ML062580234

Millstone, Unit 3, Response to Request for Additional Information on Use of a Brazed Joint Structural Integrity Methodology (10 CFR 50.55a(a)(3)(i) Request IR-2-38.)

ML062580234
Person / Time
Site:	Millstone
Issue date:	09/14/2006
From:	Bischof G T Dominion Nuclear Connecticut
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
06-557, TAC MC8893
Download: ML062580234 (41)

Text

Dominion Nuclear Connecticut, Inc. 5000 Dominion Boulevard, Glen Allen, Virginia 23060 Web Address: ~~~.dom.com September 14, 2006 U. S. Nuclear Regulatory Commission Attention: Document Control Desk One White Flint North 1 1555 Rockville Pike Rockville, MD 20852-2738 Serial No.06-557 NLOSIPRW R1 Docket No. 50-423 License No. NPF-49 DOMINION NUCLEAR CONNECTICUT.

INC. MILLSTONE POWER STATION UNIT 3 RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION ON USE OF A BRAZED JOINT STRUCTURAL INTEGRITY METHODOLOGY

/I0 CFR 50.55a(a)(3)(i)

REQUEST IR-2-38. TAC NO. MC8893) In a letter dated June 9, 2005, Dominion Nuclear Connecticut, Inc. (DNC) requested approval for the use of an alternative brazed joint assessment methodology at Millstone Power Station Unit 3 (MPS3) in accordance with the provisions of 10 CFR 50.55a(a)(3)(i). Reference ADAMS Accession No.

MLO5l6lOlOl.

The proposed methodology enables evaluation of nonconforming conditions on ASME Code Class 3, moderate energy system piping with brazed joints, as a part of the inservice inspection (ISI) program at MPS3. Attachment 1 to this letter provides a response to NRC questions received in a facsimile dated March 24, 2006, which were subsequently discussed with the NRC in a May 11, 2006 telephone conference call. Attachment 2 provides a new example with a more detailed description of the assessment methodology.

In a letter dated August 23, 2006, the NRC requested that DNC respond to the request for additional information no later than September 22, 2006. If you should have any questions regarding this submittal, please contact Mr. Paul R. Willoug hby at (804) 273-3572.

Very truly yours, ~eiald T. Bischof Vice President - Attachments:

(2) 1. Response to Request for Additional Information on Use of a Brazed Joint Structural lnteg rity Methodology

2. Detailed Explanation of Brazed Joint Evaluation Commitments made in this letter; None Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Page 2 of 2 cc: U.S. Nuclear Regulatory Commission Region I 475 Allendale Road King of Prussia, PA 19406-1 41 5 Mr. V. Nerses Senior Project Manager U.S. Nuclear Regulatory Commission One White Flint North 1 1 555 Rockville Pike Mail Stop 8 C2 Rockville, MD 20852-2738 Mr. S. M. Schneider NRC Senior Resident Inspector Millstone Power Station Serial No.06-557 Docket No. 50-423 ATTACHMENT 1 RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION ON USE OF A BRAZED JOINT STRUCTURAL INTEGRITY METHODOLOGY (1 0 CFR 5O.%a(a) (3)(i) REQUEST 1R-2-38, TAC NO. MC8893) DOMINION NUCLEAR CONNECTICUT, INC.

MILLSTONE POWER STATION UNIT 3 Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 1 of 25 RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION ON USE OF A BRAZED JOINT STRUCTURAL INTEGRITY METHODOLOGY By letter dated June 9, 2005, Dominion Nuclear Connecticut, Inc. (DNC) submitted a request for approval to use an alternative brazed joint structural integrity methodology for the resolution of nonconforming conditions on ASME Code Class 3, moderate energy system piping with brazed joints at Millstone Power Station Unit 3 (MPS3), as a part of the inservice inspection (ISI) program at MPS3. This attachment provides a response to NRC questions received in a facsimile dated March 24, 2006, which were subsequently discussed with the NRC in a May 11, 2006 telephone conference call. provides a new example with a more detailed description of the assessment methodology. In a letter dated August 23, 2006, the NRC requested that DNC respond to the request for additional information no later than September 22, 2006. NRC QUESTION 1 : Discuss the hardship in performing an American Society of Mechanical Engineers Boiler and Pressure Vessel (ASME Code) repair or replacement of a leaking brazed joint during normal operation or during a scheduled outage when an ASME Code-required leakage test was performed.

DNC RESPONSE:

DNC does not intend to use NRC staff approval of the proposed alternative structural integrity methodology as a basis for continued use of a temporary non-Code repair for a condition that is identified during a planned scheduled outage, or to support restart from such scheduled outages. In such cases, the NRC review and approval of a temporary non-Code repair will still be needed. The DNC procedure in use for non-Code repairs of ASME Class 3 piping is consistent with limitations that are described in Generic Letter 90-05, and the information to Licensees regarding resolution of degraded and nonconforming conditions in RIS-2005-20.

Specifically, the DNC procedure requires a Code repair at the earliest of the following:

Next scheduled shutdown of sufficient duration to complete repairs, or a scheduled shutdown greater than 30 days Next refueling outage Time at which flaw 1 leak size is predicted to exceed the flaw 1 leak size accepted by evaluation Leaks discovered during plant shutdown.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 2 of 25 DNC is requesting the use of this alternative methodology for conditions that reveal brazed joint leakage in ASME Class 3, moderate energy piping that remains structurally sound. With assurance of structural integrity, and the effects of leakage appropriately assessed and mitigated, these conditions have no safety consequence.

NRC approved methodologies are not currently available for establishing structural integrity in brazed joints. With no approved methodology, the operability of affected systems and components can become impacted and applicable limiting conditions for operation in technical specifications must be met. Unnecessary emergent repairs will result in associated safety equipment unavailability, forced outages, and transients to the unit that would otherwise be avoidable with an approved methodology for establishing structural integrity in this moderate energy piping. NRC QUESTION 2: Discuss the feasibility to stop the leakage by applying a fillet brazing at the end surface of the leaking brazed joint or installing a mechanical device to seal or collect the leak during normal operation or during a scheduled outage when an ASME Code-required leakage test was performed. In this case, the structural integrity of the leaking brazed joint is assumed not to be a concern. DNC RESPONSE:

Assuming structural integrity is confirmed, but leakage from the joint is unacceptable with respect to system performance, the leak may be stopped by application of adhesive sealant or soft rubber patch to the brazed joint. These measures are for leak control and provide no enhancement to structural integrity. Similarly, it is not anticipated that DNC would attempt to stop a brazed joint leak by application of a fillet. A good braze repair on a leaking joint would be extremely difficult to obtain and may adversely affect the integrity of the entire joint. More often, DNC would expect to collect and divert any leakage from brazed joints in the same manner it does for other leaks. In accordance with standard practice for leaks, nearby components sensitive to salt water such as stainless steel piping or electrical junction boxes are identified to ensure they are protected from the leakage. The leakage collection setup typically includes an inverted witch's hat and tubing to a floor drain or container.

It is noted that the leakage rate from a brazed joint is best characterized as weepage, and measured as drops per minute or often several minutes per drop. Such a leak can at times become indistinguishable from normal condensation.

There is an advantage to leakage collection in that the leak remains observable and, therefore, is a direct indicator of the condition of the joint. Periodic operator rounds can observe any significant increase in leakage and an appropriate re-assessment of structural integrity can be performed, if required.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 3 of 25 NRC QUESTION 3: Provide the bases for using the ASME Code 1998 Edition of Section XI with no Addenda for the Section XI RepairIReplacement Program activities. The NRC staff notes that in your previous relief request dated May 19, 2005, you performed a temporary non-ASME Code repair to a leaking brazed joint in service water drain line during plant operation at Millstone Unit 3. In that relief request, you referenced ASME Code 1989 Edition with no Addenda of Section XI, IWA-4000 as the ASME Code repair requirements which is different from the ASME Code edition you referenced in the current relief request (1 998 Edition with no Addenda).

DNC RESPONSE:

There was an error in the cited ASME Code Section XI Edition and Addenda for the Millstone Power Station Unit 3 (MPS3)Section XI RepairlReplacement Activities program reference in the request mentioned above. DNC requested the use of the ASME Code Section XI, 1998 Edition no Addenda, for RepairIReplacement Activities in a letter to the NRC dated June 20, 2005. NRC approval was subsequently granted in TAC Nos. MC7347 and MC7348 in NRC letter, dated September 13, 2005, (ADAMS Accession No. ML052210033).

Prior to the June 20, 2005 request and approval letter, MPS3 had used the 1998 Edition with no Addenda under the provisions of ASME Code Case N-389-1 as explained and based on clarifications provided in RIS-2004-16.

