random phenomenon where generally the WITH POWER CIRCUIT BREAKE3RS, N. B.

AKE3RS, N. B. Dillow, the magnitude and wave shape of switching

1. B. Johnson, N. R. Schultz, A. B. Were. Ibid., surges determines the insulation require-
restriking occurs at times other than vol. 71, pt. III, 1952, pp. 188-200.
ments of all equipment on high-voltage
that which would result in the maximum 6. SWITCHINO SURGES AND ARRESTRR PREFORM- systems on a more scientific basis. The
ANCR ON HIGH-VOLTAGE STATIONS, L. 0. Barthold,
voltage as a result of one restrike. There I. B. Johnson, A. J. Schultz. Ibid., vol. 75, pt. establishment of switching surge criteria
is evidence, however, to indicate more III, June, 1956, pp. 481-91. as a function of the various system charac-
severe voltage magnitudes occurring on 7. SWITCHINO OVERVOLTAGH HAZARD ELrMINATED teristics is certainly a very difficult task
IN HIGH-VOLTAGE OnL CIRCUIT BREAKHR9, L. F. and the Working Group is to be congratu-
power systems than those included in Hunt, B. W. Boehne, H. A. Peterson. Ibid., vol. lated on undertaking this assignment.
Table IV.3-5 62, 1943, pp. 98-106. The other phase of the committee assign-
A preliminary investigation describing 8. TEE INTERRUPTION OF CHARGING CURRENT ment of particular interest to utilities is
AT HIGH VOLTAGE, W. M. Leeds, R. C. Van Sickle. the system discharge requirement of ar-
the effect of lightning arresters on switch- Ibid., vol. 66, 1947, pp. 373-80.
resters when switching long cables at trans-
ing surges is described in the AIEE Trans- 9. PERFORMANCE TEST OF THE ALLOEMEINE mission voltages. Utilities are installing
actions." Further analysis of this phase ELEKTRICITATS GESELLSCHAPT FRREE-JBT AIR- miles of underground high-voltage cables
BLAST 220-KV, 2,500 MEGAVOLT-AMPBRE RE-
of the investigation by the Working CLOSING CmcuT BREAKER, Alexander
Ibid., vol. 67,
Dovjikov,
1948, pp.
for transmission of power where right of way
Clifford C. Diemond. for overhead lines is unavailable close to
Group is awaiting the receipt of an 295-306. load centers. For these conditions the
industry-wide definition of arrester char- 10. 138-KV LINE-DROPPING FIELD TESTS, D. L. thermal discharge requirements of the
acteristics under switching surge condi- Finneran, R. D. Allen, L. J. Linde, A. B. Kilgour. lightning arrester may well exceed even
Ibid., vol. 73, pt. III-A, Aug. 1954, pp. 809-16.
tions. modern station arrester discharge capa-
11. FIELD TESTS ON A 138-Kv HIGH-SPEED OIL bilities. It is to be hoped that in the near
At the present time the Working Group CIRcUIT 3REAKER AT PHuL.u SPORN POWER
on Switching Surges is devoting its atten- PLANT, Otto Naef, R. D. Hambrick. Ibid., vol. future the committee will have results to
74, pt. III, Aug. 1955, pp. 710-27. present on this phase of the investigation,
tion to a study of the type of switching including the effect of transformer satura-
surge classified items A.1(d), A.3, and tion on arrester thermal duty, on a similar
A.4 in the foregoing. basis to that given for switching long lines
in reference 2 of the paper.
References
1. SOME FUNDAMENTALS ON CAPACITANCE
Discussion I. B. Johnson: The Working Group on
SwITCEING, 1. B. Johnson, A. J. Schultz, N. R. Switching Surges wishes to thank, the
Schultz, R. B. Shores. AIEE Transac*ions, vol. H. W. Smith and H. M. Ellis (B. C. Engi- discussers, H. W. Smith and H. M. Ellis,
74, pt. III, Aug. 1955, pp. 727-36. neering Company Ltd., Vancouver, B. C., for their comments on the first report of
2. SWITCHINO SURGB VOLTAGES IN HIGH-VOLTAGE Canada): The Working Group on Switch- this group. As indicated in the report,
STATIONS, I. B. Johnson, A. J. Schultz. Ibid., ing Surges has undertaken a very impor- the question of arresters and switching
vol. 73, Pt. III-A, Feb. 1954, pp. 69-79. surges has been recognized. Papers on the
tant assignment in that the results of their
3. RESISTORS FOR 138-Kv CABLE SWITCHING, investigation will provide information re- subject are being solicited and some work
B. K. Sadler, T. M. Blakeslee. Ibid., vol. 66, quired to determine the basic insulation de- has been done. In particular, there is a
1947, pp. 39-46.
sign criteria for high-voltage transmission need for more field data to supplement the
4. SHUNT CAPACITORS IN LARGE TRANSMISSION systems, particularly where lightning surges results which so far have been obtained
NETWORKS, E. C. Starr, E. J. Harrington. Ibid., and also are now being obtained from
vol. 72, Pt. III, Dec. 1953, pp. 1129-40. are not an important factor in arriving
5. SWrnCMNG CAPACXTIVE KICLOVOLT-AMPBRES at insulation requirements. Knowledge of studies on systems in miniature.

