|
|
|
| United States Patent | 5444341 |
| Link to this page | http://www.wikipatents.com/5444341.html |
| Inventor(s) | Kneifel, II; R. William (Cincinnati, OH);
Walsh; Stephen T. (Cincinnati, OH);
Minnich; Thomas W. (Franklin, OH);
Crovella; Robert M. (Rockford, IL);
Morser; A. Harold (Oxon, GB2) |
| Abstract | Torque ripple in brushless DC motors is reduced by compensation of signals
controlling energization of motor stator windings. Motor current magnitude
is compensated in response to compensation factor values correlated to the
ratio of actual torque to ideal torque at constant current at various
relative positions of a motor rotor and stator. A motor current command
component associated with a commanded torque is compensated for cogging
torque associated with motor magnetization. Measured relative position
values are compensated to account for torque profile distortions
associated with motor current magnitude. |
|
|
|
Title Information  |
|
|
|
|
|
Drawing from US Patent 5444341 |
|
|
Method and apparatus for torque ripple compensation |
|
|
|
|
|
| Publication Date |
August 22, 1995 |
|
|
|
|
|
| Filing Date |
November 4, 1993 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Title Information  |
|
|
Claims  |
|
|
We claim:
1. A method for controlling a brushless DC motor in response to a current
command value defining a desired torque according to a predetermined motor
torque constant, the motor including a rotor moveable relative to a stator
and stator windings electrically energizable to produce torque acting
relatively on the rotor and stator, the method comprising the steps of:
a. producing position signals representing the position of the motor rotor
relative to the motor stator;
b. producing in response to the position signals a torque compensation
factor correlated to a ratio of actual torque to ideal torque produced by
constant motor current at the relative position of rotor and stator
represented by the position signals, said ideal torque being constant over
a range of rotor position;
c. producing a compensated current command value in response to the current
command value and the torque compensation factor, the compensated current
command value defining a magnitude of motor current to produce the desired
torque at the relative position of rotor and stator represented by the
position signal; and
d. controlling a magnitude of current through the stator windings in
response to the compensated current command value.
2. The method of claim 1 further comprising the step of executing a
velocity control algorithm relating a desired velocity and measured
velocity to produce a current magnitude value to which the current command
is correlated, the current magnitude value derived from the magnitude of
difference between actual motor velocity and desired motor velocity.
3. The method of claim 1 wherein the step of producing a compensated
current command value further comprises the steps of:
a. producing a torque compensation value in response to the position
signals, the torque compensation value representing the magnitude of
torque produced in the absence of motor current at the position of the
motor rotor relative to the motor stator represented by the position
signals; and
b. modifying the current command value in response to the torque
compensation value.
4. The method of claim 1 further comprising the steps of:
a. producing commutation pattern control signals in response to changes of
position of the motor rotor relative to the motor stator as determined
from the position signals, the commutation pattern control signals
defining stepwise commutation of patterns of conduction of motor current
through the stator windings to produce torque in a desired direction; and
b. controlling conduction of motor current through the stator windings in
response to the commutation pattern control signals.
5. The method of claim 4 further comprising the step of producing a
compensated position value in response to the current command value and
the position signals, the compensated position value defining a position
at which stator windings to be successively energized produce equal torque
at the current magnitude defined by the current command value.
6. The method of claim 5 wherein the compensated position value is used to
determine the torque compensation factor.
7. The method of claim 5 wherein the commutation pattern control signals
are produced in response to the compensated position value.
8. A method for controlling a brushless DC motor in response to a current
command defining a desired torque, the motor including a rotor moveable
relative to a stator and stator windings electrically energizable to
produce torque acting relatively on the rotor and stator, the method
comprising the steps of:
a. producing position signals representing the position of the motor rotor
relative to the motor stator;
b. producing a torque compensation value in response to the position
signals, the torque compensation value representing the magnitude of
torque produced in the absence of motor current at the relative position
of rotor and stator represented by the position signals;
c. modifying the current command value in response to the torque
compensation value; and
d. controlling a magnitude of energization of the stator windings in
response to the current command value as modified.
9. The method of claim 8 further comprising the steps of:
a. producing commutation pattern control signals in response to changes of
position of the motor rotor relative to the motor-stator as determined
from the position signals, the commutation pattern control signals
defining stepwise commutation of patterns of conduction of motor current
through the stator windings to produce torque in a desired direction; and
b. controlling conduction of motor current through the stator windings in
response to the commutation pattern control signals.