The RIS-2004-16 summary specified the use of a different edition of the Code without prior approval had to be addressed under 10 CFR 50.55a(g)(iv), and the September 13, 2005 approval letter addressed this issue. The request cited above in the NRC question should have been written to state that the 1998 Edition of Section XI with no Addenda was used, and that the Millstone Repair and Replacement Program had been updated to use the 1998 Edition with no Addenda, effective December 8, 2003. NRC QUESTION 4: Provide the water chemistry of the service water in the referenced piping systems and discuss its potential corrosion degradation on its adjacent components due to the leaking of the brazed joints. If the service water is seawater, the dripping of seawater on stainless steel components will cause the initiation of stress-corrosion cracking on its surface. Also, discuss the corrective action program that you will implement to inspect and clean up the dripping on the adjacent components.

DNC RESPONSE:

The proposed methodology is limited to the evaluation of structural integrity of brazed joints and is not meant as a methodology to address the extent of a condition and its associated safety significance with respect to other system interactions from the Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 4 of 25 weepage, or leakage, of a brazed joint. However, this proposed methodology will be used in conjunction with the DNC corrective action program, which independently imposes additional requirements of the operability determination process that remain consistent with considerations described in NRC Generic Letter 90-05 for evaluating such moderate energy Class 3 piping conditions. The DNC corrective action program requirements are also consistent with the information contained in RIS-2005-20 with respect to resolution of degraded and nonconforming conditions.

Therefore, the use of this proposed methodology as the basis for structural integrity of a degraded brazed joint remains conditional, as it will also require evaluation of the safety significance of system interactions that may be related to the condition. Considerations for flooding, jet spray, loss of flow, other interactions, the failure consequences and the impact to safe shutdown capabilities, are to be considered in evaluation of safety significance of system interactions that may be associated with the condition of a degraded brazed joint. Consequently, the corrosive conditions that may be present from exposure to seawater in the MPS3 service water system would be evaluated and mitigated as appropriate, as with any associated structures, systems and components that must be conservatively evaluated and mitigated in conjunction with DNC's corrective action program and the use of this proposed methodology for evaluating structural integrity of a degraded brazed joint. As a reference to the requested information, the water chemistry of the service water at MPS3 is described in many of the service water system component specifications. This is a list of properties excerpted from such a component specification.

pH Color Alkalinity (as CaC03) Phenolphthalein Methyl Orange Free CO (calculated)

Free Available Chlorine (FAC) Total Hardness (as CaC03) Nitrate (Nos) Sulfate (SO4) Chloride (CI) Phosphate (PO4) Total Solids Volatile Fixed Total Dissolved Solids Volatile Fixed Total Suspended Solids Anionic Detergent Silica (Si02) Calcium (Ca) Magnesium (Mg) Iron (Fe) Manganese (Mn) Alumina (as AI2O3) Chromium (Cr-Total)

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 5 of 25 Nickle (Ni) Less than - 0.01 PPm Copper (Cu) 0.08 PPm Potassium (K) 6.8 PPm Sodium (Na) 11,000 PPm Radioactivity Negligible NRC QUESTION 5: To support your relief request, you referenced ASME Code Case N-513-1 which permits continued operation of low energy systems with minor leakage when justified by an evaluation of system performance.

The NRC staff notes that the referenced ASME Code Case allows the continued operation of the degraded Class 3 piping only for a limited time, not exceeding the time to the next scheduled outage. However, your proposed relief request extends the time limit for the proposed alternative to exceed the next refueling outage interval with justification (Section 5.5 on page 1 I), which is not consistent with ASME Code Case N-513-1. Confirm in your response that the application of the proposed alternative is limited to the next scheduled outage with sufficient time for performing an ASME Code repair or replacement, but not beyond the next refueling outage. The NRC staff's review of your June 9, 2005, relief request on the proposed use of a brazed joint assessment methodology is based on this condition being met. A separate relief request should be submitted if this condition cannot be met. DNC RESPONSE: The intended use of the proposed methodology was for a limited period of time, allowing for a timely repair or replacement activity to be planned for discrete degraded brazed joint conditions needing evaluation, commensurate with safety.

DNC agrees with the NRC staff on the schedule limitations applied to the use of this proposal.

Specifically, the use of this structural integrity evaluation method will be conditional, in that its application for any degraded brazed joint condition will not exceed the time to the next scheduled outage of sufficient duration for performing an ASME Code repair or replacement, but not beyond the next refueling outage.

If a timely repair cannot be made, DNC will apply for a separate relief pursuant to provisions of 10 CFR 50.55a(a) to extend the duration of applicability beyond this intended limited period.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 6 of 25 NRC QUESTION 6: On page 4 of Enclosure 1, you stated that the lack of full braze bonding originates from construction or fabrication, and is not progressive over time. On page 10, you reported from a search and review of external operating experience that corrosion degradation was attributed to braze failures in closed loop and electrical cooling systems. You also stated that there was no operating experience indicating progressive failure for open loop sea water systems. To support your conclusion that no progressive failure mechanism exists in the open loop systems, you performed failure analysis on two brazed joint specimens removed from Millstone seawater service with nearly 20 years of service and no corrosion product was found. However, these specimens were taken from brazed joints that were not leaking. To adequately support your conclusion, the root cause for the leakage needs to be determined since the joint was not leaking when it was first put in service. Furthermore, failure analysis should be performed on samples taken from leaking brazed joints to determine the degradation mechanism that caused leaking. The potential degradation mechanism could be fatigue-related cracking, stress-corrosion cracking, or another mechanism and may not be limited to corrosion. These mechanisms may combine with the fabricated defects and cause leaking when it breaks the outside surface. Lacking sufficient evidence, a time-dependent evaluation, assuming the presence of a degradation mechanism progressive with time, should be performed.

DNC RESPONSE:

DNC agrees with the NRC staff on evaluating the type of degradation mechanism(s) that could be applicable to the affected service water piping. DNC's corrective action program requires that such evaluations be completed before using the proposed methodology to evaluate the fabricated defects' effects on structural integrity of brazed joints. The DNC corrective action process will evaluate the extent of condition and document a basis for conclusions regarding the type of degradation associated with each separate application of this methodology at MPS3. If progressive degradation mechanism(s) are found to be contributing to the source of leakage in the affected service water piping joint being evaluated, the proposed methodology will not be relied upon to establish structural integrity without a separate request to the NRC for review and approval of the temporary non-Code repair, pursuant to provisions of 10 CFR 50.55a(a).

It is noted that NRC approved techniques for evaluation like Code Case N-513-1, with provisions for evaluating time-dependent and progressive forms of degradation, are not applicable to the form of leakage and materials in brazed joints used in the MPS3 service water system.

DNC is, however, able to conservatively establish the applicability of appropriate evaluation techniques because the leakage from brazed joints has been extensively evaluated. The cause of leakage from service water brazed joints has been investigated several times. Each investigation shows leakage to be attributed to a defect in the braze during original fabrication. All braze joints have some areas with lack of bond. Before a Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 7 of 25 brazed joint is made, the pipe and fitting are thoroughly coated with flux material.

When the joint is heated and the braze filler is applied to the face, the flux melts and the brazed material flows by capillary action. The capillary action is effective as long as the gap is the proper size and both sides of the joint are heated to the required brazing temperature in a continuous motion around the joint. Some small round spots of lack of bond formed by either flux inclusions or voids exist in essentially all brazed joints but they do not adversely affect the performance of the joint. If, however, one or both sides of the joint are not properly heated, as can happen when one side of a joint is significantly heavier, or if heating is interrupted, larger areas of lack of bond may connect together to form a leak path.

These areas without braze filler metal are often filled with residual flux that solidifies to form a glass-like plug in the void. This glass-like plug is very hard and chemically inert and it can block the leak path for an indefinite time although it is very brittle and mechanical or thermal shock can shatter the glass plug opening the leak path at any time.

Brazed joints that are leaking when removed from service typically have the same internal appearance as those that were not leaking. The leakage flow rate is so small that there are no velocity effects, such as erosion. There is only limited surface corrosion. Other degradation mechanisms such as fatigue cracking or stress corrosion cracking are typically not credible in this application. Regarding fatigue, the leaking brazed joints have been found at various locations that are not highly stressed.

Also, the service water system inherently experiences only small temperature swings and the cyclic duty is low. Regarding stress corrosion cracking (SCC), copper base alloys are susceptible to SCC in ammonia solutions but not in the seawater environment of this piping

('I. Also, the brazing process inherently anneals the base materials in the area of the joint and results in low residual stress after the brazing process. There has been no observed stress corrosion cracking of the copper base materials at MPS3. For these reasons, and because the stated mechanism is credible and consistent with the observed behavior, it is not appropriate to assume an unobserved mechanism for the weeping of the joints. Although it has been concluded that no time-related metal degradation is involved in the weeping of brazed joints at MPS3, an essential part of the proposal's basis is that if there is such a mechanism at work, its evolution is very slow and not a factor in affecting structural integrity for an operating cycle. Generally, leaking brazed joints take many years to appear as the properties of the blockage/sealing materials from flux and trace elements in unbonded areas of the joint progress. Leaking brazed joints also do not increase their leak rate over a time that extends into the months of an operating cycle. The periodic monitoring (Section 5.4 of the submittal) is required by plant procedure for any service water leak and this monitoring will detect any significant (I) htt~://www.libraw.unsw.edu.au/-thesis/adt-NUN/~ublic/adt-NUN20040129.095303/

index.htm1.

