ML20072K560

Submits Response to NRC GL 94-03, Intergranular Stress Corrosion Cracking of Core Shrouds in Bwrs. W/One Oversize Drawing

ML20072K560
Person / Time
Site:	Oyster Creek
Issue date:	08/24/1994
From:	Keaten R GENERAL PUBLIC UTILITIES CORP.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
5000-94-0036, 5000-94-36, C321-94-2133, GL-94-03, GL-94-3, NUDOCS 9408300171
Download: ML20072K560 (19)

Text

. .

GPU Nuclear Corporation

.2M Z? Uhlhhf g

One Upper Pond Road Parsippany, New Jersey 07054 201-316-7000 TELEX 136-482 Writer's Direct Dial Number.

August 24, 1994 5000-94-0036 C321-94-2133 U. S. Nuclear Regulatory Commission Att: Document Control Desk Washington, DC 20555

Subject:

Oyster Creek Nuclear Generating Station (OCNGS)

Docket No. 50-219 NRC Generic Letter 94-03, "Intergranular Stress Corrosion Cracking of Core Shrouds in Boiling Water Reactors" _

The NRC issued Generic Letter (GL) 94-03 on July 25, 1994. The GL has two requests - licensees inspect the core shroud no later than the next refueling outage and perform an appropriate evaluation and/or repair based on the results of the inspection. In addition, each affected licensee must perform a safety analysis supporting continued operation until the inspection occurs.

This letter responds to reporting requirements 1 (a), (b), (c) and (d) and 2 (a) and (b).

GPU Nuclear, other utilities, and the Boiling Water Reactor (BWR) Owner's Group have been addressing various aspects of this issue since cracking at the circumferential beltline region welds was discovered at U.S. plants. It has been of particular concern to GPU Nuclear since Oyster Creek's 15th Refueling Outage (ISR) is scheduled to begin September 10, 1994. GPU Nuclear has been proactively planning to inspect the core shroud and developing a contingency repair. In other sections of this letter those plans will be described along with the other reporting requirements of GL 94-03. Given the proximity of ISR, GPU Nuclear anticipates working closely with NRC personnel over the next several weeks as our plans evolve.

The Generic Letter also refers to the Vessel and Internals Project (VIP) being coordinated by EPRI. GPU Nuclear has been very supportive of that effort with at least one representative on each committee including the executive lead of one committee. As a result, GPU Nuclear expects to support and generally follow any forthcoming guidelines.  !

9408300171 940824 PDR ADOCK 05000219 P PDR hI GPU Nuclear Corporation is a subsidiary of General Public Utilities Corporation (

" 1 C321-94-2133 Page 2 DESCRIPTION OF OYSTER CREEK SHROVD The Oyster Creek shroud is a stainless steel cylindrical assembly that provides a partition between the core region and the downcomer annulus to separate the downward recirculation flow from the upward flow of the coolant through the core. The core shroud is not a primary pressure boundary component. In addition, the core shroud is part of the fuel positioning structure which limits deflections and deformatiens of fuel bundles so that control rod insertion and core cooling is assured.

Unlike the other components of the core shroud structure, the shroud support ring was fabricated of a forging in the sensitized condition. Due to concern about stress corrosion cracking in this sensitized component and consequent failure of the ring, redundant bracket supports were installed in the field during original construction. These 36 supports (clevis linkage) are symmetrically located around the shroud and eliminate the need for the support ring. Amet Nent 40 to the Oyster Creek FDSAR addresses the structural capability of the redundant supports.

The dimensions of the Oyster Creek shroud, materials of construction, and weld locations are shown in Figure 1 (attached). Also attached is General Electric Company drawing #105E1413B which provides detailed information on the shroud weld material, carbon content and other fabrication related information. ,

I INSPECTION HISTORY  !

