FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
[See section 10, Rule 13]
SPEED ESTIMATING DEVICE FOR AC MOTOR, DRIVING DEVICE FOR AC
MOTOR, REFRIGERANT COMPRESSOR, AND REFRIGERATION CYCLE
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 100-8310, JAPAN
THE FOLLOWING SPECIFICATION PARTICULARLY DESCRIBES THE
INVENTION AND THE MANNER IN WHICH IT IS TO BE PERFORMED.
2
DESCRIPTION
SPEED ESTIMATING DEVICE FOR AC MOTOR, DRIVING DEVICE FOR AC
MOTOR, REFRIGERANT COMPRESSOR, AND REFRIGERATION CYCLE
5 APPARATUS
Field
[0001] The present invention relates to a speed
estimating device for an alternating-current (AC) motor
10 which estimates the velocity of an AC motor such as an
induction machine or a synchronous machine, a driving
device for an AC motor, a refrigerant compressor, and a
refrigeration cycle apparatus.
15 Background
[0002] In control of an AC motor, when load torque or
torque generated by the AC motor has pulsation, some
pulsation also occurs in the rotational speed of the AC
motor. The occurrence of pulsation in the rotational speed
20 of the AC motor may cause vibration in an apparatus on
which the AC motor is installed, which may be problematic
in occurrence of noise, mechanical strength, or the like.
In response to these problems, control for reducing torque
pulsation and speed pulsation has been considered.
25 [0003] For example, Patent Literature 1 listed below
teaches a technique for achieving control for reducing
torque pulsation and speed pulsation by a sensorless
approach without using a position sensor or a velocity
sensor in order to reduce cost or to allow application to
30 an apparatus on which it is difficult to install a sensor.
According to Patent Literature 1, a torque compensation
value is obtained on the basis of a speed ripple component
extracted from a difference between a command angular
3
frequency and a speed feedback angular frequency. The
fluctuation in speed of the AC motor is thus reduced or
suppressed without any map of correction amounts.
5 Citation List
Patent Literature
[0004] Patent Literature 1: Japanese Patent No. 6222417
Summary
10 Technical Problem
[0005] In position-sensorless control, an upper limit of
speed estimation response in conventional control systems
is several hundreds [rad/s], and so a response to highfrequency pulsation is insufficient, and it is difficult to
15 accurately estimate a pulsation. In addition, in Patent
Literature 1, because a vibration reducing unit is
configured to use an estimated speed, the performance of
the vibration reducing unit depends on a response in the
speed estimation, and is thus considered to be insufficient
20 in a high-frequency range.
[0006] The present invention has been made in view of
the above circumstances, and an object thereof is to
provide a speed estimating device for an AC motor, which is
capable of further increasing the accuracy of speed
25 estimation at high frequency in sensorless control of an AC
motor.
Solution to Problem
[0007] In order to solve the aforementioned problems and
30 achieve the object, the present invention provides a speed
estimating device for an AC motor, the speed estimating
device comprising: a model deviation computing unit to
compute a model deviation on the basis of a voltage, a
4
current, and an estimated angular velocity of the AC motor;
a first angular velocity estimating unit to compute a first
estimated angular velocity on the basis of the model
deviation; a second angular velocity estimating unit to
5 compute a second estimated angular velocity on the basis of
the model deviation, the second estimated angular velocity
differing from the first estimated angular velocity in
frequency; a compensation phase computing unit to compute a
compensation phase on the basis of a disturbance frequency;
10 and an estimated angular velocity calculator to compute an
estimated angular velocity of the AC motor on the basis of
the first and second estimated angular velocities, wherein
either one of the first and second estimated angular
velocities is computed on the basis of the compensation
15 phase.
Advantageous Effects of Invention
[0008] The speed estimating device for an AC motor
according to the present invention produces an advantageous
20 effect that it is capable of further increasing the
accuracy of speed estimation at high frequency in
sensorless control on an AC motor.
Brief Description of Drawings
25 [0009]
FIG. 1 is a block diagram illustrating a configuration
of a speed estimating device for an AC motor according to a
first embodiment.
FIG. 2 is a block diagram illustrating a configuration
30 of a speed estimating device according to a comparative
example.
FIG. 3 is a Bode plot illustrating transfer
characteristics of the speed estimating device illustrated
5
in FIG. 2.
FIG. 4 is a block diagram illustrating a configuration
of a speed estimating device according to a comparative
example different from that in FIG. 2.
5 FIG. 5 is a Bode plot illustrating transfer
characteristics of the speed estimating device illustrated
in FIG. 4.
FIG. 6 is a graph used for comparing Bode plots
illustrating transfer characteristics from a model
10 deviation ε to a first estimated angular acceleration of
the speed estimating device illustrated in FIG. 2 in an
open loop characteristic and a closed loop characteristic.
FIG. 7 is a block diagram illustrating a detailed
configuration of a second angular velocity estimating unit
15 in the speed estimating device illustrated in FIG. 1.
FIG. 8 is a block diagram illustrating a modification
of the detailed configuration illustrated in FIG. 7.
FIG. 9 is a first graph made available for explaining
the effect of the speed estimating device according to the
20 first embodiment.
FIG. 10 is a second graph made available for
explaining the effect of the speed estimating device
according to the first embodiment.
FIG. 11 is a hardware configuration diagram of the
25 speed estimating device according to the first embodiment.
FIG. 12 is a block diagram illustrating a
configuration of a speed estimating device according to a
second embodiment.
FIG. 13 is a block diagram illustrating a
30 configuration of a speed estimating device according to a
third embodiment.
FIG. 14 is a block diagram illustrating a
configuration of a speed estimating device according to a
6
fourth embodiment.
FIG. 15 is a graph illustrating an example of a
waveform of a load torque of a rotary compressor.
FIG. 16 is a block diagram illustrating a
5 configuration of a driving device for an AC motor according
to a fifth embodiment.
FIG. 17 is a block diagram illustrating a
configuration of a driving device for an AC motor according
to a sixth embodiment.
10 FIG. 18 is a cross-sectional view illustrating an
outline structure inside a refrigerant compressor
illustrated as a driven object in FIG. 17.
FIG. 19 is a cross-sectional view illustrating a
structure inside a compression chamber of the refrigerant
15 compressor illustrated in FIG. 18.
FIG. 20 is a diagram illustrating a configuration of a
refrigeration cycle apparatus according to a seventh
embodiment.
20 Description of Embodiments
[0010] A speed estimating device for an AC motor, a
driving device for an AC motor, a refrigerant compressor,
and a refrigeration cycle apparatus according to certain
embodiments of the present invention will be described in
25 detail below with reference to the drawings. Note that the
present invention is not limited by the embodiments
described below. Hereinafter, the speed estimating device
for an AC motor will be simply referred to as a “speed
estimating device” as appropriate. In addition, the
30 driving device for an AC motor will be simply referred to
as a “driving device” as appropriate.
[0011] First Embodiment.
FIG. 1 is a block diagram illustrating a configuration
7
of a speed estimating device 101 for an AC motor according
to a first embodiment. The speed estimating device 101
illustrated in FIG. 1 estimates the rotational speed of an
AC motor 2 in a technique of an adaptive observer using a
5 voltage vector applied to the AC motor 2 and a current
vector, and outputs the estimation result as an estimated
angular velocity ω^
r.
[0012] The speed estimating device 101 includes a model
deviation computing unit 11, a first angular velocity
10 estimating unit 21, a compensation phase computing unit 51,
a second angular velocity estimating unit 22, and an
estimated angular velocity calculator 23.
[0013] The model deviation computing unit 11 computes a
model deviation ε on the basis of the voltage vector, the
current vector, and the estimated angular velocity ω^ 15 r. The
first angular velocity estimating unit 21 calculates a
first estimated angular velocity ω^
r1 on the basis of the
model deviation ε. The compensation phase computing unit
51 computes a compensation phase θpls on the basis of a
20 specific disturbance frequency fd. The second angular
velocity estimating unit 22 calculates a second estimated
angular velocity ω^
r2 on the basis of the compensation phase
θpls, the model deviation ε, and the disturbance frequency
fd. The estimated angular velocity calculator 23
calculates an estimated angular velocity ω^ 25 r of the AC
motor 2 on the basis of the first estimated angular
velocity ω^
r1 and the second estimated angular velocity ω^
r2.
[0014] The speed estimating device 101 includes the
compensation phase computing unit 51 and the second angular
30 velocity estimating unit 22, and operations performed by
these components are one of the characteristics of the
present invention. Details of the compensation phase
computing unit 51 and the second angular velocity
8
estimating unit 22 will be described later.
[0015] In the first embodiment, the disturbance
frequency fd is assumed to be known. The disturbance
frequency fd may be obtained in any manner. For example,
5 in such a system in which disturbance of a particular
frequency occurs, the disturbance frequency fd can be
provided as a constant in advance. Alternatively, in such
an application as a compressor in which disturbance
depending on rotational frequency occurs, the rotational
10 frequency can be used as the disturbance frequency fd. The
rotational frequency mentioned herein can be acquired by a
rotational position sensor or a velocity sensor.
Alternatively, in a case of a device including angular
velocity estimating means as in the first embodiment, the
15 rotational frequency can be obtained from the estimated
angular velocity ω^
r. Still alternatively, the frequency of
torque pulsation may be detected or estimated by a torque
meter, an acceleration sensor, or a vibration sensor, and
used as the disturbance frequency fd.
20 [0016] The first angular velocity estimating unit 21 and
the second angular velocity estimating unit 22 both
estimate angular velocity. The difference therebetween
lies in a frequency for an angular velocity to be estimated.
While the first embodiment is directed to a configuration
25 in which the first angular velocity estimating unit 21
estimates a low-frequency component including a directcurrent (DC) component of an angular velocity and the
second angular velocity estimating unit 22 estimates a
high-frequency component of the angular velocity, the
30 present invention is not limited to this configuration.
Needless to say, an opposite configuration can
alternatively be used, in which the first angular velocity
estimating unit 21 estimates a higher angular velocity
9
frequency component.
[0017] Next, a configuration and functions of the model
deviation computing unit 11 will be described. The model
deviation computing unit 11 includes a state estimator 12,
5 a subtractor 13, and a deviation calculator 14. The state
estimator 12 calculates and outputs an estimated magnetic
flux vector and an estimated current vector on the basis of
the voltage vector applied to the AC motor 2, the current
vector outputted by the AC motor 2, and the estimated
angular velocity ω^
r. The estimated angular velocity ω^ 10 r is
an estimated angular velocity calculated by the
aforementioned estimated angular velocity calculator 23,
and is also an output of the speed estimating device 101.
[0018] The subtractor 13 subtracts the current vector
15 from the estimated current vector to calculate a current
deviation vector. The deviation calculator 14 receives the
current deviation vector as its input, extracts an
orthogonal component of the estimated magnetic flux vector
as a scalar quantity, and outputs a value of the quantity
20 as the model deviation ε. Note that the technique for
extracting an orthogonal component of an estimated magnetic
flux vector as a scalar quantity is publicly known. For
example, there are publicly known a technique of performing
coordinate transformation on a current deviation vector to
25 two rotational axes, a technique of computing the magnitude
of the cross product of the current deviation vector and
the estimated magnetic flux vector, and more.
[0019] Specifically, the state estimator 12 estimates an
electric current and a magnetic flux according to a state
30 equation of the AC motor 2. While the AC motor 2 is
assumed to be a typical embedded magnet type synchronous
electric AC motor in this example, any other AC motors can
also be used as long as a state equation similar to that
10
described below can be established in the state estimator
12. Examples of other AC motors include a surface magnet
type synchronous electric motor and an induction electric
motor.
5 [0020] In the case of an embedded magnet type
synchronous AC motor, the state equation is expressed by
the following equations (1) and (2).
[0021]
[Formula 1]
10 ∙∙∙(1)
[0022]
[Formula 2]
∙∙∙(2)
[0023] In the equations (1) and (2), Ld and Lq represent
15 inductances on a d-axis and a q-axis, respectively. R
represents an armature resistance. ω represents a primary
angular frequency. ωr represents an angular velocity. vds
represents a d-axis voltage. vqs represents a q-axis
voltage. ids represents a d-axis current. iqs represents a
20 q-axis current. φds represents a d-axis stator magnetic
flux. φqs represents a q-axis stator magnetic flux. φdr
represents a d-axis rotor magnetic flux. h11 to h32
represent observer gains. A symbol “^” represents an
estimated value.
25 [0024] Note that the primary angular frequency is given
as expressed according to the following equation (3).































































