ML20086R396

Forwards Response to NRC 950622 RAI Re Proposed Revs to TS Related to Alternate Plugging Criteria for Byron Unit 1 & Braidwood Unit 1

ML20086R396
Person / Time
Site:	Byron, Braidwood
Issue date:	07/21/1995
From:	Saccomando D COMMONWEALTH EDISON CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9507310136
Download: ML20086R396 (38)

Text

a

' Commonweattli litivm Company I 400 Opus Place Ikewners Grove, 11.60515 July 21,1995 Office of Nuclear Reactor Regulation i

U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Attn: Document Control Desk

Subject:

Response to Request for Additionalinformation Pertaining to the to Application for Amendment to Facility Operating Licenses:

Byron Nuclear Power Station, Units 1 and 2 NPF-37/66: NRC Docket Nos. 50-454/455 Braidwood Nuclear Power Station, Units 1 and 2 NPF-72/77: NRC Docket Nos. 50-456/457

Reference:

D. Lynch letter to D. Farrar dated June 22,1995, transmitting Request for Additionalinformation Pertaining to Proposed Technical

]

Specification Amendment Regarding increase in the APC Criteria j

The Reference Letter transmitted the Nuclear Regulatory Commission (NRC)

Request for AdditionalInformation (RAl) pertaining to Commonwealth Edison Company's (Comed's) proposal to amend Appendix A, Technical Specifications of Facility Operating Licenses NPF-37, NPF-66, NPF-72 and NPF-77. The proposed amendment request addresses Technical Specification changes necessary to increase the Alternate Plugging Criteria (APC) value for Braidwood Unit 1 and Byron Station Unit 1 Steam Generators from 1.0 volt to 3.0 volts. Attached is Comed's response to RAI questions 27-43.

If you have any questions concerning this correspondence, please contact this office.

Sincerely,

[3 i

I hahw

), w J O

/

Denise M. Saccomando Nuclear Licensing Administrator cc:

D. Lynch, Senior Project Manager-NRR R. Assa, Braidwood Project Manager-NRR G. Dick, Byron Project Manager-NRR S. Ray, Acting Senior Resident inspector-Braidwood H. Peterson, Senior Resident inspector-Byron Office of Nuclear Safety-IDNS D

9507310136 950721 1

x:nt.%wn.iscouper.;2 PDR ADOCK 05000454

'X P

PDR A t nicom compan3

r..

T p.

4, i

I l

l L

_l l

Response to NRC Reauest for AdditionniInformation Renardine the Proposed Revisions to the Tachnical Snacifications Related to the Alternate Pineetne Criteria Byron Unit 1 and Braldwood Unit 1 NRC Ietter Dated June 22,1995 l

8:MPOCCE93\\RAI,619,W July 21,1995

The following are the responses to the request for additional information regarding the proposed revisions to the technical specifications related to the interim plugging criteria for the Byron Unit I and Braidwood Unit I as transmitted on June 19, 1995. Question numbering is keyed to that used by the NRC in the transmittal of the RAI to Comed.

1

27. The technical bases for your pending license amendments relies, in part, on the integrity of support structures (e.g., vertical bars welded to the partition plate and the wrapper, wedges welded to the wrapper). Discuss how the integrity of these structures is ensured against such structural defects as weld cracking and other types of structural defects.

Discuss any examinations of these structures in the Byron 1 and Braidwood I steam generators (SGs) which have been performed recently or that will be performed as a result of the implementation of your pending proposal for increasing the voltage limits of the voltage-based repair criteria for hot-leg TSP intersections. Discuss any relevant inspection findings regarding the integrity of SG internals at other nuclear power plants.

RESPONSE

The integrity of the support structures, such as bars and wedges, is assured by the large margins built into the design and manufacturing of these components. The loads on these components under normal operating conditions are small. Without tube lockup, the loads are negligible. With tube lockup, the loads would increase moderately but would not challenge the integrity of the welds. In this case, however, the lockup loads would resist the SLB forces (see response to Question 36) and the bars would not be required. It should be noted that none of the bars or wedges are welded to both the wrapper and the TSP. In the Model D4 SG, the bars and wedges are welded to the wrapper or divider plate, dependent on their location.

Under an SLB event, the loads are small and the support structures have been structurally evaluated as discussed in Section 8.10 of WCAP-14273. Analyses for the bars with 0.38" full length fillet welds (minimum bar length is 1") are given in Table 8-12. The wedges, which are installed in pairs, have full length 0.38" filet welds on three sides including their 4" height and 2" width. The SLB loads on the wedges are < 100 lbs and the resulting stress intensity is < 100 psi.

No specific examinations of the bars or wedges have been performed in the Braidwood-l and Byron-1 SGs. Due to the large design margins, no inspections are planned or necessary to support implementation of the higher voltage repair limits.

Inspections of bars and wedges have not performed in any systematic manner at any plant to the knowledge of Westinghouse. Peripheral regions of the tubesheet and, occasionally the TSPs, are visually inspected as part of FOSAR inspections for loose parts. These inspections would identify a bar or wedge as a loose part. No bars are known to have 5:\\AMMIT95\\RAI,619.WP5 I

July 21,1995

been found as loose parts. Wedges that have been found as loose parts have been attributable to having been left in the SG during the manufacturing. These wedges have typically not had welds, were rejectable shapes, etc. and usually found during early operation. There are no known cases of finding a wedge attributable to a service failure, except possibly (not traceable information) in early SGs with heavy denting sufficient to deform the TSPs. These SGs have since been replaced.

In severely dented SGs, TSP cracking has been observed. Byron and Braidwood have not seen any corrosion induced denting. As noted in WCAP-14273 Section 10.7.1, a dent of

> 65 mils is required to develop stress levels above yield in TSP ligaments at dented intersections. To determine if dents of this size exist, all tubes are required to pass a 0.590 probe as is described in the response to question 35. If a tube can not pass this size probe through a dented hot leg intersection, all surrounding indications at the TSP must be repaired to the repair criteria applied to the cold leg TSPs (i.e. I volt APC) rather than the increased voltage repair criteria for the hot leg TSP (refer to WCAP-l 14273 Section 12.4).

l

28. State whether design and/or as-built conditions were used in calculating the TSP displacements under postulated accident conditions. If design specifications were used; discuss the implications of any differences between the design and as-built configuration of the SGs. Describe any effort that has been performed, or will be performed, to confirm that the as-built configurations do not significantly differ from the design configuration.

RESPONSE

1 Design conditions were used in calculating the TSP displacements under postulated accident conditions. The displacement analysis results are relatively insensitive to small l

positional variations of wedges and bars, such as would be consistent with manufacturing l

variations.

Although there are no known cases of weld failures at bars or wedges, it is intuitively more likely that a wedge could become loose than a bar. To demonstrate the minimal sensitivity of the displacements to the wedges, an analysis was performed with and without the wedges at the most limiting TSP location with tube expansion. The most limiting location for displacement that could be significantly affected by the wedges is the j

corner of TSP 3 (plate C) near the tube lane (10' wedge location). The results indicate i

that the maximum TSP displacement is increased by only 23 mils (0.051" to 0.074")

when the wedges are removed. Thus, with tube expansion, the TSP displacements are

{

not significantly dependent on the presence of the wedges.

SAAIOCCIN5\\RAI,619.WP5 2

2 1 2i.1995 7

1 Based on the general insensitivity of the displacement analyses to reasonable variations in their positions, the low likelihood of a weld failure that would cause a bar or wedge to be missing or inactive and the limited influence of the wedges on the calculated TSP displacements, no effort has been performed nor is planced to confirm that the as built configurations do not significantly differ from the design configuration.

29. In several places throughout WCAP-14273 (e.g., Pages 2-2 and 12-9) you state that the Electric Power Research Institute (EPRI) free span leak rate methodology will be used for calculating the primary-to-secondary leakage under postulated accident conditions.

Previously, for Braidwood I and Byron 1, a methodology which differs from the EPRI methodology was used as documented in the safety evaluations which were issued on August 18, 1994, for Braidwood I and on October 24,1994, for Byron 1 pertaining to the use of the " free span" voltage-based tube repair criteria. State which of these methodologies will be used. If the EPRI methodology is to be used, address the staff comments on this methodology documented in the internal NRC memorandum dated May 30,1995, from Mr. Frank J. Miraglia to Mr. Edward L. Jordan requesting review of the

" free span" voltage-base SG tube repair criteria in the proposed final version of the generic letter sent to the Committee to Review Generic Requirements (CRGR). (Refer to, Comment I-3 of this document.)

RESPONSE

The references to the EPRI free span leak rate methodology in WCAP-14273 were intended to reference this methodology in a very general sense - the use of the free span leak rate versus voltage correlation and probability of leakage correlation developed consistently with the EPRI documentation although using the latest available database consistent with NRC approved data exclusions. The Monte Carlo methodology would be that documented in WCAP-14046 Rev.1, WCAP-14277 and WCAP-14273 as modified by the Question 39 response. These Monte Carlo methods are consistent with the draft generic letter and the CRGR review comment referenced in the question.

30. The staff is continuing its review of your proposed technical specification (TS) changes.

However, we believe that you should consider adopting the proposed wording in the model TSs in the memorandum dated May 30,1995, cited above, which addresses the

" free span" voltage-based SG tube repair criteria.

RESPONSE

Comed has incorporated the model Technical Specifications from the May 30,1995 Memorandum (May 30 Memorandum) into a revised 3 volt APC submittal which supplements the February 13, 1995 submittal. The supplement was sent to the NRC on SAAFC\\0CIN5\\RAI,619EP5 3

u y 2i,i,95 L

[

l July 7,1995 (July Supplement). Requirements from the May 30 Memorandum have been incorporated into the July Supplement except for being revised for a 3 volt APC instead of a 1 volt APC as described in Attachment A of the July Supplement and shown in markups to the existing Specification in Attachments C-1 (Braidwood) and C-2 (Byron).

31. With respect to the rotating pancake coil (RPC) examinations which will be performed in implementing this voltage-based repair criteria, the staff has the following comments and questions:

a. The staff's position related to " free span" voltage-based SG tube repair criteria in the draft generic letter cited above, is that RPC examinations refer not only to the standard 3-coil RPC but also to comparable or improved nondestructive examination j

(NDE) techniques. You should consider using alternate techniques after appropriate

]

testing is performed to ensure that any alternate probe is either comparable or better than the existing 3-coil RPC. Further, you should ensure that the data obtained from any probe which you may use will meet the intent of the RPC examinations discussed in the forthcoming final version of the generic letter on voltage-based SG tube repair criteria.

RESPONSE

j i

Comed, in association with EPRI, has continued to investigate improved nondestructive examination (NDE) techniques. For examination of ODSCC at the TSPs, the standard 3-coil RPC probe has been shown to be the appropriate probe per Appendix H of the EPRI PWR Steam Generator Examination Guidelines, EPRI NP-6201. Other NDE techniques are not planned for the Fall 1995 outages at Byron or Braidwood. If other NDE techniques or other probes are shown to be comparable or better than the existing 3-coil RPC, after appropriate testing and assurance that the technique or probe meets the intent of the May 30 Memorandum, consideration will be given to using the other NDE technique or probe.

