ML20093E334

Feedwater Nozzle Fracture Mechanics Analysis

ML20093E334
Person / Time
Site:	Brunswick
Issue date:	05/31/1991
From:	Stevens G GENERAL ELECTRIC CO.
To:
Shared Package
ML20093E328	List: BSEP-95-0459, Advises of Util Revised Plans for Insp of Feedwater Nozzle Blend Radii.Rev 1 to NEDC-30633, Brunswick Steam Electric Plant Unit 2 Feedwater Nozzle Fracture Mechanics Analysis, Encl ML20093E334
References
NEDC-30633, NEDC-30633-R01, NEDC-30633-R1, NUDOCS 9510160175
Download: ML20093E334 (42)

Text

-- m --w g .

e GENuclearEnergy t

IIS ConnerAmve t l- SanJose. CA 95125

)i

'o - NEDC-30633 Revision 1 DEF B11-00491 f

Class III May 1991 GE NUCLEAR ENERGY l

BRUNSWICK STEAM ELECTRIC PLANT, UNIT 2 FEEDWATER N0ZZLE FRACTURE MECHANICS ANALYSIS G. L. Stevens i

i e

Approved:

S. Ranganath, Manager Materials Monitoring and Structural Analysis Services , ,

9510160175 951009 '

PDR ADOCK 05000324 '

O PDR

. ~ . - . .- ...

t k NEDC-30633 REY. 1 1

CLASS III l IMPORTANT NOTICE REGARDING CONTENTS OF THIS REPORT PLEASE READ CAREFULLY i

6.

{' This report was prepared by the General Electric Company (GE) solely for

,f the use of Carolina Power and Light Company. The information contained in this report is believed by GE to be an accurate and true representation of the facts known, obtained or provided to GE at the time this report was prepared.

The only undertakings of the General Electric Company respecting informa-tion in this document are contained in Carolina Power and Light Company Work Authorization No. ZS70020028 and nothing contained in this document shall be 2

construed as changing said contract. The use of this information except as defined by said contract, or for any purpose other than that for which it is a

intended, is not authorized; and with respect to any such unauthe:7. zed use, neither GE nor any of the contributors to this document makes any representa-tion or warranty (express or implied) as to the completeness, accuracy or use-fulness of the information contained in this document or that such use of such information may not infringe privately owned rights; nor do they assume any responsibility for liability or damage of any kind which may result from such use of such irformation.

j j

3 I

o 4

b

=

e

NEDC-30633 REY. 1 CLASS III CONTENTS I.} ABSTRACT 1 y

4 h

• J

1. INTRODUCTION 1.1 Background 1-i 12 Objective 1-l' 1.3 Technical Approach 1-2 l 1.3.1 Feedwater Flow Cycling 1-2 i i

1.3.2 Thermal Cycling 1-2 l 1.3.3 Nozzle Configuration 1-2 '

1-3 l 2 l

SUMMARY

AND CONCLUSIONS

< 2-1 '

3. THERMAL CYCLE DEFINITION 3-1 1

4. THERMAL ANALYSIS

~

4-1

5. THERMAL AND PRESSURE STRESSES 5-1

, 6. CRACK GROWIH ANALYSIS 6.1 Stress Intensity Factor Calculations 6-1 6.2 Crack Growth Data 6-1 6.3 Crack Growth Evaluation 6-5 6-5

7. RESULTS AND CONCLUSIONS 7-1

8. REFERENCES 8-1 APPENDIX - THERMAL BOUNDARY CONDITIONS A-1 4

e i

w

.. y

~

I 4

tti a

b s

NEDC-30633 REY. 1

CLASS III TABLES Table Title Page 5-1 Surface Stresses to choose Maximum combined Stresses 5-2

. 5-2 Limiting Stress Profile (Cross Section 3-3) 5-7

, ILLUSTRATIONS 1

q Figure Title Page 3-1 Temperature cycling for 5/3/88 Shutdown-Startup 3-4 4

j 3-2 Temperature Cycling for 5/7/88 Forced Shutdown 3-5 3-3 Temperature Cycling for 11/16/88 Scram 3-6 3-4 Temperature Cycling for 6/17/89 Scram 3-7 4-1 Nozzle Configuration 4-2 j 4-2 Thermal Boundary Conditions 4-3 A

5-1 Feedwater Nozzle-Brunswick 2 Location of Maximum Surface 1

Stresses (Steady State) 5-4 i 5-2 Feedwater Nozzle-Brunswick 2 Location of Maximum Surface j Stresses (Transient t = 3 Minutes) 5-5 4

6-1 Boundary Integral Equation / Influence Function Magnification

Factors for BWR Feedwater Nozzle 6-2 4

6-2 Stress Intensity Factor versus Crack Depth (Thermal Stresses, 3 minutes) 6-3 l 6-3 Stress Intensity Factor versus Crack Depth (Pressure

, Stresses) 6-4 6-4 Reference Fa'tigue Crack Growth Curves for Carbon and Low Alloy Ferritic Steels 6-6 l

7-1 Crack Depth versus Number of Years 7-2

.s 3' s .

• iv

W

, l$[ NEDC-30633 REY. 1 CIASS III ~

4

, ABSTRACT a

, The current revision of this report is based on actual feedwater cycling 4

.g data collected for Brunswick 2 since the original (July 1984) revision of this T

report. In addition, the triple thermal sleeve design previously analyzed has i not been implemented, and there are currently no plans to do such. Therefore,

! aj. the ring-interference fit, single thermal sleeve design (original design) was l evaluated in this report.

1 This report provides a plant-specific fracture mechanics assessment of j the Brunswick 2 feedwater nozzles to show interim compliance with NUREG-0619 and NRC Generic Letter 81-11, dated February 20, 1981. The evaluation was based upon (1) the plant operating history supplied by Carolina Power and 4

Light Company (CP&L), (2) low feedwater flow characteristics determined from actual plant feedwater cycling measurements, and (3) Moss Landing test data.