See page 5a, Table 1 .I.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 8 of 25 increase in leakage and alert the operators of a need to reassess the structural integrity of the joint.

NRC QUESTION 7: Based on your discussion in Section 5.6, "Augmented Examination" (page 12 of Enclosure I), the guidance provided is not consistent with that provided in ASME Code Case N-513-1. In ASME Code Case N-513-1, a sample size of at least five of the most susceptible and accessible locations, or if fewer than five are available, then all susceptible and accessible locations shall be examined.

To exempt the previously examined joints from re-examination could be non-conservative, since the joints may have been examined a long time ago or [with] a technique used that may not have been able to identify the degraded condition. Furthermore, if additional degradation was found in the expanded sample, the referenced ASME Code Case requires this process to be repeated until no significant degradation is detected or until 100%

of susceptible and accessible locations have been examined. Please justify this difference or revise the guidance in Section 5.6 and in the Millstone procedures or Corrective Action Program regarding the requirement to determine the extent of condition in similar brazed joints. The guidance should be consistent with that provided in ASME Code Case N-513-1. DNC RESPONSE: Augmented examination shall be consistent with ASME Code Case N-513-1. NRC QUESTION 8: Provide a copy of NAVSEA 0900-LP-001-700, "Fabrication and l nspection of Brazed Piping Systems," dated January 1, 1973. The NRC staff understands that your ultrasonic testing (UT) procedure MP-UT-45, "Ultrasonic Examination Procedure for Examination of Brazed Joints - Millstone Unit 3 Service Water Piping," Rev.

000-00 (Attachment E to Enclosure

1) was developed based on the NAVSEA procedure.

Identify the differences between the two procedures (NAVSEA vs. MP-UT-45) and discuss the reasons for the differences.

DNC RESPONSE:

DNC will not be able to provide the NAVSEA Standard per your request due to certain publishing restrictions regarding its use. The NAVSEA Standard contains the UT criteria for the technique that has been used by the U.S. Navy for many years. Although DNC referenced this Standard to develop its request, the use of the technique described in the DNC procedure MP-UT-45, and by this proposal, has been independently validated and qualified for use by DNC at MPS3.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 9 of 25 NRC QUESTION 9: Using an example, describe how the average bond level of a brazed joint was determined by UT examination.

Also, describe how the average threshold bond level of 60% was determined to be acceptable without further evaluation for brazed piping.

DNC RESPONSE:

As an example of determining an average braze bond level, refer to the original submittal Enclosure 1, Attachment C, page 3, titled "Braze Bond Measurements." There are 20 UT readings taken at even intervals around the circumference of the fitting. The boxed-in section headed "Meas. Bond" lists the actual percent bond based on the ratio of signals back reflected from the inner surface of the fitting and the inner surface of the pipe. Below the column of the 20 measured readings is a calculated average of the readings; the average is 55% in this example. The 60% threshold for acceptance of a braze bond was chosen based on test results showing that for an intentional partial bond of 60% the brazed joint develops the full bending strength load of the piping, even when the bending is extended well beyond required design levels. In fact, the testing showed even at 30% partial bond the joint developed nearly the full piping strength.

The Enclosure 1, Attachment A, Page 4, Figure 3 showing a load-deflection curve for 30% and 60% bond level illustrates these observations. Figure 8 on page 8 of the same attachment also illustrates this conclusion. The 60% figure is also consistent with the acceptance criterion in the NAVSEA standard referenced in our original submittal.

NRC QUESTION 10: Describe in detail the trial demonstrations of the UT procedure you mentioned on page 5 of Enclosure

1. Describe the samples used for the demonstrations and identify the range of data scattering and standard deviation pertaining to readings reported by qualified examiners.

Also, describe the qualification of the examiners participating in the demonstrations including a discussion as to how they were qualified and what were the qualification requirements.

DNC RESPONSE:

Trial Demonstration, Qualification: Using techniques developed from NAVSEA 0900-LP-001-7000; five UT operators (Some currently qualified level II or Ill and some with previous Navy experience) conducted a blind, round robin test on six brazed joints which had been previously installed but were removed as part of plant modifications. The qualification of examiners participating in the demonstration was as follows:

Three of the Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 1 0 of 25 UT examiners were Ultrasonic Level II or Ill technicians trained and qualified in accordance with ANSIIASNT CP-189, 1991 Edition (three of these examiners are PDI Supplement 2 & 3 qualified).

Two additional non- certified examiners familiar with UT examination of brazed joints per NAVSEA 0900-LP-001-7000 were utilized in the UT evaluation. One of the non- certified individuals is an ex-NAVSEA 7000 examiner (equivalent to a Level Ill under ANSIIASNT CP-189). The requirements for performing UT evaluation on the trial demonstration brazed samples was familiarization with the NAVSEA 7000 requirements as well as the expected UT signals from components with or without insert grooves, fitting back wall, and lack of bond signals. Trial Demonstration, Sampling

Description:

A total of 6 samples were used in the UT evaluation.

These six joints were selected as representative of the ASME section Ill service water brazed joints that have experienced weeping type leaks in service at MPS3. The selected joints included 2 and 3 inch tees, couplings and elbows. Descri~tion Quantitv Joint Identification Number 2 inch tee 2 24J and 24K 2 inch coupling 2 2511 and 25JJ 3 inch elbow 2 OS 5A and OS 5B The UT results of each individual, along with the average of all 5 operators and the maximum deviation and standard deviation from that average are presented in Table 1 on the next page. After all UT testing, each brazed joint was mechanically cross-sectioned 3 times and then polished and examined to measure the actual percentage of bond at each section.

These values and the average are also presented in Table 1 for comparison.

Trial Demonstration, Data Identification: (see Table

1) Average UT results ranged from a low of 54% bond to a high of 87% with a maximum single deviation of 9% and standard deviations ranging from 2 to 5%. This shows a good correlation and precision between the 5 operator's UT results as compared to the NAVSEA document that requires standard test specimens to be examined by 3 qualified inspectors and to average their results to set the true bond for that standard.

To qualify, operators must examine 6 test joints with no single joint deviating more than 15% from the true bond value and the arithmetic average of the six deviations shall not exceed 8%.

Serial No.06-557 Docket No.

50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 11 of 25 Mechanical sections performed to determine the actual physical

% bond range from a low of 66% to a high of 90%. These values correlate well with the UT % bond and average 10% higher than the UT% bond, showing that the UT measurements are inherently conservative.

Table 1 : UT Evaluation of % Bond vs. % Bond Determined by Cross Sectioning 24K 48 53 59 51 57 54 +5 -6 4.45 91 87 86 +5-915.29 25 77 88 87 25JJ 70 76 78 70 79 75 +4 -5 4.34 OS 5A 55 61 57 52 55 56 +5-4 3.32 0s 58 61 60 64 55 63 61 +3 -6 3.51 ' Each brazed joint was UT examined by 5 Technicians Joint ID 24J 81 70 53 95 92 74 87 88 94 75 60 64 66 65 88 72 75 14 The results of technician examinations are listed under "lndividual

% Bond by UT-~xamination." The average of the individual UT exams for each brazed joint is listed under "Avg UT Bond %" followed by the maximum deviation from the average and the Standard Deviation (by the n-1 formula).

After completion of the UT examinations, each brazed joint was saw cut to expose 3 equally spaced circumferential cross sections and then polished to reveal the braze metal and all defects or voids in the braze ring area. These were measured to determine the "actual" percentage bond at each cross section. These values are listed under "lndividual

% Bond by Section." The average of the individual sectioning examinations for each brazed joint is listed under "Average O/O Bond by Sectioning." The difference between the UT and Sectioning

% Bond, is listed under "Differential (Sectioning

% - UT %). In all cases the "actual" measured (sectioning)

% bond exceeded the percentage bond determined by UT examination with the average difference being just over 10%. lndividual

% Bond by UT Examination 74 77 74 78 78 NRC QUESTION 11 : Describe the qualification programs that you already have in-house to qualify your Level II or Ill technicians, including procedures and equipment to perform the UT examination of the brazed joints. Also, describe your in-house training program for your UT examiners to obtain adequate knowledge and skill to determine the bond in the brazed joints. Avg UT Bond YO 76 DNC RESPONSE:

The DNC written practice for qualification and certification of NDE personnel addresses the education, training and experience requirements contained in CP-189 and 10 CFR 50 for UT personnel.