I The core shroud of the Oyster Creek reactor vessel has not been inspected in its entirety since it was installed. Various parts, however, have been )

inspected during past in vessel visual inspections (IVVI) examinations. In i 1983, all 36 brackets described above were inspected visually and no relevant l indications were found. In 1986, 14 of the brackets were examined and in  ;

1991, 34 brackets were examined. The results of both latter examinations were l acceptable, that is, there were no relevant indications. In each instance the '

inspection was type VT-1 with I mil wire resolution. Because of the brackets' location (see Figure 1), the H7 weld was inspected at the same time. The areas were not cleaned prior to inspection and small cracks could have escaped notice. GPU Nuclear believes, however, that cracking of the magnitude discovered at other plants would have been detected.

Additionally, in 1991, 16 lower core plate bolts were visually examined and all were found to be acceptable. In addition, the lower core plate itself was examined for general integrity which was also acceptable. Both of these examinations were type VT-3.

1

C321-94-2133 Page 3 During Oyster Creek's 14th Refueling Outage (14R) in the fall of 1992, additional elements of the core shroud were inspected. The inspection was visual (VT-1) and the camera-to-object distance varied from 4" to 18". The extent and results of that inspection are as follows:

1. Shroud to shroud circumferential weld (H5) 360 inspected from the 0.D.

2. Lower vertical weld at azimuth 80 from the I.D.

3. Shroud to shroud circumferential weld (H5) from the I.D.

4. Lower vertical weld from the 0.D.

5. Upper vertical weld at azimuth 260 from the 0.D.

6. Conical support to shroud weld (H7) from bracket 29 to bracket 32.

7. Conical support to vessel weld (H9) from bracket 29 to bracket 32.

Review of the data showed no relevant indications. In addition, an ace.ess study was performed in the annulus to determine the necessary clearances for future ultrasonic (UT) inspections of reactor vessel welds. The probe did a complete 360 inspection in the annulus and did not reveal any displacement of the shroud.

INSPECTION PLANS The inspection plan for the 15R outage is much more comprehensive than in previous outages, and includes UT as well as visual inspections. Based upon our current knowledge of the criteria being developed by the VIP subcommittees, we believe that both the inspection scope and methodology will meet or exceed those criteria. The current plan for 15R includes using UT on welds H1,'H2, H4, H5 and H6A. Some visual inspections or (if equipment is available) eddy current testing (ECT) may be used to supplement UT if this appears appropriate. Welds H3, H6B, H9, and the clevis linkage welds will be examined visually.

GPU Nuclear, through the VIP Assessment Committee, is participating in the development of generic screening criteria to be used to determine when a repair is required. GPU Nuclear intends to utilize this criteria in its evaluation of inspection results. It should be noted that should GP'J Nuclear I determine that a repair is necessary, the inspection program will be curtailed.

t C321-94-2133 Page 4 CONTINGENCY PLANS FOR REPAIR A repair contingency plan is being developed such that it will be available for installation during the upcoming 15R refueling outage. The contingency plan is to install a stabilizer system that will restore the shroud function should unacceptable cracks be found on any or all of the horizontal welds from ,

H1 to H6B (see Figure 1). The stabilizer system is designed to maintain the structural integrity of the shroud if any one or any combination of the H1 through H6B circumferential welds are unacceptable. The stabilizers are designed to withstand normal, upset and accident conditions. This design should meet the criteria of the Repair VIP subcommittee when the criteria has been finalized.

SAFETY ASSESSMENT l A safety analysis including a plant specific safety assessment which supports  !

continued operation of the plant until the inspection in September 1994 has l been prepared. The safety analysis was approved by the Plant Review Group (PRG) and is included as Attachment 1 to this letter.

Sincerely, R. W. Keaten Vice President and Director Technical Functions

)

Enclosures RWK/DK/ pip l cc: Administrator, Region I NRC Resident Inspector Oyster Creek NRC Project Manager A

i

. = 189" O.D.