With the July 7 Supplement, Comed deleted the Upper Voltage Repair Limit for the hot-leg TSP intersections. Any crack-like indication at a hot-leg TSP identified by bobbin coil examination greater than 3 volts will be repaired regardless of confirmation by RPC.

Confirmation by RPC of TSP indications greater than 3 volts has been shown to be nearly 100%. Hot-leg TSP indications greater than 3 volts will still be inspected by RPC to determine if the indications are crack-like, axial in orientation, and within the TSP.

Indications in cold-leg tubing will follow the 1 volt APC Upper and Lower Voltage Repair Limits as defined in the May 30,1995 Memorandum.

SAAPCMrt'95\\RAI 619.WP5 4

July 21,1995

b. You propose an RPC sampling plan for intersections with bobbin dent voltages exceeding 5.0 volts and for large mixed residuals. With respect to the large mixed residuals (i.e., those that could potentially mask flaw responses at the voltage repair limits), state whether your proposal is for an upper or lower voltage repair limit. In addition, on Page 2-5 of WCAP-14273, you refer to mechanically induced dents in the context of RPC sampling. (Presumably, this has been stated in this fashion since no corrosion induced denting has been discovered.) State whether the RPC sampling requirements will be implemented at both mechanically induced dents and corrosion induced dents, should they develop. Refer to staff comments on the issue of RPC sampling contained in the memorandum dated May 30,1995, cited above, (refer to, Comments II-4 through II-8).

RESPONSE

l Indications with large mixed residuals will be inspected with RPC. In the July 7 Supplement, Attachment B states that RPC inspections are to be performed on "All intersections with large mixed residuals that could cause a 3.0 volt signal in a hot-leg tube or a 1.0 volt signal in a cold-leg tube to be missed or misread." Therefore, the Iower Voltage Repair Limit is used for cold-leg indications. For hot-leg indications, neither a Iower or an Upper Voltage Repair Limit exists. The voltage repair limit for a hot-leg tube is an indication greater than 3 volts.

TSP indications that contain dents, whether mechanically or corrosion induced, will be examined by RPC. In the July 7 Supplement, Attachment B states that RPC inspections are to be performed on "All TSP intersections that contain dents greater than 5.0 volts and a 20% sample of dents between 2.5 volts and 5.0 volts. If Primary Water Stress Corrosion Cracking (PWSCC) or circumferential cracking is detected,100% of the dents between 2.5 volts and 5.0 volts will be inspected." Distinguishing between mechanically or corrosion induced dents is not necessary since the inspection scope is for all dents.

But, distinguishing between mechanically or corrosion induced dents is addressed in the response to question 32.

J Corrosion induced dents have not been observed in the Byron or Braidwood steam i

generators. The chemical environment that is believed to cause ODSCC (i.e., a caustic crevice solution) is opposite to the environment that is believed to cause corrosion induced TSP dents (i.e., an acid crevice solution). Therefore, the probability ofinitiating corrosion induced denting is low.

l

c. Although interference in the eddy current signal attributable to the presence of copper i

is not expected since Byron 1 and Braidwood I have not operated with copper present in the secondary system as discussed on Page 12-5 of WCAP-14273, we believe that your proposed RPC sampling plan discussed on Page 2-5 and Section 12.5 of WCAP-14273 should address the need for performing such inspections in the event that copper might be introduced into the Byron 1 and Braidwood SGs.

SMPtNTP95WAl,619.WPs 5

wrzi.i995

=.

RESPONSE

Byron and Braidwcod have not operated with copper components in the secondary systems. Only trace amounts of copper have been found in samples of sludge from the steam generators. This trace amount of copper in the steam generator sludge is not believed to cause interferences in the eddy current signal. Copper components are not planned to be introduced into the secondary systems due to copper being known to accelerate steam generator corrosion. In the event that copper may accidentally be introduced into the secondary systems, the July 7 Supplement states that RPC inspections are to be performed on "All intersections with interfering signals from copper deposits."

In addition, guidance an conducting RPC inspections for interference signals due to copper has been included in both stations eddy current inspection guidelines.

32. To provide assurance of TSP integrity, TSP intersections selected for tube expansion and all surrounding TSP intersections should be confirmed to be free of corrosion induced denting. Discuss how corrosion and mechanically induced dents are distinguished from each other. If corrosion induced denting were to develop in future operating cycles, state whether additional SG tubes would be expanded. Discuss your plans for monitoring and reporting such occurrences.

RESPONSE

Corrosion induced dents can be distinguished from mechanical dents by reviewing tube inspection histories. If the dent has been present in the tube for many cycles (i.e., since pre-service) and has not grown, then it is a mechanically induced dent. Otherwise, it is assumed to be a corrosion induced dent, unless it can be proven than new debris in the area mechanically caused the dent. As stated in the response to Question 31b, corrosion induced dents at the TSP are not expected to form since the chemical environment needed is the opposite to the environment which is present in the crevice for ODSCC to form.

Monitoring for new dents is a normal part of the bobbin inspection. New dents will be reported by the bobbin inspection and evaluated for impact relative to expanded tube considerations as noted below.

As noted in WCAP-14273 Section 12.4, the requirement relative to dents and expanded tubes is that all TSP intersections selected for tube expansion and all adjacent intersections must not have corrosion induced dents > 5 volts to conservatively support TSP integrity at the expanded locations. As also noted in this section, a dent of > 65 mils is required to develop stress levels above yield in the TSP ligaments at dented intersections. The use of the 5 volt dent limit at TSP intersections to be expanded or adjacent intersections is an added measure of conservatism to enhance TSP integrity. Corrosion cracking of a tube at a dented TSP intersection (mechanical or corrosion induced dent) is less of a concern than TSP integrity due to the low plugged tube temperatures of the expanded tubes.

5:\\AIOCCI'95iRAI,619.WP5 6

My 21,1995

As discussed in Attachment B of the July 7 Supplement, redundant tubes are to be expanded during the Fall 1995 outages to provide sufficient redundancy in the unlikely event that one or more expansions fail due to degradation. If, in the unlikely event, corrosion induced dents > 5 volts were to develop in future operating cycles in a tube adjacent to the expanded tube, the structural integrity of the expanded tube TSP ligaments and its impact on limiting TSP displacements will be evaluated. Ifit is determined that additional tubes need to be expanded to maintain the necessary limited TSP displacement, then additional tubes will be expanded. Laboratory testing has shown that tube expansion at TSP intersections that essentially close the tube to TSP crevice have a lower potential for denting than tubes with the normal crevice. In addition, the low temperatures of the plugged expanded tube further reduce the potential for corrosion induced denting. Thus, new corrosion dents at expanded TSPs is even less likely than at an active tube.

However, in the very unlikely event that the periodic inspection of the expanded tubes would identify a dent, the need for expanding additional tubes would also be evaluated.

If this occurs, the NRC will be notified of the inspection findings and the corrective action taken to resolve these findings.

33. With respect to the potential for circumferential cracking in the expanded SG tubes:

a: Discuss the implications of having circumferential cracks at the expansion-transition of l

the expanded tubes. State what inspections will be performed on these tubes prior to, or subsequent to, the expansion process. Discuss the basis for these examinations, if 1

any are to be performed.

RESPONSE

The implications of having circumferential cracks at the expansion transition in the parent material of the expanded tubes has been addressed in the response to RAI Question 21.

The requirements and rationale for the periodic inspections in outages following the implementation of tube expansion have been addressed in the response to RAI Question 22 (see also Attachment B of the July Supplement) and the post-expansion inspection for process verification has been addressed in the response to RAI Question 24. This response addresses the inspection question related to inspections performed prior to the expansion process.

Prior to the expansion process, the normal 100% bobbin inspection will be performed.

Relative to selecting tubes for expansion, the bobbin data is evaluated for dents and the results of this inspection will be used together with historical data in finalizing tubes selected for expansion.

b. State whether the NRC will be notified prior to restart if circumferential cracking is found in the expanded tubes.

SSAMTC1"d5\\itAI,619.WPS 7

M y M.1995

RESPONSE

Reporting requirements are revised to the Technical Specifications with the July Supplement. The reporting requirements include notifying the staff prior to returning the j

steam generators to service (prior to Mode 4) should any circumferential crack-like j

indications or PWSCC be detected at the TSP intersections including any found in the expanded tubes.

I

c. Describe the criteria to be used for determining whether the expansion of the sampling program for expanded tubes is necessary (Refer to Page 12-8 of WCAP-14273).

RESPONSE

Supplemental information on expansion of the sampling program for expanded tubes has been provided in the response to RAI Question 22.

d. Discuss the implications, if any, of missing a circumferential indication at the threshold of detection of existing NDE techniques.

RESPONSE

For circumferential cracking at the TSP expansions to be of concern, the tube must be severed or sufficiently close to being severed such that the small SLB loads could sever the tube. Missing an indication at the threshold of detection would not be a concern for the integrity of the expansions. For the low plugged tube temperatures, it is unlikely that growth rates would lead to severing of the tube between periodic inspections. If circumferential cracking is significant, the periodic inspections will detect the cracking and expansion of the inspection would be implemented.

As an extra measure of conservatism a redundant tube expanded at the most critical areas (refer to Section 8.6 page 8-7 of WCAP-14273).

34. Indicate your schedule for submitting the analysis and listing of TSP intersections excluded from application of the SG tube repair criteria due to potential deformation of these tube locations under combined loss-of-coolant accident loads and seismic loads 1

(LOCA + SSE). (Refer to Page 12-10 of WCAP-14273.) State whether the analyses 1

performed for the previously approved " free span" voltage-based SG tube repair criteria are superseded if the limited TSP displacement repair criteria are implemented. If so, will the cold and hot leg TSP /SG tube intersections differ in that there will not be any SG tube expansion on the cold leg side.

SMICCC11950tAl,619.wl'5 8

July 21,1995

l

RESPONSE

The LOCA + SSE loads and tubes identified for exclusion from application of the tube repair criteria due to potential deformation of these tube locations are independent of the j

application of tube expansion. Thus the analyses performed for the previously approved free span repair criteria are applicable without revision for the limited TSP displacement repair criteria.

Those tubes which will be excluded from application of APC as a result of the possibility of collapse during a LOCA with SSE have been identified in WCAP-14046 Rev. 3 which was submitted to the NRC on June 19, 1995.

