IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, VOL. PAS-88, NO.

4, APRIL 1969 281

CONCLUSIONS It is believed that considerable insight for extending generation-

bulk transmission reliability models can be gained from the author's
Using the techniques described it is possible to arrive at a results. With two well-chosen examples, the author has illustrated
measure of steady-state adequacy for any point in a system, the reliability gain with shunt compensation, Fig. 5, and the negligi-
particularly those points at which major transmission terminates ble effect of the independent two-state environmental model,
and subtransmission begins. For an actual system involving Table II.
Hoping that the author may wish to offer additional comments on
many components it may not be necessary to consider all the ways to recognize maintenance considerations in the reliability
possible cases. The digital computer program can be instructed model, we offer the following question and observation. Would it be
to consider up to a fixed number of components out of service at desirable to include recognition of maintenance outage effects in
any one time. The maximum number will depend upon the com- 1) and 2) of the design criterion? The author's suggestion for includ-
ing maintenance considerations appears to cover the scheduled
ponent probabilities. Further work is presently being done at the outage events and incidents of random outages during scheduled
University of Saskatchewan, Saskatoon, Canada, in the appli- outages. The random maintenance model appears to require an
cation of these concepts to more complicated systems. additional parameter, the probability that the component is in the
maintenance outage state.
Our sincerest compliments accompany this discussion.
REFERENCES
[1] R. Billinton, "Bibliography on application of probability methods
in the evaluation of generating capacity requirements," pre-
sented at the IEEE Winter Power Meeting, New York, N. Y.,
January 30-February 4, 1966. Roy Billinton: I would like to thank Dr. Ringlee and Dr. Wood for
[2] R. Billinton, "Transmission system reliability evaluation," their comments. It was intended that maintenance outage effects
Trans. Canadian Elec. Assoc., vol. 6, 1967. would be included in the design criterion, as noted by the discussers,
[3] C. F. DeSieno and L. L. Stine, "A probability method for deter- in terms of random behavior occurring during scheduled outages.
mining the reliability of electric power systems," IEEE Trans.
Power Apparatus and Systems, vol. 83, pp. 174-181, February This presumably would not affect 1) of the criterion, if maintenance
1964. can be scheduled in a relatively light load period. Part 2) of the
[4] I. Bazovski, Reliability Theory and Practice. Englewood Cliffs, criterion would then be determined by considering the various
N. J.: Prentice-Hall, 1961. scheduled network configurations and the corresponding load models.
[51 R. Billinton and K. E. Bollinger, "Transmission system reliability The inclusion of a random maintenance model introduces additional
evaluation using Markov processes," IEEE Trans. Power Ap- complexity to the problem. It also requires the assumption that the
paratus and Systems, vol. PAS-87, pp. 538-547, February 1968. model is truly random, which may not be entirely valid. In regard
to the initial Markov model for two lines in parallel, the random
maintenance model would presumably be added only to the normal
weather side as it is unlikely that maintenance would be con-
ducted if stormy weather is likely to exist. This is an interesting
Discussion area of investigation though probably only of academic interest. I
rather doubt that the utilization of a random maintenance parameter
is valid. I would prefer to think of maintenance effects in terms of
specific configuration and applicable load models in a similar manner
R. J. Ringlee anid A. J. Wood (General Electric Company, Schenec- to that conventionally used in a generating capacity study. We have
tady, N. Y.): This refreshingly brief, yet penetrating probe of a developed a digital computer program to analyze a relatively compli-
significant contemporary problem in power system analysis deserves cated system using the approach outlined in this paper. It is expected
careful attention. The author has offered both a criterion and an that maintenance considerations will be included in these studies.
approach to the analysis of generation-bulk transmission reliability. In conclusion, I would again like to thank the discussers.

Manuscript received February 12, 1968. Manuscript received March 20, 1968.