10. The method of claim 9 further comprising the step of producing a
compensated position value in response to the current command value and
the position signals, the compensated position value defining a position
at which stator windings to be successively energized produce equal torque
at the current magnitude defined by the current command value.
11. The method of claim 10 wherein the commutation pattern control signals
are produced in response to the compensated position value.
12. The method of claim 9 further comprising the steps of:
a. producing in response to the position signals a torque compensation
factor correlated to a ratio of actual torque to ideal torque produced by
constant motor current at the position of the motor rotor relative to the
motor stator represented by the position signal;
b. producing a compensated current command value in response to the current
command value as modified and the torque compensation factor, the
compensated current command value defining a magnitude of motor current to
produce the desired torque at the position of the motor rotor relative to
the motor stator represented by the position signal.
13. The method of claim 12 wherein the compensated position value is used
to determine the torque compensation factor.
14. An apparatus for controlling a brushless DC motor in response to a
current command defining a desired torque, the motor including a rotor
moveable relative to a stator and stator windings electrically energizable
to produce torque acting relatively of the rotor and stator, the apparatus
comprising:
a. means for producing position signals representing the position of the
motor rotor relative to the motor stator;
b. means for producing commutation pattern control signals in response to
changes of relative position of the rotor and stator as determined from
the position signals, the commutation pattern control signals defining
stepwise commutation of patterns of conduction of motor current through
the stator windings to produce torque in a desired direction;
c. means for compensating the current command in response to the position
signals, the compensations effected being correlated to deviations of
actual torque from desired torque at the position represented by the
position signals, desired torque being constant over a range of rotor
position in response to constant motor current; and
d. means responsive to the commutation pattern control signals and the
current command as compensated to control energization of the stator
windings.
15. The apparatus of claim 14 wherein the means for compensating the
current command further comprises:
a. means for producing a torque compensation value in response to the
position signals, the torque compensation value representing the magnitude
of torque produced in the absence of motor current at the position of the
motor rotor relative to the motor stator represented by the position
signals; and
b. means for modifying the current command value in response to the torque
compensation value.
16. The apparatus of claim 14 wherein the means for compensating the
current command further comprises:
a. means for producing in response to the position signals a torque
compensation factor correlated to a ratio of actual torque to ideal torque
produced by constant motor current at the position of the motor rotor
relative to the motor stator represented by the position signal; and
b. means for producing a compensated current command value in response to
the current command value and the torque compensation factor, the
compensated current command value defining a magnitude of motor current to
produce the desired torque at the position of the motor rotor relative to
the motor stator represented by the position signal.
17. The apparatus of claim 14 further comprising means for producing a
compensated position value in response to the current command value and
the position signals, the compensated position value defining a position
at which stator windings to be successively energized produce equal torque
at the current magnitude defined by the current command value.
18. The apparatus of claim 14 further comprising means for producing a
current command value correlated to a current magnitude value produced in
response to a velocity control algorithm relating a desired velocity and
measured velocity, the current magnitude value derived from the magnitude
of difference between actual motor velocity and desired motor velocity.
19. The apparatus of claim 17 wherein the means for producing commutation
pattern control signals are responsive to the compensated position value
and the commutation pattern control signals effect commutation at the
position represented by the compensated position value.
20. The apparatus of claim 17 wherein the means for modifying the current
command is responsive to the compensated position value and the current
command is effective to conform torque produced to desired torque at the
position represented by the compensated position value. |
|
|
|
|
Claims  |
|
|
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to control of brushless DC motors. In
particular this invention relates to reduction of torque ripple of
brushless DC motors.
2. Description of Related Art
It is known that brushless DC motors exhibit torque variations during
operation above zero velocity and which are most pronounced at relatively
low speeds and when subject to high loads. Contributors to torque
variations include effects related to motor current as well as effects
associated with motor construction. It is recognized that DC brushless
motors which are driven by stepwise motor current commutation, for
example, six step commutation, demonstrate relatively greater torque
variation than motors designed for and driven with sinusoidal current
energization. However, stepwise commutated current energization is
relatively simpler to achieve than synthesis of sinusoidal motor current.
Consequently it is desirable to effect reduction of torque variation for
stepwise commutated brushless DC motors.
Certain torque variations associated with motor current commutation are
described in U.S. Pat. No. 4,782,272. In particular, this reference
illustrates in FIG. 9 and the associated description the effects of
deviation of current commutation from the point at which torques from
successively energized windings are equal. This reference discloses a
control for effecting commutation at a point beyond the point where the
torque of successively energized windings are equal, resulting in a torque
peak.