The evaluation considered an initial crack depth of 0.25 inch as specified in NUREG-0619. The results show that stress cycling from actual temperature and a

flow profiles results in the growth of an initial 0.25-inch crack to greater than 1 inch during the remaining life of the plant. Using the 1989 ASME Section XI fatigue crack growth curves, the analysis shows that the postulated 0.25-in. crack becomes 1 inch deep 32.3 years after the startup date.

2 l Recommendations are made for updating the plant operating history, as well as monitoring leakage flow, and using using those results to re-evaluate the crack growth analysis contained herein.

i i

I

= .

Y 4

v e

.. -- - - - - - - - - - - - ^

CIASS III -

1. INTRODUCTION This report provides a plant-specific feedwater nozzle fracture mechanics e, assessment based on the Brunswick Steam Electric Plant, Unit 2 (hereafter called Brunswick 2) plant operating history and actual feedwater cycling data.

', This is in response to the Nuclear Regulatory Commission (NRC) requirements g regarding feedwater nossle crack growth. These requirements are contained in the NRC Generic Letter 81-11, which states that a fracture mechanics evaluation must predict an end-of-design-life crack size of 1 inch or less.

1.1 BACKGROUND

The General Electric Company (CE) feedwater nozzle final report (Refer-ence 1) recommended design and operational changes to minimize both the proba-bility of crack initiation and rate of crack growth in feedwater nozzles. The low flow feedwater controller discussed in Reference 1 would not significantly reduce the probability of crack initiation, but would reduce crack growth.

The NRC (NUREG-0619) accepted the GE recommendation (Reference 1) and required that operating reactors install a low flow feedwater controller with the char-acteristics described in Reference 1 and reroute the Reactor Water Cleanup System (RWCS) flow to all of the feedwater lines. The low flow controller required above must meet strict requirements specified in Subsection 3.4.4.3 of Reference 1. The NRC later clarified its position in Generic Letter 81-11, stating that plant-specific analyses may be performed to justify not imple-

menting such modifications.

?

With respect to low flow feedwater controller installation assessment, feedwater nozzle crack growth rate analysis is required for Brunswick 2.

l 1.2 OBJECTIVE Because of the absence of Brunswick 2 low flow feodwater controller data, temperature measurement hardware was installed subsequent to the July 1984 5

revision,of this report. The data collected by that hardware for the last f

1-1 a

i

NEDC-30633 REV. 1

p. CLASS III y -

fuel cycle was utilized as a basis for defining the actual Brunswick 2 feed-water cycling characteristico used in the current analysis. The data col-1ected was considered typical for the entire design life of the reactor.

I 1

({,

i This report provides a plant-specific fracture mechanica assessment of l {, the Brunswick 2 feedwater nossles to show compliance with the requiremente of j f NUREG-0619 as amended by Generic Letter 81-11, dated February 20, 1981. The 97 purpose of this analysis is to determine whether stress cycling from actual i controller temperature and flow fluctuations will result in a final crack I depth of 1 inch or less during a 40 year plant design life. The evaluation considers an initial crack depth of'O.25 inch, as specified in NUREG-0619.

1.3 TECHNICAL APPROACH This analysis evaluates the growth of a 0.25-inch crack over a projected 40 year plant design life.

1.3.1 Feedwater Flow Cycling The feedwater flow cycling was determined from actual feedwater tempera-ture and flow data obtained from CP&L (Reference 2).

1.3.2 Thermal Cycling i

The thermal cycling of the fluid at the feedwater nozzle was determined from the actual feedwater data. The number of startup/ shutdown and scram 5

events for Brunswick 2 was linearly projected based on actual plant operating history during the first fourteen years (1975-1988) of plant operation

(References 2, 3 and 4).

l J

The thermal boundary conditions used in this analysis differed from the i

i Reference 1 document in that thermal sleeve annulus temperatures and annulus heat transfer coefficients, derived from Moss Landing test data (Reference 9),

i were used to calculate tRe-Thermal stresses in the feedwater nozzle.

1-2 6

NEDC-30633 REV. 1

f. '

CLASS III N:? ,

To evaluate the crack growth, thermal and pressure stress analyses were conduct.ed using the finite element computer code ANSYS (Reference 6). The

{,7 locations of the peak thermal stress, peak pressure stress and peak combined

'D thermal and pressure stress were determined and the crack growth was calcu-

.b

• isted using a crack growth computer code. The crack growth relationship used 1 e repreaants the 1989 ASME Section XI Code Curves. The best-fit correlation to l< E actual PWR and BWR data used in the original revision of this report was not a[

f used, since the R-ratios (K ,1,/K ,,,) were typically high for the actual cycling data; that correlation is not valid for high R-ration.

1 i 1.3.3 Nozzle Configuration The initial revision of this report analyzed the replacement triple thermal sleeve sparger and machined feedwater nozzle bore. That replacement has not been implemented as of the current date, and there are currently no plans to do such. As a result, this report evaluates the original ring-interference fit, single thermal sleeve sparger design. The nozzle bore has not been machined and, as a result, was analyzed in its original clad

, configuration.

k 4

4 .

m 1-3 e

NEDC-30633 REY. 1 CLASS III

~

p

2.

SUMMARY

AND CONCLUSIONS K

-g Application of the ASME Code,Section XI crack growth rate relationship resulted in crack growth greater than the acceptance criterion of 1 inen for a 40 yr plant life. The analysis reaulted in a 1-inch crack depth af ter 32.3 6 years of operation.

l This analysis is based on actual low flow feedwater controller character-istics obtained since the original revision of this report. Although this analysis did not show compliance with NUREG-0619, it should be noted that there are approximately 16 years of plant operation before the crack would reach the 1-inch depth.

The plant operating history is based on the initial 14 years of plant

' operation extrapolated to 40 years. Because of " learning curve" effects which are typically experienced by operating reactors during their initial years of operation, the extrapolation is most likely conservative. In addition, a l potential exists for leakage past the thermal sleeve seal as a result of possible degradation of the seal from corrosion or some other form of relaxation. Therefore, conservative heat transfer rates were used in the thermal stress analysis to accommodate the potential for such leakage.

With regards to RWCU reroute, a plant-specific analysis was performed for Brunswick 2 which demonstrates that RWCU reroute leads to only a small improve-ment on thermal cycling and fatigue usage of the feedwater nozzle region (Reference 14). Based on that analysis, it was concluded that monitoring thermal sleeve seal leakage is more important than RWCU reroute.