Creep of Transmission Line Conductors

JOHN R. HARVEY, ASSOCIATE MEMBER, IEEE

Abstract-Results of creep tests of all-aluminum conductors and INTRODUCTION

ACSR at several temperatures are presented. Elevated temperatures CREEP, the nonrecoverable deformation that occurs under
have less influence on the creep of ACSR than on the creep of the
all-aluminum conductors. load, is a physical characteristic of all solids. It increases
with both temperature and stress, but not necessarily in a linear
fashion. Creep begins the instant a load is applied and continues,
Paper 68 TP 65-PWR, recommended and aDDroved by the Trans- usually at a decreasing rate, as long as the load and temperature
mission and Distribution Committee of the IEEE Power Group for are held constant.
presentation at the IEEE Winter Power Meeting, New York, N. Y.)
January 28-February 2, 1968. Manuscript submitted September 13, Since the mid 1930's, the Alcoa graphical method fgr sag-
1967; made available for printing November 22, 1967. tension calculations has included the effect of creep. This method
The author is with Research Laboratories, Aluminum Company
of America, Massena, N. Y uses data obtained from laboratory creep tests to predict the
282 2E8E TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, APRIL 1969

----I IROOM TEMPERATURE-26CI IQOooX~~~TENSION- 20% NULS

10 100 000 I, K
t0 100 000 10000 00000
ELAPSED TIME-HOURS ELAPSED TIME- HOURS
Fig. 1. Creep of Drake at room temperature. Fig. 3. Creep of Drake at elevated temperatures.

10,00 10=
ROOMA EMPERATURE-2C
0
1000 II"-

100ELPE TM-,0OURS 00 [0,00 QO

z
cr
U)
0100
a:
Z~~~~~~~,
V_ 0
2
CL
w
W-L)
oo K)l n m MO tnnoo
,~wv 100.000
lVU,WV
ELAPSED TIME-HOURS
'--
- IA-
ELAPSED TIME-HOURS
Fig. 2. Creep of Arbutus at room temperature. Fig. 4. Creep of Arbutus at elevated temperatures.

amount of creep that will occur on a given transmission line. In ultimate strength (NUS) of the conductor. The tests were con-
recent years, the overload limits of lines have been raised and tinued for 1000 hours and the curves of log creep versus log time
lines are also being designed closer to the thermal limits of the were extrapolated to 100 000 hours. Fig. 2 shows the results of
conductors. This increase in the operating temperature of con- creep tests of 795-kemil 37-strand SAC (Arbutus) obtained at
ductors has generated interest in the effect of these higher tem- the same temperature and tensions of the same percent of
peratures on creep. Alcoa has made laboratory creep tests on nominal ultimate strength as the Drake conductor. Data from
electrical conductors at elevated temperatures since 1957. this type of test have been included in the graphical method of
sag-tension calculations since the mid 1930's.
Room Temperature Creep
Creep at Elevated Temperatures
Laboratory creep tests of conductors are made in horizontal
spans under constant tension and room temperature. A lever- The present trend toward higher operating temperatures of
loading system using ball bearings at the support points is used conductors, higher overload temperatures, and higher tensions
to load the samples and maintain constant tension. The ex- has raised questions about their effects on the creep of conductors.
tensometers, accurate to 10 microstrain (e.g., microinches per Therefore, laboratory creep tests at elevated temperatures on
inch) are mounted remotely from the conductor supporting all-aluminum conductors and ACSR were started in 1957. These
structures. The creep data are corrected for temperature vari- tests are continuing and the results of some of the tests com-
ations that occur during test. pleted to date are presented in this paper. The results of further
Creep data are typically plotted as log creep versus log elapsed tests and a method of using these results in sag-tension calcu-
time. Tests [2] made in the laboratory for the last 35 years lations will be included in a future paper.
indicate that creep of conductors tested at constant tension and A series of 1000-hour tests was made on Drake conductor at a
temperature is a parabolic function of time, that is, the relation- tension of 20-percent NUS and constant temperatures of 500C,
ship between creep and time can be expressed as Ec = mt4, where 750C, 1000C, and 1250C. Fig. 3 shows the results of these tests.
e, is creep at constant temperature and tension, m is a function A series of 1000-hour tests was also made on Arbutus conductor
of tension, temperature and the material, n is a function of the at a tension of 15-percent NUS and constant temperatures of
material, and t is elapsed time, measured from the application 500C, 75°C, 1000C, and 1250C. Fig. 4 shows the results of these
of tension. Therefore, log e, = log m + n log t is a linear equation. tests. The creep curve for each conductor at room temperature
The creep data of interest for sag-tension calculations are and a tension of 20-percent NUS is also included in Figs. 3 and 4
usually the creep after 10 years. The 10-year creep is obtained for comparative purposes.
by extrapolating data obtained from short-time creep tests Creep tests at elevated temperature have also been made on
(usually 1000 hours) to 10 years. The linear relationship between alloy conductors and ACARs. Fig. 5 shows the results of tests
log creep and log elapsed time facilitates this extrapolation and on 927.2-kcmil 37-strand 6201-T81 AAAC (Greeley). One test
improves its accuracy. One test continued for over 8 years and was made at room temperature and a tension of 20-percent NUS
several tests continued for over 5 years confirm the accuracy of and the second test was made at a temperature of 1000C and
this method of extrapolating short-time creep data. a tension of 15-percent NUS. Fig. 6 shows similar results of
Fig. 1 shows the results of creep tests of 795-kcmil ACSR tests on 1021.4-kcmil ACAR 42-EC/19-6201-T81. These tests
26/7 (Drake) tested at mean room temperature of 26°C and were made at tensions of 20-percent NUS, one test at room
tensions corresponding to 20, 40, and 60 percent of the nominal temperature and the second test at a temperature of 1000C.
HARVEY: CREEP OF TRANSMISSION LINE CONDUCTORS 283

