ML20198E459

Forwards Response to RAI Re Mechanical Nozzle Seal Assembly (Mnsa).Response Provides Sufficient Documentation to Show That Using Mnsa as Alternative Repair Method Does Provide Acceptable Level of Quality & Safety.Seminar Info Encl

ML20198E459
Person / Time
Site:	San Onofre
Issue date:	01/05/1998
From:	Rainsberry J SOUTHERN CALIFORNIA EDISON CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
Shared Package
ML20198E466	List: ML20198E459 ML20198E474
References
TAC-M99558, TAC-M99559, NUDOCS 9801090103
Download: ML20198E459 (16)

Text

--- -

I runs u,m...

ui,u,-

Ej EDISON l'f/M-,

AalNS - h7tM4ffnul= hengen January 5, 1995 i

U.S. Nuclear Regalatory Commission Attention:

Document Control Desk Washington, D.C.

20355 i

Gentleraent l

SUBJECT:

Docket Nos.56-361 and 50-362 Hechanical Nozzle Seal Assembly Code Replacement Request for Relief from 10 CFR 50.55a San Onofre Nuclear Generating Station, Units 2 & 3 (TACNa.M99558andM99559)

REFERENCES:

1) Letter from J. L. Rainsberry (SCE) to Documeni Control Desk (NRC),datedJuly 11, 1997:

Subject:

Docket Nos. 50-361 and 50-362, Mechanical Nozzle Seal Assembly Code Replacement, Request for Relief from 10 CfR 50.55a 5an Onofre Nuclear Generating Station, Units 2 & 3

2) 1.etter from Mel B. Fields (NRC) to Dwight E. Hunn (SCE), dated November 20, 1997;

Subject:

Mechanical Nozzle Seal Assembly Code Replacement for San Onofre Nuclear Generating Station, Units 2 ant: 3, Request for Additional Information (TACNos.M99558andM99559)

3) Letter from J. L. Rainsberry (SCE) to Document Control Desk (NRC),datedDecember 12, 1997; Subject : Docket Nos. 50 361

/

and 50-362, Mechanical Nozzle Seal Assembly Code Replacement, Request for Relief from 10 CFR 50.55a, San Onofre Nuclear Generating Station, Units 2 & 3 This letter provides, as an enclosure, the response to the Nuclear Regulatory Commission (NRC) request for additional inforetion concerning the mechanical noz.zle seal assembly (MNSA).

The information was requested by the NRC stat'f during the December 23, 1997, meeting between the NRC staff and representatives from the Southern California Edfson Compeny (SCE) and from ABB

. Combustion - Engineering ( ABB (,2).

The MNSA described under this submittal is

patented.by Combustion Engineering,_Inc. The patent' number is 5,619,546.

g1]g g{

j i

P PDR ~

{

Sa 6.Nwkur Gewinting Statkm.

Se Ocmente; CA 92ts7441128 714 WF-7420 -

L-

{_ f s

e r

Document Control Desk,

In puestion 21, the NRC requested that SCE evaluate cnanging from a Relief

[

RequesttoanAlternativeRepairRequestunder10CFR50.55a(a)(3)(1).

This section of the Code of federal Regulations states:

l r

- "Proposedalternativestotherequirementsofparagraphs(c),(d),(e),

i (f),(g),and(h)ofthissectionorportionsthereofmaybeusedwhen authortred by the Director of the Office of Nuclear Reactor Regulation.

The applicant shall demonstrate that l

(1) The proposed alternatives would provide an acceptable level of quality and safety."

SCE has considered this approach.

This letter and the revious SCE letters

)

(References 1and3)providesufficientdocumentationtoshowthttusingthe 1

MNSA as an alternative repair method does provide an acceptable level of quality and safety.

This is demonstrated by:

1. The MNSA is designed, fabricated, and constructed, using approved material in accordance with the rules of ASME BriPV Code, Section 111.

t

2. The MNSA is designed ',o prevent separation of the joint under all service loadings. This has been demonstrated by analysis and testing.