Therefore, based on the results presented herein, it is proposed that this analysis be used as a basis for continued operation for the next_several years until either the Brunswick 2 operating history can be updated or some form of leakage assessment can be made so that compliance with NUREG-0619 can be shown. Periodic examihaYion, required by NUREG-0619, of the feedwater nozzle will provide additional justification that continued operation is acceptable. -

2-1

NEDC-30633 REY. 1 l :? CLASS Ill

^

l

3. THERMAL CYCLE DEFINITION 9

^

The feedwater nozzle thermal cycle definitions are represented by Figures 3-1 through 3-4. These figures represent the ainlaua and maximum temperature I'

points for two startup/ shutdown events and two scram events obtained from CP&L N for Brunswick 2. The feedwater loop which has the RWCS airing (Loop B) had the most severe cycling and was therefore used throughout this analysis.

The following events appropriate for t' s analysis were identified by CP&L personnel from the last available fuel cycle.

Figure Date Event Description 3-1 5/3/88 Shutdown /Startup 3-2 5/7/88 Forced Shutdown 3-3 11/16/89 Scram 3-4 6/17/89 Scram i l

These events are depicted in Figures 3-1 through 3-4, and were digitized into computer form from microfilmed strip chart recordings of feedwater temperature. The temperature measurements were taken by hardware installed subsequent to the original issue of this report, and are symbolic of the cycling occurring at the feedwater nozzle. The events are shown exactly as digitized from the microfilm recordings; consequently, actual progress of each event is from right to left.

I A total projection of 183 startup/ shutdown events and 403 scram events over the 40-year plant life was made for Brunswick 2 based on operating data obtained from the first'14 years of operation (References 2, 3 and 4). This projection was determined using the methodology of Reference 5 as follows:

>y h

3-1

NEDC-30633 REY. 1 i

1. b CLASS III -

Number of Number of Year Time Period Startups/ Shutdowns Scranc*

1 3/75 - 3/76 5 36

,'I.. 2 3/76 - 3/77 1 13 3/77 - 3/78 3 2 22

[- 4 3/78 - 3/79 0 12

-cw

. 5 3/79 - 3/80 1 15 6 3/80 - 3/81 4 13 7 3/81 - 3/82 3 22 8 3/82 - 3/83 5 5 9 3/83 - 3/84 5 7 10 3/84 - 3/85 10 6 11 3/85 - 3/86 2 3 12 3/86 - 3/87 5 10 13 3/87 - 3/88 3 1 13.67 3/88 - 11/88 9 4

• Note: Turbine trips identified in Reference 4 were counted as scram events.

Although conservative, this method of counting is not considered to be a significant contributor to the final crack growth results.

Based on the methods of Reference 5, the following equation is used to determine the projected number of events for the 40 year design life:

i n

40 " "4 + "13.67[ 6 (1.67-0 )

where " Projected number of events at year 40 n40 n4 = number of events at year 4 "13.67 = number of events between year 4 and year 13.67 l

Thus, for startup/ shutdown events, the following is obtained:

i n

40 = 8 + 47(36/9.67) = 183 -

.s y h

3-2 e  :

NEDC-30633 REY. 1 4.. CIASS III -

(

l p.

and for scran events: l X

4,

~

n 40 = 83 + 86(36/9.67) = 403 c

i

,' For the purposes of this analysis, a single cycle is defined when tne

] nozzle fluid temperature, initially at some value TO ' " ""8** * * ** #

value Ty and then returns to TO '  !

For the purposes of defining complete startup/ shutdown cycles, the startup  ;

corresponding to the 5-7-88 forced shutdewn event (Figure 3-2) was assumed to be a mirror image of (i.e., identical to) the shutdown event. Therefore, two (2) complete startup/ shutdown cycles and two (2) complete scram cycles define j the thermal cyclic duty experienced by Brunswick 2. These events were assumed i

to be typical for the entire 40 year design life of the reactor.

I 1

a 4

e w

sy h

3-3 e

U$ inh '4 i

500

, 400 - I i

, y e

300 -

,' , s 9,_, g 8

4- l g

1 i i, H

e:

5 200 - il l' U g

. r >

i i si s i g ll ,

i si Il i Illi 100 -

o O ' ' ' ' ' ' ' ' '

0 so Tme (Hows)

Figure 3-1. Temperature Cycling for 5/3/88 Shutdown-Startup

.. .- . _ __ . . _ . . . ._ _.m... . . _ . . - . _ . . _ . _ - . . . . . , _ _ . _ . _ . . . _ _ _ _ _ . . . _ _ . _ . . . . _ . _ . . __. _ . _ . - _ _ _ . _ _ _ . _ .

• ': s$typ 500

'e 400 - is e

l C D 0 yu

• _,

e 300 - gg

= w H

E li U h, g . .

. 200- -

H II o

100 -

0 . . . , , . , , ,

O 17 >

Time (Hours)

Figure 3-2. Temperature Cycling for 5/7/88 Forced Shutdown i h

• @ g,g ,.

500

~

- 400 J

'c e Il o 300 - n m

a $

i h

- e, l s

'd I

s ie 200 - *

- I-l ' s t

i ii 100 -

fl 0

0 ' ' ' ' ' ' ' ' ' i i , , ' '

0 12 24 -

Time (Hours)

U 8"re 3-3. Temperature Cycling for 11/16/88 Scram

_ _ _ . . . . . _ . m...-~._.._

b 9. - Q t

500 400 - -

i 't U

. p "

g ,

!

• 300 - ,,

e ,

W h w O E 'l f 8 m

E !

ii O; s -

H pd g

200 - I "

Q H

100 -

• l

(

O , , , , , ,

270 -

Trne (Hours)  :

l Figure 3-4. Temperature Cycling for 6/17/89 Scram i t

NEDc-30633 REV,1

> ?