TENSION-I 5% NUS
icooc s<-
2500

20001

1500 C
_0 KO K)O 1000 K1000 IOOPO
2z / ARBUTUS
74
ELAPSED TIME-HOURS a
tc
Fig. 5. Creep of Greeley at 2600 and 100°C temperatures. imr
C
K)OO OC
I
La5001
LL
LL ELAPSEDT -RAKE
a
c
IQOA___
(I~
6C
TENSION - 20 % NUS 0 50 m O0
1
ELAPSED TIME-HOURS

Fig. 7. Comparison of creep of ACSR (Drake) and SAC (Arbutus)

during simultaneous testing of conductors through temperature
sequence.
cSlroc
s ____
need at least one year to creep as much at 1250C and 15-percent
0' In NUS tension as it would in 10 years at room temperature and
),- _ 100 -'1000) KMOi0 Xoo
ELAPSED TIME-HOURS 20-percent NUS tension.
Fig. 6. Creep of 1021.4-kcmil ACAR at 260C and 1000C A statistical model of a transmission line was developed in [3].
temperatures. Under the assumed loading conditions for this model, the con-
ductor would accumulate more than 8 hours of operation at
temperatures greater than 150°C in the 30-year life of the line.
Figs. 3-6 show the effect of temperature on the creep of the It would also accumulate more than 29 hours of operation at
various conductors. These figures also show that not much temperatures greater than 1250C in 30 years. Operating at these
difference in creep exists between any of the conductors tested temperatures for the calculated times will not increase the creep
at room temperature at a tension of 20-percent NUS. At 1000C of ACSR over that predicted from room temperature tests, but
and a tension of 15-percent NUS all conductors tested except these operating conditions would produce more creep in all-
ACSR have about the same creep. The curve for the ACAR aluminum conductors than would be predicted from room tem-
at a tension of 20-percent NUS can be converted to a tension of perature creep tests. Therefore, a thermal model of any trans-
15 percent by multiplying by the ratio of the tensions because mission line using all-aluminum conductors must be developed
creep is approximately a linear function of tension in this tension
to predict the accumulated hours of operation at elevated tem-
range.
peratures and a method of using these data in predicting the
Figs. 3 and 4 show that temperature has a much smaller effect creep of the conductors in the line must be developed.
on the creep of ACSR than it has on the creep of all-aluminum Several tests have been made in the laboratory to determine
conductors. This is to be expected for several reasons. Steel has the effect of the time of occurrence of an increase or decrease in
much lower creep than aluminum in the temperature range one of the creep test variables. Conductors were tested at con-
involved in these tests. As the aluminum strands in ACSR stant tension for 1000 hours at 100°C and then the temperature
creep, they transfer load to the steel, thus decreasing the stress was reduced to room temperature and the test continued for
in the aluminum and decreasing the creep of the aluminum. 3000 additional hours. There was no appreciable creep or creep
Moreover, as the coefficient of thermal expansion of aluminum recovery during the 3000 hours at room temperature. Another
