ML20127A570
| ML20127A570 | |
| Person / Time | |
|---|---|
| Site: | Trojan File:Portland General Electric icon.png |
| Issue date: | 01/05/1993 |
| From: | Beckjord E NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Murley T Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20127A563 | List: |
| References | |
| NUDOCS 9301110331 | |
| Download: ML20127A570 (18) | |
Text
. --_ _ ._ __
P
,' ENCLOSURE i DEC 0 91H2 HEMORANDUM FOR: Eric S. Beckjord, Director Office of Nuclear Regulatory Research ,.
FROM: LawrenceC.Shao,Direclor Division of Engineering Office of Nuclear Regulatory Research
SUBJECT:
INTERIM PLUGGING CRITERIA FOR TROJAN NUCLEAR PLANT The Division of Engineering has provided a discussion of the key technical aspects of the rationale used to support steam generator tube interim plugging criteria (IPC) for the Trojan nuclear plant and to provide independent conclusions on the viability of IPC for one fuel cycle. The IPC apply only to the specific case of outer diameter stress corrosion cracking (ODSCC) &nd intergranular attack (IGA) at tube support plate (TSP) Intersections in the steam generators. The technical rationale presented in the enclosure are based on data and analyses available from NRC research, Trojan plant operating experience, and the technical literature. The enclosure also reflects staff technical experience and opinions. The Office of Nuclear Reactor Regulation (NRR) has been consulted on technical details regarding IPC during the preparation of this document. The report endeavors to maintain a distinction between staff opinion and published data.
Based on the discussion presented in the enclosure, the Division of Engineering concludes that continued operation of the Trojan plant for one fuel cycle is justified. This justification is based on:
(1) Examination of steam generator tubes removed from service at the i Trojan plant which has revealed cracks that are generally confined l to the tube support plate intersections.
(2) Burst test results from cracked tubes removed from service at the Trojan plant which showed burst pressures well in excess of main steam line break (HSLB) pressure.
l l (3) Stress corrosion crack growth rate results which indicate that i incremental growth of the cracks to a critical length beyond the l tube support plate during one fuel cycle is unlikely.
l l
l l
l
' '}' s 1
9301110331 930107 PDR ADOCK 06000344 PDR P
o.
DEC 0 g jg;y ,. s Eric S. Beckjord 2 )
(4) The probability of a main steam line break, the key initiating event for a steam generator tube rupture is very low for one fuel cycle.
U/ .
Lawrence C. Shao, Director Division of Engineering Office of Nuclear Regulatory Research
Enclosure:
As stated
' r cc: J. Taylor -
J. Sniezek
- 1. Speis J. Heltemes -
W. Minners T. Murley F. Hiraglia W. Russell J. Richardson J. Strosnider L. Kokajko .
J. Fouchard 4
4
ENCLOSURE Discussion of Technical Rationale for Steam Generator Tube Interim Plugging Criteria (!PC) at The Trojan Nuclear Plant The purpose of this report is to provide a discussion of the key technical '
aspects of the rationale used to support steam generator tube interim plugging criteria (IPC) for the Trojan nuclear plant and to provide independent ,
conclusions on the viability of IPC for one fuel cycle. The IPC apply o,nly to the specific case of outer diameter stress corrosion cracking (ODSCC) and intergranular attack (IGA) at tube support plate (TSP) intersections in the steam generators. The technical rationale presented in this report are based on data and analyses available from NRC research, Trojan plant operating experience, and the technical literature. The report also reflects staff technical experier.ce and opinions. The Office of Huclear Reactor Regulation (NRR) has been consulted on technical details regarding IPC during the preparation of this document. The report endeavors to maintain a distinction between staff opinion and published data.
The rationale presented in this report are based on technical considernions which we believe are adequate to justify IPC for one fuel cycle. Subsequent operation with IPC would require additional review after completion of one cycle and would require consideration of additional information developed at that time. Longer term technical considerations, such as reliability and sensitivity of WDE techniques for steam generator tube inspection, are the subjects of on-going and new NRC research which is being coordinated with NRR as part of an overall tteam generator tube alternate plugging criteria (APC) action plan.
