Sag-Tension Calculation Program ror three limits will be controlling.

This
sag-tension calculating program has been
written to determine which of the limits
Digital Computer controls, and for convenience in curve
plotting, a span close to the location where
a change in limits occurs has been deter-
mined. A flow chart for accomplishing
B. M. PICKENS this is shown in Fig. 1. The range of
ASSOCIATE MEMBER AIEE
spans required is obtained by using a
shortest span (SPNX), longest span
ALCULATIONS of the sags and ten- 1. Design or final sags and tensions. (SPNM), and interval between spans
sions in conductors used in overhead
(SPI) routine. Tensions and sags are
2. Initial or stringing sags and tensions. first calculated for the shortest span
systems are a time-consuming and seem- Final sags and tensions are used to specified, then the span interval is added
ingly never-ending task. Over the past and the tension and sags are calculated
30 or more years, thousands of calcula- determine clearances to ground, other
power or communications circuits, over for this new span. This continues until
tions have been made, notably those the maximum span specified is reached
by the Copper Wire Engineering Associa- roads and railroads or other locations
which are overbuilt. when the program stops.
tion, in addition to those made by the con- Initial or stringing sags and tensions
ductor manufacturers, and there are still The sag-tension routine (R) calculates
many conductors for which sag-tension are used primarily as a guide to the actual both design and initial sags and tensions
information has not been prepared, at installation of the conductor and are used for the loading conditions and tempera-
least for some tension limits and loading in determining possible uplift at supports. tures specified for a given span, and
conditions. There are five generally This program for sag-tension calcula- punches this information in output cards.
recognized loading districts (heavy, me- tions covers both final and initial values At the same time, the routine enters into
dium, light, California heavy, and Cali- for a series of spans of the range required. a memory location the one of the three
fornia light) and special loadings to meet Usually three limits to the possible ten- tension limitations which was found to be
local conditions, hundreds of conductor sions are imposed. They are controlling. This location is identified
designs, each with several commonly 1. The maximum tension under the as- TS in the program; see Table II.
used tension limits, with the result that sumed storm loading conditions shall not As each span interval is added to the
the number of possibilities for which sag exceed a specified maximum value in previously used span length to give a new
pounds. This may be given as a per cent span, and the sags and tensions are cal-
and tension values may be needed are al- of the rated strength or as a specific value
most endless. When done by desk cal- in pounds. culated and punched for the new span,
culator and tables, or by some graphic 2. The final tension at some selected the resulting controlling tension is com-
method, each problem is a time-consum- temperature but without ice or wind load- pared with the previous controlling ten-
ing operation. ing shall not exceed a given per cent of the sion. If they are the same the machine
The use of a digital computer for this rated strength. proceeds to the next span. However, if
work seemed to offer promise of remov- 3. The initial tension at some selected the tension limits are not the same a pro-
ing the drudgery involved. Some years temperature shall not exceed a given gram is started to determine the span
per cent of the rated strength. length where the change inlimitsoccurred.
ago an attempt was made to program an
IBM (International Business Machines These last two limits are imposed to This is done by adding a fraction of the
Corporation) 604 computer for this work. protect the conductor from possible vibra- span interval to the span length used be-
While a program was worked out which tion damage. The commonly used limits fore the change occurred and checking by
could be used for copper-type conductors, are shown in Table I. means of a special routine to see if there
it lacked many desirable features. would be a change in limits. This con-
With the introduction and more com- Main Program tinues until the span length where the
mon use of the stored program machines, change would occur is determined. The
such as the IBM 650, it was possible to For any given span there is usually no determined span length is then used in
develop a program for calculating sags and way of knowing in advance which of the the routine (R) to obtain a set of sag-
tensions in which most of the work is
done by the computer.
Table 1. Limiting Working Tensions
General Initial Final
Sag-tension data fall into two main Before Maximum Load,
No lce or Wind
After Maximum Load,
No Ice or Wind
divisions:
Per Cent Per Cent
of Rated Tempera- of Rated Tempera-
Paper 58-109, recommended by the AIEE Com-
Materla Loading District Strength ture, F* Strength ture, F
puting Devices Committee and the Transmission
and Distribution Committee of the Power Appara- Copper type .all . ............... 35 . ... 60 25 .. 60
tus and Systems Division, and approved by the heavy .. 331/... . 0. 25 . 0
AIEE Technical Operations Department for pres- Aluminum ...... medium .. 33/ .. 15. .25 15
entation at the AIEE Power Industry Computer light .. 331/ 30. 25 . 30
Application Conference, Toronto, Ont., Canada. ACSRt California heavy....... 331/* .. 0.25.. 0
September 11-57, 1958. Manuscript submitted I California light.,,
..... ....