is twice the coefficient of steel, the aluminum strands expand test was made on Drake and Arbutus conductors at a tension of
more than the steel at the elevated temperatures. This increases
15-percent NUS. The conductors were tested at room temper-
the load on the steel and decreases the load on the aluminum, ature for 500 hours, the temperature was then increased to
therefore decreasing the creep of the aluminum. There is some 10000 for 1000 hours and then the temperature was increased
temperature at which the aluminum strands would completely to 15000 for 165 hours. Fig. 7 shows the results of these tests
unload onto the steel and the creep at this temperature would plotted on cartesian coordinates.
just be the creep of the steel which is very low. It is this un- Fig. 7 illustrates the great improvement in high-temperature
loading of the aluminum onto the steel that causes the 125°C creep characteristics produced by the steel core in ACSR. Many
curve of Fig. 3 to be very flat and cross the curve for 100°C.
more tests of this type are needed and they are either being
Creep itself is of very little interest to the utility industry, made or will be made as test facilities are available, before a
but the effect of creep on the sag of conductors is of very great rule can be formulated which will take the effect of creep at
interest. Fig. 4 shows that stranded aluminum conductor (SAC) high temperature into account for sag calculations.
will creep as much in about 30 hours at 125°C and 15-percent
NUS tension as it will in 10 years (87 600 hours) at room CONCLUSIONS
temperature and 20-percent NUS tension and, from the trend
shown in the curves, it would appear that SAC would creep as Creep tests at constant tension and room temperature can be
much in a few hours, probably two or three, at 1500C and 15- used to predict the creep of ACSR used in transmission lines.
percent NUS tension as it would in 10 years at room temper- The trend toward high operating temperatures of transmission
ature and 20-percent NUS. ACSR, on the other hand, would lines will have no significant effect on the creep of ACSR. This
284 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, APRIL 1969

trend, however, will have a marked effect on the creep of all- TABLE I
aluminum conductors. A future paper will further explore the DRAKE-795-KCMIL 26/7 ACSR
RATED BREAKING STRENGTH: 31 200 LB
effect of higher operating temperatures on the creep of all-
aluminum conductors and delineate a method of using this 260C
higher creep in sag-tension calculations. Percent Nominal EC Stress, Percent
Ultimate psi Ultimate
REFERENCES
15 4 650 19.4
[1] "Graphic method for sag-tension calculations for ACSR and 20 6 100 25.4
other conductors," Alcoa Aluminum Overhead Conductor Engi- 40 11 630 48.5
neering Data.
[21 R. G. Sturm, C. Dumont, and F. M. Howell, "A method of 60 15 930 66.4
analyzing creep data," J. Appl. Mech., pp. A-62-66, June 1936.
[31 G. M. Beers, H. W. Lis, S. R. Gilligan, and J. M. Schamberger,
"Transmission conductor ratings," IEEE Trans. Power Appa-
ratus and Systems, vol. 82, pp. 767-774, October 1963.