Max Dev +2-2 Std Dev 2.05 Individual

% Bond by Section 84 81 93 Average % Bond by Sectioning 86 Differential Sectioning% - UT% 10 Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 12 of 25 Qualification and certification used in the Millstone Power Station procedure, MP-UT-45, is as described in NAVSEA 7000. Equipment selection was based upon which search units would provide the resolution for discriminating the bond, insert groove and lack of bond signals. The specific equipment used was a Krautkramer USN-52L and a 5.0Mhz dual search unit. Specific training for NAVSEA 7000 examination was limited to the method of calibration, and contains requirements for examiners and inspector qualification.

Direction was provided to the examiners to discriminate the lack of bond indications from insert groove, fitting back wall and pipe back wall indications.

The DNC written practice includes instruction and practical requirements for brazed joint examiners. The following personnel requirements will be prerequisites for use of the Millstone Power Station procedure MP-UT-45:

1. Only Level II, or Level Ill personnel may independently perform, interpret, evaluate and report examination results. 2. Levels ll and Ill shall be certified in accordance with

[ANSIJASNT CP-189, 1991 Edition.]

3. The UT examiners shall have sufficient knowledge and training to determine ultrasonically the bond in brazed joints. 4. UT examiners shall demonstrate ability to recognize such technical deficiencies as insufficient beam penetration (transmission), poor transducer contact and interfering contact surface roughness from patterns displayed on the ultrasonic screen.

5. UT examiners shall maintain proficiency for examination of brazed joints by performing an examination of a brazed joint at least every six months.

6. Examiners who do not meet the requirement of [item 5 above] shall demonstrate their ability to examine brazed joints prior to performing examinations in the field. See Table 2 below for initial examiner qualification and proficiency requirements.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 13 of 25 Table 2: Brazed Joint Examiner Qualification Qualification Type No. of Samples Period of Qualification Initial Qualification 6 3 Years Proficiency 3 6 Months Requalification 6 3 Years Acceptance Criteria: The percent bond of the six test specimens as reported by the examinee shall be compared to the true bond and accepted on the following basis; the Initial Qualification: arithmetic average of the six test specimens shall not deviate by more than 8% from the true bond and no single specimen shall deviate by more than 15% from the true bond.

The percent bond of the three test specimens as reported by the examinee shall be compared to the Proficiency Maintenance:

true bond and accepted on the following basis; the arithmetic average of the three test specimens shall not deviate bv more than 15% from the true bond. Requalification:

Same as initial qualification.

NRC QUESTION 12: To ensure the performance of a qualified UT examination of brazed joints, was a performance demonstration program used? Discuss how this program was implemented to qualify the ultrasonic examination procedures, equipment and the personnel to perform the examination of brazed joints including data collection and evaluation. Discuss how this program followed the approach delineated in Appendix Vlll to ASME Code Section XI. Describe whether the sample sets prepared for the performance demonstration consisted of fabricated samples or field samples (if available, with joint configuration, pipelfitting size and wall thickness) similar to that of the brazed joints to be examined, and contained representative flaws. DNC RESPONSE:

A performance demonstration was not used as brazed joints do not fall under the jurisdiction of Appendix VIII. The round robin was conducted entirely on field-removed samples and the examiners obtained the data in a blind fashion with no access to other examiners' data. The samples were not masked, however, not masking samples in no way aids the examiner in determining the amount of bonding present.

Access to the internal diameter (ID) does not aid the examiner in determining the amount of bond present because the lack Serial No.06-557 Docket No.

50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 14 of 25 of bond exists between the component outside diameter (OD) and the backwall of the inserted pipe or fitting which is not visible. All samples contained as-found (from field removal) conditions that had varying degrees of bonding, which ranged from 0% bond to 100% bond. NRC QUESTION 13:

Describe in detail how the adjustment of bond readings (badj) to account for UT uncertainties was determined, including the database to support the UT data adjustment discussed in page 7 of Enclosure

1. DNC RESPONSE:

The adjustment to measured bond readings for use in evaluation was based on the recognition that there would be some uncertainty in bond readings. After some initial exams and comparisons among different examiners, a figure of 10% bond uncertainty at low bond readings was selected.

The database of initial exams is described in the response to Question 10.

Low bond joints typically have a patchy bond area and are thus subject to more uncertainty.

At high bond readings, i.e., when almost all the signal is reflected from the pipe wall and almost none is reflected from the fitting, the uncertainty was expected to be smaller, and is not significant anyway since the joint achieves full strength near 50% true bond conditions. Although the bond adjustment figure was determined using an approximation, the use of the selected adjustment is validated by the fit to the data when the adjustment is included.

NRC QUESTION 14: In Section 5.3.2 (page 7 of Enclosure I), you stated that for bond readings that are significantly non-uniform around the circumference of the braze, an effective (lower) bond is computed based on the equivalent moment of the adjusted bond areas. For clarification, provide an example to demonstrate this calculation and discuss the reasons and conservatism of this approach.

DNC RESPONSE:

The example given in the submittal, Enclosure 1, Attachment C, page 1 has a moderately non-uniform bond distribution. A simpler example of non-uniform bond distribution is given in Attachment 2 of this response. In the example, most of the bond is missing (cross-hatched) while on the lower portion of the joint, the bond level is 80%. The average adjusted bond is 47%. In this example, the adjusted bond bending strength about the weak axis is equivalent to a 41% uniform bond, lower than the 47% average bond level. Therefore, the accounting of bond distribution at low levels is Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 15 of 25 important to consider. The method used in the evaluation is based on straightforward geometry. The bending strength is taken as a sum of the bending strength of each small arc segment, each of which is taken to be proportional to its percent bond and distance from the bending axis. The geometry calculation is displayed on page 12 of the example in Attachment

2. The stress evaluation assumes that the piping bending loads are all in the weak axis. This is conservative because the bond shear due to torsion is not adversely affected by the bond distribution and because the actual bending loads would not always be aligned with the weak axis. NRC QUESTION 15: Assuming the worst-case scenario that a complete failure of a 3-inch pipelfitting braze joint occurred, what would be the upper-bound leakage rate at 100 psig? Discuss its potential impact on the functional requirements of the system and the reduction of the system margins of safety. DNC RESPONSE:

An upper-bound of 699 gallmin is estimated for the described conditions, although such a worst-case scenario is not a credible event with the proposed alternative. This estimate is derived using the flow formula in Crane's handbook, and assumes structural integrity has fully failed, which requires a complete failure of the bond followed by severance of the joint in 3 inch piping, and results in no system flow resistance and a discharge coefficient of 1 .O. 100 psi hL := - hL = 225ft lbf 64 - Downstream cooling to the affected components would be lost, since the piping is assumed to have no structural integrity and is separated at the joint.

However, catastrophic failure of the piping is not a credible outcome for this alternative methodology, where the non-conformance encountered amounts to weepage or some fraction of a drop a minute, and there is evidence of substantial braze bonding in the joint by UT measurement.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 16 of 25 If structural integrity of the joint is retained, joint severance does not occur.

An upper-bound maximum flow can be conservatively estimated at 6 gpm as stated in the original submittal, Enclosure I, page 6, Section 5.3.1 "System Effects." As stated, this value is very low compared to pump capacity, so the net flow to affected components would remain nearly unchanged.

In addition, the safety significance of system interactions are evaluated for each leaking brazed joint. Leakage is mitigated as appropriate to prevent adverse impacts upon structures, systems and components that are associated with, or in the proximity of, the affected brazed joint. The evaluation will not permit flow margins to be reduced below the design basis level, so there is no reduction in system margins of safety. NRC QUESTION 16: Discuss how the average UT bond readings of 60% or more, as the acceptance criteria, was determined. Discuss the conservatism of this acceptance criteria in terms of the mechanical properties of the brazing materials and the uncertainties of the UT bond readings.

DNC RESPONSE: The determination of 60% for use as acceptance criteria is discussed in response to Question 9. The conservatism of the acceptance criteria is demonstrated as follows. If the actual bond were 50%) then by Formula (3) in Attachment A, Figure 2 of the original attachment, the allowable loading of ( &,(badj) ) becomes the following:

Serial No.06-557 Docket No.

50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 17 of 25 This result is the maximum nominal longitudinal pipe stress for an actual bond of 50%, exclusive of any stress intensity factor (SIF). A SIF of (0.75)(2.1)

= 1.575 is required in piping stress analysis for primary loads. Thus, the above maximum nominal stress is equivalent to a Code stress value of 20,217 psi. Piping stress would have to reach its Faulted Code allowable of 2.4 Sh, where Sh is the piping allowable stress at operating temperature, to exceed the very conservative 5 ksi shear stress assumption for the braze material. Actual braze material in a good bond is estimated to have a shear strength of about 15 ksi, based on brazing procedure qualification tests. Brazing wire such as "SFA-5.8 BAg-1" does not have an ASME specified minimum yield or ultimate stress.