and an important segment of operating

Distribution System Characteristics
Load cost. The optimum design of the distri-
bution transformer, in terms of such things
as loss ratio, impedance, and insulation
and Their Use in Planning and Design system life, is inherently related to the
characteristics of the load to which the
transformer will be subjected.
R. H. SARIKAS H. B. THACKER Information obtained from utility-
ASOCIATE MEMBER AIEE MEMBER AIEE
load surveys can be used for many pur-
poses other than distribution system
SINCE THE SOLE PURPOSE for the This paper describes the load survey planning and optimum equipment design.
existence of an electric distribution work being done by one utility company, Often, data from the same load survey will
system lies in supplying the requirements and how the resultant data is processed also be used in studies of rate structure,
of the consumer's utilization devices, it is and used in distribution planning and cost analysis, and development of selec-
fundamental that a knowledge of these design. The paper also demonstrates the tive selling programs.
load requirements is necessary for sound importance of a knowledge of load char-
planning of the distribution system and acteristics in the design of equipment Classes of Loads
its various components. The economic which will meet the utility industry's
importance of the knowledge of load char- requirements for service with minimum To facilitate the orderly determination
acteristics has become more and more initial investment and operating cost. An of load characteristics, loads may be
recognized with the growth of load on example of the need for load information divided into the following major classes:
utility systems. to design equipment having character- residential, rural, commercial, and in-
In recent years, much work has been istics most desirable to the utility, is the dustrial. These classes usually can be
done in this field. Many utilities have pole-mounted distribution transformer. further subdivided, based on some pecu-
made and continue to make, in some form, These transformers represent a major item liarity of the particular class of load being
surveys of the load on their systems. of capital investment in the utility system served. Exanples are: residential, large

564 Sarikas, Thacker-Distribution System Load Characteristics AUGUST 1957

use, no gas available; residential, large
use, gas available; etc.

Method of Conducting Load Surveys

One of the most useful survey methods
for determining load characteristics is the
method of group-load surveys. Test CLARK

groups are selected on the basis of homo-

geneity of character and use, predom- ST
inance of one class, and convenient layout
of distribution lines for test metering. If
the test groups are selected as being repre-
sentative of their class, the load survey
data can have a broad application.
EXAMPLE OF GROUP LOAD SURVEY
The plot plan of a typical residential
test group is shown in Fig. 1. The group
is one of 48 similar residential test groups
in a current load survey. These groups
are of varying size in order to provide
coincidence information. One-fourth of
the groups has less than 6 customers, ST.
1/4 has 6 to 12 customers, 1/4 has 13 to 20 A CAODJrM
customers, and the remainder has over )P-
20 customers. An indicating, 15-minute
m

UD
interval, kilowatt demand meter is in- '4.
z
stalled at each residence. Graphic in-
tegrating meters with 15-minute interval
strip charts are installed at the master- Flg. 1. ResidentIal load survey plot plan
meter location on the primary circuit to
measure kilowatts and reactive kilovolt-
amperes for the entire group. These
tests run for a 12-month period. In addi- dential air conditioning. The annual kw- Figs. 4 and 5. The test-group coincidence
tion to daily load curves, an analysis of hr versus demand relationship will not factors, shown in the figures, are the ratio
-these data furnishes the following: provide reliable estimating results where of the maximum diversified demand of the
1. Correlation between the kilowatt-hour there is a mixture of air-conditioning and group during the 4-month peak period, to
(kw-hr) usage of an individual consumer nonair-conditioning customers and where the sum of the maximum noncoincident
and his peak demand. the yearly peak ordinarily occurs during demands, for each customer, during the
2. Coincidence factor relationship. the winter. The inaccuracy of the esti- same peak period. These curves are
3. Loss and power-factor data. mate is due primarily to the inclusion of similar to those published by others.1.2.3
4. Load factor information. the air-conditioning kilowatt-hour usage The coincidence relationship between
in the annual consumption figures. various test groups is obtained by com-
The derivation and significance of the Estimates of winter peak demands for paring the peak demand period graphic
above factors and relationships is dis- customers with large air-conditioning chart readings of the test group master-
cussed in the following paragraphs. usage, based on this relationship, will be meters.4
CORRELATION BETWEEN Kw-HR AND higher than actual demand, since the air- Surveys utilizing a single group size,
conditioning usage contributes greatly to while desirable for some purposes, do not
DEMAND
the annual kw-hr consumption and adds furnish intragroup coincidence-factor in-
This relationship has been established nothing to the winter peak demand. formation. However, the kw-hr versus
for the yearly peak demand versus annual The least-squares lines of best fit for the demand relationships discussed prev-
kw-hr usage, winter peak demand versus summer and winter peak-period relation- iously, can be derived. An example of
winter kw-hr usage, and summer peak de- ships are shown in Figs. 2 and 3. A tab- the results of such a survey, made some
mand versus summer kw-hr usage. These ulation of the pertinent statistics under- years ago, is shown in Fig. 6.
latter relationships are more significant in lying these curves is provided in Tables I
view of the increasing saturation of resi- and II. INTERCLASS COINCIDENCE FACTOR
If the test locations can, in the aggre-
Paper 57-168, recommended by the AIEE Trans- INTRACLASS COINCIDENCE FACTOR gate, be considered statistically repre-
mission and Distribution Committee and approved
by the AIEE Technical Operations Department Use of test groups of varying size pro- sentative of the residential customers as a
for presentation at the AIRE Winter General vides the data needed to evaluate intra- whole, a load curve for the entire resi-
Meeting, New York, N. Y., January 21-25, 1957.
Manuscript submitted October 25, 1956; made class coincidence. Scatter diagrams of dential class of customers can be prepared.
available for printing December 10, 1956. test-group coincidence factors for summer If this same technique is used for other
R. H. SARIKAS is with the Illinois Power Company,
Decatur, Ill., and H. B. THACKER iS With Westing-
and winter peak periods along with the classes of customers, similar load curves
house Electric Corporation, Pittsburgh, Pa. hyperbolic regression lines are shown in can be prepared. When these load