CLASS III -

4. THERMAL ANALYSIS

. s- The finite element computer code ANSYS (Reference 6) was used to develop i

. an arisymmetric model which simulated the Brunswick 2 feedwater nossle. The isoparametric heat conduction element (STIF 55) was used. The model was I developed using the nozzle configuration shown in Figure 4-1 (Raferences 7 and 44h 8). The configuration depicted in Figure 4-1 is the original nossle design without the cladding removed, as discussed in Section 1.3.3. The same model with an isoparametric stress element was subsequently used for the stress i

analysis. Further discussion of the model configuration is included in Section 3.

i The heat transfer coefficients and temperature boundary conditions were i derived from Reference 9. The method of derivation is explained in the Appen-1 dix to this report. The use of annular temperatures and heat transfer coeffi-cients removed the necessity of specifically modeling the thermal sleeve in the finite element analysis. The feedwater nozzle thermal sleeve design is a 1

single sleeve ring-interference fit to the feedwater nozzle safe end and

welded to the feedwater sparger. These heat transfer coefficients with the appropriate temperature boundary conditions are shown superimposed upon a l drawing of the finite element model in Figure 4-2.

[ The initiation of feedwater flow was modeled by varying the temperatures j

in Zones 2 and 3 from 550*F down to the temperatures indicated in Figure 4-2, over a 3-see interval. The temperatures were maintained at this level until 1

steady-state conditions were reached. The 3-sec ramp was used rather than a step change, since it is conservative and still assures numerical stability in i

the computer solution.

J 5

I

• 6 t'

1 . .

b w

4-1

+

e

, .l. '  ; ,'l i  !< 1  ! 1 !

NR,woOu ,

F' m gHH S

b S E

A N IDK C 9 AI 1 LH M 0 CT 2

yA pf o%aD 2

a

- =

=A s e 9 v 7

s. g g

2 s.

s

%I _

_M WM _

3 n n 8 f I

o S i t

N a 1

3 r 3

u g

0 S i 4 1 E f 4 H n 9 C o 1 N I

C N

L I e

E l R z

- ( A z

. E L

S o _

Z N N -

- Z O I

O_ S

. N .

E 1 3

o 5

N_ M I

8 4 D 4 _

7 .

L L

e A

r .

u g

=  :

5 E

- 4 T i

F

. O

- e N e

s 4 C e I N _

MI _

2 1 _

2

. +

. h 0 6

t e

1 6

e J 1

f S .

e:"

NEDC-30633 REY. 1 t CLASS III ,,

s, j r

l 38 2 HEAT HEAT TRANSFER BOUNDARY TRANSFER COEFFICIENT FLUID TEMPERATURE ZONE (8tu/hr ft 2.p) (.p) 1 864 550 20.0 ,, 2' 1665 126.6-373.4 3 705 373.4-550 l - .'.'. 4 0.2 120 y '

9N 4

\

'[ 13J -

N i

$ N  !

;  : @ a sneA/ l lt l l l'

74 -

i

=]

i l 5 ,

,'. I 1 I

,, ll AI

! 1 , , ,

93 ,,,.

l lIA

@ . lll: 7 l g l

, ~ 5.0 1 - 1 I

.0 10J 16.5 22 4 29.0 35.3 41.5 INCHES 4

, .. _o Figure 4-2. Thermal Boundary Conditions 4-3

NEDC-30633 REY. 1

. C' ASS III i 5. THERMAL AND PRESSURE STRESSES

p 3

.g The results of the thermal analysis were applied to the previously men-tioned finite element stress model to determine the thermal stresses. Iso-3 parametric stress elements (STIF 42) were used in the stress analysis. The

, nozzle was modeled by an axisymmetric finite element mesh with the vessel being represented as a spherical shell. This is a common approximation used

} in the stress analysis of a three-dimensional nozzle configuration in a cylin-drical shell. This was adequate for thermal stresses, but pressure stresses

, required a scaling factor based on a three-dimensional analysis. The lengths of the nozzle safe end and pressure vessel section were each modeled to at f This wa do to assure t t end effects did not i lu nce the stre ses in the nozzle corner.

Thermal stresses were evaluated during several time intervals over the course of the transient by analyzing node pair temperature differences at var-ious locations in the nozzle blend radius. Only the stresses occurring at 3 minutes were used in the subsequent crack growth analysis, since they resulted in the most limiting stress profile in the nozzle blend radius region. The highest thermal stress occurs on the inside surface of the nozzle blend radius as shown in Table 5-1. The thermal stresses which developed from a AT of 450*F were linearly scaled to the AT described in the thermal cycle definitions (Section 3). The scaled thermal stresses are subsequently used in the crack growth analysis.

The maximum thermal stresses occurred in the stainless steel cladding area of the nozzle blend radius (ends at Element 337; Figure 5-1).

1 Pressure stresses for the case of a 1000 psi vessel pressureewere also calculated; however, these stresses required application of a scaling factor.

This was necessary because of the limitation of modeling a three-dimensional component with a two-dimensional arisymmetric model. Because the three-dimensional characteristics near the nozzle cogner were not modeled, the peak

. 5-1

.I .'

t

~

NEDC-30633 REV. 1 l

l CLASS III -

l

. ,1

p. -  !

f Table 5-1

'k SURFACE STRESSES TO CHOOSE MAXIMUM COMBINED STRESSES (Steady State)

[ Thermal Hoop Pressure Hoop Pressure Ratioed Combined Element (pai) (psi) to 1.5004 (psi) l l (psi) l 169 -922 27196 40805 39883

.' 185 7217 28391 42598 49815 193 17122 29065 43610 60732

a. 209 26949 29519 44291 71239 217 35988 29725* 44600* 80588 y 233 43127 29661 44504 87631 241 49040 29295 43955 92994 257 54623 28664 43008 97631 265 59953 27663 41506 101459 281 64940 26293 39451 104391 289 68995 24462 36703 105698 305 72933 22808 34222 107155 313 76412 l 21232 31857 108269 329 79978 19821 29740 109717*

337 80273* 18380 27578 107851 353 56119 16725 25095 81213 369 52698 14798 22203 74901 l

l

• Maximum stress

=

b 5-2 e

NEDC-30633 REV. 1 CLASS III ,

Table 5-1 SURFACE STRESSES TO CHOOSE MAXIMUM COMBINED STRESSES (Continued)

(3 min after beginning of transient)

, Thermal Hoop Pressure Hoop Pressure Ratioed Combined Element (psi) (psi) to 1.5004 (psi) (psi)

I

,, 169 -3808 27196 40805 36998 l

) 185 5089 28391 42598 47688 193 15716 29065 43610 59326 209 26726 29519 44291 71017  !

