging the formula in the first line of the above
formula (6).
[0058] [Formula 7]
                0 2 L i sin 2 i R 2 L 2 sin 2
25 1 0 s2 0 (7)
[0059] Note that in the above formula (7), “δ” is set as
in the following formula.
[0060] [Formula 8]
18
  tan12L0 Rs  (8)
[0061] On the basis of the condition that the above
formula (7) is an identity, the amplitude Δi of the
pulsating current and the unknown phase angle φ are
5 obtained as in the following formulas.
[0062] [Formula 9]
 2
s2 0
0 1 R 2 L
i i 2 L
 
    
(9)
[Formula 10]
     (10)
10 [0063] Here, in the above formula (9), the resistance
component is sufficiently smaller than the reactance
component. Therefore, the above formula (8) can be
approximated as δ=π/2, and the above formula (10) can be
approximated as φ=π/2. At this time, the α-axis pulsating
15 current iα1 and the β-axis pulsating current iβ1 can be
modified as in the following formula.
[0064] [Formula 11]
 
   
  

 i isin 2
i icos 2
1
1
(11)
[0065] The above formula (11) represents that by
20 extracting the pulsating currents from the values of the
two-phase current on the stationary coordinate system and
calculating the phases of those current values, the
rotational position of the rotor 2a can be estimated as a
result.
25 [0066] Specifically, the controller 3 of the first
embodiment illustrated in FIG. 1 performs the following
operations. First, the coordinate transformation unit 20
transforms phase currents acquired from the detector 4 into
an α-axis current 20a and a β-axis current 20b, which are
19
the two-phase current on the αβ two-phase coordinate system,
and outputs the α-axis current and β-axis current. A
transformation formula used at this time is, for example,
the above formula (1).
5 [0067] Next, the pulsation extraction unit 30 extracts
an α-axis pulsating current 30a and a β-axis pulsating
current 30b on the basis of the α-axis current 20a and the
β-axis current 20b, and outputs the α-axis pulsating
current and β-axis pulsating current to the phase
10 synchronization calculation unit 40. Note that in the
following description, the α-axis pulsating current 30a and
the β-axis pulsating current 30b will be collectively
referred to as “pulsating current” in some cases.
[0068] Then, the phase synchronization calculation unit
15 40 calculates an estimated pulsation phase 40a and an
estimated pulsation frequency 40b on the basis of the α-
axis pulsating current 30a and the β-axis pulsating current
30b, and outputs the estimated pulsation phase and
estimated pulsation frequency. The estimated pulsation
20 phase 40a and the estimated pulsation frequency 40b
calculated by the phase synchronization calculation unit 40
are converted into appropriate values, and these values
obtained by the conversion are used in the steady state
estimation algorithm (not shown).
25 [0069] Next, a description will be made of an advantage
of setting the voltage command value to a value with which
none of the phase currents equals zero, that is, a value
with which none of the u-phase current, the v-phase current,
and the w-phase current equals zero. Note that in the
30 following description, the u-phase current, the v-phase
current, and the w-phase current will be collectively
referred to as “three-phase current” in some cases.
[0070] First, a PWM signal for controlling the
20
semiconductor elements of the upper and lower arms of an
inverter, not limited to the inverter 1 of the first
embodiment, has a pause period in which an off command is
given to all of the semiconductor elements of the upper and
5 lower arms. This pause period is called dead time. The
pause period is set to ensure the prevention of short
circuit between the direct current buses 15a and 15b.
[0071] The semiconductor element also experiences a
voltage drop due to the physical properties of the
10 semiconductor element. When the semiconductor element is
an IGBT, there is a collector-emitter voltage drop called
saturation voltage. When the semiconductor element is a
MOSFET, there is a voltage drop due to the resistance
between the drain and the source.
15 [0072] In order to reduce a voltage error caused by the
above factors, the voltage command value is often corrected.
These corrections are called dead time correction, onvoltage
correction, and the like.
[0073] Also, when controlling the semiconductor element,
20 it is necessary to deal with the following characteristics
peculiar to the semiconductor element.
[0074] (1) In a small current region where the current
flowing through the semiconductor element is small, the
switching transient time of the semiconductor element
25 changes in a complicated manner.
(2) In the correction of the voltage command value,
the sign of a voltage correction amount is determined using
the polarity of the current, so that chattering is likely
to occur. Note that chattering is an event in which
30 reversal of the voltage correction amount and the polarity
of the current is repeated at an unexpectedly fast cycle.
(3) When a dead zone is set to prevent chattering, the
effect of correction is reduced.
21
[0075] What can be said in common with the above items
(1) to (3) is that the smaller the current flowing through
the semiconductor element, the more difficult it gets to
correct the voltage command value. That is, the voltage
5 error increases when a first phase current, which is one
phase current among the three-phase current, is small.
