ML20134B918
| ML20134B918 | |
| Person / Time | |
|---|---|
| Issue date: | 09/11/1996 |
| From: | Mayfield M NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Strosnider J NRC (Affiliation Not Assigned) |
| Shared Package | |
| ML20134B922 | List: |
| References | |
| NUDOCS 9609200114 | |
| Download: ML20134B918 (79) | |
Text
{{#Wiki_filter:-- _ _ . ._ September 11, 1996 MEMODANDUM TO: Jack R. Strosnider, Chief ' Materials and Chemical Engineering Branch Division of Engineering, NRR l
- FROM
- Michael E. Mayfield, Chief ( 8Mdw- /o Electrical, Materials and echa ical Eng'neering Branch !
Division of Engineering Technology, RES f 1
SUBJECT:
STEAM GENERATOR TUBE FLAW DISTRIBUTIONS As promised in J. Muscara's e-mail message to you of S gtember 5, 1996, attached is a summary report on " Flaw Distribution for Example Plants." The ! report was prepared by Dominion Engineering Inc. (DEI) under a subcontract to the ANL steam generator tube integrity research program. The report is based on the extensive knowledge and data bases at DEI on steam generator tube degradation, inservice inspection (ISI) results from many plants over many operating cycles, and on DEI's extensive work on statistical analyses of steam
- generator tube degradation.
j The initial draft report and calculations from DEI conducted under this effort were reviewed by a panel of experts comprised of W. Shack (ANL), R. Kurtz (PNNL), J. begley (APTECH), C. Laire (LABORELEC, Belgium), and L. Cizelj (Jozef Stefan Institute, Slovenia). A meeting was held on August 26, 1996, where the DEI principal investigator presented results of the study to the panel members, and the members provided comment and input based on their review of the initial report and the presentation. Two key recommendations from the panel were that: a) the detected flaw distributions obtained from the ISI results should be adjusted to inclu'de the effect of inspection reliability, probability of flaw detection, and sizing accuracy, to 3rrive at estimated
" actual" flaw distributions; and b) the flaws produced by loose parts should be included. i The attached summary report provides examples of flaw distributions associated with the important degradation machanisms that would be experienced by a typical plant with Model 51 steam generators containing low temperature mill I annealed alloy 600 tubes with Wextex expansions. The recommendations of the panel members are included in the calculations for the flaw distributions in the summary report. These flaw distributions can be used by the NRR staff who are conducting evaluations of tube failure and rupture probabilities and of ,
risk in conjunction with relemaking activities underway for steam generators. ( 4 The full report will contait, detailed discussions of the methodology and assumptions used for estimating flaw dirtributions, the historical ISI and degradation data from different plants Aat is used to develop the
- distributions, and additional example haw distributions for other typical ,
l steam generator designs and flaw degradation mechanisms. We expect this full report to be completed by October 1, 1996. Attacnment: As stated Distribution w/o Attachment Granch Reading File T.Speis D.Morrison Signature R;e (LCS) B.Sheron (NRR) A.Thadani (NRR) R.C. Jones (NRR) S.Long (NRR) Tamed (NRR) E. Murphy (NRR) J.E.Donoghue (NRR) DOCUMENT NAME: G:\MEB\MUSCARA\FLWDiSTO.NRR To ,eoe6ve e copy of this document,6nsecate la the bes: *r* . Copy without attachment / enclosure *E* = Copy with attachment / enc 6osure
*N' = No copy g, / _/
WG EMEBg/p/ l E CtMEB d l DMTg /gj l l ) l 0.Mdeman V MMaited6d NAK4 J Musc'ere / DAR l/f;i 09/ y /96 ' 091 // /96 09/ // /96
//% M// ,M g OFFIC6AL LECORD COPY RES FHe Code: 1B-9
DOMINION ENGINEERING, INC. 9/11/96 Flaw Distributions for Example Plants I
- 1. Obiective The objective of this document is to provide distributions of flaws in steam l generator tubes of example plants for use in evaluations of the likelihood of Id age l or burst under severe accident and design basis accident conditions.
l l
- 2. Background L In response to ANL PO No. M3168, DEI is preparing a repon that describes l methods for estimating the distributions of flaws by size of various types in PWR l steam generator tubes. Appendices A and B of the report will provide much of the historical data regarding flaws that are needed to determine the flaw distributions for diffemnt plants. Appendix C of the repon will contain numerical examples for a l_ moderately affected, severely affected, and lightly affected plant.
This preliminary repon provides examples of flaw distributions associated with a number of important degradation mechanisms. It contains a summary of results from the full repon.
- 3. Contents and Organization of Information The flaw distribution information attached to this summary document is l organized in six pans, which are largely taken from Appendix C of the main repon for the project. Each part treats a separate degradation mechanism. The mechanisms 1 covered are t% following:
- 1. Circumferential SCC at the 'ITS (Mostly PWSCC)
- 2. Circumferential ODSCC at Dents at TSPs
- 3. Freespan Cracks
- 4. Sludge Pile IGA / SCC
- 5. Axial ODSCC at TSPs (non dented)
- 6. Flaws due to I.cose Pans For each flaw type, appropriate figures are assembled from Appendices A, B, and C of the full report and supplemented by additional figures and evaluations so as to present a coherent whole for that flaw type. The typical organization is as follows:
e Intercept time and Weibull slo;x: data, which are used to estimate numbers of cracks, are taken from Appendix A. e Monte Carlo evaluations to determine the numbers of tubes with flaws for i 1 l
i 1
- DOMINION ENGINEERING, INC.
- 9/11/96- ;
! r the three cases are taken from Appendix C. ! e Flaw distributlan data based on representative NDE and pulled tube data are taken from Appendix B. These data are adjusted as appmpriate to reflect measurement error and probability of detection (POD). e For cases where defects have only one dimension, the adjusted flaw distributions are applied to the numbers of tubes with defects to distribute the flaws by size " bins." In cases where the flaws have two dimensions (length and depth), this is not practical. In these cases analytical e.xpressions are given for the length and depth distributions.
- 4. Summary Table Some of the key results for the six flaw types are shown on Table 1.
4 i 2 l
.4 l l DOMINION ENGINEERING, INC. Table 1. Flaw Distributions for HypothetlCal Example Cases Plant Characteristics No. of Steam Generators: 3 1 No. of Tubes = 3*3388 = 10164 Tube Molenal LTMA 800 Exponeen Method Wexter Hot Lag Temperature, "F: 605 i BOC EFPY: 14 ! EOC EFPY: , 15.2 Moderately Severely Linhuv 4 Affected Affected Affected
- 5 801 flan? ElAEl
. 1. Circumferential SC0 at TTS (Mostly PWSCC) 4 Number of tubes with Cira SCC at TTS at 15.2 EFPY (Note 1) = 7.0 46.3 1.4 ! I Gemme detnbubon parerieters for crack arc length; ) arc length in degrees (bolo 2).* a= 2.84 2.84 2.84 I i- p= 28.1 28.1 28.1
*For macrocracks. Macrocracks coneet of series of 0.3
- thru-wall cracks asperated by 0.05"long ligaments l 2. CircumfWertial ODSCC at Dents at T8Ps l Number of tube: with cracks et15.2 EFPY Note 1) = 4.2 40.1 0.32 J Gamme distritx, tion parameters for crack arc length;
! are length in degrees (Note 2):' a- 34.4 34.4 34.4
- p= 3.23 3.23 3.23
*For indudual macrocracks. There are typically two near thru-wall demetncally opposed macrocracks per cracked locaten See Figures C-6 and 7 for distnbubon of combmed crack lengths.
l 3. Free Sean ODSCC Number of tubes with cracks at 15.2 EFPY (Note 1) = 5.2 80.2 1.7
- Gamme detnbubon parameters for length, in. (Note a= 0.17 0.17 0.17 l j p= 0.88 0.88 0.88
] Gemme detnbubon parameters for depth, % well (No c' = 17.0 17.0 17.0 l p= 3.80 3.80 3.80 l
!
- Crack length and depth distribubons are assumed to be independent l 4. IGA / SCC in Hot Lee Sludoe Pile
] Number of tubes with cracks at 15.2 EFPY (Note 1) = 30.8 80.2 18.9 Gemme detnbubon parameters for length, in. (Note a= 0.17 0.17 0.17
! D= 0.88 0 88 0.88 Gemme distnbuten parameters for dopin, % we8. gNo a= 17.0 17.0 17.0 0= 3.80 3 80 3.80
- Crack length and depth dWtribubons are assumed to be independent
- 5. Axial OOSCC at TSPs Number of tubes with ODSCC (0.85 voit level) at 15.2 EFPY (note 1 569.7 8024.6 131.1 Gamme detnbubon parameters for depth of 0.75"long cracks; depth in % of well: a= 0.770 0.770 0.770 0= 4.480 4.480 4 480 6 Flaws Dus to Loose Parts Number of tubes with flows at 15.2 EFPY = 0.7 0.7 0.7 Gamme detnbubon paramotors for length, in.:* a= 1.900 1.900 1.900 Q= 0.458 0.458 0.458 Gemme detnbubon parameters for depth, % wall.' a- 2.275 2.275 2.275
= 17.235 17.235 17.235
- Flow length and depth detnbubons are assumed to be independent.
Notes
- 1. Numbers of tubes are totals that reflect adjustment for detecton efficiencies.
- 2. Gamme detnbubons are for "actuar fhms, i e., after adjusenent for measurement error and POD.
9/10/96
SUMMARY
T.
September 6,1996
- 1. Circumferential SCC at TTS (Mostly PWSCC)
The distribution of times to detection of circumferential SCC at the TTS for Wextex units for a hot leg temperature of 605 F is shown on Figure A-3 (median time of 7.53 EFPY,16th percentile time of 4.56 EFPY and 84th percentile time of 10.72 EFPY). The Weibull slopes for these plants are listed on Figure A-4A, and a Weibull fit to the slopes is shown on Figure A-4B (fitted median of about 3). The distributions for intercept times and slopes were used in a Monte Carlo evaluation of the type described in Section 5 to determine a distribution of the numbers of newly detectable flaws developing during the operating cycle, i.e., between 14.0 and 15.2 EFPY. A typical run for the Monte Carlo evaluation is shown on Figure C-1. Results of five runs, representing 5000 trials, are shown on Figure C-2. Figure C-2 shows median,16th percentile and 84th percentile results for both fractions and numbers of tubes. These are taken as representing moderately affected, lightly affected and severely affected units irspectively. The results for the three cases are as follows: Case Delta F Number of Tubes with Detectable SE Flaws at 15.2 EFPY Moderate 0.045 5 Severe 0.32 33 Light 0.012 1 A typical distribution of measured crack arc lengths for TTS circumferential cracks in tubes with Wextex expansions is shown in Figures B-3A and B-3B. The gamma distribution parameters for that distribution are: at = 6.06, = 17.33. These cracks are characterized as being essentially through wall, but as being made up of segments about 0.3 inches in length separated by uncracked ligaments of 0.050 inches in length.! Based on a report with pulled tube to RPC r..easurement comparisons for two j cracks,2 RPC measurements for Wextex cracks appear to have a mean error of 6.0 (RPC overcalls the length), with a standard deviation of 5.7 Based on data in the same report, together with engineering judgment, the POD for these defects was approximated as being zero at 25 ,0.95 at 75 and higher angles, with a straight line variation between these two values. The flaw size distribution was adjusted to reflect the measurement error and POD using the method described by Heasier.3 The resulting estimated " actual" flaw size distribution is shown in Figures B-3C and B-3D. A " detection efficiency" was determined by setting it equal to the integral of the probability of detection times the actual distribution 1
September 6,1996 1 divided by the integral of the actual size distribution. The integrals were taken for all flaws greater than 20 The numbers of flaws expected at the end of the operating cycle,i.e., at 15.2 EFPY, were then distributed in size using the actual flaw size distribution of Figures B-3C and B-3D. The results of this distribution are shown in Figure C-3. Note that the total number of flaws was increased from the predicted number of detectable flaws using the efficiency factor determined above. As shown on Figure C-3, the predictions indicate that there are no large cracks, e.g., of 250 arc length or more, even for the severe case. In addition, as ! noted earlier, destructive examinations indicate that there will be ligaments present about every 0.3 inches (about 40 ) in the large cracks. 1 North Anna Unit 1 Steam Generator Ooeratine Cycle Evaluation, Westinghouse WCAP-13035, August 1991, forwarded by Virginia Electric and Power Company letter dated August 30,1991, in NRC PDR,9109100132 910880, Docket No. 50-338. 2 gt id. 3 P. G. Heasler, et al., Analysis Before Test: Estimation of Fabrication Defect Rates in Reactor Pressure Vessels. Draft PNL report for NRC, Nov.1994. 2
DOMINION ENGINEERING, INC. I Figure A-3. Industry Time to Detection of Circumferential TTS SCC (ID/OD)- ; Median Rank Analysis-Weibull Fh (Least Squares) Wextex Plants with 600 LTMA Tubing
.63 0 *50 d g , Slope b = 2.75 t i
u) e Theta = 8.60 EFPY
.20 Reference Temperature = 605*F h,
j .10 Median Fraction at Initial Detection = 0.00026 5 Y .05
- I G .02 F
t .01 2
$ .005 g !