35. Describe the basis for the 20 mil (i.e.,0.020 inches) probe centering adjustment to the size of the bobbin probe which must be capable passing through a SG tube to ensure TSP integrity with respect to denting. Your proposed criteria of using a 0.570 inch bobbin probe for a "go/no-go" test appears non-conservative (e.g., 0.664 - 0.066 - 0.020 =

0.579 inches) assuming that the adjustment for the dent size and probe centering devices are accurate. (Refer to Page 12-7 and Section 10-7 of WCAP-14273).

j

RESPONSE

The probe centering adjustment of 20 mils was developed from a laboratory probe. It has since been determined that both ZETEC and Echoram, who provide most of the probes l

used domestically, have probes from which the centering adjustment collapses to the reference probe diameter. Therefore, except for a masonable clearance, no allowance is now considered to be necessary for the probe centering adjustment. Some provisions should be included in the "go/no-go" basis for tolerances on the probe diameter and use of standard probe sizes. Standard probe sizes are typically 10 mils apart and a 0.590" probe can readily be obtained. This provides for a 9 mil allowance (e.g.,0.664 - 0.065 -

j 0.009 = 0.590), which permits use of a standard probe size and includes an adequate margin for probe diameter tolerances. The "non-conservative" margin addressed in the question is the 9 mils provided for use of a standard probe with manufacturing tolerances.

The 9 mil margin would not significantly affect TSP integrity since the analysis leading to j

the 65 mil dent allowance includes considerable conservatisms including equivalent dents j

in the four surrounding tubes, code minimum yield stress of 30 ksi (about 0.003%

probability of occurrence in Westinghouse tubing) and the assumption that exceeding yield will result in cracking of the plate. Therefore, if a probe of 0.590" diameter passes through the tube, the TSP intersection is considered to have adequate integrity for denting j

considerations.

36. With respect to your calculation of the maximum displacement of the TSP under j

postulated accident conditions:

S.\\AIOLU195\\RAI,619.WP5 9

% M.1995

^

~

~

~

~

' ~ ~ ^

H C

1 l

a. On Page 8-1 of WCAP-14273, you state that the relative TSP /SG tube displacement that is ofinterest is the one in which the SG tube and TSP position at the start of the postulated steam line break (SLB) transient is defined as the reference position. You further state that at hot standby, the TSP positions relative to axially oriented SG tube '

cracks inside the TSP are essentially the same as at cold shutdown.. Describe the I

maximum difference that can be expected assuming that the TSPs are free to move.

RESPONSE

The indications are formed at the full power positions of the tubes relative to the TSPs.

If it is assumed that the tubes are free to move relative to the TSPs (i.e., no TSP locking -

i due to crevice deposits), the maximum movement of a tube relative to a TSP between full l

power and hot standby conditions will occur due to differences in thermal growth at an upper TSP / tube intersection adjacent to a tierod. For tubes located away from the -

tierods, the full power, steady state pressure drops across the TSPs slightly displace the l

plates in an upward direction that reduces the differences in relative tube to TSP positions -

at cold shutdown and hot standby conditions. The relativ' displacements are also e

increased between full power and cold shutdown by tubesheet bow at positions away from the stayrods. Since the tierods are attached to the top of the tubesheet, the difference is-i essentially due only to the relative differences in the TSP and TS deflections as a function of their respective support conditions. The differences due to tubesheet bow between full-power and hot standby are minimal (slightly lower secondary pressure at full power than j

at hot standby would cause a modest increase in the relative displacements). The change -

y in relative positions from full power to cold shutdown are of interest in assessing the l

potential extension of a crack outside the TSP at the inspection conditions but do not

-l affect the TSP displacement analyses. The change in positions from full power to hot

-l standby are of particular interest since the SLB displacement analyses assume the indications are within the TSP at hot standby conditions.

l The differential tube to tierod thermal growth (displacement of the tube relative to the l

TSP) was calculated for full power to cold shutdown and for full power to hot standby f

conditions. The results are given in Table 36-1. The tube is displaced below the TSP i

(negative " Delta" in the table) at bc h conditions. For full power to hot standby, the displacements range from 2 mils at the FDB to 64 mils at the top TSP. For full power to cold shutdown, the displacements range from 4 mils at the FDB to 160 mils at the top j

TSP. If the plates were free to move, some indications outside the TSPs would be expected to be found at plants having significant numbers of indications at the upper TSPs (where displacements at cold shutdown exceed ~ 100 mils). At some tube locations, this l

relative displacement would be increased slightly by a decrease in tubesheet bow between l

the two conditions.

)

swocen95utAi_6:9.wr5 10 my2i.i,95

~

~

The FDB and TSP displacements from full power to hot standby conditions of 2 to 64 mils are negligible for tube burst considerations, even if it is assumed that the tubes are free to move relative to the TSPs, and the full power TSP displacements due to pressure drops are ignored. An increase of 64 mils to the TSP displacements with tube expansion would have a negligible contribution to the tube burst probability. For the lower two TSPs, for which downward displacements are calculated for a hot standby SLB, the net displacements would be reduced by thermal growth relative to the calculated TSP displacements for the hot standby position of the tube. The full power TSP displacements due to pressure drops are < 50 mils except for the top TSP which has about 93 mils of upward displacement. If the full power displacements due to pressure were included in the analyses together with assuming the tierods establish the TSP elevation at all tube locations (i.e., algebraically adding pressure drop displacements to the thermal growth displacements), the full power to hot standby displacements for the upper TSPs would be i

even smaller (< 50 mils) than the values given in Table 36-1, further reducing any consideration of the displacement effects on the probability of burst.

Overall, it can be concluded that the WCAP-14273 method of calculating TSP displacements relative to the hot standby or full ower positions is the appropriate I

methodology. For full power displacement analyses, the results are independent of whether or not the TSPs are locked since the indications are formed at the established full power position of the plates. For the hot standby displacement analyses, the WCAP methods are appropriate for the likely locked TSP condition and introduce negligible error when the tubes are free to move.

b. On Page 8-4 of WCAP-14273, you state that it is expected for the Model D4 SGs that flaw indications would be inside the TSP at both cold and hot conditions, independent of whether the TSPs are effectively clamped to the tubes as a result of crevice deposits. However, on Page 12-4 there is a discussion pertaining to SG tubes having low tube to TSP contact forces which indicates that the tube expansion may result in a small change in the hot condition TSP elevation relative to the elevation prior to expansion. Provide your basis for assuming that the TSP intersections in the Braidwood I and Byron 1 model D4 SGs have high tube to TSP contact forces. We note that a qualitative argument with respect to degradation being found somewhat centered with the TSP has been presented in WCAP-14273. However, this argument does not consider the fact that flaw indications below the threshold of detection may be at the edges of the TSP and/or extend slightly outside the TSP as has been observed on a limited number of pulled SG tubes. Ifit cannot be demonstrated that the contact forces are high at the TSPs, discuss the implications for your burst and leakage assessments in terms of the maximum crack length that may be exposed as a result of a postulated SLB.

swesect35uw ours 11 wy u. ms

9

RESPONSE

The basis for the Page 8-4 statement that the flaw indications would be inside the TSP at hot standby conditions is given in the response to Question [a] above. That is, that thermal growth, tubesheet bow and full power TSP displacements are small differences between full power and hot standby conditions, as shown by the small thermal growth differences of Table 36-1. From full power to cold shutdown conditions, the differences are larger than full power to hot standby due to increased thermal growth and tubesheet bow contributions. However, the differences remain modest as shown by the thermal growth contnbution in Table 36-1 (4 to 160 mils). This potential for indications to be outside of the TSP at the cold inspection condition if the TSPs are free to move was recognized prior to agreeing to report indications found outside the TSP. Indications in Model 51 SGs were reviewed at locations where the tube to TSP differences would be large if the plates were free to move and no indications were found outside the TSP. As shown in the following paragraph, the magnitude of the tube to TSP contact forces required to " lock" the tube to the TSP are small for the expected condition that all tubes have a TSP contact force.

The magnitude of the tube to TSP contact force required to essentially " lock" the TSPs can be estimated by calculating the average force per tube required to offset the full power and SLB pressure drops across the plate. If the actual contact forces exceed these values, one would expect minimal tube to TSP displacements except perhaps for a tube adjacent to a fixed TSP support such as the tierods and TSP support bars. The order of magnitude estimates for the full power and SLB conditions are given in Table 36-2. It is seen that an average force per tube of about I lb for the full power pressure drops and about 3 lbs per tube for the SLB pressure drops would equate to the pressure loads and essentially limit TSP displacements to negligible levels. Even if the average SLB loads are increased for a dynamic loading factor, the average force per tube would only be a few pounds per tube. A somewhat larger force would be necessary to " lock" the TSP for differential thermal growth for a tube location next to a tierod but overall forces for locking are small. These are low contact forces and can readily be obtained without denting at the TSPs. For example, Tables 5-9 and 5-11 of EPRI Report NP-7480-L, Volume 1, show tube to TSP pull forces of 80 to 700 lbs for tubes with no denting to

< 5 volt dents. When measured, forces to " break" tubes loose from TSPs have significantly exceeded these values. Although the expected condition is " locked" tube /

TSP intersections, the SLB displacement analyses are essentially independent of this assumption since the tube to TSP position differences between full power and hot standby are small even if the TSPs are free to move.

smmsmi_eie.wes 12 suiy 2i i,95 L

l The statement on Page 12-4 that there can be some change in the hot condition tube to TSP elevation with and without expansion if the contact forces are low is with regard to i

tubes adjacent to an expanded tube. Prior to expansion, small contact forces would likely.

result in the TSP moving with the tubes from the hot to the cold plant condition. When.

j the plant returns to a hot condition, the expanded tube (new tierod), which is not in service, acts to pull the plate down relative to the hotter in service tubes due to differences in thermal growth. If the contact forces are low, the active tube may " break away" from the TSP and result in a the formation of new degradation at an elevation lower than the prior degradation (by the magnitude of Table 36-1 full power to cold shutdown deltas). While these differences are negligible for tube burst considerations, the l

intent of the discussion is to advise that there is a somewhat greater chance with tube expansion of finding an indication under inspection conditions outside the TSP for low contact force conditions.

None of the analyses supporting the tube expansion APC are dependent on the assumption of low or high tube / TSP contact forces at Braidwood-1 and Byron-1. The full power tube to TSP positions are not changed significantly, except perhaps slightly adjacent to an expanded tube, and the calculated SLB TSP displacements relative to the full power positions are.the appropriate values. For a SLB at hot standby, the differences between -

locked and unlocked TSPs are small/ negligible, as shown above, and the analyses are also acceptable. Therefore, no attempt has been specifically made to determine whether the i

Braidwood-1 and Byron-1 tube to TSP contact forces are high or low. In addition, it is intended that the analyses are independent of whether chemical cleaning is applied to the SGs such as to potentially reduce the contact forces. As noted in the question, it is only qualitatively implied that the contact forces are significant since the eddy current and pulled tubes indicate that the degradation tends to be centered ~within the plate.

