ML20137X400
| ML20137X400 | |
| Person / Time | |
|---|---|
| Site: | South Texas  |
| Issue date: | 09/09/1985 |
| From: | Rodabaugh E E.C. RODABAUGH ASSOCIATES, INC. |
| To: | Hou S NRC |
| Shared Package | |
| ML20137X398 | List: |
| References | |
| TAC-57279, NUDOCS 8603060217 | |
| Download: ML20137X400 (5) | |
Text
.
ATTACHMENT ST-HL AE-10//
PAGE S OF 7 E. C. eR~LL4okoeiata, Du.
4625 CEMETERY ROAD e HILLIARD OHIO 43026 614/876-5719 September 9,1985 Dr. Shou-nien Hou Hail Stop P-522 US Nuclear Regulatory Commission Washington, DC 20555
Subject:
BTP HED 3-1, Relation Between Usage Factor and Allowable Stresses
Dear Shou:
In our telephone conversation on 9/6/85, I noted that a relationship exists between a reduced allowable stress and the corresponding reduced usage factor.
The relationship is:
= (3/8)5 (I)
Ur
]
where U = reduced (below 1.00) usage factor r
S = reduced (below Codo-allowable) stress r
S = Code-allowable stress In BTP MEB 3-1, the stress criteria is S = 0.8S; hence, the corresponding usage i
r
~
factor is U = (0.8)5 = 0.328 r
Ausagefactorof0.1correspondsto(S/S)of 0.1 2 = 0.631; a usage factor of0.4correspondsto(S/S)of 0.4 2 = 0.833.
Equation (1) follows directly from Markl's" equation:
8603060217 860228 (2) f = (490000/iS )5 I-N r
PDR ADOCK 05000490 A
]
- " Fatigue Tests of Piping Components", Trans. ASME, Vol. 74, pp. 287-303 (1952).
ATTACHMENT i
ST.HL-AE-/det/
PAGE V OF 7
, j
'i l
The g = number of cycles to failure for a cyclic stress range, S,, psi.
where N 1-factor depends upon the type of piping component but not on the number of cy-l cles. While the constant of 490,000 needs to be changed for design application to give a margin between failure and design, the key aspect here is the exponent i
of 5.
i I
The cumulative usage factor is obtained by:
U, = n /N f n /N2 * "3!
(3) g 3
2 3
Because Eq. (2) applies to each n /N M Eq. (3), it follows that limiting U, to, g 1 for example, 0.4, corresponds to reducing the allowable cyclic stress range for each pair of load sets by 0.4 2 = 0.8326.
- I Equation (2) provides a basis for Class 2 and 3 evaluation rules. However, in BTP MEB 3-1, the cumulative usage factor restriction is applied to Class 1 piping,notClass2/3 piping. Iooking at Code S-N graphe (e.g., Fig. I-9.1 and 4
f I-9.2), it appears that the slope varies with N.
For N between 50 and 10, the slope is more represented by N = (a/S)3 rather than N = (a/S)5 However, those portions of the S-N curves cannot be used directly for Class 1 piping fatigue evaluation because Sn # 30 ; see NB-3653.1, Eq. (10).
s NUREG/CR-3243 discusses how the Code S-N curves can be adjusted, wing the i
K,-factor; See NB.-3653.6, Eq. (14). The enclosed pages 14 and 17 illustrate l
how, after adjusting the Code S-N curves by the K,-factor, the slope for all N becomes reasonably representable by N * (a/S)3 Indeed, if Class 1 fatigue eval-untion is to be credible, such a correlation is necessary because Eq. (2) is bas-ed directly on a rather extensive usount of fatigue test data.,
1
e;,
J ATTACHMENT ST HL AE-/6//
PAGE 5 OF 7
,,p o
Accordingly, "or Glass 1 piping, Eq. (1) is deemed to provide an appro-priate relationship between U ands /S. If one wished to have a U, that cor-5 responded exactly by Eq. (1) to S/S = 0.8, then U m 0.8 = 0.33 r
Yours very truly, Od E. C. Rodabaug I ~
ECR/ar Enclosures I
j l
n b
b j
i i
i
-.g y
.,a,
, - -,, -, - ~
,,-r,
,,,m w---
.-c-,
,v,,---,.--
O ATTACHMENT ST-HL-AE- /fo//
_' 2 3
..s-
_PAGE fo 0F 7 l /f *) t N '.* f" - 3 2 6 14 o
n and S = 2S /K ;
for S < 3S (22) n a
a m
i After determining S by Eq. (20), the peak stress is
}
9 1.