331/s. 2 . 2.25
May 28, 1958; made available for printing Augut
4, 1958. * Degrees Fahrenheit. t Aluminum cable steel-reinforced.
B. M. PrcKxs Is with the Anaconda Wire and Maximum tensions for power line conductors of all types are usually 60(%, 50 %, or 40% of the rated trength
Cable Company, Hastings-on-Hudson, N. Y. or some specified zusimum tension such as 2,000 lb (pounds) for distribution conductors.

1308 Pickens-Sag-Tension Calculation Program for Digital Computer FTEBRUARY 1959

 SUBROUTINE 1
ULF = 1 + a2 [a2 [(a2 x CC1) + cCj] + cC] - (T/AE) - a2 [(a2 + CC4) + CC](2)

Acceau6ator
Data Instr. Upper Lower Distributor
Location Over. Address Address 8001
Sub. 1 STD EXIT
RAU SPN 200 200
HP! w 366200 .1831
SLT 0004 3662000000
DVR T .0274102 13366
RAU 8002 .0274102
MHP 8003 .0007 5131906404
SRD 0006 .0075132
STL SWT2 .00075132
RAU 8001 .0075132
MP! cci .00000284 1674359968 .0037822424
SRD 0008 Z0000028417
AIO CC2 .0109403417 .0109375000
SLT 0007 .0109403 417
R&U 8003 .0109403
HFP SWT2 .00000 8219666196 .00075132
SRD 0005 0000082197
ALO CC3 .0416748863 .0416666666
SLT 0007 .0416748 863
RAU 8003 .0416748
MPY SWT2 .00003 1311110736 .00075132
SLT 0004 .000031311 110736
AUP ONE 1 000031311 110736 lDODOOOOO
STU LF 1000031311
RAU SWT2 .00075132 .00075132
MPT CC4 .00000273 9187474956 .0036458333
SRD 0008 .0000027392
ALO CC3 .0416694058 .0416666666
SLT 0007 00416694 058
RAU 8003 00416694
MPY SWT2 .00003 1307053608 .0075132
SLT 0004 .000031307 053608
AU? HONE - .999968692 946392 1aoo000000
R&U 8003 - .999968692
MPY T - 1335 .9581738480 13360
DER AZ .0009681501 1379908.
SRT 0001 .00096850
Fig. 1. Flow chart for main program ALO LF EXIT 1 .999063161 I900031311

Fig. 2. Typical program sheet

tension values. After values for this
span are calculated and punched, the
program continues using the regular span operation is indicated by a code. These 6. Calculate the sags corresponding to the
interval until the program is com- codes are standard for SOAP I program- two sets of conditions.
pleted. ming. In a problem of this type, a number of
It should be noted that the shortest calculations are repeated over and over.
span plus n-span intervals must equal
RAU -reset add upper accumulator
MPY = multiply These can be written as subroutines and
the maximum span, otherwise the pro- SLT = shift left used whenever needed in the main pro-
gram will not operate properly, n being DVR = divide gram. A description of the subroutines
an integer. SRD =shift right and round
STL =store lower accumulator used follows. The equations for un-
PROGRAM REQUTIREMENTS ALO =add to lower accumulator stressed length factor and sag were derived
AUP =add to upper accumulator from catenary equations developed by
In selecting a method of programming, STU =store upper accumulator Woodruff and Dwight."'3
a number of considerations enter. SRT =shift right
1. The program should minimize machine UNSTRESSED LENGTH FACTOR
In the actual program the alphabetic
time. symbols are converted to numeric. The unstressed length factor is as fol-
2. The time required to write up the The numerical example shows the method lows:
problem should be minimum. used to locate the decimal point; see ULF= [1 +(l/24)al+(7/640)a4+
3. The time required to key-punch the Fig. 2 and Table III. (244/64,512)ac I -( T/AAE) X
data on cards should be minimum.
[1-(1/24)a'-(7/1,92J)a4] (1)
4. The printed results should be in tabular SAG-TENSION CALCULATIONS
form arranged for easy reading. The
In general, the sag-tension calculations where
IBM 650 computer is a widely used general-
purpose calculator. The program has involve the following steps: a=SW/T
been developed for this computer. 1. For a given span, conductor weight, and can be rewritten for computer cal-
The effective speed of the IBM 650 and tension, determine the length of con- culation as
ductor in the span.
computer can be increased by proper ULF= 1 +a21 as [(a"X CCQ) + CC21 + CCs
location of the data and instructions on 2. Divide conductor length by span length
the memory drum. To do this with a to determine a length factor. I
(T/AE) 1-as [(a'XCC4) +CC3]) (2)
minimum effort, the program was written 3. Determine unit elongation due to the CORRECTED UNSTRESSED LENGTH FACTOR
for symbolic optimum assembly (SOAP).' tension and subtract it from the length
factor; this is the unstressed length factor. The corrected unstressed length factor
In this method a special assembly pro-
gram is used and the machine assigns
4. If there is a change in temperature, is
correct unstressed length factor for the
optimum drum locations. A typical pro- change. CULF ULF[1+a(ts-tO)1 (3)
gram sheet is shown in Fig. 2.
5. For the same or a new conductor weight, SAG
In this program, which is the com- and the unstressed length factor, determine
plete program for subroutine 1, each the tension for the changed conditions. The sag is