(1)
Background:
Steam generator tube structural integrity guidance provided in Regulatory Guide 1.121 has generally translated into a 40% through-wall
" plugging limit" for flaws in steam generator tubes as part of the plant technical specifications. However, evidence from pulled steam generator tubes at several plants has revealed numerous short cracks at TSP intersections which are greater than 40% through-wall and yet can withstand pressures in excess of three times operating as required by Regulatory Guide 1.121.- It has therefore been argued by the industry that the 40% plugging limit is conservative, at least for the case of short axial ODSCC/lGA confined to TSP intersections. Burst testing of cracked tubes removed from service at the Trojan plant has resulted in burst pressures of at least a factor of two in i
excess of main steam line break (MSLB) pressure, even for through-wall cracks.' NRC research results on tubes with machined and chemically-induced .
3, flaws support the contention that the tubes retain significant structural integrity even for up to through-wall cracks, provided that the cracks are short. From this research "short" can be defined as less than 0.5 inches, which is the length of a near through-wall crack needed to burst for 7/8-inch diameter, 0.050 inch wall thickness tubing under MSLB differential pressure' (see Figure 1). The burst pressure is defined as the pressure required to penetrate the tube wall. Tube burst then, can result in either small or large leakage. Tube burst results when the differential pressure acts from the primary side. Tube rupture relates to a significant opening ender burst
!; }
e
. pressure with a consequent increase in leak rate and potential ductile crick advance. Burst failure is differentiated from collapse failure where the
(%)
differential pressure acts from the secondary sida.
Based on the arguments presented previously and supporting analyses, the industry has proposed an alternative to the traditional 40% depth-based guidance, the so-called alternate plugging criteria (APC), for steam generator tubes. The APC are based on correlations between the voltage amplitude '
recorded during eddy current tube inspections with a bobbin coil and subsequent measurements of the tube burst pressures and leak rates. The APC are also currently restricted to ODSCC at TSP intersections. A modified version of steam generator tube APC has been accepted by HRR for several Itcensees. These criteria have been termed interim plugging criteria (IPC).
(2) Trojan service Experience: Examination of steam generator tubes removed from service at the Trojan plant has revealed cracks whir.h were generally confined to the TSP intersections. Evidence from the Trojan pulled tube examinations' has shown that the outer diameter (OD) lengths of the cracks ranged almost up to the TSP thickness. Subsequent evaluation has revea-led that 2 of the 21 TSP intersections examined had cracks which extended beyond the TSP thickness; these cracks extended 0.025 and 0.110 inches beyond the TSP.
(3) Steam Generator Tube Burst Test Results: Burst test results on tubes removed from the Trojan steam generators showed no leakage under normal operating or MSLB pressures.' When pressurized to failure, the burst pressures measured for the tubes were in excess of the HSLB pressure by at least a factor of two. NRC research results from burst tests of tubes with mechined, chemically-induced and service-produced defects have also provided a significant body of data on tube integrity. Equations which have been fitted to the burst test data for electric discharge machined (EDM) slots in steam generator tubes have shown that 0.5 inches would be the length of through-wall crack that would be expected to burst at HStB pressures for 7/8-inch diameter.
0.050 inch wall thickness tubing' (see Figure 1). The equation for the through-wall EDM slot represents an extrapolation from data measured on up to 90% through-w.11 slots. NRC research has shown that the empirical equation developed fron EDM slots provides a realistic estimate of remaining margin to failure for tubes with stress corrosion cracks.* An empirical equation fitted to data from burst tests of uniformly thin.1ed steam generator tubes has also been developed.' This equation is contrasted with the EDM equation in Figure
- 1. It can be seen that the two equations are of similar form but that the uniform thinning equttion provides more conservative estimates of tube burst pressures for flaw depths greater than a/t of 0.8, where a = flaw depth and t tube wall thickness. Use of either equation to bound degradation up to a/t of 0.8 should yield similar results in tsrms of burst pressure. However, the EDM equation provides a more accurate representation of stress corrosion cracking and should be used for flaw depths greater than a/t - 0.8.
(4) Stress Corrosion Crack Growth Rate: Growth of the ODSCC tube cracks at the Trojan TSP intersections is not expected to be significant during one fuel cycle. For purposes of this report, significant can be defined as a through- Q h 1) 2
wall crack extending on the order of 0.5 inches beyond the TSP intersection.