AUGUST 1957 Sarikas, Thacker-Distribution System Load Characteristics 565

 transformer, secondary, and service losses.
Similarly the rural test group kw-hr
master-meter readings (the master meter
is usually installed on a primary lateral
Lii
serving 30 to 60 customers) less kw-hr
sales, represents losses in rural primary
0 laterals, transformers, and services.
01-J When these losses are increased in propor-
y tion to total system mileage and trans-
-j z
formers they furnish data on losses within
0
z various segments of the system. These
data, when compiled for the entire system,
0 may be used to obtain loss multipliers for
z LUJ
0
-/ SOL ID CURVE: Y:O.C004c66X3-O.7615X 2 ' 364x'+ 1.2 2467
evaluating the effect of further system
.tDASHED GURVE: Y 2.4593 x3+6.4732ox2 -1.53917x+0O.S6
596 losses caused by losses in the distribution
2 system.,,6

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 POWER-FACTOR DATA
The power factor of the various test
4- SUMMER MONTHS CONSUMPTION - MWHR groups at time of peak load during the
winter peak period is shown in Fig. 8.
20 Fig. 2 (above). Comparable information for summer
8 Residence test peak-period is provided in Fig. 9. Use
group, kw-hr con- of these data permits an evaluation of the
U,
16 sumption versus de- effect of such loads as air conditioning and
4 mand, summer-peak the improvement possible by such means
period as series and shunt capacitors. Data rel-
-Jz- 2
ative to the power factor of loads on the
I0 secondary system also permit an intelli-
~- 8 gent decision whether or not to use sec-
6 ondary capacitors after determining a
breakeven power factor for secondary
4 versus primary capacitors.7
Fig. 3 (left). Resi-
2 dence test group, LOAD FACTOR
0 kw-hr consumption
O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 versus demand, Load factor can be obtained from an
4-WINTER MONTHS CONSUMPTION - MWHR winter-peak period analysis of the daily load curves or the
load factor may also be obtained from a
comparison of demand and kw-hr readings
curves are combined, the system load Loss DATA for a comparable period. The load-
curve is obtained. Such a series of load Loss data are obtained as a by-product factor versus coincidence-factor relation-
curves along with the actual system send- of the load study. For example, in the ship is also available. This relationship,
out from dispatching records is shown in residential test group shown in Fig. 1, the which has limited application in distri-
Fig. 7. Comparison of the series pro- difference between master-meter kw-hr bution planning, is useful in developing
vides interclass coincidence relationships. readings and total sales billed represents "selective selling" programs.8

1.0 1.0
0.9 0.9
0.8 172 _ 1 1 08
0
0 0.7 A I
_ _ _I l O 07
U- 0.6 < 06
Li 0.5 a_"I%,-, ti I a 05

_ _ e- Se
0 z
z 0.4 oU 04
5 0.3 Z 03
z
0
0)
0.2 I_ - - &
0
0.2
0.1 01
0 0
0 2 4 6 8 10 12 14 16 1I 20 22 24 26 28 31vJ 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
NUMBER OF CUSTOMERS NUMBER OF CUSTOMERS

Fig. 4. Residence test group, coincidence-factor relationship, summer- Fig. 5. Residence test group, coincidence-factor relationship, winter-
peak period peak period

566 Sarikas, Thacker-Distribution System Load Characteristics AUGUST 1957

Use in Distribution Planning and Table 1. Urban Residential Electric Service.* Relationship Between Four Summer-Months
Design Kw-H r Consumption and Maximum Demand for Individual Customers During Test Period
15-Minute Demands in Kw
One of the most important uses of load
No. of Range of Tests for
characteristic data is in distribution plan- Average
Range of Four
Summer-Months Measure- Range of Middle Half of Cases
ning and design. Such data can be Kw-Hr Usedt ments Average Median All Tests Ranked by Demands
applied to local distribution design for
each customer class, as well as to over-all 346 .. Under 500 kw-hr ........ 139 ...... 1.21 ...... 0.8 . 0.3- 8.2 .......
1.5 . 0.5- 7.7 .......
0.5- 1.2
1.0- 4.0
718 .. 500- 999 kw-hr.....208 ...... 2.4 ......
system planning involving combined 1,224 .. 1,000-1,499 kw-hr..... 124 ...... 4.04 ...... 4.3 ... 1.0-12.0 ....... 2.2- 5.3
customer classes. 1,721 .. 1,500-1,999 kw-hr ..... 57 ...... 5.2 ...... 5.0 ... 2.0-10.5 . 4.5- 6.0
2,215 .. 2,000-2,499 kw-hr ..... 34 ...... 6.2 ...... 6.0 ... 3.5-10.0 . 5.0- 7.0
Loads of presently installed trans- 2,685 .. 2,500-2,999 kw-hr ..... 16 ...... 6.5 ...... 6.5 ... 3.4- 9.0 5.5- 8.1
3,289 .. 3,000-3,499 kw-hr..... 17 ...... 6.8 ...... 7.0 ... 3.0-11.0 . 6.0- 8.5
formers can be estimated by determining 9 .0 ... 7.5-11 .0... 8.5- 9.5
3,830 .. 3,500-3,999 kw-hr ..... 14 ...... 9.0 ......
the kw-hr usage of each customer served 4,182 .. 4,000-4,499 kw-hr.... 8 ...... 9.2 ...... 9.0 ... 8.0-11.0 . 8.7- 9.6
by the transformer and obtaining the 4,765 .. 4,500-5,499 kw-hr..... 7 ...... 11.2 ...... 10.0 . 6.0-15 .0... 8.0-10.2
5,847 .. 5,500-6,499 kw-hr ... 7 ...... 9.9 ...... 10 .5. 7.0-12.5 .. 9.8-10.5
individual customer demands from Fig. 2. 9,458 .. Above 6,500 kw-hr.... . 6 ...... 11.8 ...... 12.0 . 9.0-14.0 . 10.0-13.0
Application of the appropriate cinocidence * Data for all tables compiled at Illinois Power Company.
factor from Fig. 4 provides an estimate of
t June through Sept.
the peak load supplied by the transformer.
A nomograph to simplify this operation, Table II. Urban Residential Electric Service. Relationship Between Four Winter-Months
using an average of the four summer Kw-Hr Consumption and Maximum Demand for Individual Customers During Test Period
months (June-September), kw-hr usage
is shown in Fig. 10. A similar nomograph 15-Minute Demands in Kw
for estimating the winter peak-period No. of Range of Tests for
load, using an average of the four winter Average Range of Four Measure- Range of Middle Half of Cases
Kw-Hr Winter-Months Use* ments Average Median All Tests Ranked by Demands
months (November-February), kw-hr
usage is shown in Fig. 11. Similar
360 .. Under 500 kw-hr ........ 63 ...... 0.72 0.5 0.0- 3.5 ...... 0.4- 1.0
curves have been developed by others.9 760 .. 500- 999 kw-hr .. 197 ...... 1.8 1.2 0.0- 6.5 ...... 1.0- 2.0

