ML20076C767

Rev 0 to Structural Integrity Assessment of Browns Ferry Feedwater Nozzles

ML20076C767
Person / Time
Site:	Browns Ferry
Issue date:	08/31/1983
From:	Gustin H, Hsu L, Steinert L NUTECH ENGINEERS, INC.
To:
Shared Package
ML20076C746	List: ML20076C739 ML20076C767
References
XTV-02-008-R00, XTV-2-8-R, NUDOCS 8308230173
Download: ML20076C767 (57)

Text

.

- o

~

Controlled Copy No. I XTV-02-008 Revision 0 August 1983

~ File Number:

165.1202.0100 STRUCTURAL INTEGRITY ASSESSMENT OF BROWNS FERRY PEEDWATER NOZZLES Prepared for Tennessee Valley Authority Prepared by:

NUTECH Engineers, Inc.

San Jose, California d _

Prepared by: Reviewed and Approved by:

! m H. L. Gustin, P.E. L. C. Hsu, P.E.

~

Project Engineer Staff Consultant

- Issued by:

Date:

August 9, 1983 e,/

L. D. Steinert Project Manager

~

te.W PDR ADOCK 05000259 PDR p

L

-- REVISION CONTROL SHEET TITLE: Structural Integrity OOCUMENT FILE NUM8ER: 165.1202.0100

- Assessment of Browns Ferry Revision 0 Feedwater Nozzles H. L. Gustin / Senior Engineer NAME t TITLE INITIALS L. C. Hsu / Staff Consultant h INITIALS NAME/ TITLE NAME / TITLE INITIALS NAME / TITLE INITIALS DOC PREPARED ACCURACY CRITERIA "8"#"*8 APPECTED PAGElS) REV SY I DATE CHECK SY / OATE CHECK SY / 0, ATE i -

v 0 IU h f h 7/#3 1.1 through 1.3 2.1 through 2.4 3.1 through 3.5 4.1 through 4.6 1

5.1

- through l 5.27 , ,

6.1 through 6.2 7.1 \p i thr=,h 2.2 y

o #jAh+83 jajn Y Q y7 V

l I

~

PAGE I OP 1 08" 3'3 ' '

1 l.-

ABSTRACT

-- The Browns Ferry Nuclear Power Station reactor vessel feedwater nozzles were evaluated analytically to determine the rate of fatigue damage accumulation for a range of thermal sleeve seal leakage rates. Results are presented which correlate leakage rate with fatigue usage factor.

The operation of the feedwater system at low flow rates was studied to determine the extent to which low flow / unsteady flow-induced thermal cycling affects the feedwater nozzles. The conclusion is that present Browns Ferry equipment and operating procedures are adequate to minimize the crack growth concerns of

~

NUREG-0619.

A proposed Reactor Water Clean-Up system cross-tie modification was evaluated as a possible method of mitigating thermal cycling at the feedwater nozzles. The modification was shown to have no

-- significant ef fect on f atigue due to thermal cycling.

W M

m XTV-02-008 11 Revision 0 nutagh

2 TABLE OF CONTENTS Page ABSTRACT ii LIST OF TABLES iv

~

LIST OF FIGURES v

1.0 INTRODUCTION

1.1 1.1 Background 1.1 1.2 Objectives 1.2 2.0 DISCUSSION OF THERMAL CYCLING IN FEEDWATER NOZZLES 2.1

-- 2.1 Types of Thermal Cycling 2.1 2.2 Generic Mitigation of Fatigue Effects 2.3 3.0 COMPONENT DESCRIPTION 3.1 3.1 Geometry 3.1 3.2 Material Properties 3.1

_ 3.3 Functional Description 3.1 3.4 Loading Conditions 3.2 3.5 Analysis Criteria 3.3 4.0 LEAKAGE RATE CORRELATION 4.1 4.1 Introduction 4.1 4.2 Analytic Procedures 4.2 1 4.3 Use of Field Leakage Monitoring Data 4.3 5.0 PEEDWATER NOZZLE STRUCTURAL INTEGRITY ASSESSMENT 5.1

~

5.1 Introduction 5.1 5.2 Plant Specific Thermal Duty Map 5.1 5.3 Rapid Temperature Cycling 5.2 5.4 Unstable Flow / Low Flow Evaluation 5.9 5.5 System Cycling 5.11 5.6 Total Fatigue Evaluation 5.12

-- 5.7 Reactor Water Clean-up Cross-tie l 5.12

' Evaluation 5.8 In-service Inspection 5.15

6.0 CONCLUSION

S 6.1 6.1 Results 6.1

_ 6.2 Recommendations 6.2 l

7.0 REFERENCES

71 l

XTV-02-008 iii i

_ Revision 0 I

nutagh

LIST OF TABLES Table Pace Plant Specific Thermal Duty Map 5.17 5-1 5-2 Amplitude vs Frequency 5.18 5.19 5-3 Constant C3 5.20 5-4 Constant C4