When the voltage error increases, the phase current of any
of the other phase currents other than the first phase
current is increased or decreased. As a result, the
10 average value iα0 of the α-axis current and the average
value iβ0 of the β-axis current become too small or too
large.
[0076] Here, as is clear from the above formula (9), the
amplitude Δi of the α-axis pulsating current iα1 and the β-
15 axis pulsating current iβ1 is proportional to the magnitude
of the average value iα0 of the α-axis current. When any of
the three-phase current is small, the average value iα0 of
the α-axis current can be too small, and the accuracy of
estimating the rotational position and rotational frequency
20 can be reduced.
[0077] Also, although not modeled by the above formula
(2), in the SynRM, the magnetic saturation of a magnetic
member progresses as the amount of energization increases.
Moreover, in the SynRM, the ease of magnetic saturation
25 differs significantly depending on the direction in which
the rotor is magnetized. Therefore, when the average value
iα0 of the α-axis current and the average value iβ0 of the
β-axis current become excessive to cause an increase in the
degree of magnetic saturation, the pulsating current being
30 detected includes harmonics, so that the accuracy of
estimating the rotational position and rotational frequency
is reduced.
[0078] On the basis of the above points, the voltage
22
control unit 50 according to the first embodiment
calculates and outputs the voltage command value at which
no phase current of the three-phase current equals zero.
This facilitates the correction of the voltage command
5 value for setting the output voltage of the inverter 1 to a
desired value. As a result, the current of the motor 2 can
be controlled to an appropriate magnitude, and thus the
accuracy of estimating the rotational position and
rotational frequency of the rotor 2a can be increased.
10 [0079] Note that when a voltage is applied to the motor
2 that is rotating, the pulsating current due to saliency
can be observed regardless of the magnitude and phase of
the voltage. However, it is desirable that the voltage
command value output by the voltage control unit 50 be the
15 direct current voltage. The reason is as follows.
[0080] It is assumed, for example, that the voltage for
the initial estimation includes an alternating current
component with a frequency of “f”. Then, an alternating
current component with the frequency of “f” is also
20 generated in the phase current of the motor. That is, the
phase current includes a pulsating component synchronized
with the rotational frequency of the motor and the
frequency component having the same frequency as the
voltage. When a plurality of frequency components are
25 mixed in the phase current, the separation thereof is
difficult, so that the accuracy of the initial estimation
is reduced. It is thus desirable that the voltage for the
initial estimation, that is, the voltage command value
output by the voltage control unit 50, be the direct
30 current voltage.
[0081] Next, an advantage of setting the direction of
the voltage vector of the voltage command value to be the
same as or opposite to that of any of the phases of the
23
motor 2 will be described with reference to FIG. 4. FIG. 4
is a graph illustrating a relationship between the phase of
the voltage command value vector and the magnitude of the
phase current in the first embodiment. Note that the
5 “direction of the voltage vector” may be rephrased as the
“phase of the voltage vector”, and the “same or opposite
direction” may be rephrased as the “same or opposite phase”.
In this case, the same or opposite phase means that when
the phase is 60 [deg], for example, the “same phase” is 60
10 [deg], and the “opposite phase” is 240 [deg] which is
obtained by adding 180 [deg] thereto.
[0082] A horizontal axis of FIG. 4 represents the phase
of the voltage command value vector with respect to the α-
axis, and a vertical axis represents the amplitudes of
15 various phase currents normalized. Specifically, a dotted
line represents the average value iu0 of the u-phase
current, a thin solid line represents the average value iv0
of the v-phase current, a dot dashed line represents the
average value iw0 of the w-phase current, a thin broken
20 line represents the average value iα0 of the α-axis current,
and a thick broken line represents the average value iβ0 of
the β-axis current.
[0083] Moreover, the waveform of a thick solid line is
obtained by drawing a waveform portion having the smallest
25 absolute value among the average values (iu0, iv0, iw0) of
the phase currents. Here, the phase having the smallest
absolute value among the average values (iu0, iv0, iw0) of
the phase currents is defined as a “minimum phase”. Also,
the current of the minimum phase is defined as a “minimum
30 phase current”.
[0084] In the energized state of FIG. 4, the phase of
the voltage command vector corresponds to zero. At this
time, the minimum phase is the v-phase or the w-phase, and
24
the absolute value of the minimum phase current is “0.5”.
Note that it can be understood from the waveform of the
thick solid line in FIG. 4, that is, the waveform of the
minimum phase current, that the value of “0.5” is the
5 possible maximum value when the phase of the voltage
command value vector is changed.
[0085] As described above, the smaller the phase current,
the more difficult it is to correct the voltage command
value for reducing the voltage error. In FIG. 3, the
10 direction of the voltage command value vector is set to be
the same as that of the u-phase in order to “maximize the
minimum phase current”. Moreover, according to the
waveform of the minimum phase current in FIG. 4, the
maximum value occurs in 60 [deg] increments from 0 [deg].