    
 














qs qs
ds ds
32
22
12
31
21
11
qs
ds
dr
qs
ds
r
q
d
dr
qs
ds
i i
ˆ
i i
ˆ
h
h
h
h
h
h
0
v
v
ˆ
ˆ
ˆ
0 0 0
ˆ
L
R
0
L
R
ˆ
ˆ
ˆ
dt
d




























dr
qs
ds
q
d
qs
ds
ˆ
ˆ
ˆ
0
0
1/L
0
0
1/L
i
ˆ
i
ˆ
11
[0025]
[Formula 3]
∙∙∙(3)
[0026] In the above equation (3), h41 and h42 represent
5 observer gains similarly to the aforementioned h11 to h32.
[0027] While the above equations (1) and (2) are
equations based on normal induced voltage, similar
calculation can be performed with modifying the above
equations (1) and (2) to express a form using an extended
10 induced voltage. Because the above equation (1) includes
the estimated angular velocity ω^
r, an error occurs in
current estimation when the estimated angular velocity ω^
r
and an actual angular velocity ωr are not equal to each
other. Herein, the model deviation ε is defined according
15 to the following equation (4). The speed estimating device
101 adjusts the value of the estimated angular velocity ω^
r
with use of the first angular velocity estimating unit 21
and the second angular velocity estimating unit 22 so that
the model deviation ε becomes zero.
20 [0028]
[Formula 4]
∙∙∙(4)
[0029] As described above, one of characteristics of the
speed estimating device 101 is that the speed estimating
25 device 101 has the compensation phase computing unit 51 and
the second angular velocity estimating unit 22. In order
to explain this characteristic, a speed estimating device
that does not have the compensation phase computing unit 51
and the second angular velocity estimating unit 22 will
30 first be explained as a comparative example herein.
dr
41 ds ds 42 qs qs
r ˆ
i i )
ˆ
i i ) h (
ˆ h (
ˆ

  
   
dr
qs qs
ˆ
i i
ˆ


 
12
[0030] FIG. 2 is a block diagram illustrating a
configuration of a speed estimating device 101A according
to the comparative example. The speed estimating device
101A illustrated in FIG. 2 operates in accordance with a
5 sensorless vector control method in a manner similar to the
speed estimating device 101 illustrated in FIG. 1. The
speed estimating device 101A operates to adjust the model
deviation ε to zero only by using the first angular
velocity estimating unit 21.
10 [0031] In the speed estimating device 101A illustrated
in FIG. 2, the first angular velocity estimating unit 21
includes a proportional integral (PI) controller 24 and an
integrator 25. The first angular velocity estimating unit
21 operates in accordance with the following equation (5).
15 [0032]
[Formula 5]
∙∙∙(5)
[0033] In the equation (5), KP represents a proportional
gain of the entire first angular velocity estimating unit
20 21. KI represents an integral gain of the entire first
angular velocity estimating unit 21. s represents an
operator of Laplace transform, s refers to differentiation,
and 1/s refers to integration.
[0034] In the first angular velocity estimating unit 21,
25 the PI controller 24 computes a first estimated angular
acceleration ω∙^
r1 on the basis of the model deviation ε.
The integrator 25 integrates the first estimated angular
acceleration ω∙^
r1 to compute the first estimated angular
velocity ω^
r1. In the first angular velocity estimating
unit 21, the first estimated angular velocity ω^ 30 r1 is
adjusted by the PI controller 24 and the integrator 25.











   
s
K
K
s
1
ˆ
I
r P
13
The first estimated angular velocity ω^
r1 is outputted to
the outside as an output of the speed estimating device
101A. In addition, the first estimated angular velocity
ω
^
r1 is fed back to the model deviation computing unit 11.
5 As described above, the PI controller 24 operates as a
first angular acceleration estimator, and the integrator 25
operates as a first angular velocity calculator.
[0035] In addition, the transfer function Ga(s) from the
first estimated angular velocity ω^
r1 to the model deviation
10 ε is publicly known in the 226-th page of “Speed Sensorless
Vector Control Method of Induction Motor Including A Low
Speed Region”, The transactions of the Institute of
Electrical Engineers of Japan (Vol. 120-D, No. 2, 2000)
that is a Non Patent Literature. The transfer function
15 Ga(s) can be approximated by a first order lag as in the
following equation (6).
[0036]
[Formula 6]
∙∙∙(6)
20 [0037] FIG. 3 is a Bode plot illustrating transfer
characteristics of the speed estimating device 101A
illustrated in FIG. 2. The horizontal axis represents
frequency, and the vertical axis represents gain. A
transfer function of (1) illustrated by a broken line in
25 FIG. 3 is designed such that the gain in a lower frequency
range is higher. In the transfer function of (1), the gain
decreases as the frequency is higher. Specifically, the
gain decreases at a rate of -40 [dB/decade] in a lowfrequency band, and decreases at a rate of -20 [dB/decade]
30 at frequencies higher than that at a break point P1.
[0038] In addition, a transfer function of (2)
x
x
a
1 sT
A
G (s)


