ML20136H834

Justification to Defer Ultrasonic Exam of Feedwater Nozzles One Addl Fuel Cycle

ML20136H834
Person / Time
Site:	Oyster Creek
Issue date:	11/01/1985
From:	GENERAL PUBLIC UTILITIES CORP.
To:
Shared Package
ML20136H817	List: ML20136H813 ML20136H834
References
NUDOCS 8511250206
Download: ML20136H834 (23)

Text

' s.

0,-.

l t

GPU NUCLEAR CORPORATION OYSTER CREEK NUCLEAR GENERATING STATION JUSTIFICATION TO DEFER ULTRASONIC EXAMINATION OF FEEDWATER NOZZLES ONE ADDITIONAL FUEL CYCLE November 1, 1985 e

i 8511250206 851120 PDR ADOCK 05000219 P PDR

t .

l l

TABLE OF CONTENTS l

1. INTRODUCTION

2. BACKGROUND 2.1 OYSTER CREEK FEEDWATER NOZZLE CRACKING 2.2 ACTIONS TAKEN AT OYSTER CREEK TO' REDUCE FEEDWATER NOZZLE CRACKING

3. JUSTIFICATION TO DEFER ULTRASONIC EXAMINATION OF FEEDWATER NOZZLES 3.1 CRACK INITIATION CONSIDERATIONS 3.2 CRACK GROWTH CONSIDERATIONS

4. REFERENCES r

i l

9

, s e s l

l

1. INTRODUCTION Previous Oyster Creek inservice inspection plans called for ultrasonic (UT) examination of the four reactor vessel feed-water nozzles during the Spring 1986 refueling outage. This report provides the technical justification to defer UT examination of the feedwater nozzles for one additional fuel cycle, i.e., from the Spring 1986 refueling outage (llR) to the next scheduled refueling outage (12R) . The technical justification is based on the results of tests and analyses which demonstrate that any previously undetected flaws in the feedwater nozzles will not grow to an unacceptable size during one additional fuel cycle from llR to 12R and there-fore UT examination of the feedwater nozzles can be deferred to 12R without jeopardizing the integrity of the feedwater nozzles.

i

! 1-1 l

e *

2. BACKGROUND 2.1 OYSTER CREEK FEEDWATER NOZZLE CRACKING During the 1977 refueling outage, the Oyster Creek feedwater nozzles were liquid penetrant (PT) inspected from the inside surface of the nozzles with the following results.

Number of Nozzle Indications Lengths 45* 36 0.25"-4.0" 135* 3 1.0"-2.5" 225* 4 2.0"-12.0" 315* 11 1.0"-12.0" After removal of the stainless steel cladding and a thin layer of base metal from the blend radius and bore regions of the feedwater nozzles, the nozzles were reinspected by the liquid penetrant method. The remaining indications were removed by grinding. The number of indications after clad removal and the maximum crack depths (from the original clad surface) were as follows:

Number of Maximum -

Nozzle Indications Depth 45* 5 0.54" 135" 0 ---

225' 4 0.46" 315' 3 0.62" 2-1

Evaluations by GE and NRC, reported in References 1 and 2, respectively, determined that the primary cause of the feedwater nozzle cracking at Oyster Creek and other BWRs was thermal fatigue, and that there were two fatigue mechanisms present; a high cycle mechanism which initiated the cracks, and a low cycle mechanism which caused the cracks to grow.

High Cycle Fatigue The original feedwater spargers at Oyster Creek and other BWRs in the United States were installed with a clearance (10-25 mils) between the OD of the feedwater sparger thermal sleeve and the ID of the seal land surface of the feedwater nozzle bore. Bypass leakage of feedwater in the annulus between the thermal sleeve and feedwater nozzle bore resulted in high frequency thermal cycling in the nozzle bore and blend radius regions due to turbulent mixing of ' cold' feedwater bypass leakage with the ' hot' reactor recirculation water. Full scale tests and field measurements by GE determined that, with bypass leakage present, the metal temperature cycling at the feedwater nozzles was about 50 percent of the temperature difference between the feedwater and the reactor recirculation water. The observed frequencies were between 0.1 and 1 Hz. Results of fatigue analyses indicated that such high frequency thermal cycling could lead to initiation of thermal fatigue cracks in three months to three years, depending on the normal feedwater temperature and~the duration of operation at low feedwater temperature. Analyses and field experience indicate that cracks initiated by the high frequency thermal cycling mechanism described above will arrest at a depth of about 0.25" from the metal surface due to attenuation of the high frequency thermal stresses.