If the TSPs are free to move, the only impact on the TSP displacements, and the associated leak and burst analyses, would be a slight increase in the TSP displacements for the occurrence of a postulated SLB at hot standby conditions. As previously discussed, the differences would be expected to be less than the full power to hot standby deltas ( 11 to 64 mils for expanded TSPs) of Table 36-1. These differences would have a negligible effect on leakage and burst probability. The largest effects occur only for tubes near tierods and for the upper TSPs where there are few indications. Based on cycle time at hot standby and full power conditions, there is only about a 3% chance that the postulated SLB would occur at hot standby conditions.

secen5mi_m.wr5 13 wy n. ms

m

~

s-lw(

c. If the crevice deposits are the primary reason for holding the TSP in place during hot and cold conditions, state whether a postulated SLB or similar transient will result in U,

" freeing" the TSP (e.g., due to plate rotation, blowdown forces or SG tube / TSP interactions) such that the displacement should account for the position of the TSP -

{

during full power, hot standby, and/or cold shutdown conditions. If the TSP becomes free to' move as a result of a transient, would the location of the maximum TSP displacement be a result of the tubesheet bow and other factors? Specifically, would the maximum exposed crack length no longer correspond to the location of the i

maximum TSP displacement?

i RESPONSE -

i Crevice deposits are the primary reason for holding the TSPs in place relative to the tubes at cold and hot standby conditions. The packing of the crevices with hard deposits is expected to occur under the boiling conditions of full power operation. The reference

" locked" TSP positions are then the normal full power positions without tube to TSP contact forces since the plate positions are established prior to " locking" by deposits. Ifit is then postulated that some additionalload " frees" the TSPs, the plate positions at full power operation would not change. Subsequent cold and hot standby positions of the TSPs would then be the same as that for free to move TSPs as discussed in the Question

[a] response.

If it is postulated that the load to " free" the plates occurs at SLB conditions for previously j

" locked" TSPs, the TSPs would assume positicas expected for free to move TSPs as discussed above.

In all cases involving assumptions of "non-locked", " locked" and " freed" TSPs, there are j

essentially no differences in the calculated TSP disp!acements for a SLB at full power conditions. For a SLB at hot standby conditions, the differences between the calculated displacements and the positions for "non-locked" or " freed" TSPs are expected to be bounded by the full power to hot standby differential displacements of Table 36-1, which are negligible for burst and leakage considerations.

t t

i Overall, it is concluded that the SLB TSP displacement analyses, while strictly based on the " locked" TSP condition, are adequate for the "non-locked" or " freed" TSP conditions with negligible error.

1

d. Provide additional discussion describing how the change in the relative TSP /tubesheet displacements in Table 8-1 of WCAP-14273 were calculated.

i sumccm5uw 6:9.wr5 14 soir 2, i,95 l

1

RESPONSE

Three sets of displacement results are presented in Table 8-1 for each of the load cases.

The first set, the top set in Table 8-1, are the plate displacements relative to their initial installation (zero displacement) position. These displacements are taken directly from the output of the dynamic analysis. The second set of displacements, the middle set in Table 8-1, are the relative plate / tubesheet displacements (i.e. plate - tubesheet, since the tubesheet displacements also represent the tube displacements at the TSP elevation).

These displacements are calculated at each plate dynamic degree of freedom (DOF) by subtracting the tubesheet displacement at a geometric position in line with the plate DOF from the plate displacement. This subtraction process is conducted for each transient time point to arrive at the maximum variation during the transient. The third and final set of displacements, shown at the bottom of Table 8-1, are the change in the relative plate /

tubesheet displacements from the start of the transient. The calculational method is identical as for the second set of results, except that the relative plate / tubesheet displacement at time equal zero is subtracted form each of the transient disp'acen.cnts. It is the final set of displacements that represent the relative plate / tube motion and the potential for uncovering cracks that could potentially exist in the tube inside the TSP.

The algorithm for calculating the relative displacements is as follows:

AD = (D,. 7 - D,. o ),.u - (D, - r - D. o )rm,

Plate Displacement

where, D,.u.

=

Dr m = Tubesheet Displacement T = Time of maximum displacement from dynamic analysis l

1

37. With respect to the burst pressure versus crack length correlations presented in WCAP-14273:

a.

Provide the burst test data supporting Equation 9-1 of WCAP-14273. Describe any data eliminated from the database.

RESPONSE

The data supporting Equation 9-1 of WCAP-14273 are contained in a report prepared by Westinghouse (SG-95-03-010) for the Electric Power Research Institute (EPRI) entitled

" Burst Pressure Correlation for Steam Generator Tubes with Throughwall Axial Cracks,"

February,1995. Since the report contains EPRI proprietary information, EPRI has been contacted and requested to transmit the report to the NRC thorough NEI. Equation 9-1 4

in the WCAP was developed prior to the preparation of the final version of the EPRI report, hence, the database is reduced slightly. The only change in the coefficients in the final fit was a change in the first coefficient from 0.620 to 0.613. The standard s w cw u m m i_ei,.wr5 15 u y n.i995

error of the regression in the final EPRI report is 0.0172 versus 0.0176 used in the WCAP. Comed will use the analysis from the WCAP since the difference between the WCAP and the EPRI report is negligible.

The database used was a compendium of known Westinghouse, B&W, Belgian, and French data, along with data taken from NUREG CR-0718. All of the data used was either from test specimens where no sealing bladder was used or where the sealing bladder was reinforced with a foil shim. None of the data obtained from such tests were knowingly omitted from the database. Test results with a foil shim were reduced by 5%

to account for the strength of the foil. Data from specimens tested with a bladder but without a foil were not used in the correlation. The rationale for the assembly of the database used is contained in the EPRI report.

b.

In EPRI Report NP-6864-L, "PWR Steam Generator Tube Repair Limits:

Technical Support Document for Expansion Zone PWSCC in Roll Transition -

Revision 2," dated August 1993, several burst pressure versus axial crack length correlations were discussed. Describe the data used in the correlation presented in Figure 9-1 of WCAP-14273 as it pertains to the material present in EPRI report NP-6864-L. In your response, discuss the following items specifically, but also discuss any other pertinent factors:

i.

State whether the databases used in support of the various correlations have been combined. If they have not been combined, describe the basis for choosing one correlation over another. Specifically, if your basis is related to the total SG tube break size, state whether the leak rate associated with the burst has been related to safety consequences.

ii.

State how the chosen correlation compares to other correlations. Specifically, is it the most limiting correlation which could be used to assess the SG tube burst probability. State whether other data besides that in EPRI report NP-6864-L exists.

RESPONSE

i As noted in the response to RAI 37.a, an analysis of available burst data was performed to provide a specific report on the burst pressure of throughwall axial cracks to EPRI.

i The database used in the EPRI report NP-6864-L was a combination of previously available data and additional data obtained through January of 1995. The basis for i

determining whether or not burst has occurred is the development of afishmouth opening plus tearing at the ends of the crack. No assessment of the area of the opening was performed, nor is an assessment of the area used in leak rate calculations, i.e., a rupture is generally assumed to be a large leak which could potentially range from a fishmouth t

opening to a larger break approaching a guillotine break. A comparison of the final correlation for EPRI to prior correlations is provided in the EPRI report. The correlation presented is the most limiting correlation that could be used to assess SG tube burst probability based on the results of burst tests performed in Belgium at a fossil power plant. The basis for choosing one correlation over another is the goodness of the swcm95mi_6:9.wr5 16 wy ri,1995 I

fit of the regression equation (supported by the analysis of the residuals), and consideration of the number of terms in the regression equation.

The most limiting available, but non-representative, database from which a correlation could be developed is contained in Westinghouse WCAP-12522. The data therein were obtained from tests in which the ir. side of the tube was lined with a plastic bladder and burst was defined as the pressure at which the crack opened to the extent that the bladder extruded to the point that the bladder ruptured. Information presented by P.

Hernalsteen, "The Influence of Testing Conditions on Burst Pressure Assessment for Inconel Tubing," International Journal of Pressure Vessels and Piping, Volume 52, pages 41-57,1992, clearly demonstrates that the true burst pressure is not properly represented by results from tests performed employing only a plastic lining.

c.

Provide additional discussion explaining Figure 9-2 of WCAP-14273 with respect to the definition of a small gap and a large gap, including the use of a 28 mil clearance.

RESPONSE

The data points labeled small gap were obtained from tests in which the diametral clearance between a tube and TSP was 11 to 13 mils. The points labeled large gap were from tests in which the gap ranged from 19 to 23 mils. The test results labeled as having 28 mils of clearance were performed for EPRI in late 1993 to demonstrate that a 3/4" long throughwall indication in a TSP hole with significant tube to TSP clearance would exhibit a burst pressure well in excess of the RG 1.121 requirements. The data were included on Figure 9-2 for information purposes. The as-built cicarances were 28 mils as delineated on the figure. The trend of the other ' data indicates that the throughwall burst correlation would likely be followed for 0.7" long throughwall indications with exposures of up to 0.1" with tube / TSP clearances in the vicinity of 23 mils. The range of possible maximum TSP clearances for drilled hole TSP steam generators is 23 mils. The 23 mil value was obtained by applying a 95% confidence level to the upper diameter limit of a drilled support plate hole and the 95% confidence level of the minimum diameter of the tubing as discussed in Section 9.3 of WCAP-14273.

d.

You state on Page 9-4 of WCAP-14273 that the larger tube to TSP clearance over the crack exposure length of interest has the effect of reducing the burst pressure predicted by Equation 9-1 by 425 pounds per square inch (psi). Discuss the amount of data supporting this assumption. State whether this is an average reduction or maximum reduction and whether this reduction would be larger or smaller at more limited TSP displacements (i.e.,0.1 inch). State whether additional testing is either being performed or planned to determine the appropriateness of the reduction.

swcmssuw si,.wr5 17 uiyri.1995

RESPONSE

The data used to obtain the adjustment factor are illustrated as solid black circles on Figure 9-2. A total of twelve (12) data points were available for the analysis, however, the adjustment factor was obtained as the average for the six (6) data points located farthest below the predicted burst curve. Since the form of the adjustment was chosen to be a constant, the average factor is the same as would be obtained from a least squares solution. The use of a censored database is considered to be conservative since the range of interest is for exposures of about 0.1 to 0.3", i.e., the actual reduction in strength would be expected to be smaller than that used for more limited TSP displacements.

The data illustrate that no adjustment is really necessary for exposures in the range of

~0.1 to ~0.2" and that only a slight adjustment is necessary for exposures of ~0.2 to

~0.3". Thus, the selection of the data used for the determination of the adjustment factor essentially bounds the data in the range of interest. Since the use of an adjustment factor of 425 psi is conservative, no further testing is planned.

38, With respect to the probability of burst calculations presented in WCAP-14273:

a.

On Page 8-1 of WCAP-14273, you state that the number of SG tubes with a given relative SG tube / TSP displacement amplitude is determined and provided as input to the tube burst analysis. State whether your tube burst analysis considers the displacement for each group of tubes or whether the tube burst analysis assumes that each TSP intersection is exposed by the maximum TSP displacement. It appears from Section 11 of WCAP-14273 that both types of calculations have been performed but that the " official" calculation for determining the burst probability involves assuming all TSP intersections are exposed by the maximum TSP displacement.

RESPONSE

The probability of burst values reported in Table 11-1 of WCAP-14273 reflect both considerations, i.e., of grouping of indications by displacement and of all intersections being subjected to the maximum displacement. The upper portion of the table, identified as " Uniform TSP Displacements...", considers all intersections to be cracked and all cracks to be exposed by the maximum TSP displacement. The values reported in the lower portion of the table reflect consideration of displacements as a function of the l

planar location of the indications. The " official", or design objective, calculation results provided in Table Il-2 are based on assuming all indications to be exposed by the i

maximum displacement of the TSPs.

i sere 95wu_6i9.wr5 18 Joiy 21,1995

b.