Copies of design and testing reports have been submitted to l

the NRC in Reference 3.

3. The MNSA is accessible for maintenance, removal, and replacement after service.

e

4. The MNSA prototype has been subjected to additional seismic, thermal transient, and hydrostatic pressure testing to demonstrate that the joint is sufficiently leak tight (zero leakage) to satisfy an l

acceptable level of quality and safety.

Therefore,-by this letter SCE requests approval to use the MNSA as an

- alternative = repair method under 10CFR50.55a(a)(3)(1),

i i

T i

1

,r

.>p

--p-r p

y,-

- -,~- - -

s-v-,

y

+ --- -

,--e

, - +

,m-,-

ne

-m,s.

--,e,,m

e

+

Document Control Desk '

If you have any questions on this subject, pleate call me.

Sincerely, L

Enclosures cc:

E. W. Herschoff, Regional Administrator, NRC Region IV K. E. Perkins, Jr., Director, Walnut Creek Field Office, NRC Region IV J. A. Sloan, NRC Senior Resident inspector 'an Onofre Units 2 & 3 H. B. Fields NRC Project Manager, San Onofri 'Inits 2 and 3

6 i

ENCLOSURE Tile SOU rliERN CALIFORNIA EDISON COMPANY l

RESI'ONSES TO Tile NRC'S REQUEST FOR INFORhiATION hfECIIANICAL N0ZZLE SEAL ASSEhillLY (hiNSA) t e

F

,J.-

t Responses to NRC Questions From the December 23,1997 Meeting lietween SCE and the NRC 1

Regarding the Ailli Methanical Nonle Seal Assembly (MNSA) t NRC QUESTION:

1. Should Dow induced vibration of the thermowell base been considered during testing?

SCE RESPONSE:

i Flow forces on the thermowell are opposed by contact with the inside of the nozzle near the i

reactor coolant system (RCS) pipe ID. The thermowells are designed to avoid the pump rotation and blade passing frequencies and therefore do not introduce vibratory loads into the system. The natural frequency of the thermowell tip is known to be quite high and will not change as a result of a 360 degree crack of the nozzle weld The potential evitation mechanisms that are being considered for the thermowell ti? are vortex shedding and pump vane excitation. These do not.

change, The natural frequency of the nozzle body is quite low (<50 liz) and will remain low aller a 360 degree crack of the weld. Therefore, the critical naturri frequencies of the thermowell tip end of the nozzle body are, and continue to remain, decoupled, and there is no feedback mechanism between them. Addition of the MNS A to the nozzle does not alter the loading on the thermowell or the interfaces between the the,mowell and the nozzle. Since Dow induced vibration of the thermowell is not signincant, simulation or 0ow induced vibration was not introouced into the test program. And r,ince the mass (appioximately 0.6 lbs, about 3% of the total simulated mass) and stin' ness of the thermowell are insignincant, it was not necessary to include a simulated thermowellinto the program. Ailll CE is confident that the results of the testing program would not be changed with the thermowell.

NRC QUESTIONt

2. Doci thermal expansion of the base material alTect bolt preload? (For example, but not limited to, will the thermal expansion of the bolt note centerlines affect she pre-load?)

SCE RESPONSE:

Thermal expansion of the RCS base material will have an insignilicant effect on bolt preload.

Since the MNSA and the RCS components are closely coupled, there will be little, if any, differential temperature between the materials. At 600F, the mean thermal expansion coellicient for the bok material (A286, Grade 660)is approximately 9.58 x 10-6 versus 7,42 x 10 6 for the low alloy steel RCS pipe, Within the pipe wall, the thread: of the bolts tend to grow more than the female threads of the low alloy steel pipe. At the pitch diameter of the thread, the radial differential expansion between the bolt and the pipe is on the order of 0.000f 9 inches, with the

bolt growing more than the tapped hole. In the axial direction, the bolt will grow approximately 0,00011 inches more than the tapped hole. This will slightly increase the preload.