2-2

T l

I t

l l .'.

j Low Cycle Fatigue Fracture mechanics analyses indicated that cracks in the )

feedwater nozzle initiated by high frequency thermal cycling could grow to depths greater than 0.25" by the combination of low frequency pressure cycles due to reactor startup and shutdown and thermal shocks due to on-off operation of feed-water flow to the reactor vessel during periods of low feed-water demand (i.e., startups, shutdowns, and hot standby l conditions). When feedwater flow is off, the temperature of

! the feedwater nozzle will approach the temperature of the reactor recirculation water (550*F at high pressure hot standby). When feedwater flow is added to maintain reactor water level, the feedwater nozzle will be cooled by the feed-water flowing into the reactor vessel through the feedwater sparger thermal sleeve. Analyses indicated that\on-off operation of the feedwater flow to maintain reactor water level within the prescribed limits during low flow conditions could result in up to six thermal shocks per hour.

l 2.2 ACTIONS TAKEN AT OYSTER CREEK TO REDUCE POTENTIAL FOR FEEDWATER NOZZLE CRACKING During the 1977 refueling outage, the following actions were L taken at Oyster Creek to reduce the potential for additional l feedwater nozzle cracking.

Removal of Stainless Steel Cladding The stainless steel cladding (nominal thickness 0.219") and a thin layer of base metal (approximately 0.188") were removed from the feedwater nozzle blend radius and bore regions using l machining equipment designed by Combustion Engineering.

! Removal of the stainless steel cladding and a thin layer of base metal reduces the thermal stresses in the feedwater nozzle, eliminates material at the high stressed areas of the 2-3 l

1

• =g. >

. F.

feedwater nozzle which may have already been subjected to

significant-thermal fatigue damage, and improves the sensi-tivity of UT examinations from the outside surface of the nozzle.

Removal of Defects-Following removal of stainless steel cladding and a thin layer of base metal, all detected cracks remaining in the feedwater nozzles were removed by grinding. The grind-out cavities were blended into the nozzle to reduce stress concentrations.

Installation of Replacement Feedwater Sparger of Improved Design The original feedwater spargers were replaced with feedwater spargers of an improved design with respect to thermal cycl-ing (see Figure 2-1). The replacement feedwater sparger eliminates the potential for new crack initiation due to high frequency thermal cycling and improves the effectiveness of the thermal sleeve in-insulating the feedwater nozzle.from low frequency thermal cycling due to thermal shocks during low flow conditions. The main features of the replacement

.feedwater sparger which provide these improvements are:

a. Piston Rings -- The replacement feedwater sparger incorporates a stainless steel piston ring with an Inconel X-750 backup spring at the end of the thermal sleeve (see Figure 2-2). The purpose of the piston ring seal assembly is to minimize bypass leakage of feedwater in the annulus between the OD of the feedwater sparger thermal sleeve and ID of the feedwater nozzle bore.

2-4

.'.~

b. Flow-Baffles -- The replacement feedwater sparger includes two Inconel X-750 flow baffles (also shown in Figure 2-2) . The purpose of the flow baffles is to prevent-thermal mixing of the relatively cold feedwater bypass leakage with hot reactor recirculation water in the bore'and blend radius regions of the feedwater nozzle. Each flow baffle contains two leak-off holes for. discharge of the feedwater bypass leakage flow into the reactor vessel. This allows a small, continuous

' flow of feedwater bypass leakage in the annulus between the thermal sleeve and nozzle bore and stagnant conditions in the annulus are avoided. The double flow baffle arrangement, with leak-off holes between the baffles, also assures that there will be no significant leakage of feedwater bypass leakage past the periphery of the. flow baffles-(i.e., against the reactor vessel surfaces).

