ML19261A592

Fatigue Evaluation of Existing Feedwater Nozzle/Sparger Design. Concludes That Postponement of in-reactor Liquid Penetrant Exam of Nozzles Beyond Spring 1979 Refueling Outage Is Warranted

ML19261A592
Person / Time
Site:	Millstone
Issue date:	01/19/1979
From:	Edwards N, Riccardella P, Tang S NORTHEAST UTILITIES
To:
Shared Package
ML19261A588	List: ML19261A587 ML19261A592
References
86.701.0011, NEU-01-003, NEU-1-3, NUDOCS 7902050083
Download: ML19261A592 (23)

Text

'

NEU-01-003 TABLE OF CONTENTS PAGE

1.0 INTRODUCTION

1 2.0 COMPONENT DESCRIPTION 2

3.0 CRACK INITIATION ANALYSIS 4

3.1 LOADING CONDITIONS 4

3.2 ANALYTICAL PROCEDURE 4

3.3 RESULTS

.10 4.0 CRACK GRONTH ANALYSIS 12 4.1 LOADING CONDITIONS 12 4.2 ANALYTICAL PROCEDURE 14 4.3 RESULTS 15

5.0 CONCLUSION

S / RECOMMENDATIONS 20

6.0 REFERENCES

21 e

=w m

e 79020Fq)33 nutech ae

NEU-01-003

1.0 INTRODUCTION

Feedwater no :le cladding defects have been observed in the Mill-stone Unit I reactor pressure vessel during a number of reactor refueling outages.

These defects are typical of cracking which has been observed at similar locations in the large majority of operating Boiling Water Reactors (BWRs).

This problem has been the subject of an intensive program conducted by the General Electric Company (GE), the results of which are documented in Reference 1.

The program established that the nozzle cracking does not constitute a reactor safety concern as long as the crack sizes are within ASME Section X1 Code limits.

The GE program also identified several remedial measures which can be taken to miti-gate or avoid the problem.

Reference 2 presents US Nuclear Regulatory Commission (NRC) guide-lines for inspection of BWR feedwater nozzles for the interim pe-riod while remedial measures are being developed and implemented.

These guidelines call for an in-reactor, liquid penetrant examin-a ton of the no lea at the earlier of:

a)

Every other scheduled refueling outage b)

The scheduled refueling outage after 20 but prior to 40 startup/ shutdown cycles after the last liquid pene-trant examination.

These recommendations are based on early feedwater nozzle /sparger designs, which permitted significant thermal sleeve bypass leak-age, and thus were extremely prone to the nozzle cracking problem.

Reference 2 also states that, in determining the inspection fre-quency for a specific facility, the NRC staff will consider reme-dial measures a licensee has taken to mitigate the feedwater noz-le cracking problem.

During the refueling outage prior to fuel cycle 5 at Millstone Unit 1 (Fall, 1976), the feedwater spargers were removed, liquid penetrant examination was performed on the inner blend radius and bore regions of all four nozzles, and all observed cracks were re-moved by grinding.

New feedwater spargers were then installed which feature an interference fit between the thermal sleeves and no :le safe-ends to prevent thermal sleeve bypass leakage.

Addi-tiona?.y, operational measures have been taken during subsequent reactor operation (fuel cycles 5 and 6) to minimize the thermal duty imposed on the nozzle.

The nozzles were examined ultrasoni-cally during the subsequent refueling outage (Spring, 1978), and were found to have no reportable indications.

This report 3 resents an analysis of the Millstone Unit 1 Feedwater Nozzles whica specifically addresses the above remedial measures taken at the plant.

The analysis provides a basis for deferring in-reactor liquid penetrant examination of the nozzles beyond the upcoming refueling outage (Spring, 1979) as would be recommended by Reference 2 in the absence of any remedial measures.

nutech 1

NEU-01-003 2.0 COMPONENT DESCRIPTION Figure 1 presents a cross-section of the Millstone Unit 1 Feedwater Nozzle /Sparger Design, showing significant dimensions and features.

The function of the nozzle /sparger is to introduce and distribute relatively cold feedwater into the reactor pressure vessel.

The thermal sleeve portion of the sparger, which fits into the nozzle bore also serves the function of protecting the reactor vessel no :le from thermal stresses due to the temperature differential between the cold and hot water.

