ML20206H530
| ML20206H530 | |
| Person / Time | |
|---|---|
| Issue date: | 08/19/1985 |
| From: | Hazelton W, Klecker R, Koo W NRC |
| To: | |
| Shared Package | |
| ML20206H509 | List: |
| References | |
| NUDOCS 8606260171 | |
| Download: ML20206H530 (28) | |
Text
{{#Wiki_filter:- Conference on Structural Mechanics in Reactor Technology, 8/19-23/85, Brussels, Belgium D 1/9 TECHNICAL BASES FOR INTERIM LICENSING ACTIONS RELATED TO BWR PIPE CRACKING Warren S. Hazelton, William H. Koo, and Raymond W. Klecker U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Introduction During a hydrostatic test in March 1982, slight leakage was detected from two recirculation system pipe welds at the Nine Mile Point nuclear power plant. The ensuing investigation led to extensive examination of austenitic stainless steel piping for IGSCC in all operating BWR plants in the USA. Many cracks were discovered; some were significant, others were of relatively minor extent. Repair of the significantly cracked welds was usually accomplished by applying weld overlay reinforcement; whereas welds with minor cracking were evaluated for, limited continued service in accordance with IWB-3640 of Section XI of the ASME Boiler and Pressure Vessel Code. Although IhB-3640 is relatively simple to apply, it required that an . analysis for further crack growth be performed. As there was no generally accepted procedure for such IGSCC crack growth calculations, the NRC staff developed a set of bases and calculational methods that would be acceptable to the staff. 8606260171 860619 PDR MISC 8606260147 PDR
This paper describes the staff approach and provides the specific parameters used. Crack Growth Calculations Crack growth calculations are required to evaluate the continued structural
,. integrity of a weld with known cracks, if it is desired to continue operation without repair or reinforcement. The rate of growth of IGSCC is not easy to predict, because the several important factors are usually imperfectly known.
Research work in this area has been helpful in defining the general effect of these factors, but a large uncertainty in crack growth predictions still remain. g Nevertheless, crack growth calculations can be performed within certain limits l-with enough confidence to ensure plant safety without excessive conservatism. i i Crack growth calculations are based on the fundamental concept that the crack f growth rate of a specific material in a specific environmeat will be a function I j of the applied stress intensity factor K7. Laboratory crack growth data are usually presented in this manner. Details of the calculational methods used to calculate Ky are provided later in this paper, but an important point 1 to note here is that K depends g on the crack depth, therefore it changes continuously during crack growth. I
-5 Crack growth analysis methods are, therefore, iterative in nature. Given an initial crack depth, the Ky is calculated for the particular stress distribution of interest. Knowing the yK , the amount of growth for a specific time is cal-culated, the growth is added to the initial crack depth, a new Ky is calculated, and the process is repeated. Time intervals selected can vary from 1 hour to 1000 hours, depending on the rate of growth and rate of change in Ky with crack depth.
Selection of Crack Growth Rate Parameters Although only two parameters, crack growth rate and K 3
, are used, they are both highly dependent on several factors. Crack growth rate is affected by the degree of sensitization of the material and by the severity of the environment.
Our interest as it relates to BWR piping is primarily in a degree of sensi-
! tization normally caused by welding, and in an environment similar to normal i
BWR water conditions.
) }
l Most formal crack growth studies are carried out with standard fracture mechanics specimens, which makes Ky determination easy. These specimens are not readily machined from pipe welds, so the material is given an artificial l sensitization treatment, intended either to simulate the effect of welding or, in some cases, the more severe effect of furnace sens ' ue9 0n. Tests to ascertain whether the intended degree of sensitization nas been obtained are still inexact, causing significant scatter in laboratory test results intended to apply to a similar metallurgical state. l l
m Tests to simulate the BWR environment are usually run at operating temperature in high purity water containing 0.2 ppm oxygen. This is generally accepted to be a representative condition, although higher oxygen levels could occur locally for short periods of time. Tests are also often run in water con-taining up to 8 ppm oxygen, usually to achieve accelerated comparisons of materials or conditions. In addition to these standardized tests for crack growth rate, results of actual pipe tests are available. Many hundreds of welds have been tested in General Electric's pipe test facility. These tests, although generally more relevant in terms of material condition and environment, are more difficult to evaluate. Ky is more difficult to calculate, and accurate crack growth rates are also more difficult to measure. Nevertheless, this body of data has been used to augment those data from the more standard laboratory tests, to select appropriate crack growth rates. Figure 1 (from NUREG/CR-3292, reference 1) shows much of the relevant labora-tory data in the conventional form, where measured rates are plotted against K.y This plot clearly shows the large scatter resulting from a wide variation in material condition and environment. This information, together with additional information from actual pipe tests, was used to select a crack growth curve that is appropriate for use in safety evaluations. Note that if the fastest crack growth rate shown in Figure 1 is used, cracks would be predicted to grow completely through pipe walls in a matter of days. Clearly this would not reasonably represent reality.