The accuracy of the two demand estima- 1,231 .. 1,000-1,499 kw-hr ..... 117 ...... 3.4 4.0 0.1- 7.5 ...... 1.5- 5.0

1,735 .. 1,0500-1,999 kw-hr ..... 78 ...... 4.98 5.0 1.0-12.0 ...... 4.0- 6.0

tion nomographs is shown in Fig. 12. 2,240 .. 2,000-2,499 kw-hr ..... 49 ...... 5.99 6.0 1.5-10.0 ...... 5.0- 7.0

A number of companies are using similar 2,709 .. 2,500-2,999 kw-hr ..... 31 ...... 6.3 6.0 3.2-11.0 ...... 5.3- 7.0

3,242 .. 3,000-3,499 kw-hr ..... 17 7.98 ... 7.5 4.5-12.2 ...... 6.4- 9.4

schemes to monitor the loading of distri- 3,729 .. 3,500-3,999 kw-hr ..... 10 ..... 8.3 8.4 6.0-10.8 ...... 8.0- 9.0

bution transformers.'011 At least one 4,243 .. 4,000-4,499 kw-hr ... 11 ..... 9.0 8.7 5.5-14.5 ...... 8.4- 9.7

4,977 .. 4,500-4,999 kw-hr ..... 14 ..... 9.1 . 9.5 6.0-10.9 ...... 8.0-10.0

company has mechanized the operation, 9,414 .. 5,000-above kw-hr..... 6 .14. 14.6 .13.0 8.5-25.0 ...... 9.5-13.0

using punched card techniques employing * Nov. through Feb.

the computing machines normally used
for billing.'2
Safe overloads for distribution trans- drops at various points along the feeder. be adjusted by the ratio of measured to
formers can be determined from an anal- The estimated load on each feeder can be estimated demand. If voltmeter readings
ysis of the daily load curves of the various compared to the load recorded on demand at the regulated bus of each substation
test groups. The frequency of occurrence meters installed on the respective feeders, bus are available for the peak period, the
of the peak-load condition is of great and the calculated voltage drops may then sending-end voltage at the feeder may be
significance.'3 The relative magnitude
of the daily peak loads throughout the 3.6 NOG S SEVICI 3.6
year, for a typical distribution trans- * NO GAS SERVICE AVAILABLE
former, is provided in Fig. 13. These data * GAS SERVICE ALSO AVAILABLE
3.2 A APARTMENT BUILDINGS 3.2
have been regrouped in the form of a *@A PRIMARY MEASUREMENT
cumulative frequency polygon in Fig. 14. OOL SECONDARY MEASUREMENT
2.8 2.8
This figure permits an estimate of the w

probability of obtaining the true peak-

load of a distribution transformer using 2.4 2.4 O
cn ()
field measurements of a given duration. 0

Transformer-secondary system voltage- a:

w
2.0 2.0 r

drop calculations based on the resulting a- a-

load estimates, will furnish data useful Z

1.6 1.6 Z

from an operating as well as an engineer- 0

ing standpoint. If the loading and regula- z

r5C
1.2
z

o 1.2
tion of an existing transformer-secondary
combination exceeds design limits, data 30.
0.8
are available to plan an adequate rebuild. 0.8

PRIMARY CIRCUIT LOADING AND DESIGN 0.4

0.4

Fig. 6. Group co-

Since the load at each distribution incident demand
transformer along with the diversity be- versus average an-
n)
u
0

0 1000 2000 3000 4000 5000 6000 7000 8000

tween transformer loads can be estimated, nual use per cus-
it is thus possible to calculate the voltage tomer KWH CONSUMED IN YEAR