To verify that differential thermal expansion will not impact the sealing quali ies of the MNSA, t

the SCE MNSA test hardware was constructed of the same material types as are used at the San Onofre Nuclear Generating Station (SONGS). Low alloy steel was used for the simulated pipe or

t tank test pieces, all clamps were constructed of stainless steel, and bolts were unde from A286, No loosening of the bolts was noted during the thermal cycle testing and no leakage was obsen ed from the seals.

NRC QUESTION:

3. Does ADB CE have any experience installing other mechanical seals on a curved surface such as is proposed at SONGS 7 What is the acceptance criteria for surface condition of the hfNSA installation locations? (Since the Grafoil is being compressed into a fixed volume, what size imperfections, such as gouges or protrusions can be accommodated?) llow is this controlled in the field?

SCE RESPONSE:

ADil CE has experience installing Grafoil seals over curved surfaces with a much smaller radius and more surface irregularities than the RCS pipe and the pressurizer bottom head. An example is the Canopy Seal Clamp Assembly designed by ADD CE to address leaking seal welds in Westinghouse plants. Approximately 70 such seals have been installed in the United States and Europe by ADD CE. This device has a Grafoil packing seal which begins as an "L" shaped cross section (see attached figure). The Grafoil sealis compressed around multiple small radii curved surfaces into the semi toroidal contour of a typical canopy seal, as the seal is loaded. The area where the seal is made is an "as welded" surface having the usual ripples and ridges associated with a welded seam. For the h1NS A, the compression force will casily cause the Grafoil to conform to the slightly curved surface of the RCS pipe or vessel.

The surface is required to be clean and free of dirt, loose scale, or other foreign material. There is no specified acceptance criterion for surface irregularity. Ilased on ADD CE's extensive experience with Grafbil seals, surface imperfections such as small ni& or scratches which do not clean up casily can most likely be dispositioned as acceptable. The compressed seal volume is approximately 70%

ofits original volume. This is higher than the minimum compression required to seal against RCS pressure and it can accommodate expected surface conditions based on the insignificant volume of the imperfections.

A procedural control is included during installation of the h1NS A alerting the installer to check the sealing smface for irregularities. Light cleaning with Scotchbrite or equivalent is an acceptable part of'this procedure. The procedure does not allow any machining of the pressure boundary. Any protmsions or indentations are required to be brought to the attention of engineering and they are evaluated on a case-by-case basis by engineering prior to installatian of the h1NS A.

NRC QUESTION:

4. Provide a more complete explanation for why the auto; lave heatup rate of 150 Fr -is suflicient, given the SONGS pressurizer heatup rate limit is 200 F/hr. SONGS has had several excessive cooldown even's in the past; could the h1NS A handle these eveuts?

l 2-l

SCE RESPONSE:

The 150 F/hr heat up rate applied in the test program is an operational limit for the laboratory autoclave. Ilowever, the h1NSA design can accommodate thennal transients at a much higher rate.

The portion of the bolts engaged in the vessel would closely track the vessel temperature because they are in intimate contact. Outside the vessel, the clamped par 1s and the bolts are in intimate contact with each other and the diflbrential expansion coeflicients are nearly identical for the stainless steel flanges, seal retainer, and compression collar and for the A286 bolts. The instantaneous value for stainlessis 10.38 x 10-0 and for A286, it is 9.58x10 6. Because the par 1s are in good contact and clamped to the pipe, it is not likely that any significant thermal lag could occur and very little diffbrential expansion could occur between the chmped stainless pieces and the bolts. Tharefore, any efTect on bolt preload is very small and would not afibet the sealing ability of the 'dNSA.

For the case of a very rapid cooldown, such as the event which occurred during Cycle 8, the 6herent springback of the Grafoilis suflicient to accommodate whatever slight increase in seal volume might occur due to the small diffbrential contraction of the h1NSA parts and the nozzle. This springback characteristic was shown ouring the seismic tests during which significant lateral nozzle movement was seen without compromising the scaling ability of the Grafoil.

