United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/SR-96/008
May 1996
EPA Project Summary
Efficiency Optimization Control of
AC Induction Motors: Initial
Laboratory Results
M.W. Turner, V.E. McCormick, and J.G. Cleland
A fuzzy logic, energy optimizing con-
troller has been developed to improve
the efficiency of motor/drive combina-
tions which operate at varying loads
and speeds. This energy optimizer is
complemented by a sensorless speed
controller which maintains motor shaft
revolutions per minute (rpm) to pro-
duce constant output power. Efficiency
gains of from 1 to 20% are obtained
from laboratory demonstration with
commercial motors and drives. Motor
shaft rpm is controlled to within 0.5%.
The energy optimizing controller used
for "vector control" adjustable speed
drives is complemented by a torque
pulsation control scheme to rapidly
damp vibrations.
This Project Summary was developed
by EPA's National Risk Management
Research Laboratory's Pollution Pre-
vention and Control Division, Research
Triangle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Electric motors use over 60% of the
electrical power generated in the U.S. The
U.S. population of approximately 1 billion
(109) motors consume over 1700 billion
kWh per year. Each year, 140 million new
motors are sold. More than 80% of the
electricity used by motors is consumed by
less than 1% of the motor population [mo-
tors greater than 20 hp (142 kW)]. Each
1% improvement in motor efficiency could
result in savings of over $1 billion per
year in energy costs, 6-10 million tons
(5.4-9.1 million tonnes) less per year of
combusted coal and approximately 15-20
million tons (13.6-18.1 million tonnes) less
carbon dioxide released into the atmo-
sphere.
Adjustable speed drives (ASDs) are
power electric devices which allow control
of the speed of rotation of electric motors.
ASDs can provide a significant savings in
energy for motors which spend a portion
of their duty cycle operating at less than
their rated speed and torque. Prior to the
introduction of ASDs, control of motor-
driven devices such as fans and pumps
were always controlled by valves, vanes,
dampers, and other mechanical devices,
which are inherently inefficient.
Conventional ASDs do not optimally
minimize motor input power at any given
motor speed and load torque. The objec-
tive of the research described in this re-
port has been to improve ASDs by adding
controls which optimize the ASD on the
basis of energy efficiency. The research
and development program was defined by
the following precepts:
New controls must work with com-
mercial ASDs. Controllers should be
able to be added to existing ASDs
and/or integrated into new ASD de-
signs.
Motors of interest are those rated from
5 to 5000 hp (3.7 to 3730 kW).
Steady-state operation of large mo-
tors is emphasized.
Simple, low-cost design is emphasized.
The controllers must reduce energy
consumption "significantly." A reduc-
tion in overall energy use of 2% is
targeted for motors with rated effi-
ciencies above 85% (large motors).
-------
Controls are physically integrated into
the motor/ASD configuration as shown in
Figure 1. The qualitative interactive perfor-
mance of these components is the same
at essentially any scale. Energy optimiz-
ing controllers interface with the ASD to
minimize line power consumption. In ac-
tual applications, the direct-current (dc)
brake is replaced by an actual load such
as a pump or conveyor belt. The dyna-
mometer is a research tool used to mea-
sure and maintain a specific simulated
load.
Controllers that have been developed
under this project for two kinds of adjust-
able speed drives used with ac induction
motors are
1) ASD Type 1: These ASDs control
the frequency and voltage supplied
to a motor by maintaining a ratio of
voltage to frequency that is the same
as the ratio at the motor's rated con-
ditions (e.g., 208V at 60 Hz, 104V at
30 Hz). These drives are best for
steady-state operation (where load
or speed fluctuations are only occa-
sional or the dynamics are slowon
the order of minutes rather than mil-
liseconds or seconds).
2) ASD Type 2: These ASDs control
frequency and current by indirect
vector control. These drives are bet-
ter for dynamic operation and con-
trolling speed under rapidly chang-
ing load conditions, such as control
of smaller motors in manufacturing,
positioning, and computer-aided de-
sign/manufacturing machining.
All efficiency optimization under this
project is based on a fundamental ap-
proach-the voltage (for ASD Type 1) or
Fuzzy Logic
Energy Optimizer
Feedback
* Voltage
* Current
* Frequency
* Power
Developed Torque
Motor Load Signal
Control Signals
* Voltage/ Current
* Frequency
Stator
Temperature
Line
1 Power I (230 VAC)
Dynamometer
Controller
Adjustable
Speed Drive
Adjust AC
Voltage,
Current,
Frequen.
T
AC
Induction
Motor
Optical Shaft
Encoder
Flexible
Coupling
DC Power
Dynamometer/
DC Brake
Figure 1. Motor research laboratory main component layout.
-------
the current (for ASD Type 2) is perturbed in
a manner that decreases the motor's input
power while the motor output power is main-
tained constant. By these means, the core
losses of the machine decrease while the
copper losses increase until the combined
core and copper losses reach a minimum
value, as shown in Figure 2.
Fuzzy Logic and Control Designs
The mathematical technique called fuzzy
logic offers a new approach to improving
ASD voltage/frequency/current control.