The sparger design illustrated in Figure 1 is installed with a nominal 0.010 in. diametral interference, and the thermal sleeve end is Inconel to minimize loss of interference due to differen-tial thermal expansion between the sleeve and safe-end.

Earlier

.sparger designs were installed with only a nominally tight fit

(.005 to.015 in, diametral clearance) and were entirely stainless steel.

This permitted significant bypass leakage of cold water, and thus did not provide satisfactory protection of the nozzle from thermal cycling.

Interference fit there.a1 sleeves of the type illustrated in Figure 1 have been shown to significantly re-duce nozzle thermal cycling, both by experiment (Reference 1) and by improved field performance as compared to earlier designs.

The installed interference fit is expected to degrade with time, how-ever, due to corrosion of the carbon steel safe-end at the seal seat.

Thus the actual nozzle /sparger configuration analyzed ir one which varies from c;. initial, non-leaking interference fit to a loose-fitting sparger later in life which permits bypass leakage.

m 4

y-M m

6 nutech 2

i i

i i

8.75" 13.0" 5.5"

~

MIN.

N.

N "R

0.22" NOM.

6.0 SS CLAD CARBON STEEL SAFE-END

- LOW ALLOY 5

STEEL NOZZLE CAR 80N STEEL g

\\

PIPING

\\

6.0" N

2.0" R

/.I 0.82"

\\

d,

[

VA 4

ASME SECTION XI

~

n x

C ' ' '

T n

ALLOWABLE BORE INCONEL END REGION FLAW OF SPARGER 10.875"5 li.44"p THERMAL SLEEVE

g,97"p 12.81"O 13.8t"0 22.06"O

/_

1 f)NLESS STEEL SPARGER 2

FIGURE 1 - EXISTING MILLSTONE UNIT 1 5

3 FEEDNATER N0ZZLE/SPARGER CONFIGURATION k

C 7

,+

c u

O 3"

NtlE-01-003 3.0 CRACK INITIATION ANALYSIS 3.1 LOADING CONDITIONS A plant specific duty map was constructed for Millstone Unit 1,

fuel cycle 5, to perform the crack initiation analysis.

This duty map was generated using feedwater temperature and flow rate data taken periodically throughout the fuel cycle (Reference 3).

The data are presented in Figure 2 as a plot of feedwater temperature vs. feedwater flow rate expressed in percent of rated feedwater flow.

Each dot in Figure 2 repre-sents a period of time at the defined plant operational condi-tion.

The duty map was ther divided into ten regions repre-senting steady state reacto. operation.

Power level, feedwater temperature and reactor temperature are defined for each region in the inset tabic in Figure 2.

The number of hours for each map region was obtained by summing the total number of hours for each data point in each region of the duty map.

Since fuel cycle 5 was from December 1, 1976 through ?! arch 10, 1978 and the crack initiation analysis is performed on a yearly basis, the hours determined for regions 1 through 10 were scaled to a period of one year, and the scaled values are also tabulated in the inset table in Figure 2.

In accordance with General Electric Generic Feedwater Duty Maps (Reference 1), additional regions were added to account for reactor startup/ shutdown transients, maneuvering and cold shutdown.

A total of twenty-six duty map regions result as shown in Table 1.

The plant specific duty map defined in Table 1, along with alternating stress and amplitude / frequency data for rapid cycling from Reference 1 were used to perform the crack initiation analysis which follows.

3.2 ANALYSIS PROCEDURE A computer code DAMSUM has been developed to analyze the high cycle fatigue usage factor for BWR feedwater nozzles.

The al-ternating stress produced by rapid cycling for every duty map region is calculated using the following equation:

EaAT

=

o alt 2(1-v) where.

AT = A x C xC (T

-Tpg) 3 4

R E

= Youngs Modulus a

= Coefficient of Thermal Expansion Poisson's Ratio V

=

nutech 4

NEll 0 0 3

~

/"W 358 sr.

., g. -

,...v g

... ~

. g.

.p 3M

5....

n

{

. $r 10

4. g.

u o.,

VN w

5 1 Flow g 254 T

T

'[

Map Rate Fcedwater Reactor W

llrs/Yr Region t

F F

m g

1 100 363 546 5883 2W H

2 84

.552 544 1243 m

3 56 323 540 149 W 2M -."

g 4

37 293 538 50 3

5 33 252 537 35 o

d 6

23 106 536 22 o

g 7

10 106 536 40 8

10 160 536 66 154 'y. * '

9 20 50 536 119

• f' 10 90 280 544 473 6

IN

.. g 4..

t.

