IEEE Transactions on Energy Conversion, Vol. 14, No. 3, September 1999
419
-
EQUIVALENT LOAD TEST FOR INDUCTION MACHINES THE FORWARD
SHORT CIRCUIT TEST
D.H. Plevin
C.N. Glew
J.H. Dymond, SM IEEE
GEC ALSTHOM LARGE
MACHINES LIMITED
Leicester Road, Rugby
Warwickshire CV21 1BD
England
GEC ALSTHOM LARGE
MACHINES LIMITED
Liecester Road, Rugby
Warwickshire CV21 1BD
England
GE Motors and Industrial Systems
107 Park Street North
Peterborough, Ontario
Canada K9J 7B5
ABTRACT : Often full load testing of induction machines with
outputs greater than 1 MW present difficulties when attempting
to verify the rated performance, especially the temperature rise.
This is because of limitations imposed by the manufacturer's
power supplies and load equipment. The Forward Short Circuit
Test, as an equivalent load test, is an alternate loading method
which most manufacturers and service facilities could employ
with a minimum capital expenditure. The theory of the method is
described and temperature rise test results from a 1960 kW 4 pole
machine are compared with test results obtained using several
other test methods.
equivalent loading, heating, forward short circuit,
losses, temperature
Keywords:
I. INTRODUCTION
The factory testing of induction machines at full rated
conditions fulfills the needs of manufacturers and the
customer alike in so far as:1. The test results demonstrate to the purchaser that the
contractually agreed performance has been achieved.
2. The test results demonstrate to the manufacturer that
his predictive performance calculation methods are
accurate thus lowering the business 'risk during the
tendering process.
Established business patterns have demonstrated that
the number of motors produced decreases nearly
exponentially as the machine power output increases. It is
also an established fact that the cost of provision of
PE-196-EC-0-2-1998 A paper recommended and approved by the
IEEE Electric Machinery Committee of the IEEE Power Engineering
Society for publication in the IEEE Transactions on
Energy
Conversion. Manuscript submitted July 17, 1997; made available for
printing March 2 , 1998.
motor test plant follows the opposite trend.
The
combination of very high cost of test plant with the lower
utilisation of test resources presents a significant business
dilemma to motor manufacturers. As a consequence, many
test methods [l, 2, 31 have been devised, which attempt by
various means to produce accurate performance verification
without the need for prolonged operation at rated power.
For large-output induction machines, non-standard
frequencies, or vertical arrangements, the two alternate
methods that are commonly used are superposition and
equivalent loading. Variations of these principles are
included in International and National Standards in varying
degrees for instance IEC 34-2 [4]and IEEE 112 [ 5 ] .
II. SUPERPOSITION
The principle of test superposition relies on the premise
that full loss and the corresponding temperature rise can be
created in each relevant part of the machine separately,
using the same cooling conditions. Hence, the full rated
loss and its corresponding temperature rise can be
accurately derived by superimposing the individual results.
Its advantage, as a low power simulation, is countered by
the fact that it is a multi-test method which in practice
requires scaling of the results as a minimum to obtain the
answer. This principle is commonly used in testing
generators.
111. EQUIVALENT LOADING
In contrast to the previous method, and offering distinct
advantages, the equivalent load test is designed to permit
the temperature rise of the stator winding at rated load to be
determined by a single test conducted at less than rated
load. The intent of the test is to replicate as far as possible
the rated load loss distribution in the motor using
equipment and a test set-up much smaller and less
expensive than would be required for a conventional rated
load test.
Such a test is the Forward Short Circuit Test (FSC)
0885-8969/99/$10.00 0 1998 IEEE
420
sometimes called the Forward Stall Test and the subject of
this paper. The test as described has been carried out by the
authors’ companies for many years and in consultation with
world-wide customers has found specific acceptance. The
test method is not currently included in any International or
National Standard but has been briefly described in an
unpublished internal subcommittee document of IEC SC 2G
numbered as (2G(Secretariat)62 dated July 2, 1993).
induction generator at a speed well beyond the rated pullout
point are shown in Figure 2.
A. TEST ANALYSIS
The behaviour of an induction machine subjected to these
tests is described with specific reference to the distribution
of losses. The results of tests are compared with those from
direct loading and other methods.