Fuzzy logic has evolved from an esoteric
branch of mathematics into a useful engi-
neering tool. By virtue of its adaptability, it
can be applied to problems whose non-
linearity and dynamic nature makes them
intractable to solution via classical control
methods. Motor control has all of the at-
tributes of this class of problems. Fuzzy
logic has been implemented in this devel-
opment of improved motor control because:
1) Fuzzy logic overcomes the math-
ematical difficulties of modeling highly
non-linear systems;
2) Fuzzy logic responds in a more
stable fashion to imprecise readings
of feedback control parameters, such
as the dc link current and voltage;
and
3) Fuzzy logic control mathematics and
software are simple to develop and
flexible for each modification.
Three interactive efficiency-optimizing (in-
put power minimizing) controllers have been
developed for Type 1 ASDs. These control-
lers are 1) voltage perturbation for input
power minimization, 2) speed correction,
and 3) slip compensation.
The voltage perturbation controller is
based on changes in input power and sta-
tor voltage. Fuzzy logic control has been
emphasized for voltage perturbation. The
fuzzy logic membership functions for both
inputs and the output are partitioned using
five fuzzy sets. The input variables are
AVsold and APin; the output variable is
AVS . Triangular fuzzy sets are used for
(using speed corrector control)
Torque
Speed
DC Link Power
Stator Voltage
Total Machine
Loss
Copper
Loss
Converter
Loss
Iron Loss
Efficiency Optimized Time
Operating Point
Figure 2. Changes in core and copper losses with changing flux.
-------
both inputs and outputs, with a restriction
that the output fuzzy sets must be isosce-
les to simplify defuzzification.
Membership functions and the associ-
ated rule set are shown in Figure 3, where
input and output values are represented
linguistically (i.e., NM=negative medium,
NS=negative small, ZE=zero, PS=positive
small, and PM=positive medium). The rule
base table can be read according to the
following example: If the last voltage change
(AVSo|d ) is a "positive small" value and the
measured input power change (APin) is a
"negative small" value, then AVsnew is "posi-
tive small."
Speed correction control is needed be-
cause the perturbation approach alters mo-
tor speed and output power. The motor's
output rotor speed should be maintained
as constant as possible. For Type 1 ASDs,
a fuzzy logic speed corrector controller was
designed to correct for the speed change
with voltage perturbation. The fuzzy speed
controller uses voltage, commanded speed,
measured frequency, and measured volt-
age to estimate the best new frequency
setting.
Slip compensation, has also been devel-
oped to further reduce motor power con-
sumption. For many motor ASD applica-
tions, whenever the frequency is set, a
higher than desired rotor speed results,
using more power. For example: If an op-
erator wishes to reduce speed to 50% of
the rated value, the operator sets the fre-
quency from 60 to 30 Hz. However, with
-0.16-0.12-0.08-0.04 0
AV,.
0.04 0.08 0.12 0.16
-0.06 -0.04 -0.02
0.02 0.04 0.06
'old
A P.
in
AV,.
'old
A P.
in
0
-0.4 -0.3 -0.2 -0.1 0
AV,,
0.1 0.2 0.3 0.4
new
NM
NS
ZE
PS
PM
NM NS ZE PS PM
NM
NS
ZE
PS
PM
NS
NS
ZE
PS
PS
ZE
ZE
ZE
ZE
ZE
PS
PS
ZE
NS
NS
PM
PM
ZE
NS
NM
Figure 3. Membership functions and rule set for fuzzy voltage perturber.
-------
the frequency change, the slip, s, of the
motor also changes. Slip is defined as
Rotor speed/Frequency = 1-s. In this re-
search, the slip compensation control
mathematically estimates the slip which
will result from a given change in fre-
quency, and adjusts frequency to give
the desired percent speed.
An indirect vector controlled induction
motor drive with an efficiency optimiza-
tion controller is shown in Figure 4. All
the control functions indicated by the
dashed outline are implemented in real
time by a single digital signal processor.
The feedback speed control loop gener-
ates the active or torque current com-
mand iqs*. The vector rotator receives the
torque and excitation current commands
iqs* and ids* from one of the two positions
of a switch: the transient position (1) or
the steady-state position (2). The fuzzy
controller becomes effective at steady-
state condition; i.e., when the speed loop
error Acor approaches zero.
A feed-forward pulsating torque com-
pensator has been developed to prevent
speed ripple and mechanical resonance
during transient operation. As the excita-
tion current is reduced in adaptive steps
by the fuzzy controller, the rotor flux de-
creases exponentially. The decrease of
flux causes loss of torque, which nor-
mally is compensated for slowly by the
speed control loop.
Efficiency optimization control is effec-
tive only at steady-state conditions. A dis-
advantage of this control mode is that the
transient response becomes sluggish. For
any change in load torque or speed com-
mand, fast transient response capability
of the drive can be restored by establish-
ing the rated flux. Therefore, for ASD
Type 2, the system starts in efficiency
optimization and then switches to tran-
sient response optimization in the event
of a load disturbance or a change in set
speed. During non-steady-state condi-
tions, the system establishes the rated
magnetizing current (Figure 4 switch in
Position 1).