FE,BRUARY 1 959 Pickens-Sag-Tension Calculation Program for Digital Computer 1309

 Table II. Program Symbols Table 111. List of Symbols
Symbol SOAP Code Description Symbol SOAP Code Description Example

a. SW/T ...... (span X weight)/tension WEWB ... bare weight, lb/ft (lb per foot) . 0.1831
ASE ..AE ...... area X modulus WRW ...... WR.resultant weight, lb/ft 1.2350
AE.. ..AEI ...... areaXinitial modulus W. WI* ......... .... iced weight, lb/ft ........................... 0.7719
BY( ) ....... instruction address A .. A .......... cross-sectional area, square inches............. 0.1219
a. COEF ... coefficient of expansion Tp. F T .tension, final limit, lb. 1,336.0
CULP(N)... corrected unstressed length TR .TR ......... tension, resultant limit, lb ............... 2,672.0
factor . T TI. tension, initial limit, lb. 1,781.0
D. D( . ) .. finalsag
DIX( ).....initial sag Conductor Data, Type 6/1 ACSR
EXIT( ) .... instruction address mE.! . MES .... steel modulus......... 4,000,000
Ecs . ECS ... secant modulus nEa.N NEA.. aluminum modulus .7,320,000
fa . PA .. .. stress in aluminum EF . EP... final modulus .11,320,000
fam. FAM ...... apparent maximum stress E lEI...... Initial modulus.10,830,000
fc. FC .. .. critical stress (Zs+Za) as . COEPS .... coefficient of linear expansion, steel.0.00000640
fm. FM .. maximum stress
..
ar .COEFI ..... ...... coefficient of linear expansion, initial .......... 0.00001032
FIN ........ instruction address .COEFF
a .... coefficient of linear expansion, final. . 0.00001053
H . H .. horizontal component of gcr ... GCR .......... c...constant from elastic curve ................... 0.00015300
tension t. TEMB . data temperature, P. 70
K1. Ki1 .. . calculated constants .C1 ...c........... constant, stress-strain equation ............... 0.000165
K. - mEs X as(IB-SR) C2 .. C2.................. constant, stress-strain equation ........... 0.003450
K2 . K2.. calculated constant C., ..... 0C21.... C21 ................. constant, stress-strain equation ............. 0.532000
Kg gcr + (tB-IR) X C . ......... C3 . ... constant, stress-strain equation ............... 0.0000061
(2as-ai) C .C4 .... constant, stress-strain equation ............... 0.0000551
Ks . K3 .. calculated constant C ..C5 . ... constant, stress-strain equation ...... ......... 0.0094420
Ks - EF/nEA C . C6..... constant, stress-strain equation ............... 0.01530
K4 . K4. calculated constant Variable Data, Temperatures, Heavy Loading
K- asmEt
K5.
K. . calculated constant . TEMP .... temperature for TF, F ....................... 0
K. nEA/(mEsEP) R. TEMR. temperature for TR, P . 0
Ks K6 .. calculated constant tt TEMI ....................................... temperature for TI, P.
tn........... TEM32 ............ temperature, F ............................
0
32
K.- WB(AmEs)
LFP.........length factor In .TEMN(x) ....... temperature, P ........ -20, 0, 30, 60
LOP( ).. . instruction address (x 1 to 6) 90,120
LOOP( ) .... instruction address Variable Data, Spans
P- - .. data to be punched
Q- --........data to be punched S . SPN .. span being used in program
R(ROUT) .. main routine address SPNA............... span being used in program
RI ......... instruction address SPNY............... span being used in program
SKP( ) .. instruction address S. SPNX.... shortest span, ft. 200
SUB( )...subroutine Sm .SPNM .... l ongest span, ft ............................. 2,400
a . SWT2 .. (SWT) (SWT) Si .SPNI.... span interval, ft ...................... 100
T. T .. tension SN
.SPNN ... SPNM+SPI ............................... 2,500
TS..........controlling tension Constants
TEMP......temporary storage CC. C . . 224/64,2............................ 0.0037822420
TEM(. ).....temperature CC . CC2 . . 7/640 ... .............. 0.0109375000
TH( ) ....... instruction address CCs .... .. CC3 . . 1./24 .......... 0.0416666667
TIX( ) ...... trial tension CC4
C..
. ..7/1, 920 ......... . .. 0.0036458333
TmT..TM.... maximum tension CC CC5. 241/46, 080. 0. 0052300347
TMX( ). . trial tension . CC . . CC6. 7/384 .............................. 0.0182291666
TEMD... temperature difference CCM ..... CC7.. 1/8 ............................. 0.1250000000
ULF.... ULF ... unstressed length factor M. M2 ..... 1/40, 320 ..... 0.0000248016
Un . UN . .. length factor correction M. M1 . 1/720 0.0013888889
W . W( ) .. weight of conductor 1/2 . HALF... 1/2 0.5000000000
Z. . Z1 .. calculated constant ONE ..... 1 1.0000000000
Z1 faKu
................................