IV. THE FORWARD SHORT CIRCUIT TEST (FSC)
Referring to Figure 1, the machine under test, TM, is
driven at rated speed, N, by an auxiliary driving motor DM1 of a size which is approximately 10 percent of the rated
output of TM. The motor, TM, is connected electrically to a
separately excited alternator, G, which is driven by a drive
motor DM-2 at a speed designed to produce a frequency of
80 to 85 percent of the rated frequency of TM. If the
excitation of the alternator, G, is gradually increased and its
frequency, Ft is adjusted by changing the speed of its
driving motor, DM-2, the motor under test, TM, will begin
to operate as an induction generator at a negative slip
greater than its normal pullout point. Under these
conditions of operation, the rotor resistance is relatively
small compared to the starting resistance. The machine
impedance consists predominately of the reduced frequency
leakage reactance which results in only a small alternator
voltage being required to circulate rated current through the
motor, TM.
Field
Winding
Figure 2. Torque and current as a function of speed
I I
-
In particular, compared with a conventional rated load test:-
Generator
DM-2
Test Motor
r
TM
G
V,Ft
DM-I
Figure 1. Test equipment set-up for FSC Test.
As the voltage applied to TM is increased .from zero, the
input power to the alternator drive motor, DM-2, should
first reduce and then reverse. If the opposite occurs, the
phase sequence should be changed.
The torque and current relationships of a 1960 kw 4 pole
6600 volt machine machine under test operating as an
1. The friction and windage loss is the same.
2. The cooling is the same.
3. The fundamental stator copper loss is the same,
4. The rotor copper losses are higher.
5 . The harmonic or pulsation losses are higher.
6. The fundamental frequency iron loss is lower.
In addition to providing a reliable measurement of the
rated stator winding temperature rise, the FSC test electrical
measurements can be used to derive an approximate value
of the full load stray load loss and equivalent circuit data for
other performance calculations.
B. CHOICE OF SUPPLY FREQUENCY (Ft)
The supply frequency for the FSC test is determined by
consideration of two factors. First, for the motor under test,
421
TM, the magnitude of the rotor copper loss increases as the
test frequency decreases and secondly, for the drive motor,
DM-1, the required output decreases as the test frequency
decreases (see Figure 3).
2
wri
secondary resistance of low speed motors is so large that
shallow bars are used, with relatively small ACDC
resistance coefficients. Thus under FSC conditions the
increased rotor 12Rloss in high speed motors is mainly due
to rotor frequency skin effect, while in low speed motors,
the increase in rotor loss is mainly due to increased rotor
current.
The increased IzR loss in the rotor does not generally give
rise to any damaging thermal or mechanical problems,
because in general, motors are designed to accept much
higher temperatures during starting and for generators the
cage is designed with very low 12R loss to give high
efficiency. However, care must be taken in the choice of the
FSC test frequency for if it is too low, the increased rotor
12R loss may unduly affect the measured stator winding
temperature.
9
c1
Y
From Figure 2 it can be seen that operation as an
induction generator at a slip in excess of the speed
corresponding to maximum torque can lead to unstable
operation.
C. CHOICE OF DRIVING MOTOR @M-1)
m
Y
Y
36
37
38
39
40
41
42
Figure 3. Tested motor output and losses and driving
motor output as functions of test frequency.
Under FSC test conditions the rotor 12R loss is higher
than at full load due to increased rotor current and the skin
effect in the slot portion of the rotor bars. In fact he FSC
test condition is similar to the normal locked rotor condition
but with reversal of the small in-phase component of
current associated with operation as a generator. The rotor
current is mainly demagnetising and nearly equal to the
stator current and therefore higher than full load current.
This increase in rotor current is consequently small for high
speed (low pole number) machines with their low per unit
magnetising currents, and large for slow speed (high pole
number) motors with their higher per unit magnetising
currents.
For example, if a 50Hz motor is given a FSC test at
4OHz, then, the FSC Test slip is -0.25 based upon the 40Hz
synchronous speed and the rotor current frequency is 0.25 x
40 = 1OHz. High speed motors employ deep bar effect
rotors to achieve satisfactory torquehpeed characteristics, so
that at a frequency of lOHz the ACDC resistance
coefficient can be significant. By contrast, the per unit
Experience suggests that the size of the test drive motor,
DM-1, is usually less than 10 percent of the rated output
for the machine under test. Reference to Figure 3
demonstrates how the choice of the test drive motor is
directly linked to the choice of the test frequency. Rigorous
rules can not be made about these dependant choices,
suffice to say that it is one of the important advantages of
the FSC test that test parameters and equipment can be
choicen to suit the available equipment, motor size and
design.
Although a small same speed induction motor can be used
as the driver, a shunt-excited DC motor is preferred for the
duty because of the ease with which the speed can be
controlled and the input and/or output is measured and/or
calculated.