As described previously in (2), the Trojan cracks were generally confined to the 1SP thickness; hence, growth beyond the TSP on the order of 0.5 inches would be required for these cracks to be considered critical from a MSLB pressure standpoint. Upper bound laboratory ODSCC growth rate data' indicate that crack growth of this magnitude would not be expected to occur during one fuel cycle. Uhlle a through-thickness, fu'l TSP length crack would be expected to fail at MSLB pressure, the opening or rupture would be constrained by the tube support plate. True rupture for the portion outside of the TSP would be expected to occur at MSLB pressure only if the crack had grown on the order of 0.5 inches beyond the TSP intersection. Further, little or no movement of the TSP which could potentially " uncover" the cracks is predicted to occur for the HSLB condition.
(5) Probability of Main Steam Line Break: The probability of a MStB, the key initiating event for a steam generator tube rupture, is very low. The MSLB would cause approximately a 2600 psi pressure differcntial across the steam generator tubes. A HSLB has never occurred in a U.S. nuclear plant. Quoting from reference 5, "Under the [yJ1uation and imorovettnt of NDE for Inservice insoettion of licht Water Pea 11 ors Proaram sponsored by the HRC, a team.of experts estimated the median frequency of a MSLB to be 1.7 x 10" per reactor year for a volume of 50 gallons per minute. This extrapolates to a frequency estimate of 6.8 x 10" per reactor year for a four loop plant such as Trojan.
(6) Sumary and
Conclusions:
Based on a review of Trojan steam generator tube operating experience, on destructive examinations of tubes removed from the Trojan plant, stress corrosion crack growth rates and expert opinion concerning MSLB frequency, it is concluded that operation of the Trojan plant with steam generator tube IPC for one fuel cycle does not constitute a significant threat to public health and safety. Subsequent operation with IPC would require additional review after completion of one cycle and would include consideration of infortnation developed at that time. In summary, the above conclusion is based on:
(1) Examination of steam generator tubes removed from service at the Trojan plant which has revealed cracks that are generally confined to the tube support plate intersections.
(2) Burst test results from cracked tubes removed from service at the Trojan plant which showed burst pressures well in excess of main steam line break (MSLB) pressure.
(3) Stress corrosion crack growth rate results indicate that incremental growth of the cracks to a critical length beyond the tube support plate during one' fuel cycle is unlikely.
(4) The probability of a main steam itne break, the key initiating event for a steam generator tube rupture is very low for one fuel cycle.
3 e
References
}
}- " Trojan Nuclear Plant Steam Generator Tube Repair Criteria for Indications at Tube Support Plates,
- Westinghouse Energy Systems, WCAP-13129,: Revision 1, December,1991, WESTINGHOUSE PROPRIETARY CLASS 2.
2- NUREG/CR-0718, Steam Generator. Tube Integrity Program, Phase I-Report, USNRC, September, 1979. .
3- NUREG CR/2336, Steam Generator Tube Integrity Program, Phase.!! Final Report,,USNRC, August, 1988. ,
4- NUREG CR/5117 Steam Generator Tube Integrity / Steam Generator Group Project, Final Project Summary Report, USNRC. May,- 1989.
5 -- Memorandum, C.J. Heltemes to F.P. Gillespie, GI-163, Multiple-Steam Generator Tube Leakage, September 28, 1992.
p W
)
9
.-3
'j .e4 4
P
+
-. 4 m .-a... _J-.' .+,--.._..# , - -.. , ., , ,,wy ,,,%.y~ , , .,_, , . .7., ,, ..,3, ,.__ 9.
ji $*' .. ~
,'.ijjy ' i 1
. 1 k Detta P = DIfferentIel Pressure 0.9 -
Po = surst Pressure foe thftewed Ttee (Ap realantely 10,000 pet) l'
.{
0.8 -
i w 0.7 - .
o
\
- $- 0.6 -
w Q
o_
. 0.5 -
L EDM, M= 0.8 g,
_ O.4 -
< .W -
UNIFORM THIN O g = o,e
, 0.3 - . . ,,,,, 4 ,
N
' l' ao ni _
___ 4 i 0.2 - UNIFORM THIN M = 0.9 EDM. M O.9
. 0.1 -
"""""""""LLL: :
1 =a :,: : : : : : -
~ EDM, M=1.0 (tirough-waf) 0
' ~
';^:: e--------
.g i i i . . .