S =KS w
(23) g p
an Figure 4 shows the Code 1 S, vs N curve, Code Fig. I-9.1, for UTS I
= 20 ksi, m = 3, n = 0.2, I = 2.0,* )uch a s k
as dashed lines, are the adjusted (for K curves 80 ist. Al so shown, s
4 for a carbon steel with S SA106 Grade B st temperatIres up to 400*F.
An example of the development k
of the dashed curves, for N = 108, K, = 2.0, is:
{s 1.
At N = 108, S, = 83 kai (from Code Table I-9.1).
{
2.
F4aation (20) with I
.0, K, = 2.0, and 3 S, = 60 kai give s S, =
[2 + (4 + 32 x 83/60)
]/(8/60) = 67.1 ksi.
3 3.
B e c a u s e S, ) 3 S,,
Eq. (20) applies; not Eq. (21) or (22).
j 4.
Equation (23): S 2 x 67.1 = 134 ksi, s.'.
=
is plotted at h = 108 to establish a point on the dashed line h
t 5.
S
!!beledI,=2.
f This procedure is repeated to obtain other points on the dashed lines in f
Fig. 4.
The I adjustment stops to the right when S = 3S and there is a e
n a
f portion of the dashed curves at the extreme lef t where Eq. (21) controls.
(
1, il d !
.C ORNL-DWG s3-4598 ETD l
4 l
1 l
8 l
jSM CODE 1
[
~
w CODE 1 ADJUSTED FOR K2 = 2.0 l
- *.,, ' ' ~ ~..,, '
a:
$ i00 EO. (24) 2 5 = 490,000 N-0.2 4
CODE 1 ADJUSTED FOR K2 = 1.0 a 50 j
N
\\
s' CODE 1 e
I I
I e
e i
l i
,o to 102 103 104 105 106 f
DESIGN CYCLES, N l
I Fig. 4.
Comparison of Eq. (24) with Code 1 including K adj us tme nt, SA106 Grade B up to 400*F.
y, y
h 4
e-n-
Al I ACHMENT ST HL AE-/6 //
' r, ',-
PAGE r/ OF 7 t-
?*
17 o / N'" cM./ N W i1
(
test s were run on piping components made of raterials with SHowIver,ging f rom ran l
62.4 to PS.3 ksi, and S ranging f rom 38.9 to 56.2 ksi.
fatigue y
failures in girth butt welds may be related to the properties of the weld metal or heat-affected zone rather than the base metal. We wonid specu-l ",
late that girth butt velds in SA672-J100 pipe would not be much better-than girth butt welds in SA106 Grade B, and in this particular respect.
Code 1 is probably more accurate than Code 2.
Note, however, that ratios l
is Table 1 are to the basic piping product fatigue curve, Eq.(24), and that Code 2 allowable f atigue design stresses, as illustrated in Fig. 3 I
contain a large margin for a low number of cycles.
A somewhat analogous situation exists in ANSI B31.3-1930, Chemica!
Ff arre and Petrotsum Refinery Piping, where changing the allowable stress basis free a f actor of 1/4 to 1/3 on S increased allowable de sign f atigue stresses by a f actor of 4/3 for some mIterials and temperatures (e.g.,
for A106 Grade B from 15 to 20 kai for temperatures up to 400*F).
1 1
4.4 Austenitic Steels. A11ovs 600 end 800 l
l Available f atigue test data on piping components made of type 304 anstenitic stainless steel and dimensional equivalent piping components made of SA106 Grade B carbon steel are abstracted in Appendix C.
These dets indicate that SA106 Grade B components are slightly stronger than
(
type 304 components. Accordingly, it is pertinent to continue compari-se n s w i th Eq. (24).
000 W ",
F1gure 5 show: comparisons be tween Eq. (24), 2iS
=
d KS.
Figure 5 is f or an austenitic steel with S
=
s ad Eq. (23 ), S
=
a o
an 20 kai (e.g., SA312 type 304 at 100*F). For sustenitic steels (and alloys 600 and 300), m = 1.7, n = 0.3, and I = 3.333.
ORNL-oWG s3-4599 E TD I
l 1
[
l l
5 l
l CODF1 7
r CODE 1 ADJUSTED FOR K2=20 3i f-E O. (24)
-I..
/
2.S = 490.000 N-o 2 d
g t o)
,~~~~
y 7
CODE 1 ADJUSTED FOR K. = 1.0 g
r i
l I
t I
i l
I
,o in 107 103 108 105 108 DESIGN CVCLES, N
(
Fig. 5.
Comparison of Eq. (24) with Code 1 including K, adj ustment, SA312 type 304 a t 100*F.
-