NRC QUESTION:

5. Was bolt preload checked aller each of the thermal cycle tests and;or seismic tests?

SCE RESPONSE ADH C'3 reviewed the test files and determined that bolt preload was not checked between thermal cycle and seismic tests. In the test installation, no locking tabs were used on the bolts to prevent loosening. Ilowever, while not documented in the report, there was no loosening of the bolts observed by the test technicians aller thermal cycle and seismic tests No leakage was observed from the seal during these tests.

NRC QUESTION:

6. In the seismic test, the thermowell wa; not installed; a test weight was installed on the top of the nozzle. Why is this a satisfactory method of testing?

SCE RESPONSE:

Care was taken to simulate the distribution of masses and stiflhesses during the test program. For RTD applications, the significant items include the RTD head and housing (which were simulated with test weights), the nozzle, and the clamp itself. The thermowell itselfis a light weight object (approximately 0.6 lbs) located inside the nozzle adding little or no stillhess and an insignificant mass,

' especially compared to the exterior test weights which exceeded 20 lbs. Therefore, it is acceptable to model the test arrangement without the thermowell included.

3

ADD CE also tested the lower pressurizer MNSA and simulated the mass of the valve which is

. Installed in the plant.

NRC QUESTION:

7. NRC expressed concern that in the event oficakage past the Grafoil seal, water in the annulus will be replenished and some boric acid corrosion could occur.

SCE RESPONSE:

Ifleakage past the Grafoil seal should occur, the crevice between the cracked nozzle (Alloy 600) and the pipe /shell (Iow alloy or carbon steel) would fill with water or (depencSg on the size of the crack in the Grafoil seal) a steam / water mixture. In the first case, any aerated water in the annulus will be flushed with deaerated borated water at operating temperatures from the bulk reactor coolant.

WCAP.7099 presented data for A302D pressure vessel steel immersed in borated water (2000 2500 ppm 11 as boric acid) at temperatures of 70 to 500F. Under deaerated conditions, the maximum corrosion rate was 0.001 in/ year (1 mpy). Galvanic coupling of A302D to type 304SS resulted in zero to cery minor increase in corrosion rate (less than a factor of 2). The presence of oxygen had a much greater effect than galvanic coupling. In addition, there was no indication of hydrogen pickup in the A302D (isolated or coupled to 304SS) which indicated hydrogen embrittlement as a esult of absorption of corrosion hydrogen or RCS hydrogen is not a concern.

If the annulus is filled with a steam / water mixture, there vill be little corrosion of the carbon / low alloy steel over most of tne annulus. Most of the corrosion will be where steam / water mixture exits the anr.ulus (rear the Grafoil seal). The Combustion Engineering Owners Group (CEOG) funded a laboratory corrosion study that included heated rnockups with cracked Alloy 600 tubes welded into the mockups. liigh temperature simulated reactor coolant (800 ppm D, I ppm LI) leaked through the cracks into a narrow annulus and out the mockups. There was essentially no corrosion over most of the 4 inch long annulus es indicated by the presence of original machining marks and severe corrosion at the point where the coolant exited the annulus, The consequences of the corrosion relative to integrity of the low alloy steel components were addressed by two CEOG tasks, the results of which were previously presented to the NRC (see CEN 607, " Safety Evaluation of the Potential for and Consequence of Reactor Vessel llead Penetration Alloy 600 ID Initiated Nozzle Cracking", for example). In addition, the corrosion test report CE-NPSD 648 P was previously provided to the staff by another CEOG member.

The bolts used in the MNSA are fabricated from a corrosion resistant material, SA453 grade 660 (A-286 stainless steel). If the coolant leaks through the seal, it will exit as wet steam which may impinge onto the bolts forming wet deposits of boric acid and cooling the bolts to 220 300F. In carbon / low alloy steels, these conditions cause high corrosion rates (>l in/ year) as indicated by a ABB CE laboratory conosion study in which heated laboratory fasteners were exposed to wet steam from borated water. In the same tem, fasteners of 17 4Pil stainless steel and Alloy 718 were exposed to the same conditions for up to 2500 hours0.0289 days <br />0.694 hours <br />0.00413 weeks <br />9.5125e-4 months <br /> with no corrosion occurring. Although A-286 was not indaded in this test, its composition, which includes about the same level of Cr as 17 4Pil, indicates it will have the same corrosion characteristics as 17-4Pil. Other studies support this conclusion. For a description of the fastener corrosion test and summaries of similar tests see J.F.llall,