230 V. 3<|>
60 Hz
IM
Figure 4. An indirect vector controlled induction motor drive incorporating the proposed efficiency
optimization controller.
-------
Results and Recommendations
Figure 5 is a captured screen from a
real-time demonstration of the optimizing
and speed controllers for a Type 1 ASD.
The motor load being measured and con-
trolled simulates a pump or fan running at
90% of rated speed and 81% of rated
torque. At each step, a speed correcting
controller compensates for changes in
speed with changes in input frequency. Ul-
timately, the input power is reduced from
about 81 to about 78% of rated input power.
The speed controller has been shown to
hold speed during efficiency optimization to
within 0.5%. Figure 6 illustrates controller
behavior over several pump-fan load con-
ditions tested in the laboratory. Slip com-
pensation was not active in these tests.
0.825
0.765
0.00
1.50
3.00 4.50
Time (Mins)
6.00
7.50
Figure 5. Efficiency optimization results for 90% speed and 81% torque.
40:16 50:25 60:36 70:49 80:64
% Rated Speed : % Rated Torque
Figure 6. Percent change in motor speed from initial motor speed, without slip compensation.
6
90:81
-------
Typical power savings due to slip com-
pensation for a 10 hp motor are shown in
Figure 7. A total of 1-2% of rated power is
saved.
The operation of the pulsating torque
compensation scheme for the indirect vec-
tor control drive system is illustrated in Fig-
ure 8. An ASD initially operating in a steady-
state mode has its command speed
changed from 450-900 rpm. At 3 seconds,
the system has already reached a new
steady-state mode with rated current rees-
tablished. A new search for the optimum
efficiency point is initiated. The drive speed
response demonstrates the adequacy of
the method for fast transient applications.
40:16 50:25 60:36 70:49 80:64
Speed : Torque
90:81
95.90
Figure 7. Power reduction in watts due to slip compensation (rated input power = 8477W).
OA
0 rpm
OA
OA
(a)
(b)
Figure 8. Drive performance in time domain with sudden changes in command speed.
a) Top: current (3.33 A/div.); Bottom: speed (3.5 rpm/div.).
b) Top: ids* (3.33 A/div.); Bottom iqs* (3.33 A/div.).
Time scale: 5 sec./div.
-------
Figure 9 contains efficiency curves for
the Type 2 ASD, where the dotted curves
are results obtained by optimal control with
the fuzzy controller and the solid curves
represent standard drive control. At light
load torque, efficiency gains on the order of
15% may be obtained.
For ASD Type 1, Figures 10 and 11, it is
seen that Motor A exhibits less gain from
efficiency optimization in the 50 to 60%
speed range than at the highest and lowest
output powers. The test data also show
that, in this operating range, voltage pertur-
bation for optimization reverses its direc-
tion; i.e., at the higher speed/torque combi-
nations, voltage perturbation results in volt-
age increases until Fj>n is minimized. At
lower speed/torque combinations, voltage
decreases to optimize. Motor B results are
more common during efficiency optimiza-
tion, with almost no improvement at rated
conditions (100:100). Motor A behavior sug-
gests that the optimum slip of a motor does
not necessarily occur at rated conditions.
The finding is significant because it implies
that, for some motors, input power can be
reduced significantly near rated output
power operation.
Recommendations for continuing efforts
related to the efficiency optimization con-
trollers include:
Final hardware implementation of a
microprocessor integrating all control-
lers into ASD circuitry, and testing
of the hardware configuration.
Further investigation of the influ-
ence of motor design and fabrica-
tion on the operating conditions
where maximum efficiency gains
are found during application of the
new controllers.
Demonstration of the hardware-
implemented controllers in an in-
dustrial setting.
90
80-
s-
0
'o
e eo-
LU
E
to ._
> 5
CO
40-
30
U>r = 0.75 pu
Rated Flux
Optimum Flux
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Load Torque (pu)
Figure 9. Experimental system efficiency curves.
92
90
g88
& 86
0)
o 84
Si
82
80
78
40:16 50:25 60:36 70:49 80:64 90:81 100:100
% Rated Speed : % Rated Torque
40:16 50:25 60:36 70:49 80:64 90:81 100:100
% Rated Speed : % Rated Torque
Figure 10. V/Hz and optimum efficiencies-Motor A.
Figure 11. V/Hz and optimum efficiencies-Motor B.
-------
M.W. Turner, V.E. McCormick, and J.G. Cleland are with Research Triangle
Institute, Research Triangle Park, NC 27709.
Ronald J. Spiegel is the EPA Project Officer (see below).
The complete report, entitled "Efficiency Optimization Control of AC Induction
Motors: Initial Laboratory Results," (Order No. PB96-153 424; Cost: $21.50,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
National Risk Management Research Laboratory (G-72)
Cincinnati, OH 45268
Official Business
Penalty for Private Use $300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
EPA/600/SR-96/008
------- |