D. LOSS DISTRIBUTION DURING THE FSC TEST
(1) FRICTION AND WINDAGE LOSS
The test method requires that the motor under test is
driven at rated speed thus ensuring that not only are the
friction and windage losses identical with those of the rated
load test, but also the cooling is identical.
(2) STATOR COPPER LOSSES
With rated stator current flowing the d.c. value of the
stator copper loss is the same as that during a rated load
test, However, because of the reduced frequency, if there
are any measurable eddy losses in the stator copper, then
they will be slightly less. To a first approximation, bearing
422
in mind that this is an equivalent load test, it is usually safe
to assume that the stator copper loss is the same.
(3) FUNDEMENTAL IRON LOSSES
As the fundamental frequency of the power supply to the
motor under test is 80 to 85 percent of the rated freauenw
-~
and as stated in the “Test Dekiption”, the applied voltage
is comparatively low, thus the associated iron loss is very
small and will have minimal effect on the stator
temperature rise. The temperature rise for the missing iron
loss, designated as the difference between no-load heat runs
at rated
and frequency and reduced
and
frequency, can be added to the FSC rise to arrive at a
conservativetotal temperature rise for the machine..
(4) ROTOR CAGE LOSSES
As explained earlier under ‘‘Choice of Supply Frequency”
in this paper, the rotor cage loss is higher than at a rated
load test and its magnitude is a function of the choice of test
power supply frequency.
(5)
be almost identical for the rated load and the FSC test
conditions. The harmonic fields generated by the stator
contain both permeance harmonics and MMF harmonics.
Because the permeance harmonic loss is voltatre deoendant.
it follows that during the FSC test its magnitude will be
small.
- .
Tooth pulsation losses have a significant effect on the
stator temperature rise, particularly for high speed machines
where it is possible for the total stator tooth harmonic loss
at rated load to be many times the fundamental frequency
tooth iron loss. Christofides and Adkins [6,7] state that
stray load losses are mainly due to the MMF harmonics and
how the MMF harmonic field strength is limited by the
saturation of the iron due to the main field. Therefore, in
the FSC test, the losses in the rotor teeth caused by t h e G
harmonics are larger than those present during fuil load due
to the negligible saturation of the core as a
of the low
applied voltage, The rotor slot harmonic pole pairs, ph, and
the corresponding stator referred frequencies, Fh, are given
by :-
STRAY LOAD LOSSES
With the lower than rated frequency of the test power
supply, the stray load losses in the stator teeth supports and
the clamping plates together with the stator winding eddy
current losses are less than their value at rated load.
However, as these losses are calculable and minimized in
the design of most modern machines, a further reduction in
an already small quantity will have negligible effect on the
stator temperature rise.
The loss in the rotor bars due to stator winding MMF
harmonics are only slightly different from those at full load.
For a given stator current, these losses are proportional to
the bar ACDC resistance coefficient at the induced bar
current frequency. For example, a 50 Hz motor with rated
slip of O.Olp.u., FSC tested at a supply frequency of 40 Hz
and driven at rated speed, the comparative frequencies of
the MMF induced bar current are as shown in Table 1
below.
Ph = [mf 1]P
(1)
*
Fh = [(m1)( 1 -S) +s]F~
(2)
where P = Stator winding pole pairs
n = 1 and 2 (higher orders can be ignored).
Ft = Stator supply frequency (Hz).
R = Number of rotor slots.
s =Per unit slip.
By inspection of the above tabulation it can be clearly
seen that the resulting ACDC resistance coefficients would
Table 1. Rotor Bar Current Frequencies
Phase
Harmonic
-5
Bar Current Frequency Hz
Belt I F.L. 50 Hz
I FSC 40 Hz
I 297.5
I 287.5
296.5
594.5
593.5
306.5
584.5
603.5
-0.2375
n
Pole
Pairs
Frequency, Hz
Full load
1
-48
1187.5
52
1
1287.5
2 I -98
-2425
21
102
2525
Slip p.u. 0.01
-
FSC Test
- 1197.5
1277.5
- 2435
2515
-0.2375
423
V. DERIVATION OF LOSSES AND
EQUIVALENT CIRCUIT PARAMETERS
Assuming the iron loss of the machine being tested by the
FSC test is negligibly small as discussed earlier, then the
losses of the machine under test are:-
Stator copper loss
Is2RS
Rotor cage loss
12RR
For example, FSC test results on a 12.5 MW 4 pole
induction motor gave the results summarized in Table 3
which also include the results for the same parameters
calculated from the tests required by the IEEE 112 Method
F.