• Corrosion of 4

l l

)

Low Alloy Steel Fastener Materials Exposed to Borated Water", Proceedinns or the Third

, Internailemal Symoosium on Environmcatal Degradation of Materials in Nuclear Power Systems-l Water Reactors.-

l i

NRC QUESTION:

8. Provide additional rationale for the self limiting galvanic corrosion phenomenon.

t

.i SCE RESPONSE:

~

i The response to Question 7 included a discussion of galvanic corrosion between the vessel material-l and another corrosion resistant alloy (304 stainless steel) which noted only minimalincrease in the.

. corrosion of the vessel material. With respect to crevice conditions, leakage through a crack will fill.

l" the crevice, aller which there will not be any replenishment of the solution in the crevice other than perhaps by very limited diffusion through the crack. Given the tightness of stress corrosion cracks, any difRision will be minimal. The coolant chemistry will thus be the same as the reactor coclant except that it could be aerated immediately following an outage. Any oxygen in the coolant will be quickly consumed by corrosion with the shell material leaving a deaerated condition. The corrosion of the shell material will also result in the formation of metallic hydroxides which will elevate the

[

crevice pH resulting in a further reduction in the corrosion rate. In summary, there will not be significant galvanic corrosion.

During the midcycle outages, SCE expects to install MNS As on inaccessible portions of the RCS hot legs as an Interim replacement. These will be removed during the next refueling outages. To provide

.t confirmation of the above conclusions, the nozzle bore area will be inspected for corrosion when these interim MNSAs are removed.

NRC QUESTIONt

~

9. Provide a summary of Grafoil characteristics. Ilow is seal compression controlled? Provide an explanation of how the coeflicient of friction of Grafoil in steel was used.

SCE RESPONSC:

A copy of the Union Carbide technical pi.per on Grafoil is attached. This paper describes the material L

properties and composition. Grafoil was selected for the MNSA seal applications because ofits ability to be defonned and to flow into confined spaces under pressure at ambient conditions, and to seal against surfaces which are less than perfect sealing surfaces. The MNSA clamp components.

provide the mechanism for squeezing the Grafoil into the desired shape and to the desired pressure.

Although the published technical information provides guidance, it does not address such things as irregular surfaces, the need for split seals, unusual angles, etc. Design of the MNSA started with engineering judgment based on past experience with Grafoil seals of similar sizes and applications.

Next, the design process involves testing to optimize the way in which the Grafoil seal material is i

f y

S.

- i

.m,u._,.,

c.

,,..-m.

procured, loaded, deformed and contained in the clamp. Testing is also used to confirm acceptable

. perfonnance of the as-instaled condition. In the case of the hiNS A, the sealing material is squeezed a predetermined amount until metal to metal contact is established between the clamping parts. In so doing, the confined volume for the compressed seal remains stable during operation.

With regard to the application of the coeflicient of fiiction of Grafoil on steel and the manner in which it was applied in analyses, the following rationale is provided. In an " ideal" and worst case break scenario,the 360 degree crack would be complete, instantaneous, and oriented such that no base or weld me',al could interfere with the motion of the nozzle. In this case the only resistance to the nozzle motion would be provided by the attached piping and by the Grafoil seal in reality, this condition would uot exist and, instead there would be significant resistance offered to the motion of the nozzle by the cracked surfaces and by the integral material, if motion would be allowed at all.