. :3...

J 58 1

11 21 31 48 51 Il 78 M

H IN FEECWATER FLON RATE (% OF RATED)

FIGURE 2 - MILLSTONE UNIT 1, FUEL CYCLE 5 FEEDh'ATER FLOW VS. TEMPERATURE DATA nutech 5

NIiU-01-003 TABLI! 1 MII,I.STONI! Fl!!!DWATIIR NO2ZLII DUTY MAP FOR CRACl; TNITIATION ANAI,YSIS FW FLOW T

REGION RATli (%)

IlPS/YR Rl! ACTOR pg 1

100.00 5,888.00 546.00 363.00 2

84.00 1,248.00 544.00 352.00 3

56.00 149.00 540.00 323.00 4

37.00 50.00 538.00 293.00 5

33.00 35.00 537.00 252.00 6

28.00 22.00 536.00 106.00 7

10.00 40.00 536.00 106.00 8

10.00 66.00 536.00 160.00 9

20.00 119.00 536.00 50.00 10 90.00 473.00 544.00 280.00 11 1.00 24.00 480.00 80.00 12 1.00 24.00 430.00 80.00 13 1.00 24.00 380.00 80.00 14 1.00 24.00 330.00 80.00 15 1.00 24.00 280.00 80.00

~

16 1.00 24.00 230.00 80.00 17 1.00 24.00 180.00 80.00 18 1.00 24.00 130.00 80.00 19 0.00 43.00 340.00 300.00 20 1.00

.40 360.00 350.00 21 2.00 1.10 350.00 190.00 22 2.00

.80 340.00 125.00 23 2.00

.40 330.00 70.00 24 2.00 1.10 400.00 190.00 25 3.00

.20 340.00 200.00 26 0.00 431.00 70.00 70.00 e

nutech 6

NEU-01-003 Reactor Temperature T

=

g Feedwater Temperature T

=

pg C3, C4 and A are empirical coef ficients developed from test data reported 11. Reference 1, and are presented in Figure 3 and Table 2.

Since C3 depends on the thermal sleeve bypass Icakage rate through the interference fit seal, an analysis was per-formed to determine the Icakage rate (in gpm) as a function of assumed gap size at the thermal sleeve seal.

The resulting leakage rate versus gap size is also shown in Figure 3.

This analysis was perforraed for a sparger pressure drop of 10 psi, which corresponds to 100% feedwater ficw.

The leakage rate for a given gap size is assumed to vary linearly with feedwater flow rate.

TABLE 2 AMPLITUDE / FREQUENCY SPECTRUM FOR RAPID CYCLING Index Amplitude Frequency I

A Cycles /Ilr 1

1.00 15 2

0.98 15 3

0.95 15 4

0.93 30 5

0.84 75 6

0.75 120 7

0.65 150 8

0.55 180 9

0.45 450 10 0.35 1200 11 0.20 7500 A specially developed fatigue curve for sensitized stainless steel in a BWR primary coolant environment is incorporated into the DAMSUM code, and is illustrated in Figure 4.

The nozzle fatigue usage factor is determined from this fatigue curve us-ing a multiple damage summation, as follows:

I J

K U=

E (E E

U.p)1 3

i=1 j=1 k=1 th jk Usage {actorduetoj amplitude and frequency where:

U

=

for kt flow, temperature and time.

()1 =

Usage factor for the i year.

nutech 7

NEU-01-003 C3 C4 a

n

.30 2.0

.20

.10 G

1.0

=

0 50 100 0

.5 1.0 1.5 FW FLOW RATE (% OF RATED)

BYPASS LEAKAGE (GPM) h 30 BYPASS LEAKAGE h

)

s m.m_

g 2.0

-GAP SIZE s

m

{

1,g SPARGER AP = 10 PSI 0

0

.961

.002

.003

.004 GAP SIZE (INCHES)

FIGURE 3 - N0ZZLE THERMAL CYCLING FACTORS C AND C 3

4 AND BYPASS LEAKAGE RATE USED IN CRACK INITIATION ANALYSIS nutech a

S O

i4}4

' 4 ', p;I 4'!44'+.

p'

>.y ; '2 L':.s u u

.t..