y .002 16th Percentile = 4.56
}2 ,001 - -
EFPYs 50th Percentile = 7.53 j .000E EFPYs 84th Percentile = 10.72
.0002 EFPYs ,
.0001 , - ' - - ' ' ' ' ' ' ' ' ' ' ' '
I ' I'I I ' ' 1 10 100 Service Time (EFPY) i 9/4/96 Wextex.xis i l
DOMINION ENGINEERING, INC. i Figure A-4A. Weibull Slopes for Wextex Plants with Circumferential TTS SCC (before peening) , Median Slope = 2.82 - 16% Slope = 2.13 84% Slope = 4.16 , No. Plant Slope % Error Points Median Rank 1 A 2.00 9.4 4 0.130 2 B 2.78 32.6 6 0.315 3 C 2.82 6.6 3 0.500 ' 4 D 3.35 8.2 3 0.685 5 E 4.32 21.2 5 0.870 5 < Numberof Plants Updated: Jul-96 r i i i i i 9/(v% Weuex Slopes.XI.S
)
l i m Mq q 5 6 5 5 5 5 E o oo o o o 6 6 6 6 e o a
, _ . _ . . O n
o
+- i e I i
i e
+
-~ _ . . . ___ N r , l
- .t _
i i 1 i i i i, , i
- o ;_ -. e 9
7 ___._ I o: n . I i 4 ! Il 33 g i I
! Si m 5
.i - l l I f
! to i ! ! l 10 !N I i !' 5 l i l M i
' ! i i '
iG i l i , iM l l _ i G I , I : l l 1 lo i-l i, i : i
'I o
a9 l !.t
~
!iL ! ! 1 -, ._. - O '
!= ! l i I r l3 i i . t , , T r ,gl3 i
l l r i W l ! ,
! i j ! '
f I 1
. i 1 ; ,
4 g ! i i , ! T ' 4 N. 3
, 1 1
cn ;
+>\
> ' i
.*= l , l ,
i i i
.lk j 4 i i i e + i j
! i i i w i
- 1 m i 3
g-l j 4. i
~
i j ! ! oI i - 8' d E a
. N C _3 l 'N C o ,
y' -
* 's l .c a
3 J i N r ,
-~w C
w stueld ;o uo!pe;3 e ,_ O v-
'C **
g CD
'E $ "6 5 6 6 8 m
DOMINION ENGINEERING,INC. i I Figuce C-1. Weibull Distribution Monte Carlo - Wextex TTS Circumferential Flaws i. Distributions Times of Interest Time to Susceptible 14.0 EFPY i Slope b O.026 % tO% Fraction S 15.2 EFPY Failures Type III Weibull Type ill Weibull Type [1l I Slope 3.58 Slope 2.75 B 1 . Theta 3.39 Theta 8.604 10 Results error (1000 trials) error Men b 1 Min tO% 0.00 Min S 0.001 Max b 8 Max tO% 1751.6 Max B 1 Median 16%ile 84%ile Normalization 1.01 Normalization 1.00 Normalization Delta F Delta F Delta F ! Start Trunc O.0126 Start Trunc O.0000 Start Trunc O.000446 0.000114 0.00296 i~ l i E i i t 9/6/96 3ParBnds.xis Sheett i
DOMINION ENGINEERING, INC-Figure C-2. Wextex Delta F Runs Median 16%ile 84%ile Delta F Delta F Delta F 0.000483 0.000119 0.00399 0.000473 0.000112 0.00346 0.000446 0.000121 0.00292-0.000408 0.000115 0.00314 0.000420 0.000114 0.00259 mean 0.000446 0.000116 0.00322 delta N 4.5 1.2 32.7 o l i i i i 9/6/96 Wextex delta F.xis I
- = ,
.__m. . . _ _ _ _ . . . _ _ . _ _ . _ _ - . _ . . - _ _ . _ .
DOMINION ENGINEERING, INC. Figure B-3A. North Anna 1 1991 TTS Circumferenzial Crack Arc Lengths i 120 100 -- -- ---- - - - - - - - - - - - - - -- -- 80 -- - - - - --------- - - -- - -- - -
- ----= = = = ~ -
MNA1 1991 Data i Gamma Dist. 3 o N . . . _ c
$ 60 - - - - - -- - - -------- ----- - - -- ---
[ Alpha = 6.06
- Beta = 17.33 40 -- - -- -- - - - - - - - - - - -
i l 20 - - --- - - - - - -- - - - - l 0 ' 15 45 75 105 135 165 195 225 255 285 315 345 360 I Crack Arc Length, Degrees [ i i 9/6/96 TTSCirc.XLS ObsSizeDens
DOMINION ENGINEERING, INC. Figure B-3B. North Anna 1 1991 TTS Circumferential Crack Arc Lengths 1.0 . : : : 0.9 -- --- - -- -- - - - - - - - -
-- Alpha = 6.06 - -
+ l ,
0.8 - - - - - - -- - - - - - - - -- l 0.7 l O.6 --- - - - - - A- --- -- - - - - - - e
+ Observed O.5 ---
-----}-,---- '--- -- -- -- ---- - - - - - - - - - - - - - - -
Fit 2 ; labels O.4 - - - - - - - - - - O.3 - - - - - - - t 0.2 O.1
+
0.0 - O 60 120 180 240 300 360 Circumferential Extent (deg) TTSCirc.XL5 ObsSizeCum 9/6/96 ,
DOMINION ENGINEERING, INC. Figure B-3C. TTS Circumferential Crack Arc Lengths 0.0120 - - - - --- -- - - - - - - - - - - - - - - - I; O.0100 - - --- -
- j nq /
I g i i g
~~~_ Distribution Fitted to - _ _
0.0080
-- t f --- -~-g- - - - -
g g Observed Data ._
; i Alpha = 6.06 Estimated Actual
>- t Beta = 17.33 --- Fit to Observations g
g Mem39Pariance = t,33 l0.0060 g -
---} , -
g g --- - 42.7
~~~ ~ ~ ' ' ~
e g \ Lab 2 l l
- Lab 3 I g Estimated Actual Size ~ ~ ~ ~ ' ~
1 [ \
---A Distribution a=2 0.0040 ,I h _ p8.5 [
[ <
\ Mei$$thariance = ,
g s 48.1 i i 'g l l O.OO2O . -
-s Detection Efficiency -
I
\ 71.2 % ,
/ x -
/
/ 'N O.0000 l
O.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 Degrees i 9.'6/96 TTSCirc.XLS
Y DOMINION ENGINEERING,INC. 1 Figure B-3D. TTS Circumferential Crack Arc Lengths 1.00 ,
, I
- t O 90 - - - - - - - - - ' - - - - - - - ' ~ ~ ~ - ' ~ ' - - -
-,/ Observed Data
/ I
/ Alpha = 6.06 0.80 --
/ Beta = 17.33 f
/ Mean = 105.0 0.70 - - - - - - - -
--/ ~ ~
' ~~ Square Root Variance =
~' ~~~' - -
! 42.7 I
c - 3 0.60 E
- -/ --- -- -- --
/
Estimated Actual
--- Fit to Observed l
E / Estimated Actual Size i
! 0.50 - I- ~~ ~~'
Distribution
-- -- - ~ - ~ ~ ~
j ,/ ! Lab 2 i f Alpha = 2.84 Lab 3 E 0.40 -- - ~ ~ ~ - - 1 U Beta = 28.57
/ ,
I Mean = 81.1
# -Square Root Variance -
'J.30 f 48.1 i
0.20 ---- l- - -- - - - - - - - - - I Detection Efficiency = I
/
0.10 - --
/ -- -- - ~. - ---
71.2 % --
/
/
' / l 0.00 O.O 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 Angle (deg) 9/6/96 TTSCirc.XLS l t
4 DOMINION ENGINEERING,INC. ; Figure C-3, Binned Flaw Distributions - Circumferential Crack Arc Lengths at TTS
- Case: Moderate Severe Light I
Detected Flaws: 5.0 33 1 Detection Efficiency: 0.712 Total Flaws: '7.0 46.3 1.4 Alpha 2.84 Beta 28.57 i Arc Length ! Degrees F No. in Bin No. In Bin No. In Bin j 25 0.074 0.2 1.4 0.04 l
- 30 0.110 0.3 1.7 0.05 l
< 35 0.151 0.3 1.9 0.06 40 0.195 0.3 2.1 0.06 45 0.243 0.3 2.2 0.07 l 50 0.292 0.3 2.3 0.07 ! 55 0.341 0.3 2.3 0.07 l 60 0.389 0.3 2.3 0.07 65 0.437 0.3 2.2 0.07
]'
70 0.483 0.3 2.1 0.06 0.06 I 75 0.528 0.3 2.1 80 0.570 0.3 1.9 0.06 j 85 0.609 0.3 1.8 0.06 90 0.646 0.3 1.7 0.05 ; 95 0.681 0.2 1.6 0.05 l 100 0.712 0.2 1.5 0.04 105 0.742 0.2 1.4 0.04 . 110 0.769 0.2 1.2 0.04 115 0.793 0.2 1.1 0.03 120 0.815 0.2 1.0 0.03 . l 125 0.836 0.1 0.9 0.03 , 130 0.854 0.1 0.8 0.03 135 0.870 0.1 0.8 0.02 )' 140 0.885 0.1 0.7 0.02 145 0.898 0.1 0.6 0.02 l 150 0.910 0.1 0.5 0.02 i I
, 155 0.921 0.1 0.5 0.01 ;
160 0.930 0.1 0.4 0.01 l 165 0.939 0.1 0.4 0.01 i 170 0.946 0.1 0.3 0.01 175 0.953 0.0 0.3 0.01 180 0.958 0.0 0.3 0.01 185 0.964 0.0 0.2 0.01 1 190 0.968 0.0 0.2 0.01 195 0.972 0.0 0.2 0.01 200 0.976 0.0 0.2 0.00 205 0.979 0.0 0.1 0.00 210 0.981 0.0 0.1 0.00 215 0.984 0.0 0.1 0.00 220 0.986 0.0 0.1 0.00 9/6/96 Page1 Binned TIS Circ Flaws
DOMIN10N ENGINEERING,INC. Figure C-3. Binned Flaw Distributions - Circumferential Crack Arc Lengths at TTS 2 Case: Moderate Severe Light Detected Flaws: 5.0 33 1 Detection Efficiency: 0.712 Total Flaws: 7.0 46.3 1.4 Alpha 2.84 , Beta 28.57 } Arc Length Dagrees F No. in Bin No. in Bin No. in Bin 225 0.988 0.0 0.1 0.00 ! 230 0.989 0.0 0.1 0.00 235 0.991 0.0 0.1 0.00 240 0.992 0.0 0.1 0.00 245 0.993 0.0 0.0 0.00 250 0.994 0.0 0.0 0.00 255 0.995 0.0 0.0 0.00 ! i 260 0.995 0.0 0.0 0.00 I 265 0.906 0.0 0.0 0.00 270 0.997 0.0 0.0 0.00 275 0.997 0.0 0.0 0.00 2 280 0.997 0.0 0.0 0.00 285 0.998 0.0 0.0 0.00 290 0.998 0.0 0.0 0.00 295 0.998 0.0 0.0 0.00 300 0.999 0.0 0.0 0.00 305 0.999 0.0 0.0 0.00 310 0.999 0.0 0.0 0.00
, 315 0.999 0.0 0.0 0.00 320 0.999 0.0 0.0 0.00 ;
i 325 0.999 0.0 0.0 0.00 330 0.999 0.0 0.0 0.00
- 335 0.999 0.0 0.0 0.00 340 1.000 0.0 0.0 0.00 j 345 1.000 0.0 0.0 0.00 l l
350 1.000 0.0 0.0 0.00 355 1.000 0.0 0.0 0.00 360 1.000 0.0 0.0 0.00
- sum = 6.7 44.3 1.3 l
1 i l l 9/6/96 Page 2 Binned TIS Circ Flaws
- a
September 6,1996 .
- 2. Circumferential ODSCC at Dents at TSPs The amount of denting at TSPs of Model 51 steam generators with Wextex expansions varies from just a few tubes to 100% of the tubes being affected, and several plants have about 25% of their tubes affected. A denting rate of 10% is considered to be a moderate rate for this group of units, with 100% dented at severely affected units and about 1% at lightly affected units.
