Vous êtes sur la page 1sur 9

# Stress-Strain, Creep, and Temperature Dependency of ADSS (All Dielectric Self Supporting) Cables Sag & Tension Calculation

Cristian Militaru

## Alcoa Fujikura Ltd., Spartanburg, SC Abstract

It has been common in the industry to calculate sag & tension charts for ADSS cables without taking into consideration the influence of creep, coefficient of thermal expansion (CTE). and the difference between the initial and final modulus. In some applications where the sag and tension performance of the cable is not critical. the presentation of data in this manner is appropriate. However, the great majority of applications require very exact determination of sag and tension, and the influence of the above factors is important. There is also confusion between the final state (after creep) and the loading condition (wind+ice). which are 2 different cases. Following thorough and repeated stress-strain and creep tests, this paper will show that ADSS cable has both an initial state and a final state, each state having an unloaded (bare cable) and a loaded (ice and/or wind) case with resulting sag 8 tension charts as a function of creep and CTE. Additionally, the results of this work were compared and validated by common industry sag &tension software, including Sag10 and PLS-CADD.

Y 1
w

.x + k,

## Catenary Curve Analytic Method

Fig.1 presents an ADSS cable element under the extrinsic (wind and ice) stresses and intrinsic (cable weight) stresses, with a length, On the curve y(x), given by the formula: I= I,/=. dr (1); yields: g = ,/q XI Also,the equilibrium equations results in: (4) considering rel. (2). the derivative of rel. (4) yields: (2)

## 4-Z = .(f+) y+ l + y (12) which has as solution: y = sinh&.x+k,)

(13) ; integrating rel.(13) results: (14);for x=0 results:

## y=f.cosh(\$.*+k,)+k, ,,=c(15)and: y=O

(16),so:

k, = k, = 0 (17).

H, =H, =H (3); dV dl

V,-V, =dV=w.dl

## - = W.-g= w -F l+y,,, (5); also, the slope in any du

point of the catenary curve is defined as the first derivative of the function Y(~) of the curve: V=H.tancp=H.%=H.y(6);yields:

dV -= H 4 a!x Yz
and (8) results:

H. y = w .
integrating rel.

## (9) , and then:

(10),

y = sinh? (20). a In Fig.2 the designations are: S= span length B=S/2= half span length (assuming level supports) D= sag at mid-span H= tension at the lowest point on the catenary (horizontal tension) - only for level span case, it is in the center of the span T= tension in cable at structure (maximum tension) P= average tension in cable L/2=arc length of half-span I= arc length from origin to point where coordinates are (x,y)

## International Wire & Cable Symposium Proceedings 1999

605

a and C respectively = distance of origin (of support respectively) from directdx of catenary t= angle of tangent at support with directrix k= angle of tangent at point (x,y) w= resultant weight per unit length of cable E = cable strain (arc elongation in percent of span).
n

## At the limit shown in Fig.2, for: x =: = becomes:

B. rel. (18)

yielding the parabolic equation: 3 terms formula for rel. (28) results in:

(36)

D=u.[l++(+)2

+\$@-*]

(37)

which, using B=S/2 and rel. (19) yields the approximate catenary formula : 1 D

= g+&:;31

(38)

For sags larger than 5% of the span, ret. (36) the parabolic equation gives erroneous results, while rel. (38) provides a more accurate solution, the exact solution being given by rel. (28).

## Example for an ADSS cable on transmission lines

C=a.o&(21); a and:sinhX=6.5. a I = jm.dx= 0 [= where:c&z=OS. (22)

The following variables, for the same span, have the same values for any material (ACSR, AAC, EHS, ADSS, etc.) as long as they respect the cetenary equations.
Span: S=1400 [ft]; w=1 [Ibs/ft]= constant value; H=9038 [Ibs]=assumed value; a=H/w=9038 [fl]; B=S/2=700 [ft];

(23),

so,

jJq.dr 0

(24),or,

C=mosh~=9065.12

[fit

D=C-a=27.12 [ft];

## jcosh; . cln (25), resulting from rel. (25) the cable

(28). At the limit, for (20 -1

L=2.a.sinhB=1401.4 Ml; E= i-1 .100=0,1M a ( 1 \$. 100 = 1.9372 [%I; T= IV. c = 9065 [Ibs]; H+T ?=9Ocj5 [Ibs]; PC---=9052 Dbsl: p = 9052 Ws1 2
s,me of the ADSS characteristics, presenteWd in this example, are:d=0.906 [in]; ,.f = a.& = 0.6447 [in2];

## length equation: x =: = B, rel.(26) becomes

wc=0.277 [Ibs/ft].
Loading Curve Type: B= ADSS w/o ice or wind Cbare unloaded cable): resultant weight: wr=wc=0.277 [Ibs/ft] H= ADSS plus heavy loading: according to NESC: Regular ice density: rice = 57 [lbs/ft3]; Ice radial thickness: t=0.5 [in]; Temperature: 0 = Wind velocity: VW=40 [mph]; NESC factor: k=0.3; Wind pressure: Pw=0.0025,Vw2 =4 ]ps6; Ice weight:

0 [OF]:

is defined as arc elongation in percent of span: A Taylors series for cash yields:

w,, = - . 4

[Ibs/ft]

606

## International Wire & Cable Symposium Proceedings 1999

Resultant weight: wr

[lbs/ft]
0.277 1.615

## S=1400 [ft] Stress [psi] +(pj (509g+9 @q.g=m

B H

This catenary table is transformed in a Preliminary SagTension Graph, in Fig.4. This graph has 2 y axes: left side: stress [psi], og and OH: B-bare cable: H-heavy load. and right side: sag: D [ft]. Also, it has 2 x axes, strain. E [%] (arc elongation in percent of span) and temperature, 8 WI.

Stress-Strain Tests
Stress-strain tests on ADSS cable performed in the lab show (see Fig.3) that they tit a straight line, characterized by a polynomial function of 1st degree, whereas metallic cables (conductors. OPT-GW, etc.) are characterized by a polynomial function of the 4th degree (5 coefficients). From ail the tests performed, results show, that differences exist between the initial modulus. E i (slope of the charge curve) and the final modulus, E f (slope of the discharge curve) and their values depend upon the ADSS cable design. There is also a permanent stretch, E p (also referred to as set"), at the discharge, which depends on the ADSS design. Creep Tests According to the ADSS cable draft standard, IEEE P12221, the creep test must be performed at a constant tension equal to 50%. MRCL for 1000 hours at room temperature of 60 F. in general, for ADSS cables. MRCL=[%MIN...%M~RBS therefore the test is done at T=[%MIN/2...%MAX/2] RBS=ct. (see Fig.3). Considering a nominal value of MRCL=50% RBS, the default constant tension for the test would be: T=25%. RBS. The creep test on the cable design analyzed was performed at a constant tension: T=50%, MRCL=28Yc RBS, because for this cable: MRCL=56%- RBS. The values ware recorded after every hour (see Fig.8CreepTest, Polynomial Curve and Fig.G-"Creep Test: Logarithmic Curve). The strain after 1 hour, defined as initial creep, was 42.89 [pin/in]. After 1000 hours (41.8 days) the strain was 314.10 [pinlin]. So the recorded creep during the test, defined as strain at 1000 [h] minus strain at 1 [h], is 271.41[pin/in]. The extrapolated value after 87380 hours (10 years, 364 days/year) was 1142.69 [pin/in]. Therefore, the IO years creep, which is defined as strain at 87380 [h] minus strain at 1 [h]. was 1100 [pin/in]=O.ll [%I. Other creep tests performed on other ADSS cable designs showed IO years creep values in the same range. The curves on the stress-strain and tension-strain graphs are identical. The only difference is that on the ordinates (y) axis, when going from tension [ibs] to stress [psi], there is a division by the cross-sectional area of the cable, A [in2]. The values on the strain (x) axis remain the same. For the stress-strain graph (Fig.3) at a tension (stress) equal with the value for which the creep test was performed, T=50%MRCL (a=50%. MRCL/A), a parallel to the x axis intersects the initial modulus curve in a point of abscise, 0.5. E MRCL, and from that point, going horizontally. adding the IO years creep value of 0.11 [%I, is obtained the point on the IO years creep curve corresponding to that tension for which the creep test was performed. Drawing a line from the origin through that point gives the slope (the modulus) for the IO years creep, E 0 Always,

Tensions Limits: a) Maximum tension at Oo F under heavy loading not to exceed 51.35% RBS: MWT=51.35%RBS [Ibs]. I I Note: M& (max. working tension) was selected less than MRCL, in order for this ADSS cable to cope with limit c) presented below. b) initial tension (when installed) at 600F w/o ice or wind (bare cable) not to exceed 35% RBS:

TEos, = 35%RBS
c)

[Ibsl;

Final tension at 600F w/o ice or wind (bare cable) not to exceed 25% RBS:

Guide to Columns: 1 and 2 are the same for any span, any material. 3,4,5,6 are the same, for the same span for any material: ACSR, AAC, EHS, ADSS, etc. 7 and 8 are different, from one material to another (ACSR, AAC, EHS, ADSS, etc.)

607

Anqle:

a'>d

>p

tnOl

.A0

608

I 100

## I 1000 TIME: t 1 hours 11

/ 10000

.: 100000

Fig. 5- Creep test for a particular ADSS cable : Polynomial Curve log E = 0.2889 .log(t)+ log42.69

## 87360 h (10 yn)

1 year=364 days

Fig. 6 - Creep test for a particular ADSS cable : Logarithmic Curve International Wire & Cable Symposium Proceedings 1999 609

for any ADSS design, the relation between the 3 moduli is Ef>Ei>Ec.