In Figure 9-3 of WCAP-14273, a plot of the probability of burst for free span cracks and for exposed cracks with large clearances was presented. Describe how the effect in the reduction in burst pressure as a result of large clearances was accounted for, if at all, in these calculations. Shouldn't the probability of burst for a crack exposed 0.7-inches outside the TSP approach the probability of burst for a 0.7-inch free span crack? State how the SG tube material properties were accounted for in these calculations.

RESPONSE

Section 9.5 of WCAP-14273 provides a detail discussion of the model employed for the calculation of the probability of burst (PoB) accounting for the reduction in burst pressure due to the clearance between the tube and TSP. The PoB was calculated as the probability of obtaining a random Student's t variate of the magnitude as calculated using Equation 9.10. These results are considered to be bounding because of the conservative approach used to develop the adjustment factor of 0.425 ksi. The reason that the calculated probability of burst for a large clearance indication remains significantly above that of a free span indication as the exposed length increases to near the length of the indication is an artifact of the selection of a constant adjustment factor for the TSP indications instead of a graded adjustment factor that would diminish to zero when the exposed length equals the total length. The burst correlation with the clearance adjustment is intended to support assessments of tube expansion requirements with TSP displacements < 0.5 inch. There is no need to refine the correlation for larger indications.

Consideration of the material properties is discussed in detail in Section 9.3 of the WCAP. Specifically, the variance of the burst pressure is estimated using Equation 9.6 to account for variation about the regression curve and variation in the material properties. Equation 9.6 contains a typographical error in that af should read S at f

all locations in the equation, thus V(S ) is the variance of the flow stress.

f c.

On Page 9-5 of WCAP-14273, you state that the distribution of burst pressure does not follow a Student's t-distribution but that a comparison of Monte Carlo results with individual upper bound 95 % confidence limits indicate some conservatism in the results. Describe how this Monte Carlo analysis was performed, including a discussion of how you accounted for the distribution of SG material properties.

State whether similar Monte Carlo analyses have been performed for the limited TSP displacement case. If so, describe the analyses and the results.

RESPONSE

Equation 9.5 of WCAP-14273 describes the relation between the burst pressure, P, and the normalized burst pressure, Pu, and the flow stress of the material, S,.

For the simulation of the burst pressure of a single indication, prediction bounds about the mean of the normalized burst curve and about the mean of the-distribution of flow stresses are each simulated by assuming a Student's t distribution. Because of the large number of degrees of freedom in the database, the Student's t distributions are approximately swoccase_mwn 19 wyn m5

O normal, thus, the third moment, i.e., skew, of the burst pressure distribution, M, is 3

related to the distributions of the normalized burst pressure and the flow stress as, Af (P,) oc P'S' V(P ) V(S ),

(1) 3 f

y f

where V indicates the variance of the respective variable. Since all of the terms comprising the skew are positive, the skew is likewise positive. Hence, the distribution of the burst pressures will be skewed right, i.e., the tail.will be higher / longer for high burst pressures and lower / shorter for low burst pressures. This means that deterministic calculations performed based on the assumption of a symmetric distribution will predict conservative, higher probabilities of burst for short indications than would be expected from Monte Carlo simulations of the burst pressure. Figures 13 and 14 of the throughwall burst correlation report (SG-95-03-010) prepared for EPRI illustrate this effect. For 3/4" tubes, the deterministic and Monte Carlo predictions of the probability of burst of throughwall cracks are approximately the same for a crack length of 0.7".

For a crack length of 0.5", the deterministic estimate of the probability of burst is approximately 4 times that obtained from the Monte Carlo simulation.

A comparison of deterministic calculation and Monte Carlo simulation results of the limited TSP displacement case would be expected to be similar to those obtained for the throughwall crack comparison, since the reduction in burst pressure is accounted for as a constant adjustment to the throughwall crack burst correlation. The adjustment is considered to be conservative, hence, specific simulation of the TSP displacement case is not necessary. Monte Carlo analyses would only result in burst probabilities smaller than the 10~ value for the maximum TSP displacement of 0.1" obtained for the limiting SG tube applied to all TSP intersections and further assuming all TSP intersections expose a 0.1" throughwall crack.

d.

On Page 1.1 of WCAP-14273, you state that the SG tube burst probability assessment is consistent with the previously published draft generic letter on voltage-based tube repair criteria. However, your calculation of the SG tube burst probability does not include a consideration of the parameter uncertainty as the draft generic letter states that it should. Your discussion in WCAP-14273 ic, which you justify not performing the probability of burst calculation considering the uncertainties in the regression parameters, is primarily qualitative. The staff believes that the issue of SG tube burst probability could be more definitively addressed by directly accounting for parametric uncertainty in this calculation.

Provide your assessment (i.e., a calculation which includes parametric uncertainty) of the probability of burst for an indication extending outside a TSP by an amount equal to the maximum TSP displacement.

sw c wc95m ui9 m3 20 my 2i, 995

RESPONSE

The calculation of the probability of burst of an indication as a function of the bobbin amplitude of the indication is performed in accordance with the guidelines of the draft generic letter. Parameter uncertainties are explicitly included in the Monte Carlo simulations "using the correlation between burst pressure and voltage." The details of the methods employed in performing the calculations are provided in the Westinghouse methods report (WCAP-14277).

For a single indication extending outside of a TSP intersection, the deterministic calculation includes consideration of the uncertainties in the parameters. The effective variance of a predicted normalized burst pressure about the regression curve is V(P ) = s { g + {f9 }r[grF)~' {fo }

(2) 2 y

where s is the standard error of the regression, {fo} is the coefficient derivative vector and [F F]'8 is the normalized covariance matrix. The term containing the covariance T

matrix accounts for the variance of the coefficients of the regression equation. This is used in equation 9.6 of WCAP-14273 to obtain the variance of the burst pressure.

The discussion on page 9-5 of WCAP-14273 that refern to omission of consideration of uncertainties in the parameters, is with respect to the effect on the combined probability of burst as afunction of crack length of all of the indications in the SG, not on the probability of burst of any one indication. For this calculation, it is considered that the use of the deterministic estimate of the probability of burst is justified because the individual calculated probabilities of burst are an order of magnitude higher than obtained from Monte Carlo simulations, the number of intersections with indications is overestimated by at least an order of magnitude, and the assumption that the maximum TSP displacement occurs at all intersections in the SG hot leg likely overestimates the number of indications exposed to that displacement by two orders of magnitude. Thus, the effect of simulating uncertainties in the parameters would have to be sufficient to result in a change in the probability of burst of about four orders of magnitude to increase it above the values stated. The simple fact is that the probability of burst of an indication which is exposed on the order of 0.1 to 0.3" is extremely small. The development of a Monte Carlo code to simulate the distribution ofindications over the TSPs, the distribution of indications at a TSP over the area of the TSP, and the distribution of indication lengths in order to simulate a distribution of exposed lengths, in addition to simulating the parameters of the correlation and the material properties is not justified in light of the very low probabilities obtained from the deterministic estimation.

e.

On Page 9-6 of WCAP-14273, you state that since no indications at the FDB have been found at either Braidwood 1 or Byron 1, the FDB is not included in the assessment of the SG tube burst probability. Since your proposed voltage-based i

repair limits will not be applied to the FDB during this cycle, it is reasonable to not perform a burst probability analysis at this location. However, state whether your burst probability calculation will combine the contribution of the burst probabilities 1

swesece95mi_6iv.wr5 21 u y 2i i995

E

]

for cold leg indications and hot leg indications. If these two separate SG tube burst probabilities are to be combined, explain how this will be done. (Depending on the j

final analysis, the staff recognizes that the hot leg contribution to the burst probability analysis may be negligible. However, this has not yet been j

demonstrated.)

j i

RESPONSE

The analyses of Section 11 of WCAP-14273 using the Section 9 methods as discussed above demonstrate that, with tube expansion, the hot leg burst probability is negligibly small and < 10'80 As noted in the response given above, this analysis assumes that all hot leg TSP indications have throughwall indications and maximum TSP displacements.

Since these extremely conservative assumptions bound all conceptual degradation and the i

resulting burst probability remains negligible, it is not necessary to calculate the hot leg burst probability on a cycle dependent basis. The negligible burst probability obtained envelopes the potential hot leg burst probability for any operating cycle. Therefore, the hot leg burst probability would not be calculated for operating cycles.

It is only necessary to calculate the cold leg burst probability on a cycle dependent basis.

The analyses will be consistent with the NRC draft generic letter and apply WCAP-

)

14277 methodology.

f.

Discuss whether the potential contribution to the burst probability due to circumferential failure at the TSPs should be included. In responding, discuss this in terms of both short-term and long-term implementation of this voltage-based repair criteria.

RESPONSE

As the repair limit approaches the full APC repair limit (about 10 to 15 volts) permitted with tube expansion based on the structural limit (about 35 volts) for axial tensile tearing, it may become necessary to consider burst probability analyses as discussed below. For the conservative 3.0 volt repair limit requested for Braidwood-l and Byron-1, it is not necessary to calculate the tube burst probability.

1 The primary loading mechanisms which could lead to circumferential failure and axial separation of a tube, if cellular corrosion is present, is the end cap pressure load in the tube itself. A second potential loading mechanism is the load applied by the TSP (s) if the tube is locked at one or more TSPs. However, when TSP locking occurs on active I

tubes, a large number of tubes are involved and the tubes share the load such that the resulting loads on any one tube are negligible. Consequently, only the end cap loading needs to be considered in the burst probability for axial separation of the tube. For a SLB pressure differential of 2560 psi, the end cap load is 886 lbs for 3/4 inch diameter tubing. This compares to an axial load capability for undegraded tubing of 14,000 to 15,000 lbs (see, for example, Plant P-1 data of Table 41-1 in the response to Question 41). This magnitude comparison demonstrates the large margins for small indications.

swcweis5mi_6 9.wr5 22 Juir 21.1995

[

i Until the voltage repair limit is much higher than the 3.0 volt limit requested for Braidwood-l and Byron-1, the burst probability for axial loading can be ignored. This can be shown by calculating the probability of burst as a function of voltage for axial tensile rupture. The correlation for the tensile load capability of WCAP-14273 is updated in the response to Question 41f and this updated correlation is applied for this analysis. The burst probability for an axial tensile burst ranges from about 3x10-5 at 10 4

volts to about 10 at 30 volts. These burst probabilities are about two orders of magnitude lower than obtained for axial free span burst at 10 volts and even'a lower fraction at 30 volts. With a 3 volt repair limit, the likelihood of indications in the 10 to 15 volt range is small and the resulting indications would have an insignificant axial tensile burst probability. Consequently, there should be no need to calculate burst probability until the repair limit approaches about 10 volts.

39. As a result of the on-going discussions between Comed and NRC staff, several models for predicting leakage under postulated accident conditions have been developed.