15 That is, according to FIG. 4, it can be seen that the
voltage command value vector need only be oriented in the
same direction as or in the opposite direction to any of
the u-phase, the v-phase, and the w-phase.
[0086] For example, when the phase of the voltage
20 command value vector is 60 [deg], the voltage command value
vector is in the opposite direction to the v-phase. Also,
for example, when the phase of the voltage command value
vector is 120 [deg], the voltage command value vector is in
the same direction as the w-phase.
25 [0087] However, when the phase of the voltage command
value vector is other than zero, the phase angle φ defined
by the above formula (5) is different from the value
represented by the above formula (10). Therefore, when the
phase of the voltage command value vector is other than
30 zero, appropriate correction is required in the processing
by the phase synchronization calculation unit 40.
[0088] As described above, the voltage control unit
according to the first embodiment calculates and outputs
25
the direct current voltage command value such that the
phase of the voltage command value vector is in the same
direction as or opposite direction to any phase of the
motor. This facilitates the correction of the voltage
5 command value for setting the output voltage of the
inverter to a desired value. As a result, the motor
current can be controlled to an appropriate magnitude, and
thus the accuracy of estimating the rotational position and
rotational frequency of the rotor can be increased.
10 [0089] Second Embodiment.
In a second embodiment, detailed configurations and
operations of the pulsation extraction unit 30 and the
phase synchronization calculation unit 40 illustrated in
FIG. 1 will be described.
15 [0090] FIG. 5 is a block diagram illustrating a
configuration of the pulsation extraction unit 30 according
to the second embodiment. As illustrated in FIG. 5, the
pulsation extraction unit 30 according to the second
embodiment includes two high pass filters (HPFs) 301 and
20 302 having the same characteristics. The α-axis current
20a and the β-axis current 20b, which are the two-phase
current in the stationary coordinate system, are input to
the high pass filters 301 and 302, respectively.
[0091] The time required for removing the direct current
25 component included in the α-axis current 20a and the β-axis
current 20b depends on a cutoff frequency of the high pass
filters 301 and 302. The higher the cutoff frequency, the
shorter the time required for removing the direct current
component, and the estimation calculation by the phase
30 synchronization calculation unit 40, which will be
described later, can be started quickly. However, the
frequency of the pulsating current that needs to be
extracted changes along with the rotational frequency of
26
the rotor 2a. Thus, if the cutoff frequency is too high,
even the amplitude of the pulsating current is attenuated
to possibly reduce the S/N ratio, so that attention is
required in the design.
5 [0092] Next, the phase synchronization calculation unit
40 will be described. FIG. 6 is a block diagram
illustrating a configuration of the phase synchronization
calculation unit 40 according to the second embodiment. As
illustrated in FIG. 6, the phase synchronization
10 calculation unit 40 according to the second embodiment
includes a phase error calculation unit 401, an amplifier
402, and an integrator 403.
[0093] The phase error calculation unit 401 receives the
α-axis pulsating current 30a (iα1), the β-axis pulsating
15 current 30b (iβ1), and the estimated pulsation phase 40a
(θ^2). The estimated pulsation phase 40a is the output of
the integrator 403. The notation “θ^2” is an alternative
notation to one in which a hat symbol “^” is attached to
the top of the character “θ” in “θ2”. In the present
20 specification, the alternative notation is used except for
the mathematical formulas inserted as images. The similar
applies to “ω^2” described later.
[0094] The phase error calculation unit 401 calculates a
phase error 40f (Δθ2) according to the following formula.
25 [0095] [Formula 12]
2 1 2 1 2  2   2   : i sinˆ  i cosˆ  isin 2  ˆ  i 2  ˆ (12)
[0096] The amplifier 402 amplifies the phase error 40f
(Δθ2) and outputs the estimated pulsation frequency 40b
(ω^2). As indicated in FIG. 6, the amplifier 402 is
30 preferably a proportional integral (PI) controller that
performs proportional integral.
[0097] The integrator 403 integrates the estimated
27
pulsation frequency 40b and outputs the integrated value as
the estimated pulsation phase 40a. The estimated pulsation
phase 40a is fed back to the phase error calculation unit
401.
5 [0098] In the above formula (12), when 2θ>θ^2, “Δθ2” is
positive, so that the estimated pulsation frequency ω^2 and
the estimated pulsation phase θ^2 are corrected to be
increased. On the contrary, when 2θ<θ^2, “Δθ2” is negative,
so that the estimated pulsation frequency ω^2 and the
10 estimated pulsation phase θ^2 are corrected to be decreased.
Eventually, 2θ becomes equal to θ^2 and thus the phase and
frequency of the pulsating current are estimated. The
phase synchronization calculation unit 40 thus takes the
form of a phase locked loop (PLL).