U.S. Pat. No. 4,868,477 describes a control for energizing windings of a
variable reluctance motor to reduce torque ripple in positioning servos.
The control of this reference includes storage for data defining the
magnitude of current for each motor winding at each rotor position to
produce a desired torque. The stored data is based on measured motor
performance. The current control method of this reference contemplates the
simultaneous energization of winding phases to achieve a desired torque
magnitude at a particular position and is not suitable for effecting
velocity control.
Motor control techniques for reducing torque variation associated with
certain current related effects and effects produced in the absence of
motor current in axial field permanent magnet motors are described in U.S.
Pat. No. 5,223,775. These control techniques are directed particularly to
modifying the shape and magnitude of applied sinusoidal motor current
where each winding phase is controlled independently.
The known control techniques do not achieve satisfactory reduction of
torque variation for brushless DC motors driven with stepwise commutated
motor current while retaining comparatively simple current commutation.
SUMMARY OF THE INVENTION
It is an object of the present invention to control energization of
windings of a brushless DC motor to reduce torque fluctuations.
It is a further object of the present invention to control a brushless DC
motor to reduce torque fluctuations which are attributed to torque
variations at constant motor current by compensation of motor current
command values.
It is a still further object of the present invention to control a
brushless DC motor to reduce torque fluctuations which are attributed to
motor magnetization by compensation of motor current command values.
It is a still further object of the present invention to control stepwise
commutation of motor current of a brushless DC motor to reduce torque
fluctuations by effecting motor current commutation at the relative
position of rotor and stator at which the constant motor current torques
of successively energized motor windings are equal.
Further objects and advantages of the present invention shall become
apparent from the following description and the accompanying drawings.
A motor control periodically produces motor current commands defining a
motor current magnitude and produces commutation pattern control signals
defining the instant of motor current commutation from one to another of
successively energized motor windings. Each motor current command includes
a current component derived from a torque command value and, for velocity
control, includes a component derived from a velocity control algorithm.
An adjustment angle is used to correct the commutation pattern control
signals to conform the instant of motor current commutation to the instant
when the rotor and stator will have a relative position at which constant
motor current torques of successively energized motor windings are equal.
The adjustment angle is calculated as a function of commanded motor
current magnitude. The control effects compensation of current command
values in two respects. First, a torque command value included in the
motor current command is modified by a compensation component accounting
for torque fluctuations attributable to motor magnetization. Compensation
components are derived from measured motor torque in the absence of motor
current at plural relative positions of the rotor and stator. Stored
values of compensation components are recalled according to measured
relative position of rotor and stator. A second compensation of motor
current command values is effected using compensation factor values
correlated to ratios of actual torque to ideal torque produced by constant
motor current. The control includes data storage of compensation factor
values associated with plural relative positions of motor rotor and
stator. As the relative position of rotor and stator change, compensation
factors are recalled and current commands are modified in response to the
recalled compensation factor values. The measured relative position value
used to recall the stored compensation factor data is modified by the
adjustment angle.
BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1a though 1c illustrate ideal and actual torque characteristics at
constant motor current.
FIG. 2 is a block diagram showing a brushless DC motor and control
therefore.
FIG. 3 is a flow chart of a procedure effected by the control of FIG. 2 for
producing motor current control signals.
FIG. 4 is a flow chart of a procedure effected by the control of FIG. 2 for
producing values used in the procedure of FIG. 3.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
To illustrate the invention, a motor control developed by Cincinnati
Milacron Inc., the assignee of the present invention, shall be described
in detail. While this motor control represents a preferred embodiment of
the invention, the description thereof is not intended to be in any way
limiting in respect of the scope of the invention which is defined by the
appended claims and all equivalents thereof. The motor control to be
described effects correction for torque variation related to motor current
as well as variations associated with motor magnetization.
To describe the effects of motor current related torque variations
corrected by the motor control of the subject application, reference shall
be made to FIGS. 1a-1c which illustrate trapezoidal torque characteristic
for motor windings energized with constant current, i.e. an unchanging
current magnitude. The control of this invention is applicable to stepwise
motor current commutation whether applied to motors having trapezoidal or
sinusoidal torque characteristics.
FIG. 1a illustrates full cycles of ideal torque produced by three phases of
stator windings energized with constant current in response to passage of
opposing poles of a rotor. Motor construction and arrangement of the
windings is designed to produce the staggered torque profiles illustrated.