As discussed in Appendices A and B, the experience of North Anna 1 is used as a basis for this prediction, since the experience at North Anna I was the best characterized. The results for North Anna 1 are adjusted to reflect the lower temperature of the example plant (605 F versus 615 F). As indicated in Appendix A, the North Anna experience indicates that this type of degradation is expected to reach a level of 1% of the tubes with dents at about 7.6 EFPY at a hot leg temperature of 616 F. Adjusting this to 605 F using an activation energy of 54 Kcal/ mole results in this mode reaching 1% of dented tubes at the example plant at 12.2 EFPY. This is taken as the median intercept time. A distribution for the intercept time for this mode is not available. However, experience with other modes indicates that it is reasonable to assume that a standard deviation change occurs at about a factor of 1.5 in time, such that 16% of the plants are expected to experience this mode at 12.2/1.5 = 8.1 EFPY and 84% of the plaats by 1.5 x 12.2 = 18.3 EFPY. As discussed in Appendix A, a Weibull slope of 4 is considered to be a median slope for this mode. A distribution for slopes for this mode is not available. Based on variations in slopes for other modes, a standard deviation of about 2 is consid: red reasonable. This results in the 16% slope being 2 and the 84% slope being 6. In summary, the parameters controlling this mode are as follows: Pementile Susceptible Fraction EFPY to 1% of Slope (% Tubes with Dents) Dented Tubes 16 % 1 8.1 2 50 % 10 12.2 4 84 % 100 18.3 6 The above variables were modeled using log normal distributions for intemept times and susceptible fractions and a normal distribution for slopes, and then these distributions were used in a Monte Carlo evaluation of the type described in Section 5 to determine a i distribution of the numbers of newly detectable flaws developing during the operating cycle, i.e., between 14.0 and 15.2 EFPY, A typical run for the Monte Carlo evaluation is l shown on Figure C-4. Results of five mne representing 5000 trials, are shown on Figure 1
5eptember 6,1996 C-5. Figure C-5 shows median,16th pcreentile and 84th percentile results for both fractions and numbers of tubes. These are taken as representing moderately affected, lightly affected and severely affected units respectively. The msults for the three cases are as follows: Cass Delta F. % Number of Tubes with Detectable (of whole tube bundle) Flaws at 15.2 EFPY Moderate (50%) 0.035 4 Severe (84%) 0.386 39 Light (16%) 0.003 0.3 The distribution of measured crack are lengths for tids mode of cracking found at North Anna 1 is shown on Figures B-14A and B-14B. The gamma distribution parameters for that distribution are: a = 16.71, = 6.02. These cracks are characterized as being essentially through wall, with most ligaments having been corroded away.1 , Based on a report with pulled tube to RPC measurement comparisons for three cracks,2 RPC measurements for ODSCC cracks at dented TSPs appear to have a mean error of-10.9 (RPC undercalls the length), with a standard deviation of 39 . Based on data in the same report, together with engineering judgment, the POD for these defects was approximated as being zero at 25 ,0.95 at 75 and higher angles, with a straight line i variation between these two values. The flaw size distribution was adjusted to reflect the
- measurement error and POD using the method described by Heasier.3 To accomplish this adjustment, the standard deviation for measurement error had to be reduced to 20 because the estimated 39 value led to an unrealistic actual crack size distribution. The resulting estimated " actual" flaw size distribution is shown in Figures B-14C and B-14D. A
" detection efficiency" was determined by setting it equal to the integral of the probability of detection times the actual distribution divided by the integral of the actual size distribution. j Tne integrals were taken for all flaws greater than 20 .
Based on pulled tube results from North Aana 1,it was assumed that two diametrically opposed cracks am present at the same elevation of each tube with a detected crack. It was further assumed that the sizes of the two cracks are independent, and that I each independent crack size is controlled by the distribution shown on Figures B-14C and B-14D. To develop a distribution for the combined cracks, the sizes of each of the two , cracks were determined using independent Monte Carlo sampling for each crack; these two I sizes were then added to determine a total crack length for the trial. This was done for i 2
September 6,1996 4 1000 trials. The results of the distribution of total crack lengths for the two independent cracks combined are shown in Figure C-6. The numbers of flaws of various total are lengths expected at the end of the operating cycle, i.e., at 15.2 EFPY, were then determined. This was done starting with the numbers of tubes with detected flaws determined earlier. The numbers of detected flaws were divided by the detection efficiency to detemiine the total number of flaws. These flaws were then distributed into size bins using the distribution shown on Figure C-6. As shown I on Figure C-7, the predictions indicate that some large cracks, i.e., of 270 are length or ; more, are predicted as being present for the severe case. Since destmetive examinations indicate that there few uncorroded ligaments in this morphology, this type of crack may l involve the potential of causing significant leaks under accident conditions. I North Anna Ur41 Steam G _rator Ooeratine Cycle Evaluation. We"inghouse WCAP-13035, August 1991. forwarded by Virginia Electric and Power Company letter dated Augua 30,1991, in NRC PDR,9109100132 910880, Docket No. 50-338. 2 lbid. 3 P. G. IIcasler, et al., Annivsis Befora Test: Estimation of Fabrication Defect Rates in Reactor Pressure Vessels. Draft PNL report for NRC. Nov.1994. l l l 3
DOMINION ENGINEERING. INC. t i a Figure C-4. Weibull Distribution Monte Carlo - Circumferential Cracks at TSP Dents j Distributions Times of interest Time to Susceptible 14.0 EFPY Slope b 1.000 % 11 % Fraction S 15.2 EFPY Failures j Type [1] Normal Type ill Log Normal Type 111 Log Normal 3 3 Mean b 4 00 t1% @ mean 12.2 S @ mean 0.1 { i I Std.Dev. b 2.00 exp(Std.Dev.) 1.5 exp(Std.Dev.) 10 Results Mean Ir"1 %) 2.50 Mean in(B) -2."30 (1000 trials)
.Dev. hdt1 % 0.41 S.Dev. in(S) 2.30 5 Min b 1 Min t1 % 3.61 Min B O.001 Max b 8 Max t1% 41.2 MaxB 1 Median 16%ite 84%1e Normalization 1.10 Norrnatization 1.00 Normalization 1.22 Delta F Delta F De'ta F !
i Start Trunc O.0668 Stert Trunc 0.0013 Start Trunc 0.0228 0.000367 0.000031 0.00343 i [ I i \ I I I i T I L 3ParBnds. mis Sheet 1 9/6/96 i
DOMINION ENGilJEERING, INC. Figure C-5. TSP Circ Delta F Runs Median 16%ile 84%ile Delta F Delta F Delta F 0.000419 0.000031 0.00436 ! 0.000320 0.000025 0.00384 , 1 0.000362 0.000027 0.00354 0.000361 0.000033 0.00405 0.000311 0.000027 0.00350 ; mean 0.000354 0.000029 0.00386 ; delta N 3.6 0.3 39.2 I 1 l l i l I 9/6/96 TSP Circ delta F.xis
t DOMINION ENGINEERING, INc. t Figure B-14A. North Anna 1 1991 Dented TSP Circumferential Crack Arc Lengths 60 t 50 - - - - - - - --- - - -- - - - - - - 40 - - - - - - - - - - - - - - - - - - - - - - MNA1 1991 Data Gamma Dist. E
$ 30 - - - - - -- - - - - -- -
[ Alpha = 16.71 Beta = 6.02 i 20 - - - - - - - - - - - 10 - - - -
- - - - ~ ~ ~ ~ ~ - ~ ~
O 15 45 75 105 I" 135 165 195 225 255 285 315 345 360 i Crack Arc Length, Degrees 9/6/% TSPCirc.XI.S ObsSizeDens
DOMINION ENGINEERING, INC. Figure B-14B. North Anna 1 1991 Dented TSP Circumferential Crack Arc Lengths - 1.0 , : : : : :
+
- O.9 - -- - ----
- Alpha = 16.71 - - - --
' ?
O.8 7 0.7 -- - -- - - - - - - - --- - - -
---t--
0.0 - - - - - - - - - - -- - - - - -- c
- Observed O.5 -- - - --- - - - - - - --
Fit E
- labels
' ~ ~
O.4 - - - - - - - - -
+
0.3 0.2 - - - - - - - - - - - - 6 0.1 0.0 : O 60 120 180 240 300 360 Circumferential Extent (deg) 9/6/9; TSPCirc.XLS ObsSizeCum
DOMii4!ON ENGINEERING, INC. ! Figure B-14C. Dented TSP Circumferential Crack Arc Lengths 0.0250 ----- - - --- -- O.0200 - -- - - - - Distribution Fitted to , 3 Observed Data
\ Alpha = 16.71 - - - -
0.0150 ------4 -- f 1 -- - --" Beta = 6.02-- Estimated Actual - f
--- Fit to Observations
[ [ \ kiihe RdNariance = 24.6 L ebl l 3 g l 1 g l l Lab 2 E i Estimated Actual Size Lab 3 0.0100 - %-- Distribution -- - - - - -- - - - - - -- - - - - -- i 'g Alpha = 34.40
; g Beta = 3.23 '
' l
[ g k'Oe"Rddt Yariance = i f g 18.9 0.0050 l- L - f- --f---- ----- ' I g Detection 5fficiency = f
\ 94.8%
\
I J
/
O.0000 - - - - O.O 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 .
+
Degrees TSPCirc.XLS 9/6/96
DOMINION ENGINEERING,INC. Figure E-14D. Dented TSP Circumferential Crack Arc Lengths r 1.00 , , 0.90 - - - - - - - ----- - - r-- - - --- -- - - - ," -- - ---
/ I 0.80 -
[ Beta = 6.02 , I 0.70 anance =
-- Square Roo y .
# 24.6 c I - _ ,
Estimated Actual 3 0.60
--- t --- - -
o I
'11 --- Fit to Observed 2 s
- Estimated Actual Size g Lab 1 j 0.50
~ ~ - - - -
Distribution I Lab 2 i e 1 l Alpha = 34.40 Lab 3
}s O.40 - - -- - -
i l -- 1 Beta = 3.23 o I Mean = 111.1 [ g O.30 - - - - - - - j-Square Root Variance . i i 18.9 I ' O.20 -l- - - I I I Detection Efficiency = 0.10 -- f- 94.8 % -- - - - - - - -
/
/
/
o.co ; o.o 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.c Angle hieg) k 9/6/96 TSPCirc.XI.S 1 i
DOMINION ENGINEERING,INC. . Figure C-6. Total Arc Length of Two C.acks at Dent . 100 go . _ _ . _ _ _ _ _ . _ ._ . _ _ _ . _ _ _ _ _ _ . .. . _ _ _ _ _ . _. _ _ _ _ _ _ _ _ _ _ _ _ ._. ._ _ _ _ . 80 -]-------
- - - - - - - - -- 5 ^ - - - - - - - -- -
70 --- - --- - -- - - - - - - - - - Parameters for Individual Cmcks Alpha = 34.4 [
~'~~~~' ~~~ ~ ~---~ ~~
60 - - - - - - Beta = 3.23 ~ ~ ~ - -~ 1 3 50 - -
-- -f - - -
- - -- - - j - - - ----
E l e E 40 - - - - - - - - - - - - - - - - - --- --~-------- --
-----t----- - - - - - - -
4 30 - -- 20 10 - - ~ - - - - - - l 0 0 50 100 150 200 250 300 350 Arc Length Two Crack MC 9/6/96
DOMINION ENGINEERING,INC. Figure C-7. Binned Flaw Distributions Total Circumferential Crack Lengths at Dents (Two independent Cracks per TSP Location) Case: Moderate Severe Light Detected Flaws: 4.0 38 0.3 Detection Efficiency: 0.948 Total Flaws: 4.2 40.1 0.32 Arc Length Count in 4 Degrees 1100 trials F No. in Bin No. in Bin No. in Bin 25 0 0.000 0.0 0.0 0.00 30 0 0.000 0.0 0.0 0.00 35 0 0.000 0.0 0.0 0.00 40 0 0.000 0.0 0.0 0.00 45 0 0.000 0.0 0.0 0.00 50 0 0.000 0.0 0.0 0.00 55 0 0.000 0.0 0.0 0.00 60 0 0.000 0.0 0.0 0.00 65 0 0.000 0.0 0.0 0.00 70 0 0.000 0.0 0.0 0.00 75 0 0.000 0.0 0.0 0.00 80 0 0.000 0.0 0.0 0.00 85 0 0.000 0.0 0.0 0.00 90 0 0.000 0.0 0.0 0.00 95 1 0.000 0.0 0.0 0.00 100 2 0.000 0.0 0.0 0.00 i 105 3 0.000 0.0 0.0 0.00 110 3 0.000 0.0 0.0 0.00 j 115 4 0.000 0.0 0.0 0.00 120 9 0.000 0.0 0.0 0.00 125 17 0.000 0.0 0.0 0.00 130 28 0.000 0.0 0.0 0.00 135 42 0.000 0.0 0.0 0.00 140 64 0.000 0.0 0.0 0.00 145 86 0.000 0.0 0.0 0.00 150 109 0.000 0.0 0.0 0.00 . 155 142 0.000 0.0 0.0 0.00 160 173 0.000 0.0 0.0 0.00 165 213 0.000 0.0 0.0 0.00 170 266 0.000 0.0 0.0 0.00 175 324 0.000 0.0 0.0 0.00 180 374 0.000 0.0 0.0 0.00 185 452 0.053 0.2 2.1 0.02 190 499 0.099 0.2 1.8 0.01 195 568 0.148 0.2 2.0 0.02 200 626 0.200 0.2 2.1 0.02 205 677 0.283 0.3 3.3 0.03 210 736 0.369 0.4 3.5 0.03 215 796 0.440 0.3 2.9 0.02 220 842 0.519 0.3 3.2 0.02 225 873 0579 0.3 2.4 0.02 Binned TSP Cire Flaws 9/6/%
.. _- . . . . . - = _ _ _ -. . . .. . . __ . . _ .