15 [0099] Now, when the phase of the voltage command value
vector is set to zero as in the first embodiment, the
average value (iβ0) of the β-axis current 20b equals zero
as illustrated in FIG. 4. Therefore, at first glance, it
seems unnecessary to use the high pass filter 302 to
20 extract the β-axis pulsating current 30b (iβ1) from the β-
axis current 20b. However, when there is a difference in
the presence/absence and characteristics of filtering for
pulsating current extraction between the α-axis and the β-
axis, the amplitude and phase of the pulsating current to
25 be extracted differ between the α-axis and the β-axis.
Therefore, the two high pass filters 301 and 302 are
required regardless of the phase of the voltage command
value vector. It is also desirable that both have the same
characteristics. Note that the same characteristics in
30 this case do not mean that the physical characteristics are
completely the same, but mean that they are designed and
configured with the expectation of having the same
characteristics.
28
[0100] A method is also conceivable in which either the
α-axis pulsating current 30a or the β-axis pulsating
current 30b is used to calculate the pulsation frequency
from the zero crossing interval and the pulsation phase
5 from the zero crossing timing. However, since the signal
from the detector 4 contains noise that enters the circuit
of the detector 4, the zero crossing timing cannot always
be detected accurately. Moreover, during high-speed
rotation of the motor 2, the number of samplings per cycle
10 with which the current pulsates decreases, so that an error
of the zero crossing timing becomes more remarkable.
Furthermore, since the pulsating current includes low-order
harmonics due to the influence of the magnetic saturation
and spatial harmonics of the motor 2, the relationship
15 between the zero crossing timing of the pulsating current
and the position of the rotor 2a becomes more complicated.
[0101] On the other hand, the phase synchronization
calculation unit 40 according to the second embodiment
includes a circuit corresponding to a PLL including a
20 feedback path. When the circuit corresponding to the PLL
is configured as in FIG. 6, the integrator 403 is included
in the path before the estimated pulsation phase 40a is
obtained, so that it is not easily affected by the noise
entering the signal path of the detector 4. Moreover, the
25 estimated pulsation phase 40a is continuously calculated so
as to follow the true pulsation phase. As a result, even
if the number of samplings per cycle with which the current
pulsates is small, it is easy to correct an error caused by
discretization. Furthermore, even when disturbances such
30 as the magnetic saturation and spatial harmonics are
contained in the pulsating current, the estimated pulsation
phase converges to the true phase on average.
[0102] As described above, the pulsation extraction unit
29
30 according to the embodiment includes the two high pass
filters that remove the direct current component from the
two-phase current in the stationary coordinate system and
output the pulsating currents. The two high pass filters
5 have the same characteristics. Moreover, the phase
synchronization calculation unit according to the second
embodiment includes the phase error calculation unit that
calculates the phase error on the basis of the pulsating
current and the estimated pulsation phase, the amplifier
10 that amplifies the phase error and outputs the estimated
pulsation frequency, and the integrator that integrates the
estimated pulsation frequency and outputs it as the
estimated pulsation phase. This makes it less susceptible
to various disturbances, and thus the rotational position
15 and rotational frequency of the rotor can be estimated with
high accuracy.
[0103] Third Embodiment.
In a third embodiment, switching of the operation mode
when starting the inverter 1 will be described.
20 [0104] First, as described in the second embodiment, the
pulsation extraction unit 30 requires time corresponding to
the cutoff frequency of the high pass filter in order to
remove the direct current component from the two-phase
current in the stationary coordinate system.
25 [0105] Moreover, the estimated pulsation frequency does
not converge while the direct current component remains in
the signal input to the phase synchronization calculation
unit 40. When the amplifier 402 and the integrator 403 of
the phase synchronization calculation unit 40 start the
30 calculation while the direct current component remains, the
estimated pulsation frequency 40b (ω^2) and the estimated
pulsation phase 40a (θ^2) diverge or oscillate to
inaccurate values. As a result, the time required for the
30
initial estimation gets long. It is thus desirable that
the amplifier 402 and the integrator 403 start the
calculation at a point when a required time elapses after
the inverter 1 starts energization.
5 [0106] A phase synchronization calculation unit 41
according to the third embodiment is configured as
illustrated in FIG. 7. FIG. 7 is a block diagram
illustrating the configuration of the phase synchronization
calculation unit 41 according to the third embodiment.
10 Comparing FIG. 7 with FIG. 6, a gain switching signal 40c
that controls switching of the gain of the amplifier 402 is
added in FIG. 7.
[0107] In the phase synchronization calculation unit 41
of FIG. 7, the gain representing an amplification factor of
15 the amplifier 402 is set to zero immediately after the
inverter 1 starts energization, and after a first time
elapses since the inverter 1 starts energization, the gain
of the amplifier 402 is switched to a non-zero value, that
is, a value larger than zero. When the gain of the
20 amplifier 402 is set to zero, the estimated pulsation
frequency 40b remains zero no matter what phase error 40f
is input to the amplifier 402. As a result, the input to
the integrator 403 also equals zero, and the estimated
pulsation phase 40a also remains zero.