Provide a detailed description of the methodology that is currently being proposed for calculating the primary-to-secondary leakage under postulated accident conditions (i.e.,

how the SG tube leakage from the hot leg side will be calculated and how it will be combined with the SG tube leakage from the cold leg side and FDB baffles).

RESPONSE

The SG tube leakage from the hot leg side is calculated using a modification of the free span leakage methodology described in the Westinghouse methods report, WCAP-14277, which is in concert with the guidelines provided in the draft Generic Letter (95-XX), as a basis. This method involves Monte Carlo simulation of the distribution of indications, uncertainties in the measurement of the indications, the growth of the indications, etc. A modification to the methodology used for the calculations has been effected which increases the leak rate for hot leg indications which are simulated as being restricted from bursting (designated as IRBs). If the simulation of the burst pressure results in a value which is less than the differential pressure during a postulated SLB, indicating a probability of burst (POB) of one instead of zero, the probability of leak (POL) of the indication is set to unity and the leak rate from that indication is assumed to be the bounding IRB leakrate, of 5.5 gpm'. All indications simulated to be IRBs are assigned the bounding leak rate regardless of the bobbin amplitude of the indication. If the indication is predicted to not burst, the probability of leak is found from the logistic correlation of the POL to the common logarithm of the bobbin amplitude. A random uniform deviate is then generated to determine if 'he indication leaks, i.e., the POL is j

either zero or one. If the indication leaks, the leak rate is calculated from the correlation of the common logarithm of the leak rate to the common logarithm of the bobbin amplitude. The simulation is conservative in that the estimated leak rate obtained from i

the correlation is used even ifit exceeds the bounding leak rate for an IRB. Thus, the total leak rate obtained from one simulation of the SG, with n indications in service, is given by, 8 This is an upper bound value determined from leak testing of indications which were deformed by internal pressurization to simulate an attempted burst within the TSP.

swesecasmi_6i,.wr5 23 u y n,i,95

i

[POB 5.5 + (1 - POB) POL 10*** M *'M }

Oso 7,a

=

where #3, #4, and #3 are the simulated population intercept, slope, and standard error of the leak rate correlation, and Z; is a random normal deviate. The estimated EOC total leak rate for a SG is calculated as a 95% confidence bound on 95% of the possible population of leak rates based on the results of many, i.e., greater than 100,000, Monte Carlo simulations of the total leak rate.

The APC repair limits of the NRC Generic Letter apply to the cold leg indications and the FDB indications if the APC is implemented at the FDB. In the July Supplement, Comed stated that APC will not be applied to the FDB. The repair limits for the hot leg TSPs with tube expansion do not apply for the cold leg indications. The free span leakage methodology of the generic letter, as implemented in WCAP-14277, applies to these indications. The leakage calculated for cold leg indications is added to the leakage calculated for the hot leg TSP indications and the total leakage is compared to the allowable limits. To minimize the computational effort when there are only a modest number of cold leg indications, it is acceptable to include the cold leg indications in the hot leg leakage analyses and directly calculate the combined leakage. This inclusion in the hot leg analyses is a conservative option since the leak rate including the IRB analyses is equal to or slightly larger than the free span leakage.

40. Explain the statement on page 12-12 of WCAP-14273 which states, in part, that "...the SLB leak rate shall be compared to the Braidwood-l or Byron-1 allowable limits as given in the Technical Specifications and potentially modified by administrative controls,..."

RESPONSE

The statement in the WCAP is intended to provide some' outage dependent flexibility in establishing site allowable primary to secondary SLB leakage limits. If at a given inspection, the projected EOC SLB leakage would exceed the allowable limits as given in the Technical Specifications, a reduction in reactor coolant activity limits can be implemented for the associated fuel cycle to increase the allowable leakage limit. In the July 7 Supplement, Comed conservatively reduced the Technical Specification Dose Equivalent Iodine-131 limit from 1.0 pCi/gm to 0.35 pCi/gm for Byron-1 and Braidwood-1. This results in an increase of the site allowable SLB leakrate from 12.8 gpm to 36.5 gpm (Byron-1) and from 9.4 gpm to 26.8 gpm (Braidwood-1).

41. With respect to the basis for voltage limits to prevent axial tensile serving of a SG tube, respond to the following staff requests:

a.

The bobbin coil voltage is being used to assess the potential for a SG tube to fail circumferentially. Since the bobbin coil is relatively insensitive to SG tube circumferentially oriented cracking, an inherent assumption in the methodology is that there is a significant axial component in the SG tube flaw such that it can be detected with the bobbin coil. Describe the basis for this assumption.

SSAIWXT95\\RAI,619.WP5 24 uiy ri,i995

RESPONSE

The degradation morphology that has been found in pulled tubes for ODSCC at TSP intersections is dominantly axial SCC with varying extent of cellular patches including the absence of cellular corrosion. Cellular corrosion is a combination of axial and oblique angle cracks that form small cells of undegraded tubing within the crack pattern.

The patterns of cellular corrosion have been characterized by radial grinds from the tube surface through the tube wall. The axial cracks are consistently deeper than the oblique cracks. That is, as the radial grinds progress through the wall, the crack pattern changes from cellular to multiple axial microcracks with the oblique cracks typically less than 50% to 60% of the tube wall when the axial crack are near throughwall. The oblique cracks in the cellular pattern are typically < 90* (i.e., not circumferential) and the bobbin coil voltage responds to both the axial and oblique cracks.

It can be noted that the bobbin voltage responds primarily to the deepest and longest axial cracks. In correlating bobbin voltage with the axial tensile capability of the tube, it is assumed that the cellular pattern will increase in circumferential extent in some close proportion to the length and depth of the dominant axial cracks. Since this is not generally the case, it can be expected that the spread in the correlation (range of tensile force capability at a given voltage) will be significant and greater than that found for the axial burst pressure correlation with voltage. Indications,vith negligible cellular corrosion would have very high tensile force capability at a given voltage level. To minimize obtaining high tensile capability at high voltages when cellular patches are small or non-existent, a data exclusion criterion will be developed by EPRI to exclude indications with negligible cellular corrosion from the tensile correlation. While this data exclusion criterion has not yet been developed, it is expected that a minimum azimuthal cellular patch involvement would be required for the indication to be included in the tensile force correlation such as requiring cellular patches to encompass a fraction of the tube circumference. Cellular corrosion can generally be seen with modest magnification on the OD of the specimen following the axial burst test since the pressure tends to expand the tube diameter and increase the visibility of the tube degradation.

The absence of cellular corrosion to eliminate the tensile test can be identified in this manner although the final criterion for inclusion in the tensile force correlation would be based on metallographic specimens.

b.

On Page 9-11 of WCAP-14273, you state, in part, that significant intergranular attack (IGA) depths have not been found at TSP intersections in pulled SG tubes which have not been plugged for at least two years prior to the SG tube pull. You further state that the SG tube structural limit should be based on cellular corrosion.

State how many specimens are in the voltage-based repair limit database, and which ones, were plugged prior to the SG tube pull. Discuss the amount of IGA found in these specimens. Discuss any restrictions that need to be placed on recovering (i.e., unplugging and placing back in service) previously plugged tubes based on the current proposal that the structural limit should be based on cellular corrosion.

Discuss the process for verifying that significant IGA will not occur at TSP elevations for the remaining lifetime of the Byron 1 and Braidwood 1 SGs.

swecet95mi 6i9.wr5 25 wy 2i. i,95

RESPONSE

The number of tubes in the EPRI database that were plugged prior to the tube pull is very small. A detailed review of the data for previously plugged tubes was not conducted as this information is not readily available. In general, however, previously plugged tubes are not pulled due to the additional plug removal options and uncertainty on the degradation. Two tubes are known to have been previously plugged and the total number is not expected to be significantly different. These two tubes resulted in four indications that are included in the database. Tube R12C8, TSP 1 from Plant L had reported IGA depths up to about 60% and IGA involvement over about 39% of the tube cross section. This indication could have had cellular corrosion that was called IGA but radial grinds were not performed to characterize the corrosion. The second previously plugged indication (Tube R21C22, TSP 1 from Plant A-2) was a single axial crack that had IGA width of about 20 mils at each crack face. It is speculated that the IGA grew at the crack face after the tube was plugged. The pattern of the IGA from Tube R12C8, TSP 1 is consistent with this assumption although the large number of axial indications precluded judgement on the formation of the IGA.

No restrictions need to be placed on recovering previously plugged tubes other than that the bobbin voltages measured after tube plugging must be less than the applicable repair limit. Voltages measured at the time of tube plugging are not an acceptable basis for returning the indications to service. If significant IGA were to occur in the plugged tube condition, it can be expected that the voltage would increase significantly and likely exceed the applicable repair limit. With the tube expansion based APC, consistency with the database would be based on the tensile force and leak rate correlations.

If significant IGA is found at pulled tube TSP intersections in the future, the indications would be included in the database. The recommendation that IGA specimens not be included in the tensile force correlation is meant to exclude laboratory specimens that have been prepared to obtain IGA. The correlation will include all pulled tube morphologies. In this case, it is noted that only significant cellular corrosion has been found to date with the exception of Plant L tube R12C8. The latter tube has been included in the correlation of tensile force capability given in Figure 9-9 of WCAP-l 14273. The process to be followed by Braidwood-l and Byron-1 to confirm that tube morphologies are acceptable is to pull tubes based on the EPRI tube pull guidelines, if approved by the NRC, or the NRC generic letter guidelines on the frequency of pulling tubes.

c.

The database supporting your proposed SG tube structural limit to prevent axial tensile tearing contains a limited number of specimens. In addition, only 9 specimens from plant E-4 involved measuring the force required to sever the tube.

j Describe the program you plan to implement to supplement this database along with the schedule for this implementation. Describe any adjustments made to the data (i.e., voltage adjustments to account for calibration differences and the use of a transfer standard). Describe how the SG tube morphology of the plant E-4 degradation corresponds to what has been observed in the Braidwood I and Byron 1 SGs. It appears that SG tube specimens from nuclear power plant SGS with both 3/4-inch and 7/8-inch tubing have been combined. Provide the basis for combining swome95mi_sie.wr5 26 w y n.i995

i these databases in light of the fact that there currently exists separate databases for the free span correlations for 3/4-inch and 7/8-inch SG tubes.

RESPONSE

The program to supplement the tensile force correlation database is to include the tensile force test on pulled tube intersections as an addition to the axial burst tests that support the EPRI free span burst correlation. Guidance on performing the tensile force test will be included in the EPRI pulled tube examination guidelines. As noted in the response to part [a] above, a criterion based on the degree of cellular corrosion will be developed to identify specimens for the tensile test. In the longer term, the tensile force database will be based only on specimens for which the axial tensile force test was performed. In the interim, data points based on residual, undegraded cross sections of the tube, such as applied for the Braidwood-1 and Byron-1 pulled tubes, will be used to supplement the tensile test data. Schedularly, the plan to enhance the database follows the tube pull frequencies of the EPRI or NRC generic letter guidelines. Sufficiently high voltage indications to add significantly to the database are likely to be obtained only after the repair limit is increased to three or more volts.