Each torque profile includes two constant torque portions of opposite
polarity, and each constant torque portion occupies one sixth of the
torque profile. It will be understood that in a motor constructed with six
rotor poles, a complete cycle of a torque profile is completed with each
120.degree. of relative rotation of rotor and stator. Torque
characteristics of sinusoidal type motors exhibit substantially narrower
regions of constant torque but do produce regions of relatively small
torque variation within the corresponding one sixth portion of the torque
profile.
It will be appreciated from FIG. 1a that the motor may be advantageously
operated within the regions of constant torque of the torque profiles by
controlling the selection of stator windings to be energized, and the
direction of current through the windings. To achieve this type of
operation, it is desired to effect commutation of current from one winding
to another at the points 20 where transitions occur from constant torque.
With constant motor current, the ideal characteristics of FIG. 1a would
produce constant torque as phases are successively energized, that is,
torque would remain constant through motor current commutation. In a motor
having a sinusoidal torque characteristic, stepwise commutation of motor
current at the same positional intervals will result in reduced torque
regions between the narrower constant torque portions of the torque
characteristic.
In contrast to the ideal torque profiles of FIG. 1a, actual motor torque
characteristics exhibit a continuous change of torque at the transitions
to and from constant torque. In FIG. 1b curves 14 and 16 are portions of
two torque profiles, corresponding to curves 10 and 12 of FIG. 1a,
spatially overlaid. While commutation will desirably occur at a location
where the torque for phases to be successively energized is equal,
commutation of constant motor current can not produce constant torque
because of the loss of torque in the vicinity of the transitions. This
effect shall be referred to as torque constant error. A torque variation
likewise occurs for sinusoidal motors in the transition regions between
constant torque portions of the torque characteristic. To correct for
torque constant error, the control of this invention effects compensation
of motor current commands in accordance with the ratio of actual torque to
ideal torque at constant current. The compensated current commands
overcome the loss of torque at constant current in the transition regions.
A second current related effect is illustrated by the torque profiles of
FIG. 1c. There, the curves 13 and 15 correspond to the curves 14 and 16 of
FIG. 1b at one constant current magnitude and the curves 17 and 19
similarly correspond at another constant current magnitude. As shown, the
shape of torque characteristic profiles change with current magnitude. As
a consequence of these profile changes, the points 24 and 26,
corresponding to point 22 of FIG. 1b, occur at different relative
positions of rotor and stator. That is, the commutation points are
translated as a function of motor current magnitude. The direction of
translation depends on the direction of current flow producing the torque.
This positional translation shall be referred to herein as torque profile
shift. To correct for this effect, the control of this invention effects
adjustment of the position of occurrence of commutation of current from
one winding to another. The magnitude of adjustment of commutation
position is determined from a linear function joining the desired
commutation points at different torque magnitudes. The linear function is
defined by torque measurement data characterizing actual motor
performance.
In addition to effects on torque profiles attributable to current
magnitude, torque fluctuations attributable to motor magnetization occur
in the absence of motor current. The physical arrangement of rotor magnets
and stator poles are such that magnetic reluctance varies with relative
positions of the rotor and stator. Reluctance variations are manifested in
positions of the rotor and stator from which positional change is resisted
and to which positional alignment is aided by magnetic forces. The
resultant torque variation, is known as "cogging". To correct for this
effect, the control of this invention effects compensation of a motor
current command component having values determined in response to
commanded torque. Compensation values are determined from measurement of
motor torque in the absence of motor current.
Motor Controller
A motor controller according to the present invention shall be described
with reference to FIG. 2. Motor controller 30 includes a signal processing
module 32 and a motor and transducer interface module 34. Current
switching devices located in power block module 36 control delivery of
electrical current to motors 100 through 106 in response to control
signals output by interface module 34. Motor controller 30 is responsive
to inputs received from a host system (not shown) to perform position and
velocity control of motors 100 through 106.
SIGNAL PROCESSING MODULE
Signal processing module 32 includes a host interface 37 for data
communication with a host system. Host interface 37 includes a memory 78
to which data are written for exchanges between processing module
processors or between the processing module and the host system. Host
interface 37 also includes address and data bus interfaces which serve as
signal buffers between digital signal processors of signal processing
module 32 and the memory. Two digital signal processors 38 and 42 execute
programs recorded in respective program stores 40 and 44 to implement
motor control algorithms, perform diagnostic routines and effect
communication with the host system. Applicants have chosen the TMS 320C25
digital signal processors available from Texas Instruments Inc. Servo bus
interface 50 provides signal buffering for signals to be transferred from
the digital signal processors 38 and 42 to motor interface module 34.