4 DOMINION ENGINEERING,INC. I i i Figure C-7. Binned Flaw Distributions Total Circumferential Crack Lengths at Dents l (Two independent Cracks per TSP Location) l l Case: Moderate Severe Light Detected Flaws: 4.0 38 0.3 Detection Efficiency: 0.948 Total Flaws: 4.2 40.1 0.32 Arc Length Count in j Degrees 1100 trials F No. in Bin No, in Bin No in Bin 230 913 0.638 0.2 2.3 0.02 235 954 0.704 0.3 2.7 0.02 240 971 0.762 0.2 2.3 0.02 245 997 0.820 0.2 2.3 0.02 250 1014 0.874 0.2 2.2 0.02 255 1033 0.914 0.2 1.6 0.01 260 1047 0.938 0.1 1.0 0.01 265 1063 0.953 0.1 0.6 0.00 270 1069 0.959 0.0 0.2 0.00 275 1077 0.965 0.0 0.2 0.00 280 1082 0.971 0.0 0.2 0.00 285 1086 0.977 0.0 0.2 0.00 290 1087 0.983 0.0 0.2 0.00 295 1090 0.989 0.0 0.2 0.00 300 1092 0.995 0.0 0.2 0.00 305 1093 1.000 0.0 0.2 0.00 310 1094 1.000 0.0 0.0 0.00 315 1094 1.000 0.0 0.0 0.00 320 1096 1.000 0.0 0.0 0.00 325 1096 1.000 0.0 0.0 0.00 l 330 1097 1.000 0.0 0.0 0.00 335 1097 1.000 0.0 0.0 0.00 340 1097 1.000 0.0 0.0 0.00 345 1097 1.000 0.0 0.0 0.00 l 350 1097 1.000 0.0 0.0 0.00 I 355 1099 1.000 0.0 0.0 0.00 j 2 360 1100 1.000 0.0 0.0 0.00 sum = 4.2 40.1 0.3 4 9/6/% Dinned TSP Circ Flaws
September 6,1996
- 3. Freesnan Cracks The distribution of times to detection of free span SCC for Model 51 steam generators for a hot leg temperature of 605 F is shown on Figure A-17 (median time of 12.8 EFPY, 16th percentile time of 5.80 EFPY and 84th percentile time of 22.3 EFPY). The Weibull 4
slopes for preheater plants are used since slopes are not available for feedring units; the preheater slopes are listed on Figure A-16A, and a Weibull fit to the slopes is shown on Figure A-16B (fitted median of about 3.1). The distributions for intercept times and slopes were used in a Monte Carlo evaluation of the type described in Section 5 to determine a distribution of the numbers of newly detectable flaws that develop during the opemting cycle, i.e., between 14.0 ard 15.2 EFPY. A typical run for the Monte Carlo evaluation is shown on Figure C-8. Results of five runs, representing 5000 trials, are shown on Figure C-9. Figure C-9 shows median,16th percentile and 84th percentile results for both fractions and numbers of tubes. These are taken as representing moderately affected, lightly affected and severely affected units respectively. The results for the three cases are as follows:
.Q s Delta F Number of Tubes with Detectable fa Flaws at 15.2 EFPY Moderate 0.032 3 Severe 0.34 35 Light 0.007 1 As discussed in Appendix B, freespan cracking at Palo Verde 2 has been the best characterized, and is taken as being typical for the industry for cases where detailed inspections for freespan cracks are being performed. Consistent with the Palo Verde results, crack length and depth distributions are taken as being independent. A distribution of measured crack lengths for freespan cracks in Paio Verde 2 is shown in Figures B-10A and B-10B. The gamma distribution parameters for the freespan crack length distribution are: a = 1.85, = 0.48. A Palo Verde submittal to the NRC contains data relating RPC measured length to the structural length determined by destructive examination.1 Based on evaluation of these data, the mean error for the length measurements was zero, and the standard deviation for length measurements was 0.5 inches. The POD as a function of length was approximated as being zero at 0.1 inches and 0.95 at 0.3 inches and longer, with a straight line variation between these two values. The flaw size distribution was adjusted to reflect measurement error and POD using the method described by Heasier.2 The resulting estimated " actual" flaw size distribution is shown in Figures B-10C and B-1
September 6,1996 10D. A " detection efficiency" of 0.66 was determined by setting it equal to the integral cf the probability of detection times the actual distribution divided by the integral of the actual size distribution. The integrals were taken for flaws in the range of 0.1 and 4.1 inches in length. I 2 A distribution of estimated crack depths for freespan cracks in Palo cerde 2 is shown in Figures B-11 A and B-11B. The gamma distribution parameters for the freespan crack ) depth distribution are: a = 11.04, p = 3.99. Based on analysis of data scatter presented in ' a Palo Verde submittal to the NRC,3 he t mean error was estimated at zero and the standard j deviation as 14% of wall. The POD was taken from a curve in the same submittal except that the maximum POD was set at 0.95 (reached at 68% of wall) (Figure B-1IF). The flaw size distribution was adjusted to reflect measurement error and POD using the method described by Heasier.4 To accomplish this adjustment, the standard deviation for measurement error had to be reduced to 12% of wall because the estimated 14% of wall value led to an unrealistic actual crack size distribution. The resulting estimated " actual" 4 flaw size distribution is shown in Figurer, B-11C and B-1ID. Note that the depth distribution indicates that there is essentially no chance of a defect being through wall. . Also note that a " detection efficiency" of 0.66 was determined by setting it equal to the integral of the probability of detection times the actual distribution divided by the integral of the actual size distribution. The integrals were taken for flaws in the range of 20 to 100% of wall. del performed a check of the probability of the above length and depth distributions causing a tube burst. This check was performed using Monte Carlo sampling and showed that there were no cases out of 10,000 trials where a defect had both a length over 1.8 inches and a depth over 90% of wall, i.e., no cases occurred where burst would be predicted under normal operating conditions. While this is considered reasonable for cases where extensive inspections for freespan defects are being performed, it is not considered realistic for the Wextex units being considered here, as discussed below. The flaw size depth distribution based on the Palo Verde data is for a unit where a rupture had occurred and very thorough and detailed inspections for free span cracking are being performed. This is not the case for the Wextex plants, where no ruptures due to freespan cracks have occurred and relatively limited freespan cracking has been detected by ECT. At these plants routine bobbin coil inspections are typically relied upon for scanning for freespan defects. With this level of inspection, there exists some likelihood of short through wall flaws being present, and there probably also is some chance of burst occurring, say one per every few thousand tubes with free span defects. To reflect these possibilities (i.e., possibilities of through wall defects and very occasional mptures) the 2
d September 6,1996 depth distribution was adjusted by trial and error to result in a about a 2% chance of a defect being through wall (most will be very shon and thus cause negligible leakage) and about a 0.1% chance of a defect causing a burst, where burst was considered as occurring if the flaw has both a length over 1.8 inches and a depth over 90% of wall (for 3/4" diameter tubes). The depth distribution providing these results, when combined with the l 3 length distribution of Figure B-10C and 10D, is shown on Figure B-1lE. Distributing the numbers of flaws expected at the end of the operating cycle into size bins is not practical since there are two independent size distributions. The user will have to determine the probability of having a defect with both a large length and a large depth , using Monte Carlo sampling or equivalent of the two distributions. Note that the total number of flaws should be increased from the numbers of detectable flaws determined above by dividing by the product of the detection efficiencies of the length and depth distributions, i.e., by dividing by 0.66 x 0.88 = 0.58 or multiplying by 1/.58 = 1.72. I Palo Verde Nuclear Generatine Station Unit 2 Steam Generator Evaluation. August 1995, in NRC PDR,9509120077 950907, Docket No. 50-529. 2 P. G. Heasier, et al., Analysis Before Test: Estimation of Fabrication Defect Rates in Reactor Pressure vessels. Draft PNL report for NRC, Nov.1994. 3 Palo Verde Nuclear Generatine Station Unit 2 Steam Generator Evaluation. August 1995,in NRC PDR,9509120077 950907, Docket No. 50-529. 4 P. G. Heasler, et al., Analysis Before Test: Estimation of Fabrication Defect Rates in Reactor Ptsnute Vessels. Draft PNL repon for NRC, Nov.1994. a 3
DOMINION ENGINEERING. INC. Figure: A-W. Industry Time to Detect Freespan IGA / SCC - Median Rank Analysis - All Domestic Westinghouse Design Model 51 Plants - Thot >= 600 deg F Weibull Fit (Least Squares) 0.9 - - L- -- -lz_:_-1-- _- i =l- L J-" - - - - -- - -- -
-- i - --- --i---- i Reference Temperature = 605.0 F 0.63 ~I~ ~
~~* ~' ~ ~~
Activetion Energy = 0 kcal/ mole 0.5 Median Fraction at First Detection = 0.0010 l i j i f . O 0.2 - I Slope b u 1.74 O
- a 0.1 - - - - -- t- --- ' --
--r--- -- -- --- - - - - - - - - - - + - - - - --
e [;' Theta = 15.79 EFPY j 8 0.05 !
$ i E l .
I - l-i - - E' O.02 1 g l y 0.01 m _!. 0.005 n.
'O
+- - - - --
j f
.[o 0.002 --
i 16th Percentile = S.80 M EFPYs E - - - - 0.001 -- -- -
! I i i 50th Percentile = 12.80 EFPYs ercen He = 22.3 0.0002 J-- -- - - -
L EFPYs
-, -i-- T r r-- m- r , T r r 1- -v i
-f r - l } l f, 0.0001 r r ,
10 100 1 Sennce Time (EFPY) W Feedring.xis 9W'M
DOMINION ENGINEERING,INC. Figure A-16A. Weibull Slopes for Freespan IGNSCC in Westinghouse Preheater LTMA Plants No. Plants = 7 Median Slope = 2.93 16% Slope = 2.33 84% Slope = 3.70 _ No. Plant Slope Median Comment b Rank 1 A 2.19 0.095 Cycles 8 through 10 for tubes repaired due to freespan IGNSCC 2 B 2.48 0.230 Cycles 8B through 10 for tubes repaired due to freespan IGNSCC 3 C 2.86 0.365 Cycle 8 for tubes repaired due to freespan IGNSCC 4 D 2.93 0.500 Cycles 6 through 10 for tubes repaired due to freespan defects 5 E 3.46 0.635 Cycles 6 through 8 for tubes repaired due to freespan defects 6 F 3.67 0.770 Cycles 58 and 6 for tubes repaired due to freespan IGNSCC 7 G 3.93 0.905 Cycles 5 and 6 for tubes repaired due to freespan IGNSCC Updated: Aug-96 W Preheater Slopes.xis 9/6M6 : i
i 4 ! i 4 O OO N O Q O 4 O Q O Q O Q O Q O 3 G) j o c. O
,- -O
+-
en u
' i l i - ~ - ~ ~ .--, _ . . _ _ . _ j
" l l i 0 { l
. h. ,
i l i ! e1 g Oi S. . li ' l T *"*
, to l i t
- H u o l
! .O I i
'(/) l i .
4 C ! , ; , m , i O. - 4 g + hw i l l i 1 5 l l t , 6 i 1 , . I iW i t
, ! i tW '
i
<W .
*E ! ! i i ! t e e a
,o O w 7 ; j i -
. ! 55
' O. _;
i i O eg.e .%
.t i l ! t
;lk i i i
' ( !
i i j {- ig _4 4 .Q I - J W + + ) lh ' l l i 5 l i i i i i l
. s! '
} i ,
- E 4 . i
=
.c .x+ i' I -
e i d _! Ng\ ' ' e w i
' i e s$~N 4
i > i l s 2e eA I a l , 3 , y' I *g h h,x
~
3 15
! NN i
. a@ !,
I \ ! i l
- l 1
8 t. , e , i l, , .
- L. i en ,
h r ! ( - e - i cn h s\ - C i l uJ ' e stueld 10 uonceJ3 .