- A A A O HI f 37j~

/ CORE SPRAY SPARGERS L Y TOP GUIDE FLANGE - ,,

r H3 i REACTOR PRESSURE VESSEL

\ / H4 CORE SPRAY PIPE 18 2I" y ali 3" I

N

• n

// -N'

-Y' 90f" 238"

/

/

= l79" O.D ,

g V CORE PLATE FLANGE / _._, _ _. i l2n H6A 4,

/ ) ,

SUPPLEMENTAL CLEVIS LINKAGE (36) '

7 H6s H7 22*1" INCONEL LUG 304 S.S.

CLEVIS & PIN /

(BRACKETS) sO'

(/ -

N Y

/

/

s _

H8 5"

1 i

- SUPPORT RING 304 S.S. FORGING

/

/

g H9 INCONEL SUPPORT CONE FIGURE I OYSTER CREEK CORE SHROUD CROSS SECTION NOTE: ALL SHROUD MATERIAL MADE FROM 304 S.S.

PLATE UNLESS OTHERWISE NOTED

ATTACHMENT 1 OC CORE SHROUD EVALUATION 1.0 Introduction The purpose of this document is to address the Oyster Creek specific shroud concerns outlined in the GE assessment of core shroud cracking. This document in conjunction with the GE assessment will constitute an operability determination for Oyster Creek until the September 1994 refueling outage.

The Oyster Creek Shroud is a stainless steel cylindrical assembly that provides a partition between the core region and the downcomer annulus, to separate the upward flow of coolant through the core from the downward recirculation flow. In addition, the shroud provides a support for the separator assembly as well as lateral stability for the core geometry.

The shroud is not a primary pressure boundary. The Oyster Creek shroud welds and their designations are depicted in Figure 1. Cracking of similar shroud welds have been identified in a number of BWR plants.

2.0 Backaround The BWROG has been requested by the NRC to undertake an assessment of BWR susceptibility to shroud cracking. General Electric under contract to the owners group developed a document in response to the NRC request (Reference 5). Oyster  ;

Creek - a BWR 2 - is identified as somewhat of an outlier in i this evaluation. Therefore, GPU Nuclear has performed plant I specific evaluations and identified plant specific design features to be used to assess the consequences of Oyster Creek shroud cracking.

l 3.0 Shroud Welds The shroud is composed of a number of cylindrical sections welded together. Those welds in the horizontal plane represent an important feature in the structure of the shroud.

Vertical welds are also present in the structure assembly, but do not provide any significant role in terms of the loads imposed upon the shroud. In addition, the industry experience ,

shows that the vertical welds have not exhibited any l significant cracking (Reference 9). Therefore, the focus of this evaluation is on the horizontal welds depicted in Figure 1.

l

4.0 Oyster Creek Issues Identified By GE work The GE document (Reference 5, page 42) identifies several design differences of the BWR 2 plants. These differences listed below are addressed for Oyster Creek in this document.

1. Recirculation flow enters the reactor vessel from the bottom

2. ECCS for large breaks are 2 redundant, double capacity, core spray systems

3. short and long term cooling responses for large recirculation line breaks rely on core spray, as the vessel will not flood

4. failure of the H8 weld (without consideration of the in place brackets) results in the shroud having little vertical support

5. the reactor vessel has two instead of four steam lines

6. accident plus seismic is not a design or licensing basis 4.1 Oyster Creek Response To GE Report Description Of BWR 2 As outlined above there are differences between the BWR 2 plant and the remaining plant designs. Specifically these dif f erences have an ef fect on the response to the design basis accidents.

Item 1 is that the recirculation flow enters the reactor ,

I vessel from the bottom of the vessel. The main result of this is that the vessel is completely drained and does not re-flood )

following a design basis recirculation loop failure. j Therefore, effective core spray cooling is required until the l containment can be flooded to a level above the TAF. As a l result, it is necessary to demonstrate that any cracking in i the shroud will not lead to core spray system damage following '

l a DBA LOCA. The discussion provided in section 4.1.1 shows that this is not a concern for Oyster Creek.