Local address and data bus interfaces 46 and 48 provide signal buffers for
address and data signals transferred between program stores 40 and 44 and
the respective digital signal processors 38 and 42.
Motor control routines of interest for the present invention include
position detection routines 56, velocity loop processing routines 52,
commutation control routines 54, torque command processing routines 58,
and gain adjustment routines 60. Of particular interest in respect of
velocity loop processing routines are motor current calculation 62 and
torque constant compensation 64. Of particular interest in respect of
torque command processing routines are torque command compensation
routines 66 and torque command conversion routines 68 which produce a
component of motor current. Processor 1 program store 40 includes a data
table of torque constant compensation factors 70. Processor 2 program
store 44 includes reluctance compensation data table 72.
Signal processing module 32 effects motor control by periodic execution of
routines implementing motor control algorithms at a predetermined loop
closure interval. With each execution of the routines, new values of motor
current commands CMD.sub.I and commutation pattern control words
SEXT.sub.I are produced (the subscripts I identify the motor for which the
commands and control words are effective). The commands and control words
are employed by motor interface module 34 to control the delivery of
electrical current to the motors 100 through 106 during the loop closure
interval. In addition, the commands and control words are made available
to the host system through host interface 37.
MOTOR & TRANSDUCER INTERFACE MODULE
Interface module 34 includes motor drive interface 74 and position
transducer interface 76. Motor drive interface 74 includes digital to
analogue conversion and processing circuitry 80 which is responsive to the
motor current commands ICMD.sub.I, commutation pattern control words
SEXT.sub.I, and measured motor current to produce control signals for the
current switches of power block module 36. The switches are controlled by
pulse width modulation to achieve the desired motor current magnitudes.
Motor drive interface 74 includes input/output interface circuitry 82 for
exchange of signals with power block module 36. A specific implementation
for generating pulse width modulated switch control signals and for
exchanging signals with power block module 36 is disclosed in U.S. Pat.
No. 5,187,417.
In addition to generation of switch control signals, interface module 34
includes signal interfaces for position transducers 110 through 116.
Position transducer interface 76 includes excitation signal interface 84
and transducer output signal receivers 86. In the preferred embodiment,
position transducers are resolvers. Resolver excitation signals, fixed
frequency sinusoids, are produced by resolver excitation generator 88 and
conditioned by excitation signal interface 84 for output to the position
transducers. Resolver output signals, SIN.sub.I and COS.sub.I, are sampled
at the loop closure interval and the sampled values converted from
analogue to digital representations by analogue to digital converter 90.
The resolver interfaces of the preferred embodiment are of known design
and further details thereof shall not be provided herein.
Velocity Loop Control Algorithms
Digital processing module 32 effects velocity control to determine a motor
current command value. It will be appreciated that velocity control may be
effected as a subsisting control algorithm in a system effecting position
control. The motor current command value produced by the velocity control
defines a desired motor torque. The desired motor torque includes a
component calculated to achieve a desired velocity and a component
associated with torque that may be otherwise commanded. The velocity
dependent component is calculated in response to a velocity error VE,
i.e., the difference between a velocity command VCMD defining a desired
motor velocity and actual motor velocity. In the preferred embodiment,
actual instantaneous motor velocity is derived from the measured change of
position occurring during the immediately preceding loop closure interval.
Velocity control is selectively effected according to either proportional
plus integral or pseudo derivative feedback type control algorithms.
Proportional plus integral velocity control is characterized by a control
transfer function having the form:
KP*[(s+KI)/s]*VE
Where:
s represents complex frequency
KP=proportional gain factor
KI=integral gain factor
##EQU1##
Where: VCMD=velocity command
VACT=actual velocity (scaled and compensated for velocity loop control)
Pseudo derivative feedback velocity control is characterized by a transfer
function having the form:
KPDF*[KI/s]*VE-(KPDF,VACT)
Where:
s represents complex frequency
KPDF is a gain factor equivalent to the velocity gain factor in
proportional plus integral control and the other quantities have the same
definitions as in the case of proportional plus integral control.