25 [0108] The time from when the inverter 1 starts
energization until the gain of the amplifier 402 is first
switched is determined on the basis of the cutoff frequency
of the high pass filters 301 and 302 in the pulsation
extraction unit 30. More specifically, when the cutoff
30 frequency of the high pass filters 301 and 302 is high,
relatively short time is required for removing the direct
current component, so that the amplifier 402 can start the
calculation within a relatively short time after the
31
inverter 1 starts energization. On the contrary, when the
cutoff frequency of the high pass filters 301 and 302 is
low, relatively long time is required for removing the
direct current component, so that a relatively long grace
5 period needs to be set before the amplifier 402 starts the
calculation.
[0109] The gain of the amplifier 402 needs to be
relatively large from the time when the amplifier 402
starts the calculation until the estimated pulsation phase
10 40a converges to a first value close to the true value. On
the other hand, the required gain is not that large once
the estimated pulsation phase 40a reaches the first value.
The required gain of the amplifier 402 depends on whether
or not the approximation of sin (2θ-θ^2)≈2θ-θ^2 holds in
15 the definitional equation of the phase error represented by
the above formula (12). That is, when 2θ being the true
value of the phase of the pulsating current and the
estimated pulsation phase θ^2 are approximately close to
each other, a relatively small gain is required to cause
20 the estimated pulsation phase θ^2 to follow the true value
2θ. On the contrary, when the difference between the true
value 2θ and the estimated pulsation phase θ^2 is large, a
relatively large gain is required to converge the estimated
pulsation phase θ^2 to the true value 2θ.
25 [0110] On the other hand, as described in the second
embodiment, the α-axis pulsating current 30a and the β-axis
pulsating current 30b include the low-order harmonic
components due to the influence of the spatial harmonics of
the motor 2 and the magnetic saturation of the motor 2.
30 Since such harmonics are also amplified by the amplifier
402, the estimated pulsation phase θ^2 becomes slightly
oscillatory. When the estimated pulsation phase θ^2 is
oscillatory, an error from the true value can be large
32
depending on the timing of holding an estimation result.
It is thus desirable that the gain of the amplifier 402 is
minimized in terms of improving the accuracy of the initial
estimation.
5 [0111] Therefore, in the phase synchronization
calculation unit 41 of FIG. 7, the characteristics are
switched such that the gain of the amplifier 402 is reduced
after a second time elapses since the inverter 1 starts
energization. The second time is longer than the first
10 time. However, the gain after switching is assumed to be
larger than zero.
[0112] Note that there may also be used a method of
holding a plurality of constants in the amplifier 402 and
selecting any of the constants on the basis of the gain
15 switching signal 40c. Alternatively, the gain switching
signal 40c itself may be a signal including the constant
itself that determines the gain of the amplifier 402.
[0113] Next, the operation when the gain of the
amplifier 402 is switched according to the elapsed time
20 will be described with reference to FIG. 8. FIG. 8 is a
graph for explaining the operation of switching the gain of
the amplifier according to the third embodiment.
[0114] FIG. 8 illustrates various waveform examples when
the gain of the amplifier 402 is switched. More
25 specifically, a first tier of FIG. 8 illustrates the uphase
current by a dot dashed line, the v-phase current by
a broken line, and the w-phase current by a solid line. A
second tier of FIG. 8 illustrates the α-axis current by a
solid line and the β-axis current by a broken line. A
30 third tier of FIG. 8 illustrates the α-axis pulsating
current by a solid line and the β-axis pulsating current by
a broken line. A fourth tier of FIG. 8 illustrates the
estimated pulsation frequency by a solid line and the true
33
frequency by a broken line. A fifth tier of FIG. 8
illustrates the estimated pulsation phase by a solid line
and the true phase by a broken line.
[0115] In FIG. 8, gate start is performed first at time
5 t0 in which the inverter 1 starts outputting the voltage.
Once the gate start is performed by the inverter 1, the
three-phase current starts to flow. At this time, as
represented by the waveforms in the first tier, the
pulsating currents proportional to the magnitude of the
10 direct current component are superimposed. Moreover, the
three-phase current in the first tier is transformed into
the two-phase current in the stationary coordinate system,
which is represented by the waveforms in the second tier.
Furthermore, the two-phase current in the second tier is
15 input to the pulsation extraction unit 30, and a result of
extracting the pulsating currents is represented by the
waveforms in the third tier.
[0116] Referring to the waveform of the α-axis pulsating
current in the third tier, it can be seen that the direct
20 current component is removed after a lapse of a certain
time between time t0 and time t1. At time t1, the gain of
the amplifier 402 is switched from zero to a positive value,
and the estimation calculation by the phase synchronization
calculation unit 41 is started. The time from time t0 to
25 time t1 is the time corresponding to the first time
described above. At the time when the estimation
calculation is started, the direct current component of the
two-phase current is sufficiently removed, so that the gain
of the amplifier 402 is set relatively high. Therefore,
30 the estimated pulsation frequency 40b and the estimated
pulsation phase 40a quickly converge to values close to the
true values as illustrated by the waveforms in the fourth
and fifth tiers.