ONE ..... -1 .................................-1.0000000000

Z. Z2 .. calculated constant ONE I . . 1.1 1.1
Zs'fm-Zi
................................

TWO ......... 2 2
Zs . Z3 .. calculated constant
..............................

MONE I ..... -1 -1
ZsJ (tn-SR) K
.............................

NINE .. .. 0.9 0.9

STX 6 ... CODEt ................................ 666
* For light loading WI - WB and values for this loading are not calculated.
D Sa [(1/8) +(7/384)a2+
(241/46,080)a4. .. I (4) t If 066 is entered for TEMN 6, values for this temperature are not calculated.

which can be rewritten for computer cal-

culations as TENSION Ti-1 in the new calculation. This con-
D=Sa I a' [(al X CC) + CCs I + CC71 (5) One recurring problem is to determine tinues until Ti+1=Ti and the required
a cable tension for a given corrected un- value of T is determined.
Equation 4 is not accurate for long stressed length factor. This is done by Formula 8 has a number of pitfalls for
spans. For more accurate results, a computer operation. The initial correc-
correction can be made using the formu- an iteration process.
tions may result in minus, zero, or very
las small values for Ti+. The program must
H= T- WD (6) + ULFi- ULFi._)(i be arranged to keep the swings in T values
(8) within limits. A flow diagram for deter-
and In this formula the only known value mining T (subroutine 2) is shown in Fig.
is CULF. Two values of T are assumed. 3.
D = (HIW) [b'(1/2)+b'{ (1/24) + In subroutine 2 the tension used in
b2 [1/720+(1/40,320)b2J)} (7) T =_i=TW2/W1, Tj=1.1T1_i
determining the unstressed length factor
where Using equation 1 the corresponding values is entered and multipled by the ratio of
b= WS/2H of ULFt-, and ULF, are computed and the weight per foot for the new condition
used in equation 8. Each time Tt+, is over the weight per foot of the conductor
In the program the additional correction determined it becomes T, for a new cal- for which the unstressed length factor
is only used where SW/T exceeds 0.5. culation and the previously used T5 is was determined. This gives a trial value

1310 Pickens-Sag-Tension Calculation Program for Digital Computer FIEBRUARY 1 959

 (A)

(B)
Fig. 4. Flow charts for initial tension calculations
A-Subroutine 4, ACSR B-Subroutine 4C, copper
Fig. 3. Subroutine 2: Row chart for determin.
ing tension corresponding to given corrected Fig. 5 (right). Typical ACSR
unstresd length factor sd*s-sain chart 30

1-Initial composite, equdtion

of T which is first compared with 2TR 10
to see that it is not excessively large, 2-Initial aluminum, equation CL204<X
then with 1/2TF to see that it is not too 11
small. If it is larger than 2TR then 2TR 3-Initial steel
is used for Ti-1 or, if it is smaller than 1/2 4-Final composite
5-Final aluminum -10
Tj,, 1/2T7 is used for Ti-I. T1 is then 6-Final steel
taken L.1Ti-, and the unstressed length
factors corresponding to T1 and Ti-1 are
determined using subroutine 1. These are Fig. 6 (below). Flow chart 0
entered into equation 8 and the value of for Rnal tension calculation Q3 0.4
Ti+i determined. If Tt+1 should be (ACSR) STRAIN- PER GENT
zero or have a minus value, a new Ti-i is
selected that is one half of the previously
used value. If Ti+1 has a plus value it
is compared with Ti. If they are differ-
ent, Ti becomes Ti-I and T+1 becomes Ti.
T is then used to obtain a value of
SWIT.
If T is positive but a very small value,
SWIT will be large. For most calcula-
tions values of SW/T do not exceed 1.0.
The largest value shown in Martin's
Tables' is 1.2, so a value larger than 2.0
would indicate that the Tt+i value is too
small. If this occurs a new start is made
using one half of the previous values for