For the specimens included (laboratory 1GA not included) in the tensile force correlation, the only specimens that required a voltage adjustment are the Belgian Plant E-4 data.

The correlation applied to adjust the Belgian voltages to the IPC/APC calibrations is the same as applied for the axial burst data from Plant E-4 in the EPRI database and is given as Figure 3-2 in EPRI Report NP-7480, Volume 2. The correlation includes the cross-calibration of the Plant E-4 field standard to the reference laboratory standard.

The tube morphology of the Plant E-4 degradation has been found to be dominantly axial SCC with significant cellular corrosion. In general, the extent of cellular corrosion has been found to be greater than that found in the Braidwood-l and Byron-1.

Data from both 3/4" and 7/8" diameter tubes have been combined to develop the tensile force correlation although the data is dominantly 3/4" data (Plant L data is only 7/8" data in WCAP-14273). The combined data have been used to enhance the database although it is recognized that the voltages are not directly relatable (3/4" voltages need to be increased to improve but not resolve all voltage differences). Combining the data tends to increase the spread of the data, increasing the uncertainties and tending to decrease the lower 95% confidence level used to define the structural limit. It is believed that a large spread in the data is inherent to this correlation (see discussion in part [a] response) and the data can be acceptably combined, at least until there is sufficient data to consider separate correlations for the two tube sizes. For the EPRI axial burst correlation, sufficient data exists for 7/8" and 3/4" tubing to statistically conclude that they from two populations for the axial burst correlation. Due to the limited data for tensile tearing, a similar determination cannot be made at this time.

Therefore, all data is combined for establishing the tensile force correlation.

swecmsmi_onwr5 27 wy n, ms

4 d.

Describe how the tensile tests were performed.

RESPONSE

The axial tensile tests are performed following the axial burst testing of the specimens.

The following describes the tensile tests performed on Plant P-1 pulled tubes (performed since WCAP-14273 - see Table 41-1). The burst tested specimens are typically about 10 inch long sections of steam generator tubing with the TSP crevice regions centered in the sections (the location of the axial burst opening) and with Swagelok fittings attached to each end of the sections. Simply pulling a burst tested specimen using its attached Swagelok fittings would not work, as the local stresses associated with the Swagelok ferrule would be the source of the tensile fracture. In order to bypass the stresses concentrated at the ferrules, the Swagelok fittings are welded onto the tube above the ferrules using a buttering type weld with a nonuniform weld front that would diffuse local stresses at the weld front. A tensile test gripper mandrel (metal plug) small enough to pass through the compressed Swagelok ferrule regions of the tube is also utilized.

The burst tested specimens are then tensile tested following guidelines in Section 6.9.1 and Figure 11 of ASTM E8. The snug-fitting metal plugs are inserted far enough into the ends of the specimens to permit the testing machine jaws to grip the specimen properly. The plugs do not extend into the gage length portion of the specimens, which is 4 inches long with the burst opening centered in the gage lengths. The specimens are then pulled at a crosshead speed of 0.05 inch per minute and the load to failure recorded on the load-time recorder chart. Only the loads to failure from the tensile tests are considered meaningful.

e.

Table 9-3 in WCAP-14273 appears to be incomplete. For example, there is no plant E-4 pulled tube data as indicated in Figure 9-9 and 9-10. Provide all data used in the development of your proposed repair limit along with a discussion of any assumptions used in determining the residual cross-sectional area.

RESPONSE

Replacement Figure 41-1 and Table 41-1 are provided that are consistent in content and that clarify the nomenclature of Figure 9-9 and Table 9-3 of WCAP-14273. For plant E-4, all tubes represented are pulled tubes; however, for the tubes identified as " Tensile",

the residual cross section (RCS) was determined based on the pull force for rupture and the measured material tensile properties (i.e., rupture force divided by the ultimate tensile strength times the undegraded tube cross sectional area), and for the tubes identified as " Met.", the RCS was determined during destructive examination by measuring the undegraded fraction of the tube cross section. It should be noted that the ultimate tensile strength is used in this conversion while the smaller flow stress is used to develop the structural limit (see question [f] response). At low levels of degradation, the rupture strength is governed by the ultimate tensile strength while at larger degradation, the rupture strength is governed by the flow stress. Thus for the larger Plant E-4 voltage indications, it would be appropriate to use the flow stress to obtain the RCS, which would increase the RCS by about 35%. The conservatism of using the ultimate strength to obtain RCS from the rupture force is being retained pending obtaining a larger data base to develop the correlation.

swocce9smi_6:9.wrs 28 ur 2i. i995

Additional data are included in Table 41-1 as obtained from recent Plant P-1 (7/8" diameter tubing) pulled tubes. The pull load to fracture the tube (methods described in question [d] response) was determined for the TSP indications and for an undegraded free span section of tubing. The RCS was also obtained directly from the metallographic data for the tensile rupture, transverse section of the tube by measuring the undegraded tube area. This provided a comparison of the actual load capability of the tube with the indicated capability based on the measured RCS. The actual load to fracture was within about 5% of the indicated load based on the measured RCS for three of the four measurements with the measured RCS overestimating the tensile load capability by about 15% for R22C38, TSP 2. This suggests that the metallurgical examination tends to provide adequate estimates (with some judgement on the appropriate non-axial depths that affect the tensile rupture capability) of the RCS for degraded tubes, although the preferred method is to measure the tensile rupture force for both the degraded tube section and an undegraded section of the same tube. The latter method defining the RCS as the ratio of the degraded to undegraded (freespan) rupture forces is applied in Table 41-1 to obtain the Plant P-1 residual cross section. This method is typically within 1%

to 2% of the ratio of the rupture force to the ultimate tensile strength times the tube cross sectional area as applied for the Plant E-4 data. For Plant P-1, the tensile strength is appropriate for converting the rupture force to RCS since the degradation is shallow for the pulled tubes. However, for large indications, use of the flow stress leads to better agreement with test result then use of the tensile strength. The use of the ratio of undegraded to degraded rupture forces eliminates the need to apply a judgement on use of ultimate or flow stress to convert the degraded rupture force to the RCS. Application of the ratio to the tensile strength eliminates the need to perform the free span rupture test but requires judgement on the use of ultimate or flow stress to obtain the RCS.

For the Byron and Braidwood tubes, the degradation cross section area was obtained by measuring the circumferential extent of the degradation on the pulled tube, metallographic transverse sections. The angle and a conservative average depth of corrosion were estimated and converted to residual cross section by averaging the depth over 360*.

For plant L, R12C8, the available circumferential profile of the degradation depth was utilized to develop the arc-weighted average undegraded section of the tube as the RCS for the tube. For the other available pulled tubes from plant L, uniform depth of penetration at the mean crack depth for multiple cracks around the tube circumference was assumed to calculate the RCS since the tube examination did not provide sufficient detail for a more accurate RCS estimate. This is conservative since the undegraded tubing between cracks is ignored, and the calculated RCS is less than the actual RCS.

For the IGA specimens, the original Table 9-3, and Figures 9-9 and 9-10 included an adder of 5% in the depth calculation. Since these specimens were destructively examined, the depth of the degradation is known. Therefore, the updated Table 9-3 and Figure 9-9 include the corrected values without the 5% depth uncertainty adder. For the EPRI Library specimens, the depth of penetration is based on the NDE determine <1 values plus a 5% adder to account for uncertainties. For these IGA specimens, uniform degradation around the circumference at the values shown was assumed in the calculation of the RCS and the RCS is adequately obtained as one minus the percent depth.

swcm95uw m.wr5 29 m y n i995

f.

Describe how the minimum required residual cross-sectional area was calculated (i.e., the structural limit). If lateral constraint was not assumed, describe the effect on the minimum required residual cross-sectional area and the resultant structural limit.

RESPONSE

With tube expansion, the structural limit is based on the potential for an axial tensile rupture of the tube at the location of the TSP degradation. With oblique angle degradation, such as the cellular corrosion found on some pulled tubes, the i.mlication can rupture by axial tensile tearing since an axial burst of the indication is prevented by the TSP constraint. The minimum required residual cross section (RCS) for the structural limit is based on the Reg. Guide 1.121 guideline for 3APuo as described below. The constraint of the TSP prevents an axial burst of the tube as described in WCAP-14273 and also tends to increase the burst pressure for a bending instability burst of a circumferential crack within the TSP. The bending instability mode of burst applies to long throughwall circumferential cracks (up to about 270* TW), above which the area reduction is sufficiently large that a tensile rupture would occur. Uniform or variable depth cracks would rupture by tensile loading. The lateral constraint of the TSPs also increases the burst pressure for a circumferential burst from bending instability to equal to or higher than that for axial tensile tearing of the remaining tube area. Thus, the appropriate structural limit is based on a tensile rupture which is proportional to the residual cross section of undegraded tubing.

Per Reg. Guide 1.121, the axial load on the tube results from the 3APuo pressure force acting on the inside area of the tube. The residual cross section is then obtained from the following:

/" =.F" = 3AP" A, RCS A*

f A,

a A*

RCS =3AP g

w l

tw where:

4"

= Flow stress at lower tolerance limit (LTL) properties APuo

= Primary to secondary pressure differential at normal operating conditions Am

= Area of the ID of the tube A,

= Area of the undegraded tube wall RCSst.

= Structural limit for the residual cross section of the tube wall -

undegraded fraction of the tube wall area 1

swecce95e_ooms 30 w y 2i.i,95

For Braidwood-1 and Byron-1, 3APuo is about 4020 psi, 'the lower tolerance limit (lower 95/95 confidence) flow stress is 65.3 ksi, the tube diameter is 0.750 inch and the wall thickness is 0.043 inch. These parameters result in a RCS structural limit of 23%

remaining cross sectional area. Based on Figure 41-1, this corresponds to a structural limit of approximately 35 volts.

g. State whether axial forces other than internal pressure were included in the calculation of the minimum required residual cross-sectional area. Specifically, address the axial forces arising as a result of (1) SG tubes being locked in place; (2) plate rotation; (3) bending forces; and (4) incipient denting. If other axial forces were not considered, provide justification for not including these forces.

RESPONSE

As described in the response to question [f], the minimum required RCS (structural limit) is based on the Reg. Guide 1.121 burst margin guidelines. The actual axial loads due to pressure loading, locked tubes, bending forces, etc., would be less than 3APuo loads.

The structural limit could be based on the actual axial loads or the 3APuo loads, but it would be inappropriate to add the loads from locked tubes, etc., to the bounding 3APuo loads.

42. You state in WCAP-14273 on Page 10-4 that SG tubes with axial outside diameter stress corrosion cracking (ODSCC) indications can be used for your proposed expansion into the TSPs since axial cracking does not affect the joint stiffness. As discussed in EPRI report NP-7480-L Revision 1, " Steam Generator Tubing Outside Diameter Stress Corrosion Cracking at TSPs - Database for Alternate Repair Limits, Volume 1: 7/8 Inch Diameter Tubing", application of modest axial forces such as from sleeving on indications of stress corrosion cracking or cellular corrosion, can be expected to slightly open oblique crack faces as was observed at the Trojan facility. Discuss the implications of this observation with respect to selecting tubes for expansion with existing degradation.