Discussion

J. B. Roche (Kaiser Aluminum and Chemical Sales, Inc., Newark,

Ohio): Mr. Harvey is to be complimented for a very well-written
paper. I do have comments or questions, however.
As defined by the author, creep, the nonrecoverable deformation
that occurs under load, is a physical characteristic of all solids. It
increases with both temperature and stress.
The data presented by the author for single metal conductors,
such as "Arbutus" (795-kemil EC) and for "Greeley" (927.2-kcmil TIME IN HOURS
6201 alloy) would seem to follow the author's definition of creep-
nonrecoverable deformation. The data for these conductors is pre- Fig. 8. Creep of 2493-kemil-54/37 ACAR at room temperature.
pared on the basis of a percent of the rated strength of the con-
ductors which does establish a consistent stress level in the single
metal conductors. gated. It would follow that if a composite conductor is loaded to a
The data for both the ACSR "Drake" and the ACAR construc- tension which is held constant, the elongation measured would be
tions, however, should be examined more critically since the stress not only creep but would include elastic elongation. The elongation
in the component materials which, as the author defines, is responsi- thus plotted would depend on the modulus of the core material as
ble for the creep, is not dependent upon any calculated nominal well as the core metal area.
ultimate strength but is dependent upon the area of each metal in What rightly may be concluded by testing 26/7 ACSR "Drake"
the composite conductor and the modulus of elasticity of the separate may not be true of another construction, such as 45/7 or 18/1
materials. This is defined in the graphic method for sag-tension ACSR, because of the lesser amount of steel and the higher stress
calculations. in the aluminum portion of the conductor.
For example, if we use "Drake" as the author suggests, Table I At this point I believe that the author's conclusion must be ques-
shows the true metal stress in the aluminum component as affected tioned. The author states that "many more tests of this type are
by the conductor temperature for each percentage of the rated or needed and they are either being made or will be made as test
nominal ultimate strength at the test temperatures shown. The facilities are available before a rule can be formulated which will
data assumes that the aluminum and steel lengths are fixed in take the effect of creep at high temperatures into account for sag
relation to each other at room temperature. calculations."
Note that when the conductor is stressed to 15 percent of the Until the author accomplishes this, how can he conclude that
nominal strength, the EC portion is stressed to 19.4 percent of its high operating temperature will affect one type of conductor more
strength. In all cases, the EC is stressed to a higher percent of its than another?
strength than the conductor. It has been reported by Winkleman of Bonneville Power Adminis-
Since the author indicates in the paragraph before the conclusion tration and others, as well as tested by ourselves, that if conductor
that more tests of this type are needed, I wonder if the data for creep is subjected to a high mechanical load and relaxed to a lower ten-
could be better analyzed if the stress level of component materials sion, the conductor does not continue to elongate until some elapsed
were considered since this does determine creep rather than basing time when the conductor approaches its "true" creep curve for the
the data on a calculated strength which in no way indicates the applied tension.
component material stress level. The same graph can be drawn for temperature increased creep
Creep data normally is a plot of elongation increase over some as reported by the author in Fig. 7. Here he reports creep in con-
initial length at the start of a test. I ask the author if this is what he ductor tested for 1000 hours at 1000C at constant tension. After
has plotted in Figs. 1, 3, and 6. From tests we have made, even this, the temperature was reduced to the room temperature with
simple ones such as stress-strain tests of composite conductors and the results that no appreciable creep occurred for 3000 hours.
the associated core tests, it is apparent that during the hold period Refer to Fig. 8. Our data shows this to be true. Fig. 8 shows
there is a greater elongation in the total cable than there is in the conductor stressed in one case to 45 percent of its calculated strength
separate core test. The method for reconciling core test to the and in another case, stressed to 35 percent of its rated strength.
complete conductor is to stress the core to the same elongation at If, after a period of time, the stress is lowered from 45 to 35 percent
the beginning of the hold period. This is assumed to cause the same of the rated strength, the conductor will not creep until after a
psi load as the core would have in the complete conductor at the period of "dead" time. It is apparent that either mechanical or
tension and elongation for the beginning of the hold period. temperature overload will cause prestress in conductor. The amount
The explanation, as verified by other tests, is that the aluminum of creep or prestress depends upon the severity and time of mechan-
creeps and gives up load to the core whether the core is steel or alloy. ical load or the temperature overload and time.
To support this additional load, the core mustfbe elastically elon- The real difference in an aluminum or composite aluminum con-
ductor and ACSR is that the aluminum construction will experience
the creep due to temperature overload and therefore have limited
Manuscript received February 12, 1968. creep when returned to room or operating temperature. ACSR.
HARVEY: CREEP OF TRANSMISSION LINE CONDUCTORS 28 5