The signal processing module 32 effects the selected type of control using
effective values of the gains KI, KP, and KPDF derived from user supplied
values which may be transferred to the motor control 30 from the host
system. Adjustment of the gain values is accomplished through adjustment
parameters PADJ, PDFADJ,and KIADJ in accordance with the following:
KP(eff)=PADJ*KP(nominal)
KI(eff)=KIADJ*KI(nominal)
KPDF(eff)=PDFADJ*KPDF(nominal)
To provide increased velocity loop gain at and near zero velocity, a
further gain modification factor KRO is applied in accordance with the
following:
KP=(1+KRO)KP(nominal)
KI=(1+KRO)KI(nominal)
KPDF=(1+KRO)KPDF(nominal)
The change of gain during low speed operation is effected at a constant
rate. The value of KRO is a user specified friction compensation
multiplier which is input from numerical control 10.
A current command is produced in response to the selected type of control
taking into consideration the axis moment of inertia, the motor torque
constant and the current scaling associated with the pulse-width
modulation control signals to be applied to the power block 36. In
addition, a current component produced in response to a direct torque
command input is summed with the current command component derived from
the velocity loop control algorithm. Inclusion of a torque command
component ITCMD in the algorithm permits entry of values from a host
system thereby providing direct control to achieve a desired torque.
Torque commands other than null may be used, for example, at speeds where
velocity error based control can not be relied on exclusively to produce
satisfactory results, for example, where velocity errors are quite small.
Applicants have chosen to effect compensation for torque variations
related to motor magnetization by including a compensation component in
torque command TCMD.
The calculations performed to produce the current command implementing
proportional plus integral control are:
ICMD=IKP*VE.sub.k +IKI*I.sub.k +ITCMD.sub.k (1)
Where:
##EQU2##
Where: I.sub.k-1 =integrator evaluation of the k-1 update interval
##EQU3##
Where: C1=constant accounting for conversion of units including conversion
of time to time per sample interval and the multiplier of KIADJ
KIADJ=integral gain adjust factor
##EQU4##
Where: AS=power block current scale factor
C2=constant accounting for conversion of units and the multiplier of KPADJ
JA=axis moment of inertia
JADJ=moment of inertia adjust factor
KPADJ=proportional gain adjust factor
KRO=low speed gain modification factor
KT=motor torque constant (torque/amp)
ITCMD=current component associated with direct torque command
VE.sub.k =VE at the k update interval
The calculations performed to produce the current command implementing
pseudo derivative feedback control are:
ICMD=-IKPDF,VACT+PKI*I.sub.k +ITCMD.sub.k (2)
Where:
##EQU5##
and all other components have the definitions stated above.
Values of the axis moment of inertia JA, the torque constant KT, and the
current scale factor AS may be supplied from the host system. Processors
38 and 42 perform integer arithmetic calculations to execute the velocity
control algorithms. For clarity, the expressions of the algorithms are
given without scale factors used to accommodate integer arithmetic. It
will be understood that the expressions define relationships among
variables and that calculations according to these relationships may be
performed using floating point arithmetic whereby scaling required for
integer arithmetic would be eliminated.
Torque Compensation
Applicants have chosen to effect compensation for torque constant error as
described with reference to FIG. 1b by applying a factor to the current
command value produced by evaluation of the selected velocity loop control
algorithm. The compensated current command is computed according to
relative rotor and stator position and an empirically determined
compensation factor according to the following expression:
ITCMP=ICMD+ICMD*TCF.sub..crclbar.CMP (3)
Where:
ITCMP is the value of a current command compensated for torque constant
errors
TCF.sub..crclbar.CMP =torque compensation factor at angular position
.crclbar., where .crclbar. has been compensated for motor current
magnitude
A value for the torque compensation factor TCF.sub..crclbar.CMP is recalled
from a table using position information indicating the relative position
of rotor and stator.
Compensation of commanded current for a calculated ratio could be effected
by multiplication of the uncompensated current value by the inverse of the
ratio of actual torque to ideal torque. However, deviations from ideal
torque to be compensated are relatively small so that the significant
difference between values will reside in fractional portions of the
values. Consequently, expression (3) has been chosen by applicants to
implement the compensation calculation so that the significance of these
small differences can be realized with reasonable data word lengths. Each
stored factor value is computed by subtracting one from the inverse of the
ratio of actual torque to ideal torque. Expression (3) is the equivalent
of multiplying ICMD by the sum of one and the compensation factor value.
Since the compensation factor value is computed by subtracting one from
the inverse of the ratio, the effective addition of one according to
expression (3) restores the one previously subtracted. The result of
expression (3) considering the previous subtraction is multiplication of
ICMD by the inverse of the ratio.
Actual torque values used to compute compensation factor values are
determined by measurements made at constant current at different relative
positions of rotor and stator. Statistical methods are used to eliminate
anomalies in the m | | |