AUGUST 1 957 Sariktas, Thacker-Distribution System Load Characteristics 567

 ~~~~~~4 0

0~~~~~~~~~~~~~~~~~~~~~I1

LU~~~~~~~~~~~~~~~~~~~~~~~~oC
0~~~~~~
U
(D1 N0 o 0 0 0 0 0 0
_r 0 SIS0NINOSN±~ B

0~~~~~~~~~~~

0 ~~~~~~~~~~~~~~000
_
0
0~~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

CN0 -J~~~~~~~~~~~~~~~~~~~
m 0

0 ~~ ~ ~ ~ ~ ~
0
o 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 U,~~~~~~~~~~~~~~~~~~
1- ~~~~~~~~~~~~~~~0
00

z W z~~~~~~~~~~~~~~~~~
0C
Li) 0l
0~~~~~~~~~0
(1) ()g

oLU
0
o~~~~~~~~~
U

< 0 0 0 0 0 00 0 0 0 00
0
Ld a: 0 CD 0 L 0 0
pD001 )jOd8 10 4U83JOd PO 18 0lOJdL

ZD 0
0 0~~~~~~~~~~~~~~~
0 r-, F- c

LL w~~~~~B WU) LU
cli

wn N.1<w0G

V L
TIIlIInS (.......ID 1]:Iz
~
~~~~~~C
cIn

LU

m~~~~~~~~~~~~~~~~~~ 0
0 (0 () 0 0 0 0 0 L

°
0 X 0 0 O 0 0
0 N 0 0 0 0 _ 0

PDOo MD0d 10 0U9D,0d pDOo1MaDd 10 IUO3iJd Z

Q VSSI9S8V NI NMOHS 0130i- 3MOd
SNIAVH SdnlOdo iS31 JO iN33d1d

568 Sarikas, Thacker-Distribution System Load Characteristics AUGUST 1957

40
15000
10000 This latter information is obtained from INDUSTRIAL LoAD SURvEYS
35 billing data. The coincident demand for Since most customers in this class have
30 6001 5000 each area is then estimated by using the graphic-type meters for billing purposes it
4000
I
previously developed kw-hr versus de- is possible to obtain their characteristics
25- mand relationship, and intraclass and
3000 r2 both individually and collectively without
20- 30- interclass coincidence factors. Using this special surveys. To assist in future
15
50 2000-
0 method it is no longer necessary to assume planning, the characteristic-load curve of
1500
LL
2 that the entire load served by a substation various types of industrial customers
20- S
2
2 or a feeder is growing at the same rate. A was prepared with a composite-load curve
I0- cr 0
2 study using kw-hr consumption trends, of all industrial customers; see Fig. 7.
9- 2
0 1000
0
pinpoints the area where load is growing
8-
7- 10- a
4
900 0
2 most rapidly. Further, the load growth SYSTEM LOAD CURVES
0L
6- 0 B-
cr 800 I
picture of an area is not lost when load An analysis of system load character-
5 - Z-
2
6- 2 is transferred between substations or istics coupled with an appropriate cost
5.- 0 700 circuits.
2
D
4.
0
2 analysis based on a study of the contri-
4- z
3.:
4Ia: 600- z COMMERCIAL-GROUP LOAD SURVEYS butions of each class of load at various
3-
2
levels of the system; i.e., transmission, sub-
2- 500- A study of commercial groups provides transmission and primary, provides the
0Itr the necessary data to estimate trans- answer to many puzzling problems. For
4
2- 400- former loads and the necessary interclass example; when shall a customer be
4
I.0-
0.9- and intraclass coincidence factors to esti- served at 34.5 kv and when should he be
0.8-
0.7- 300- mate circuit loading. The duty cycles for served at 4.16 kv? What is the break-
0.6- 250- transformers serving this class of load even load, assuming equal accessibility
05-
are also obtained and may be used as a of both types of circuit? A complete dis-
basis for design and allowable loading. A cussion of the subject is more properly
FiS. 10. Demand esmation nomograph for group of typical commercial-load curves the material of another paper, but one of
urban reidential sevce, four summer months
kw-hr consumption are shown in Fig. 15. A tabulation of the the most important factors in the solution
pertinent statistics is provided in Table to the problem is an adequate knowledge
III. The pattern of these load curves is of load characteristics.
used as an "anchor-point," and the similar to those which have been de-
voltage at any customer served by the veloped by others."4 Load Studies by Other Utilities
feeder may be calculated. This would RURAL-GROUP LOAD SURVEYS
eliminate the need for routine secondary- Recognizing the importance of obtain-
voltage testing in order to keep informed What has been said for residential and ing load characteristic information, other
of customer voltage levels. Knowledge commercial load surveys also applies to utilities have been conducting load surveys
of the primary voltage existing at each rural group load surveys. A rel.ationship and utilizing the data obtained in many
distribution transformer would permit a between individual customer annual kw-
different design transformer-secondary hr usage and demand, has been developed; 40- 15000-

voltage drop for each transformer sta- see Fig. 16. The underlying data are shown 10000-
in Table IV. These data were obtained by 30-
tion, depending upon its location on the 8000-
feeder. The transformer-secondary com- installing demand meters at each residence 6000-:
ldr
2

bination nearest to the substation may and at the totalizing master-meter loca- 20-
100- 5000- I
be permitted more voltage drop than a tion. A large number of test groups were 80- ler
4000- 2
transformer-secondary combination near installed without individual demand 15. 60- 0

the end of the feeder. meters to obtain further data on group 50-
3000 1U'I-0
40-
Since iMdividual transformer loads are demands and losses. Since generally 0.
most rural customers are served by indi- 30-.
known, the loads to be connected to pro- 10- I,-
0:

posed feeders can be estimated with a vidual transformers, individual customer 20-
2000- 0

high order of accuracy. This eliminates load information is very significant. A is-
0
0
2
41
lz
1°0
difficulties which might arise if more load test has been initiated to determine the 0
2
0
1500-
1
is transferred to a new feeder than was individual characteristics of a carefully 6- 2
0
10-
intended. selected sample of our rural customers. 5. 8- 1000-
2 0
This will provide data on the duty cycle 4. 2 6-.
0
2
LoAD FORECASTING of rural distribution transformers serving 5- 41 800- 2
a:
A further use of these data is in the prep- single- or multiple-customer secondaries 3-
4-
600-
2