INITIAL TENSIONS
In calculating initial tensions for copper
and similar conductors the procedure is
to use a virtual initial modulus and a

FiEBRUARY 1959 Pickens-Sag-Tension Calculation Program for Digital Computer 1311l

 4
CZ

Nt

O00400
N0 82W'
000 p000
ON

"400.. 0.- OI
0 ~ ~ .0
00 Nt..8000 tNo 00 00 0
N
N
0
4m
CO
e0
2eq I- go t- 00000o
0m0t-
N
la 0 C ot..
00
o0
0
cq
N0 0"4
00~~400000 etoco 0.

00 eqCI00 W e t. 00c "40 000000

0 co C440.
00
"400 "40c000 00C
a. co Q 0006 C
2. 00't 0i

-1
"40
88-1 ic
bNNO 4
-
00
, 001 ............

80c~ 0
0000000~~c - 0 000z
coo 000000 0 0
8888...... ..
0
U
C'4
0000 04 "0 00
S 00 04Ui"4ko "44 0 cN

E
S co Neqe
VW
0 0
0. 00
0 0 "D00
000 0 C- coo VI
"a 0 a0
C
i 00
-40 0000 4m NI
8 1,-

40 ~0
II 00 S
0 le
V0
00 4
0000 00 ~~~~0000- 0 0b
b 004 000q
0 0.: coNN
000
OO8
0
O
000 0o

I 0

Ut
0
0.
40 eq0
04
"4NO 80
0
0
0
0 0

0
ONO

0
"
0

OW~
,S.
N

M4
.1 8

'Ill. C

:3 N 6010.0.000C ;C
3- C

..............~~~~~~~~..
3Et'
0000~~~~~ 00 O
49
t.
ooeq08Q
0
00
0000N
N

0; 93 400..N 00

I "400 NO0 eq. 0

88

P.*
t
r-4 -0 00

NU
V4

000oLf 00
00000
0 cq
00w 0-

N
'o~~~"
"-40

a C; aC! "40..
00I 0000.000"O0
-.4

p
co 8 08000 8000
1312 Pickens-Sag- Tension Calcukzion Program for Di-gital Computer FEBRUARY 1959
 fan =fm+(mEsas) (tB -ti)
The initial stress-strain curve (Fig. 5) is
represented by equation 10 for computer
operation. The secant modulus used
for E in subroutine 1 to obtain the initial
unstressed factor is
=
Ec,s fin
Ba-(TB -tB) (2as-at) -gcr
The initial unstressed length factor is then
corrected for temperature and used in
subroutine 2 to obtain initial tension
values. El is used in this operation.
FINAL TENSIONS
Calculation of final tensions for ACSR
conductors requires a special routine.
The method follows that described by
Jordan' except that values for the alu-
minum stress are obtained from an em-
pirical equation
fa = CJfam +CJifam"+ C2ifam (11)
A flow diagram for determining final
tensions for ACSR is shown in Fig. 6.
Above some critical temperature the
aluminum wires of ACSR lose their ten-
sion and the steel core carries the entire
load. In this range the relation between
stress and strain is that for the steel core.
A critical tension is determined for each
temperature and compared with the ten-
sion obtained assuming the cable is op-
erating as a unit. If the critical tension
Fig. 7. Flow chart For ACSR sag-tension routine is larger, than the corrected unstressed
length factor is reduced by a factor U and
a new tension, calculated by using mE,
permanent set value read from a stress- pirical equation instead of being read rather than Ef in subroutine 1 where used
strain curve. from a curve. This equation has the in subroutine 2.
The Copper Wire Engineering Associa- form The critical tension To corresponds
tion published a number of permanent set to a critical stressfc
(PS) curves for copper and copper- Ba=astrain = Cfa,n' + C4fam" + Ctfai + C6
(10) TC-fcA
covered steel conductors. These curves
were an approximate average of several Constants for a number of ACSR con- and
test curves. They can be closely repre- structions are shown in Table V. Table
sented by an empirical equation of the V also shows constants for equation 11. fc Z2 +Z.
form Flow diagrams for determining initial where
tensions are shown in Fig. 4.
PS= Cl1.f+Clep+C1Lp+C14f (9) The apparent maximum stress used in Z3 (tn -tr) co8mEs
-