The replacement feedwater sparger provides two essentially independent methods of eliminating significant high frequency' thermal cycling which resulted in high cycle fatigue crack initiation in feedwater nozzles fitted with the original loose-fit feedwater spargers. First, results of full size - N piston' ring seal leakage tests show that the single piston ,

rings will limit feedwater bypass leakage to an extent that high cycle fatigue cracking is mitigated, even without the flow baffles. Second, full size two temperature tests show that the flow baffles will also prevent significant high

frequency thermal cycling, even at feedwater bypass leakage flow' rates that are far greater than expected even if the piston rings were in a degraded condition. These feedwater sparger: design features and test results are discussed in more detail in Section 3 of this report. As a result, the combination of the piston' ring and flow baffles on the 2-5

replacement feedwater sparger essentially eliminates any new crack initiation by the high frequency thermal cycling mechanism. Liquid penetrant inspections of one feedwater nozzle at Nine Mile Point Unit 1 in 1981 confirmed this; no new crack initiation occurred after about four years of operation with an identical replacement feedwater sparger.

With regard to low frequency thermal cycling due to normal operating transients, while the replacement feedwater sparger significantly improves the effectiveness of the thermal sleeve in insulating the feedwater nozzle from thermal shocks due to on-off operation of feedwater flow, results of analyses presented in Section 3 of this report indicate that such thermal shocks would still propagate any undetected, pre-existing flaws in the feedwater nozzles to some extent.

The rate of growth of any such flaws and the estimated maximum size after an additional fuel cycle are addressed in the following sections of this report.

I i

i 2-6

x Reactor Vessel

\, 7 Sparger Thermal Sleeve

\1 A A a> h

:,.,k

_N: ~ =_ s, D.D

(_ f""

\ Feedwater Nozzle

' 'N,' 1- Flow Baf fle

.b., U,

~

Piston Ring

.']'

~ . ,.

[a . l\,.

%' .. I .. . -.

s

. - g, .,r._,r.,,. .;.: -

_rz, A s-( '.','N l\. N L.'.3 Section A-A OYSTER CREEK REPLACEMENT FEEDWATER SPARGER FIGURE 2-1

l I

l IFeedwaterNozzle ~

~

f Piston Ring l

I .

.x a

N ANN'NNNNNNN'NL3 spring Thermal sleeve ,

Platon Ring Detail Feedwater Nozzle Oh Mh Inner Baffle See Detail Above

.s s .

y ]3 s- s s s s s s s s s s

/s s/s s/T\QgM t - ,

N s '% -

Sparger s l s

' Thermal Sleeve Flow Baffle Detall l

OYSTER CREEK REPLACEMENT FEEDWATER SPARGER PISTON RING AND FLOW BAFFLE FIGURE 2-2

3. TECHNICAL JUSTIFICATION TO DEFER ULTRASONIC EXAMINATION OF FEEDWATER NOZZLES This section of the report provides the technical justifica-tion to defer. ultrasonic (UT) examination of the Oyster Creek reactor vessel feedwater nozzles for one additional fuel cycle, i.e., from the Spring 1986 refueling outage (llR) to the next scheduled refueling outage (12R). The technical justification is based on the following considerations.

a. No new crack initiation by high frequency thermal cycl-ing is expected in the feedwater nozzles as a result of the 1977 repairs to the feedwater nozzles and the installation of the improved feedwater sparger design with piston rings and flow baffles. On this basis, the largest undetected flaw in the feedwater nozzles would be a flaw inadvertently left in the nozzles during the 1977 repairs. This flaw size is conservatively taken to be 0.062" deep based on the 1977 liquid penetrant examination acceptance criteria.

b. Results of fracture mechanics analyses indicate that an assumed 0.062" initial flaw left in the feedwater nozzles in 1977 would grow to 0.095" by the Spring 1986 refueling outage (llR) and would grow an additional ,

0.012" to a total depth of 0.107" by the next scheduled-refueling outage (12R). Flaw sizes of this magnitude are considered acceptable from the standpoint of feed-water nozzle integrity and are well within existing nozzle reinforcement and wall thicknesses. They are also significantly less than the flaws discovered and repaired in 1977.

3-1

These considerations are discussed in detail in the following sections of this report.

3.1 CRACK INITIATION CONSIDERATIONS As discussed previously in Section 2.2, above, the replace-ment feedwater spargers. installed in 1977 contains single piston rings and double flow baffles. The piston rings reduce bypass leakage to a low value and the flow baffles prevent thermal mixing of the " cold" feedwater bypass leakage with the " hot" reactor recirculation water in the bore and

. blend radius regions of the feedwater nozzle.