aration of load-density maps which show and further improve our knowledge of 3-
13
It

the coincident load in each unit of area. kw-hr versus demand, and coincidence 2-
500-
This unit area is usually a square mile for relationships. The meters to be used for 2- 400-
rural areas and 1/4 of a square mile for this test are of a special design and have a
urban areas. The load-density maps meter trough that can be plugged into the I '~~ ~~-
300d
are prepared by counting the number of customers' outdoor meter socket. This
customers in each class within an area, trough will in turn accommodate the test
I 250]
using distribution maps, and assigning meter and the customers' kw-hr billing Fig. 11. Demand estimation nomograph for
these customers the average kw-hr usages meter. Such an arrangement greatly urban residential service, four wintef months
of customers in their respective area. reduces the cost of such a survey. kw-hr consumption

AxJGUSTr 1957 Sarikas, Thacker-Distribution System Load Characteristics 5,69

 0
1. Curves showing the daily average 100 3 4 5 7
10 demand per customer for an average week-
day, and for the day when the test group
90

20_
reached its maximum coincident demand. z
w 80

z
30
40 -
2. The daily average load characteristic
0 per customer shown as a percentage of 0ir

(A
z
50_ maximum demand per customer, both Sz

60- for an average weekday and for the group 3L 0

f--
70 peak-load day. J

80_ 3. Coincident maximum demands per 40

90 customer and group load factor.
00 - 4. Noncoincident demand per customer as
100 90 80 a function of individual customer's monthly
PERCENT ACCURACY
kw-hr usage.
0 10 20 30 40 50 60 70 80 90 100

5. Coincident demand per customer as a PERCENT OF DAYS HAVING LOAD GREATER THAN
Fig. 12. Accuracy of transformer load function of average monthly kw-hr usage LOAD SHOWN IN ORDINATE
estimation, kw-hr method per customer for the group.
Fig. 14. Daily load magnitude frequency oF
6. Diversity factors related to number of occurrence, group N-1, eight customers
ways. A representative list of problems customers and presented as absolute
to which load data are being applied fol- diversity versus number of customers and
lows: also in per cent of full diversity versus
number of customers. load survey data, a demand meter has
1. Load forecasting by determination of been developed that records 15-minute de-
class and system peak loads. Another company installs primary and
secondary metering to obtain: mands on a magnetic tape. The tapes
2. Determination of distribution trans- containing the 15-minute demand read-
former, primary feeder, and substation load- 1. Fifteen minute maximum kva (kilo-
ing for use in distribution system planning. ings can be fed into a machine which
volt-amperes) and cumulative kw-hr. punches the data on cards that in turn
3. Economic determination of appliance- 2. Kilowatt-hour and kilovar-hour usage
promotion policies based on effect on are fed into data-processing machines to
per hour and per day. Readings are taken
system load caused by load characteristics on primary feeders, at the distribution
obtain various desired information. This
of various appliances. transformers, and at the individual home. new demand meter will considerably
4. Development of load curves in various These data is used in obtaining the following: speed up operations in interpreting load-
portions of the system as a basis of design a. Daily load curves. survey data and may well be justified
for specific facilities. b. Determining the effect of tempera- both for load testing and for demand
5. Periodic survey of distribution trans- ture on hourly load variation of billing purposes.
former loading and development of change- appliance groups.
out schedules. c. Determining kva demand versus IMPORTANCE OF LOAD CHARACTERISTICS
6. Determination of "cost to serve" kw-hr usage of various appliance IN DISTRIBUTION TRANSFORMER
individual-load groups as a basis for evalu- groups. It is planned to use a program
for developing this information in DESIGN
ating current rate schedules.
which data-processing machines will Over the years, many changes have
Many different methods and procedures eliminate a major part of the manual taken place in the make-up of the distri-
are used in making load surveys and apply-
calculations.
bution transformer. The basic com-
ing the information.'0 One company has Another electrical utility uses a continu- ponents of copper, iron, insulation, and
selected customer groups of about 40 ing field check on certain selected trans- oil have been improved by new process
customers having some common electric- formers to keep statistical data up to date techniques. New materials, tanks, finish
load characteristic such as an average for use on their entire system. Recording appearance, and protection have all been
peak-month kilowatt-hour usage between a eters and voltmeters are in contin- points of design improvement. How-
100 and 150 kw-hr. Individual graphic uous service, measuring the loads on this ever, since the introduction of grain-
recording meters are installed on each sample group of transformers, and sta- oriented steel, there has been very little
customer's service to record 1/2-hour tistical data are thus kept up to date by change in the basic operating character-
kilowatt demands for a 16-day period. periodically correcting according to data istics of the distribution transformer.
From these demand data and from kw-hr obtained from the sample group. This is rather surprising in view of the
usage obtained either by test data or from As a means of simplifying and reducing significant changes in system loads and
meter books, the following is developed: the cost of tabulating and interpreting operating practices which have occurred

100
x
4 90

X 80
i : f ilA!gflunl.1

1~ ~ ~ ~ ~ ~ ~ ~11~ ~ ~ ~ ~ ~ I
4
P
Z 70
z

c 60
0
50

3 40
a. 30
Z 20
o 10

0
JAN. FEB. MAR. APR. MAY JUNE JULY AUG SEPT OCT NOV DEC. Fig. 13. Daily maximum loads, group
MONTH AND DAY N-1, eight customer
570 Sarikas, Thacker-Distribution System Load Characteristics AUGUST 1 957
 YEAR 1954 - 55 Fig. 15. Com-
mercial test
CHAMPAIGN CHAMPAIGN groups, daily
SUMMER PEAK-WEEK WINTER PEAK - DAY load curves
10 o 0T
° 80 60 1 T1
a-
- so
'20 i F- N *20

12 2 6 T( 2 4 12 2 4 6 18 10 12 2 4 6 6 10 12 12 2 4 6 8 10 12 2- 4 6 8 10 12
AM PM AM PM
a-

DANVILLE DANVILLE

o° mf
SUMMER PEAK- DAY SUMMER PEAK-WEEK
100
0.