Items 2 & 3 cover the fact that the ECCS for large breaks are two redundant, double capacity, core spray systems which provide the sole short term cooling following a DBA LOCA.

Thus the ability to cool the core following a DBA LOCA is directly linked to the effective operation of these systems.

These items are linked to item 1 above. The evaluation of the DBA LOCA (Section 4.1.1) shows that this is not a concern for Oyster Creek.

l i

4 Item 4 is that the failure of the H8 weld will result in downward displacement of the core shroud and can damage the core spray system piping following a DBA LOCA. As a result of prefabrication structural issues Oyster Creek had installed 36 support brackets (Figure 1) which are symmetrically located around the base of the core shroud. These brackets provide additional support to structurally replace furnace sensitized components in the original design. The brackets were designed to fully support the shroud under original design basis loads.

The brackets are welded to the shroud with material which is compatible with the 304SS of the shroud (Reference 1). In addition, the bracket is welded to the inconel cone with a compatible weld material . These brackets will provide support given the loss of either the H8 or H7 welds (or of H8 and H7).

Therefore, the downward displacement of the Oyster Creek shroud will not occur if the H8 or H7 weld fails.

The support bracket welds are considered to be in good condition. They were visually inspected albeit without cleaning. In addition, their location in the lower part of the downcomer affords them good H2 water chemistry protection and little if any radiation assisted stress corrosion cracking concern.

l Oyster Creek also has an H9 weld shown in Figure 1. This weld !

connects the inconel support cone to the RPV and provides  !

vertical shroud support. The weld was included as part of. the original vessel fabrication and as such it received the same '

stress relief treatment as the vessel. Therefore, the  ;

residual stress associated with its fabrication was reduced. i The stress relief minimizes the possibility of crack development and growth. In addition, the H9 weld location -

bottom of the downcomer - receives good H2 water chemistry protection removing concern for irradiation assisted stress corrosion cracking. As a result, significant crack development is not expected in this weld. An additional consideration regarding H9 is that it is a back groove weld that is much thicker than the rest of the support cone.

Therefore, substantial crack growth is necessary before it becomes of structural significance. The combination of the .

stress relief, H2 water chemistry, minimal radiation influence l and weld geometry lead to high confidence in this weld's  !

integrity.

Item 5 is that the reactor vessel has two steam lines instead of four which has a direct impact on the Main Steam Line Break assessment. With fewer steam lines, the upward force on the shroud following a break will be greater due to the higher break flow through the one larger line. This issue is i addressed by GPU Nuclear using plant specific analyses as discussed in section 4.1.2 below.

Item 6 is that accident plus seismic is not- a design or licensing basis condition. The brackets discussed under item 4 were designed to withstand an SSE in conjunction with a main steam line break or an OBE in conjunction with a DBA LOCA. In addition, an SSE in conjunction with a DBA LOCA was  !

considered. For the SSE/LOCA evaluation, allowable stress limits were exceeded with permanent deformation that did not affect the ability to mitigate the LOCA.

4.1.1 DBA LOCA Evaluation The GE report (Reference 5) states that with a failure in weld H8 in conjunction with a large recirculation line break, the shroud will experience a vertical displacement of several inches downward. This displacement will damage the core spray system virtually eliminating its ability-to provide short term cooling. This is a result of items 1, 2, 3 & 4 listed above.

However, Oyster Creek is equipped with additional shroud support capability in the form of 36 brackets attached at the base of the core shroud (refer to Section 4.1, Items 4 & 6 above). The existence of these brackets prevents any vertical displacement of the shroud and thus prevents damaging the core spray system piping. For the design basis event GE states-that shroud displacement is not expected at any but the vertically unsupported H8 weld. For Oyster Creek this displacement is prevented by the aforementioned brackets. The structural integrity of the H9 - weld is assured also, as discussed above. Therefore, the failure of any other welds in or directly associated with the shroud in conjunction with a DBA LOCA will not endanger the core spray system. It is concluded that the core spray system will function as designed.