The piston ring seal assembly was designed to have sufficient compliance to accommodate the expected worst case differ-ential thermal expansion between the OD of the thermal sleeve and ID of the seal land on the feedwater nozzle bore during conditions of varying.feedwater temperature. Results of fell scale tests indicate the single piston ring seal assembly will limit bypass leakage to less than 0.3 gpm for the expected worst case conditions. Tests simulating degraded seal. conditions (i.e., a failed piston ring) resulted in a maximum bypass leakage of 1.5 gpm. These bypass leakage flow rates are.substantially less than the bypass leakage without piston rings (estimated to be in excess of 50 gpm) .

The effectiveness of the flow baffles in limiting thermal cycling in the blend radius region of the feedwater nozzle.is shown in Figures 4-32 and 4-33 of Reference 1. These figures show that for feedwater bypass leakage flow rates up to 15 gym, the metal temperature cycling at the nozzle blend radius is less than 17 percent of the available delta T (i.e., the difference between reactor recirculation water temperature and feedwater temperature) for a single flow baffle with gaps up to 0.020" at the periphery of the flow baffle. For a 3-2

1 l

.* )

= 1 single flow baffle with zero gap, the metal temperature cycling is less than 12 percent of the available delta T.

The zero gap condition is attained in the replacement feedwater sparger by the use of double flow baffles with leak-off holes between the baffles to prevent leakage at the periphery of the baffle. Without flow baffles, the metal temperature cycling would be about 20 percent of the available delta T at low bypass leakage flows (less than 1 gpm) and up to 50 percent of the available delta T at high bypass leakage flows (greater than 15 gpm). . Figures 4-32 and 4-33 of Reference 1 also show that with flow baffles, the metal temperature cycling is relatively insensitive to the bypass leakage flow rate in the O to 15 gpm range.

High cycle fatigue usage factors were calculated for the feedwater nozzle metal temperature cycling shown in Figure 4-33 and the feedwater time / temperature / flow map given in Table 4-29a of Reference 1. The method of calculation was in accordance with Sections 4.7.2.1 through 4.7.2.4 of Reference 1.

The high cycle fatigue evaluation was based on the random temperature cycling trace shown in Figure 4-125 of Refer-ence 1 using the Ordered Overall Range (OOR) cycle counting approach. The OOR load spectra and envelope spectrum for the high cycle fatigue evaluation are given in Figure 4-127 and Table 4-25 of Reference 1, respectively. The design fatigue curve used in the evaluation is shown in Figure 4-128 of Reference 1. This curve is the same as the design fatigue curve given Section III of the ASME Code up to 10 6 cycles but has been extended by GE out to 10 11 cycles.

3-3

The calculated high cycle fatigue usage factors are sum-marized below.

High Cycle Fatigue Usage Factor Gap at Periphery of Flow Baffle Per Year In 30 Years In 40 Years Zero 5.18E-5 0.00155 0.00207 0.02" 4.55E-3 0.137 0.182 The calculated fatigue usage factors indicate that initia-tion of new fatigue cracks due to high frequency' thermal cycling is not expected over the remaining life of the plant.

It should be noted that the above results are valid for feedwater bypass leakage flow rates up to 15 gpm compared to the maximum expected bypass leakage during worse case condi-tions of 0.3 gpm. Thus, these results indicate no new crack initiation in the feedwater nozzles due to high frequency thermal cycling even with a degradation in piston ring effectiveness by a factor of 50. Because of this, it is not necessary to demonstrate piston ring integrity with the Oyster Creek replacement feedwater sparger to assure feed-water nozzle integrity. Ehis feature of the Oyster Creek replacement feedwater sparge: was recognized by the NRC in l Section 4.3.2.4 of Reference 2 in which Oyster Creek and l Nine Mile Point Unit 1 (khich has an identical replacement feedwater sparger) are e.iempted from the periodic piston ring leak tests that are-required for other feedwater spargers (see Table 2 of Reference 2).