l The curve selected for use by the NRC staff is shown on Figure 2. Note that it is a curved line on the semilogarithmic chart. On log-log coordi-nates, it plots as a straight line. For conversion or in calculations, it is expressed as: da 2.181
= 3.590 x 10 8 . Ky inches per hour E
As can be seen, the crack growth rate is a very strong function of Ky . In laboratory tests, Ky is easily determined with good accuracy. This is not the case for real pipes and real pipe cracks. There are two major sources of uncertainty: knowledge of the actual crack size and shape, and the actual stress distribution in the area of the crack to be evaluated. The service distribution at a pipe weld is made up of the stress caused by the service loading and the residual stresses caused by the welding process. Of these, knowledge of the residual stress is the more uncertain. Nevertheless, a residual stress distribution through the pipe wall must be defined, if realistic crack growths are to be calculated. Although this is covered later in more detail, several comments are in order here. The residual stress distribution caused by welding is the major stress component causing IGSCC. Welding causes a high tensile residual stress on the inside surface of the pipe near the root of the weld where the material is sensitized.
~. ,
1 I l This residual stress level has been calculated and measured to be up to or above the yield strength of the material. It typically is four or five times as high as the service-induced stress. In fact, without this very high residual stress at the sensitized area, IGSCC would not be a problem in BWR piping. This fundamental observation is helpful; wherever this combination of stress and sensitization occurs, cracking occurs. In actual cases, if there are significant cracks, there must be significant tensile residual stresses, and this should be accounted for in the crack growth analysis. The method used by the staff is described below. Stress Intensity Factor Calculations There are several analytical methods available for calculating the stress [ intensityfactor(K)causedbystressdistributionsofthetypefoundat y BWR pipe welds. The method using influence functions is the one used by the staff and will be summarized here. Other methods, such as those described in the ASME Boiler and Pressure Vessel Code, Section XI, Appendix A, may also be used where appropriate. Stress Analysis The total stress state, including residual stress, pressure stress, and other stresses caused by normal operation must be known or assumed. Note that factors such as stress indices used for purposes of other stresses should not be used when calculating stress levels that apply to Ky calculations. l
Residual Stress The laboratory-measured throughwall axial residual stresses on pipe wall thickness > 1 inch are presented in Figure 3, taken from NUREG/CR-3292 (reference 1). The solid line in Figure 3 is the axial residual stress distribution used for the calculation of stress intensity factors for pipe sizes of 12" diameter and larger. The residual stress distribution is . handled by fitting the curve of residual stress distribution through the wall by an analytical expression. For this particular residual stress 1 distribution, the nondimensional expression given below is used. 4 E_ .1 Uj& oj J:0 , where og = 1.0 01 = -6.910 oz= 8.687 03 = -0.480 l 04 = -2.027 I
& = x/t og = stress magnitude at & = 0 (inner surface) ,
The above formula permits calculation of the residual stress value at any point (x) through the vessel wall thickness (t) as a function of the peak residual stress value at the inside diameter (ID), oj.
