ML20141N045

Safety Evaluation Supporting Deferment of Feedwater Nozzle Insps Until Cycle 12R Outage Scheduled for 1988

ML20141N045
Person / Time
Site:	Oyster Creek
Issue date:	02/24/1986
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20141N044	List: ML20141N043 ML20141N045
References
NUDOCS 8603030479
Download: ML20141N045 (4)

Text

[

UNITE 3 STATES g

,c 7,

NUCLEAR REGULATORY COMMISSION r,

j wasHmorou.o.c.zonsa

%,e...*/

SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION DEFERMENT OF FEEDWATER N0Z2LE INSPECTION FROM CYCLE 11R OUTAGE GPU NUCLEAR CORPORATION JERSEY CENTRAL P06iER AND LIGHT COMPANY 0YSTER CREEK NUCLEAR GENERATING STATION DOCKET N0. 50-219 1.0 INTPODUCTION, In a submittal (Ref,1) dated November 20, 1985 GPUNuclear(thelicensee) l requested a deferment of the feedwater nozzle inspection at the O ster Creek Ngclear Generating Station, This would postrone the Ultrasonic ( T? examina-tico from the Spring 1986 refueling (Cycle 11R) outage to the 1988 (Cycle 12R) outEge. The licer.cee stated that this deferrent is based on the results of analysis (Ref. f) which shows that no previous undetected flaw in the feedwater nozzle will grow to an unacceptable size during one additional operating cycle.

The staff met with the licensee on this issue at NRC/NRR headquarters in Bethesda, Maryland, on Cecember 13, 1985.

The licensee provided the material listed in References 2 and 5 to the staff at that time. This material is attached to the meeting sum.ary issued by the staff on January 2,1986.

2.0 DISCUSSIOJ In the period between 1974 and 1980, inspection of the feedwater nozzle /sparger systems disclosed some degree of cracking in the bcre and inner radius (thickest cross-section is about 7,4 in.) of reestor vessel feedwater nozzles in 18 of the 23 commercially operated boiling water reactor (BWR) plants in the United States, The staff reviewed this issue as part of Generic Technical Activity A.10. The staff's review and recomendations are docurented in NUREG-0619(Ref.3).

In NUREG-0619, the staff corr.luded that the cracks in feedwater nozzles were initiated by high cycle thermal fatigue, resulting from turbulent mixing of cold feedwater bypass 1eakage with hot reactor recircu-lation water, From analyses and experience in repairing feedwater nortles, it is known that high-cycle thermal fatigue cracks grow to a depth of about 0.25 in, before the cyclic thermal stress amplitude attenuates to an insignt-ficant level. Analyses also indicate that stainless steel cladding contributes to high-cycle themal fatigue crack initiation, l

The staff also concluded that propagation of the cracks, once initiated, could

~

result from low frequency but high amplitude stresses, which are caused by the intermittent flow of cold feedwater into the vessel during startup and shu'.down and during hot standby conditions when feedwater is added to saintain reactor water level, The frequency and magnitude of the stresses depend to a large GbO3030479 g40224 PDR ADOcK 05000219 l

G POR wv"---

.m -

...m

. degree on whether such additions are modulated smoothly or are made by an on-off control system. The licensee identified cracks along the bend radius and bore regions of the four feedwater nozzles during the 1977 refueling outage. Subsequently, the licensee removed the stainless steel cladding and a thin layer of base metal from the affected regions of the feedwater nozzles.

The licensee also replaced the thennal sleeve /sparger with an improved design, which included a pistrn ring seal and flow baffles to prevent the thermal mixing at the feedwater nozzles.

According to NUREG-0619, Oyster Creek is scheduled to have an Ultrasonic examination every two refueling cycles, a visual inspection every four refueling cycles, and a dye-penetrant test every six refueling cycles.

The licensee performed the liquid-penetrant testing immediately after the nozzle /sparger modification in 1977 (Cycle 7R outage). Two visual inspections were performed during the Cycle 8R and 9R refueling outages. No cracks were reported. The Ultrasonic examination scheduled for the Cycle 9R outage was deferred at that time with the staff's approval.

3.0 EVALUATION The staff eval e ted the licensee's feedwater nozzle crack analyses' (Refs. I and 2)'to determine (A) whether the cracks would initiate under the high cycle thermal fatigue environment; (B) whether the cracks, if they exist, would grow to an unacceptable depth during an additional operating cycle that might jeopardize the feedwater integrity; and (C) whether the licensee's analytical method is comparable to that contained in a General Electric Company (GE) report (Ref. 4), which the staff has approvd.

The licensee's analysis is divided into two parts:

(1)ahigh-cycle fatigue analysis which predicts the crack initiation, and (2) a low-cycle fatigue analysis which calculates the crack growth rate.

For the high-cycle fatigue analysis, the licensee used the GE-developed thermal load spectra, Design Basis fatigue curves and feedwater time / temperature / flow maps. The licensee also considered the effect of the flow baffles and piston rings in mitigating the turbulent mixing. The baffles prevent turbulent mixing of the cold feedwater by-pass leakage from the hot reactor recirculation flow. The piston rings minimize cold feedwater leakage in the annuius region.

Thelicenseeperformedapistonringsealleakagetest(Ref.5),whichshowed that the maximum leakage is about 0.3 gpm. Without the piston ring the leakage could be in excess of 50 gpm.