For the DBA LOCA there is also a concern associated with i asymmetric loads on the shroud. The failure of ~ a ,

recirculation loop suction line will produce an asymmetric l depressurization of the vessels annular region. The annular '

region nearest the break _ will depressurize faster than the region 180* away. This produces a potential concern that the shroud will have tipping loads imposed upon it by the <

unbalanced pressure forces. However, the asymmetry associated )

l

with the annular region is an acoustic effect, and as such is very short lived (milliseconds). The short duration is a direct result of the pressure wave propagation around the annulus at the speed of sound rapidly equalizing the annular pressure. This very short loading interval will not be  ;

sufficient to tip the shroud if it were detached.

4.1.2 Main Steam Line Break Evaluation The GE report (Reference 5) points out that the BWR 2 is equipped with two rather than four main steam lines. With fewer steam lines, the break flow through a broken line will be greater for a BWR 2 as compared to a four steam line plant.

The above core main steam line break produces forces which cause upward motion of the shroud. Therefore, with a failure in any of the welds below the top core guide (welds H1 and H2 are excluded) the potential to raise the guide up above the top of the fuel assemblies exists. This will cause damage to the core spray piping inside the vessel and could allow fuel geometry to go askew preventing complete insertion of the control rods. The GE report does point out that for the scenario where the break is above the core, the vessel will reflood and effective spray cooling is not critical to ensuring adequate core cooling. This conclusion is consistent -

with the core cooling assessment for Oyster Creek following a failure of the main steam line. The only issue of concern is then that the rods do not fully insert. .The GE report goes on to say that more detailed accident analyses would probably show that the top guide lif t would be limited to below the top of the fuel assemblies.

GPU Nuclear has performed detailed thermal hydraulic analysis ,

(Reference 4) of this condition using the RELAPS computer code. The thermal hydraulic analyses perfonned are used to establish an upper plenum to lower downcomer pressure difference.

l The Oyster Creek Reactor Vessel model used to assess this '

condition employs a dynamic representation of the shroud displacement. The dynamic representation (Reference 4) accounts for the forces on the detached section. As the l detached shroud section rises it opens a gap in the shroud which relieves the upward forces acting to lift the shroud. )

The purpose of this evaluation is to determine if the upward I displacement would cause the top core guide plate to clear the )

fuel assemblies. For Oyster Creek an upward displacement of 13.31 inches is required to clear the top of the fuel (Reference 11).

l

l I

A With the dynamics of the shroud displacement modeled a full instantaneous guillotine failure of a main steam line is assumed. The instantaneous failure is a very conservative assumption since even a fraction of a second break development time would offer relief in the maximum shroud displacement.

The evaluation shows that the maximum displacement is 13.285 inches. This displacement is below the top of the fuel assemblies (approximately 13.31 inches or 13.51 inches if irradiation effects are considered). Since the shroud displacement is below that required to clear the fuel channels, there is no concern associated with this event. In addition, the dynamic modeling of the displaced shroud shows that the shroud would return to rest with a force of no greater than its weight. This is a result of the shroud dropping down closing the gap and allowing the lif ting forces to increase within the upper plenum. Therefore, the return to rest of the separated chroud section will not impart significant loads to the reactor vessel.

1 l

. . _ - - . - - - .-- . . - - - _ . . = _ _. . . - . .

5.0 Plant Response To a Weld Failure Under Normal Conditions The failure of a shroud weld while operating the plant under

• normal conditions may cause the detached shroud section to separate and rise. The height to which the section will rise is related to the balance between the lifting force and the weight of the detached member. As a gap in the shroud develops due to the elevation change of the detached member the lifting forces are relieved through the gap. Once the  :

lifting force equals the detached shroud members weight the ,

member will no longer rise. The GPU Nuclear calculation (Reference 5) shows that this is on the order of 1.0 inch (H3 weld for lower welds the value is smaller).

i With the detached shroud section floating 1.0 inch above the remaining section there exists several things to consider.