14
illustrated by a dotted line in FIG. 3 corresponds to the
transfer function Ga(s) of the above equation (6). Because
the transfer function Ga(s) has a first order lag
characteristic from the first estimated angular velocity
ω
^ 5 r1 to the model deviation ε, the gain decreases at a rate
of -20 [dB/decade] in a frequency range higher than a
cutoff angular frequency f1. Addition of the two transfer
functions results in a transfer function of (3) having an
open loop characteristic illustrated by a solid line.
10 [0039] If PI control gains in the above equation (5),
that is, the proportional gain KP and the integral gain KI
in the first angular velocity estimating unit 21 can be set
to be sufficiently large, speed pulsation at high
frequencies can be accurately estimated. These gain values,
15 however, are constrained by the estimation computation
period and the influence of an error in a motor constant.
A forced increase in the gains enhances vulnerability to
high-frequency noise, which makes appropriate estimation
processing impossible. For this reason, the speed
20 estimating device 101A according to the comparative example
has a problem in that it is difficult to capture highfrequency speed pulsation.
[0040] Next, another comparative example will be
explained. FIG. 4 is a block diagram illustrating a
25 configuration of a speed estimating device 101B according
to a comparative example different from that in FIG. 2.
Hereinafter, for distinction from FIG. 2, the comparative
example of FIG. 2 will be referred to as a “first
comparative example”, and the comparative example of FIG. 4
30 will be referred to as a “second comparative example”. As
compared with the speed estimating device 101A of FIG. 2,
the speed estimating device 101B according to the second
comparative example illustrated in FIG. 4 is provided
15
additionally with a second angular velocity estimating unit
22B.
[0041] In the speed estimating device 101B of FIG. 4,
the second angular velocity estimating unit 22B includes a
5 second angular acceleration estimating unit 30B, and an
integrator 31. The second angular acceleration estimating
unit 30B computes a second estimated angular acceleration
ω
∙^
r2 on the basis of the disturbance frequency fd and the
model deviation ε. The integrator 31 integrates the second
estimated angular acceleration ω∙^ 10 r2 and outputs a second
estimated angular velocity ω^
r2.
[0042] In addition, the second angular acceleration
estimating unit 30B includes a Fourier coefficient
calculator 26, PI controllers 27 and 28, and an AC
15 restoring unit 29.
[0043] The Fourier coefficient calculator 26 converts a
specific frequency component of the model deviation into
direct current, and extracts the obtained DC component. A
cosine coefficient Ec and a sine coefficient Es are
20 outputted from the Fourier coefficient calculator 26, which
correspond to the specific frequency component obtained by
the DC conversion.
[0044] In this process, the cosine coefficient Ec of the
model deviation ε and the sine coefficient Es of the model
25 deviation ε are calculated by the following equations (7)
and (8), respectively, on the basis of the model deviation
ε and the disturbance frequency fd.
[0045]
[Formula 7]
30 ∙∙∙(7)
[0046]
[Formula 8]
cos( 2 f t) dt
T
2 Td
0
d
d
c      
16
∙∙∙(8)
[0047] In the formulas (7) and (8), t represents time.
In addition, Td represents a cycle period of disturbance,
which is the reciprocal of the disturbance frequency fd.
5 That is, Td=1/fd.
[0048] The cosine coefficient Ec of the model deviation
is subjected to PI control by the PI controller 27 as
expressed by the following equation (9). In addition, the
sine coefficient Es of the model deviation is subjected to
10 PI control by the PI controller 28 as expressed by the
following equation (10).
[0049]
[Formula 9]
∙∙∙(9)
15 [0050]
[Formula 10]
∙∙∙(10)
[0051] In the equations (9) and (10), KP2 represents a
proportional gain of the entire second angular velocity
20 estimating unit 22B. KI2 represents an integral gain of the
entire second angular velocity estimating unit 22B. A dot
over each character refers to differentiation, and the
number of dots represents the order of differentiation.
[0052] The AC restoring unit 29 performs computation of
25 the following equation (11) on the basis of the cosine
coefficient Ec of the model deviation and the sine
coefficient Es of the model deviation. The equation (11)
is an arithmetic expression for calculating the second
estimated angular acceleration ω∙^
r2.
sin( 2 f t) dt
T
2 Td
0
d
d
s      
c
I2
c P2
s
K
K
ˆ





   
s
I2
s P2
s
K
K
ˆ





   
17
[0053]
[Formula 11]
∙∙∙(11)
[0054] FIG. 5 is a Bode plot illustrating transfer
5 characteristics of the speed estimating device 101B
illustrated in FIG. 4. The horizontal axis represents
frequency, and the vertical axis represents gain. A
transfer function of (1) in FIG. 5 is the same as the
transfer function of (1) in FIG. 3. A transfer function of
10 (2) in FIG. 5 is the same as the transfer function of (2)
in FIG. 3. A transfer function of (3) in FIG. 5 represents
a transfer function of the second angular velocity
estimating unit 22B illustrated in FIG. 4. Addition of the
three transfer functions results in an open loop
15 characteristic of (4) illustrated by a solid line.
[0055] In FIG. 5 as compared with FIG. 3, the gain in a
specific frequency band is higher in the open loop
characteristic of (4) illustrated by the solid line. More
precisely, the speed estimating device 101B according to
20 the second comparative example increases the gain in a
specific frequency band in which occurrence of speed
pulsation due to periodic disturbance can be predicted, by
using the first angular velocity estimating unit 21 and the
second angular velocity estimating unit 22B, thereby making
25 it possible to increase the accuracy of speed estimation.
As a result, the speed estimating device 101B according to
the second comparative example enables estimation of highfrequency speed pulsation with high accuracy, which is
difficult for the speed estimating device 101A according to
30 the first comparative example.
[0056] As described above, the speed estimating device
101B according to the second comparative example can
 sin(2 f t)
ˆ
cos(2 f t)
ˆ ˆr2  c  d s  d
  