3-4

3.2 CRACK GROWTH CONSIDERATIONS

. Crack growth rate analyses were performed for the following conditions:

a. An initial flaw size of 0.062". As stated previously,

-this flaw is the estimated maximum defect that may have

~

been left in the feedwater nozzles during the 1977 re-

. pairs based on the liquid penetrant examination acceptance criteria that were used. The replacement

- feedwater sparger installed in 1977 is expected to eliminate any new crack initiation by the high cycle fatigue mechanism as discussed in the previous section.

b. The reference BWR duty cycle shown in Piqures 4-143,

• 4-144, and 4-145 of Reference 1. The reference BWR duty cycle consists of 130 startups/ shutdowns (Figure 4-143), 349 scrams to low pressure hot standby and return to power (Figure 4-144), and 62 scrams to high presssure hot standby and return-to full power (Figure 4-145) in 40 years of plant life. The reference BWR duty cycle includes cyclic pressure loads due to normal' reactor startup/ shutdown and postulated scrams to high

and low pressure hot standby conditions, and thermal shocks-(6 per hour) due to on/off operation of the feedwater flow during the time the reactor is held at i high and low pressure hot standby conditions.

l

c. The best' estimate crack growth curve given in l Figure 4-137 of Reference 1. As illustrated in Figure 4-137, the best estimate curve essentially

( ' envelopes the BWR data points in that 38 out of 43 BWR

! data points fall below the best estimate curve. Thus, the best estimate curve can be considered as an upper

bound curve for crack growth rates in BWR environments.

l 3-5 1

i

Analyses performed by GE and documented in Reference 1 demonstrate that calculated crack growth rates based on the reference BWR duty cycle and the best estimate crack growth curve adequately predict the actual crack growth rates in BWR feedwater nozzles (see Figure 2-1 of Reference 1) .

Based on a review of control room logs, instrument strip charts, scram logs, and plant operating procedures, an Oyster Creek plant specific duty cycle was developed (Figures 3-1 and 3-2). The Oyster Creek plant specific duty cycle differs from the reference BWR duty cycle mainly in the number of thermal cycles as discussed below.

Startup/ Shutdown Cycle The Oyster Creek Plant specific'startup/ shutdown cycle assumes a total of 9 thermal cycles during startup and shut-down, whereas the reference BWR startup/ shutdown cycle assumes a total of 12 thermal cycles during startup, and 27 thermal cycles during shutdown.

Scram Cycle The Oyster Creek plant specific scram cycle considers a single scram to high pressure hot standby and return to -

power with a total of 12 thermal cycles during the scram.

The reference BWR scram cycle considers two scrams; a scram to low pressure hot standby and return to power with a total of 72 thermal cycles and a scram to high pressure hot standby and return to power with a total of 12 thermal cycles.

In addition, the Oyster Creek plant specific scram cycle assumes a pressure spike of 1050 psig whereas the reference BWR scram cycle assumes a pressure spike to 1200 psig.

3-6 1

E~ l

~

l

.- l Thus, crack growth rate analyses based on the reference BWR duty cycle can be expected to yield conservative results for l actual Oyster Creek operation.

The stresses in the feedwater nozzle due to cyclic pressure and thermal loads were obtained from a finite element analy-sis of the specific Oyster Creek feedwater nozzle geometry.

The finite element model included the final machined con-

~

figuration of the feedwater nozzle due to removal of stain-less steel cladding and a thin layer of base metal during the 1977 repairs.

Pressure stresses were determined for a unit 1000 psi internal pressure load. Thermal stresses were determined for a unit'450*F step change in feedwater temperature (i.e.,

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 temperature of the water in the feedwater sparger thermal sleeve and.feedwater nozzle will approach the temperature of the reactor recirculatic.: water-(assumed to be 550*F at high pressure hot standby). Initiation of l feedwater flow to maintain reactor water level will result i in a rapid decrease in the feedwater temperature flowing _

through the feedwater sparger thermal sleeve (assumed to be

'100*F when feedwater heaters are out-of-service).

In the thermal stress analyses, a conservative' heat transfer i coefficientof155 Btu /hrft.Fwasusedontheportionof 2 the inside surface of the feedwater nozzle insulated by the

! ' replacement feedwater sparger thermal sleeve. This value of heat transfer coefficient is based on an assumed feedwater bypass leakage flow rate of 3 gpm compared to the maximum expected bypass leakage of 0.3 gpm during worst case condi-tions. Thus, the crack growth rate analyses are considered

, 3-7 l

l l

l

, .- . - _ . . - , _ - - - . - - . . _ . - - . - . - - r

valid even with a degradation in piston ring effectiveness by a factor of 10.