Table IV shows values for the constants obtaimnng the stress and strain values Z2 fm -(faEF)I/nEa
Cui through C14 for several commonly used from the stress-strain curve for a given
conductors as well as elastic modulii and ACSR is The aluminum stressfa is obtained from
coefficient of thermal expansion values.
The permanent set value PS is sub-
tracted from the unstressed cable length
factor corresponding to fi, the maximum
stress in the conductor, giving the cor-
rected unstressed length factor to be used
in calculating initial sags and tensions.
For calculating initial tensions of
ACSR, a "secant modulus" and apparent
maximum stress as described by Jordans Fig. 8. Subroutine cTx
are used. Here again, the strain value 7: Row chart for
corresponding to a given apparent maxi- determining control-
mum stress fai is obtained from an em- ling tension

FEBRUARY 1959 Pickens-Sag-Tension Cakulation Program for Digitl Computer 1313

 ANACONDA WIRE & CABLE COMPANY
Engineering Department the stress-strain diagram, Fig. 5. For
Hastings-on-Hudson, New York the program it is obtained from equation
REQUEST FOR ACSR SAO-TENSION CALCULATIONS 11.
iBM 65o
The correction factor is
s210 6 AL / ST ACSR Z.763 LBS(50j%) HEAVY
Size, Awg or Mcm Strand, Al/St Max. Tension (TRY Loading Dirsrict
U (fcnEa) /(mREs)
RATED STRENCrH_545j
CARD DRUM
SAG PROGRAM (R)
CODE V A L UE NO. LOCATION DATA
31 --40 13-16 Z3-26
For a given span the following steps
are used in calculating the sags and
0842 0448 Cable Weight, Lbs/Ft
WR .1,22.0.0 0843 0398 Resultant Loading, Lbs/Ft tensions
WI . 77.Z 0844 0948
. 2.L i
Iced Weight, Lbe/Ft
A o845 0224 Cross Sectional Area, Sq.In. 1. Calculate unstressed length factor for
Tr o846 0040 Final Tension Limit, Pounda the limiting final unloaded tension using
0847 0365 Temp. at Final Limit, F
TR .216.7t,25 0848 0235 Maximum Tension, Lbs equation 1.
0849 0597 Initial Tension Limit, Lbs 2. Correct unstressed length factor to
o850 0415 Temp. at Initial Limit, F
SPNX ,, Q.0, o851 o848 Shortest Span, Feet temperature of the maximum tension limit
SPFNM Q0 0852 0058 Longest Span, Feet using equation 3.
SPI . z.0.Q o853 o656 Span Interval, Feet
3. Calculate a tension based on the cor-
Use Deck A For Temperatures (7 Cards) rected length factor and the resultant load-
Use Deck a For Stress-Strain (16 Cards)
ing We using subroutine 2.
STANDARD TEMPEATURES AND LOADING DISTRICTS
4. Compare the tension obtained under
CARD DRUM SPN item 3 with maximum allowable tension
OODE HEAVY MEDIUM LIIGHT CAL,LT. SPECIAL NO. LOC. VALUES TR, and select the smaller value.
im -0 15 -30 - -SP S14
TEKNl -20 0 20 25 0813 0210 ro -
-7Z 5. Using the tension selected under
TEN2 0 15 30 50 08114 0535 100 600 50 item 4, calculate the initial tension at the
TWMN3 30 30 60 70 o815 0736 200 1000 100
TD411N4 60 60 90 90 0816 1037 1400 1600 200 temperature at which the initial limit
TEHN5 90 90 120 130 0817 0288 600 2400 applies using subroutine 4 or 4C.
TEMN6 120 120 150 666 0818 0989 3200
DECK XQ B C D 6. Compare the initial tension value
obtained in item 5 with the initial limit
STRESS-STRAIN (CARD NOS. 0826-08141)
T,. If it is smaller than T, then the
Strand 6/1 1 7/1 214/7 26/7 26/7 30/7 45/7 54/7 514/1930/19 8/1 12/7 tension obtained under item 4 is the maxi-
Size 8-2 1-0 14-2 All To605 605UP ALL ALL ALL ALL ALL
Deck a 1 J K L N N 0 P Q A T mum tension for this span.
ALT. VALUES TR & SFNLWAD: )INO CODE 7. If the initial tension obtained under
ITR % SPNX SPNM Location Instruction item 5 is larger than T1 then the maximum
(2) I - 1 69- 19541 1953 tension must be reduced until an initial
21 - 22 214 tension value equal to Tr is obtained.
27-30 8000
140 and 80 12 Pch. This is done by assuming an arbitrarily
(5) ; ; ..- = smaller value of TV and calculating the
By h Date
t X-l 8 corresponding initial tension by an iteration
process.
Fig. 9. Form for sg-tension request Using the formula