The "Grafoil Engineering Design hianual, Volume one: Sheet and Laminate Products,"

11. A. Iloward, Union Carbide Corp.,1987, lists static friction coeflicients of 0.05 to 0.20. In laboratory hydro tests, nozzle motion has been observed at near 1900 psi, near 900 psi, but also at pressures near 100 psi. While some of the conditions for the observed motions may not be representative of the final installation, the pressure in the fully compressed seal is approximately 3,000 psi Assuming a 1 inch nozzle diameter,0.15 inch seal height, the seal circumferential contact force is 1,413 lb. Assuming a friction coeilicient of 03, the force resisting nozzle ejection is then 0.3x1,413= 424 lb. Using an aperating pressure of 2,250 psi and the nominal cross sectional area for a 1. inch nozzle, the ejection force is 1,766 lb. The resistive force is approximately 24% of the ejection force. To simulate this resistive force and the physical nature of an actual 360 degree crack, a 30% reduction in the ejection force was used in sevcral h1NS A calculations.

While based on an " ideal" break scenario the applied assumptions may not necessarily appear conservative, they were judged appropriate to offset the very conservative break assumption. It should also be kept in mind that the impact load is a 'one-time' occurrence and that the derived loads were actually conservatively treated as steady, normal operating loads.

NitC QUESTIG A:

10. The NRC had questions about the 1985 ABB CE A 286 corrosion study, especially about the conditions for stress corrosion cracking (SCC) and the hiNS A stress levels. Also, there were questions about the fabrication controls-SCE RESPONSE-CE.NPSD.305, " Stress Corrosion Cracking of A.286 Stainless Steel" presents the results of a corrosion test in which various types of A 286 SCC specimens were immersed in 550F and 650F simulated reactor coolant. Uniasial unnotched tensile specimens were loaded in 304SS fixtures to stresslevels of124 to 189 KSI. 8 of 8 specimens survived 9000 hours0.104 days <br />2.5 hours <br />0.0149 weeks <br />0.00342 months <br /> of test with nn indications of cracking. Ten notched specimens (Kt = !, or 4.3) were similarly loaded tr nominal stresses of 118 to 181 KSt. One specime,(Kt = 1.7) at a nominal stress of 124 KS1 cracked afler 5000 hours0.0579 days <br />1.389 hours <br />0.00827 weeks <br />0.0019 months <br /> at 550F. All other specimens, which included higher nominal stresses and Kt values, survived 9000 hours0.104 days <br />2.5 hours <br />0.0149 weeks <br />0.00342 months <br /> without cracking.

6-

t l

t in addition, this test included 16 pacracked bolt loaded CT specimens for measuring crack

(

, propagation at 550 and 650F. These were loaded to 55,45,30 and 15 KSI(in. Crack growth

- occurred in all specimens in the first two groups, but not in the two latter groups. Any indications

. ofgrowth in these specimens were within the scatter obtained in crack length measurements.

Results from these tests are not appl 2able to the hiNSA bolt application since the specimens were immersed in the coolant. The h1NSA belts are exposed to air and may be sprayed with steam. Since the bolts and the hiNSA parts are corrosion resistant materials, there is little likelihood that an autoclave could develop around the bolt threads as a result orcorrosion product buildup in the thread region as has occurred on occasion with carbon / low alloy steel fasteners.

hiNSA bolts are procured to SA 453 Grade 660, as Quality Class I components. The threads are

-[

machined.

The bolts receive a final dimensional and visual inspection consistent with the AShiE BPV Code for bolts of this size.

=i NRC QUESTION:

11. What are the details of the insulation which will be installed aller the h1NSAs are installed?

SCE RESPONSE:

The insulation configuration will be modified for hiNSA installation. The base insulation opening on i

- the piping or vessel will be enlarged to accommodate the lower flange. Pieces ofinsulation will be fabricated to fit between the tie rods. ~These pieces will be fastened to the nozzle with latches which will allow easy removal for future inspections. The insulating mat # iis reflective and will not create an enclosure around any threaded connection. Thereibre, moisture will not accumulate on or around the threads and an autoclave environment will not be created.

NRC QUESTION:

12. - Can the bolts be inspected ultrasonically for loss of preload and SCC?

SCE RESPONSE:

SCE has determined that UT is i at practical for detection orloss of preload. The hiNSA design

- employs retainer washers with bending tabs to prevent bolt rotation. A visual inspection of the l'olts and retainer washers each refueling outage will ensure that bolt rotation does not cause loss of

- preload.