~
Thestressintensityfactorcausedbytheresidualstressfromwelding(KIR)* is calculated using influence functions taken from NUREG KR-3384, page A.19, Table (7),-(reference 2). The influence functions, ij, given in this Appendix are for a 360 circumferential crack in a cylinder with an R/t ratio of 10. In view of other analytical conservations and uncertainties (i.e., assumed crack geometry and initial depths), it is believed that they may be used for cylinders with R/t ratios of from 9 to 11 to obtain reasonable and conserva-tive estimates of crack growth versus time. For R/t ratios significantly different from 10, other influence functions or other analytical methods should be used. The specific formula used by the staff is: K 4 IR
=Jna I oj ai i j og V t j=o L where:
I og,...o4 and og are as above 2 3 io = 1.1220 + 0.3989 a + 1.5778 a + 0.6049 a 2 3
- i. = 0.6830 + 0.1150 a + 0.7556 a + 0.1667 a3 1 2 iz = 0.5260 + 0.1911 a - 0.1000 a + 0.5802 a3
- 2 i3 = 0.4450 + 0.0783 a + 0.0556 a + 0.3148 a3 2 i4 = 0.3880 + 0.1150 a - 0.1333 a + 0.3519 a a = a/t a = crack depth t = wall thickness l
u - ._ _g_ Membrane Stress The membrane stresses are assumed constant through the wall thickness, so a, =a p where o = membrane stress (o,) from pressure p The stress intensity factor for a 360 circumferential crack from pressure, Kyp, is calculated by 2 3 Kyp = 'PR 4 t V na (1.122+0.3989a+1.5778a +0.6049a) l ZE where a, t are as above i P = pressure R= radius to center of pipe wall The total stress intensity factor, KIT, is given by KTT = Kyp + KIR where l h Kyp and KIR are as above.
~- e Correlation with Service Experience Although the residual stress is assumed to be the same for all welds, the applied stresses, primary and secondary, vary from weld to weld; therefore, calculations must be performed for each weld evaluated. Figure 4 shows the results of K ycalculations for several pipe sizes using a nominal applied stress of 7500 psi. Note that at relatively shallow depths, the Ky is high; therefore, the crack growth rate will be relatively fast. However, the K y actually diminishes as the crack grows to about half way through the wall. This prediction is consistent with service experience; very few, if any, actual cracks of significant circumferential extent have been found deeper than about 50% of the wall thickness. References (1) Shack, W. J., et al., " Environmentally Assisted Cracking in Light Water Reactors: Annual Report, October - 1981 - September 1982, "NUREG/CR-3292, Washington,D.C.: U.S. Nuclear Regulatory Commission, October 1982. (2) Stevens, D. L. , et al. , " VISA - A Comp" uter Code for Predicting the Probability of Reactor Vessel Failure, NUREG/CR-3384, PNL-4774, AM, Washington, D.C.: U.S. Nuclear Regulatory Commission, September 1983.
e ..
-3 10 _
NRC CURVEl 9 - l in./yr /, 10-4 _. e -
, S -
/ .
~
-5
? 10 j e -
O -
,! Y T 0.2 ppe 0 ;2sensitized at 1150*F/2 h _
2 (EPR = 15 C/cm ) - GE
%r A 0.2 ppe 0 5 5'"55tized at 1150*F/2 h -
2 E (EPR=10C/cm)-GE 2 z- 0 0.2 ppm 0,; sensitized at iiS0*r/24 h h - GE N O 0.2 ppe 0 2; severely sensittred S 10-6 0 8 **= 02 5'a$5tir'd at '250'r/24 h O -
~
GE @ HITACHI ~ y _ , ,. GENED @ ANL o ECR0 M 8 ppe 0 ;2senstttred by welding;
<f g - 0.04 in./yr tr5 at ,32.r/24 h (sri)
@ 8 ppe 0 2; sensitized at (1292*F/10 min)'
2
- + (932'F/24 h) (EPR = 4 C/cm ) """