.0 0.4 i 0.8 1.2 1.6 : 2-Y .l
' ' ' *5 '"d**
LENGTH (inches)
Figure 1 . Burst Pressure Parameter curves for EDM Slot and Uniform Thinning Specimens
4 ENCI.05URE 2 NOTE TO: G. Burdick FROM : J. Hopenfeld
SUBJECT:
Repty to your request for comments on Draft 'R$S POSITION ON STEAM GENERATOR TUBE INTEGRITY' by LC. Shao Based on certain data, discussed in items 18, the document concludes that *it is reasonable to continue operation for one fuel cycle with flaws greater than 40% through-wall at TSP intersections.' The document further suggests that ' subsequent operation will require additional review after completion of one cycle and willinclude consideration of information developed at that time'.
GENERAL COMMENT
The information provided in items 18, of the subject document does not address the main issue conceming steam generator tube integrity which arose from recent operating -
experience with ODSCC. The issue is as follows:
Is it safe to operate plants where an accident such as steam or feed line break may open existing but previously undetected cracks, which will result in a significant primary-to- secondary leakage. Whether the leakage is significant or not would depend on whether the operator can stop the leak before the RWST is depleted.
Degraded tubes also may cause a significant increase on risk from severe accidents.
The f act that cracks within the TSP can vathstand the MSLB pressure and that their length will not become critical during one fuel cycle is not an indication that they also will not l leak. The Trojan burst test results show that three oest of the 21 test specimen developed leaks at pressures, of 3300 psi, 7500 psi, and 5500 psi,. with an average depth of penetrations of 38%,58% and 72% respectively.
Iterr. (7) points out that the above specimens *have shown no leakage under normal operating or MSLB pressure conditions * . IT FAILS TO POINT OUT, HOWEVER, THAT THERE IS NO DATA WHICH WOULD ALLOW ONE TO RELATE THE ABOVE LEAKAGE-WITH THE OBSERVED DEGREE OF DEGRADATION. In other words,if these specimen had undergone a more severe wa!! penetration would these specimens have leaked at 2600 psi.7. Considering that the 21 specimeh represent a sample of a population of I 13,000, the conclusions in (7) above are questionable.
The document ignores two other tubes which were pu!!ed out of two US plants and developed leaks at SLB pressures. The leakage was at least an order of magnitude higher than under normal delta ps'. A third tube from a Belgian plant indicates a factor of eight increase in leakage under SLB conditions, (see Mar. _23 memo) . Theoretical
.c - - .
considerations also indicate a factor of 1000 increase in leakage under SLB conditions.
In conclusion, the absence of a deterministic and empirical models for these newly observed cracks precludes the conclusions reached in the subject document. The claim that the conclusions in item 9 are supported by items 1-8 could be considered valid only if one ignores the available data which indicato that higher than normal leakage will occour at SLB pressures even if the tubes do not rupture.
Finally the justifications for arry plant operation should not be based on staff opinions or published data.on SCC.
SCC is a semi ompirical art, in the absence of applicable database other routes of approaching the problem should be considered.
Ele justification for operating with cracked tubes should be based on what procedures would the operator follow given certain primary to secondary leak and a MSLB between the containment and the MSIV. These justifications should clearfy demonstrate compliance with 10CFR100.
I beleive that the staff can more properly judge operator action than predict localized - .
corrosion behavior.
SPECIFIC COMMENTS:
ttem 1.
The EDM initiated grooves studies provide some measure of the ability of the tube to resist rupture given certain known wall Imperfections. It bears little relation to how ODSCC form, propagate and leak in steam generator environments.
Item 4 This definition of *significant* is questionable. It makes no difference whether the cracks extend beyond the TSP if they leak at the gap. it appears that operators rely on such leakage because they lowered the leakage requirements during normal operations.
Unless one can show that the TSP will cause cracks to plug and they will remain plugged under MSLB pressures the above definition may lead to confusion.
The statement that
- upper bound laboratory ODSCC ....
- would not be expected to occur during one fuel cycle
- is not supported by data. The document should compare and present plant and laboratory data with regard to stress
- intensities and environments before making such claims.
Item 5 The high frequency quoted,6.8x 10-4/RY contradicts the statement that
- it is reasonable *
, item C , because this frequency would result in a core melt probability of 6.8x 10-3/RY with containment bypass as discussed in the March 27 Memo. The above number is consiocrably higher than present safety goals.
The statement that the key initiating event for SGTR is MSLB is incorrect when taken in the context of the entire documsnt. Item 6 contradicts this statement.
Item 7 Although this item is correct, as stated, it presents only part of the data. As already discussed, three tubes leaked at Trojan. Three tubes from other plants also leaked at MSLB pressures. Rudimentary consideration dictate that leakage increases when detta
- p across the wa!! is increased.