In the joint strength evaluation model proposed, the joint strength is directly proportional to braze shear strength and the adjusted percent bond. At low bond readings, the factor of conservatism introduced by the bond reading adjustment increases.

For example, a 30% bond reading would be adjusted to 22%, a reduction by a factor of 1.36. The combined conservatism of the braze bond adjustment and assumed bond shear stress capacity in actual practice is best illustrated by the submittal Enclosure 1, Attachment D, page 3, Table 1. The table shows a margin factor of at least 1.52 when the specified evaluation parameters are used, and that value is conservatively low, as has been described in the request.

NRC QUESTION 17: Identify the Construction Code qualification stress analysis reports that were reviewed to determine the design-basis loadings at the subject braze joint (page 7 of Enclosure 1, 5.3.3). Confirm that these are NRC-accepted piping stress analysis reports.

DNC RESPONSE:

The Construction Code for the MPS3 service water piping is ASME 111, 1971 Edition with Summer 1973 Addenda. For Code analysis, the piping is divided into separable portions and stress analyzed as described in licensing basis.

Refer to the Updated Final Safety Analysis Report (UFSAR), in Section 3.9, Mechanical Systems and Components, and in Table 3.9B-11, Stress Analysis Requirements.

The analyses are documented in stress calculations, one for each portion of the piping system. In application of the proposed alternative, the stress calculation specific to the brazed joint being evaluated would be used as the source of piping loads and stress. The stress calculation on record is not being modified. The methodologies and stress limits that are employed in the analyses remain consistent with what are documented in the UFSAR.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 18 of 25 NRC QUESTION 18: It appears that no stress intensification factors were applied in the stress analysis of the subject brazed joints (page 8 of Enclosure 1, 5.3) as discussed in 2.0 of Attachment B to Enclosure

1. Is this approach consistent with the applicable Construction Code requirements for stress analysis of the brazed joints?

If not, please elaborate.

DNC RESPONSE:

The theoretical and testing bases for the proposed alternative were derived from applied forces and moments. The testing applied a load in a three point bending configuration resulting in an easily calculated moment at the brazed joints.

As a convenience for evaluation purposes these are converted to equivalent nominal pipe stress, however the strength correlation to braze bond is based upon empirical analysis of the loads testing. Local stress concentration effects at the joint, if any exist, were inherent in the tests on actual brazed joint fittings.

The stress intensification factors (SIF) as required for Code stress analysis of the brazed joint configuration does not enter into the strength correlation.

When existing Code stress analysis of piping is used as input to the evaluation, DNC can either access the detailed piping loads that are available as computerized output, or use the summarized pipe stress output that includes the effects of the detailed piping loads. The Code stress results include the effect of an SIF that is required when comparing stress results to Code allowable stress limits. To get nominal stress from the Code results the SIF must be factored out. This allows the actual joint loading, in terms of nominal stress, to be compared directly to joint strength, also in terms of nominal stress. NRC QUESTION 19 (i): For equation (3) in Figure 2 of Attachment A to Enclosure 1 : Describe briefly how these equations were developed and identify the references.

DNC RESPONSE:

Equation 3 was developed from first principles Strength of Materials, in which stress or load at a point is proportional to its distance from the bending axis. The strength (bending moment capacity) of a brazed joint is, therefore, the integration of the strength of each bond area times its distance from the neutral axis.

In torsion all the incremental areas are the same distance from the axis of rotation, so the strength of the joint in torsion is twice the strength in bending. It is, therefore, conservative to combine the torsion and bending together, and to compare the result Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 19 of 25 with the bending strength. A formal derivation of the formula is included in Attachment

2. NRC QUESTION 19 (ii): For equation (3) in Figure 2 of Attachment A to Enclosure 1

For the applicable braze materials, provide the ASME Code-allowable mechanical properties including the allowable shear stress, the certified mechanical test data of the braze alloy used in the fabrication of the referenced components and the minimum mechanical properties based on the material specifications referenced in purchasing.

DNC RESPONSE:

The ASME Code does not define allowable mechanical properties for braze material.

For ASME Ill, Class 2 and 3, the Code does not require Certified Material Test Reports. The fabricated example brazed joints were fabricated from materials taken from station stock and are, therefore, representative of actual joints in service. Since the failure of a brazed joint occurs at the interface and not through the braze metal, mechanical properties of the braze metal do not directly determine the strength of the joint. NRC QUESTION 20 (i): The NRC staff notes that all brazed joints in the program were tested by three-point bending with the brazed fitting in the middle of the configuration.

Provide the technical bases for the selection of this testing method to evaluate the strength of the subject braze joints. Also, discuss the limitations and uncertainties of using this testing method to evaluate the bond strength of the braze joints given that the test sample is a composite of fitting, piping and braze materials and a bending load is applied to the sample. In AWS C3.2, "Standard Method for Evaluating the Strength of Brazed Joints," a tensile testing method is recommended. Describe in detail, including sketches as applicable, how the three-point bending test was performed and provide a sample calculation to show how the test data was collected and evaluated.

You stated in page 1 of Attachment B to Enclosure 1 that the load transfer between pipe and fitting is primarily by shear through the braze filler. Provide a discussion of why the three-point bending test is an acceptable method to evaluate the bond strength of the brazed joints. DNC RESPONSE:

The three point bending test was utilized because the most significant design loads experienced by the joints are bending due to deadweight and seismic loads. Also, testing in torsion or direct pullout would have required a complicated test fixture, and these loadings are not the most severe when there is any non-uniformity of the bond.

Figure 1 is a diagram of this testing configuration.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 20 of 25 Load Application LAF w Load Beam w Flexible Chain

-: ;2 LL~ Brazed Joint w w Three Point Bending Test Apparatus FIGURE 1 I w w I A test machine of very large load capability would be required for direct pullout testing. Uncertainty on the loads and moments applied to the joint are also reduced with the three point bending testing fixture that was used. The testing load cell is calibrated and the accuracy of the moment arm is known to within a fraction of an inch. Therefore, accuracy of test loading is reasonably adequate.

The testing configuration that was used, which includes the pipe, fitting and braze, is an optimal model for simply and predictably determining the structural integrity of the joint. The fact that the piping deformation and deflection sometimes governs the limit load means that the piping cannot exert enough load to exceed joint capacity.

I : w NRC QUESTION 20 (ii): I Pipe w w I You stated in Section 4.1 [Attachment D of Enclosure 11 that all test items (refer to Figure 4 in Attachment A) achieved their test collapse load at a load well above that which would be predicted for a 5-ksi braze shear strength.

Provide details regarding how the test collapse load is calculated from an assumed 5-ksi braze shear strength.

Show how the test collapse load, piping collapse load and the bond failure load was determined. In the three-point bending test, explain how the failure of the bond, as to both local bond failure and total bond failure, was determined.

DNC RESPONSE: The test collapse load is derived from the load-deflection curve.

The collapse load is defined in ASME Ill, Appendix II, Section 11-1430. Refer to Figures 2 and 3 below, which are annotated versions of the Figures 6 and 7 of the original request.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 21 of 25 IEaplacemenf tin) TWO INCH FIELD SAMPLES FIGURE 2 Initial bond failure is detectable by a discontinuity or knee in the load-deflection curve. This is indicated by LC in the figures. The discontinuity or knee is the ASME-defined test collapse load. La on the figures indicates the allowable joint load based on the submitted methodology. The difference between the La and LC illustrates the margin in the overall methodology.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 22 of 25 ~splixrmtrrt till) THREE INCH FIELD SAMPLES FIGURE 3 The LC and La load values are the "Test Collapse Load" and "Shear Capacity Load" (based on 5 ksi shear) that were listed in Attachment Dl Table 1 of the original request. The collapse load force was converted to a moment at the joint. The moment was converted to an equivalent nominal pipe stress by dividing it by the piping section modulus. The test collapse load does not depend on assumed joint shear strength.

Referring to Table 1 in Section 4.3 of Attachment D, test joint 36 had a measured test collapse load of 2,025 Ib. With a moment arm of 12.64 inch, the computed moment at that load was 12,799 in-lb, and the nominal pipe stress for that moment was 22.8 ksi based on a piping sectional volume Z of 0.561 in3. The minimum joint strength predicted from a 5 ksi braze shear strength is calculated using Formula (3) from Figure 2 in the request. Continuing the above example, with L.insert = 0.656 inches, D = 2.375 inches, and Z.pipe = 0.561 in3, the lower bound of joint strength for 100% bond is 25.9 ksi. For 61 % adjusted bond in the joint 36 example, the joint strength is reduced to 15.8 ksi as listed in Table 1 for that joint. In the testing there was no differentiation between local and total bond failure. Under progressive loading the initial bond failure is expected to be local, and additional loading results in additional bond failure. Since all tests were continued up to a deflection limit in order to determine an ultimate load capability, subsequent bond failure beyond the initial failure occurred. However, the brazed joint is considered to have failed at the initial indication of bond failure, and the subsequent additional bond disruption is of no consequence to the determination of credited joint strength.