. To create a SCC environment with a h1NSA installed, leakage must occur past the seal and moisture must accumulate around the susceptible materials. Inspections for evidence ofleakage will be-

. performed every refueling outage. If any evidence ofleakage is identified, the hiNSA will be -

disassembled and all hiNSA components will be inspected and replaced as necessary.

r 7

1

)

- a.

J

f i

NRC QUESTION:

j

13. - Referring to the 1985 ADB CE report, the tests were performed on tensioned bolts rather than -

i torqued bolts. Since the MNSA bolts are torqued, would there be any difference in the conclusions?

SCE RESPONSE:

l No. The crit cal factor in the development,of cracks was the stress applied to the material. Test i

i results showed that the cracks did not develop until stress approached the material yield stress. The j

torque value used in the installation of the MNSA,30 fi lbs, resulting in only 22.5 ksi stress to each j

bolt, compared to a minimum yield of 85 ksi for the material.

NRC QUESTION:

i

14. Compare the test fluid to the RCS in regards to chemistry additives (boric acid, hydrogen, i

lithium, etc.). Are there any signincant din'erences? Do they have any effect on the results?

SCE RESPONSE:

s A nominal boron concentration of 2000 ppr.r was specined for MNSA testing. Using boric acid is useful in laboratory testing because any leakage will appear as a crystalline deposit. (If deionized-water is used, it may never become visible because any leakage might simply evaporate.) Other i

chemical variations would have no impact on the corrosion resistant clamp components or the Grafoil, which is used in much more hostile industrial environments.

NRC QUESTION:

15. The MNSAsubmittalidentined SCC as requiring stresses greater than 45 KSI, but the ABB 1

CE report noted cracking at 151(St.- Elaborate?

6 SCE RESPONSE:

There were not any specimens tested at stress levels below 118.5 KSI. The one uniaxial tensile specimen that did crack had a total stress of 124 KSic The ABB CE test concluded that:

i

1. Uniaxial specimens, which assess potential for crack initiation, will fail in high temperature water, but only when highly stressed. Available data from this program and other -

investigations indicate that peak stresses exceeding 100.000 psi are required to produce stress corrosion cracking in A-286-

2. Limited test data from another vendor suggests that the threshold r. tress for crack initiation of machined fastener is signincantly higher than 100,000 psi.

'The test also indicated that precracked specimens will show crack growth when the stress intensity factor, K(1),(not the same as stress level) exceeds 45 KS1 (in. This r.lso agrees with the 6ndings of 8

O-3

-. _ _ _, _ - ~. -.. _. _ -, -. _.

I f

other investigators. At IS KSl/m, there was not any clear indication of crack growth. The

-- measurement ofone specimen at this level which was not zero crack growth (6 mils on one side after

' 7000 hours0.081 days <br />1.944 hours <br />0.0116 weeks <br />0.00266 months <br />) was within the scatter obtained with optical measurements of crack length.

NRC QUESTION:

16. What was the lubricant (if any) used on the studs in the 1985 test? Is it the same as used on the MNSA bolts? Does it have any efTect on the SCC 7 i

SCE RESPONSE:

The lab specimens were not lubricated, and thus there was not any lubricant effect on the findings. _

The MNSA bolts, like many other fasteners, are lubricated with Neolube, which is a mixture of graphite and mineral oil. Graphite is benign with respect to SCC in fasteners. Only if significant f contaminants mch as sulfur compounds, chlorides, lead, etc. are present in the mineral oil will there j

f be an effect. I abricants have been implicated in the SCC of fasteners in other applications, the most j

notable examples involving the use of molybdenum disul6de based lubricants where the reaction of 1

the lubricant with high temperature water formed 1125, which will cause cracking in a wide range of materials. The use of Neolube, especially in an application where the bolts will not be immersed and where an autoclave condition will not likely develop, will not increase the potential for SCC.