3, = t 5.. '.

., '3...., } p!

4 t H.+f. f *4N9d 4

i l

.. 4-p M

-.j a t i,

+,.

7

+

+. -..

1 a.7

-7 7

_._..u

.gj vLyr. clu g... -

.a+

.,4,y.

+ ;

n 4

i g

1 w

1

.e 3

m 7

_. +

ritj + { { t-tjt' t -f f-t{2 j ljit f

t f

f

[

• t.,t - +

t i

• H

• t'

.F

• t t. i

+:tt

+

+

t/)

l't I

2 l

t c

s

+ t t

+'

y O

,a o

n o.,4 q

.4, l'

+

+

't tyy{ tIQ tiYf}; I } {- I-y ^ tt{} 4[17 Z .,7 3 4 j 4-4 + g% H t ,g -f-t* t t p c cc s y crn Z ~+ i T-t-ttti [*I + T t i_ +:tr _ti,-. ,i i t co +

ttt O

_y W t/) 1 s e 4 itii 4 % a -{- [i1. l iff' 't!' g Uj f i t. j t 1:tt .i t ' I t i '+ t-* o o,-a o 6 e 1 + t.s o gg +. -+ J -t

j.

o O Ov g +, 7 4 + s e,o e p .I l i f 1- - I- !{[4 !114 ,Ii! I t 1-Qgmg - y!!! ! ! l-j l { ll [r ! ! '

l 4

r 1 1 1: g i ,,'i + t gg a m m ter+ 4 o O O M WH o .' p1 ,D.C.b . W c1._.. A_. - m A-O . g F t 1- , H

I,i j

r i-t, tjf h j g <? T _4 a ,.a g). ;t f. {t 'h { lf1 i ffe1 t g t'4t I h. t i t 'f +1 1 j

d. t'it

+ T t- + + + .4 e> tl ti l.) ,c .t I., t +- U.I tj i-t t' y 1 -t m , t, t t-t m m di +tn cu = _, O T- ...-o . i V t . Q: / H O i. O ,ir . C .,71 i i '-~' = .f t 't' 8 i t s, 3 - z. s ,, o.,,rt, ,,t g m a.

i. n t.

t, i t i-n, ,1,,, NA- - y .,: t. 7,, r, i, j l

e.. t, t

1

1. 1,.t

.m.. -4. 4 i i i e-i 1 _r. _m o ce / y 4 o. F

1 -

t:- / y.

Iri f!-

i

t.j !t,i.j ;.

lt

i. ' r l j j

,t-p-l t. t i + i r r + + J .1 - / i t., t j t, l t

! ! i.

+ T. t.t '. t, t, t t tyi ,1Jr t-j

t..

+ = + m i .t / f 4.. / I a 1 f i t t t-tg p

4 m

/. o

. r

l to 1.

1.,1,,1,r i

l t t-1 i i r i

t. t.,.-

a n., .w 1 1 ..t 4 i ,,1, = y 7/ u-t3 J.1 m i /... r F u // iB 5m t-r ittt t4 14 itit !;ir = z q to itit hii t e/ i' a ;: t t1, .t w w 1 n ..,s,,s ,/ w n <o a .u! ce r .. 4......ut, .a. = g 4 e m c O M M ,~ ISd 'SS381S ON11VNH3f1V i i i i i