does not have this characteristic because at elevated temperatures Although this paper is primarily about conductor creep, another
there is limited, if any, stress in the aluminum portion because of important consideration in operating conductors at elevated tem-
the difference in temperature coefficient of expansion. Since there peratures is the effect of these high temperatures on the strength
is either no stress or limited stress in the aluminum, there is no of the conductors. Were tension tests made after the completion of
prestress effect of the conductor at elevated temperature. If this is creep tests?
true-that there is no prestress of the aluminum in ACSR-ACSR After reading this paper one gets the feeling we should use only
conductor may be expected to creep after a much shorter period of ACSR for transmission lines. Does this mean that all-aluminum or
"dead" time than does an aluminum construction. aluminum alloy combinations should not be considered where creep
Line design must allow for a given magnitude of creep regardless is a factor?
of whether it comes from normal creep at operating temperature,
high stress from mechanical load, or from high temperature from
overload. From discussions we have had with power companies,
we find that in the life of the line there are very few hours when the
line is subjected to temperatures over 100°C. The author cites a H. W. Adams (Reynolds Metals Company, Richmond, Va. 23218):
statistical model of a transmission line that will accumulate 8 This paper represents, to my knowledge, the first published discus-
hours above 150°C and 29 hours above 125°C. By examining the
creep curves as reported by the author for these times and tempera-
sion and data on creep of aluminum conductors at elevated tem-
perature. As the author states, this is a matter of particular concern
ture, one may judge that the effect of this creep generated will be in relation to the trend toward higher operating temperatures of
approximately equal to normal creep at operating temperature. conductors.
Since even the author's data indicates that the creeps are not Reynolds Metals Company, Richmond, Va. has also been con-
cumulative, there is no economical justification for proposing ACSR ducting laboratory creep tests at elevated temperatures for a num-
on the basis of high-temperature creep. ber of years, and we can testify to the frustratingly long periods of
One short word about tests on conductor as long ago as 30 years. time required to collect interpretable and meaningful data. The data
The ordinary sample was prepared at that time with compression presented by the author corresponds rather well with data from
fittings and since the author even now accepts the idea of total Reynolds' laboratory, and we suspect that some of the differences
pounds load in the complete conductor as being the dominant factor, result from differences in test procedure. I should like to ask the
I question the usability of the data. Compression fittings or strain author how his tests were initiated. Were the test specimens first
clamps disturb the true component stress in the conductor. In loaded to the test tension and then heated to the test temperature,
this case the data does not truly represent what takes place in a or was it vice versa?
line. The reason, of course, is that in a test length of 50 feet, the I should also like to ask the author's comments regarding the
slight length addition of aluminum strands caused by compression interpretation of elevated temperature creep data after short periods
fittings is a significant percentage when compared to the same length of elapsed time. Reynolds' investigations indicate that the rate of
if it is added to a field length in excess of 5000 feet. Such a length creep under any condition of temperature and tension is a function
addition in a test length can cause a redistribution of stresses be- of the amount of prior creep that has been imparted-regardless
tween the EC and core and result in questionable data. I ask the of the combinations of time, temperature, and stress that may have