217 37160 29725* 44600* 81760 l 233 45956 29661 44504 90460 241 53733 29295 43955 97688 l 257 61359 28664 43008 104367

! p. 265 68870 27663 41506 110376 1 281 77127 26293 39451 116577

'3 289 84835 24462 36703 121538

!j 305 92060 22808 34222 126281

313 98704 21232 31857 130561 329 105297 19821 29740 135036 337 108208* 18380 27578 135786*

l 353 88351 16725 25095 113445 369 88634 14798 22203 110837

)

j

• Maximum stress s j i .

1 i

5-3 f

I

- - - - - ~ ~ ~

)

NEDC-30633 REV. 1 CLASS III .

i 1

f l\/\

\1 L 1 IIlilifff i

l-

" 2 a

3 1

1 lx 1\

\ \\\

/ ~-

x I

\ \\ \ \

\ \\ \ \ \

[ N

\ \\ \ \\

f

/ N

\ \\ \ \\

i \\ \ \\

\J /\ /\ g (

V 169

'58 i

I \\\ \ \ \ \ \ \

t l

/ 185 i$$ \ \ \ \ \ \ 209

' t 233 447 434 421 408 382 353 \329 305 281 257 241 h 395 369 337 313 289 205 1 3 MAXIMIM PRESSURE STRESS: 44,600 psi I

i N MAXIMUM M^X' MUM COMBINED THERMA STRESS: 109,717 psi STRESS: 80,273 psi

>y Figure 5-1.

Feedwater Nozzle-Brunswick"2 Location of Maximum Surface S$ress (Steady State) '

5-4 e

NEDC-30633 REY. 1 CLASS III ,

a I

A i LJilljjj;j;

2. '

2

{ 1. 3 s

1

/

lx \\

\ \\\

/ -x

\ \\ \ \

[ x

\\\\ \ \

\ \\ \ \ \

f x i \\ \ \\

\ \\ \ \\ G

'58

/ \i \\\

li A \ \A /\ O

\ xxx 185 SS 1,\\ \ \ \ \

'1 209

' i 233 447 434 421 408 395 382 353 \329 305 281 257 389 337 313 289 265 241 h 1, 3 MAXIMIM PRESSURE STRESS: 44.600 psi

, M xlMUM

/ MAXIMUM COMBINED STRESS: 133,786 psi hep l

STRESS: 108,208 psi i _

i

.6 ,\

• Figure 5-2.

Feedwater Nozzle-Brunswick 2 Location of Maximum Surface , Stress (Transient t = 3 minutes) 5-5 e

NEDC-30633 REY. 1 CIASS III -

stresses at the nozzle corner were not accounted for accurately. Therefore, a generic three-dimensional model developed by Gilman and Rashid (Reference 10) was used to scale the pressure stress. The scaling factor for the pressure stress is given by the ratio of the peak pressure stress on the inside surface

. reported by Gilman and Rashid to the peak pressure stress on the inside sur-face from the finite element model used in this reporr. The peak pressure stress of the finite element model was 29,725 psi, while the peak pressure stress reported by Gilman and Rashid is 44,600 psi. This resulted in a scal-ing factor of 1.5004. The scaled peak pressure stress on the inside surface is shown in Figure 5-1.

The combined charsal and scaled peak pressure stresses were examined to determine the area where the combined peak stress on the inside surface occurs, as shown in Table 5-1.

1 '

The stresses at the cross section associated with the limiting stress profile (see Table 5-2, and cross section 3-3 on 1

Figure 5-2) were used to calculate crack growth.

d i

1 i

i i

. Ni h

i 5-6 l

NEDC-30633 REV. 1 CLASS III ~

Table 5-2 LIMITING STRESS PROFILE (CROSS SECTION 3-3)

.; Distance from Pressure Hoop Inside Surface Thermal Hoop Ratioed to

/ (in.) 1.5004 (psi) 3 minutes

, (psi)

, 0.0 27578 135786 0.075 27029 130367 0.225 26325 96223 0.400 25242 88215 0.600 24161

79987 O.850 23049 69259 1.150 21849 59016 1.500 20587 47598 1.885 19843 36964 2.384 18253 26363 3.007 16571 16994 3.631 15084 10689

4.255 13705 6691 4.879 12360 4260 5.503 10976 I 2782 i

l a

i

• l

.s y 1

i

=

5-7

? .

I

NEDC-30633 REY. 1 l CLASS III

6. CRACK GROWTH ANALYSIS 6.1 STRESS INTENSITY FACTOR CALCULATIONS

Stress intensity factors were determined using solutions for standsrd stress distributions (e.g., constant, linear, quadratic, and cubic variations) and applying the superposition principle. The stress intensity solution for these unit load cases was expressed in terms of the crack length and appropri-ote magnification factors for the specific crack geometry (Figure 6-1). The

ctress intensity for an arbitrary stress distribution was then obtained by fitting a third-order polynomial of the forms

+ +

o=A0+Al 2 3

)

! cud applying the principle of superposition. Once the curve fit parameters g AO , A y, A ,2 and A were 3 known, the stress intensity factor was determined as a function of crack depth using the equations in Figure 6-1.

Magnification factors for several common two-dimensional geometries are f cve11able in References 11 and 12. For the feedwater nozzle, a set of three-5

{ dimensional magnification factors is presented in Reference 1. As illustrated in Figure 6-1, the nozzle corner factors (0.706, 0.537, 0.448, and 0.393) were

! obtained by averaging the magnification factors developed for circular surface I l crack geometries in half and quarter spaces. This expression (labeled FUN 11) l was used to calculate stress intensity factors in the fracture mechanics eval-uation which follows.

The pressure and thermal stress distributions were fit to third-order polynomials using a standard least squares procedure. Overall accuracy of the polynomial representations is considered more than adequate.