NEU-01-003 The DAMSUM program determines the mean gap size for each year of reactor operation, based on initially installed interference fit, assumed yearly corrosion rate, and expected loss of fit due to differential thermal expansion between the thermal sleeve and the nozzle safe-end, all of which are program input parameters. The program then determines leakage rate, C3 and C4 for each map region using the appropriate feedwater flow rate for the region (Table 1). Fatigue usage factors are then calculated for each map region using the amplitude / frequency spectrum (Table 2) and the sensitized stainless steel fatigue curve (Figure 4). The usage factors for all map regions are then summed to obtain total high cycle fatigue usage factor for the year. This pro-cess is repeated for each year of reactor operation, updating the mean gap size due to expected corrosion during the year. The program output consists of a table of interference fit, leakage rate and total cumulative usage factor versus years of reactor operation. 3.3 RESULTS Table 3 presents the DAMSUM program input and output for the Millstone Unit 1 plant specific crack initiation analysis. As noted above the material was assumed to be sensitized stainless steel (SSS). An initial radial interference fit of.009 inches was determined from measurements taken during the sparger in-sta11ation during the refueling outage preceding fuel cycle 5 (References 5 and 6). It was also determined through clastic-plastic analysis that the seal can sustain this level of inter-ference fit without yielding and attendant loss of fit. General Electric design basis corrosion rates of 0.000675 inches / year and.000397 inches / year were assumed for the carbon steel safe-end in the non-leaking (creviced) and leaking (uncreviced) con-ditions, respectively. Based on thermal calculations, it was determined that the maximum differential thermal expansion ex-pected under the most severe operating transient is 0. 0 0 5 inches, and this value was conservatively assumed as the " fit to cause Icakage" in the DAMSUM analysis. The output portion of Table 3 lists mean interference fit, leak-age rate, and cumulative high cycle fatigue usage factor for 11 years of Millstone Unit 1 operation, starting at the refueling outage prior to fuel cycle 5. The interference fit progres-sively decreases, due to corrsion, but does not degrade below the.005 inches required to cause leakage until year 7. Con-sequently, fatigue usage factor accumulates at a constant rate of 0.037 per year up to this point. The rate of accumulation of fatigue usage begins to accelerate in year 7 due to the ini-tiation of thermal sleeve bypass leakage, and surpasses the ASME Code allowabic usage factor of 1.0 somewhere between years 8 and 9, at which point reinitiation of the high cycle fatigue cracking observed in prior refueling outages may be expected. riutech 10

NEU-01-003 TABLE 3 RESULTS OF MILLSTONE UNIT 1 CRACK INITIATION ANALYSIS INPUT SSS Material = Initial Interference 0.009 in. = Corrosion Rate: 0.000675 in/yr Before Leakage = After Leakage 0.000397 in/yr = Fit to Cause Leakage 0.005 in. = OUTPUT Interference Leakage Cumulative Fatigue Year Fit (Inches) (GPM) Usage Factor 1 .0087 0.000 .037 2 .0080 0.000 .074 3 .0073 0.000 .111 4 .0066 0.000 .149 5 .0060 0.000 .186 6 .0053 0.000 .223 7 .0048 .110 .293 8 .0044 .477 .657 9 .0040 .972 2.459 10 .0036 1.500 8.466 11 .0032 1.500 14.473 It should be noted that the predominant difference between the present analytical prediction and past performance of the Mill-stone Unit 1 feedwater nozzles is the presence of a substantial interference fit, and therefore, the absence of thermal sleeve bypass leakage for several years. Prior feedwater spargers in-stalled in the Millstone reactor did not have this feature, and thus led to crack initiation in one to two years, as would be predicted by the analysis. The DAMSUM analysis presented here-in is thus consistent with field experience at Millstone to date. nutech 11

NEU-01-003 4.0 CRACK GROWTH ANALYSIS 4.1 LOADING CONDITIONS Plant specific operational transient curves were obtained for Millstone Unit 1, fuel cycles 5 and 6 in order to conduct the crack growth analysis (Reference 7). These curves show, in detail, the reactor tenperature, pressure and feedwater tem-perature versus time for each reactor shutdown and subsequent return to power. They also provide a conservative estimate of the number of on/off feedwater flow cycles (jogs) which oc-curred during each transient. The following data reduction scheme was used to transform the data from Reference 7 into a suitable format for crack growth analysis-1. The overall number and magnitude of pressure cycles and the corresponding temperature fluctuations are determined for each transient. 2. The number of feedwater jogs are then determined, and the 'emperature fluctuation (AT) associated with each jog. 3. The feedwater jogs are grouped in accordance with the re-actor pressure at the time Cae jog occured, and the maxi-mum AT of all the jogs in r group is used. 4. If a group of feedwater jogs occurs during a pressure ramp, the average pressure for the ramp is assigned to that group of jogs. Figure 5 shows a comparison of a General Electric generic feed-water nozzle thermal duty transient (scram to lower pressure hot standby and return to power) with a typical Millstone plant specific transient (fuel cycle 5, ramp #15). Table 4 presents a comparison of the number of cycles involved for each transient and the magnitude of the associated temperature and pressure fluctuations, derived in accordance with the data reduction scheme described above. This comparison shows that the two transients are in reasonable agreement, with the Millstone plant specific cycle involving fewer total cycles, but slightly larger temperature fluctuations for each cycle. Nozzle bore region pressure and thermal stress profiles for t:te crack growth analysis were obtained from Reference 1. The ~ Reference 1 pressure stress profile is for 1000 psia and was scaled up or down depending on the actual pressure fluctuations for each operational transient. Two Reference 1 thermal stress profiles for a clad nozzle were selected, one for a leaking sparger case, in which a heat transfer coefficient of 2000 nutech 12