l 8 8 ppe 0 ; 2sensitized at (1292'F/10 min) 2
+(932'F/24h)(EPR=4C/cm)
I f = 0.1 Hz, R = 0.94 lQ'7 N 8 ppe 0 ; 2sensittred at (1292*F/10 min) l 2
+ (842*F/257 h) (EPR = 15 C/cm )
I @ 8 ppa 0 ;2 sensIttzed at (1292'F/10 min) r RET l +(842'F/257h)(EPR=15C/ cal AE f = 0.1 Hz, R = 0.94 l $ 8 ppe 0 ; sensitized at 1292*F/14 h DATA 2 2 (EPR = 20 C/cm ) f = 0.008 Hz, l R = 0.95 i X 8 ppe 0 ;2 sensitized at 1292'F/is n 2 (EPR = 20 C/cm ) f = 0.08 Hz, l R = 0.95 s 8 1 1 0 10 20 .30 40 50 60 70 STRESS INTENSITY,K (ksi/in.) l Figure 1 i
- 104-da vs K for intergranular Stress Corrosion Cracking 1
359 x 104 K2.tst 1 i r
~
8
- =
)
45 104 - 3 x 104 ~l 1 I ~~ 1~~ ~1 1- y j g i j
~ 10 15 20 25 '30 40 '50 60 70 80 90100 i
Ki (ksi VTni j i 4 Figure 2 l i I . i! .. _ .__ _ _ . _ . . . _ . . _ _ _ . _ _ . _ . . _ _ . . . . . . _ _ . . . . . . . _ . . _ , . . - . _ . _ _ _ _ _ _ _ _ _ _ _ _ _ . _
INSIDE WALL OWSIDE WE 50 I I l l l l l l l GE 26 40 - o GE 26 (4 ozimuths) o ANL 26 (2 ozimuths) N { o ANL 26 (IN-SERVICE FROM KRB)
- ANL 20 20 -
o _
} 10 -
y oo g. M oo w 0 -.-- _ o. _o_o_.,g_ _ , _e_O Y 8 3_ o_ _E o *eo *
- 10 - o 8g ,
q .
-20 -
. o o
b; -
-30 -
a.a , - I I
-40 I I I I I I l 0 0.2 0.4 ~ 0.6 0.8 1.0 alt Fig. 3 Through-wall Distribution of Axial Residual Stress in Large-Diameter Pipes (t > 1 in.).
l
\
\
I % i
~ -
- g -
4 i l l 30 - 28", t = 1.3" i A 24", t =1.1"
} . ] 20", t = o.90" Z
j 12",' t = 0.65" . 20 - m . l 1 10 i ! i i i -i 1 i i -t i s ._ . . j o.1 o.2 o.3 o.4 o.s - o.s o.7 o.s o.s Figure 4 ' l e 5 _.---,--n - - - - ,
---.--,--_..---.,-.-----.---------,---,,.---.-----+_,----.---.--~,..-,----,.,.,.--,I
u Conference on Structural Mechanics in Reactor Technology, 8/19-23/85, Brussels, Belgium D 1/9 TECHNICAL BASES FOR INTERIM LICENSING ACTIONS RELATED TO BWR PIPE CRACKING Warren S. Hazelton, William H. Koo, and Raymond W. Klecker U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Introduction During a hydrostatic test in March 1982, slight leakage was detected from two recirculation system pipe welds at the Nine Mile Point nuclear power plant. The ensuing investigation led to extensive examination of austenitic stainless steel piping for IGSCC in all operating BWR plants in the USA. Many cracks were discovered; some were significant, others were of relatively minor extent. Repair of the significantly cracked welds was usually accomplished by applying weld overlay reinforcement; whereas welds with minor cracking were evaluated for , limited continued service in accordance with IWB-3640 of Section XI of the ASME Boiler and Pressure Vessel Code. Although IWB-3640 is relatively simple to apply, it required that an analysis for further crack growth be performed. As there was no l generally accepted procedure for such IGSCC crack growth calculations, the NRC staff developed a set of bases and calculational methods that i would be acceptable to the staff. I
e _ o This paper describes the staff approach and provides the specific parameters used. Crack Growth Calculations Crack growth calculations are required to evaluate the continued structural 1
-integrity of a weld with known cracks, if it is desired to continue operation without repair or reinforcement. The rate of growth of IGSCC is not easy to predict, because the several important factors are usually imperfectly known.