0 60
° 60
° 40 20

20
u 100 40

2 4 6 P M0 12 2 4 6 6
A2 12 12 2 4 6 6 10 12 2 4 6 8 1o 12
60
AM PM AM PM
f-J

--I
DANVILLE DANVILLE
I
KEWANEE
WINTER PEAK- DAY WINTER PEAK -WEEK SUMMER PEAK - DAY
100 0. I0 1 I I/
00 1I
0
-J
so , soC 0 0
a
eol-ll r rvrI 1 T1
I 1 n
so
1 11
0
4o80
40
-r-l °l I- I III---l- -
I
°ITTIA- 1
I t
S 0
+i40- a-

0.
,,20 1 201 1 I I--
0
12 2 4 6 8 10 12 2 4 6 6 10 12 12 2 4 6 6 10 12 2 4 6 8 10 12 i2 2 4 6 8 i0 12 2 4 6 8 IC
AM PM AM PM AM PM
I

KEWANEE
SUMMER PEAK -WEEK 11 T of 6

0.

a 4-
20
-11111AM
m
1P

0L2 2 4 6 8 10 12 2 4 6 8 10 12 Le J.,I
i i
AMl pM

Table 111. Urban Commercial Electric Service. Identity and Characteristics of Test Groups of Customers
Maximum 15-Minute
Demand for Group Annual-Load Factor Kw-Hr Per Kw of Demand Intraclass Coincidence Factor
No. of Summer Winter Kw-Hr Used Kw-Hr Summer Winter Summer Winter
Group Customers in Peak Peak in Year by Per Kw of Peak Peak Peak Peak
Location Test Group Period Period Test Group Demand Per cent Period Period Annual Period Period

Kewanee ...... 72 . 297 . 233 . 839,362 2,859 .. 32.6 1,175 ....... 1,006 ... 0.76 .... 0.81 . 0.68
Champaign .. 74 .. 6 i 23 .. 2,195,350 2,535....... 28.9
1,0. 996 ... 0.82 .. 0.87.. 0.76
Danville ..... 154 . . 724 574 2,110,427 1,122
2,917 .... 33.3...
1,089.. 0.67 . 0.73 0.66
Granite City. 74 . .477 . 297 1,468,179 3,081 .... 35.2
1,008 . 1,016 . 0.72 O.. 0 .. . 0.66
Mt. Vernon.. 60 402 306 1,212,050 3,020 .... 34.5....5
1,193 .. .... 1,180 0.63.7... , 72
Average ..... 87 . . .2,882..... 32.9..9 1,058.. 1,102 ... 0.72 . .78 . 0.69
Note: Summer-peak period includes the months of June-Sept. Winter-peak period includes the months of Nov.-Feb.

AUGUST 1957 Sarikas, Thacker-Distributxon System Load Characteristics 571

 Table IV. Rural Farm, Residence Electric Service. Relationship Between Annual Consumption
and Maximum Demand for Individual Customers During Test Peiod

15-Minute Demands in Kw
6 1 No. of Range of Tests for
Average Measure- Range of Middle Half of Cases
54 Kw-Hr Range of Annual Use ments Average Median All Tests Ranked by Demands
-z
4-4
> < 793 . Under 1,000 kw-hr . .....4....... 1.4 . 1.6 . 0.2-2.1 ..... 1.5-1 7
a 1244 . 1,000-1,499 kw-hr ...... 12. 1.9 . 1.8 . 0.3-3.9 ..... 1.3-2.5
1908 ...... 1,500-2,499 kw-hr ...... 5 ...... 2.8 .......2.4....... 1.4-3.4 ........ 2.0-2.9
3296. 2,500-4,499kw-hr ...... 14 4.2 . 4.3 . 1.5-7.2 ..... 2 8-5 2
5297 . 4,500-5,999 kw-hr . .....9 5.8 . 5.7 . 4.2-8.1 ..... 4.5-6 6
6720 . 6,000-7,999 kw-hr . .....8....... 6.1 . 6.4 . 3.4-8.1 ..... 5.4-7 1
0- - 9649 . 8,000 kw-hr and over.... 6 . 7.2 . 7.1 . 5.4-9.5..... 6.4-7.7
0 2 3 4 5 6 7 8 9 10
CONSUMPTION IN MWHR
Fig. 16. Rural test group, annual kw-hr made along these lines at probable in- 3. SYNTHETIC OR EQUIVALENT LOAD CURVBS.
R. F. Hamilton. AIEE Transactions, vol. 61,
consumption versus demand creases in manufacturing costs. Justifica- 1942, pp. 369-81.
tion for such changes in the light of in- 4. TEE SUMMATION OP LOAD CURVES, R. P.
creased costs must be based on studies of Hamilton. Ibid., vol. 63, Oct. 1944, pp. 729-35.
in recent years. Load growth on the 5. Table III of discussion by Robert H. Sarikas
distribution system has resulted in greater economical transformer loading, recogniz- of WHAT Do LOSSES COST IN HYDRO, TEBRMAL,
load densities, and the advent of new ing the added cost of increased losses and AND COMBINED SYSTEMS, V. W. Ruskin. Ibid.,
vol. 75, pt. III, June 1956, p. 336.
loads such as air conditioning and the tele- greater voltage regulation resulting from
6. ELECTRICAL DISTRIBUTION ENGINEERING
vision set have resulted in a demand for heavier transformer loading. This could (book), H. P. Seelye. McGraw-Hill Book Com-
better voltage regulation. Such load be done, were sufficient information pany, Inc., New York, N. Y., first edition, 1930,
chap. 33.
growth has also been accompanied gen- available, by developing the relationship 7. ECONOMIC MBRITS OP SECONDARY CAPACITORS,
between the annual costs of the various R. A. Zimmerman. AIEE Transactions, vol. 72,
erally by lower power factors and by factors involved in the application of dis- pt. III, Aug. 1953, pp. 694-97.
significant changes in the daily and tribution transformers and the average 8. LOAD RESEARCH, AS A GUIDE TO SELECTIVB
seasonal load curves. Perhaps the main SELLING, G. M. Turner. Electric Light and Power,
reason for there being little basic change load over the expected life of a trans- Chicago, Ill., May 1952.
in the operation characteristics of the former. Development of such relation- 9. DISTRIBUTION TRANSFORMER LOAD SURVEY
ships, based on reliable information as to AND TRANSFORMER-SECONDARY DESIGN, H. E.
distribution transformer, has been the Carpenter. Missouri Valley Electric Associatiox
difficulty of obtaining load-characteristic load factor, cost of losses, cost of voltage Engineering Conference, Kansas City, Mo., Mar.
1955. (Available from the author.)
information which is sufficiently represen- regulation, etc., provides a sound basis 10. MBTER BooKc GIVES TRANSFORmER LOAD,
tative for the entire electrical utility in- for determining economical transformer L. J. Weed. Electrical World, New York, N. Y.,
dustry to serve as the basis for evaluating loading, and evaluating the desirability Sept. 6, 1954.