First, will the free floating shroud section rotate under the influence of separator flow induced forces? "econd, are there  ;

any forces which will act to shift the shroud laterally? '

Third, what will be the impact of a seismic event during normal operation with one of the shroud welds failed?

Finally, will there be any indications that the operators will have to forewarn them of a problem? The lateral or rotational movement of the shroud can disrupt the core geometry and inhibit the insertion of control rods. This is a concern for the welds below H2 only since the core support plates are  ;

below H2 and will not shift if the shroud separates at or above H2. j The first issue associated with the rotation of the shroud caused by reactionary forces within the steam separators is not a concern. The act of turning the fluid imposes a force  ;

on the fluid which imposes a torque on the separator which is ,

translated to the shroud through the standpipes. This is not l expected to induce a significant enough load to overcome the  !

inertia of the shroud. In addition, the detached shroud l member under this condition will resist any rotation through .i the existence of the separator guide rods attached to the l vessel as well as the core spray piping which penetrates the l shroud above the weld H3 between H1 and H2.

The second issue associated with lateral displacement of the shroud due to normal operating forces is not a concern. There are no asymmetric flow patterns within the reactor vessel under full loading conditions. ' Asymmetric forces which may develop due to turbulent flow patterns within the vessel are expected to be minimal. It is concluded that the forces are symmetrically balanced or are of a minimal differential size that no displacement or tipping is expected. If, however, asymmetric loading were to be encountered the core spray piping as well as the separator guide rods would prevent.

lateral movement.

l l

l The third issue is that of seismic loading under normal l operating conditions. The primary concern associated with a j seismic event is the impact that shroud movement will have upon control rod insertion. Lateral movement of the detached I

shroud section may change the core geometry and prevent rod insertion. This concern is associated with welds H3 and below; the upper welds - H2 and H1 - are not required to provide lateral stability to the core geometry.

If the design basis earthquake (OBE) is considered with a complete through wall failure in a shroud weld a bounding assessment of shroud movement can be perf onned . First consider the insertion of control blades during OBE seismic events considering a separated shroud at H3. Dynamic test results (Reference 12) show that a control rod insertion can occur with large cyclic fuel channel full-span displacements.

The maximum differential displacement of a parted core shroud can not exceed the maximum seismic ground displacement. For Oyster Creek the maximum seismic ground displacement is 7.92 inches for SSE or 3.96 inches for OBE. The actual differential displacement is less than this due to .

1. The lateral displacement of the vessel pedestal, skirt, and internals

2. Lateral displacement due to water in the vessel

3. Lateral displacement due to the attached core spray piping

4. Lateral displacement due to the fuel contacting the top guide For an OBE, which is a higher probability earthquake, the actual dif ferential displacement of a parted core shroud at H3 will not exceed 3.96 inches. The aforementioned test results show that this displacement, considered dynamically, will not prevent control blade insertion for an OBE. For the SSE the argument is more qualitative in terms of the displacement being limited by the four items listed above. In addition, the SSE has a lower probability of occurrence than the OBE.

For the lower weld locations the separated shroud section is not lif ted upward by the normal operating pressure conditions.

Therefore, considering the insertion of control blades during an OBE seismic event with a failed joint at a lower weld location (below H3), credit can be taken for friction to transmit shear through the joint especially since only rigid -

body motion will occur at the very low frequencies at which the maximum seismic ground displacement occurs. Due to the l

friction as well as the aforementioned displacement limiting features listed above, little or no differential displacement will develop.

To date no plant has identified a complete 360 through wall crack in any shroud weld. It is more likely that Oyster Creek may have some lesser crack to contend with during any postulated seismic event. Calculations (Reference 10) show that a single weld with a through wall crack of up to 360 inches in length (64% of the circumference) can withstand the combined seismic and normal operating loads. A seismic event in combination with a through wall crack which is larger than 360 inches could shift the shroud. However, the displacement would be less than that previously discussed for the complete weld f ailure. Therefore, complete control rod insertion is to be expected and no fuel damage would occur.