## Coefficient of Thermal Expansion

The values for CTE (designated here as a) were determined by the individual material properties in a mixture formula: a = ,i,

## where oi, At, Ei are

the CTE, cross-sectional area and modulus of each one of the t elements in the ADSS construction respectively. For the great majority of ADSS cable designs, the influence of CTE is smaller than that of creep. Designs with a low number of aramid yarn ends (typically for short spans) will yield larger differences in sags due to temperature than designs with a high number of aramid yam ends. This is due to the fact that the aramid yarn is the only element with a negative a , while the rest of the elements have a positive a. To appreciate the impact of the contribution of aramid yam to the ADSS CTE, designs with low number of aramid yam ends have a CTE typically in the range 2.10-6 [l/OF] to 9.10-8 [l/OF which is relatively close to aluminum, CTE=l2.8.10- tl [l/OF], and sometimes larger than steel, CTE=8.4.10-6 [l/OF]. The CTE for cables with higher numbers of aramid ends are often 100 to 1000 times smaller, 2.10-8 [l/OF] to 8.10-g [IloF]. and so, for those designs, the influence of CTE on sag is negligible.

Sag-Tension Charts
The well known general equation of change of state: (40)

## shows- that the change in slack Is only equal to the

change In elastic elongatlon + change In thermal elongation, and does not include the change In plastic elongation (the creep). Therefore, the above relation is true only if the 2 states of the cable are in the same stage, initial or final. When viewing sag charts (Fig.1 1 & Fig.12). this equation will allow a user to go only vertically from one case to another case, but it will not allow him to go horizontally (same temperature, same loading conditions, from initial stage to final stage) due to the influence of creep. A simplistic way of solving this issue which is still used in some European countries is the following: the creep influence is considered to be equivalent with an off-set temperature, 8creep. g iven by the ratio (conductor IO yrs. creep-initial elongation)/ CTE. But this is not an exact method, because it only calculates an INITIAL sag&tension chart, with the FINAL sag&tension chart being identical with the initial chart, the only thing is that the Initial chart,

(Fig.8) exceeds the flnal sag at OoF after heavy loading=16.35 ft (Fig.7), creep is the governing case.
For this case, users of Sag10 will see the flag, CREEP IS A FACTOR. SAG10 will print only the final chart afler creep (not the final chart afler heavy load). Users of PLSCADD will see the same results in the chart called FINAL AFTER CREEP (see Fig.1 1 & Fig. 12). The final sag and tension at OoF must now be corrected using the revised stress-strain curve. For this purpose, the temperature axis will have an off-set of 0.01992 [%] to the lefl (Fig.8) to provide the values at OoF. Therefore, the stress-strain graph is moved to left (Fig.8) until OoF on the temperature scale coincides with reference point R (Fig.9). The corrected final tension at OoF, bare cable, (after creep for 10 years at 8OoF)=5720 psi (3887 Ibs) is found at the intersection between curves 3b and B. The corresponding final sag (18.40 ft.) is found vertically on curve D. The final tension at OoF under heavy loading, (after 10 years creep at 8OoF)=lO900 psi (7027 Ibs) is found at the intersection between curves 3b and H. Its corresponding resultant final sag (58.07 fl) its on curve D (Fig.9). When

is moved to align It with the new corresponding temperature. Therefore, the final sag at temperature 8 is
equal with the initial sag at temperature 9+9creep. The most accurate and exact solution is the graphic method. In this method, which was developed by Alwa2, 3 , the stressstrain graph (Fig.3) of the ADSS cable is superimposed on the ADSS preliminary sag-tension graph (Fig.4), so their abscissas coincide and the whole system of curves from Fig.3 are translated to the left, parallel with the x axis. up until the initial curve. noted 2, in Fig. 3 (and also in 610

## International Wire & Cable Symposium Proceedings 1999

I.,.: ,.. ..I... ,..I./,<., 0.r Y.,..... . O.F ..d..... L.., ,... ,,. ,,.*. . . . . . . . I. *A*,*

pg:::

## International Wire & Cable Symposium Proceedings 1999 611

Fig. 10- Final Trial. for IZOOF, after adjustment for 10 years creep correction

612

## International Wire & Cable Symposium Proceedings 1999

References 1.
2.

3.

IEEE 1222P- Standard for All Dielectric Self-Supporting Fiber Optic Cable (ADSS) for use on Overhead Utility Lines - Draft, April 1995 Aluminum Electrical Conductor Handbook. chapter 5- third edition, 1989 Alcoa Handbook, Section 8:Graphic Method for Sag Tension Calculation for ASCR and Other Conductors-1970

received MS degree (1980-1985) and Ph.D. degree (1990-1995) in Electrical Power Engineering from Polytechnic University of Bucharest, Romania. He worked for 11 years as a Transmission Design & Consultant Engineer in the power utility industry in Europe, Middle East and SouthEast Asia. Since 1996 he has been employed with Alcoa Fujikura Ltd., USA, as a Development Engineer in the OPT-GW 8 ADSS cable and hardware department. Mailing Address: Alcoa Fujikura Ltd. P.O.Box 3127, Spartanburg, SC 29304-3127.
Mr. Cristian Militaru

613