After full deflection the piping had ovalized and some joints were distorted.

There were no complete severances of the joints. A post-test UT was not performed, which would Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 23 of 25 have measured bond levels after the joint was destructively loaded well beyond its collapse load. The bond readings following testing are not used for a correlation with the collapse load, which is the failure of the joint. Consequently, the post-test UT is not required.

After testing, the two joints 37 and 4A were separated and visually examined by the independent testing laboratory. In un-bonded areas superficial corrosion of the pipe and fitting materials was noted. There was no observed corrosion of the braze filler metal. A portion of joint 37 was sectioned and "significant lack of braze bonding and detachment" was noted. This result was consistent with the measured pre-test bond level of 27%. NRC QUESTION 20 (iii): Provide a detailed description of how the data in Table 1 were derived. Also, describe how the bond failure, test collapse load and shear capacity loads were determined from the mechanical testing. Discuss whether the specimens were destructively examined or re-UT examined to determine the level of bond failure resulting from the testing. DNC RESPONSE: The response to Question 20 (ii) addresses this question.

NRC QUESTION 20 (iv): The NRC staff notes that there is significant data scattering in the test results of field samples of brazed joints as shown in Figure 7 of Attachment A to Enclosure

1. Provide reasons for the observed data scattering and discuss its impact on the reliability of bond level determined by UT examination.

DNC RESPONSE: Figure 7 shows load deflection curves for the several samples, and it exhibits expected variations that are based on percent bond.

Figure 9, related to Figure 7, shows data scatter at adjusted bond levels below 50%; however, all data points are above the acceptance criteria.

The figure shows some 3 inch samples with data well above acceptance threshold. The 3 inch samples have a relatively large meniscus fillet at the face of the joint. The scatter for 3 inch samples that is shown in the figure appears to reflect how this fillet characteristic is not credited for strength in the evaluation methodology.

Also, the percent adjustment at low bond readings is an almost 10% reduction on measurements that is an added conservative adjustment to the true bond. The aggregate of such conservatisms result in a relatively large data scatter while still validating the intended margin factor of 1.5.

Serial No.06-557 Docket No.

50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 24 of 25 NRC QUESTION 21 (i): Figures 8 and 9 in Attachment A to Enclosure 1 - Describe how the shear (Sh) limit was determined and what the 2.4 Sh limit means. DNC RESPONSE:

The "Sh" value is the ASME Ill Code allowable stress of the piping for operating temperatures.

The 2.4 Sh value is the maximum stress for faulted conditions (Level D) permitted by ASME Ill. This value is shown because, by Code, the maximum piping stress, including stress intensification effects, must be less than 2.4 Sh. Therefore, the 2.4 Sh stress level represents the maximum loading of interest in piping systems. NRC QUESTION 21 (ii): Figures 8 and 9 in Attachment A to Enclosure 1 - Describe how the equivalent pipe stress was determined from the test load.

DNC RESPONSE:

See the response to Question 20 (ii). NRC QUESTION 21 (iii): Figures 8 and 9 in Attachment A to Enclosure 1 - Describe how no bond failure was determined and whether destructive examination to support the determination was performed. How was the local bond failure differentiated from the gross bond failure?

Was it based on the shape of the test curves or the appearance of the test samples?

DNC RESPONSE:

See the response to Question 20 (ii). NRC QUESTION 22: For Attachment C to Enclosure 1 (Example Structural Assessment), a more detailed description of the assessment methodology should be provided. It would be helpful to provide sketches to show the dimensions.

For Part 2, it is not clear how the effective bond was calculated and its relationship to the bond level determined by UT. Further explanation is needed for the plot, and the definition and calculation of Dxx, Dyy, Doffset, Alpha, Bxx, Byy Bbend and Bpress. Provide a sketch to show the locations of the 20 UT readings at joint 1A and how the average reading at each location was Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 1, Page 25 of 25 obtained.

For Part 3, additional explanation is needed for the definition and calculation of Sxx, Syy, Sallow and the use of the Lookup Table. For Part 4, describe in detail the method, the input data and the equations used in the calculation of the max nominal stress of 4370 psi. Discuss whether this stress calculation shown was based on data taken from NRC-approved piping stress analysis reports.

DNC RESPONSE:

The response to Question 17 addresses the question on stress calculation approval. With respect to the NRC staff request for a more detailed description of the assessment methodology, a new example of the evaluation using this methodology is included in Attachment

2. NRC QUESTION 23: The NRC staff notes that your proposed use of the alternative brazed joint assessment methodology in lieu of an ASME Code repair or replacement, when leakage is found in a brazed joint resulting from the performance of a system leakage test performed in a scheduled outage, is not consistent with the purpose of the ASME Code-required system leakage testing. The ASME Code-required system leakage test should be scheduled in such a manner that there is sufficient time to perform an ASME Code repair or replacement of the affected component. Allowing a plant to start up with known leakage will not provide an acceptable level of quality including defense in depth in the operation of the plant.

The proposed alternative should not be implemented on a generic basis during a scheduled outage. Therefore, given this, it is not clear to the staff what the acceptable level of quality, including defense in depth in the operation of the plant, that this proposed change represents.

Please clarify the NRC staff's understanding of this. DNC RESPONSE: The request will not apply to leakage identified during a scheduled ASME Code-required system leakage test and it will only be applicable to leakage associated with brazed joints during system operation outside of a refueling outage.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology ATTACHMENT 2 DETAILED EXPLANATION OF BRAZED JOINT EVALUATION (1 0 CFR 50.55a(a)(3)(i)

REQUEST 1R-2-38, TAC NO. MC8893) DOMINION NUCLEAR CONNECTICUT, INC. MILLSTONE POWER STATION UNIT 3 Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 1 of 12 DETAILED EXPLANATION OF BRAZED JOINT EVALUATION (1 0 CFR 50.55a(a)

(3) (0 REQUEST IR-2-38, TA C NO. MC8893) This attachment provides a detailed explanation of an example brazed joint evaluation.

The example evaluation is on pages 9 through 12 of this attachment. Other than a simplified dataset and minor formatting, the example uses the same evaluation as provided with the original submittal. In the example evaluation on pages 9 through 12, the dashed boxes indicate input fields; all other fields are calculated or referenced from others. The following items are associated with the call outs on the example evaluation. "Part 1" of the evaluation on page 9 lists basic data that identifies the brazed joint, the basic system function of the piping, and the relevant design drawings and stress analysis. The only specific numerical input used in the evaluation is the pipe diameter and wall thickness, 2.375 and 0.156 inches in this example. The numerical values of the lower bound bond shear strength (5,000 psi) and percent bond adjustment parameter (1 0%) are a constant for all evaluations. The average of the unadjusted bond measurements is listed on page 9 and is 52% in this example. This value is linked to its calculation on page

11. As per the alternative, if this value is above 60% then the bond is acceptable without further evaluation. In this example, the average of 52% is less than 60% so a detailed evaluation is necessary.

Dxx and Dyy are the offsets of the weighted center of bond strength from the nominal centerline on the joint. Bond strength is best evaluated in a coordinate system aligned with its center of strength. Its local strength is simply the shear load capability of a good bond times the local percent bond.

When the bond is non-uniform, the purely axial loads of pressure thrust develop a bending moment with respect to the center of strength axis, with the Dxx and Dyy being the components of the effective moment arm, the resultant of which is denoted as Doffset. The Dxx and Dyy are calculated on page 12. In this example, the bond is symmetric about the vertical axis, so Dxx is 0.000 in, while for the horizontal axis there is more bond below the centerline, resulting in the negative value

-0.1 93 inches for Dyy. Alpha is another parameter representing the non-uniformity of the bond. It is the angular offset of the principle axes of the bond relative to the coordinate system used for bond measurements.

It is also calculated on page

12. In this example the bond is symmetric about the measurement axes so the Alpha value is zero. The braze bond data is presented in two columns. The first contains values based on measured, or "Actual" readings; the second contains values based on adjusted bond readings. The data based on "Actual" readings are presented for information only and are not used in the evaluation.

The more conservative "Adjusted" values Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 2 of 12 are used for evaluation.

The listed parameters include Bxx and Byy. These are the effective bond for bending loads about the principle axes, calculated on page 12. The resultant of these, Bbend, is used for evaluation of bending loads. The Bpress value is the effective bond for axial loading such as pressure.