N1:U 0 0 3 FULL POWER FULL POWER OPERATION OPERATION w z EH ON/OFF FW FLOW @C 500 cycles PER hour 430 425 CYCLING AT SIX oL 425 h y 4H Eb 300 2e 3H $ 2H r 9N I 10 0 3o 1H i uz IO hr. j TIME (hr.) a) GE GENERIC TRANSIENT (SCRAM TO LOW PRESSURE HOT STANDBY AND RETURN TO FULL POWER - REF.1) EE Ill ~ @E R L 508 i g i g g 4H / o / e $ 388 ,/ 3 ~ 9 y$ 208 l g lH O I 24 48 72 II TIME (hr.) b) TYPICAL MILLSTONE, UNIT I PLANT SPEClFIC TRANSIENT (FUEL CYCLE 5, RAMP NUMBER 15 - REF. 7) FIGURE 5 - COMPARISON OF GENERAL ELECTRIC GENERIC AND MILLSTONE PLANT SPECIFfC FEEDNATER NOZZLE TIIERMAL DUTY TRANSIENTS 13 nutech

NEU-01-003 2 BTU /hr ft F was assumed, and one for a non-leaking sparge; case, in which a heat transfer coef ficient of 100 BTU /hr f t'

• F was assumed.

In both cases, the Reference 1 stress profile is for a step change temperature transient with AT = 450 F, and was scaled to the actual magnitude of the temperature fluctua-tions for each operational transient. TABLE 4 TilERMAL DUTY TRANSIENT COMPARISON No. of Pressure Cycles (psig) Temperature General Electric 1 1200-0-1200 AT = 325 F Generic Transient 60 120 AT = 200 F (Sc ram to Low Pressure llo t Standby - Ref. 1) 11 930 AT = 330 F Millstone Plant 1 1010-10-1010 AT = 300 F Specific Transient 4 510 AT = 220 F (I ucl Cycle 5 Ramp #15 - Ref. 7) 5 10 AT = 350 F 6 225 AT = 320 F 3 720 AT = 250 F 4.2 ANALYTICAL PROCEDURE Calculation of fatigue crack growth was performed using the NUTECH computer code NUTCRAK. Complete documentation and description of this program is given in Reference 8. A brief explanation of the capabilities of the program is provided below. Three different computations related to crack growth are performed by NUTCRAK. They are; curve fitting of any arbi-trary defined stress profile, calculation of stress intensity factor, and determination of crack growth through integration of the appropriate material fatigue crack growth law. For any stress distribution across a component wall thickness, a least square polynominal curve fitting routine is employed to obtain a third order equation of the form: o=A0+A1X+A2X +AX (1) 3 Using the principle of supe

• position and the above stress equa-tion, stress intensity factors ict the specific crack medel are generated for the constant, lineat, quadratic and cubic stress terms.

The resulting stress intensity factor is obtained by the following equation: 2 3 2a a 4a 0 1 1 2+ IA F + n A F) (2) ~ K = /ia(A + A F 2 3 3 4 nutech 14