34
[0117] Next, at time t2, the gain of the amplifier 402
is switched to be decreased. This as a result reduces the
pulsation of the estimated pulsation frequency 40b and the
estimated pulsation phase 40a. The time from time t0 to
5 time t2 is the time corresponding to the second time
described above. Finally, at time t3, the estimation
result is held, and the steady state estimation algorithm
is started using the result.
[0118] As described above, the phase synchronization
10 calculation unit according to the third embodiment switches
the gain of the amplifier from zero to the value larger
than zero after the first time elapses since the inverter
starts energization. As a result, the time required for
the initial estimation can be reduced. Moreover, the phase
15 synchronization calculation unit according to the third
embodiment switches the gain of the amplifier to be
decreased after the second time elapses since the inverter
starts energization. As a result, the degree of
amplification of the harmonics included in the pulsating
20 current is reduced, and the accuracy of the estimation
result is improved.
[0119] Fourth Embodiment.
Next, a drive apparatus of a synchronous motor
according to a fourth embodiment will be described. FIG. 9
25 is a diagram of a configuration of a motor drive apparatus
according to the fourth embodiment. A motor drive
apparatus 101 illustrated in FIG. 9 is obtained by adding a
correction calculation unit 60 to the controller 3 in the
configuration of the motor drive apparatus 100 according to
30 the first embodiment illustrated in FIG. 1. Also, due to
the addition of the correction calculation unit 60, the
controller 3 is illustrated as a controller 3A. Note that
the other configurations are identical or equivalent to
35
those of FIG. 1 and are thus denoted by the same reference
numerals as those in FIG. 1, whereby a redundant
description will be omitted.
[0120] The output of the pulsation extraction unit 30 of
5 the controller 3 has the phase that is advanced as compared
with the pulsating current included in the original twophase
current. The degree of advancement of the phase
depends on the cutoff frequency of the high pass filter and
the frequency of the pulsating current, that is, the
10 rotational frequency of the rotor 2a. Therefore, the phase
synchronization calculation unit 41 according to the third
embodiment performs the estimation calculation on the
signal whose pulsation phase is advanced. As a result, an
error is included in the rotational position of the rotor
15 2a obtained by converting the output of the phase
synchronization calculation unit 41. In the fourth
embodiment, a method of eliminating the error will be
described below in detail.
[0121] In FIG. 9, the correction calculation unit 60
20 calculates and outputs an estimated rotor phase 60a and an
estimated rotor frequency 60b on the basis of the estimated
pulsation phase 40a and the estimated pulsation frequency
40b being the output of the phase synchronization
calculation unit 41.
25 [0122] Next, a detailed configuration of the correction
calculation unit 60 will be described with reference to FIG.
10. FIG. 10 is a block diagram illustrating a
configuration of the correction calculation unit according
to the fourth embodiment. As illustrated in FIG. 10, the
30 correction calculation unit 60 according to the fourth
embodiment includes a low pass filter (LPF) 601, a look-up
table 602, a subtractor 603, and a conversion unit 604.
[0123] The low pass filter 601 has a high frequency
36
cutoff characteristic and smooths the estimated pulsation
frequency 40b to output a smoothed pulsation frequency 60d.
The look-up table 602 outputs a phase correction amount 60e
on the basis of the smoothed pulsation frequency 60d. The
5 subtractor 603 subtracts the phase correction amount 60e
from the estimated pulsation phase 40a, and outputs a
result of the subtraction as a corrected pulsation phase
60c. The conversion unit 604 converts the corrected
pulsation phase 60c with a constant to obtain the estimated
10 rotor phase 60a, and converts the smoothed pulsation
frequency 60d with a constant to obtain the estimated rotor
frequency 60b.
[0124] The estimated pulsation frequency 40b pulsates
due to the influence of the magnetic saturation and spatial
15 harmonics of the motor 2. In the third embodiment, the
method is described in which the pulsation is reduced by
switching the gain of the amplifier 402 to be decreased
according to the elapsed time. However, switching the gain
cannot completely remove the pulsation. Thus, in order to
20 further improve the accuracy, the estimated pulsation
frequency 40b is smoothed by using the low pass filter 601.
[0125] The low pass filter may be of the type such as a
filter in which a transfer function is a temporary lag, or
may perform an operation of averaging the input signal over
25 a set time. Note that the higher the high frequency cutoff
performance of the low pass filter, the more powerfully the
pulsation of the estimated pulsation frequency 40b can be
removed, but the settling time for the smoothed pulsation
frequency 60d gets longer. In other words, the
30 characteristics of the low pass filter 601 need to be
determined such that the smoothed pulsation frequency 60d
is settled within a target time for the initial estimation.