~

5-5 Rapid Cycle Fatigue Usage:

1.0 GPM Leakage 5.21

- 5-6 Rapid Cycle Fatigue Usage 1.25 GPM Leakage 5.22 5-7 Rapid Cycle Fatigue Usages

~

1.50 GPM Leakage 5.23 5-8 Rapid Cycle Fatigue Usages Leakage Increasing with Time 5.24 5-9 Rapid Cycle Fatigue Usage: -

Contribution by Thermal Duty Region 5.25 Total Fatigue Usage 5.26 5-10 l

l l

[ _

! XTV-02-008 iv

! - Revision 0 nutggh l

i i

i - . - . _ _ __ , __ _ -_.,.. . __ ..

LIST OF FIGURES Figure Page

_ 3-1 Feedwater Nozzle / Thermal Sleeve

Configuration 3.4 3-2 ASME Section III Design Fatigue Curve (Extended) 3.5 4-1 Analytical Predictions for Normalized External Temperatures 4.6 5-1 Fatigue Usage vs Leakage Rate 5.27 e

1 i

l l

l

~

1 i

I l ...

l l XTV-02-008 v

- Revision 0 I O I

-%-. ._7, ,.-.- ,-- , - , ,-.- - - . - - - ,,.- e.-- ---y-, _.-,-------,-w ~ - ,

---

w- --- - - - , - - y.-_. ,,,-,-..wi-,, .. w- w----,a .- -. - = -

1.0 INTRODUCTION

1.1 Backcround-In the late 1970 's Tennessee Valley . Authority (TVA) replaced the original-equipment loose _ fit thermal sleeve

~

feedwater spargers at Browns Ferry Nuclear Power Plant with interference fit thermal sleeve spargers, and later with triple sleeve spargers. The purpose of these modifications is to minimize . feedwater leakage past the

-- thermal sleeve, and thereby ninimize thermal cycling in the vicinity of the feedwater nozzle. Rapid thermal cycling-induced fatigue is a major cause of feedwat nozzle cracking. Rapid cycling fatigue damage is a strong function of feedwater leakage rate past .tha thermal sleeve seals.

Thermocouples are being installed on- the ~ cutside of the feedwater lines at the junctions with the feedwater

-- nozzles. These thermocouples wi 1 be used to monitor feedwater leakage past the thermal sleeve interference t

~~

seals. By monitoring leakage v'eraus time,' TVA is taking action to anticipate any cracking occurrence.

~

s A second f3tigue process which may af fcct the service lifyok,feedwaternozzles is that of low frequency ,

v s ,

XTV-02-008 o 1.1 -

Revision,0 ,

, O n.

g ~, * a

'\  % ,,

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

W temperature cycling at the feedwater nozzles due to

~

unsteady operation of the feedwater system at very low power levels. Although low frequency cycling does not contribute significantly to crack initiation in BWR feedwater nozzles, this phenomenon can have a strong v

effect on crack growth. TVA has collected temperature data in the feedwater lines immediately upstream of the thermal sleeve seal to determine the extent of low

-- frequency thermal cycling experienced by the feedwater nozzles.

1.2 objectives The NUTECH scope of work on the Browns Ferry feedwater nozzle evaluation program consists of four tasks.

1. Evaluation of feedwater nozzle modifications

-- performed on Browns Ferry Units 1, 2, and 3.

NUTECH has evaluated the effects of rapid thermal l

~

cycling-induced fatigue on the feedwater nozzles for plant-specific operational characteristics and configurations. In addition, fatigue effects due to system cycling have been studied. Predictions of the effects of thermal sleeve seal bypass leakage on feedwater nozzle fatigue life have also been developed.

~

XTV-02-008 1.2

-- Revision 0

-. 2. Study of low flow controller requirements.

NUTECH has studied the details of Browns Ferry Feedwater System operation at low flows, to determine the effects of low frequency thermal

~

cycling due to system operating characteristics.

3. Reactor water clean-up crosstie study.

NUTECH has evaluated the ef fectiveness of a proposed Reactor Water Clean-Up Crosstie modification in reducing nozzle fatigue.

~

4. Feedwater nozzle inservice inspection requirements.

NUTECH has prepared predictions of fatigue usage vs time for several postulated bypass leakage rates.

Since nozzle cracking is traceable to thermal fatigue which in turn is related to leakage rate,

_. continuous monitoring of leakage coupled with a fatigue / leakage correlation such as presented herein should provide an acceptable alternative to in vessel nozzle inspection. However, since the

~

leakage monitoring system is not yet operational at Browns Ferry as of the date of this report, the i results presented are parametric in nature, rather than quantitative.