Item 8 The lengthy discussion of uniform thinning only confuses the main issue. There are
[
several ways that the reouction load bearing capabilities of a component due to corrosion can be accounted for there is nothing special about these equations. The ASME j
code takes this into account. The main problem here is LOCAllZED corrosion with an
! UNKNOWN ATTACK RATE.
1 L lte m 9 A discussion should be added of the type of new information which is required for.the l
- additional review" to justify subsequent operations.
1 l
- ATTACHMENT 1 i; -
Second item : Dr. instead of Mr. or Just Hopenfeld The following is missing:
On Sept 11,1992 J. Hopenfeld filed an addendum'to the March 27,1992 concluding -
that " strong coupling exists between hot leg mass flow, SG tube Icakage and crack
On Sept 11,1992 J. Hopenfeld filed an addendum to the March 27,1992 concluding
. that ' strong coupling exists between hot leg mass flow, SG tube leakage and crack propagation. If confirmed, such a relation between system behavior and undetected tube defects may cause small leaks to quickly enlarge and results in a MULTIPLE TUBE RUPTURE BEFORE THE RCS IS DEPRESSURIZED BY FAILURE OF THE SURGE LINE.
THE RESULTANT CONTAINMENT BYPASS WILL INCREASE E S URCE TERM.*
J. H9pd cc; _
P. Norian, Gm. Mazetis, W. Minne , . Beckjor e
-'S
! ~.
ENCLOSURE 3 l DIVISION OF ENGINEERING RESPONSES TO COMMENTS OF J. HOPENFELD l
The Division of engineering has compiled responses to the comments of J.
Hopenfeld on steam generator tube interim plugging criteria (IPC). The responses are restricted to the Division's areas of expertise and refer to the note from J. Hopenfeld to G. Burdick of December 9,1992 (attached). The comments by J. Hopenfeld, in turn, refer to a previous draft of the Division of Engineering memorandum which was issued in final form on December 9,1992.
(1) COMMENT: Is it safe to operate plants where an sccident such as steam or feed line break may open existing but previously undetected cracks, which will result in a significant primary-to-secondary leakage? Whether the leakage is significant or not would depend on whether the operator can stop the leak before the refueling water storage tank (RWST) is depleted. Degraded tubes may also cause a significant increase on risk from severe accidents.
RESPONSE: Plants operating with the interim plugging criteria (IPC) will have an increased probability of leakage under HSLB or FLB. As a result these plants have more restrictive limits on leak rate, and operators have received additional training in the handling of steam generator leakage problems. The Division of Engineering has estimated the overall leakage under Main Steam -
Line Break (MSLB) to be 145 gallons per minute (gpm) for the Trojan plant.
With consideration of a 428,000 gallon RWST, this provides the operators with significant time in which to respond to steam generator tube leakage or rupture.
(2) COMMENT: The fact that cracks within the TSP can withstand the HSLB pressure and that their length will not become critical during one fuel cycle is not an indication that they will not leak.
RESPONSE: As stated previously, plants operating with IPC will have an 4
increased probability of leakage under MSLB or FLB. The Division of Engineering has attempted to estimate this leakage as described in the response to comment (1).
(3) COMMENT: ... There is no data which would allow one to relate the above leakage with the observed degree of degradation.
RESPONSE: A leak rate model (reference 1) has been used to relate crack size with the amount of leakage under normal operating conditions and MSLB. The leak rate model has been used to determine the estimate presented in the response to comment (1). The model has been benchmarked to a previous EPRI leak rate model (reference 2) and to leak rate data from both U.S. and foreign sources. The leak rate data were for cracks in. laboratory-degraded tubes that were intended to simulate tubes with actual service-induced cracks. While the leak rate data exhibit significant scatter, the overall trends of leak rate versus crack length were consistent with the leak rates predicted by the model.
- ,-n e- -- -,-e w.--, e , , .---ne , - - - -
,1-, -
--wm ra--m-w,w-----,---org-~- g ,,-,,-,.,w - , -v -m -
(4) COMMENT: The document ignores two other tubes which were pulled out of two U.S. plants and developed leaks under SLB pressures. The leakage was at least an order of magnitude higher than that developed under normal delta p's.
A third tube from a Belgian plant indicates a factor of eight increase in leakage under SLB conditions (see Mar. 23 memo). Theoretical considerations also indicate a factor of 1000 increase in leakage under SLB conditions.