Research work in this area has been helpful in defining the general effect of these factors, but a large uncertainty in crack growth predictions still remain. Nevertheless, crack growth calculations can be performed within certain limits with enough confidence to ensure plant safety without excessive conservatism. l Crack growth calculations are based on the fundamental concept that the crack growth rate of a specific material in a specific environment will be a function i j of the applied stress intensity factor Ky. Laboratory crack growth data are I usually presented in this manner. Details of the calculational methods used to calculate Ky are provided later in this paper, but an important point to note here is that K depends y on the crack depth, therefore it changes continuously during crack growth. l l
o I 1 1 l Crack growth analysis methods are, therefore, iterative in nature. Given an initial crack depth, the Ky is calculated for the particular stress distribution of interest. Knowing the gK , the amount of growth for a specific time is cal-culated, the growth is added to the initial crack depth, a new Ky is calculated, and the process is repeated. Time intervals selected can vary from 1 hour to l 1000 hours, depending on the rate of growth and rate of change in Ky with crack
- depth, t
Selection of Crack Growth Rate Parameters Although only two parameters, crack growth rate and Kg ,: are used, they are both highly dependent on several factors. Crack growth rate is affected by the degree of sensitization of the material and by the severity of the environment. Our interest as it relates to BWR piping is primarily in a degree of sensi-tization normally caused by welding, and in an environment similar to normal BWR water conditions. Most formal crack growth studies are carried out with standard fracture - mechanics specimens, which makes Ky determination easy. These specimens are not readily machined from pipe welds, so the material is given an artificial l sensitization treatment, intended either to simulate the effect of welding or, in some cases, the more severe effect of furnace sensitization. Tests to 1 ascertain whether the intended degree of sensitization has been obtained are still inexact, causing significant scatter in laboratory test results intended l to apply to a similar metallurgical state. l !
- i. !
i 1 o
3 Tests to simulate the BWR environment are usually run at operating temperature in high purity water containing 0.2 ppm oxygen. This is generally accepted to be a representative condition, although higher oxygen levels could occur locally for short periods of time. Tests are also often run in water con-taining up to 8 ppm oxygen, usually to achieve accelerated comparisons of . materials or conditions. In addition to t'ese h standardized tests for crack growth rate, results of actual pipe tests are available. Many hundreds of welds have been tested in-i General Electric's pipe test facility. These tests, although generally more relevant in terms of material condition and environment, are more difficult to evaluate. Ky is more difficult to calculate, and accurate crack growth rates ! are also more difficult to measure. Nevertheless, this body of data has been used to augment those data from the more standard laboratory tests, to select L appropriate crack growth rates. Figure 1 (from NUREG/CR-3292, reference 1) shows much of the relevant labora-tory data in the conventional form, where measured rates are plotted against K.y This plot clearly shows the large scatter resulting from a wide variation in material condition and environment. This information, together with additional information from actual pipe tests, was used to select a crack growth curve that is appropriate for use in safety evaluations. Note that if the fastest crack growth rate shown in Figure 1 is used, cracks would be predicted to grow completely through pipe walls in a matter of days. Clearly this would not reasonably represent reality.
n The curve selected for use by the NRC staff is shown on Figure 2. Note that it is a curved line on the semilogarithmic chart. On log-log coordi-nates, it plots as a straight line. For conversion or in calculations, it is expressed as: da 2.182
= 3.590 x 10 8 . Ky inches per hour dt l
As can be seen, the crack growth rate is a very strong function of Ky . In laboratory tests, Ky is easily determined with good accuracy. This is not the case for real pipes and real pipe cracks. There are two major sources of uncertainty: knowled 0 e of the actual crack size and shape, and the actual stress distribution in the area of the crack to be evaluated. The service distribution at a pipe weld is made up of the stress caused by the service loading and the residual stresses caused by the welding process. Of these, knowledge of the residual stress is the more uncertain. Nevertheless, a residual stress distribution through the pipe wall must be defined, if realistic crack growths are to be calculated. Although this is covered later in more detail, several comments are in order here. The residual stress distribution caused by welding is the major stress component causing IGSCC. Welding causes a high tensile residual stress on the inside surface of the pipe near the root of the weld where the material is sensitized. I j
l This residual stress level has been calculated and measured to be up to or above the yield strength of the material. It typically is four or five times as high as the service-induced stress. In fact, without this very high residual
+
stress at the sensitized area, IGSCC would not be a problem in BWR piping. This fundamental observation is helpful; wherever this combination of stress and sensitization occurs, cracking occurs. In actual cases, if there are significant cracks, there must be significant tensile residual stresses, and
' this should be accounted for in the crack growth analysis. The method used by the staff is described below.