.9 - O to j C v= O O Cn th l
'["* d d d d d iB 8 &
DOMIN!ON ENGINEERING,INC. . I i f Figure C-0. Weibull Distribution Monte Carlo - Free Span Cracks Distributions Times of heterest Time to Susceptible 14.0 EFPY ; Slope b O.103% tO% Fraction B 15.2 EFPY Failures Type 111 Wobull Type [1] Weibull Type [1] ' Slope 4.99 Slore 1.74 6 1 Theta 3.34 The a 15.79 10 Results i errt r (1000 trials) erre . Min b 1 Min to% 0.00 Min B 0.001 Max b 8 Max tO% 6850.1 Max B 1 Median 16%ile 84%ile i Normalization 1.00 Normalization 1.00 Normalization Delta F Delta F Delta F Start Trunc O.0024 Start Trunc O.0000 Start Trunc O.000342 0.000068 0.00409 - i i 3ParBnds.xis Sheet 1 9/6/96
i i DOMINION ENGINEERING, INC. Figure C-9. Freespan Delta F Runs Median 16%ile 84%ile Delta F Delta F Delta F 0.000323 0 000069 0.00357 0.000360 100065 0.00341 0.000325 J00072 0.00390 0.000277 0.000060' O.00256 0.000320 0.000065 0.00356 mean 0.000321 0.000066 0.00340 delta N 3.3 0.7 34.6 i j l 2 i I 9/6/96 FS delta F.xis
? l J DOMINION ENGINEERING, INC. Figure B-10A. Palo Verde 2 Free Span Defect Length 18.0 16.0 - - - - - - 14.0 - - - - - - - - - - - - - - - - - - - - - - - - - 1 12.0 ===~= M NA1 1991 Data ' Gamma Dist. ) y - - - o 10.0
- E o
AIPha = 1.85 8.0
' Beta = 0.48 i
6.0 - - - - -- - - - - l - i l 4.0 l 2.0 - -- - - - - -- l l j O.0 E O.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3 2.5 i Crack Length, Inches l l 9/6/% FSLngth.XLS ObsSiteDens l
DOMINION ENGINEERING. INC. i Figure B-10B. Palo Verde 2 Free Span Defect Length 1.0 i e - 0.9
! +
t
-L---- - - - - - - - - --
0.8
- - - - -- - - - - - ~ - - - - --- -- - -- -
+
* +
i
~
0.7 -- - - -- -l- - -
+ .
i O.6 -- --- - - - - - ----- i e
- O.5 - - - - - - - - - - ' - - - -' -- - $- - - -- --- - - -- - - -
Fit 2
* , labels
' '~ ~
O.4 --- --- - b * - - - - - -- --- - -- -- - - O.3 - - - - - - - - - - - I --- ---~- - - --- - -
+
0.2 -- - j t O.1 -
- - t -- -- - --
i 0.0 i O O.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 3 Crack Length, Inches 9/6/96 FSimgth.XLS ObsSizeCum ( t l
DOMINION ENGINEERING,1NC. 4 Figure B-10C. Free Span Defect Length 1.2000 ,-- - ~ T 7-
-L--- 4 -- ------ --
1.0000 + - - - - - -- - - - -
- 1- - - ---- - - - -
l Distribution Fitted to , Observed Data l Alpha = 1.85 ~~ ~~ ~~
-/ N - - - - - - - - - --Beta = 0.48 ~ ~ ~
0.8000 t I i I a Estimated Actual I \ Udi a"re"Rh6$ Variance = I \ 0.7 --- Fit to Observations
?e r s s j i
' -a I Lab 1 I O.6000 --
- d Actual Size Estimate g --,I \ Lab 2 2 g Distribution f Lab 3 1
,\
g Alpha = 0.17 i x Beta = 0.88 O.4000 - I - een Ro6k O
\s hquare. Variance =~ ~ ~ - ~ ~ ~ ~ ~ ~ ~ ~ ~
N O.4 N l O.2OOO -- - -- -~--~- --
\- - --- -
Detection Efficiency == _ , _ . N 66.0 %
's
~~~ ~- -- ---
0.0000 - - - - - -~---- - H t-- O.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Crack Length (in) 9/6/9' FSt.ngth.XI.S _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ ms_ . _ . _ _ _ _ . _ _ ____.____.____________.______________________.___________m__ _ _ _ _ . _ _ _ _ . _ _ . _ _
DOMINION ENGINEERING,INC. t Figure B-10D. Free Span Defect Length 1.00 , ____
, - " _ ,, , _ _ _l
** *
- I 0.90 ---
t----- Observed Data
/
e' Alpha = 1.85 1 i 0.80 - ----- --- --,'--- - - - - - - - - - -- - - - --
. Beta = 0.48
/ i
/ Mean = 0.9 0.70 - --- j ,/ - - - ---- ~ ~ ~ -
Square Root Variance = ~
/ O.7 c /
O.60 - - ----I- -- -- - - -- -- -- ---- - - - - - - -- Estimated Actual u I *
!! I --- Fit to Observed
- u. / Estimated Actual Size -
l-- Lab 1 j O.50
~
Distribution
--- - ~ ~ - ~
f l Lab 2 _lii /
/ Alpha = 0.17 Lab 3 g ----- -- ' ----
O.40 - - - ---/- - - + - - - ----- -- --
- s Beta = 0.88 o /
f a Mean == 0.1 0.30 - -- / - - - Square Root Variance = - - -- - - - - - - - - -
/ o,4 j i
I O.20 --- f - I - t
/ Detection Efficiency = !
O.10 - /- - --- - - - - - - ------ 66.0 % f
/
/ ,
0.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Crack Length in ! 9/6/s- FSLngth.XI.S i i _ _ _ _ . _ . . _ . . _ . _ _ _ _ . _ _ _ _ _. ._ ___m.____m. ____m_____ _ _ _ _ _ _ _ __ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ . _ . _ . _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ ___________. ____ _ _ _ _
DOMINION ENGINEERING, INC. l l i Figure B-11 A. PV 2 Free Span Defect Depths ; 4 f 70.00 , ; i 60.00 - --- - -- -- -- - 50.00 - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - MNA1 1991 Data (i Gamma Dist.
>3 40.00 - --- - - - - - - -
E i
- s i CT e Alpha = 11.04 i 2 30.00 --
- - - - - - - - - - - ~ - - - - - - - -
Beta = 3.99 i 20.00 - - - - - 10.00 i i O.00 ' - - E h - - o m m m e o m o o - s a m m m e m m o
- - m a m e e m o e o 6 o m m m o 4
Crack Depth, % of Wall 9/6/% I:SDepth.XLS ObsSireDens
. ._. .- w. - . .- ..- . . - . - . - . . - - - . . . . --- . . --. - . _ . . . _ . . - . . . . i I DOMINION ENGINEERING, INC. Figure B-11B. PV 2 Free Span Defect Depths 1.0 . I !
* )
O.9 -- - - - - --
+
+
0.8 -- --- - - - - - - - -- -.-- _ O.7 - - - -- - - - - - - - - - - - - - - -
+ - - - -
t 0.6 - - - - - - - - - - - - - - - - - - - I j c *
'ho,5 . _ _ _ _ _ _ _ _ . _ . . . __ _____}_._______.._ _ . _ _ _ . _ _ _ _ _ _ _
- Observed Fit a
sa. Alpha = 11.04 labels O.4 -- -- -- Beta = 3.99 -
+ -- --- --- - F 0.3 -- -
t L O.2 - - - - - - - -
+
0.1 --- - i
?
- O.O O.0 20.0 40.0 60.0 80.0 100.0 Crack Depth, % of Wall t
FSDepth.XLS ObsSizeCum 9/6/% _ . _ _ _ _ . . . _ . _ _ . . _ _ _ _ . . . ..______._._.____m.__ . _ .__ . . _ . _ _ _ _ _ _ _ . _ . _ . _ _ _ _ _ _ _ . _ _ - -
_ _ _ _ _ _ _ = . - _ - _ _ _ _ _ _ - _ _. DOMINION ENGINEERING, INC. Figure B-11C. Free Span Defect Depths 0.0600 - --- - - - - - -
---------q--- - - - - - -
3 i I 0.0500 ,- - --- - ----- - - - - - --
---j - - - ----f-----
l , j i Distribution Fitted to L-O.0400 - - - - - -- - - - - - - - --- - - - - - - 94
' - ~
Beta = 3.99 Estimated Actual g YJ3Aei!!!bariance = _ _ _ pi, to 03,,,y,, ion, ,
~
0.0300 - - - - + - - ,-
'N L --4---~ -
- - - - - - Lab 1 >
Lab 2 t Estimated Actual Size
# Lab 3 Distribution +
\ Alpha = 29.09 l .
0.0200 --
-/-
- %- -- Beta = 1.42_ - ; -- -. .._-- _
\
p g Ud80eihofkariance =
/ \ 7,7
/ \
l-
/ \ Detectionl Efficiency l
=
O.0100 ___ f f 66.1 % ,
,1 'Q i N
'~~-
O.0000 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 1 Depth (% TW) ; l i 9/6.96 FSDepth.XLS I (
DOMINION ENGINEERING,INC. Figure B-11D. Free Span Defect E epths 1 00 i i . l *
. A' _ I
+- - - - - - - - - -
O.90 ----------F------ J, Observed Data
/
/ t Alpha = 11.04 i O.80 - - - - -- --
- - -/ ,
-F-- -- - - --
j
' Beta = 3.99 f
Mean = 44.1 ~ - - - - O.70 - - - - - - - - -~ - - - - <-- f Square Root Varia,ce = ~
/ 13.3
- c ; '
h---- 5 - --- Estimated Actual
- -/- L - - -- -- t---
.$ O.60 t o #
j -I --- Fit to Observed E # Estimated Actual Size j O.50 - - - - - - - - -- d- f
#- - ~ - - - - -
Distribution
-~ '~ - - - _ . Lab 1 I 5 Lab 2
/
_e
/ Alpha = 29.09 ! Lab 3 !
g -R !
--- i O.40 -- - - --
o s Beta = 1.42 t i i Mean = 41.3 ; O.30 -- 1 --
/ Square Root Variance =-- - - - - -
/ 7.7
/ i O.20 -- -
/- --
f - - -- -- 1 - - I t r Detection Efficiency = 0.10 - -- - - -
- /!r$ -- - - - - - - - - - -
66.1 %
/
/ l 0.00 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 Depth (% TW)
FSDepth.XLS 9/6/96
DOMINION ENGINEERING. INC. Figure B-11E. Cumulative Depth Distribution for Free Span Cracks 1.00 , l l ! l
-L +
0.90
-- - - - -- - } -
0.80 - - - - -- - - - - - - - - -
----t--~------ - - - - - - - -
I I t-O.70 - - - - - - - - - - - - t Mean = 64.60 ! Alpha = 17.00 e
---------t----- - -- -} - Beta = 3.80 - ---
-S 0.60
$ Sq. root of var. = 15.67 l i
b- -- - I - - - -- - O.50
---t-------
5 3 .
$ O.40 - - - ---
U O.30 - - - - - -
-}- - --- - -
i I ,
- f-O.20 - -
- - - - - - -' - - - ~ -- - - - --
+
0.10 - t i O.00 ! 0 10 20 30 40 50 60 70 80 90 1 03 Depth, % of Wall ; r I Free Span Distr. 9/6/%
t DOMINION ENGINEERING,INC. Figure B-11F. Axial Free Span Defects - Probability of Detection 1.0 0.9 - 0.8 0.7 0.6 O.5 O.4 ' 0.3 0.2 I O.1 0.0 0 10 20 30 40 50 60 70 80 90 100 Defect Depth (%TW) , FSDepth.XLS , 9/6/96
September 6,1996
- 4. Sludge Pile IGA / SCC The distribution of times to 1% cracking of free span SCC for Westinghouse feedring steam generators for a hot leg temperature of 605 F is shown on Figure A-22 (median time of 14.7 EFPY,16th percentile time of 10.9 EFPY and 84th percentile time of 18.2 EFPY).
The Weibull slopes for sludge pile IGA / SCC are listed on Figure A-23A, and a Weibull fit to the slopes is shown on Figure A-23B (fitted median of about 3.4). The distributions for intercept times and slopes were used in a Mon.e Carlo evaluation of the type described in Section 5 to determine a distribution of the numbers of newly detectable flaws that develop during the operating cycle, i.e., between 14.0 and 15.2 EFPY. A typical run for the Monte Carlo evaluation is shown on Figure C-10. Results of five runs, representing 5000 trials, are shown on Figure C-11. Figure C-11 shows median,16th percentile and 84th percentile results for both fractions and numbers of tubes. These are taken as representing moderately affected, lightly affected and severely affected units respectively. The results for the three caws are as follows: fast Delta F Number of Tubes with Detectablg fc Flaws at 15.2 EFPY Moderate 0.22 23 Severe 0.89 90 Light 0.10 11 As discussed in Appenua B, distributions for sludge pile IGA / SCC are taken to be the same as for freespan SCC. Crack distributions for freespan cracking are covered in the previous section. Gamma distributions for flaw length are shown on Figures B-10C and B-10D, and for flaw depth are shown on Figure B-1lE. The detection efficiencies of 0.66 determined for the length distribution of freespan flaws and of 0.88 for the depth distribution of freespan flaws likewise apply. Distributing the numbers of fiaws expected at the end of the operating cycle into size bins is not practical since there are two independent size distributions. The user will have to determine the probability of having a defect with both a large length and a large depth using Monte Carlo sampling or equivalent of the two distributions. Note that the total number of flaws should be increased from the numbers of detectable flaws determined above by dividing by the product of the detection efficiencies of the length and depth distributions, i.e., by dividing by 0.66 x 0.88 = 0.58 or multiplying by 1/.58 = 1.72. I 1
DOMINION ENGINEERING,INC. ._. . - -- _. _ -- -- -_ Figure A-22. Industry Average Time to 1% HL SP IGA / SCC - Median Rank Analysis - West. Feedring Plants with FDE and no FDBs Weibull Plot
.90 -- -- -- -j -- - - - - - - h - -- -- I - - A - I 1 --- - - -
j j j
--- - --] - -
Reference Temp. = 605.0 F j j , l
. i i ; ;
- I i !