Here, the “target time for the initial estimation” is the
37
time from when the inverter 1 starts energization to when
the steady state estimation is started.
[0126] Moreover, the look-up table 602 is determined
according to the phase characteristics of the high pass
5 filters 301 and 302 in the pulsation extraction unit 30.
The look-up table 602 holds data indicating how much the
phase of the signal that has passed through the pulsation
extraction unit 30 changes in accordance with the frequency
of the signal. In the subtractor 603, the phase correction
10 amount 60e is subtracted from the estimated pulsation phase
40a. As a result, the corrected pulsation phase 60c
exactly matches the pulsation phase of the two-phase
current before being processed by the high pass filters 301
and 302.
15 [0127] Finally, a purpose of the conversion unit 604
will be described. The corrected pulsation phase 60c and
the smoothed pulsation frequency 60d are the phase and
frequency of the pulsation component superimposed on the
two-phase current. As represented by the above formula
20 (11), the two-phase current pulsates at a frequency twice
the angle of rotation θ of the rotor 2a. Therefore, in
order to obtain information on the rotational position and
rotational frequency of the rotor 2a, the phase and
frequency of the pulsating current each need only be
25 multiplied by 0.5.
[0128] Moreover, in the fourth embodiment, the units of
the estimated rotor phase 60a and the estimated rotor
frequency 60b are not particularly limited. Furthermore,
the conversion unit 604 may perform conversion between a
30 mechanical angle and an electrical angle, conversion
between degree measure and radian measure, and the like at
the same time depending on the configuration of the steady
state estimation algorithm (not shown).
38
[0129] As described above, the controller according to
the fourth embodiment includes the low pass filter that
smooths the estimated pulsation frequency. As a result,
the harmonics included in the estimated pulsation frequency
5 are further reduced, and the estimation accuracy is
improved. The controller according to the fourth
embodiment further includes the look-up table to reference
the phase correction amount on the basis of the output of
the low pass filter, and corrects the estimated pulsation
10 phase with the phase correction amount by referring to the
look-up table. The look-up table is determined on the
basis of the frequency-phase characteristics of the high
pass filter included in the pulsation extraction unit. As
a result, the advancement of the phase of the signal in the
15 pulsation extraction unit is corrected, and the phase of
the pulsating current is accurately obtained. Therefore,
accurate information on the rotational position of the
rotor can be obtained.
[0130] Note that in the fourth embodiment, the example
20 is described in which the method of smoothing the estimated
pulsation frequency and correcting the estimated pulsation
phase using the look-up table is applied to the third
embodiment, but the present invention is not limited
thereto. Similar correction may also be applied to the
25 second embodiment, which can obtain a similar effect.
[0131] Moreover, in the first to fourth embodiments, the
description has been made on the assumption that the motor
2 is the SynRM, but another type of motor may be used. For
example, the motor 2 may be an interior permanent magnet
30 synchronous motor (IPMSM). As described above, in the
first to fourth embodiments, the information on the
rotational position and rotational frequency of the rotor
2a is estimated by using the fact that when the direct
39
current voltage is applied to the motor 2 that is rotating,
the motor current pulsates at twice the frequency due to
saliency. Therefore, the IPMSM that is designed to be able
to obtain not only magnet torque but also reluctance torque
5 can apply the initial estimation algorithm according to the
first to fourth embodiments.
[0132] Fifth Embodiment.
Next, a hardware configuration for implementing the
arithmetic functions of the controller 3A of the fourth
10 embodiment will be described with reference to FIGS. 11 and
12. FIG. 11 is a block diagram illustrating an example of
the hardware configuration that implements the arithmetic
functions of the controller of the fourth embodiment. FIG.
12 is a block diagram illustrating another example of the
15 hardware configuration that implements the arithmetic
functions of the controller of the fourth embodiment.
[0133] When some or all of the arithmetic functions of
the controller 3A of the fourth embodiment are implemented
by software, as illustrated in FIG. 11, the configuration
20 can include a processor 90 that performs an arithmetic
operation, a memory 91 that saves programs to be read by
the processor 90, and an interface 92 that inputs and
outputs signals.
[0134] The processor 90 may be arithmetic means such as
25 an arithmetic unit, a microprocessor, a microcomputer, a
central processing unit (CPU), or a digital signal
processor (DSP). The memory 91 can include, for example, a
non-volatile or volatile semiconductor memory such as a
random access memory (RAM), a read only memory (ROM), a
30 flash memory, an erasable programmable ROM (EPROM), or an
electrically EPROM (EEPROM (registered trademark)), a
magnetic disk, a flexible disk, an optical disk, a compact
disc, a mini disc, or a digital versatile disc (DVD).
40
[0135] The memory 91 stores programs for executing all
or some of the arithmetic functions of the controller 3A.