It is important to note that no plant has found cracks of this size. It should also be noted that, the larger the crack the more likely the detection of it. This holds true for all weld locations with the exception of H4, H5, and H6A. At these locations, as stated in Reference 5, the net force is downward and the crack will not open up.

Based upon the above, it can be said that a substantial crack would be required for a seismic event to be a concern. The crack would need to be larger than any detected in the industry to date. In addition, a crack of this size would result in a decrease in reactor performance and is potentially detectable. Finally, the probability (5.51E-5 per two months) that an earthquake would cause significant movement is considered to be small.

The final consideration is that of the ability of the operator to recognize a failed weld. As is pointed out in the Appendix A of the GE report, the operator would observe power changes due to core bypass flow being diverted to the downcomer. This power reduction would occur without a corresponding reduction in recirculation flow. The discussion in the appendix to the GE report is consistent with what would be expected at Oyster Creek.

l 9

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

)

l 6.0 Additional Consideration In addition to the above evaluation Oyster Creek considerations are provided which include inspections, H 2 I water chemistry, and probability of the event. l P

6.1 Oyster Creek Weld Inspection History i

In 14R Oyster Creek performed in-vessel visual inspections (IVVI) of weld HS (Reference 6) without any surface preparation. This inspection showed that both the ID and OD of this weld did not have any reportable indications. In addition to weld H5 the inspection also included several vertical welds as well as the weld to support ring on the brackets 29 and 32. As with weld H5 no reportable indications were observed.  !

This inspection technique is not conclusive since the weld was not cleaned prior to inspection and was not of fine enough resolution to detect very small cracks. However, it did  ;

demonstrate that there was no gross cracking present.

In 14R a submarine carrying a camera was used to perform an access study in the annulus to determine necessary clearances for future vessel weld UT inspections. This 360* visual-inspection of che. annulus-did not reveal any displacement of the shroud. Therefore, as of that inspection these welds were providing the necessary vertical support to the shroud.

6.2 Benefit of Hydrogen water chemistry for non jet pump BWR2 (

Oyster . Creek installed a hydrogen water chemistry system during cycle 12. Although the hydrogen water chemistry system was not installed until later in plant life and therefore cannot be credited for prevention of crack initiation it certainly is available to minimize continued crack growth.

In-core measurement (Reference 7) and radiolysis modeling (Reference 8) for a non-jet-pump plant like Oyster Creek shows that when mitigation is achieved in the. recirculation system piping, the reactor vessel' lower plenum also is in the IGSCC.

protection regime and that oxidizing conditions (oxygen and peroxide concentrations) in the lower core. bypass region are reduced to the point where cracking and crack growth rates would be significantly -reduced. For Oyster Creek this protection would apply to the H8, and H9 weld in both the downcomer region and - the lower plenum. In addition, the shroud support bracket welds located in the lower downcomer benefit from the H2 water chemistry.

9 4

6.3 PRA evaluation of Piping Failure In calculating the probability of a large loss of coolant accident which can result in core shroud movement it is not appropriate to use the frequency given in the Oyster Creek Probabilistic Risk Assessment (OCPRA). The OCPRA large LOCA initiator group includes recirculation suction and discharge breaks and main steam line breaks inside containment as well as feedwater line breaks and core spray over-pressurization.

In addition, this group evaluates medium as well as large breaks. Less conservative and more current frequencies are given in " Pipe Break Probabilities in Boiling Water Reactors",

a BWR Owners Group Report, BWROG-93149, November, 1993.