The results of crack growth rate analyses are presented in Figure 3-3 and summarized below.

Refueling Flaw Size Outage Year (inch) 7R 1977 0.062 llR 1986 0.095 12R 1988 0.107 Estimated As shown in tt.e above table, the additional growth of the assumed 0.062" initial flaw due to one additional fuel cycle from IIR to 12R is calculated to be only 0.012". The total depth of the assumed 0.062" initial flaw is calculated to be 0.107" at the. time of the next scheduled refueling outage (12R). These crack depths are considered acceptable from the standpoint of feedwater nozzle integrity.

l l

3-8 k- _ _ _ - _ _ ,, _ . , .,_ . _ . - _ , . _ - _ - - -- ,

i 700 -

600 -

w L.

485 9 CYCLES 9 CYCLES

$g

430 425 430

,f

$ 400 = '

)

WU gO 300 -

eE 200 -

100 100 --

TNE (HR) 1400 -

1200 -

1045 1045 ff w

U 1000 - 930 [ \ 930 m

a. -

sOO 8

h 600 -

e 400 -

200 -

TNE 04R)

OYSTER CREEK' PLANT SPECIFIC STARTUP/ SHUTDOWN CYCLE FIGURE 3-1 2

i i

800

=

700 D 800 "

12 CYCLES

~

E 425 425 is s ,

. g 4= -

E 8

g [ 300 -

200 -

100 -

1 100 1

1 TNE (HR) 1400 -

1200 1050 g 1000

=

A .30 ,

g s00 -

8 e 800 -

N e

400 -

200 -

TNE (HR)

OYSTER CREEK PLANT SPECIFIC SCRAM CYCLE FIGURE 3-2

3

. p_____._ _.

. _ _ . . .g

, e < --.-n-.n t -.+ --~~ . . - - -

I %g -*- - ..+.--m--,ee- w- 4.-

- + .*m.-.+ e e e e-a.,**+=s==+e...we- M aw

---,=.=e-.- .g%%-+-+-w---e- ww ._=.e.9

.eM--e-

= =miO-M W'-+

=peh9-W---I a -G Mmem - e me.W m'+4-9"*>h-m MMw

..mmipM>.e.my-eemp.--e--q.--e-- .-e-4--mme-e m.

%W '.

.,4 +w-..-

.p 4 ,eq W=-i-- W g g-e ee>W e.mpme w-i m ex.er pg _

+.--.. :_+. . - ~ , . -

--..d+,

~ ~ - , - - -

,, , #-  % % . e  %% +

-4 4-

=-P===+W4- W4* M .

• .'"'"9 - ,- f

+. ..W M iw,W4-ee>< w g.-me+m+.4 W' .

g_ _j -

O O -.-

N -

O =- .-

I to /

w ' /

$6/f- f -

r N'~DOD ~ /

k:s

-0.69s' g f

-+ +

3 , , y4- {

e -.- .-.-

& f- -

4g I

~T?]

g y - -' ' . - i .e .

. ,. . -. -.u/r!

, ,e.

~ ~$Ob - -

x 4

Z r h d"

._,4,g

>b I

s

- . Vk_. r

/6 .. //$ r

'AA o i-

~T.Cyg W r

h.--+  ! k .

h I n - Lt73 w M4iW(

• fs4E:6 i

Y .- Nh !

i -

L

q. N/_ _ * { 6 k (O N

_---- - -_......--.a M MN T. ._.

a 4 ... -

_ . ~ . . _ .__ _ _ .-. ( g g p ' m g g g , @ p -

'~ - _.- ._

. . _ .J

._+____s__

- n -, . .

1

4. REFERENCES

1. NEDE-21821-02, Boiling Water Reactor Feedwater Nozzle /

Sparger Final Report, General Electric Company (Proprietary), August 1979.

2. NUREG-0619, BWR Feedwater Nozzle and Control Rod Drive Return Line Nozzle Cracking, U.S. Nuclear Regulatory i Commission, April 1980.

i b

=

9 6

4-1

- _ _ - _ _ _ _ _ _ _ - _ _ _ _ _