RESPONSE: The Division of Engineering memorandum considered Trojan plant-specific data (reference 3). Pulled tube data from other U.S. plants, also documented in reference 3, showed four cases where leakage rates increased by a range of from 7X to 2,040X in going from normal operating differential pressure to MSLB. However, in all of these cases the leakage rate went from completely insignificant to a barely measurable leakage, e.g. 0.0041/hr (0.000017 gpm) to B.161/hr (0.035 gpm), and not to a large leak or rupture.
The leak rate data and modeling from reference 1 indicate an expected increase of 10 times in going from normal operating differential pressure to MSLB for a 0.7 inch long flaw. Trojan pulled tube burst tests revealed no leakage under normal operating or MSLB conditions. A deterministic model for leakage from an axial crack in a oressurized tube, described in reference 4, indicates a possibility of up to a factor of 1000 increase in leak rate in going from
~
normal operating conditions to MSLB. However, to obtain such a large increase, the model assumes tight (10ym wide at the center) cracks which would have completely insignificant leakage under normal operating conditions.
Belgian test data provided in reference 3, show approximately a factor of 2 increase in leakage for MSLB versus normal operating conditions. Our conclusions from the examination of this data indicate that while a factor of 1000 increase in leak rate is possible in going from normal operating differential pressure to MSLB, such an increase is not likely. Further, on an absolute basis, the resulting amount of leakage for a given tube would likely be insignificant.
! (5) COMMENT: In conclusion, the absence of deterministic and empirical models for these newly observed cracks precludes the conclusions reached in the subject document. The claim that the conclusions in item 9 (draft memorandum) are supported by items 1-8 (draft memorandum) could be considered valid only if one ignores the available data which indicate that higher than normal leakage will or. cur at SLB pressures even if the tubes do not rupture.
I RESPONSE: The conclusions in the memorandum were reached primarily on the basis of the limited incremental crack growth expected for one fuel cycle and the low probability of MSLB. The emphasis in the memorandum was also on steam generator tube rupture as opposed to leakage. Subsequent work by the Division of Engineering has provided a best estimate of tube leakage under MSLB based on a probabilistic flaw distribution (using Trojan pulled tube data as the basis) and a leak rate model (reference 1). This analysis indicates a best estimate leak rate of 145 gpm under MSLB. This analysis, coupled with the Trojan pulled tube data and the discussion in the response to comment (4) above, do not support the claim for a drastic increase in leakage under MSLB.
(6) COMMENT: Finally, the justifications for any plant operation should not l
be based on staff opinions or published data on SCC. SCC is an empirical art, l
in the absence of an applicable database, other routes of approaching the l
1
, _ . . _ . _ _ . _ - , _ , , - , -_..r
problem should be considered. I believe that the staff can more properly judge operator action than predict localized corrosion behavior.
RESPONSE: Justifications for plant operation are commonly based on consensus staff opinion regarding the state-of-the-art in a specific technical area.
SCC growth rate data specific to the Trojan plant steam generator tubing and environment were not available for this analysis. However, SCC growth rate data for mill-annealed inconel 600 in a NaOH environment (49/1 concentration) .
at 350*C (662'F) were available from reference 5. Utilizing this data, several growth rates were applied to a probabilistic flaw distribution based on the Trojan pulled tube examinations to yield a final leak rate determination. A 'best estimate" growth rate of 0.1 pm/hr was used which .
yields total growth over a twelve enonth cycle that would be conservative based on comparisons with Trojan operating experience.
(7) COMMENT: Item 1, draf t memorandum - The EDM initiated grooves study provides some measure of the ability of the tube to resist rupture given certain known wall imperfections. It bears little relation to how ODSCC form, prepagate and leak in steam generator environments.
RESPONSE: The EDM data were utilized to ascertain the length of crack thht would be expected to " burst" at various pressures, including MSLB. Burst is defined as through wall penetration of the tube wall with pressure from the -
primary side. No claims were made for EDM slots simulating initiation and propagation of SCC cracks. Equations fitted to the EDM slotted tubes data were found to predict burst pressures for tubes with stress corrosion cracks of similar lengths and depths (reference (1)).
(8) COMMENT: Item 4, draft memorandum - The definition of significant is questionable. It makes no difference whether the cracks extend beyond the TSP
-if they leak at the gap. It appears that operators rely on such leakage because they lowered the leakage requirements during normal operations.