The processor 90 transmits and receives necessary
information via the interface 92 and executes the programs
5 stored in the memory 91 to be able to perform PWM control
on the inverter 1 and the initial estimation that estimates
the rotational position and rotational frequency of the
motor 2.
[0136] Moreover, the processor 90 and the memory 91
10 illustrated in FIG. 11 may be replaced with a processing
circuit 93 as in FIG. 12. The processing circuit 93
corresponds to a single circuit, a complex circuit, an
application specific integrated circuit (ASIC), a field
programmable gate array (FPGA), or a combination of those.
15 [0137] As described above, in a fifth embodiment, the
hardware configuration for implementing the arithmetic
functions of the controller 3A of the fourth embodiment has
been described, but the present invention is not limited
thereto. Needless to say, the controller 3 of the first to
20 third embodiments can also be implemented with a similar
hardware configuration.
[0138] The configurations illustrated in the
aforementioned embodiments merely illustrates examples of
the content of the present invention, and can thus be
25 combined with another known technique or partially omitted
and/or modified without departing from the scope of the
present invention.
Reference Signs List
30 [0139] 1 inverter; 2 motor; 2a rotor; 3, 3A
controller; 4 detector; 10A, 10B, 10C leg; 10a
transistor; 10b diode; 12, 13, 14 connection point; 20
coordinate transformation unit; 20a α-axis current; 20b
41
β-axis current; 30 pulsation extraction unit; 301, 302
high pass filter; 30a α-axis pulsating current; 30b β-
axis pulsating current; 40, 41 phase synchronization
calculation unit; 401 phase error calculation unit; 402
5 amplifier; 403 integrator; 40a estimated pulsation phase;
40b estimated pulsation frequency; 40c gain switching
signal; 40f phase error; 50 voltage control unit; 60
correction calculation unit; 601 low pass filter; 602
look-up table; 603 subtractor; 604 conversion unit; 60a
10 estimated rotor phase; 60b estimated rotor frequency; 60c
corrected pulsation phase; 60d smoothed pulsation
frequency; 60e phase correction amount; 90 processor; 91
memory; 92 interface; 93 processing circuit; 100, 101
motor drive apparatus; 110 power source.
15
42
We Claim:
1. A motor drive apparatus comprising:
an inverter to drive a synchronous motor with
5 saliency;
a controller to control an operating state of the
inverter; and
a detector to detect phase currents of the synchronous
motor, wherein
10 the controller comprises:
a voltage control unit to determine an output voltage
of the inverter;
a coordinate transformation unit to transform the
phase currents into a two-phase current in a stationary
15 coordinate system;
a pulsation extraction unit to extract a pulsating
current from the two-phase current; and
a phase synchronization calculation unit to estimate
and calculate a frequency and a phase of the pulsating
20 current, and
the voltage control unit outputs a voltage command
value at which none of the phase currents equals zero.
2. The motor drive apparatus according to claim 1,
25 wherein
the voltage command value is a direct current voltage.
3. The motor drive apparatus according to claim 2,
wherein
30 a direction of a voltage vector of the voltage command
value is same as or opposite to any phase of the
synchronous motor.
43
4. The motor drive apparatus according to any one of
claims 1 to 3, wherein
the pulsation extraction unit includes two high pass
filters to remove a direct current component of the two5
phase current, and
the two high pass filters have a same characteristic.
5. The motor drive apparatus according to claim 4,
wherein
10 the phase synchronization calculation unit comprises:
a phase error calculation unit to calculate a phase
error on a basis of the pulsating current and an estimated
pulsation phase;
an amplifier to amplify the phase error and output an
15 estimated pulsation frequency; and
an integrator to integrate the estimated pulsation
frequency and output a result of integration as the
estimated pulsation phase.
20 6. The motor drive apparatus according to claim 5,
wherein
a gain of the amplifier is switched from zero to a
value larger than zero after a first time elapses since the
inverter starts energization.
25
7. The motor drive apparatus according to claim 5 or 6,
wherein
a gain of the amplifier is changed to be decreased
after a second time elapses since the inverter starts
30 energization.
8. The motor drive apparatus according to any one of
claims 5 to 7, wherein
44
the controller comprises:
a low pass filter to smooth the estimated pulsation
frequency; and
a look-up table to reference a phase correction amount
5 on a basis of output of the low pass filter, and
the estimated pulsation phase is corrected by the
phase correction amount to obtain an estimated rotor phase,
and the output of the low pass filter is multiplied by a
constant to obtain an estimated rotor frequency.
10
9. The motor drive apparatus according to claim 8,
wherein
the look-up table is determined on a basis of a
frequency-phase characteristic of the high pass filters.
15
10. The motor drive apparatus according to any one of
claims 1 to 9, wherein
the voltage control unit outputs the voltage command
value at which none of a plurality of the phase currents
20 equals zero during a period of initial estimation that is
executed at start-up of the inverter and estimates a
rotational position and a frequency of the synchronous
motor.