NEU-01-003 where: K = Stress Intensity Factor a = Crack Length M gnification factors for specific crack F F,F I 3 4 geometry being analyzed. Fatigue crack propagation is then evaluated by numerical inte-gation of the fatigue crack growth law: hk = C (Ke f f) (3) where: Keff = K (1-R)m max R=Kmax/Kmin C,m, n = Material Constants The procedure described abave is identical to that used in Reference 1 to perform a generic crack growth analysis of BWR feedwater nozzles. The results of this generic analysis are summarized and compared to maximum crack depths observed in a large number of BWRs in Figure 6. This figure presents two curves, one for the nozzle blend radius region, which agrees with the cracking observed in the large majority of reactors, and a second, more rapid crack growth curve for the nozzle bore region, which bounds the crack depths observed in the re-latively few BWRs in which bore region cracking occured. This correlation serves as a qualification of the analytical proce-dure used in this report. It is noteworthy, however, that even though the Millstone data point in Figure 6 falls right on the blend radius curve, and bore region cracking has not been ob-served at Millstone, the more conservative bore region analysis was used in this report. 4.3 RESULTS The resulting Millstone crack growth estimates for fuel cycles 5 and 6 are presented in Figures 7 and 8. Figure 7 presents the calculated growth of a 0.25 inch initial crack under the assump-tion that the sparger was leaking significantly right from the start of fuel cycle 5. Crack growth curves are presented for both the General Electric generic operational transients (from Reference 1) and for the Millstone plant specific operational transients described above. Note that the Millstone curve essentially parallels the generic curve, only at a lower pre-dicted rate of growth. This difference, particularly in fuel cycle 6 is indicative of the operational measures taken at the plant to reduce the thermal duty imposed on the nozzle. Figure 8 presents similar crack growth curves, but for the expected case of a non-leaking sparger, since the analysis in Section 3 of this report indicates that no sparger bypass nutech 15

NEU-01-003 OBSERVED CR ACK DEPTHS: 6 BORE REGION O BLEND RADIUS h PLANTS WITH PARTIAL FIXES 1.5 - (WE LDE D SPARGE R, INTE R-O FERENCE FIT SPARGERS. OR NINE MILE HIGH FE RRITE CLADDING) POINT BEST ESTlf IATE 1.4 - ANALYTIC AL CURVE (BORE RE ilON) 1.3 1.2 1.1 OVERSE AS BWR j 1.0 = f 0.9 O 2y 0.8 PILGRIM c: b .J BEST ESTIMATE HUMBOLDT BAYo 0.7 AN ALYTICAL CURVE g F (DLEND RADIUS) 0.6 OYSTER CREEK O MILLSTONE 0.5 MONTICE LLO O O DRESDEN 2 IT ES.2 O QUAD CITIES-1 0.4 PE ACH BOTTOM 2 O o OORESOEN 3 O VERMONT 0.3 YANKE E 0.2 COOPER h 0.1 BWRS R810WNS FERRY 1 OVE RSE AS PE ACH h OVERSEAS BOTTOM 3 BWRS ""o*tS "aa72 L io i i , ocfc"' i i i 2 0 y y y yy y 0 2'O 40 60 80 100 120 NUMBER OF STARTUP/SHUTOOWN CYCLES FIGURE 6 - GENERIC FEEDWATER N0ZZLE CRACK GROWTil ANALYSIS CORRELATION WITil FIELD DATA (REF. 1) nutech 1,

NEU-01-003 l.4 1.2 1.0 2 GENERIC hg .O ASME SECTION II d ALLOWABLE CRACK DEPTH O .6 MILLSTONE PLANT '4 / SPECIFIC /'& s END OF FUEL CYCLE 5 y ART OF FUEL CYCLE 6 (THRU DEC.,1978) 0 ~ 0 10 20 30 40 50 60 NUMBER OF OPERATIONAL TRANSIENTS FIGURE 7 - MILLSTONE FEEDNATER N0ZZLE CRACK GROWTH ANALYSIS (LEAKING SPARGER CASE, U = 2000 BTU /IIR-FT2 op) ~ nutech 17

NEll 0 0 3 .8 ASME SECTION XI ALLOWABLE CRACK DEPTH ,7 .6 / GENERIC ~ ~~' / I .5 --- HILLSTONE PLANT g SPECIFIC h .4 l l o x .3 2 r l END OF FUEL CYCLE 5 ,j /-START OF FUEL FUEL CYCLE 6 (THRU DEC.,1978) / CYCLE 5 0! 0 10 20 30 40 50 60 NUMBER OF OPERATIONAL TRANSIENTS FIGURE 8-MILLSTONEFEEDWATERN0ZZLECRACKGRONTyANALYSIS (NON-LEAKING SPARGER CASE, U = 100 BTU /IIR-FT F) nutech 13