Unless one can show that the TSP will cause cracks to plug and they will remain plugged under MSLB pressures, the above definition may lead to confusion.
RESPONSE: It does make a significant difference if cracks extend beyond the TSP both from the standpoint of potential tube rupture and increased leakage.
If the cracks extend far enough beyond the TSP (on the order of 0.5 inches),
full rupture could occur under HSLB. While a through-thickness, full TSP length flaw would be expected to fail it MSLB pressure, the opening would be constrained by the TSP. This constraint would serve to limit a rupture and any consequent leakage.
(9) COMMENT: Item 4, draft memorandum - The statement that " upper bound laboratory ODSCC ... would not be expected to occur during.one fuel cycle" is not supported by data. The document should compare and present plant and laboratory data with regard to stress intensities and environments before making such claims.
J
.e
-r
-' m g
- ys3.=-
J RESPONSE: The statement referred to above is specifically based on data that ;
was referenced in the memorandum (reference (5)). The discussion in response to comment (6) is applicable here.
i (10) COMMENT: Item 7, draft memorandum - Although this item is correct as stated, it provides only part of the data. As already discussed, three tubes leaked at Trojan. Three tubes from other plants also leaked at MSLB pressures. Rudimentary considerations dictate that leakage increases when delta p across the wall increases.
RESPONSE: The aulled tube data from Trojan showed that none of the tubes that were tested leaced at MSLB pressure. Three of the pulled Trojan tubes eventually leaked beyond MSLB pressures (3300 psi, 5500 psi and 7500 psi) before rupturing (reference 6). The tube that leaked at 3300 psi was also noted to have been damaged during removal. Pulled tube data from other plants have shown leakage for MSLB pressure as described previously in the response to comment (4). For a through-wall flaw, leakage should generally increase with increasing differential pressure across the tube wall. However,_as pointed out previously in the response to comment (4), the leakage may be completely insignificant.
(11) COMMENT: Item 8, draft memorandum - the lengthy discussion of uniform thinning only confuses the main issue. There are several ways that the reduction in load bearing capabilities of a component due to corrosion can be accounted for (and) there is nothing special about these equations. The ASME Code takes this into account. The main problem here is localized corrosion with an unknown attack rate.
RESPONSE: The discussion of uniform thinning was included at the request of the Olvision of Safety Issue Resolution (DSIR) to clarify differences with the EDM slot equation. While the growth rate of the localized corrosion (ODSCC and Intergranular Attack (IGA)) is uncertain, it can be reasonably estimated with date available from reference 5.
(12) COMMENT: A discussion should be added of the type of information which is required for the " additional review" to justify subsequent operations.
RESPONSE: Information for additional review would include, but not be limited .
to, eddy current inspection data at the end of the cycle, projected versus /
actual crack growth based on destructive analyses of pulled tubes,
metallographic examination of pulled tubes, leak and rupture-tests of pulled tubes and additional secondary side inspections.
References:
1- NUREG CR/2336, Steam Generator Tube Integrity Program, Phase 11 Final ,
Report, USNRC, August, 1988.
2- Griesbach, T., R. Cipolla and J. Lang, ' Analysis Nethods for Evaluating Leak-Before-Break in U-Tube Steam Generators," Proceedings of the ASME Winter Meeting, Miami Beach, FL, 1985. . . .
3- " Trojan Nuclear Plant Steam Generator Tube Re)atr Criteria for Indications at Tube Support Plates,
- Westingiouse Energy Systems, WCAP-13129, Revision 1, December,1991, WESTINGHOUSE PROPRIEIARY CLASS 2.
4 - EPRI NP-6864-L, LICENSABLE KATERIAL, 'PWR Steam Generator Tube Repair Limits: Technical Support Docun.ent for Expansion Zone PWSCC in Roll Transitions (Revision 1)," Electric Power Research Institute, Palo Alto, CA, December, 1991.
S- NUREG CR/5117, Steam Generator Tube Integrity / Steam Generator Group Project, final Project Summary Report, USHRC, Hay,1989.
6 - Magee, T.P J.F. Hall and R.S. Maurer, ' Trojan Steam Generator Tubing Destructive Examination Interim Report Update," TR-MCC-186, ABB --
Combustion Engineering, Windsor, CT. NOTE: THIS REPORT IS NOT DATED AND MAY BE CONSIDERED PROPRIETARY.
I e - - , , w