NRC QUESTidN:

17, Discuss the applications of the weld pads on the half nozzle installation and on the side shell of the pressurizer.

SCE RESPONSE:

There are two applications of weld pads with regard to nozzle repairs. The Pressurizers were originally provided with low alloy steel reinforcement weld pads on the lower side shell penetration to meet the requirements for area reinforcement of the nozzle bore in the vessel. These pads are the only low a!!oy rein %rcement pads installed for Alloy 600 penetrations and there is one on each Pressurizer When a. weld repair is made on a pressure vessel (Pressurizer or Steam Generator), an

. automated temper bead technique is employed. This technique has been used to form Alloy 690 pads l

associated with three repairs on the Unit 2 Pressurizer. One of these Alloy 690 pad buildups was placed over the original low alloy reinforcement pad on the lower shell RTD nozzle.

MNS A installation will_ occur on a reinforcement pad for the Unit 3 lower shell RTD nozzle only.

1 No new weld pads will be created for installation of MNSAs and none of the other pressure vessel jpenetrations have existing wcld pads. The process ofcreating an Alloy 690 weld pad is time and dose

. intensive. The MNSA provides a method to quickly seal the potential leak site with a signi6 cant savings in time and dose.

Y F

' c Ej.

9 d=

4 o..

,,__a

,, - m -,,.m

_,,,_.4,-_..

_a; a -

t NRC QUESTION:

18. Discuss how the potential stress risers from the holes drilled in the pipe and vessels are handled in the analysis.

SCE RESPONSE:

Stress concentration factors of 3.5 to 5 are applied in the analyses to account for the presence of the holes drilled in the pipe and vessels.

NRC QUESTION:

.i

19. What will be done with the holes for any temporary h1NSAs which are removed? Will they prevent installation of a permanent nozzle?

SCE RESPONSE:

- A threaded round head plug will be installed snug tight at the bolt hole locations for any temporary MNSA removed._ The temporary installation of the htNSA will not prevent the installation of a permanent half nozzle con 0guration. There is adequate space between the nozzle and bolt hole location to allow for the installation of a permanent half nozzle which is welded at the outside wall j

of the vessel / pipe. The design adequacy of the permanent half nozzle installation and the plugged bolt holes will be documented in a design report NRC QUESTION:

20, Discuss a Technical Speci0 cation bases change to identify how a hiNSA is treated like the pressure boundary.

SCE RESPONSE:

The first sentence in the badpound section of Bases B 3.4.13 ' RCS Operational Leakage" states:

" Components that contain or transport the coolant to or from the reactor core make up the RCS Component joints are made by welding, bolting, rolling, or pressure loading, and valves isolate connecting syste.us from the RCS."

The above statement encompasses the htNSA, and no change to this statement is required for the hiNSA to be considered the RCS pressure boundary.

In addition, the following'DASES change will be made prior to installation of a h1NSA:

-10.

o

-. L

LCO RCS operational LEAKAGE shr.ll be limited to:

Pressure lloundary LEAKAfiE No pressure boundary LEAKAGE is allowed, being indicative of material deterioration.

LEAKAGE of this type is unacceptable as the leak itself could cause further deterioration, resulting in higher LEAKAGE. Violation of this LCO could result in continued degradation of the RCPil. With the exception of LEAKAGE past a mechanical nozzle seal assembly, LEAKAGE past seals and gaskets is not pressure boundary LEAK AGE.

NRC QUESTION:

21. Consider changing from Relief Request to Alternative Repair Request under 10CFR50.55a(a)(3)(i).

SCC RESPONSE As stated in the cover letter, SCE has changed the method by which NRC approval is requested.

SCE now requests approval to use an Alternative Repair under 10CFR50.55a(a)(3)(i).

i

-Il-

• I CSCA Mechanical Joint

~

j

..m.,

,<_,m

)

/

/j/j/

/[

""E' I

h&

W

$n$

Q$

x4-N u_ imesies i

/z w xp f

xx x

y

\\,

- ommii semi p

hg'j g79 s,i, mc.,_

q h

'\\n Spare CRD tr Dxl i

i