18
estimate high-frequency speed pulsation with high accuracy,
but it is envisaged that the control system may be instable
depending on the magnitude of phase error. In the
circumstances, the inventors of the present application
5 have considered the necessity of phase compensation for a
proposal of the present invention, and such consideration
will be described below.
[0057] FIG. 6 is a graph used for comparing Bode plots
illustrating the transfer characteristics from the model
10 deviation ε to the first estimated angular acceleration
ω
∙^
r1 of the speed estimating device 101A illustrated in FIG.
2 in an open loop characteristic and a closed loop
characteristic. The open loop characteristic is a transfer
characteristic in a state in which the first estimated
angular velocity ω^ 15 r1 is not fed back to the model deviation
computing unit 11. The closed loop characteristic is a
transfer characteristic in a state in which the first
estimated angular velocity ω^
r1 is fed back to the model
deviation computing unit 11 as illustrated in FIG. 2. Note
20 that FIGS. 3 and 5 are Bode plots from the model deviation
ε to the first estimated angular velocity ω^
r1. In contrast,
FIG. 6 illustrates the transfer characteristics from input
of the model deviation ε to output of the first estimated
angular acceleration ω∙^
r1, and attention should be paid to
25 a first-order differential characteristic being added to
FIGS. 3 and 5. As illustrated in FIG. 4, when the second
angular velocity estimating unit 22B computes the second
estimated angular acceleration ω∙^
r2 on the basis of the
model deviation ε, the characteristic from input of the
30 model deviation ε to output of the second estimated angular
acceleration ω∙^
r2 illustrated in FIG. 6 can be considered
as a characteristic to be controlled by the second angular
velocity estimating unit 22B.
19
[0058] In feedback control, a technique of design in
view of the open loop characteristics of a controlled
object and a controller is often used. In this situation,
what is considered first is a case of designing the second
5 angular velocity estimating unit in view of the open loop
characteristic of the controlled object.
[0059] According to the open loop characteristic in FIG.
6, the gain decays at -20 [dB/decade] in a low-frequency
band. The phase is -90 [degrees] in the low-frequency band,
10 lags as the frequency is higher, and converges to -180
[degrees]. While the used frequency band depends on an
application, the phase can be considered as being
substantially constant at -90 [degrees] in a case where a
low-frequency band is mainly used. In such a case, if the
15 open loop characteristic is a controlled object in
designing the second angular velocity estimating unit,
phase compensation seems to be unnecessary.
[0060] In fact, however, the PI controllers 27 and 28
serving as second angular acceleration estimators perform
20 computation in a state in which the PI controller 24
serving as the first angular acceleration estimator
operates. For this reason, it should be considered that
the first estimated angular velocity ω^
r1 estimated by the
first angular velocity estimating unit 21 including the PI
25 controller 24 is fed back to the model deviation computing
unit 11. Accordingly, the second angular velocity
estimating unit 22B needs to be designed in view of the
closed loop characteristic to be controlled by the second
angular velocity estimating unit 22B.
30 [0061] The closed loop characteristic in FIG. 6 has a
first-order integral characteristic in a high-frequency
band, and a first-order differential characteristic in a
low-frequency band. The gain has a characteristic that
20
lowers at a gradient of -20 [dB/decade] in consistency with
the open loop characteristic in a high-frequency band, but
a characteristic that increases at a gradient of +20
[dB/decade] in a low-frequency band and makes the gain
5 lower as the frequency is lower. In addition, the phase
converges to -180 [degrees] in a high-frequency band, which
is consistent with the case of the open loop, but is +90
[degrees] in a low-frequency band. The phase thus changes
significantly from +90 [degrees] to -180 [degrees] in the
10 frequency band therebetween.
[0062] Note that, as described above, the angular
acceleration computation in the second angular velocity
estimating unit 22B of the speed estimating device 101B
according to the second comparative example is performed
15 using the above equations (7) to (11). The equations (7)
and (8) are arithmetic expressions for obtaining the cosine
coefficient and the sine coefficient that correspond to DC
components, the equations (9) and (10) are arithmetic
expressions for PI control, and the equation (11) is an
20 arithmetic expression for restoring an AC component by
getting back the DC components into the AC component. In
the series of computation processes, the phase is not
considered. This means that the phase is assumed not to
change depending on frequency in a range from input of the
25 model deviation ε to output of the second estimated angular
acceleration ω∙^
2.
[0063] For example, as described above, in the case
where only the open loop characteristic is deemed to be a
controlled object and operation is performed in a low30 frequency range, the phase change may be considered as
being sufficiently small and it is contemplated to adopt
the control design as described above. In fact, however,
the phase characteristic of a controlled object greatly
21
changes depending on frequency as illustrated in the Bode
plots in FIG. 6. For this reason, unless an estimated
angular acceleration is calculated in view of the fact that
the phase characteristic of a controlled object changes
5 depending on disturbance frequency, an error with respect
to an estimated phase of appropriate angular acceleration
is caused.
[0064] The second angular acceleration estimating unit
30B in the second comparative example computes an angular
10 acceleration with use of the PI controllers 27 and 28.
With this configuration, when a phase error is sufficiently
small, the phase may converge to an appropriate value as a
result of adjustment of a control quantity by the PI
controllers 27 and 28. In a frequency band in which a
15 phase error is significantly large, however, control may
become instable.
[0065] In the circumstances, in the first embodiment,
the second angular velocity estimating unit 22 includes a
control system for performing phase compensation such that
20 an angular velocity can be computed with an appropriate
phase. FIG. 7 is a block diagram illustrating a detailed
configuration of the second angular velocity estimating
unit 22 in the speed estimating device 101 illustrated in
FIG. 1. As illustrated in FIG. 7, the second angular
25 acceleration estimating unit 30 includes a Fourier
coefficient calculator 52, integral (I-) controllers 53 and
54, and an AC restoring unit 55. In the first embodiment,
the second angular acceleration estimating unit 30 operates
as a specific-frequency angular acceleration estimator. In
30 addition, in the second angular acceleration estimating
unit 30, the Fourier coefficient calculator 52 operates as
a specific frequency extractor that extracts a specific
frequency component, and the I-controllers 53 and 54 and
22
the AC restoring unit 55 operate as a specific-frequency
angular velocity estimator.
[0066] In FIG. 7, the disturbance frequency fd is
inputted to the compensation phase computing unit 51, the
5 Fourier coefficient calculator 52, and the AC restoring
unit 55. The compensation phase computing unit 51
determines a compensation phase θpls in view of the closed
loop characteristic to be controlled. Specifically, the
compensation phase θpls is stored as a map associated with
10 the disturbance frequency, and the compensation phase θpls
can be determined by reference to the map. Alternatively,
an approximation formula that is changed depending on
disturbance frequency may be held, and the compensation
phase θpls may be determined using the approximation formula.
15 The compensation phase θpls is inputted to the Fourier
coefficient calculator 52.
[0067] The Fourier coefficient calculator 52 obtains the
cosine coefficient Ec’ and the sine coefficient Es’ of the
model deviation on the basis of the disturbance frequency
20 fd and the compensation phase θpls using the following
equations (12) and (13).
[0068]
[Formula 12]
∙∙∙(12)
25 [0069]
[Formula 13]
∙∙∙(13)
[0070] The cosine coefficient Ec’ of the model deviation
is subjected to I-control by the I-controller 53 as
30 expressed by the following equation (14). In addition, the
cos2 f t  dt
T
2
E d pls
T
0
d
c
d      


sin2 f t  dt
T
2
E d pls
T
0
d
s
d      


23
sine coefficient Es’ of the model deviation is subjected to
I-control by the I-controller 54 as expressed by the
following equation (15).
[0071]
5 [Formula 14]
∙∙∙(14)
[0072]
[Formula 15]
∙∙∙(15)
10 [0073] In the above equations (14) and (15), Krpl_i
represents the integral gain of the I-controllers 53 and 54.
Note that the cosine coefficient Ec’ and the sine
coefficient Es’ that are control inputs in the Icontrollers 53 and 54 are in a dimension of an angular
15 velocity, whereas control outputs of the I-controllers 53
and 54 are in a dimension of an angular acceleration. In
addition, conversion from an angular velocity to an angular
acceleration has a relation of differentiation, whereas a
controlled object naturally has an integral characteristic.
20 Thus, in a coordinate system resulting from conversion to
direct current, the cosine coefficient Ec’ and the sine
coefficient Es’ that are control inputs each appear as a
gain of a multiple of conversion frequency. The cosine
coefficient Ec’ and the sine coefficient Es’ are therefore
25 regarded as a proportional characteristic in this
coordinate system rather than an integral characteristic.
In this situation, control can be performed only by
integrators, and the I-controllers 53 and 54 are thus used.
Needless to say, PI controllers may be used where necessary
30 as in the second comparative example in order to improve

 








  c
rpl_i
c E
s
ˆ K 

 








  s
rpl_i
s E
s
ˆ K 
24
the responsiveness.
[0074] The AC restoring unit 55 performs the computation
of the following equation (16) on the basis of the cosine
coefficient Ec’ and the sine coefficient Es’. This equation
5 (16) is an arithmetic expression for calculating the second
estimated angular acceleration ω∙^
r2.
[0075]
[Formula 16]
∙∙∙(16)
10 [0076] FIG. 9 is a first graph used for explaining the
effect of the speed estimating device 101 according to the
first embodiment. FIG. 10 is a second graph used for
explaining the effect of the speed estimating device 101
according to the first embodiment. FIGS. 9 and 10 both
15 illustrate an example of a result of simulation in which
the AC motor 2 is driven with being provided with speed
pulsation and the rotational speed of the AC motor 2 is
estimated. In this simulation, the second angular velocity
estimating unit 22 is activated and speed pulsation
20 estimation is started five seconds after the AC motor 2 is
driven.
[0077] In addition, FIG. 9 illustrates waveforms
obtained by plotting cosine components Ωc
∙^ of an angular
acceleration estimated with the estimation response being
set to 1 [rad/s], and sine components Ωs
∙^ 25 of the angular
acceleration. Upper part thereof illustrates waveforms in
a case without phase compensation, which corresponds to a
result obtained by the configuration in the second
comparative example. Lower part thereof illustrates
30 waveforms in a case with phase compensation, which
corresponds to a result obtained by the configuration of
the first embodiment.
     sin 2 f t 
ˆ
cos 2 f t
ˆ ˆ
r2 c d s  d
  
     
25
[0078] In the case without phase compensation, because
some error is caused in an estimated phase of speed
pulsation as described above, the value of angular
acceleration does not converge but diverges, and a set
5 response cannot be obtained. In contrast, in the case with
phase compensation, it can be seen that the estimated
angular acceleration converges and the operation is stable.
In addition, it can be seen that the response speed rises
by 63% in about one second from the start of estimation,
10 and desired response is thus achieved.
[0079] In addition, FIG. 10 illustrates waveforms of an
estimated angular velocity, upper part thereof illustrates
a case without phase compensation, and lower part thereof
illustrates a case with phase compensation. A left part of
15 each of the upper and lower parts illustrates waveforms
before activation of the second angular velocity estimating
unit 22, and a right part of each of the upper and lower
parts illustrates waveforms after the second angular
velocity estimating unit 22 is activated and the value has
20 converged. Each thick curve represents an actual angular
velocity, and each thin curve represents an estimated
angular velocity.
[0080] Before the activation, the phase of the estimated
angular velocity lags behind the actual angular velocity,
25 and also the amplitude of the estimated angular velocity is
smaller than that of the actual angular velocity. After
the activation, in the case without phase compensation, the
control diverges and the amplitude of the estimated angular
velocity is significantly larger than that of the actual
30 angular velocity, and also the phase thereof has a
difference from the latter. In contrast, in the case with
phase compensation, the estimated angular velocity is equal
to the actual angular velocity, and it is seen that the
26
control is performed satisfactorily.
[0081] The integrator 31 integrates the second estimated
angular acceleration ω∙^
r2 computed by the AC restoring unit
55 according to the following equation (17) to obtain the
second estimated angular velocity ω^ 5 r2. The integrator 31
operates as a second angular velocity calculator. The
second estimated angular velocity ω^
r2 is computed as a
specific high-frequency component of the actual angular
velocity.
10 [0082]
[Formula 17]
∙∙∙(17)
[0083] Note that it is obvious for a person skilled in
the art that a block diagram of a control system can be
15 modified. For example, a configuration may be realized as
in FIG. 8. FIG. 8 is a block diagram illustrating a
modification of a configuration of a detail part
illustrated in FIG. 7. For example, the configuration of
the integrators 25 and 31 in FIG. 7 may be modified to be a
20 configuration such that the estimated angular accelerations
are added before passing through an integrator.
Specifically, while two integrators 25 and 31 are located
on input sides of the estimated angular velocity calculator
23 in the configuration of FIG. 7, one integrator 32 may be
25 located on an output side of the estimated angular velocity
calculator 23 as in a speed estimating device 101-1 of FIG.
8. This configuration produces an advantageous effect that
it is possible to reduce the number of integrators.
[0084] An equation for estimating a final angular
30 velocity is expressed by the following equation (18).
Specifically, in the estimated angular velocity calculator
23, the second estimated angular velocity ω^
r2 calculated by
r2 r2
ˆ
s
1
ˆ  
27
the integrator 31 is added to the first estimated angular
velocity ω^
r1 computed by the first angular velocity
estimating unit 21, thereby obtaining the estimated angular
velocity ω^
r expressed by the following equation (18).
5 [0085]
[Formula 18]
∙∙∙(18)
[0086] While an example of using an adder is described
in the above equation (18) and in the estimated angular
10 velocity calculator 23 illustrated in FIG. 7, the present
invention is not limited to this example. In such a case
where the definition of positive or negative of the
compensation phase in the compensation phase computing unit
51 and the definition of an output of the AC restoring unit
15 55 are opposite in phase, a subtractor is used. In other
words, the configuration of the estimated angular velocity
calculator 23 is determined depending on the definition of
positive or negative of the compensation phase, the
definition of an output from the AC restoring unit 55, and
20 the like.
[0087] A difference between the equation (18) and the
equation (5) lies in that the second estimated angular
velocity ω^
r2 is used in the equation (18). The second
angular velocity estimating unit 22 converts a given
25 harmonic wave of the model deviation ε into direct currents
by dividing the harmonic wave into a sine wave and a cosine
wave, extracts the sine wave and the cosine wave, and
performs I-control so that the sine wave and the cosine
wave become zero. The second angular velocity estimating
30 unit 22 then restores the output of the I-control to an
alternating current to estimate a high-frequency component
r2
I
r P
ˆ
s
K
K
s
1
ˆ   