6. This figure on page 9 is a figurative representation of the bond distribution.

The angular location of the bond measurements is shown. As a geometric necessity in a 2D plot, the cylindrical area of the bond surface is represented as an annular area. The x and y axes show the coordinate system used for measurement and the indicated angle is the sense of the Alpha parameter described in item 4 above. The grey portion of the annular area of the plot represents the effective degree of full strength bond and the cross-hatch portion represents the degree of disbondment. A 100% bond would be all grey. For an alternative representation of the same data, a linear graph of the same bond distribution is shown on page

11. 7. On page 10, Part 3 of the evaluation is devoted to determining the load capacity of the brazed joint. These geometric inputs include piping dimensions (linked from page 9), the piping section modulus (calculated from D and t), and the socket depth of the braze fitting, Linsert. This latter value is linked to the spreadsheet lookup table on the right for the socket depth.

The dimensional standard for the fittings (MIL-F-1183, called "Milspec" ) is what is listed in the piping specification used for procurement and construction.

8. The strength of the bonded joint is represented by the equivalent piping bending stress; i.e., the bending stress in the pipe for the same bending moment that the braze joint can withstand. The joint strength is presented as an allowable stress so as to permit simple comparisons with pipe stress levels available in pipe stress calculations. The derivation of the Formula (3) of Figure 2 in Attachment A of the original submittal is as follows: C adj =effective uniform percent bond da =.R.L .do insert Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 3 of 12 The "S,,(100%)" value in the example represents the strength of a fully bonded joint. It is based on the formula in the box on the right side of page 10, in the example evaluation, which is basically the same formula as derived above. Thus, in the example the Smax(lOOO/')

value is calculated as: / * \ 9. For values of badj less than 100%, the braze joint strength is simply badj Smax. These values of Sxx, Syy, etc are based on this relation, using bdj = Bxx, Byy, etc. from page 9 (see item 5). Thus Syy = 25,662 41% = 10,421 psi. The Sallow is the minimum of Sxx and Syy and represents the worst case joint strength. Thus, Sallow = 10,421 psi. 10. Once the joint load capability is known the actual loads in the piping are required for comparison.

The Pipe Stress Data part of the evaluation summarizes stress results from the piping analysis and converts it to be in the same terms as the joint load capability so as to permit a direct comparison.

This top part identifies the applicable pipe stress analysis.

It is the same analysis that documents the ASME Code and licensing basis qualification of the piping and is therefore a valid basis for determining design basis loading on the brazed joint.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 4 of 12 11. The piping ch aract eristics here are copied from earlier portions, except for "A.pressure", which is the effective area for pressure thrust load calculations.

It is based on the outside piping diameter (2.375 inches) rather than the inside diameter of the pipe because the brazed joint socket leaves the end of the pipe exposed to system pressure. 12. This section identifies the piping analysis node corresponding to the brazed joint. It also lists the ASME Code stress intensification factor (SIF) that was used for stress calculations.

This is needed because for calculating ASME Code pipe stress results, an SIF multiplier was applied to moment loading, so if nominal stresses are used to represent moment loading the effect of the SIF must be divided out. Thus, The SIF is listed at its nominal value (2.1 for all brazed joints) and also the ASME Code value used for primary loading, "Primary SIF" = 0.75 2.1 = 1.575. 13. These formulas are used to convert the Code pipe stress results to nominal results. The first formula calculates the stress, "Sp-offs~", representing the additional nominal stress resulting from the product of the pressure thrust (pressure times area) and the lateral offset of the braze bond center of effort, Doffset, as discussed in item 3 above. In the formula, P,, is the maximum pressure and A,,,, is the area for pressure thrust (the same as "A.pressurefl described above). The second formula in the box calculates the nominal piping stress, S', equivalent to the braze joint loading. The only new term in the formula is SI,, which is the ASME longitudinal pressure stress that is included with the moment stress term in ASME Code stress results.

The ASME Code formula for SI, is Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 5 of 12 The derivation of the formula for S' is as follows. With 100% bond, the maximum shear on the braze filler metal due to pipe bending and pressure loading is 'bond = 'bend 'press With less than 100% bond, the maximum shear on the remaining braze filler metal is =bend 'press --+- 'bond - *bend Bpress Since the acceptance criterion is 'bond < 'max and defining 'allow= &bendmax the acceptance criterion can alternatively be written as - -- .- Bbend"bond

' 'bend + =press < 'allow press Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 6 of 12 Since the bond shear stress is directly proportional to pipe stress, the above criterion becomes

%end 'bend + 'lp'- ' Sallow %press Since 'code = + psif"bend then 'code - 'lp Sbend = psif Adding in the offset pressure bending term, the criterion becomes 'code - 'lp Bbend psif + Sp-oi~set

+ Slp.r press < 'allow The left side is the required stress for comparison to the allowable

14. This block lists the ASME Code pipe stress input on the left and the conversion of the pipe stresses to nominal stresses on the right. The SIP value on the left is calculated based on the design pressure and piping dimensions using the formula stated above. The 64 value of Sp-offset due to the non-uniform bond is based on the formula in the box above: The conversion of the ASME Code stresses into nominal stresses using the formula derived above is illustrated by the conversion of the 2500 psi value for Eq. 8: 2500 - 76 1 40.0!! + 64 + 761.- = 1830 psi 1.575 46.7% The same formula applied to the Eq. 9 value (3500 psi) and Eq. 9F value (4500 psi) results in the listed nominal stress values of 2465 psi and 3100 psi respectively.

Since the 3100 psi value bounds the others it is used for comparison to the allowable.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 7 of 12 Part 5 simply compares the results of Parts 3 and 4 to determine whether the brazed joint has adequate structural integrity to withstand all design basis loadings. In this case 3,100 psi joint loading is less than the 10,421 psi joint loading capability, so the joint is structurally adequate for design basis loads pending its repair. (If the comparison were not successful, the joint would be declared inoperable and appropriate action would be taken.) The "Braze Bond Measurements" sheet records and summarizes the braze bond UT readings.

The UT readings are those recorded by procedure MP-UT-45 as provided with the original submittal.

This sheet performs the bond level adjustments and also develops tables of values for plotting the percent bond around the circumference of the joint.

The "Meas. Bond" column is an input field for listing the measured percent bond at each of the joint locations around the circumference. In this example, the readings are all 40% except for ones at locations 7-9 and 13-1

5. The "Adj. Bond" column is the result of adjusting each of the bond readings according to the formula stated in the submittal.

For example, the 40% bond is adjusted to The columns of measured and adjusted bond values are used on page 12 of the evaluation.

The Average, Minimum and Maximum bond readings are summarized here. The average for the measured bonds listed here is the source of the "Measured Ave. Bond" value reported on page 9. The linear graph plot shown here is an alternative representation of the bond readings to the polar plot shown on page 9. The "Braze Bond Calculations" on page 12 is used to determine the effective bond characteristics such as its principle coordinate system and the effective bond for bending about the principle axes. The top half of the evaluation shown on page 12 calculates results for the as-measured bond readings from page 11. The results are shown for information only, because the results are not used in the formal evaluation.

The lower half of page 12 calculates results for the adjusted bond readings from the evaluation shown on page 11. The adjusted bond values are more conservative (lower), and the effective bond values calculated are used in the formal evaluation of bond strength. The geometric formulas presented here are used in the evaluation shown on page 12. The outline of the bond calculations is presented in item 24.

Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 8 of 12 The evaluation shown on page 12 first calculates the first, second and cross-product moments of the bond distribution (variables ry, rx, Bpyy, Bpxx and Bpxy respectively).

From these the offsets of the center of bond effort (Yoffset, Xoffset) are computed, and then the bond moment terms are translated to the offset coordinate system (Byy, Bxx, Bxy) using formulas from the second formula box. From these three latter terms the angle of the principle axes, Alpha, is determined in a standard manner in the third formula box. With this series of formulas Alpha is forced to be in the range -45 to +45 degrees.

With Alpha now available the bond moment terms in the original coordinate system can be rotated into the principle coordinate system, resulting in Byy-p and Bxx-p. The equivalent bond levels for bending are reported as twice these values, since even a 100% bond would have a bond moment for bending of 0.5. The bond characterization values Dxx, Dyy, Bxx, Byy and Alpha listed in the evaluation shown on page 9 are linked to these calculated values at the bottom of the spreadsheet. Again, the results based on measured bonds are for information only and the results based on adjusted bonds are used for evaluation. The following pages, pages 9 through 12 of this attachment, show an example evaluation described in items 1 through 25 above.

Serial No.06-557 Docket No.

50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 9 of 12 Part 1 Basic Data ( dashed boxes are inputs) i?E!S_---l-l--.*-*l-lI inputs \ ~ine NO. !3~WP-~2-999-3 1 r--.-.------..

t PipeDta, 23751n Sys Functmn: !A supply to ACUS-1A 1 I Nom.~aH~hki 0.156rn Rpi'ng Iso; :CP-0123456 Rpe Mat?:SB 466 CDA 706 Joint. 1NRC 1 I 1 Filfing 11/1atY~SB

&-----------a 61or 62 Stde of Joint ;Upstream

Ref. Bond Strength:

5,000 psi Jf Orientailon.