the over-all economics of transformer de- of further changes in transformer design. 11. LOAD CRECK TEAT SMALL TRANsFORMER,
M. J. Urner. The Line, Sept. 1955.
sign changes. For this reason, load- 12. COMPUTERS GAUGE TRANSFORMER LOAD,
survey work such as described in this Conclusion H. E. Stratton, R. H. Joerden, B. M. Gal}aher.
Electrical World, New York, N. Y., Mar. 21. 1955.
paper is highly desirable. The work being done in the electrical 13. ECONOMIC LOADING OP DISTRIBUTION TRANS-
A significant change in the load char- utility industry to obtain a more com- PORMERS, C. F. Mitchell. Proceedings, American
Power Conference, Chicago, 11., Mar. 1956.
acteristics of many areas of the United plete and realistic knowledge of load 14. LOAD CHARACTBRISTICS OP "SMALL LIGHT
States, having an important bearing on characteristics is producing valuable re- AND POWER" CUSToMERS-THEIR E CONOMIC SIG-
distribution transformer design and oper- sults. A knowledge of the character- NIPICANCB, Constantine Bary. Electric Light and
Power, Chicago, III., Oct. 1955.
ating practice, has been the trend towards istics of individual and group loads of 15. MANUAL OF PROCBDURE FOR LOAD SURVEYS,
a summer-peak demand which approaches different classes and how they combine Load Research Committee. Association of Edison
or exceeds the winter peak. These Illuminating Companies, New York, N. Y.,
in various parts of the system is essential Feb. 1951.
summer peaks, brought about mainly by to efficient, economical system design.
a growing use of residential air condition- Many other benefits can accrue from a
ing, occur at a time when ambient tem- greater familiarity with load character-
peratures and their effect on distribution istics. Not the least of these are possible
transformer loading are at an extreme.
Such loads also appear to have resulted in
changes in apparatus design to provide
greater operating economy to the utility.
Discussion
a higher-load factor with attendant re- Because of the importance of such informa-
duction in the ability of the transformer A. H. Kidder (Philadelphia Electric Com-
tion to the industry, it is the authors' pany, Philadelphia, Pa.): This discussion
to carry short-duration loads. Reduction opinion that load surveys and the free was prompted by my inference for Mr.
in allowable transformer loading is exchange of resulting data should be Sarikas' paper that he proposed to use a
certainly not to be desired from the stand- strongly encouraged by both utilities and curve similar to Fig. 6 with one similar
point of obtaining the most use from the to Fig. 4 as a moderately reliable basis for
manufacturers. estimating the maximum load on any given
investment in distribution transformer distribution transformer installation. My
capacity. Recognizing this, an improved References experience has taught me that the vagaries
insulation system and more efficient cool- of customers' meter-reading schedules, the
ing has been incorporated in recent trans- 1. COINCIDENCE-FACTOR RBLATIONsHIPS OF ELHC- dissimilarity between the loads of appar-
former designs, which has resulted in sub- TtIC-SERVICB-LOAD CHARACTR.1STICS, Constantine ently identical residences, etc., make kw-hr
Bary. AIEE Transactions, vol. 64, Sept. 1945, sales an unreliable substitute for distribu-
stantial improvement in life expectancy pp. 623-29.
tion-transformer load testing; although the
and has allowed a greater margin of 6ver- 2. EVALUATION OP INCOMPLBTELY DIVEERS'IED information illustrated by the authors
LOADS, W. L. Tadlock. Electrical Engineering,
load capability. Other changes could be vol. 62, Nov. 1943, pp. 485-92. does provide a practical basis for estimating

572 Sarikas, Thacker-Distribution System Load Characteristics AUGUST 1957