I l l
._e _ _ .. _ 4 ___ ; .l {0.6'
-+ 0.5 i 1 1 i ;
I i !
! l I Slope b = 4.57 i !
i' l i Theta = 15.93 EFPY I ! m ! ! ! c -
- - - 6 i' j 102 j i .E!
Q- ; I j E I 1 i l l t c i S l l I l
~ ~ " ~~ I- I-T ft !
- - ~
b.10
- ~ ~ - "- l 'l Activation Energy 54 kcal/ mole i
( i i ! 0.0: 16th Percer10.87 EFPY f i 50th Percen 14.70 EFPY ! I f t 84th Percent 18.18 EFPY f l0.0: i l i 1 i l lf 1
- - - - - - - - ,- 3 r - ,- r i r t ,-j
, j- l l :
.01 f 10 100 1
Time to 1% K. SP IGA / SCC (EDY) IIL SP MR t to 1%FDR No FDli 9/e/96
f b DOMINION ENGINEERING, INC. Figure A-23A. Weibull Slopes for HL SP IGA / SCC - LTMA 600 No. Plants - its Median Slope = 4.02 16% Slope = 1.07 84% Slope = 6.62 No. Plant Slope Median Comment b Rat k f t 1 A 0.41 0.038 2 B 0.89 0.092 l 3 C 1.02 0.147 4 D 1.21 0.201 5 E 1.33 0.255 6 F 1.74 0.310 7 G 2.43 0.364 8 H 3.27 0.418 L 9 1 3.87 0.473 ' 10 J 4.16 0.527 l 11 K 4.23 0.582 t 12 L 4.58 0.636 13 M 4.76 0.690 14 N , 4.81 0.745 15 O 6.30 0.799 16 P 6.72 0.853 i 17 O 7.61 0.908 18 R 14.60 0.962 Updated: Aug-96 i Fig A-23 SPSlopes 9/6/96
L ; ! > l
- ! : s a i ti 9 6 5 2 0 0 O. o, 0 0 f 0 0 0 0 0 0 O O 0 0 0 e p
- ,i jt! lj. ;!i' !!l!
0 o _ 1
+ !f !f !lI ls _
C
!' l,! i? i-C _
S _ !i i!l!il iii'
/ -
F;li A -
.\!f t
G I P - S
!!. ' ij . !!i!
_ ~ _
~
+I t lI _
4 t!I1l '
- ~ 1
_ 5 ' .
~ - ~
i'
~
- 0. ~
~
7 !I 4
- iil 2
0-1 t a _ s -
' ~
- = = e ko -
~
0~ 1 b ht S l 4
~
~
~
- ~
C ~ C - - S
/ .
~
_ A ~ G I
~
l e i -
~
P
~
e g _ ~
~
- ~
_ e d
~
~
- 0 p i
u F!l.I
~
~
_ ~ 1 lo S _ S I ' i _
# ._ ~ ~
o ~ t .
*/
' 4 1 ~ _
t - ' _
~ - - .
t - - - 1 I I! tI F
- O' -
~ - ~
l l 4 -
~
+
jI u ~ LI-i e il" _ g - 4 - 8
- - ~ .
E T , ~
/.
- - ~- '
Me '
~ i
/
. ~
~ _
~ .
E _ - ~ '
) _
s - e + - _ ~
/
r a .
~
~
u ._ - q . I S _ -
~ '
t _ s + a e _ - r c L ( n - I g n
, Fit l
lu
*p -
- i ir b e
e ie _ ,/ n W _
, i1 ig n
E n 2EE o gBEu. 1 o 1 0 0 0 6 in 1 0 0 0 9 i 9 1 / m 0 0 0 0 0 6 o / 9 D
. _ _ _ _ _ _ _ _ _ _ - . _ . _ _ . . _ _ . _ - . _ _ . _ .. . . _ . . - _ _ _ _ _ . ~ _ . ._ r DOMINION ENGINEERING,INC. L
.t Figure C-10. Weibull Distribution Monte Carlo - Sludge Pile IGA / SCC Distributions Temos of Interest ,
Time to Susceptible 14.0 EFPY I S! ope b 1.000 % t1 % Fraction a 15.2 EFPY j
~
Failures
! Type 111 Weibull Type III Weibull Type III Slope 1.27 Slope 4.57 a 1 Theta 4.51 The's 15.93 10 Results erro (1000 tnals) ;
erro , Min b 1 Min t . % 0.00 Min S 0.001 Max b 8 Max 11 % 18474.1 Max B 1 Median 16%ile 84%ile Normalization 1.36 Normahzation 1.00 Normalization Delta F Delta F Delta F Start Trunc O.1373 Start Trunc O.0000 Start Trunc O.002069 9.000984 0.00839 l f h e I 9/6/96 3ParBnds.xis Sheet 1 I
l DOMINION ENGINEERING, INC. l Figure C-11. Sludge Pile IGA / SCC Delta F Runs i l Median 16%ile 84%ile l Delta F Delta F Delta F l 0.002199 0.001024 0.01045 l l 0.002221 0.001079 ' O.00841 l 0.002306 0.001077 0.00800 l 0.002386 0.001023 0.00862 0.002060 0.001022 0.00886 l mean 0.002234 0.001045 0.00887 delta N 22.7 10.6 90.1 l l l l l
- i l
9/6/96 SP delta F.xis
i i September 6,1996 !
- 5. Axial.ODSCC at_TSPs (non dented) -
Evaluation of this mode needs to account for the fact that data for development of i
- i. these flaws is mainly for flaws that were repairable using the older 40% of wall l l l plugging criterion, while current practice is to repair in accordance with 2-volt alternate repair criteria (ARC). To accommodate this situation the numbers of tubes with detectable indications using the older 40% of a all criterion is first determined, and then ,
the number of tubes removed from service as a result of plugging will be estimated. The distribution of times to 1 % of tubes with detectable axial ODSCC at TSPs for Westinghouse feedring steam generators for a hot leg temperature of 605*F is shown on Figure A-20 (median time of 13.4 EFPY,16th percentile time of 7.49 EFPY and 84th percentile time of 20.2 EFPY). Per Appendix A, this value is for pluggable flaws using a 40% of wall criterion, with bobbin coil (BC) voltage of about 0.85 or higher. The Weibull slopes for these units are listed on Figure A-21 A, and a Weibull , fit to the slopes is shown on Figure A-21B (fitted median of about 4.5). The distributions for intercept times and slopes were used in a Monte Carlo evaluation of the type described in Section 5 to determine a distribution of the numbers of detectable flaws that have developed at 15.2 EFPY (sorae of these flaws will have been removed because of large voltage signals, as discussed later). A typical run for the Monte Carlo evaluation is shown on Figure C-12. Results of five nms, representing 5000 trials, are shown on Figure C-13. Figure C-13 shows median,16th percentile and 84th percentile results for both fractions and numbers of tubes. These are taken as representing moderately affected, lightly affected and severely affected units respectively. The results for the three cases are as follows: 1 l 1 1
l 1 September 6,1996 l l Case F Number of Tubes with Detectable _% Flaws at15.2.EFRY2 l Moderate 1.49 151 Severe 15.4 1568 Light 0.32 33
- Includes repaired tubes.
Over time, the voltages of tubes with ODSCC tend to increase at an average rate of about 0.2 volts per EFPY. A relatively high fraction of tubes with voltages over the 2 volt plugging criteria are usually confinned by RPC and thus require repair. After one or two more cycles, this fraction is expected to reach essentially 100%. To simplify the calculation for tubes repaired, it is therefore assumed that all tubes require repair when their voltage reaches 2.2 volts. With this assumption, tubes typically require repair at (2.2 - 0.85)/0.2 = 6.75 EFPY after reaching 0.85 volts. Using this value, tubes that developed detectable flaws at the 0.85 volt level at 14.0 - 6.75 = 7.25 EFPY or earlier are likely to have developed voltages high enough to result in plugging by 14.0 EFPY. Monte Carlo evaluations similar to those described above were performed to determine these values, with the results shown in Figures C-14 and 15. These values were then subtracted from those of the 15.2 EFPY case to determine the number of tubes with detectable flaws in service at 15.2 EFPY, as follows: l l 2 l
. . _ __ _.___~___ _. ._ _ _ _ _ _ _ _ _ _ _ _ __
l September 6,1996 l l Case Number.of. Tubes.with Number of Tubes with Number _ofln- ! Detectable. Flaws at DetectableElaws at Service _ Tubes _with { 15.2EFPY 7.25EEELand DetectableElawsat Plugged by 14.0EFPJ 15.2 EFPJ l Moderate 151 12 139 Severe 1568 98 1470 Light 33 1 32 As discussed in Appendix B, a crack size distribution for this mode of cracking at Farley 1 is taken as being typical. Figure B-13A shows a crack size distribution , developed fmm voltage data for a recent Farley inspection, for assumed 0.75 inch long cracks. A cermlation between field bobbin coil voltages and the depth of 0.75 inch long cracks was developed ucing burst pressure - voltage data for pulled tubes to estimate the burst pressure as a function of signal voltage, and then using the ANL failure pressure correlation to convert these burst pressures to the depths of 0.75 inch long cracks. Based on evaluation of these data, the mean error for the depth measurements was zero, and the standard deviatien was 15% of wall. The POD as a l function of depth was approximated as being zem at zero defect depth and increasing linearly until it became tangent with the POD curve for freespan defects discussed in l Section 3 (a POD of 0.71 at a depth of 40%); the resulting curve is shown in Figure B-l 13E. The reason for mnning the POD all the way to zero is to allow treatment of the numerous equivalent cmcks of shallow depths shown in Figure B-13A. The presence 1 2 3 i
1 1 September 6,1996 I of the shallow cracks is a result of assuming that all cracks are 0.75 inches long. In 1 actual fact, many small voltage signals are for shoner cracks that are deep enough to be detected. The Gaw size distribution was adjusted to reDect measurement error and l l POD using the method described by Heasier.' To accomplish this adjustment, the standard deviation for measurement error had to be reduced to 5 % of wall because the estimated 15% of wall value led to an unrealistic actual crack size distribution. The resulting estimated " actual" flaw size distribution is shown in Figures B-13C and B- l 1 13D. A " detection efficiency" of 0.244 was determined by setting it equal to the integral of the probability of detection times the actual distribution divided by the integral of the actual size distribution. The integrals were taken for Daws over 10% of wall in depth. The numbers of unrepaired flaws expected at the end of the operating cycle, i.e., at 15.2 EFPY, were then distributed in size using the actual Haw size distribution of Figures B-13C and B-13D. The results of this distribution are shown in Figure C-16. Note that the total number of flaws was increased from the predicted number of detectable Daws using the efficiency factor determined above. I P. G. Heasier, et al., Analysis Before TestLEstimation affabrication Defect Ratesin Reactor.PressureYessels, l Draft PNL report for NRC, Nov.1994. 4
l i DOMINION ENGINEERING, INC. l Figure A-20. Industry Time to 1% HL TSP IGA / SCC - Median Rank Analysis - All Westinghouse Design LTMA Drilled Hole Feedring Plants - No Prior Phosphate l l Weibull Fit (Least Squares) 0.9 - - - ---
-} _ ; . _{ 1n[_ _ j , - p-- L - - - - - -
~t [ 9 ,
Reference Temperatore = 605.0 F I i ff i 1 i
-l
+
0.63 0.5 . _ _ . _ _ _ _ Activation Energy = 54 kcal/ mole
/ ,
j i l l
! ll
- - * - -- + -
{ !I 0.2 O j Slope b = 2.37 l l , in
+ I Theta = 15.65 EFPY l 4
4' - R 0.1 !' 12 l O F 0.05 [ - i i :
# 0.02
-} -
-f -l i
f'
+ ! i m l If
^
I 4 N 0.01 i !! n
!{
$ 0.005 i o
8 [ I-l' 3 i j l t 0.002 y 16th Percentile = 7.49 EFPYs
!l
- t - --- -- -- -- -
p- j 0.001 . -- -
-- - + - } - f ' ' i l l l ! 50th Percentile = 13.41 ; j 0.0005! l ! EFPYs ; j i
ll ercentHe = 20.20 l 0.0002: EFPYs t l b ' k 4-, 0.0001r 1 10 100 Service Time (EFPY) N W Feedring.xk
DOMINION ENGINEERING, INC. ,. i Figure A-21 A. Weibull Slopes for HL TSP IGA / SCC in Westinghouse Feedring LTMA Plants Using Latest WC (Plants with greater than 1% tubes affected only) No. Plants = 9 Median Slope = 3.41 16% Slope = 2.64 84% Slope = 9.15 i No. Plant Slope Median Comment b Rank 1 A 2.34 0.074 Cycles 6 through 12 for all repairable defects (includes pre- and post- boric acid) 2 B 2.71 0.181 . Last 8 cycles for all repairable defects (includes pre- and post- boric acid) 3 C 3.24 0287 Cycles 13 and 14 for all BC indication . 4 D 3.26 0.394 Cycle 12 for BC indications v1voit 5 E 3.41 0.500 Cycles 6 through 12 for di repairable defects (includes pre- and post- boric acid) 6 F 4.22 0.606 Cycles 9 and 10 for BC indications v1 volt 7 G 4.31 0.713 Cycles 7 through 10 for all repairable defects using reanalyzed data assuming no IPC 8 H 8.75 0.819 Cycle 15 for BC tubes affected (Cycle 16 excluded) 9 l 10.81 0.926 Last 3 cycles for all repairable defects Updated: Aug-96 W Feedring Sbpe: .xis 9/(V96 4
- - ,- --r , . -e __._ _ _ _ , . . . _
* %% N # % @ @ # @ 3 o oo o o o o o o o o e
_ . . . _ . _ . . . . _ . . _ _ . _ . _ . _ . . . . _ _ _ _ _ _ _ _ _ _ . _ . . _ , . . . _ . _ _ _ . _ _ _ . . _ _ _ _ . _ . , o a 3 , c. m i
.._ _ . _ _ _ .__.__. _ . . . _ _ . . . , _ . _ _ _ . _ . _ . , . _ _ _ . _ . _ . _ _ . . . l l
l
- u. , u. u.