 




   
28
of the actual angular velocity, and increases the gain only
in a part with a determined frequency. Thus, a speed
pulsation component due to periodic disturbance can be
estimated as the second estimated angular velocity ω^
r2 with
5 high accuracy. Note that the second angular velocity
estimating unit 22 described above has a structure of a
sort of iterative controller or a learning controller.
Another sort of iterative controller or a learning
controller may therefore be used instead of the second
10 angular velocity estimating unit 22.
[0088] FIG. 11 is a hardware configuration diagram of
the speed estimating device 101 according to the first
embodiment. Although not illustrated in FIGS. 1 and 7, a
voltage applying unit 3 and a current detecting unit 4 are
15 illustrated in FIG. 11. The voltage applying unit 3 serves
as voltage applying means for applying a voltage to the AC
motor 2. An example of the voltage applying means is a
power converter. The voltage vector corresponds to a
voltage command generated by the voltage applying unit 3.
20 A voltage generated on the basis of the voltage command is
applied to the AC motor 2, and information on the voltage
command is inputted to the speed estimating device 101. In
addition, the current vector is generated by the current
detecting unit 4 and inputted to the speed estimating
25 device 101. The current vector is vector information on an
AC current flowing in the AC motor 2. An example of the
current vector is a detected value of a dq-axis current
obtained by converting an AC current detected by the
current detecting unit 4 into a value on dq-coordinate axes.
30 [0089] The speed estimating device 101 includes a
processor 901 and a memory 902. The memory 902 includes a
volatile storage device, which is not illustrated, typified
by a random access memory, and a nonvolatile auxiliary
29
storage device, which is not illustrated, typified by a
flash memory. Note that the memory 902 may include an
auxiliary storage device of a hard disk instead of the
volatile storage device and the nonvolatile auxiliary
5 storage device. The processor 901 executes a program
inputted from the memory 902. Because the memory 902
includes the auxiliary storage device and the volatile
storage device, a program is inputted from the auxiliary
storage device to the processor 901 via the volatile
10 storage device. The processor 901 may output data on a
computation result to the volatile storage device of the
memory 902, and may save the data into the auxiliary
storage device via the volatile storage device.
[0090] Various systems have been considered for the
15 voltage applying unit 3 and the current detecting unit 4,
and basically any system may be used therefor. The voltage
applying unit 3 and the current detecting unit 4 may be
provided inside the speed estimating device 101. In
addition, the speed estimating device 101 may include
20 voltage detecting means for detecting the voltage vector
outputted by the voltage applying unit 3. In this case,
the voltage applying unit 3 may be configured to transmit a
voltage vector command value to the processor 901, for a
numerical value relating to the voltage detected by the
25 voltage detecting means to be transmitted to the processor
901. The current detecting unit 4 may also be configured
to transmit a detected numerical value to the processor 901.
[0091] The processor 901 computes the estimated angular
velocity ω^
r on the basis of the current vector and the
30 voltage vector of the AC motor 2. By the processor 901
performing the computation of the second angular velocity
estimating unit 22 described above, the speed pulsation due
to periodic disturbance can be estimated with high accuracy.
30
Note that the processor 901 may also serve as the driving
device for the AC motor 2. Specifically, the processor 901
may be configured to not only perform the speed estimation
but also calculate a voltage command vector such that an
5 estimated speed has a desired value. There are publicly
known various methods for performing position-sensorless
torque control, including that of Non Patent Literature
mentioned above.
[0092] As described above, the speed estimating device
10 for an AC motor according to the first embodiment can
estimate speed pulsation of the AC motor with an
appropriate phase regardless of frequency, and can realize
higher accuracy of speed estimation.
[0093] In addition, the speed estimating device for an
15 AC motor according to the first embodiment enables speed
estimation with high accuracy even in a case of high
pulsation frequency, which has been a problem in the
conventional art, and can also estimate pulsation in a
higher frequency range than in the conventional art by
20 virtue of provision of a computation unit for increasing
the estimation response in a specific frequency band even
if a special storage means is not provided therefor.
Furthermore, because the compensation phase is obtained in
view of the closed loop phase characteristic of angular
25 velocity estimation of the first angular velocity
estimating unit that is an adaptive magnetic-flux observer,
the angular velocity estimation can be performed at a
desired response speed, and stable control can be achieved.
[0094] Second Embodiment.
30 FIG. 12 is a block diagram illustrating a
configuration of a speed estimating device 101C according
to a second embodiment. In FIG. 12, in the speed
estimating device 101C according to the second embodiment,
31
the second angular velocity estimating unit 22 in the
configuration of the speed estimating device 101 according
to the first embodiment illustrated in FIG. 7 is replaced
with a second angular velocity estimating unit 22C. In the
5 second angular velocity estimating unit 22C, the second
angular acceleration estimating unit 30 is replaced with a
second angular acceleration estimating unit 30C. In the
second angular acceleration estimating unit 30C, the
Fourier coefficient calculator 52 is replaced with a
10 Fourier coefficient calculator 52C, and the AC restoring
unit 55 is replaced with an AC restoring unit 55C. While
the compensation phase θpls computed by the compensation
phase computing unit 51 is inputted to the Fourier
coefficient calculator 52 in FIG. 7, the compensation phase
15 θpls is inputted to the AC restoring unit 55C in FIG. 12.
Note that the other configuration is the same as or
equivalent to that in FIG. 7, and the same or equivalent
components are denoted by the same reference symbols and
redundant description thereof will be omitted.
20 [0095] The Fourier coefficient calculator 52 of the
first embodiment illustrated in FIG. 7 computes Fourier
coefficients by using the equations (12) and (13). On the
other hand, the Fourier coefficient calculator 52C of the
second embodiment illustrated in FIG. 12 computes Fourier
25 coefficients by using the following equations (19) and (20).
[0096]
[Formula 19]
∙∙∙(19)
[0097]
30 [Formula 20]
cos  2 f t dt
T
2
E d
T
0
d
c
d    