author, then, if his data was based on compression or strain clamp been involved. In actual service a conductor is not likely to be ex-
fittings or if the data has been verified by comparison with data posed to elevated temperatures until after having been in the air
that may be accumulating today using more realistic test fittings. for a substantial period of time. On Fig. 4, for example, what does
75°C represent? Is the
the data for creep in 1 hour at 15 percent and
time of one hour measured after the specimen was first raised to
tension and then heated to temperature, or after it was first heated
and then raised to tension? It seems possible that some creep may
R. L. Retallack (American Electric Power Service Corporation, not have been recorded. Again, what would the data represent in
New York, N. Y.): Some years ago we, at American Electric Power a conductor that had previously been operated at 20 percent and
Service Corporation, questioned the major suppliers of aluminum 262C for, say, 1000 hours?
conductor on research work concerning creep at elevated tempera- With reference to Fig. 3, the data showing the amount of creep
beyond about 3000 hours to be greater 100°C
tures. We did not get much of a response and were contemplating
making our own tests. We are delighted to see that Alcoa has under-
at
125°C
than at
seems implausible. I should like to ask whether these curves were
taken to make such creep tests. We believe the tests, as reported in not extrapolated from data from a shorter test period. If so, it seems
this paper, to be an important contribution to the technology of possible that the extrapolation indicates the curves to cross each
aluminum conductors. other, whereas they may oniy approach each other asymptotically.
Our investigations have shown that creep at elevated tempera- Finally, there is no question but that the creep of ACSR is less
tures is a major consideration in operating conductors under emer- affected by elevated temperatures than are all-aluminum conduc-
gency overload conditions. Certainly, clearances to ground objects tors. Even so, it should not be overlooked that economic and other
or other lines will be reduced by the accelerated creep and by thermal considerations will usually result in normal temperatures that are
expansion of the conductor. This was very dramatically demon- fairly low. The higher temperatures would then be expected to be
strated by the blackout in the eastern region last summer. This associated with emergency operating conditions that would prevail
trouble was initiated by the sagging of a transmission line under for relatively short periods of time. In any event, the prospect is
heavy electrical load into another line. The loss of a line is serious, that in the not too distant future data will be available to permit
but not so serious as the loss of that line during emergency loading the estimation of the amount of creep in any type of conductor
conditions. that may be expected for any assumptions of time, temperature,
However, surviving the emergency may not mean all is well. We and tension over the life expectancy of the line.
may find we have "sacrificed" the line. Because of the increased
creep, the new clearances may be too low for normal operation of
the line. We may be faced with replacement of the conductor-a Manuscript received February 19, 1968.
very costly proposition.
Mr. Harvey's paper provides information for us to avoid such
costly mistakes. We hope this work will be continued by Alcoa and
by others. Temperature levels 150°C
of or higher should be in- J. R. Harvey: It was not the intent of the paper to show that ACSR
vestigated. is a superior conductor for all lines. Transmission line conductors
In the paper, nominal ultimate strength is used as a base of refer- should be selected on the basis of economy and for many lines the
ence for conductor tensions. Others have recommended the use of most economical conductor is one of the all-aluminum conductors
stress levels for the strands as the criteria of comparison. Would The intent of the paper was to show that, under operating conditions
the author comment on this? that include periods of conductor temperature substantially higher