Stress Intensity Factor Calculations There are several analytical methods available for calculating the stress intensityfactor(K)causedbystressdistributionsofthetypefoundat y BWR pipe welds. The method using influence functions is the one used by the staff and will be summarized here. Other methods, such as those described j in the ASME Boiler and Pressure Vessel Code, Section XI, Appendix A, may also be used where appropriate. Stress Analysis The total stress state, including residual stress, pressure stress, and other stresses caused by normal operation must be known or assumed. Note that factors such as stress indices used for purposes of other stresses should not be used when calculating stress levels that apply to Ky calculations. i t
m i l Residual Stress The laboratory-measured throughwall axial residual stresses on pipe wall thickness > 1 inch are presented in Figure 3, taken from NUREG/CR-3292 (reference 1). The solid line in Figure 3 is the axial residual stress distribution used for the calculation of stress intensity factors for pipe sizes of 12" diameter and larger. The residual stress distribution is handled by fitting the curve of residual stress distribution through the wall by an analytical expression. For this particular residual stress distribution, the nondimensional expression given below is used. 4 O_ = .I Uj h "i 3"O ! where i o= g 1.0 01 = -6.910 a= z 8.687 7 03 = -0.480 a4 = -2.027
& = x/t aj = stressmagnitudeat&=0(innersurface)
The above formula permits calculation of the residual stress value at anypoint(x)throughthevesselwallthickness(t)asafunctionofthe peakresidualstressvalueattheinsidediameter(ID),aj.
m _ _ _ _ - The stress intensity factor caused by the residual stress from welding (KIR)> is calculated using influence functions taken from NUREG KR-3384, page A.19, Table (7),(reference 2). The influence functions, ij, given in this Appendix are for a 360 circumferential crack in a cylinder with an R/t ratio of 10. In view of other analytical conservations and uncertainties (i.e., assumed crack geometry and initial depths), it is believed that they may be used for cylinders with R/t ratios of from 9 to 11 to obtain reasonable and conserva-tive estimates of crack growth versus time. For R/t ratios significantly l [ different from 10, other influence functions or other analytical methods should be used. The specific formula used by the staff is: K 4 IR
=Jna I aja i i j ajdt j=o where:
og,...a4 and oj are as above 2 3 io = 1.1220 + 0.3989 a + 1.5778 a + 0.6049 a 3 2 i.1
= 0.6830 + 0.1150 a + 0.7556 a + 0.1667 a 2 3 k
iz = 0.5260 + 0.1911 a - 0.1000 a + 0.5802 a 3
-2 is = 0.4450 + 0.0783 a + 0.0556 a + 0.3148 a 3 2
i4 = 0.3880 + 0.1150 a - 0.1333 a + 0.3519 a
. a = a/t f a = crack depth t = wall thickness l
l
es_ . . Membrane Stress The membrane stresses are assumed constant through the wall thickness, so 4 m p where o = . membrane stress (o,) from pressure p The stress intensity factor for a 360 circumferential crack from pressure, Kyp, is calculated by 2 3 K yp = PR V t V na (1.122+0.3989a+1.5778a + 0.6049 a ) . 2I where a, t are as above P = pressure R= radius to center of pipe wall The total stress intensity factor, KIT, is given by - KTT = Kyp+KIR where i K yp and KIR are as above. 1 1
m Correlation with Service Experience Although the residual stress is assumed to be the same for all welds, the applied stresses, primary and secondary, vary from weld to weld; therefore, calculations must be performed for each weld evaluated. Figure 4 shows the results of K ycalculations for several pipe sizes using a nominal applied stress of 7500 psi. Note that at relatively shallow depths, the Ky is high; therefore, the crack growth rate will be relatively fast. However, the K y actually diminishes as the crack grows to about half way through the wall. This prediction is consistent with service experience; very few, if any, actual cracks of significant circumferential extent have been found deeper than about 50% of the wall thickness. References (1) Shack, W. J., et al., " Environmentally Assisted Cracking in Light Water Reactors: Annual Report, October - 1981 - September 1982, "NUREG/CR-3292, Washington, D.C.: U.S. Nuclear Regulatory Commission, ! October 1982. (2) Stevens, D. L., et al., " VISA - A Computer Code for Predicting the Probability of Reactor Vessel Failure," NUREG/CR-3384, PNL-4774, AM, Washington, D.C.: U.S. Nuclear Regulatory Commission, September 1983.