Table 3 Comparison of Test Parameters
Parameter
FSC test
results
IEEE 112
Method F
Is
1360.3
0.01196
0.0379
0.769
66.4
211.6
147.2
1358.3
0.01196
0.0382
0.772
66.2
207.1
135.0
Windage & friction loss PNL
Stray losses
WS
The sum of these losses is the difference between the
output power, Po, of the test machine drive motor, DM-1,
and the generated power, Po, of the machine under test,
TM, and as shown in (3):PO
- PO = 12%+ I’RR + WS + PNL
(3)
RS, ohms
Rz, ohms
XO,ohms (XI + X z )
I&, kW
I’RR, kW
Ws, kW
The power transmitted across the air gap of the machine
under test is the sum of the generated power output, PO,and
stator copper losses, I?&.
This total power when
multiplied by the slip, s, is equal in magnitude to the rotor
cage loss, 12RR,hence, (4):-
I ~ =RSPT~ = S ( P ~ + 1 2 ~ s )
(4)
If (4)is substituted into (3), then the single unknown, the
stray load loss, WS, is calculated from the result (5)
-
Ws = Po Po(l+s) - I2Rs ( l + ~-)PNL
(5)
The rotor resistance for the test condition can be calculated
from the rotor cage loss, 12RR,and stator current, IS, if the
small numerical difference between stator and rotor currents
is neglected. The resistance, R2, derived by (6) is usually
higher than that at full load due to skin effect and higher
bar temperature.
RZ= 12RR/312
(6)
In order to achieve adequate accelerating torque margin,
particularly in the region of 80 percent speed, the design of
cage bars in large induction motors must take account of
supply limitations (allowable voltage dips) and regularly use
bars with a decreasing skin effect with speed. For such
machines, the test-derived value of referred rotor resistance
is sufficiently close to the full load value to enable its use in
the calculation of running performance.
With full-load current flowing through the windings, the
leakage fieldcurrent ratios correspond to full load, except
for a small reduction due to rotor bar skin effect affecting
the embedded bar portion. Therefore, reactance calculated
from generated power, voltage and current, corrected to
rated frequency, agrees closely with the total leakage
reactance at full load.
0
m
_.
0 _.
l-
o _.
\o
0
v)
0
d
..
__
a ..
8
190
I Amperes
I
I+
195
200
205
210
Figure 4. Heating test temperature rises.
215
220
424
VIII. ACKNOWLEDGEMENTS
From Figure 4 it is clearly seen that the line drawn
through the FSC test Embedded Temperature Detector
(ETD) temperature rises agrees very closely with the line
drawn through the direct load test temperature rises. The
line drawn for the mixed frequency tests is approximately
6OC higher and is not an untypical result for that test.
The lines drawn through the cooling air temperature rises
for the three tests show a similar pattern to that shown by
the ETD temperature rise curves.
VII. CONCLUSIONS
This paper has demonstrated that the Forward Short
Circuit Test is an economical equivalent loading test that
can be applied to establish accurate temperature rise
performance for induction machines of any size. In
particular, it offers a unique method for testing those
products whose size is such that full load testing is
impossible. In addition to providing accurate thermal
performance, it is possible to derive data from the test from
which some machine equivalent circuit parameters can be
obtained. In the example chosen for test, the FSC test is not
only more accurate than the mixed frequency test but
because of the steady power supply conditions, the vibration
can be monitored throughout the test to determine the
sensitivity of the rotor to thermal effects created by the
change in temperature of the rotor cage as discussed in
Reference 8.
The FSC Test has been used for over 40 years in the
United Kingdom by GEC Alsthom and its constituent
companies on induction motors up to 21 MW. Extensive
investigation into the method was also done by GE Motors
in Canada on a range of machines from 225 kW to 12.5
MW and a speed range from 128.57 to 3600 rpm. Some of
these latter investigations are reported in reference 8.
Suppy Frequency
I DL
I 50
I DL
I 50
In 1988, Denis Plevin prepared a paper on this subject but
his premature death prevented its completion. The current
authors have edited, revised and rewritten parts of the
original manuscript and prepared this paper as a record of
the theory of the technology, giving due acknowledgement
to the original work of Denis Plevin in this field of
induction motor testing.
Messers Dymond and Glew would like to thank the
Directors of their respective companies for permission to
publish this paper. Mr. Glew would particularly like to
acknowledge the help and assistance provided by Eur Ing
Graham Le Flem and Mike Tarkanyi.
E.REFERENCES
1. Ytterherg. A., Ny metod for fnllhelasting av electriska maskiner utan
drivmotor eller avlastningsmaskin, s.k. shakprov. Technisk lidskrifl 1921, p
42-46,79.