The above information was used to calculate the fatigue usage factor to estimate the possibility of the crack initiation. The resulting usage factor is 0.137 for a gap of 0.02 in, at flow baffles in 30 years. This usage factor is well below the ASME Code limit of 1.0 for crack initiation.

Therefore, the calculated fatigue usage factor shows that crack initiation due to the high-cycle thermal fatigue is unlikely. This methodology for high-cycle fatigue analysis is in accordance with Sections 4.7.2.1 and 4.7.2.4 of the GE report.

n

. The purpose of the low-cycle fatigue analysis was to show that a 0.0625 in.

deep crack in the feedwater nozzle would not grow to a size requiring repair (i.e., 0.5 in.).

It is assumed that cracks deeper than 0.0625 in, and longer than 0.25 in, would have been found by the penetrant inspection. The analysis was divided into two subtasks: (1) thermal and pressure stress analyses, and (2) crack growth rate analysis. A finite element model of the Oyster Creek feedwater nozzle was constructed using a 2-dimensional axisymmetric model of a nozzle in a sphere. The 2-dimensional axisymmetric model will result in peak hoop stresses and stress intensities in the nozzle that are equivalent to the stresses that would be obtained from a more exact 3-dimensional model.

The thermal and pressure stress analyses used a finite element analysis to obtain the nozzle stress distributions which, in turn, were used as the input l

for the crack growth rate analysis.

The thermal stress analysis considered the worst thermal transient case, where the feedwater nozzle temperature was assumed to undergo a step change from 550'F to 100*F. This transient represents the low frequency thermal shock due to on-off operation of feedwater flow during low flow conditions.

When feedwater flow is off, the feedwater nozzle temperature will approach the temperature (550*F) of the reactor recirculation water during high pressure l

hot standby. As the reactor water level is reduced to a point where I

feedwater flow is demanded, the initiated feedwater is considered to be at l

100*F assuming feedwater heaters are out-of-service. Tomodelthethermgl transients at the nozzle, heat coefficients of 150, 575, and 1000 BTV/ft

-br *F were applied to various nozzle regions. The temperature distribution l

or isotherms in the feedwater nozzle at 5,10, 30, and 310 seconds and at steady state were calculated. Thermal stresses were determined for each l

temperature distribution. The hoop stress and stress intensity in the blend l

radius were found to be a maximum at steady state.

I The pressure stress analysis used an internal pressure of 1000 psi at the i

nozzle of the finite element model. An axial stress of 2820 psi was applied at the nozzlo/ pipe junction edge to represent the pressure load from the attached feedwater pipe. The maximum hoop stress and stress intensity were found to occur on the inside surface in the blend radius area of the feedwater nozzle. The hoop stress is the dominant stress; therefore, the maximum stress intensity is approximately equal to the hoop stress.

The results of the thermal and pressure stress analyses were used in the crack growth analysis. The crack growth analysis was based on Section 4.7.3 of the GE report. The thermal and pressure stress distributions were sub-l stituted into the polynomial equations to calculate the stress intensity factors. Other inputs used in the crack growth analysis were as follows:

(A) the modified GE gcneric duty cycles, which include 130 startup/ shutdown, 349 scrams for low pressure hot standby and 62 cerams for high pressure hot standby and (B) an initial flaw of 0.0625 in. These GE generic duty cycles are more conservative than the plant-specific cycles at the Oyster Creek station.

i

. Using these inputs, the crack growth per cycle was calculated from the stress intensity factor data in conjunction with the " fatigue crack growth law."

The licensee used.two laws; one is the upper bound curve from Section XI of the ASME Code, and the other is the best estimate curve from the GE report. The results showed that an assumed 0.0625 in, deep flaw in 1977 the 1986 refueling (Cycle 11R) outage and 0.107 would grow to 0.095 in, by(Cycle 12R) outage.

in. by the 1988 refueling l

4.0 CONCLUSION

The staff concludes that, based on the review of licensee's analyses, any previous undetected flaws in the Oyster Creek feedwater nozzles will not grow to an unacceptable size during one additional operating cycle. Moreover, the licensee has already modified the nozzles by removing the stainless steel cladding and existing flaws, and installing an improved thermal sleeve /sparger design with piston rings and flow baffles.

Based on these findings, the staff judges that the UT examination of the Oyster Creek feedwater nozzles can be deferred for an additional refueling cycle to the Cycle 12R outage without degrading the structural integrity of the nozzles.

5.0 REFERENCES

1.

Letter from R. F. Wilson of GPU Nuclear to J.A. Zwolinski of NRC, subject: "0yster Creek Feedwater Nozzle Internal Inspection Defennent,"

November 20, 1985.

2.*

GPU Nuclear report, "0yster Creek Nuclear Generating Station: Evaluation of Low Flow Feedwater Control System," MPR-783, MPR Associates, Inc.,

l August 1983.

3.

NUREG-0619. "BWR Feedwater Nozzle and Control Rod Drive Return Line Nozzle Cracking," Nuclear Regulatory Comission, November 1980.

l 4.

NEDE-21821-02, " Boiling Water Reactor Feedwater Nozzle /Sparger Final Report," General Electric (proprietary), August 1979.

i 5.*

MPR Associates, Inc. report,"Feedwater Sparger Piston Ring Seal Test," MPR Associates, Inc., May 24, 1977.

Principal Contributor:

J. Tsao.

Dated: February 24, 1986.

• Material provided to the staff in the meeting with the licensee dated December 13, 1985. See meeting sumary issued January 2,1986.

n

- -