It&$? &,up _ _ ] bnd Adjustment 10% A Measured Ave Bond 52% (calculated. For bond measurements, see sheet 'UT Readings')

52 X *= $0 % ? No, Detailed assessment required - Part 2 Bond Data Summary (data from sheet 'Bond Calcs') \ \p had on adjusted bnd: 0.000 In WY -0.193 in Doffset 0.193 ~n (16% of pipe radius) --+ Alpha 0.0 degrees - rotation angle af principal axes kalculated effective bond data are I in principal axes system, and are based an adjusted bnd I Actual Adjusted Bxx 55% 50% BYY 47% 41 % B bend 47% 41 % Bpress 52% 47% Note: Plot is figurative only, actual brae bond 1s cyllndrlcal, not El- m- m- Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology Attachment 2, Page 10 of 12 Part 3 Calculated Band Load Capability D 2.375 ~n tnom 0 156 in Pipe Z 0.566 lnn3 Linsert 0.656 in (from lookup table at right) smax(100%)

25,662 psi (from formula at right) Load Capability (Allowable Nominal Pipe Stress) (Based on bond tevels from Part 2) Lookup TM: L.lnsert per MtlSpec 0.nom D od Linsert 34 1.05 11/32 'I I ,315 7/16 1.5 1.9 38 2 2.375 21/32 2.5 2.875 25/32 3 3 5 53/64 Actual Adjusted Sxx 14,175 12,898 psi stress hsed on shear allow.

and percent bond SYY 12,054 10,421 psi Salfow 12,054 10,421 psi Part 4 Pipe Stress Data m- (data from Part i) (data from Part 1) Stress Calc NP-XI901 Pipe Dia 2 375 ~n Rev / CCN Rev. 5 CCN 4 Nom. Wall Thk 0 156 ~n Line No: 3SW-002-999-3 Pipe Mat'l SB 466 CDA 708 Sys Function:

A supply to ACUS-1 A Flttlng Mat'l SB 61or 62 Piping Iso CP-0123456 A.pressure 1 865 inA2 Joint. NRC I z.P@= 0 566 inA3 Eq. 9F (Design Basid( _ 4500 psi J Faulted 9F' 31 00 ps~ Max Pipe Nominal Stress 31 W ps~ Part 5 Structural Integrity Determination Joint NRC I -Joint Load Capability 10,421 psi (from Part 3) Design Basis Load 3, I00 psi (from Part 4) Check: 3,100 < 10,421 ===s Braze is adequate for design basis loads Monitor until repairlreplac&nent Serial No.06-557 Docket No. 50-423 Response to RAI on Brazed Joint Assessment Methodology , Page 1 1 of 12 \ Braze Bond Measurements

/ Joint NRC 1 Reading 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 N readings 20. dTheta 18 degrees Bond Adjustment

' 10% Angle M_la&_Bo~d-kdj Bond PloWalue Adj Plot 40%i 33% 0.850 0 833 40x1 33% 0.850 0 833 Aue 52% M in 40% Max 800h Adjusted Bond Readings R I Max 1 1 1 I I 1 1 1 1 I 1 I 1 1 I 1 1 1 1 1 Rmin 0 75 Min 0 75 0 75 0 75 0 75 0 75 0 75 0 75 0 75 0 75 0 75 0 75 0 75 0.75 0 75 0.75 0 75 0 75 0 75 0 75 0 75 Serial No.06-557 Docket No. 50-423 Attachment 2, Page 12 of 12 pjJ 1 Braze Bond Calculations Joint NRC 9 Bofbet Nreadhgs Equivalent bond based an measured bond readings, without adjustment 4 0% 20 D Angle Meas. Bard 3.375 0 40% AOffset 18 40% Inplt 36 40% 0 degrees 56 40% 0.mO rad 72 40% 90 40% 208 80% 126 80% 144 80% 162 40% 180 40% 198 40% 216 80% 234 80% 252 80% 270 40% 288 40% 306 40% 324 40%

342 40% cosflheta) db'cos dbacus*2 db'sln'cos sin(iheta) db'sln db'sin"2 1.000 0.400 0.400 0.DW 0.000 0,000 0.000 0.951 0.384 0.362 0.118 0.309 0.124 0.038 0.809 0824 0262 0.1W 0.688 0.235 0.138 0588 0.236 0.138 0.190 0.809 0.324 0.262 0300 0.124 01338 0.118 0.951 0.380 0.362 0.000 0.000 On00 D.WI 1.000 0.400 0.400 -0.309 -0.247 0076 -D.235 0.951 0.761 0.724 -0.588 -0.470 0276 -0.380 0.809 0.647 0.51 -0.809 -0.M7 0824 -0.380 0.588 0.470 0 276 -0.051 3 0.32 -0.118 0.309 0.124 0.038 1,000 -0AW 0.400 0.m 0.WO 0.000 0,000 -0.951 -0.380 0.362 0.118 -0.309 -0.124 0.038 -o.aag -4.647 0,524 0.380 -0.~8 -0.470 0.276 -0.588 -0A7U 0176 0.m -0.809 -0.647 0.524 -0.3~ -0.247 om 0.235 -0.951 -0,761 0.724 0.000 0.o~ oaoo o.m -1.000 -woo 0.400 0.304 0.124 0038 -0.118 -0.951 -0.380 0 362 odes 0235 0.138 -0.190 -0.809 -0.324 0.262 0.809 0.324 0262 -0.190 -0.588 -0.235 0.138 0951 0.380 0.362 -0.118

-0.309 -0.124 0 038 Rolrset Y&ei Bxy Byy+Bxr Xoffset Bxx 0.156 -0.156 0235 O.OM D.511 0.000 0.276 BBYY 47% Bave 5134 BBxx 55% 7.- - 4736 55% Byy-Bxx-0 Bxy- tan 2alpha cos 2alpha sln 2alpha tan check Alpha -0.041 0.000 0.000 1.000 OQOO 0.OW DDOOrad FALSE TRUE 0.0 deg I YLLIC., Angle Adj. Bond 0 33% 18 33% 36 33% 54 33% 72 33% 9D 33% 108 78% 126 78% 144 78% 162 33% 180 33% 198 33% cosqheta) db"cos dbocoP2 db'slKcos sln(thet6) db'sln db%lW2 1,000 0.333 0.333 0.003 0.000 0.000 0.000 0.951 0.317 0.302 0.098 0.309 0.103 0.032 0.803 0270 0218 0.159 0.588 0.196 0.115 0.588 01 0.115 0.158 0.809 0.270 0.218 0309 0.103 OD32 0.098 0,951 0.317 0.342 0.000 0.000 OQOO 0.003 4.000 0.333 0.333 -0.300 -0.240 0.074 -0.229 0.951 0.740 0.7W -0.588 -0.457 0269 .0.370 0.809 0,629 0.509 -0.809 -0.629 0.509 -0370 0.588 0.457 0.269 - -0,951 -4.317 0,302 -0.098 0.309 0.103 0.032 -1.000 -0.333 0633 O.OM 0.000 0.000 0.000 -0.951 -0.317 0.302 0.098

-0.309 -0.103 0.032 216 78% -0.8W -0.629 0309 0,370 -0.588 -0.457 0.269 234 78% -0.588 -0.457 0269 0.370 -0.809 -0 629 0.509 252 78% -0,309 -0240 0,074 0.223 -0.931 -0 740 O.7W 270 33% O.OW oam oaoo 0.003 -1.ooo -0.333 0.333 288 33% 0.309 o.im 0.032 -o.ose .o.ss~ -0.317 0.302 306 33% 0588 0.196 0.116 -0,159 -0.809 -0.270 0.218 324 33% 0.804 0270 0218 -0.159 -0.588 -0.196 0.115 342 33% 0.951 0.317 0302 -0.098 -0.309 -0.t03 0.032 0.0W -0.076 4607 0.000 0,000 0.000 5.026 Bprass cnetk=0 by Bpyy Bpxy check=O rx Bpxx 47% -0.162 0215 0.0m om7 0.000 0.251 -- Rdlsd Y&et Byy Bxy BW+Wo( Xoffd Bxx 0.193 -4.183 0203 0.W 0,454 0.000 0.251 41% Bave 45% BBxx 50% -- ," B -B Eyy-Bxx-0 Bxy=O tan alpha CDS 2alpha sin 2alpha tan check Alpha E~_~=B,~w&

a)- =--2 mntitq) -0048 O.O@O OOOt) 1 .OW ODOO 0.W 0.000 rad FALSE TRUE 09 deg