Tm+'= + (Tm+I- Tm) +Tm+a (12)

2/0 6/1 ACSR MAX T 50% HEAVY
the maximum tension value is obtained.
IC 1/2 1/2 0 0 0 0 0 0 8. Once the maximum tension and corre-
WIN4D 4 0 0 0 0 0 0 0
TEN 0 32 20- 0 30 60 90 120 sponding unstressed length factor are
SPAA0 established, the tension values for other
2C 293 263 57 69 97 147 212 241 temperatures and loading conditions can
2C 20848 14676 16131 13364 9442 6219 4318 3805 be obtained. Subroutine 3, subroutine 2,
2C 49 57 74 102 151 217
and subroutine 6 are used to obtain final
2C 18645 16117 12410 8940 6076 4224
4C 917 832 298 357 463 579 660 710 tension values, subroutine 6 being used
4C10 26725 18632 12295 10274 7920 6340 5559 5174 only for ACSR while subroutines 4 or 4C
4C10 201 231 290 372 473 582 are used for initial tension values. Sub-
4C10 18197 1 5869 12624 9861 7757 6299 routine 5 is then used to obtain all sag
2194 508 453 115 138 189 265 350 386
2134 242 82 17220 16003 1 3 363 9768 6979 5284 478 5 values.
2834 90 1u3 19 sb9
16 3 ,3 n
3.1U
12 . A
9. Eight tension values are stored in the
2134 20473 1 7979 14337 10940 8052 5965 . ..
6C 1o 2077 1974 1410 1497 1621 1716 1777 1838 first eight positions of a 10-word punch-
6C 1o 26725 17778 5872 5534 5117 4839 4674 4523 out band of the computer and the corre-
60 1o 1152 1245 1379 1506 1625 1739 sponding eight sag values are located in
6C 1o 7174 6644 6006 5505 5104 4774 the first eight positions of an adjacent
8CD0 3726 3616 3091 3170 3284 3355 3417 3479
SCDO 26725 17406 4805 4689 4531 4438 4360 4285 band. With this arrangement the eight
SC30 2860 2942 3061 3176 3287 3395 sag values are punched in the first output
SCDO 5182 5042 4851 4680 4526 4387 card and the eight tension values in the
1O(00 5890 5778 5274 5350 5455 5517 5581 5644 following card. Span lengths are punched
10%00
10(
1G00
26725
37220 5061
4630
5136
4566
4312
4475
4266
4387
4220
4306
5568
4229
into the left four digits of the first word
location. (Each word location has ten
digits.)
12200 7721 1831 1219 26725 13363 17817 0
0 165 3450 532000 61 551 94420 1600 10. Initial values are placed in locations
4 7320000 1320000 330000 640 1032 1053 15300 so that they will print under the corre-
sponding final valtues, i.e., the 120 F initial
Fig. 10. Sag-tension data for Fig. 9 values will be located under the 120 F