In this reference the probability of a large pipe break is given as 7.5x10-6 per reactor year. The probability of this LOCA during the two month exposure interval (time before planned inspection) is given by the equation:

Probability of Large LOCA x Exposure Interval 1

7 . 5x10 -6 x 2 months /12 months per year 1.25x10~6 per 2 month interval i

This probability represents the occurrence of a large pipe  !

break (greater than 6 inches effective pipe diameter) in the I reactor coolant system. The case of the main steam piping  ;

breaks can be estimated by the multiplication of the main ,

steam piping sections (2) divided by the total RCS piping l sections (12) (i.e., recirculation piping sections (10) and.  ;

main steam piping sections . (2) ) . The probability of a main I' steam line break over the two month interval is:

2/12 x 1. 25x10-6 Large LOCA Probability 2.08x10 4 per 2 month interval The probabilities above reflect the occurrence .of the initiating event not necessarily the occurrence of core shroud movement. This phenomena (i.e., core shroud movement) requires the shroud to have a 360 and greater than 90%

through wall circumferential crack which is unlikely and further reduces the probability of an undesirable outcome.

This probability is sufficiently low to justify continued operation until the planned inspections can be performed.

\

7.0 Conclusions Weld inspections which have been performed show that gross weld cracking is not present in those locations inspected.

Although past inspections are not conclusive nor complete, they do provide some degree of confidence that severe cracking is not present.

A probability assessment which does not credit either OC specific structural features or thermal hydraulic analyses shows that the probability of a LOCA with shroud failure is very low. Based upon these probabilities it is considered acceptable to continue to operate until the September 1994 refueling outage.

Finally, the BWR 2 differences raised in the GE report were addressed. It is concluded that those design differences raised in the GE report in regard to BWR 2's are not service limiting factors to Oyster Creek. In particular the H8 weld issue is not applicable to Oyster Creek. With H 2 water chemistry in use at Oyster Creek any cracks which may have initiated early in plant life would have minimal growth following its full implementation. Even if the weld were to fail, the existence of the additional shroud support brackets will prevent the vertical displacement of the shroud following a recirculation loop LOCA.

The remaining weld concerns attributed to the main steam line failure are addressed by the plant specific analysis performed by GPU Nuclear. For all Oyster Creek design basis accidents the ECCS will perform as required to ensure adequate core cooling.

The combination of the GE report (Reference 5) as well as plant specific evaluation outlined in this document provide justification for the continued safe operation of Oyster Creek for the remainder of the present fuel cycle.

r '

l l

8.0 References

1. OC FDSAR, Amendment 40.

2. OC Probabilistic Risk Assessment (Level 1), page 4.6-19. )

3. Database for Probabilistic Risk Assessment of Light l Water Nuclear Power Plants, PLG-0500, Volume 7, Rev. O, page 2-3. j l

4. GPU Nuclear Calculation, C1302-222-5450-006, Rev. O, )

"OC Shroud Displacement Calc".

l

5. GENE-523-A107P-0794, "BWR Shroud Cracking Generic )

Safety Assessment".

6. GPU Nuclear Memo 6142-93-065, " Core Shroud and Shroud Bolting", R.T. Nademus to J.D. Abramovici, October 21, 1993.

7. EPRI NP-7200-M, " Measurement of In-Core and Recirculation System Response to Hydrogen Water Chemistry at Nine Mile Point 1", March, 1991.

8. EPRI NP-6386, "Modeling Hydrogen Water Chemistry For BWR Applications", June, 1989.

9. GENE-523-110-0794 DRAFT REPORT, " Justification For Not Postulating a Full Plate Length Axial Flaw in a Shroud Evaluation".

10. GE-NE-523-173-1293, " Evaluation and Screening Criteria for the Oyster Creek Shroud".

11. GPU Nuclear Calculation, C1302-226-5411-315, Rev. O,

" Comparison of Relative Vessel Positions Between the Fuel & Top Core Guide".

12. GENE 771-44-0894, " Justification of Allowable Deflection of the Core Plate and Top Guide Shroud Repair".

s OVERSIZE DOCUMENT l PAGE PULLED l SEE APERTURE CARDS f

NUMBER OF OVERSIZE PAGES FILMED ON APERTURE CARDS i 74090/ w g 1

APERTURE CARD /HARD COPY AVAILABLE FROM RECORDS AND REPORTS MANAGEMENT BRANCH l

l l

l i