RESPONSE

Crack faces at cellular corrosion can open up from axial forces to increase the eddy current response of the oblique crack components such as occurred at Trojan. This effect increases the crack resistance and the voltage response but does not cause an increase in the crack length. The same effect was seen on before and after tube pull RPC responses for some of the Trojan pulled tubes. Cellular corrosion on pulled tubes to date has been generally small patches on up to 11 volt indications such as the Braidwood-1 and Byron-1 pulled tubes. Opening up the crack faces for these cellular patches by expanding the tubes would have no influence on the effectiveness of the expansion for resisting TSP displacernents. There would be no reduction in the pull force resistance of the expansions for these small indications. Total severing of the tube is required to obtain significant TSP displacements at the expansion locations.

swecci35WA!_619.WP5 31

>=iy 2:.i,95

Given that the extent of cellular involvement at TSP intersections is limited to patches of corrosion and that total severing of the expansion is necessary to obtain significant TSP displacements, it is acceptable to expand at TSP intersections with ODSCC indications.

No restrictions on the degradation are necessary other than that the indications are typical of axial ODSCC responses found at TSP intersections for IPC applications.

43. Information regarding TSP degradation has recently been published. Specifically, articles contained in the June 15,1995, issue of Nucleonics Week and the Inside NRC issue dated June 12,1995, described instances of TSP degradation. Discuss the significance of these findings. Describe the potential for similar TSP degradation to occur in the Byron 1 and Braidwood 1 SGs.

RESPONSE

The response to this question was discussed with the staff in a proprietary meeting on July 20,1995.

swenusuva_6i9.wr5 32 u y n.i995

E'

.5 Table 36-1 Calculations to Evaluate Differential Tube \\

Tierod Thermal Growth i

Full Power to Cold Shutdown j

Plate location T(stm]

Tube Temp Tierod Tube Delta l

A 6.375 544.9 581.4 0.0216 0.0258

-0.0042 C

36.375 544.9 579.6 0.1232

-0.1465 0.0234 F

72.375 544.9 576.5 0.2451

-0.2906

-0.0455 J

108.375 544.9 573.0

-0.3670 0.4336

-0.0667 L

151.375 544.9 570.3

-0.5126

-0.6036

-0.0910 M

194.375 544.9 568.4

-0.6582

-0.7729 0.1147 N

237.375 544.9 566.6 0.8038

-0.9416 0.1378 P

280.375 544.9 564.7

-0.9494 1.1096 0.1603 i

Cold Shutdown to Hot Standby l

Plate Location T(stm)

Tube Temp Tiernd Tube l

A 6.375 547.0 547.0 0.0217 0.0239 j

C 36.375 547.0 547.0 0.1237 0.1362 F

72.375 547.0 547.0 0.2461 0.2710

/

J 108.375 547.0 547.0 0.3686 0.4056 L

151.375 547.0 547.0 0.5148 0.5668 M

194.375 547.0 547.0 0.6611 0.7278 N

237.375 547.0 547.0 0.8073 0.8888 P

280.375 547.0 547.0 0.9536 1.0499 Differential Displacement: Full Power to Hot Standby

(

Plate location l Tierod l

Tube l

Delta l

A 6.375 0.00010 0.00188

-0.00198 C

36.375 0.00054

-0.01034

-0.01088 F

72.375 0.00108 0.01957

-0.02065 J

108.375 0.00162

-0.02782

-0.02944 L

151.375 0.00227 0.03677

-0.03903 M

194.375 0.00291

-0.04508

-0.04799 N

237.375 0.00355

-0.05275

-0.05630 P

280.375 0.00420

-0.05979

-0.06399 Tiermi Alpha 550 7.13E-06 Tube Alpha 600 7.90E 06 550 7.85E-06 DISK 225A BRDWD\\TUBEXNTBL228 07/13/95 4.s.

w L

~:,

Table 36-2 Maximum Force Per Tube to Equilibrate Full Power and SLB Pressure Drop Loads Full Power lead Max SLBIoad Plate Imad loadfrube Imad load / Tube A

795 0.1737 11568 2.5269 C

1208 0.2639

-9702

-2.1193 F

2111 0.4611 5726

-1.2508 J

2878 0.6287 3122

-0.6820 L(hot) 3928 0.8580 5146 1.1241 M

3212 0.3508 9394 1.0260 N

3987 0.4355 16523 1.8046 P

4561 0.4981 23647 2.5827 No. Tube Intersections Plates A - L 4578 Plates M - P 9156

[

l l

l i

l I

l i

l DISK 225A BRDWD\\TUBEXP\\TBL229 07/13/95

F s%

Table 41-1 (WCAP-14273 Table 9-3: Revised July,1995)

Summary ofCellular/ IGA Corrosion Data Plane Tutpo/i$P Boten C,

Degradeton Sareg tnformaton UTS Rgtes Reessual Commeme Spamrt Scuste vofenge Type Widet

(%TW)

Bees Force Croce Sec8an (in )

(k-Ese)

(%)

E-4 R16G31/3 15 7 Coeuker 0.2 80 %

Met 112 8745 830%

Resegues CSA re-from Foran Date O2 50 %

Met R26C34/3 88 Cetular 0.12 100 %

Met 100 7 758 81.1 %

Resoluel CSA Csicadened from Force Data R28C47/2 11.7 Ceauser 0 18 80 %

Met 98 8 832 72.3 %

Residual CSA Calculated from Force Data R45C54/2 18 2 Coeuler Met 97 7.419 80.1 %

Residual CSA r*w from Force Date R45C54/3 14 Capular 0 12 50 %

Met 97 7.844 82.5 %

Resefuel CSA Calculated from Force Date 0.28 50 %

Met 02 80 %

Met R47C88/2 18 0 Ceeuler Met 97 5 171 55 8%

Residual CSA Calculated from Force Data R47C88/4 0.3 Coeuler O 79 50 %

Met 97 7.014 75 7 %

Rosaluel CSA eair.ama=d from Force Date 0 28 50 %

Met R33/C98/2 80 Ceduler Met 97 7.194 77.7 %

Residuni CSA ch=eed from Force Date R43C87/2 20 Coeuler Met 97 8.812 73.5%

Residual CSA Calculated from Force Date E-4 R42C49/2 30 Cellular N/A twA Met 70 0 %

Rosesual CSA from Meteturgas Exam R42C49/4 7.1 Coeuler WA N/A Met 800%

Residual CSA from Metesurgical Exam R8C47 34 0 Co#uler N/A N/A Met 41.0%

Reeefuel CSA from Mosesurycol Exam R17C58 40.1 CeRuler '

N/A N/A Met 39 0 %

Residual CSA from MeteAurycol Exam P-1 R22C38/2 0 b4 Comuner til dog.

8% (380" seg )

Met 11.420 77.3 %

Resegues CSA = Reen Degrade @Fresepen Rupture Forca R22C38/3 0 82 Coeuler 197 dog 15% (380* evg )

Met 12.000 81.2 %

Resedual CSA = Rego Degredo6Treespen Rupture Force R22C38/FS 1118 14 783 R28C42/1 0 58 Ceeuler 270 dog 12% (380* eve )

Met 13.750 98.0 %

Residual CSA = Reto C.v.. ^-

. Rupture Force R28C42/2 1.08 Celluter 250 dog 18% (380* evg )

Met 12.500 87.3 %

Resefuel CSA = Roeo Degrade @Freenpen Rupture Force R28C42/FS 108.1 14.320 Bredwood.1 R37C34/5 10.3 Geomer 80 dog.

37% (80* seg )

Met 91.8%

Roonfuel CSA from Matasuryces Exem, Average Dep8i R42C44/5 2.05 Cohdor 90 dog.

19% (90* ovg )

Met 95.3 %

Residual CSA from MeteAurycel Exem, Average Depth R42C44/7 0 21 Casuier 380 dog 32% (380* evg )

Met 880%

Reeddual CSA fromm=gical Exem. Average Depth l

Byron.1 R3C101/3 5 98 Ceewer 80 dog.

51% (80* ovg )

Met 91.5 %

Reassual CSA from ^^ -, - Exem. Average Depth R20C7/3 10.9 Ceauser 80 dog.

80% (80* egg )

Met 867%

Residual CSA frorn MeteAurgical Exem, Average Depth R20C102/5 1.11 - CeEuler 135 dog.

51% (135* ovg.)

Met 80 9 %

Residual CSA from Mete 8urgical Exem, Average Depth L

R12C8/1 3 83 CeE1LA Cec. Profile of Degraded Area Met 111 NA 61.0%

integrate rematual weg cross secean R12C8/2 1.38 Creds Multiple (50)croche - 29.9% mean Met 111 NA 89 0 %

Assume un8erm weg lose et moon crack depth R12C8/3 0 85 Cracks Mulliple (95) cracks 27.7% moon Met 111 NA 71 0 %

Aneume ur# arm was lose et rnoon crack dep83 IGA 45 IGA Uniform Circumt 9.5%

Met NA NA 90.5 %

Specimens 45 IGA l

10 7 %

Met NA NA 89 3 %

4.5 IGA l

14 7 %

Met NA NA 85.3 %

89 IGA l

40 4 %

Met NA NA 59 8 %

9 IGA

{

50.3 %

Met NA NA 49 7%

EPRIIGA.

1-3 2.1 IGA Un8erm Circumf.

11 %

NDE NA NA 84 0 %

RCS includes 5% edder for NDE uncertemly Ubrary 1-3 4.7 IGA l

17 %

NDE NA NA 77 0 %

l Specimens 1-4 3.1 -

tGA l

12 %

NDE MA NA 83 0 %

l 14 20 IGA l

8%

NDE NA NA 880%

l 1-8 40 IGA l

8%

NDE NA NA 80 0 %

{

1-15 09 IGA I

8%

NDE NA NA 80 0 %

[

RAITBL9.XLSTMS.

s, Figure 41-1 (WCAP-14273 Figure 9-9): Residual Strength of Tubes with Cellular Corrosion o

Plant E-4 Tensile (Data Fit Excludes IGA and Burst Data) piani s.4 pr a

Plant L

~

o Byron 1 O

o o

o Braidwood 1 90%

x Plant P-1

%x o

Regression Curve y

0

---

95% Prediction

^

80%

M

- - - - - -95% Pred. @ LTL x

- NOp @ LTL o

v x

EPRIIGA, NDE Area

~0%

h

+ IGA, Measured Area

,3 U

A Plant E-4, Burst 60%

Plant J-1, Burst o

^

m ct)

~.'

+

Plant P-1, Burst o

F-m o Plant D-2, Burst o 50%

~

y 8

~.. _:

x

+

~.

E-40% -c_.

t sl

..,..,.s A

n

.s.

cs c

n e

e nu

..- -::. 'j. s -

g 30%

w

.. ).' [.

20%

?..

~

l 10%

1 0%

O 5

10 15 20 25 30 35 40 45 Bobbin Volts RFK: 7/20/95,9:04 PM

[ CELLULAR.XLS] Fig 9.9