1314 Pickens-Sag-Tension Calculation Program for Digital Computer FEBRUARY 1959

final values, etc., when listed. Fig. 7 is to clear the memory to minus zeros amount of engineering time in setting
shows a flow diagram for the complete then to set the storage entry switches to up the information to be punched. The
ACSR tension-sag operation. program is in machine language which
70-1951-1953. When loaded the pro-
11. Subroutine 7 is used to determine gram is started by setting the storage permits rapid calculation. Results can be
which of the three possible tension limits entry switches to 00-0000-0002 and push- printed out in a form convenient for use.
would apply for a given span. It is used ing the program-reset and program-start One program deck is used for ACSR
in locating the span where there is a change
in tension limits. The flow diagram is buttons. conductors, a separate but similar pro-
shown in Fig. 8. Fig. 9 shows forms used The program is assembled with the gram is used for copper and similar con-
to request sag-tension calculations. Only subroutines first, then the main routine, ductors.
nine cards are usually punched for a given followed by the data cards.
request, one word to a card. The standard
temperature values, stress-strain data cards, References
and span-length cards are assembled with SAG-TENSION DATA
the main routine cards as well as the nine Sag-tension data corresponding to the 1. 650 PROGRAMMING. BuUetin 1, Symbolic
data cards and loaded with the main Optimum Assembly Programming (SOAP), Inter-
program. After the program is loaded it
ACSR request shown in Fig. 9 is shown in national Business Machines Corporation, New
Fig. 10. The first data line across shows York, N. Y., 1956.
may be started without the addition of any 2. PRINCIPLES OF ELECTRIC POWER TRANS-
more cards and it will calculate and punch final sags. This is followed by final MISSION (book), L. F. Woodruff. John Wiley &
out sag and tension values for the range of tensions, initial sags, and initial tensions Sons, Inc., New York, N. Y., second edition, 1938,
pp. 204-05.
spans covered. At the end of the calcula- for the span lengths shown at the left.
tion the data used will be punched so that 3. SAG CALCULATIONS POR TRANSMISSON LINES,
a record will be made which can be checked. H. B. Dwight. ArEE Transactions, vol. 45, May
To calculate additional values a new maxi- Conclusions 1926, pp. 796-05.
mum tension limit or change in span data 4. MARTIN'S TABLES (pamphlet), J. S. Martin.
can be entered in the console. From 1/2 Copperweld Steel Company, Glassport, Pa., 1931,
A method of calculating sags and ten- 64 pp.
to 1 minute of machine time is required sions has been programmed which re- 5. A SIMPLIFIED SAG-TBNUON MBTHOD FOR
for each span length calculated. STEEL-REINFORCID ALUMINum CoNDUCToRS, C.
quires a minimum number of cards to be A. Jordan. AIEE Transacions, vol. 71, pt. III,
The procedure for loading the machine punched for each problem, and a minimum Dec. 1952, pp. 1108-17.

Digital Computation of Synchronous terminal voltages regardless of system

conditions, and must on occasions be
restrained from reducing the generator
Generator Pullout Characteristics field below the value required to main-
tain stable coupling between the turbines
and the electric system.
Methods of making the necessary com-
W. F. CAWSON H. E. BROWN putations of the relation between mvar
ASSOCIATE MEMBER AIEE MEMBER AIEE
(megavars) versus mw (megawatts) and
field current versus mw are well estab-
THIS PAPER describes a digital com- when taking transmission facilities out of lished and have been described in previous
puter program developed to carry out service and, in preliminary planning, to literature.'14 The method adopted by
the necessary computations in developing determine the adequacy of present trans- the authors is that of determining the
generator stability limit curves. The mission lines or the need for future trans- excitation required to maintain an
use of this program has resulted in a con- mission lines with respect to stability operating power angle between the
siderable saving of engineering time. requirements. assumed internal voltage es and the volt-
The large amount of line-charging age of an infinite bus ea, that will provide
kilovolt-amperes of the cables of a metro- The Problem the desired margin of stability.
politan network and the installation of Where it is desirable to take account
capacitor banks on the distribution sys- The magnetic fields of the generators of magnetic saturation it is possible to
tem both in urban and rural areas reduce to a great extent control the voltage of the use the reciprocal of short-circuit ratio
the excitation requirements on the gen- system and provide the magnetic coupling in place of xd in the calculations. This
erators and thereby weaken the margin that transmits power from the turbines will give slightly less conservative
of stability during light load periods. To to the electric system. At times there is values, but can be justified in a metro-
assure adequate excitation at all times, an area of conflict between these two politan system where the generator can
minimum field current versus load curves functions. The more efficient units, carry- be considered directly connected to an
must be obtained to facilitate proper ing heavy loads during system light-load infinite bus.
setting of the voltage regulator lower periods, as dictated by economy, require
limits. Operating personnel should be relatively strong fields to maintain syn-
supplied with curves showing permissible chronism, while the demand for voltage Paper 58-1091, recommended by the AIEE System
Engineering Committee and approved by the
reactive loadings as a guide to the opera- support to the system is low due to the AIEE Technical Operations Department for
tion of the system. The limits imposed permanently connected capacitive equip- presentation at the AIEE Power Industry Com-
puter Application Conference, Toronto, Ont.,
on the generators, as shown by these ment providing leading reactive some- Canada, September 15-17, 1958. Manuscript
curves, are calculated for special or times in excess of the customers' reactive submitted June 5, 1958; made available for printing
August 21, 1958.
contemplated system conditions and are demand. The generator voltage regula- W. F. CAWSON and H. E. BROWN are with Common
used to indicate the liability involved tors tend to maintain the specified wealth Edison Company, Chicago, Ill.

FEBRUARY 1959 Cawson, Brown-Digital Computation of Generator Pullout 1315