32
∙∙∙(20)
[0098] In addition, the AC restoring unit 55 of the
first embodiment illustrated in FIG. 7 computes the second
estimated angular acceleration ω∙^
r2 by using the above
5 equation (16). In contrast, the AC restoring unit 55C of
the second embodiment illustrated in FIG. 12 computes the
second estimated angular acceleration ω∙^
r2 by using the
following equation (21).
[0099]
10 [Formula 21]
∙∙∙(21)
[0100] If the relation of the phases used for
computation by the Fourier coefficient calculator 52 (52C)
and the AC restoring unit 55 (55C) is maintained, the
15 effect of estimating speed pulsation with an appropriate
phase regardless of frequency, which is mentioned in the
first embodiment, can be similarly produced. Thus,
computation as expressed by the equations (19) to (21) can
produce the same effect.
20 [0101] Note that, in a manner similar to the first
embodiment, a configuration in which the cosine coefficient
Ec
’ of the model deviation ε and the sine coefficient Es
’ of
the model deviation ε are obtained from the I-controllers
53 and 54, respectively, can be adopted also in the second
25 embodiment, and the arithmetic expressions for it are still
the equations (14) and (15) with no change.
[0102] Third Embodiment.
FIG. 13 is a block diagram illustrating a
configuration of a speed estimating device 101D according
30 to a third embodiment. In FIG. 13, in the speed estimating
sin  2 f t dt
T
2
E d
T
0
d
s
d    


r 2  c  d pls  s sin 2 f
d
t pls 
ˆ cos 2 f t
ˆ ˆ   
     
     
33
device 101D according to the third embodiment, the first
angular velocity estimating unit 21 and the second angular
velocity estimating unit 22 in the configuration of the
speed estimating device 101 according to the first
5 embodiment illustrated in FIG. 7 are replaced with a first
angular velocity estimating unit 21D and a second angular
velocity estimating unit 22D, respectively. In the first
angular velocity estimating unit 21D, the integrator 25 is
omitted from the configuration of the first angular
10 velocity estimating unit 21, and in the second angular
velocity estimating unit 22D, the integrator 31 is omitted
from the second angular velocity estimating unit 22. Note
that the other configuration is the same as or equivalent
to that in FIG. 7, and the same or equivalent components
15 are denoted by the same reference symbols and redundant
description thereof will be omitted.
[0103] A PI controller 24 included in the first angular
velocity estimating unit 21D performs computation
processing expressed by the above equation (5). In other
20 words, the first angular velocity estimating unit 21D
generates the first estimated angular velocity ω^
r1 only in
PI control without using an integrator, and outputs the
first estimated angular velocity ω^
r1 to the estimated
angular velocity calculator 23. Similarly, a second
25 angular acceleration estimating unit 30 included in the
second angular velocity estimating unit 22D also generates
the second estimated angular velocity ω^
r2 without using an
integrator, and outputs the second estimated angular
velocity ω^
r2 to the estimated angular velocity calculator
30 23. Subsequent operations are as described in the first
embodiment.
[0104] Because the speed estimating device 101D
according to the third embodiment includes the second
34
angular velocity estimating unit 22D, the speed estimating
device 101D can estimate high-frequency speed pulsation
more accurately than the speed estimating devices of the
first and second comparative examples. The reasons thereof
5 are as described in the first embodiment.
[0105] The accuracy of speed estimation in the third
embodiment, however, is lower than that in the first
embodiment. In terms of the computation amount required
for computation for estimation, however, the third
10 embodiment is more advantageous because the computation of
integration is omitted. Therefore, in a case where the
computation performance of the processor 901 illustrated in
FIG. 11 is low and the calculation amount is to be as small
as possible, the configuration of the third embodiment is
15 more preferable. Although details will be provided later,
in a case where speed pulsation suppression control
described in a fifth embodiment is performed, however, the
configuration of the speed estimating device 101 according
to the first embodiment is more preferable.
20 [0106] In addition, for an example similar to the third
embodiment, a configuration may be realized in which the
first angular velocity estimating unit 21D includes the
integrator 25 and the second angular velocity estimating
unit 22D does not include the integrator 31. Alternatively,
25 another configuration may be realized in which the first
angular velocity estimating unit 21D does not include the
integrator 25 and the second angular velocity estimating
unit 22D includes the integrator 31.
[0107] Fourth Embodiment.
30 FIG. 14 is a block diagram illustrating a
configuration of a speed estimating device 101E according
to a fourth embodiment. In FIG. 14, the speed estimating
device 101E according to the fourth embodiment includes a
35
second compensation phase computing unit 56 and a third
angular velocity estimating unit 33 in addition to the
configuration of the speed estimating device 101 according
to the first embodiment illustrated in FIG. 1. In addition,
5 the estimated angular velocity calculator 23 is replaced
with an estimated angular velocity calculator 23E. In
other words, two angular velocity estimating units are
provided in the first to third embodiments, but the fourth
embodiment is directed to a configuration having three
10 angular velocity estimating units. Note that the other
configuration is the same as or equivalent to that in FIG.
7, and the same or equivalent components are denoted by the
same reference symbols and redundant description thereof
will be omitted.
15 [0108] Typically, the characteristics of angular speed
pulsation included in the rotational angular velocity of an
AC motor vary depending on an application applied to the
motor or on a load device connected to the AC motor. Now
in a case where a connected load device has periodic torque
20 fluctuation, a rotary compressor is considered by way of
example.
[0109] FIG. 15 is a graph illustrating an example of a
waveform of load torque of a rotary compressor. The
horizontal axis represents rotation angle, and the vertical
25 axis represents load torque. Herein, the number of
compression chambers in the rotary compressor is
represented by k. 0 to 360 degrees of the rotation angle
corresponds to one cycle period of a mechanical angle, that
is, a mechanical angle period.
30 [0110] First, in a case where only one compression
chamber is present, that is, in a case of k=1, the load
torque vibrates greatly with the mechanical angle period as
illustrated by a solid curve in FIG. 15. Although second
36
and third harmonics are also included in the load torque
waveform, the first-order vibration is the greatest.
Therefore, in a case where a configuration in the first to
third embodiments is applied, the largest first-order
5 angular speed pulsation can be estimated with high accuracy
by setting the disturbance frequency fd used for
computation of the second estimated angular velocity ω^
r2 to
a primary frequency of the mechanical angular frequency.
[0111] In the fourth embodiment, a plurality of angular
10 velocity estimating units are provided in parallel.
Therefore, speed pulsation due to second-order and thirdorder torque fluctuations included in the load torque
characteristics can also be estimated with high accuracy.
In the example of FIG. 14, the frequency of speed pulsation
15 to be estimated is used as a second disturbance frequency
fd2, which is estimated by the third angular velocity
estimating unit 33 and outputted as a third estimated
angular velocity ω^
r3.
[0112] The same is applicable to a case where the number
20 of compression chambers is two or three, that is, a case of
k=2 or k=3. As the number of compression chambers is
larger, the structure is more complicated and the cost is
higher, but the waveform has smaller pulsation as
illustrated in FIG. 15. Specifically, in the case of k=2,
25 a second harmonic component of the mechanical angular
frequency is large, and in the case of k=3, a third
harmonic component is large.
[0113] In the case of k=2, for example, the second-order
vibration in the mechanical angle period is dominant as
30 illustrated in FIG. 15. Therefore, the disturbance
frequency fd inputted to the second angular velocity
estimating unit 22 is set as a secondary frequency of the
mechanical angular frequency. Then, if higher-order
37
frequencies than the secondary frequency are to be further
estimated, these higher-order frequencies may be inputted
as the second disturbance frequency fd2 to the third
angular velocity estimating unit 33.
5 [0114] In the case of k=3, for example, the third-order
vibration in the mechanical angle period is dominant as
illustrated in FIG. 15. Therefore, the disturbance
frequency fd inputted to the second angular velocity
estimating unit 22 is set as a tertiary frequency of the
10 mechanical angular frequency. Then, if higher-order
frequencies than the tertiary frequency are to be further
estimated, these higher-order frequencies may be inputted
as the second disturbance frequency fd2 to the third
angular velocity estimating unit 33.
15 [0115] It is noted that a plurality of angular velocity
estimating units are provided in parallel and each of the
angular velocity estimating units performs phase
compensation in the fourth embodiment, but the present
invention is not necessarily limited by this example.
20 Phase compensation may be performed by at least one of the
angular velocity estimating units, and this single
compensation is sufficient for producing the effects
peculiar to phase compensation described above.
[0116] Fifth Embodiment.
25 FIG. 16 is a block diagram illustrating a
configuration of a driving device 102 for an AC motor
according to a fifth embodiment. The driving device 102
according to the fifth embodiment is a driving device
configured to control the AC motor 2 with use of the speed
30 estimating device 101, 101C, 101D, or 101E described in the
first to fourth embodiments. FIG. 16 illustrates a
configuration to which the speed estimating device 101
according to the first embodiment is applied, as an example.
38
[0117] As illustrated in FIG. 16, the driving device 102
according to the fifth embodiment includes a speed
controlling unit 5, an adder 7, a torque controlling unit 6,
a compensation torque command computing unit 8 that is a
5 compensation amount computing unit, and the speed
estimating device 101. The compensation torque command
computing unit 8 operates as a “compensation command
computing unit”.
[0118] First, the operation of the compensation torque
10 command computing unit 8 will be explained. Note that by
way of example, the following description is given for a
configuration in which the compensation torque command
computing unit 8 performs computation on the basis of
information on angular acceleration computed by the second
15 angular velocity estimating unit 22.
[0119] The compensation torque command computing unit 8
computes a compensation torque command τ*
rip by using the
following equations (22) to (24).
[0120]
20 [Formula 22]
∙∙∙(22)
[0121]
[Formula 23]
∙∙∙(23)
25 [0122]
[Formula 24]
∙∙∙(24)
[0123] In the equations (22) and (23), Ksi_rip represents
an integral gain of the compensation torque command
  c
si _ rip
c 0 ˆ
s
K
T 
  