Substituting these polynomial coefficients (A O

, yA , A and A ) into the

' 2 3 FUN 11 stress intensity fac, toy expression of Figure 6-1 leads to the stress intensity factor versus crack depth data shown 1,n Figures 6-2 and 6-3. (These stress inte'nsity factors apply to cross section 3-3.)

6-1 e

_ _ _ _ _ _ - _ - - - - - _ - - - - - - - - - - - - - - - - - ~ ~ ~'

NEDC-30633 REY. 1 CIASS III a _

I t- CX I

_ -- j -

FUN 9 - SEMI CIRCULAR CRACK IN HALF SPACE 2 3 Kg* /Te 10.884 A, + 0322 (2 ale) Ag + 0 434 (4 /21 A2 + 0.377 Ide ns) A 31 i

I X

a L

FUN 10 - OUARTER CIRCULAR CR ACK IN QUARTER SPACE j {A.

2 K, = /73Il0 723 A, + 0 ssi ca/el Ag + 0 as2 u /21 A2

• 0 808 'd* O' ^3I l

l.

! I l

l l I

i

\  ! x i \

\ _ .___ _ __ l l

\

\ -

4 FUN 11 - SIMULATED 3 0 NOZZLE CORNER CRACK 2 3 K, = S10.70s,A,,+ 0.537 Ga/el Ag + 0.448 te /21 A2 + 0.393 44e /3el A31 l,

l Figure 6-1. Boundary Integral Equation / Influence Function Magnification Factors for BWR Feedwater Nozzle l 6-2 l .

l 90

60. -

_ 70 -

r E w 80 - i e- M i

" ts AT = 450 F (550 - 10(fF) y, 0, ,

]

I 50 - 3 Minutes After Transient Cross Section 3-3 n"

j 40 -

g 2 '

g 30 -

5 20 -

o I

10 -

0 , , , , , , , , , , , , , ,

. 0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 Crack Depth Along Section (inches)

Figure 6-2. Stress Intensity Factor versus Crack Depth (Thermal Stresses, 3 minutes)

i 45 40 --

'e _ 35 -

c E .5 l

-3 30 - R 6 9 g o&

I $' 25 - Resswe - 1000 psi . O I s

'E 20 -

Cross Section 3-3 0 g; c ,

• e

% i g 15 -

D r u) 10 -  !

o 5-

' i O i i i i i i . . . . . . . i i

O O.4 0.8 1.2 1.6 2.0 2.4 2.8 Crack Depth Along Section (niches)

Figure 6-3. Stress Intensity Factor versus Crack Depth (Pressure Stresses)

MEDC-30633 REV. 1 CLASS III -

6.2 CRACK GROWTH DATA

! Figure 6-4 represents the fatigue crack growth data for low alloy steel from Section II of the ASME Code (Reference 13). The R-ratio (K ,g,/K )

dependence of this data is built-in by representing three cases

(1) R-ratio less than 0.25, (2) R-ratio between 0.25 and 0.65, and (3) R-ratio greater than 0.65. These data were used to deteratine the growth of an assumed 0.25-inch initial depth crack.

The best-fit compilation of fatigue a ack growth data used in the origi-j nal revision of this report was not used in the current analysis. That rela-S tionship is not valid for high values of R-ratio. Much of the data obtained for Brunswick 2 yielded high R-ratios (>0.9). As a result, unrealistic crack growth would result from the use of this relationship.

l h 6.3 CRACK GROWTH EVALUATION

!I g

The thermal cycle definitions are represented by Figures 3-1 through 3-4

{

for startup/ shutdown and scram / return to full power events. A projected total of 183 startup/ shutdown events and 403 scram events was made for Brunswick 2 over the 40 year design life of the plant as described in Section 3.

l The analysis conservatively assumed that the initial crack depth of 0.25 inch included the cladding thickness. Since the thermal stresses are j higher in the stainless steel cladding region, the corresponding stress intensity factor would also be greater, thereby resulting in a more rapid i crack growth propagation.

II The procedure for calculating the crack propagation is as follows: for

]l sach cycle, the maximum and minimum stress and the number of occurrences were

~

a calculated. From this, the stress intensity factor range and corresponiting~

R-ratio weta calculated for the cycle being analyzed. Using this and the selected crack growth rel'azionship, the incremental. crack growth was

.l =

6-5 i

i

NEDC-30633 REY. 1 i

l CLASS III ,

I

) 1000 /

i

/

l *LINE AR INTE RPOLATION il f l

[

700 "

RECOMMENDEO TO ACCOUNT &

) FOR RATIO DEFENDENCE OF E, [ I

~ WATER ENVIRONMENT CURVES. '

FOR 0.25 < R < 0.86 FOR

[/

2 SHAL LOW SLOPE 185

&2

, ele /dN

• 11.01 s 10-1) O= A K .& 4D

• 3.78 R + 0.06 & /

v R"K menIKmes h

I /

I $

i j

100 =

SUS $URFACE FLAWS (AIR ENVIRONMENTI  %

g de/dN * (0.0267 a 10-3) Ag K . 6

/ [g 5

,0 -

/ er ee I 4

• 4 I N ~ OETERMINE THE AK AT WHICH j $ THE LAW CHANGES SY CALCULATION [ f j

Z OF THE INTERSECTION OF THE TWO CURVES g

i i

$URF ACE F LAWS l $ (WATER RE ACTOR f i w ENVIRONMENTI

) N 20 - APPLICABLE FOR j 5 a < 025

/

j +0.25 < R < 0 65 R > 0 65 l 10 g.g fg mm mas

( 5 f

( 7 =

f 5 - LINE AR INTE RPOLATION l$  !

(

J f ( RECOMMENDED TO ACCOUNT )

l i g- FOR R RATIO DEPENDENCE OF j 0

g WATER ENVIRONMENT CURVES.

I = "

" FOR 0.26 < R < 0A5 FOR STEEP i

! R

- / ~

o SLOPE.