NEU-01-003 leakage would be expected following the installation of the existing spargers in the Fall of 1976. The predicted crack growth of a 0.25 inch initial crack under this more realistic assumption is much slower than in Figure 7. In fact, a rela-tively slow rate of crack growth is predicted in Figure 8, even if a 0.5 inch deep initial crack is assumed, which is equal to the deepest crack ever found in the Millstone feed-water nozzles. The ASME Section XI allowable crack depth for a bore region flaw is indicated at the innermost clad section of the nozzle in Figure 1. This allowable base metal flaw depth is conservative-ly taken to be 10 percent of the minimum metal path to cause a through-wall flaw at the flaw location. The nominal clad depth is then added to yield a total allowable crack depth of 0.82 inch. (The technical basis for this method of es'.ablishing Section XI allowable crack depth for BWR feedwater nozzles, as well as a quantitative assessment of the safety margins which it provides can be found in Reference 9.) The 0.82 inch allowable crack depth is indicated by a horizontal dashed line in Figures 7 and 8. It is apparent from these figures that even a very conservative extrapolation of the nozzle crack growth curves for the balance of fuel cycle 6 and 7 does not lead to a concern of exceeding the Section XI allow-able crack by the end of fuel cycle 7. w w nutech 19

NEU-01-003

5.0 CONCLUSION

S / RECOMMENDATIONS Based on the analysis results presented in this report, the following conclusions can be drawn regarding the expected performance of the existing feedwater nozzle /sparger design at Millstone, Unit 1: 1. Loss of interference fit and attendant thermal sleeve bypass leakage is not expected for approximately 6 years following the sparger installation in the Fall of 1976. 2. In view of 1 above, and considering Millstone plant specific operational nozzle thermal duty, reinitiation of the blend radius cracking observed in prior nozzle examinations is not expected for approximately 8 years following the sparger instal-lation. 3. Even conservatively assuming that a crack was present at the time of sparger installation, and that bypass leakage initiated immediately, crack growth beyond the ASME Section XI allowable crack depth is not expected loring the current or the next re-actor fuel cycle (fuel cycles 6 and 7). On the basis of these conclusions, it appears that a postponement of in-reactor liquid penetrant examination of the nozzles, beyond the upcoming Spring, 1979 refueling outage, is well warranted. 4e nutech 20

NEU-01-003

6.0 REFERENCES

1. H.T. Watanabe, et.al., Boiling Water Reactor Feedwater Nozzle / Sparger Final Report, NEDE-21821, General Electric Co., March, 1978. NUREG-0312, Interim Technical Report on BWR Feedwater and Control Rod Drive Return Line Nozzle Cracking, U.S. Nuclear Regulatory Commission, July, 1977. 3. Letter, W. F. Lenz to P. C. Riccardella, Millstone Unit No. 1 Feedwater Temperature Data Corrected at Low Feedwater Flog and at the End of Cycle 5, dated October 30, 1978. 4. D. A. Hale and D. H. Inhoff, Reactor Prinary Coolant System _ Pipe _ Rupture Study Progress Report No. 33, GEAP-10f07-33, General Electric Co., October, 1975. 5. Special Report Addendum No. 3, Feedwater Nozzle Cladding Defects, Millstone Nuclear Power Station Unit 1, November, 1976. 6. General Electric Co., Thermal Sleeve Machining Measurements, Special Process Control Sheet No. RPVR 13.0-2, October, 1976. 7. Letter W. F. Lenz to P. C. Riccardella, Millstone Unit No. 1 Shutdown /Startup Nozzle Temperature and Reactor Pressure Curves, dated November 6, 1978. 8. P. C. Riccardella and S. S. Tang, NUTCRAK User's Manual, NUTECH, 08.039.0005. May, 1978. 9. P. C. Riccardella and W. E. Cooper, Safety Evaluation of Reactor Vessel Nozzle Cracks, ASME paper 78-PVP-90, presented ~ at ASME/CSME Pressure Vessels and Piping Conference, Montreal, Canada, June, 1978. 36 1 nutech

ATTACHMENT 3 Indications Zone 1 and Zone 2 Zone 3 (Blend Radius) (Bore) (Cylindrical Section) Recordable Indications which exceed 40% Indications that exceed 25% Level FSH at primary refe ;nce FSH at the primary reference level. level. Repor table Indications that are recordable, travel in time position, Level and can be identified as the same indication when the nozzle is scanned in two directions (clockwise and counter-clockwise) unless acceptably dispositioned by a Level III UT inspector.}}