)

xTv-02-008 1.3 Revision 0 nutagh i

.2.

2.0 DISCUSSION OF THERMAL CYCLING IN FEEDWATER NO22LES 2.1 Types of Thermal Cycling In the early 1970's, Boiling Water Reactors world-wide began to experience problems with cracking of feedwater nozzles and related components such as safe-ends.

Cracks in the nozzles, and to a lesser extent the safe

-- ends, are traceable to fatigue derived from thermal and pressure cycling. Three different types of cycling have

~~

been shown to be significant contributors to the cracking problem. These are

1. System Cycling: This category includes major operational transients such as start-ups, shut downs, etc. The fatigue effects of these transients were considered in the original

-- feedwater nozzle stress report (Reference 5).

i l

~

2. Rapid Thermal Cycling: This type of thermal cycling is caused by the turbulent mixing of hot reactor water with relatively cold feedwater during steady state operation. The feedwater nozzle blend radius has historically been susceptible to fatigue

_ caused by cycling of this sort. Feedwater leakage which bypasses the thermal sleeve produces signifi-XTV-02-008 2.1 l

-- Revision 0 nutaq, b 1

~

cant thermal cycling in the vicinity of the nozzle,

.i which has been shown (Reference 2) to be the dominant cause of crack initiation in the feedwater

-- nozzle.

~

3. Unstable / Low Frequency Thermal Cycling: This phenomenon is experienced primarily at low power and hot standby conditions, when the feedwater system is operated in an unsteady manner. When a plant is in hot standby, some steaming still occurs due to decay heat. As a result, although the full flow capabilities of the feedwater system are not

- required, *t is occasionally necessary to add water to maintain reactor water level. In addition,

~

since the feedwater heaters are usually not available durinq such operation, the feedwater temperature is low. The cycling occurs as the feedwater nozzles, which contain essentially stagnant reactor water (high temperature), are

- flushed with cold feedwater, at a frequency of a few cycles per hour.

1 -

Low frequency cycling of this type has been shown (Reference 2) to have minimal effect on crack

~

XTV-02-008 2.2 Revision 0 nutgg;b

,- +w,- --wr--y- ---,,s ---,,~--,--r--,r-vp>e -,----ern--*y------ww-- +- -w-- n -,

initiation (that is, it contributes little to

-- fatigue usage factors), but to have appreciable effect on crack growth.

2.2 Generic Mitication of Fatique Ef fects The effects of system cycling are unchangeable without making major operational changes. Consequently, the designers and owners of af fected Boiling Water Reactors have emphasized mitigation of the effects of rapid and unsteady (low frequency) thermal cycling in their responses to feedwater nozzle cracking.

Leakage of feedwater past the thermal sleeve is the principal cause of rapid cycling. New thermal sleeve designs have been developed to minimize such leakage.

] TVA has installed double piston ring seal, triple thermal sleeve /sparger equipment to replace their -

original loose-fit equipment. In addition, thermo-couples are being installed on the outside surface of the feedwater line at the pipe-to-nozzle junction.

~

Experience has shown that feedwater leakage past the thermal sleeve seals can be detected and quantified by monitoring feedwater nozzle temperatures in such a manner. Significant leakage past the thermal sleeve seals can be anticipated in time to prevent accumulation I

XTV-02-008 2.3 Revision 0 nutggb

- , , , , , ..--,~,,m-..-w-- - - . . , - . , - . , _ , . .-e.----,.,,,,.m.----mm.m_.,.v,.n..- __.,% .m.-.,-,.,--,,,,-..a ...-,..mw---

W of excessive rapid thermal cycling-induced fatigue

_ damage by use of a leakage monitoring system of this type.

The extent of low frequency thermal cycling is

~~

determined by the degree to which the feedwater flow can be accurately controlled at extremely low flow rates (e.g., 1-10%). Reviews of the flow controller equipment at several operating U.S. plants have shown a wide disparity in the acceptability of feedwater low flow

- controller equipment. Some plants require no modifications whatsoever, while others require extensive

~

equipment and operational changes to minimize the Concerns above.

s l

l _.

XTV-02-008 2.4 Revision 0 nutggb l __

x; a

3.0 COMPONENT DESCRIPTION 3.1 Geometry Figure 3-1 presents the detailed nozzle / safe-end/ thermal sleeve configuration which was analyzed. The nozzle is SA-508 Class 2, low alloy steel, and the safe end is SA-105, Grade II carbon steel. A triple thermal sleeve is inserted in the nozzle.