Manuscript re'ceived February 12, 1968. Mnsrp

Manuscript received
eevdJn June 17,
7 1968.
98
2a86 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, VOL. PAS-88, NO. 4, APRIL 1969

than room temperature, the creep of an all-aluminum conductor is predicted from room temperature tests that it has to be accounted
greater than the creep determined from room temperature creep for in sag-tension calculations.
tests. The data are based on tests using compression fittings applied
Creep, as presented in the paper, is the measured increase in by a carefully controlled procedure. Bolted come alongs are used
elongation of the conductor after the load is applied. Elevated tem- to prevent strand movement, and the fittings are compressed in the
perature tests are made by first bringing the conductor to the test reverse direction. Many comparative tests have been made between
temperature and then applying the load. Tests are presently being samples using specially applied compression fittings and the so-
made at a temperature of 1500C, and future tests will be made at an called more realistic fittings, and the results were identical. The
even higher temperature. Tensile tests are made on all samples compression fittings are readily available and more convenient to
after the tests are completed, and these results will be published use than the epoxy-type fittings.
when sufficient data have been accumulated and analyzed. The creep curves used in the paper, as well as the creep data used
The creep datum needed for the design of lines operated at ele- by all conductor manufacturers for sag-tension calculations, are
vated temperatures is the increase in creep due to operation at ele- based on the results of extrapolation of 1000-hour creep test data
vated temperatures over the creep predicted from room tempera- using the parabolic formula for creep. This formula is based on the
ture tests. The creep predicted from room temperature tests is results of long-time creep tests made by Alcoa in the 1930's. The
presently included in sag-tension calculations. Therefore, the fact curves presented are typical of conductors for which the individual
that ACSR has limited creep at elevated temperature and follows strands were drawn from rod that was hot rolled from large size
the room temperature predictions after a much shorter time than the ingots. Wire manufactured by other methods may have different
all-aluminum conductors, is unimportant. As long as the amount of creep. For example, Alcoa Research Laboratories have found that
creep that occurs during operation at elevated temperature is less EC alloy wire made from Properzi rod has lower creep than wire
than the creep predicted from room temperature tests, then this produced from hot-rolled rod. The relationship is not necessarily
creep is already accounted for in sag-tension calculations. It is only the same for other conductor alloys.
when the creep at elevated temperature is greater than the creep I wish to thank the discussers for their fine comments.

A Method for Reducing Dead Time for

Single-Phase Reclosing in EHV

Transmission
HAROLD A. PETERSON, FELLOW, IEEE, AND NARAYAN V. DRAVID, MEMBER, IEEE

Abstract-Most faults on transmission lines are single line to INTRODUCTION

ground (SLG). For such faults single-phase reclosing would result THE ADVANTAGES of high-speed automatic reclosing of
in an improvement in transient stability over the alternative of
three-phase reclosing. The improvement is offset by the capacitive circuit breakers on transmission lines following nonperma-
coupling from the other two energized conductors because it tends nent faults are well understood and documented in the technical
literature [1].1 A significant gain in transient stability is the
to sustain current in the original fault arc path, thereby lengthening
the time required for arc deionization. An increase in dead time is principal benefit to be derived. The importance of this gain in
therefore required which decreases the net improvement. This transient stability varies from one case to another. In some
paper presents a novel method of nullifying the capacitive coupling cases high-speed automatic reclosing may be necessary to achieve
effect. The method requires the addition of a capacitor, connected the desired margin of stability consistent with the criteria of
across the terminals of each breaker pole, proportional to the par- loading and fault conditions to be met.
ticular line being switched. Most applications of high-speed automatic reclosing have been
three phase. That is, for any kind of fault on the transmission
line, all three poles of both circuit breakers (one at each end)
Paper 68 TP 72-PWR, recommended and approved by the Trans- are opened after two or three cycles. After a further delay of 10
mission and Distribution Committee of the IEEE Power Group for
presentation at the IEEE Winter Power Meeting, New York, N. Y., or 12 cycles to permit the fault arc path to deionize (dead time),
January 28-February 2, 1968. Manuscript submitted September 18, both circuit breakers reclose thus restoring the system to its
1967; made available for printing December 11, 1967.
H. A. Peterson is with the Department of Electrical Engineering, original capability provided the fault is nonpermanent.
University of Wisconsin, Madison, Wis. 53706.
N. V. Dravid was with the Department of Electrical Engineering,
1 Following [1] is an exhaustive bibliography pertaining to arc
University of Wisconsin, Madison, Wis. He is now with the Bonne-
ville Power Administration, Portland, Ore. 97208. deionization time and high-speed reclosing in its various aspects.