-3 10 -
NRC CURVE' S - 1in./yr /, 10"4 _.- e _
, G -
Alf
~
-5 j
, Ak -
% 10 _
- / V 0.2 ppe 0 ;2 sensftized at 1150'F/2 h (EPR = 15 C/cm ) - GE A 0.2 ppe 0 ;2sensitized at 1150*F/2 h -
T 2 (EPR = 10 C/cm ) - GE z- 0 0.2 Ppm 0 ;2 sensitized at 1150*F/24 h
~ -
H . GE 3 D 0.2 ppe 0 ;2 severely sensitized O 8 ppm 0 2; sensitized at 1150'F/24 h CD l0-6 - GE HITACHI _ g GENE 0 ANL o GECR0 hor 48 ppe 0 ; sensitized by welding;
<:t g - 0.04 inlyr 2
uS at 932 r/24 h (Su)
@ 8 ppe 0 ; sensitized at (1292'F/10 min)'
2 2
+(932*F/24h)(EPR=4C/cm) ~
8 8 ppe 0 ; 2sensitized at (1292*F/10 min) l 2
+(932'F/24h)(EPR=4C/cm) 4 I '- ' " ' " ' ' '
10 k 8 ppe 0 ; 2sensitized at (1292*F/10 min) l 2
+(842'F/257h)(EPR=15C/cm) l @ 8 ppe 0 ;2 sensitized at (1292*F/10 min)
' 2 RECENT l +(842'F/257h)(EPR=15C/cm) f = 0.1 Hz, R = 0.94 )ANL 9 8 ppe 0 ; 2sensitized at 1292*F/14 h DATA l 2 (EPR=20C/cm)f=0.006Hz, l R = 0.95 l X 8 ppe O g; sensitized at 1292'F/14 h 2
l (EPR = 20 C/cm ) f = 0.08 Hz, 8 R = 0.95 s 1 0 10 20 .30 40 50 60 70 STRESS INTENSITY,K (ksi/in'.) Figure 1
.x .. _
104 .- a vs K for intergranular Stress Corrosion Cracking 9
~
3.59 x 104 K 2.161 , 1 1 l
.c 8
- =
45 . 104 - i . l 3 x 104 ~l ' I I ~~ ~ l ~ 'l' l' I I I i 1
~ 10 15 20 25 '30 40 50 60 0 ~80 90100 K (ksi . 4 -
Figure 2 1 l l l e
I INSIDE WALL OUTSIDE WALL 50 y ; ; ; ; ; ; ; ; oGE26 90 % o GE26 (4 orimuths) t F a ANL 26(2 ozimuths) M o - B o ANL 26 (IN-SERVICE FROM KR8)
- ANL 20 20 -
a g - 3 10 - C 0 8 8*
)
t M o o O 0 0 -s-- " o *-es-o fg i 5 8 5---{- o o
.{ o e
* -10 - o o E '
oo @ . o
-20 -
g, -
)
-30 -
64 e -
- e
~ I I I I I i l l I 0 0.2 0.4 0.6 0.8 1.0 o/t Fig. 3 Through-wall Distribution of Axial Residual Stress in Large-Diameter Pipes (t 3,1 in.).
i i 9
-.-------e ' ' ' ' ' ~
j g - l l 30 - 28", t = 1.3" 24", t =1.1"
~
~
6 20", t = 0.90"
- I f
l i i 12", t =0.65" . i 20 - l k-I . f l . 1 10 1 1 1 1 -1 3 3 ....y ~" ' ' ' 0.1 0.2 0.3 '0.4"~0.5~ ~ 0.6 ~ 0.7 0.8 09 f
~
att6 Figure 4 t
-- ,- ,.----m..
e-- ,-s--- - wc--.-- _ - , - - . - - , - , _ . _ - , , --,vw,---,.w. -,-.,~,,,-,y.,--,----w.,-----,,we.wemm.mrw--&r
-}}