! f da/dN * (1.02 s 104)Og AK5M f

)

E / 2 Og

• 2ssR - s.725 o E j

• f R = K,g/K ,

\

I \

l , i i i / i e i

~ ~

1 2 s 7 10 20 s0 70 100 1

STRESS INTENSITY F ACTOR RANGE (AK g ks4 68

>y l

r Figure 6-4. Reference Fatigue Crack Growth Curves for Carbon and l Low Alloy ferritic Steels l

l 6-6 i

l

NEDC-30633 REV. 1 CLASS III -

i

calculated for that event. The crack size was updated and the procedure

} repeated. This continued for every cycle until the entire life was analyzed.  !

I The pattern of events was assumed to bet

~

25 sets of: 8 startup/ shutdown events followed by 16 scram / return to full power events Note that 25x8 = 200 startup/ shutdowns and 25x16 = 400 scrams.

The effects of modeling the seventeen extra startup/ shutdown events and three less scran events is considered small.

i l

l The eight startup/ shutdown events were further broken down as follows:

i 4 5/3/88 Shutdown-Startup events (Figure 3-1)

~

4 5/7/88 Forced Shutdown events (Figure 3-2) j (including the mirror-image startup)

Total = 8 1

4 >~: The sixteen scram / return to full power events were further broken down as s

follows: 1 L

i 11/16/88 Scram events (Figure 3-3) 8 l

8 6/17/89 Scram events (Figure 3-4) i l Total = 16 '

One crack propagation calculation was made corresponding to the limiting stress profile shown in Table 5-2. '

l t .v  ;

1 i . .

m 6-7 l

=

lDJ l

NEDC-30633 REY. 1 CLASS III ,

7. RESULTS AND CONCLUSIONS Because of the recent acquisition of plant-unique data, the feedwater nozzle thermal cycle definitions defined in Section 3 were assumed to be representative of all startup/ shutdown and scran events. The number of events over the plant life as projected using plant specific data (References 2, 3 and 4) was utilized. A plant-specific finite element stress analysis was

• performed for the feedwater nozzle.

i The fracture mechanics analysis was based on the thermal stresses l obtained from the finite element analysis, the thermal cycle definitions derived from actual plant feedwater data, and the historical frequency of the

, number of startup, scram and shutdown events.

4 P The results of the fracture mechanics analyrf.m are given in Figure 7-1 1

for the limiting location (cross section 3-3) as a function of the number of

{

years since initial plant startup.

I Using the 1989 ASME Section XI fatigue crack growth curves, the analysis shows that the postulated 0.25-in. crack becomes 1 inch deep 32.3 years after the startup date.

! Even though compliance with NUREG-0619 cannot be shown at this time, there are approximately 16 years in which to update the Brunswick 2 operating history. The operating history used in the current analysis was based on an extrapolation of the initial 14 years of plant operation. Because of

" learning curve" effects which are typically experienced by operating reactors during their initial years of operation, this extrapolation is most likely conservative (as recognized by the large number of startup/ shutdown events).

Therefore, thie cycle counting information should be continually. updated so that any conservatism present in the 40 year projection can be identified and eliminated. If a significant conservatism is identified, it may be possible to demonstrate compliancg with the requirements of NUREG-0619.

h 7-1 e

,ll mB7uoe0 h - .

D>sn = pH

• o s

. a $ o n ,,

• J 6 j o E n E t !!!!!!!l-lijiiii !Ij.!I.! 0 s

- 4 r

- a e

Y f

o r

-- e

..lli:l';..::l  ::.l'  :..- b

- m u

a. s--

e-0 N g h-3 s c u n- s s i- r r

- a e 0.- e v 1- s Y

. - hr f h

t a o t h-=- Pe ey r p

, t p- e e e- D 3 b D

,20 mu D- k k c c k-c-

a2 r3 N a r

a-r- Cne i C

C- -

l d e- be .

ah

^ lb-a- owca 1 7

w- l e B

o- AR e r

A- ,oi u

g i

F

~ -

^

o 2 1 0 9 8 7 6 5 4 3 2 1

. . . . - . o.

1 1 1 O 0 o 0 0 0 0 0 O o g E 8 5aeo f 8O

e,~

nA

~

NEDC-30633 REV. 1 CLASS III

• l In addition, conservative heat transfer coefficients were used to accommodate the potential for possible leakage flow past the thermal sleeve seal (see Appendix). Although the Brunswick 2 configuration is a ring-interference fit, degradation of the seal is assumed to occur with time as a result of corrosion and other forms of relaxation. To account for this degradation, leakage flow was assumed in the analysis. Since this assusp- l tion has a significant impact on the resulting thermal stresses (as recognized by Brunswick 2 stresses being higher than Brunswick 1), it is recommended that ,

some form of leakage assessment or monitoring be performed. This monitoring would provide an actual measurement of the leakage flow rate which cculd then be factored into this analysis as appropriate.

Based on the results of the above recommendations, the crack propagation j could potentially be re-evaluated to show full compliance with NUREG-0619 requirements. During the interim period, the NUREG-0619 required examinations can provide justification for continued plant operation.

f

!i i

lk' i

h j 4 l l

t l

I i

% a 1

• 6 f l .

7-3 e

, NEDC-30633 REV. 1 CLASS III -

8. REFERENCES

1. NEDE-21821-A, " Boiling Water Reactor Feedwater Nozzle /Sparger Final Report," General Electric Company, August 1979.

2. Letter EGE-146, File BG0029A-AA-A500, J. M. Brown (CP&L) to G. L. Stevens (GE), " Brunswick Steam Electric Plant, Unit Nos. 1 & 2 NUREG-0619 Analyses, PCN G00029A Information Transmittal," November 17, 1989.

3. NEDO-22196, " Reactor Pressure Vessel Thermal Cycle Fatigue Evaluation for Brunswick Steam Electric Plant Units 1 and 2," General Electric Company, March 1983.

I

4. Letter EGE-164, File BG0029A-AA-A500, A. M. Lucas (CP&L) to G. L. Stevens (GE), " Carolina Power & Light Company, Brunswick Steam Electric Plant Units 1 & 2, Feedwater Nozzle NUREG-0619 Crack Growth Analysis Update,"

December 10, 1990.

5. GE Services Information Letter (SIL) Number 318 "BWR Reactor Vessel Cyclic Duty Monitoring," December 1979.

g 6. G. J. DeSalvo and J. A. Swanson, "ANSYS Engineering Analysis System User's Manual," Swanson Analysis Systems, Inc., May 1, 1989, Revision 4.4.