i w' w w E E e E.
= = = i
._7__
o ' W , O l Q . . _ _ _ . . . _ _ _ . . . _ . . _ _ _ _ . _ _ _ _ _ _ . . _ . . _ _ _ _ _ .._ _ - _. _ . O l
- 13. i U) ,
J H
, i !
t g e , i o y _ . - _o j m w O a k- "
\
, y _ __._ _1 . g , , - N i 3 F i \l i ,
.O . T-l g 1 >
g ._. r i N _ d y-
** \, ,
N e
\
\
( g g i i e u
~
3 '\'+ l C) g \ - a $ \ + ! M N I 3 \
.. _, . . _ _ a ua G
u C e I
. iI o>
C - c n
"_5 o .
o G
.5, o =
3__ -. -- w e stueld jo uo!pwj . O e
'E - O C
- 1. m e o O .
O)
-E 6 6 6 6 o @
8 e
DOMINION ENGINEERING,INC. i Figure C-12. Weibull Distribution Monte Carlo - Axial ODSCC at TSPs at 15.2 EFPY ' Distributions Times of Interest r Time to Susceptible 0.0 EFPY Slope b 1.000 % 11 % Fraction G 15.2 EFPY l Failures Type [1] Weibutt Type [1] Weibull Type 111 Slope 1.88 Slope 2.37 a 1 Theta 5.52 Theta 15.65 10 Results error (1000 trials) error Min b 1 Min t1 % 0.00 Men S 0.001 Max b 8 Max t1% 9084.3 Max B 1 Median 16%Ie 84%1e Normalization 1.21 Normalization 1.00 Normalization Delta F Delta F Delta F : Start Trunc O.0395 Start Trunc O.0000 Start Trunc O.015250 0.003091 0.12065 r l L i
)
9/6/96 3ParBnds.xis Sheet 1 , i i
_ _ _ . . _ _ ___ _ . . __ _ ._ . _ _ _ _ _ _ _ . _ _ = _ ._ DOMINION ENGINEERING, INC. l Figure C-13. HL TSP ODSCC 15.2 EFPY F Runs l Median 16%ile 84 %Ie l Delta F Delta F Delta F l 0.013685 0.003322 .0.13613 1 ! 0.014797 0.003092 0.16142 , l 0.014207 0.003123 0.13901 l l 0.015874 0.003207 0.15420 l 0.015896 0.003483 0.18080
- mean 0.014892 0.003245 0.15431 delta N 151.4 33.0 1568.4 l
l l i l l l
\
l l I { l l l 9/6/96 TSP 152dF.xis
DOMINION ENGINEERING,INC. i i Figure C-14. Weibull Distribution Monte Carlo - Axial ODSCC at TSPs at 7.25 EFPY i l Distributions Times of Interest f Time to Susceptible 0.0 EFPY , Slope b 1.000 % 11 % Fraction B 7.25 EFPY Failures ! Type [1] i Type [1] Weibull Type 11) Weibull Slope 1.88 Slope 2.37 8 1 Theta 5.52 Theta 15.65 10 Results error (1000 trials) error , Min b 1 Min t1% 0.00 Min B O.001 Max b 8 Max t1% 9084.3 Max B 1 Median 16%ile 84%ile Normalization 1.21 Normalization 1.00 Normalization Delta F Delta F Delta F Start Trunc O.0395 Start Trunc O.0000 Start Trunc O.0012^5 0.000099 0.00953 5 9/6/96 3ParBnds.xis Shect1
DOMINION ENGINEERING, INC. l l l Figure C-15. HL TSP ODSCC 7.25 EFPY F Runs l l Median 16%ile 84%ile l Delta F Delta F Delta F 0.001164 0.000073 0.00825 0.001237 0.000080 0.01085 0.001151 0.000084 0.00868 f 0.001126 0.000091 0.00876 , 0.001383 0.000075 0.01174 , ! 1 mean 0.001212 0.000081 0.00965 delta N 12.3 0.8 98.1 l l l I i l l l l 'i l l r I l l t + 9/6/96 , TSP 725dF.xis I r
I DOMINION ENGINEERING, INC. 3 i Figure B-13A. Farley 1 1995 - Estimated Crack Depth Due to Axial ODSCC at l TSPs 120 l i I 100 - - - - - - - - - 1 i j f 80 m -- --- i m NA' s991 Data ! I i Gamma Dist. y . h , L ! C
$ 60 Alpha = 1.61 i
[
"- Beta = 5.93 !
40 - - - - - - -- - - - - - - - - - - I i l ' 20 i
?
l g _ m E E - - _ _ _ o 4 - 9 C. 4 " 9 *
- N 9 h *. N 9 h 4 N 9 h 4 N !
o N e m s m - N e e. m m - m m e m o N m m o m !
- - - - - - m m n a m m m m m m m ,i Through Wall Depth (%) !
l 9/679- TSPAxial.XLS ObsSizeDens ;
DOMINION ENGINEERING, INC. Figure B-13B. Farley 1 1995 - Estimated Crack Depth Due to Axial ODSCC at TSPs 1.0 , l
+
0.9 -+l------ -- - - - - Alpha = 1.61 '
+ l
- - - - - - - 5- - -
O.8
- - - = - - -
]-
1 0.7 - - - - - - -- - -- - - - - ----- - 0.6 - - - - - C
- Observed
-{O.5 -- - - - - - - - - - - __. . _ .
Fit 2
' labels l 0.4
~ -- -
O.3 O.2 - - - - - - - - - - - I -- - - - + - - - - - - -
+ --
O.1 , O.O O.0 20.0 40.0 60.0 80.0 100.0 Through Walt Depth (%) 9/6/96 TSPAnial.XIS ObGiicCum _m - - . + - , -
DOMINION ENGINEERING,INC. Figure B-13C. Estimated Crack Depth Due to Axial ODSCC at TSPs 0.0500 -- l 0.0450 I A ~~' ~~ ~ '~~ ^^~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ l 1 1 .s. 0.0400 -g i
) Distribution Fitted to 1
0.0350 -- 1 - ~ ' ~ ~ ~ ~ ~ ~ ~ - Observed Data _ _ _ _._. Alpha = 1.61 f g Beta = 5.98 l O.0300 -- \ -- .._ g
+ -
--a Mean = 9.7 ;-
Square Root Variance = 7.6 i;/ \ ; --Estimated Actual ; 0.0250 - - - ' -~ ~ ~ ~ ~ ~ '
---- Fit to Observations 1
g g EstimatM Actual Size - ^ g g [ Distribution 0 0200 ~ ~ ~ ~ ~ Alpha = 0.77 g Beta = 4.48
\
_.__ .__ _ . _ _ Mean = 3.4 __. . _ . _ _ _ _ . _ _ _
\ l g 1-0.0100
\ Detection Efficiency = 24.4%
\ >10% TW
\
0.0050
' ' ~ ~ ~ ~ ~ ~ ~ ~ ~ ' ~~ ~
N ~ s 0.0 20.0 40.0 60.0 80.0 100.0
% TW Depth 9/6/96 TSPAxial.XLS ,
DOMIMlON ENGINEERING,INC. Figure B-13D. Estimated Crack Depth Due to axial ODSCC at TSPs 1 1.00 p__-
-_ s t o _ _ _ _,____ _ _ _
0.90 -------/ -- - - - - - - --- -- Observed Dati f
/ l 0.80 - -- ~- - /{ - -- - - . - -
Alpha = 1.61 ]_ _ _ _ _ _
/ 3 Beta = 5.98 0.70 -
- - - - - - - - l- #" "
f-- Square Root Variance = 7.6 f c I '
- I - - - -- - -- -- - - N -- --- --- - - - - - - - - - - - - - - - - - - ~ - - -
&u 0.60 i
Estimated Actual Estimated ActualSize f ---- Fit to Observed
- 0.50 f Distribution
.5 I i -'
Lab 3 g 1 l Alpha = 0.77
- 0.40 1
O Beta = 4.48 I i Mean = 3.4 1--- 0.30
,I-- dRAN~MM~6~~ ~~~ ~ ~-
I I - - - - - - - - - - - - -- -- -- 0.20 i-I-1 J Detection Efficiency = 24.4% 0.10 1 - -- -- - -
--- - -4 >10% TW l i i !
0.00 O.0 20.0 40.0 60.0 80.0 100.0
% TW Depth 9/6s6 TSPAxial.XLS k
DOMINION ENGINEERING,INC. I i Figure B-13E. Axial TSP Defects - Probabil ty of Detection , I t 1.0 l 0.9 O.8 ! O.7 [ O.6 O o 0.5
- n. .
i O.4 O.3 t O.2 O.1 t 0.0 O 10 20 30 40 50 60 70 80 90 100 Defect Depth (%TW) 9/6/96 TSPAxial.XLS
-e-* -
e-- -v -e.r ar- - -
+,m____ _ _ _ _ _ _ . _ _ _ _ _ , _ , . _ _ _ _ ________m _ _ _ _._mm_m_ _ . _ , . .