2. Romeira, M.P., The superimposed Frequency Test for induction motors,
AIEEE Trans 1948 Vol. 67, pt. 11 pp 952-955.
3. Fon& W, New temperature test for polyphase induction motors by phantom
loading, Proc. IEE 1972 Vol. 119, pp 883-887.
4. IEC 34-2 , Standard Part 2: Methods for determining losses and efftciency
of rotating electrical machinery from tests (excluding machines for traction
vehicles), International Electrotechnical Commission, Geneva, Switzerland,
1972
5 . IEEE 112-1996, IEEE Standard Test Procedure for Polyphase Induction
Motors and Generators, IEEE, New York, 1996
6. Christofides, N., Origins of load losses in induction motors with cast
aluminium rotors, hoc. IEE 1965 Vol. 112, p 2317.
7. Christofides, N. and Adkins, B., Determination of load losses and torques
in squirrel-cage induction motors, Proc. IEE 1966 vol. 113, pp 1995-2004.
8. Dymond, J.H.,
Forward stall test an alternate method of rotor and stator
loading for temperature and vibration verification, IEEE IAS Trans. Vol. 31 ,
SeptIOct 1995 pp 1153-1158.
I MF I MF I MF IMF I FSC I FSC I FSC I FSC I FSC
I 50/40 I 50/40 I 50/40 I 50/40 I 40 I 38
I 38 I 38 I 38
Table 4. Heating test results. DL = Direct Loading, MF = Mixed Frequency., FSC = Forward Short Circuit.
425
X. BIOGRAPHIES
D. H. Plevin joined The English Electric Company Stafford,
England as a trainee in October 1943. He progresses via the roles
of Engineering Assistant and AC Design Engineer to be the Chief
Designer of Induction Machines in 1968. He was elected a
member of The Institutionof Electrical Engineers (MIEE) in 1956.
He transferred to GEC Large Machines Ltd. at Rugby, England in
1969 where he was employed as Section Engineer and Principal
Engineer up tp his premature death in 1990.
He devised the application of the Forward Short Circuit Test as an
economical alternative to full load testing of large induction
machines in the Company and this paper contains his contribution
to the theory of the test and is presented by his co-authors as a
memorial to his work.
C. N. Glew completed an apprenticeship with Metropolitan
Vickers Electrical Company, Manchester, England, following
graduation from Leeds University in 1959 with an honours degree
in electrical Engineering.
He joined AEI Motor and Control Gear Division in 1961 as a
design engineer and has since occupied the positions of Section
Engineer, Chief Development Engineer, Design Manager and
Engineering Manager with AEI and its sucessors GEC Machines
Ltd. and GEC Alsthom Large Machines Ltd.
He was elected a member of The Institution of Electrical
Engineers (MIEE) in 1967 and a Fellow (FEE) ~ f1990,
l
a Fellow
of The Institution of Mechanical Engineers (MlMechE) in 1994
IEC and Canadian standards and working groups. He has
written and presented a number of papers on testing, insulation
systems, rotor failures, transient heating, shaft currents and
machine sparking.
Mr. Dymond is a senior member of IEEE Power Engineering
Society and the Industry Applications Society. He is a
Registered Professional Engineer in the Province of Ontario.
and is currently Chairman of The Rotating Electrical Machines
Association (REMA) Technical Committee in the United Kingdom
and Chairman of CENELEC Technical Committee 2 in Europe
covering rotating electrical machines,
He has held responsibilities for the design of all types of and sizes
of low and high voltage induction machines. He has written and
presented technical papers on the subjects of machine noise,
standards, environment, new products, insulation systems,
machines for hazardous areas and application criteria.
J. H. Dymond received the B. Sc. degree in mathematics from
Memorial University of Newfoundland,Canada in 1965 and the B.
Eng. degree in electrical engineering (power) from the technical
University of Nova Scotia, Canada, in 1967.
In 1967, he joined General Electric Canada, Inc., Peterborough, in
their Engineering Training program in induction motor design.
From 1968 to 1972, he worked on a postgraduate project on lasers
at Memorial University. In 1972, he returned to General Electric
Canada, Inc. where he has worked on indiuction motor design and
development. He has worked on projects covering air-to-air cooler
design for motor, induction and synchronous machine ventilation,
electromagnetic noise and resonance, windage noise, transient
heating and cooling of induction machines, insulation systems,
product structuring, harmonic heating effects, starting calculations
and methods, induction machine test methods, national and
international standards, and general induction machine design. He
holds a number of Canadian and U. S. pat and GE patent awards.
He is a contributor, member or chair of a number of IEEE, MI,