7. GE Drawing 767E723, Rev.1, "Feedwater Nozzle Safe End," July 1975. l l

4 8. Chicago Bridge and Iron Drawing Number 31, Rev 17. "12-In. Diameter i Feedwater Nozzles," October 1971.

k 9. NEDE-21659-1, C. M. Kwong and H. Choe, " Moss Landing Feedwater

{ Nozzle /Sparger Test Data Files," February 1979.

10. J. D. Gilman and Y. R. Rashid, "Three-Dimensional Analysis of Reactor Pressure Vessel Nozzles," Proc. lat Inc. Conf. on Structural Mechanics in Reactor Technology, Vol. 4 Part C, September 1971.

11. ASTM-STP-590, C. B. Buchalet and W. H. Bamford, " Stress Intensity Factor j

Solutions for Continuous Surface Flaws in Reactor Pressure Vessels,"

Mechanics of Crack Growth, American Society for Testing and . Materials, 1975.

, 12. ASTM-STP-590, R. Labbens, A. Pellissier-Tanon and J. Heliot, " Practical t i

' Method for Calculating. Stress Intensity Factors Through Weight Functions," Mechanics of Crack Growth, American Society for Tes, ting and Materials, 1975. ,

1 l' l , 13. ASME Boiler and Pressure Vessel Code,Section XI, 1989 Edition.

I

,u

14. NSEO-75-882, " Effects of Reactor Water Cleanup Reroute on Feedwater Nozzle Fatigue Usage, Brunswick Steam Electric Plant Units 1 and 2,"

August 1982. ,

I 8-1 t

l

l NEDC-30633 REV. 1 CLASS III ,

7 APPENDIX THERMAL BOUNDARY CONDITIONS A.1 HEAT TRANSFER COEFFICIENTS The annular heat transfer coefficients were developed from the data of Reference A-1 as follows. An allowance was made for potential leakage flow past the thermal sleeve seal as a result of degradation of the ring-interference fit. As a result, the data base consisted of two tests which i

were run at low feedwater flow and approximately 4 GPM leakage flow. The heat transfer coefficients during the tests were determined from eight heat flux meters mounted circumferential1y around a section of the nozzle blend radius.

i j The highest heat transfer coefficient measured was taken and corrected to 2

account for the difference in cozzle blend radius between the test sparger and the Brunswick 2 sparger. The Nusselt number, Nu, is proportional to the i

Reynolds number to the nth power, where n is typically 0.8. The Reynolds num-l ber is in turn directly proportional to the nozzle blend radius, R. There- l l fore, the Nusselt number is proportional to the nozzle blend radius raised to

$ the 0.8 power. In equation form, 4)

Nu a R

• i The heat transfer coefficient, h, is given by h = Nu (k/R) l where k is the thermal conductivity of the fluid. Thus l -0.2 I

haR This proportionality is used to correct the heat transfer coefficient. In the I

tests of Reference A-1,>h a?. 1767 Btu /hr-ft *F and R = 2 inches. For the Brunswick 2 sparger, R = 2.69 inches. Therefgre, b = 1665 Btu /hr-ft, *F.

A-1 e

NEDC-30633 REY. 1 CLASS III A.2 BOUNDARY TEMPERATURE CONDITION Boundary (or annulus) temperatures were taken from both of the aferemen-tioned two tests. The second test had two data samples taken, so a total of three data samples were available. The test data is expressed in terms of a normalized temperature which is equal to the difference of the annular fluid temperature and the feedwater temperature divided by the difference of the reactor temperature and the feedwater temperature. Readings are available at several circumferential locations at four sections of the nozzle. At each section, the lowest readings for each test were averaged to produce the final result.

The annulus fluid temperatures, as determined from these Moss Landing tests, are given in Table A-1 as a function of position of the ring- i

, interference fit nozzle configuration. The expression for obtaining the

{ annulus fluid temperature is as follows:

1 l

T=TFW + C1 (Ty -TFW) j f where h

j j T = ar.nulus fluid temperature (*F)

FW = feedwater temperature (*F)

T

! Ty = vessel temperature (*F) l Cy = coefficient from Table A-1 i

A.3 REFERENCE I

ii A.1 NEDE-21659-1,. C. M. Kwong and H. Choe " Moss Landing Feedwater j; Nozzle /Sparger Test Data Files," February 1979. -

'l .

m I A-2 f

• _ . _ _ _ _ . . . - - . - . - . . - . - ... . . - - - - - - - - - - - - --- - -- - ^^

i

NEDC-30633 REV. 1 CLASS III ,

i i

Table A-1 C1 COEFFICIENT j (GE Proprietary) l <

Location 1 l F2 0.0257 i

A 0.0592

. B 0.1578 4

C 0.6075 i

i 1

1 ,

{ Use linear interpolation between locations as

! illustrated in Figure A-1.

E l

I h i r l Y 4

i t

f i

l ~

t

>y j . .

f

' 1 A-3

i.

. I

, lh

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

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

THICK WALL $8MULATOR Q METALSURFACE SENSOR -e-O FLuiosENson 8 C+ e - VEME L WALL

=

F2+ A ==-

1. (J ()

DtREdT)ON OF FLOW e

( --

.- _ g O 8

} La V

en so ta

\ () H

()

. F2+ A+ .

^

p e .

C+

--e- -o-1.12 SN.

o 5.42 IN.

• T 17 87 tN. r t

25 si sN.  :

Figure A-1.

Instrumentation of Feedwater Nozzle and Vessel Wall (from Moss Landing Tests, Reference A-1)

1 NEDC-30633 REV. 1 i

CLASS III t

t i

DISTRIBUTION 1

i Name 4 M/C i

K. F. Cornwell 732 S. Ranganath 747 2

i W. Yee BRU A. D. Ketcham BRU 1

{

G. L. Stevens (15) 747 4

GE-NE Library (2) 1 528  !

4 1

4 i

i f

.i i

i l

l l

l i

i I

4 l

4 .

i

.s y

.i j = e i

a e ,

i

= -