 s
si _ rip
s
ˆ 0
s
K
T 
  







T sin2 f t T cos2 f t rip  s  d  c  d 

39
computing unit 8. In addition, Tc in the above equation
(22) represents the amplitude of the compensation torque
command τ*
rip corresponding to the cosine component of the
pulsation of the angular acceleration, and Ts in the
5 equation (23) represents the amplitude of the compensation
torque command τ*
rip corresponding to the sine component of
the pulsation of the angular acceleration. As expressed by
the above equations (22) and (23), a compensation torque
command τ*
rip is computed such that each of the cosine
10 component of the pulsation of the angular acceleration and
the sine component of the pulsation of the angular
acceleration is zero. Control using this compensation
torque command τ*
rip can reduce the pulsation of the angular
acceleration, and as a result, can reduce the speed
15 pulsation as well.
[0124] The reason why integral control is used in the
equations (22) and (23) is that the characteristic of the
controlled object is a proportional characteristic when the
compensation torque command τ*
rip is obtained on the basis
20 of the angular acceleration. In addition, it is because an
ideal closed loop characteristic can be achieved by causing
a controller to have an integral characteristic and
performing feedback control. In the case of a
configuration like the second angular velocity estimating
25 unit 22 in which the compensation phase θpls determined by
the compensation phase computing unit 51 is used,
compensation can be made for the phase of the equation (24)
on the basis of the compensation phase θpls. Alternatively,
when the compensation phase θpls is a compensation phase
30 based on the disturbance frequency fd, the computation may
be performed using an equation other than the equation (24).
Although the concerned explanation is not provided because
the principle thereof is the same, the computation of the
40
compensation torque command described above can be
similarly performed with use of the second estimated
angular velocity ω^
r2 computed by the second angular
velocity estimating unit 22.
5 [0125] Next, the operations of the speed controlling
unit 5, the torque controlling unit 6, and the adder 7 will
be explained.
[0126] The speed controlling unit 5 computes a basic
torque command τ*
ω on the basis of an angular velocity
command and the estimated angular velocity ω^ 10 r. Speed
control performed by a typical PI controller can be applied
to the computation of the basic torque command τ*
ω.
[0127] The adder 7 adds the compensation torque command
τ
*
rip to the basic torque command τ*
ω to compute a torque
command τ* 15 according to the following equation (25).
[0128]
[Formula 25]
∙∙∙(25)
[0129] The torque controlling unit 6 includes the
20 voltage applying unit 3 illustrated in FIG. 11. The torque
controlling unit 6 determines a voltage vector to be
applied to the AC motor 2 on the basis of the torque
command τ*. The voltage vector may be of a type computed
through electric-current control such as PI control on the
25 basis of a current command value computed on the basis of
the torque command τ*. Alternatively, an appropriate
voltage command value depending on the torque command τ*
may be stored in the memory 902 and obtained directly on
the basis of the torque command τ*.
30 [0130] The driving device 102 according to the fifth
embodiment can obtain a compensation torque command to
reduce the speed pulsation on the basis of information on
 