DOMINION ENGINEERING. INC Figure C-16. Binned Flaw Distributions Axial Crack Depths at TSPs (Assumed Crack Length of 0.75 inches) Case: Moderate Severo Light Detected Flaws: 139.0 1470 32 Detectm Effnency: 0.244 Total Flaws: 569.7 6024.6 131.15 Alpha: 0.770 Beta: 4.480 Depth, % of Wa F No. in Bin No. In Bin No. In Bin 10 0.00000 0.0 0.0 0.0 11 0 21324 121.5 1284.7 27.97 12 0.38028 M2 1006.3 21.91 13 0.51136 74.7 789.7 17.19 14 0.61437 $8.7 620.6 13.51 15 0.69543 46.2 488.3 10.63 16 0.75928 36.4 384.7 8.37 17 0.80963 28.7 303.3 6.60 18 0.84936 22.6 239.4 5.21 19 0.88074 17.9 189.1 4.12 20 0.90554 14.1 149.4 3.25 21 0.92515 11.2 118.2 ' 57 22 0.94067 88 93.5 2.04 23 0.95295 7.0 74 0 1.61 24 0.96268 5.5 $8.6 1.28 25 0.97039 44 46.4 1.01 26 0.97650 3.5 36.8 0.80 27 0.98134 2.8 29.2 0.64 28 0.98518 2.2 23.1 0.50 29 0.98823 1.7 18.4 0.40 30 0.99065 1.4 14.6 0.32 31 0.99257 1.1 11.6 0.25 32 0.99410 0.9 9.2 0.20 33 0.99531 0.7 7.3 0.16 34 0.99627 0.5 5.8 0.13 35 0.99703 0.4 4.6 0.10 36 0.99764 0.3 3.7 0.08 37 0.99812 0.3 2.9 0.06 38 0.99851 0.2 2.3 0.05 39 0.99881 C.2 1.8 0.04 40 0.99905 0.1 1.5 0.03 41 0.99925 0.1 1.2 0.03 42 0.99940 0.1 0.9 0.02 43 0.99952 0.1 0.7 0.02 44 0.99962 0.1 0.6 0.01 45 0.99970 0.0 0.5 0.01 46 0.99976 0.0 0.4 0.01 47 0.99981 0.0 0.3 0.01 48 0.99985 0.0 0.2 0.01 49 0.99988 0.0 0.2 0.00 50 0.99990 0.0 0.1 0.00 51 0.99992 0.0 0.1 0.00 52 0.99994 0.0 0.1 0.00 53 0.99995 0.0 0.1 0.00 54 0.99996 0.0 0.1 0.00 55 0.99997 0.0 0.0 0.00 56 0.99998 0.0 0.0 0.00 57 0.99598 0.0 0.0 0.00 58 0.99998 0.0 0.0 0.00 59 0.99999 0.0 0.0 0.00 60 0.99999 0.0 0.0 0.00 61 0.99999 0.0 0.0 0.00 62 0.99999 0.0 0.0 0.00 Page1 Binned TSP Axial Flaws 9/6/%
. . . .. - _- . ~ , ._. .. - . _._ . . . . = - . - . ~ -
DDMIN10N ENGINEERING,INC Figure C-16. Binned Flaw Distributions Axial Crack Depths at TSPs (Assumed Crack Length of 0.75 inches) l I Case; Moderate Severo Light Detected Flaws: 139.0 1470 32 Detecten Effcency- 0.244 , Total Fiaws: 589.7 6024.6 131.15 Alpha: 0.770 Beta: 4.480 Depth, % of Wall F No. in Brt No. in Bin No. in Bin ! 63 0.99999 0.0 0.0 0.00 64 1.00000 00 0.0 0.00 l 65 1.00000 0.0 0.0 0.00 66 1.00000 0.0 0.0 0.00 67 1.00000 0.0 0.0 0.00 68 1.00000 0.0 0.0 0.00 69 1.00000 0.0 0.0 0.00 70 1.00000 0.0 0.0 0.00 71 1.00000 0.0 0.0 0.00 72 1.00000 0.0 0.0 0.00 73 1.00000 0.0 0.0 0.00 74 1.00000 0.0 0.0 0.00 75 1.00000 0.0 0.0 0.00 76 1.00000 0.0 0.0 0.00 77 1.00000 0.0 0.0 0.00 78 1.00000 0.0 0.0 0.00 79 1.00000 0.0 0.0 0.00 80 1.00000 0.0 0.0 0.00 81 1.00000 0.0 0.0 0.00 82 1.00000 0.0 0.0 0.00 l 83 1.00000 0.0 0.0 0.00 84 1.00000 0.0 0.0 0.00 85 1.00000 0.0 0.0 0.00 86 1.00000 0.0 0.0 0.R 87 1.00000 0.0 0.0 0.00 88 1.00000 0.v 0.0 0.00 89 1.00000 0.0 0.0 0.00 90 1.00000 0.0 0.0 0.00 l 91 1.00000 0.0 0.0 0.00 92 1.00000 0.0 0.0 0.00 93 1.00000 0.0 0.0 0.00 94 1.00000 0.0 0.0 0.00
;5 1.00000 0.0 0.0 0.00 96 1.00000 0.0 0.0 0.00 97 1.00000 0.0 0.0 0.00 I 98 1.00000 0.0 0.0 0.00 l 99 1.00000 0.0 0.0 0.00 l 100 1.00000 0.0 0.0 0.00 sums 569.7 6024.6 131.1 i
l l 9/6/% Page 2 Binned 15P Axial Flaws l
September 6,1996
- 6. Flaws Due to Looce Parts - Wextex Units As discussed in Appendix D, the frequency ofloose parts events for units that are thoroughly inspected on a frequent basis is about 0.046 per steam generator cycle, and about 5 tubes are involved in each event. For a Wextex unit with three steam generators, this amounts to an average of 3 x .046 x 5 = 0.7 tubes per cycle.
The depth and length distributions suggested for loose parts flaws are shown in Figures D-2 and D-3. Distributing the numbers of flaws expected at the end of the operating cycle into size bins is not practical since there are two independent size distributions. The user will have i to determine the probability of having a defect with both a large length and a large depth 3 using Monte Carlo sampling or equivalent of the two distributions. i 1 4 1 i i l l l l 1 l l
September 6,1996 > Appendix D Flaws Due to Loose Parts The objective of this appendix is to quantify the likelihood of flaws due to loose parts (LPs) of various sizes being present in steam generators at the end of an operating cycle. The effects of reduced or less intense inspections on the probability ofloose part events (LPEs) is addressed in a preliminary fashion. i Loose pans in steam generators (SGs) have required repair of substantial numbers of i , tubes, and have resulted in occasional leaks, and two ruptures. The occurrence of LP ) damage seems to be largely a random occurrence. Somewhat surprisingly, there does not I seem to be a noticeable correlation with SG inspection history, since newer plants that have been subjected to less intense inspections (less frequent and/or panial inspections) have not experienced higher rates of leakage or mpture events than the more intensively inspected units. Nevertheless, concern remains that the probability oflarger flaws due to loose pans being present could increase as inspection intensity is decreased. This appendix is an initial effort directed at developing a model for assessing this concem. In the longer term, it may be desirable to develop a more detailed mechanistic model that takes into account the likely growth characteristics of LP flaws, and that also models the likelihood of detection as a function of flaw size and inspection practices. A review of EPRI tube plugging data through 19941 indicates the following for recirculating PWR steam generators with alloys 600 or 690 tubes covered by the EPRI data: Total plant years: 2165 Total SG Years: 6340 Total tubes plugged for LP wear: 972 For this same set of plant experience, two ruptures and 18 leakage events were identified.2 Since complete data on leakage events was not available,it is assumed that the number ofleakage events probably was actually in the neighborhood of 30. Based on the depth and length of flaws required to cause rupture at typical normal operating primary to seconda.y differential pressures, the flaws involved in the two ruptures are considered to have been equivalent to axial cracks of 2.15 inches or longer with wall penetrations of 90% i B. L. Dow, Jr., Steam Generator Procress Reoort. Revision 11. EPRI Nov.1995. 2 Annual EPRI Steam Generator Progress Reports. AECL reports 4449,4753,5013,5242,7251, 7689. 8179, & 9107, periodic EdF reports at EPRI rnectings, and NRC (Murphy) draft review of rupture expenence. D-1
September 6,1996 or more. The 30 leakers are assumed to have had 100% of wall penetrations, and variable lengths. Measured LP flaw depths are available for a set of four units (proprietary data). A gamma distribution was fitted to these data, with the results shown on Figures D-1 and D-
- 2. This gamma distribution was taken as being typical for the industry. It indicates that 3% of the LP flaws will be 100% of wall in depth, which is consistent with the industry leakage history discussed above (i.e.. 03 x 972 = 29 leakage events).
l The above LP flaw data were approximated using the depth distribution of D-1 and 2 and the length distribution shown on Figure D-3. The two distributions were assumed to be independent, i.e., depth was assumed to be independent of length. The parameters of , the gamma distribution for flaw length shown on D-3 were selected by trial and error using Monte Carlo methods to result in an average of two bursts per 1000 flaws. Burst was defined as occurring when the flaw length exceeded 2.15 inches and the flaw depth exceeded 90% of wall in the same trial. Adjustment of the flaw distributions to account for measurement error and POD was not performed since the distributions were normalized to provide results consistent with industry leakage and burst experience. Thus, the distributions represent actual and not measured sizes without adjustment. Based on the above, it is recommended that loose part flaw sizes be modeled using independent gamma distributions for depth and length, with the following parameters: Distribution Deoth. % Wall Length. in. Alpha 2.275 1.900 Beta 17.235 .458 In order to estimate the number of flaws present at the end of an operating cycle, it is suggested that the following approach be taken: 4 For moderately and severely affected units, the tube plugging, leakage and mpture data cited above are assumed to apply since these units are intensively inspected. For 't these units, the probability of an LPE is approximated by the following: n Each LPE is assumed to involve 5 tubes. Using this assumption, there have I been about 972/5 = 194 LPE. D-2
September 6,1996 l Since typical operating cycles are about 1.5 years long, the number of steam generator cycles (SGC) ~ 6340/1.5 = 4227. The frequency of LPE per SGC = 194/4227 = .046 In summary, for units inspected on a thorough and frequent basis (applicable to most plants with 600MA tubes), it is suggested that the probability of LPEs be modeled as being 0.046 per SGC, and that each LPE be taken as involving 5 tubes. It is l suggested that the size and depth distributions of these Daws be deled as described earlier.
- For SGs that are inspected at less frequent and complete rates the likelihood of a LPE is likely to increase as time progresses, since more chances of entry of LPs occur without them being detected. For example, for a SG that is inspected using a 20%
random sample per cycle, the probability of a LPE will be approximately as follows: limg Prob. of No LPE Prob. of LPE EOC1 .954 .046 BOC2 .2 + .8(.954) = .963 .0368 EOC2 (.2 + .8(.954)).954 = .919 .0811 BOC3 .2 + .8(.919) = .935 .0649 EOC3 (.935).954 = .892 .1079 BOC4 .2 + .8(.892) = .914 .0863 EOC4 .954(.914) = .872 .1284 i BOC5 .2 + .8(.872) = .897 .1027 EOC5 .954(.897) = .856 .1440 It is suggested that the increasing probability of a LPE in a SG subjected to partial inspections be modeled using the above approach, with each LPE involving 5 tubes. It is further suggested that the Daw distributions described earlier be used. This amounts to increasing the number or likelihood of tubes being affected by LP I damage, but not changing the Daw size distributions, such that the likelihood of a leak l or rupture is controlled by the number of affected tubes and not by the growth in size l ofindividual flaws. It could be argued that the average sizes of LP Daws should also l increase, as well as the likelihood of occurrence. However, a qualitative review of D-3
i September 6,1996
; plant experience indicates that, in fact, SGs subjected to partial inspections have not experienced deeper or bigger LP flaws.
i i i s r I J e i e i a 3 i i
, t i
4 i 9 h i I l i d l D-4
.= . - -
DOMINION ENGINEERING,INC i Figure D-1. Measured Depths of Loosa Part Wear 20 18 t 16 - - --
- - - - - - - - - - - - - - hidasur~ed Data - ~ - - - - -
Gamma Dist. i 14 --
- - - - - - - - - - - - - - - ~
a = 2.2746 i ! i p = 17.235 i 12 - --- - - - - i
>- t o -
c E 10 tr 2 u. 8
-~-- - --
6 l 4 2 I t 0 [ m m m o m m m m m
- m o e o e 8 m o>
Measured Depth, % 9/6/% 1 oow Part Lkpths f i i t
DOMINION ENGINEERING,INC. Figure D-2. Cumulative Depth Distribution for Loose Part Flaws 1.00 0.90 - - - - - - - --- 0.80 -- -- - - -- - - - - - - - - - - - - - - - - - - 0.70 .] Gamma Distribution Parametem ___
-- -4 ------t--
Mean = 39.210 c Alpha = 2.275
-E 0.60
-- Beta = 17.235 - - - - - - - - - - -
$ Sq. root of var. = 25.996 u.
--]-----
y 0.50
}
3 ! i E i s 0.40 0 0.30
---[-
0.20 j t------ 0.10 --l - - - - - - - - - --- - - - - 0.00 0 10 20 30 40 50 60 70 80 90 100 Depth, % of Wall 9/6/% loose l' art Distr
DOMINION ENGINEERING,INC. ; Figure D-3. Cumulative Length Distribution for Loose Part Flaws 1.00 1 0.90 - - - -- - - - - - - - I- - --- -- -- ---- --
- - -- -- - l -
0.80 j . Gamma Distribution Parameters- - _ _ . 0.70 Mean = 0.870 Alpha = 1.900 e -- - - - -
- Beta = 0.458 - - - - - - - -
j
.@ 0.60 Sq. root of var. = 0.631 u.
- - - - - - ~ - - - - - - - - - - - ' - - - - - - - - -
? 0.50
-~
ii 3
- -- - - ~ - - - - - - --
h 0.40 0 0.30 l 0.20 l 0.10 l 0.00 0 0.5 1 1.5 2 2.5 3 Length, in. I 9/6/% loose Part Distr - _ - _ - . _ - - _ _ _ . _ _ _ _ _ _ _ _ - _ _ - - _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ - - _ _ - _ _ - _ _ _ _ _ _ _ - _ _ _ - _ - _ - _ . - _ _ _ _ _ - _ _ _ _ _ _ _ _ - _ _ _ - _ _}}