    rip
41
the angular speed pulsation obtained by the speed
estimating device 101. This produces an effect of reducing
uneven rotation of the AC motor 2.
[0131] While FIG. 16 illustrates a configuration
5 including the compensation torque command computing unit 8
that computes the compensation torque command τ*
rip, the
present invention is not limited to this configuration.
Another configuration may be realized in which a
compensation current command computing unit that computes a
10 compensation current command is employed instead of the
compensation torque command computing unit 8. In the case
of this configuration, an adder and a current controlling
unit are provided on a subsequent stage of the torque
controlling unit 6. The adder adds a basic current command
15 generated by the torque controlling unit 6 to the
compensation current command computed by the compensation
current command computing unit to generate a current
command. The current controlling unit determines a voltage
vector to be applied to the AC motor 2 on the basis of the
20 current command outputted from the adder. Subsequent
operations are as described above.
[0132] Sixth Embodiment.
FIG. 17 is a block diagram illustrating a
configuration of a driving device 102A for an AC motor
25 according to a sixth embodiment. In FIG. 17, the AC motor
2 illustrated in FIG. 16 is replaced with a refrigerant
compressor 2a including the AC motor 2. The driving device
102A according to the sixth embodiment is configured using
the speed estimating device 101 according to the first
30 embodiment in order to reduce the speed pulsation of the
refrigerant compressor 2a. FIG. 17 illustrates a
configuration to which the speed estimating device 101
according to the first embodiment is applied, but the
42
present invention is not limited to this example. The
driving device 102A may be configured using any of the
speed estimating devices 101C, 101D, and 101E described in
the second to fourth embodiments. Note that the
5 configurations and the functions of the speed estimating
devices 101, 101C, 101D, and 101E are as described above,
and so description thereof is omitted in this part.
[0133] Next, the structure of the refrigerant compressor
2a and a load torque in the refrigerant compressor 2a will
10 be described in detail with reference to FIGS. 18 and 19.
FIG. 18 is a cross-sectional view illustrating an outline
structure of the interior of the refrigerant compressor 2a
illustrated as a driven object in FIG. 17. In addition,
FIG. 19 is a cross-sectional view illustrating a structure
15 of the interior of a compressing unit 202 of the
refrigerant compressor 2a illustrated in FIG. 18. Note
that a refrigerant compressor called a rolling piston type
rotary compressor will be described herein, but the present
invention is not limited to this example. The refrigerant
20 compressor may be another type of compressor such as a
scroll compressor.
[0134] The refrigerant compressor 2a includes an
airtight container 211, the AC motor 2 housed in the
airtight container 211, a shaft 201 having one end passing
25 through a rotor 2-1 constituting the AC motor 2, the
compressing unit 202 through which the other end of the
shaft 201 passes and which is fixed to the inside of the
airtight container 211, an inlet pipe 203 provided to the
airtight container 211, and an outlet pipe 204 provided to
30 the airtight container 211.
[0135] A stator 2-2 of the AC motor 2 is attached to and
retained by the airtight container 211 by shrinkage fitting,
freeze fitting, or welding. Electric power is supplied to
43
a coil 2-3 on the stator 2-2 via an electric wire, which is
not illustrated. The rotor 2-1 is disposed inside the
stator 2-2 with a gap 2-4 therebetween, and rotatably held
by a bearing, which is not illustrated, via the shaft 201
5 situated at the center of the rotor 2-1.
[0136] In the refrigerant compressor 2a having the
configuration as described above, upon driving of the AC
motor 2, refrigerant sucked into the compressing unit 202
via the inlet pipe 203 is compressed, and the compressed
10 refrigerant is discharged out from the outlet pipe 204.
The refrigerant compressor 2a often has a structure in
which the AC motor 2 is immersed in the refrigerant, and it
is difficult to attach a position sensor to the AC motor 2
because of wide swing in temperature thereon. For this
15 reason, in the refrigerant compressor 200, the AC motor 2
has to be driven in a position sensorless driving manner.
[0137] As illustrated in FIG. 19, the compressing unit
202 includes an annular cylinder 212, a piston 205 formed
rotatably and integrally with the shaft 201 and located
20 inside the cylinder 212, and a compression chamber 213
provided in an inner circumferential part of the cylinder
212.
[0138] The cylinder 212 includes an inlet 206
communicating with the inlet pipe 203 illustrated in FIG.
25 18, and an outlet 207 through which the compressed
refrigerant is discharged outward. The inlet 206 and the
outlet 207 communicate with the compression chamber 213.
The cylinder 212 includes a vane 210 that partitions the
compression chamber 213 into a low-pressure sub-chamber
30 communicating with the inlet pipe 203 and a high-pressure
sub-chamber communicating with the outlet 207, and a spring
209 set to energize the vane 210.
[0139] The shaft 201 connects the AC motor 2 and the
44
piston 205 with each other. The piston 205 is eccentric,
so that the capacities on the suction side and the
discharge side change depending on the rotation angle. The
refrigerant sucked through the inlet 206 is compressed by
5 the piston 205. When the pressure in the compression
chamber 213 increases, a discharge valve 208 opens, and the
refrigerant is discharged through the outlet 207. When the
refrigerant is discharged, refrigerant flows into the
suction side at the same time. As the rotation of the AC
10 motor 2 is continued, the refrigerant is discharged once
per revolution in the mechanical angle of the piston 205.
[0140] The load torque pulsation of the refrigerant
compressor 2a corresponds to periodic disturbance for the
AC motor 2, and so becomes a factor for speed pulsation.
15 It is generally known that greater speed pulsation in the
refrigerant compressor 2a results in greater noise and
vibration.
[0141] There is a significant point that the frequencies
of the load torque pulsation and the speed pulsation are
20 determined by the structure of the refrigerant compressor
2a, and so are known in advance. In the refrigerant
compressor 2a according to the sixth embodiment, the
control system illustrated in FIG. 17 is built under favor
of that point. The refrigerant compressor 2a estimates a
25 specific frequency component of the speed pulsation with
high accuracy in the second angular velocity estimating
unit 22, and computes such a compensation torque command
τ
*
rip that suppresses the pulsation in the compensation
torque command computing unit 8. As a result, the speed
30 pulsation can be reduced without preconditioning. Because
preconditioning becomes unnecessary, the cost for
conditioning prior to shipment can be significantly reduced,
then resulting in great usefulness.
45
[0142] Seventh Embodiment.
FIG. 20 is a diagram illustrating a configuration of a
refrigeration cycle apparatus according to a seventh
embodiment. The refrigeration cycle apparatus 300
5 illustrated in FIG. 20 includes the driving device 102 for
an AC motor, the refrigerant compressor 2a, a condenser 301
connected with the refrigerant compressor 2a via piping 305,
a liquid receiver 302 connected with the condenser 301 via
the piping 305, an expansion valve 303 connected with the
10 receiver 302 via the piping 305, and an evaporator 304
connected with the expansion valve 303 via the piping 305.
The evaporator 304 is connected to the inlet pipe 203.
[0143] By the refrigerant compressor 2a, the condenser
301, the liquid receiver 302, the expansion valve 303, the
15 evaporator 304, and the inlet pipe 203 being connected by
the piping 305, the refrigerant compressor 2a, the
condenser 301, the liquid receiver 302, the expansion valve
303, the evaporator 304, and the inlet pipe 203 constitute
a refrigeration cycle circuit 306 in which the refrigerant
20 circulates. In the refrigeration cycle circuit 306,
processes of evaporating, compressing, condensing, and
expanding of the refrigerant are repeated, and heat is
transferred while the refrigerant repeatedly changes from
liquid to gas and from gas to liquid.
25 [0144] The functions of the devices constituting the
refrigeration cycle apparatus 300 will be explained. The
evaporator 304 evaporates refrigerant liquid in a lowpressure state, draws heat from the surrounding, and thus
has a cooling effect. The refrigerant compressor 2a
30 compresses refrigerant gas into high-pressure gas in order
to condense the refrigerant. The refrigerant compressor 2a
is driven by the driving device 102A according to the sixth
embodiment. The condenser 301 releases the heat to
46
condense the high-pressure refrigerant gas into refrigerant
liquid. The expansion valve 303 subjects the refrigerant
liquid to throttle expansion into low-pressure liquid in
order to evaporate the refrigerant. The liquid receiver
5 302 is provided for adjusting the amount of refrigerant to
be circulated, and may be omitted in a case of a compact
apparatus.
[0145] Typically, improvement in quietness and reduction
in cost are required for a refrigeration cycle apparatus.
10 In a household refrigeration cycle apparatus, there are
particularly high demands for lower cost, and so a single
rotary compressor is often used. A single rotary
compressor is a rotary compressor described in FIGS. 18 and
19, which is a compressor of a type including only one
15 compression chamber 213. A rotary compressor involves
significantly great load torque pulsation, and thus tends
to cause large vibration and loud noise. On the other hand,
in a feedforward control system in the conventional art,
complicated control adjustment has needed for reducing
20 vibration and noise.
[0146] The refrigeration cycle apparatus 300 according
to the seventh embodiment performs feedback control so that
the driving device 102A automatically makes the speed
pulsation be zero. As a result, the cost for adjustment
25 before shipping can be significantly reduced. In addition,
according to the seventh embodiment, speed pulsation is
reduced by feedback control, thereby enabling flexible
response to variations in manufacture, fluctuation in motor
constant, and changes in compressor load condition.
30 Consequently, the refrigeration cycle apparatus 300 that
has higher environment resistance can be achieved.
[0147] The configurations presented in the above
embodiments are examples of contents of the present
47
invention, and can each be combined with other publicly
known techniques and partly omitted and/or modified without
departing from the scope of the present invention.
5 Reference Signs List
[0148] 2 AC motor; 2a refrigerant compressor; 3
voltage applying unit; 4 current detecting unit; 5 speed
controlling unit; 6 torque controlling unit; 7 adder; 8
compensation torque command computing unit; 11 model
10 deviation computing unit; 12 state estimator; 13
subtractor; 14 deviation calculator; 21 first angular
velocity estimating unit; 22, 22B, 22C, 22D second angular
velocity estimating unit; 23 estimated angular velocity
calculator; 24, 27, 28 PI controller; 25, 31, 32
15 integrator; 26, 52 Fourier coefficient calculator; 29, 55
AC restoring unit; 30, 30B, 30C second angular
acceleration estimating unit; 33 third angular velocity
estimating unit; 51 compensation phase computing unit; 53,
54 I-controller; 56 second compensation phase computing
20 unit; 101, 101-1, 101A, 101B, 101C, 101D, 101E speed
estimating device; 102 driving device; 200 refrigerant
compressor; 201 shaft; 202 compressing unit; 203 inlet
pipe; 204 outlet pipe; 205 piston; 206 inlet; 207
outlet; 208 discharge valve; 209 spring; 210 vane; 211
25 airtight container; 212 cylinder; 213 compression
chamber; 300 refrigeration cycle apparatus; 301
condenser; 302 liquid receiver; 303 expansion valve; 304
evaporator; 305 piping; 306 refrigeration cycle circuit;
901 processor; 902 memory.
48
We Claim :
1. A speed estimating device for an AC motor, the speed
estimating device comprising:
5 a model deviation computing unit to compute a model
deviation on the basis of a voltage, a current, and an
estimated angular velocity of the AC motor;
a first angular velocity estimating unit to compute a
first estimated angular velocity on the basis of the model
10 deviation;
a second angular velocity estimating unit to compute a
second estimated angular velocity on the basis of the model
deviation, the second estimated angular velocity differing
from the first estimated angular velocity in frequency;
15 a compensation phase computing unit to compute a
compensation phase on the basis of a disturbance frequency;
and
an estimated angular velocity calculator to compute an
estimated angular velocity of the AC motor on the basis of
20 the first and second estimated angular velocities, wherein
either one of the first and second estimated angular
velocities is computed on the basis of the compensation
phase.
25 2. The speed estimating device for an AC motor according
to claim 1, wherein
the first estimated angular velocity has a frequency
lower than a frequency of the second estimated angular
velocity,
30 the first angular velocity estimating unit computes
the first estimated angular velocity on the basis of the
model deviation, and
the second angular velocity estimating unit computes
49
the second estimated angular velocity on the basis of the
model deviation, the compensation phase, and disturbance
frequency.
5 3. The speed estimating device for an AC motor according
to claim 2, wherein
the compensation phase computing unit computes the
compensation phase in view of a phase characteristic of the
first angular velocity estimating unit.
10
4. The speed estimating device for an AC motor according
to any one of claims 1 to 3, wherein
the second angular velocity estimating unit includes:
a specific frequency extractor to extract a specific
15 frequency component of the model deviation on the basis of
the disturbance frequency and the compensation phase; and
a specific-frequency angular velocity estimator to
compute the second estimated angular velocity on the basis
of the specific frequency component.
20
5. The speed estimating device for an AC motor according
to any one of claims 1 to 3, wherein
the second angular velocity estimating unit includes:
a specific frequency extractor to extract a specific
25 frequency component of the model deviation on the basis of
the disturbance frequency; and
a specific-frequency angular velocity estimator to
compute the second estimated angular velocity on the basis
of the specific frequency component and the compensation
30 phase.
6. The speed estimating device for an AC motor according
to any one of claims 1 to 3, wherein
50
the first angular velocity estimating unit includes:
a first angular acceleration estimator to compute a
first estimated angular acceleration from the model
deviation; and
5 a first angular velocity calculator to compute the
first estimated angular velocity from the first estimated
angular acceleration, and
the second angular velocity estimating unit includes:
a specific frequency extractor to extract a specific
10 frequency component of the model deviation on the basis of
the disturbance frequency and the compensation phase;
a specific-frequency angular acceleration estimator to
compute a second estimated angular acceleration on the
basis of a specific frequency component of the model
15 deviation; and
a second angular velocity calculator to compute the
second estimated angular velocity from the second estimated
angular acceleration.
20 7. The speed estimating device for an AC motor according
to any one of claims 1 to 6, comprising:
at least one other second angular velocity estimating
unit configured similarly to the second angular velocity
estimating unit, wherein
25 at least one of the second angular velocity estimating
units computes the second estimated angular velocity on the
basis of the compensation phase.
8. A driving device for an AC motor, the driving device
30 comprising:
the speed estimating device for an AC motor according
to any one of claims 1 to 7, wherein
the driving device determines a voltage to be applied
51
to the AC motor on the basis of a current flowing in the AC
motor and the estimated angular velocity computed by the
speed estimating device.
5 9. The driving device for an AC motor according to claim
8, further comprising: a compensation command computing
unit to compute a compensation current command or a
compensation torque command on the basis of an angular
velocity or an angular acceleration computed by the second
10 angular velocity estimating unit.
10. A refrigerant compressor comprising:
the driving device for an AC motor according to claim
8 or 9;
15 an AC motor to which a voltage is applied by the
driving device; and
a compressing unit in which refrigerant is compressed
by the AC motor.
11. A refrigeration cycle apparatus comprising the
refrigerant compressor according to claim 10.