3.2 Material Properties Minimum required ASME Section III material properties for the above materials are assumed to apply. In addition, due to the high frequency nature of the rapid cycling loading condition, the ASME Section III design fatigue curve for carbon and low alloy steels was

__ extended (Figure 3-2).

3.3 Functional Description The function of the feedwater nozzle /sparger system in BWRs is to introduce and distribute relatively cold feedwater into the reactor. The nozzle penetrates the reactor pressure vessel just above the active core level. The sparger serves the dual purpose of distrib-XTV-02-008 5.1 Revision 0 nutagh

uting the feedwater uniformly around the periphery of the reactor and mitigating the ef fects of thermal stresses in the nozzle caused by the introduction of the cold water. The Browns Ferry nozzle /sparger design addressed in this report incorporates a seal arrangement which is used to minimize the potential for bypass leakage of the cold feedwater around the thermal sleeve.

3.4 Loadino Conditions

- 3.4.1 Design Loads The design pressure and temperature for the nozzle are 1250 psig and 575'F. Normal operational pipe reaction loads on the nozzle, including seismic loads are given in Reference 1.

i 3.4.2 Operating Conditions The normal operating pressure and temperature for the

! nozzle are approximately 1000 psig and 546*F,

' ~

respectively.

l XTV-02-008 3.2 Revision 0 nutgrJ)

3.5 Analysis Criteria The criteria used in this report for stress analysis of the Browns Ferry feedwater nozzles are the stress limitations of the ASME Boiler and Press'are Vessel Code,

~

Section III, Article NB-3000 (Reference 3).

An estimated design life was determined using informa-tion describing system operational conditions and thermal cycling provided by TVA (Reference 4). In addition, f atigue usage was also determined on a per operational condition basis for various assumed seal

~

leakage rates. This permits TVA to monitor the actual system operational conditions and thermal sleeve seal leakage rates as described in Section 4.0 of this report. The governing criterion for establishing B

fatigue design life is that the ASME Code allowable fatigue usage factor of 1.0 may not be exceeded during l the design lifetime.

h M

4WW XTV-02-008 3.3 l

Revision 0 I

nutaq,h

i t

A 3

s I

y T

x F

R E

I T

AE WL DZ EZ EO 7

v-l FN i

N O

I T

A R

l U

E G V I E F E N L O i

S C L E A V

._ M E R E

_ t E

i 1 L i - S W -

T 3 L

O L e A F ' r M I

u R

_ S g E S i I I

A F T P i /

I S

E u Y B

E L

Z Z

O L N I

P U

' R O E C T O A M W R D

_ I E E H E T ef F I

D N

I E

E F

A S

_ W

_ \ O I

g L F

I N? AS 51" 8 o

e I

t

6 10 Sa (psi) N,11eq 550,000 1 10 200,000 2 10 86,000 3 10 39,000 4

~

10 20,000 5 10 13,000 6 10 g 10,000 10 7

$. 105: 8,000 10 8

7,000 9 en 10 vi E

=

c

- Z -

Q E

- =

$ 10 :

6 7 10 12 10 1 0 2 3 4 5 8 9 10 10 10 10 10 10 10 10 10 10 10 10 ALLCWABLE NUMBER OF CYCLES (Nallow}

Figure 3-2 ASME SECTION III DESIGN FATIGUE CURVE EXTENDED BEYOND 10 6 CYCLES (CARBON / LOW ALLOY STEEL) 3.5 nutg.q. b

-.- g-- - ,.-, ,_ , . -..,--v---w-y,. , , - - . - - - . - , -

4.0 LEAKAGE RATE CORRELATION 4.1 Introduction

~.

The rate of accumulation of fatigue usage at the

~

feedwater nozzles is related to the rate of feedwater leakage past the thermal sleeve seals. Leakage will also affect local temperatures, since the feedwater temperature is much lower than reactor water tempera-ture. If there is no seal leakage, there will be some

_ dif ference between top and bottom temperatures within the thermal sleeve annulus, due to natural thermal stratification. When the seals are leaking, the difference will be appreciably greater because of the

~

difference in density between hot reactor water and much cooler feedwater (leakage) water. That is, because of the density gradient, the cooler leakage water will tend l

_ to collect at the bottom of the annulus.

To determine the actual seal leakage rate through the feedwater thermal sleeve seals, TVA is installing a leakage monitoring system (LMS) on the three Browns Ferry units. By comparing the LMS data with analytical predictions, an assessment of actual seal leakage and its effects on the fesdwater nozzles can be made.

XTV-02-008 4.1 Revision 0 nutggh

4.2 Analytic Procedures Generic finite element models of feedwater nozzle configurations similar to Browns Ferry's have been developed using the ANSYS computer code. Analyses based upon these models were used to generate information relating temperature predictions (for leaking and non-leaking ' cases) to axial distance from the thermal sleeve seals (Figure 4-1).

The outside surface temperatures calculated by the analysis should correspond to thermocouple-measured temperatures at the thermocouple locations. The analytically predicted temperatures were normalized as follows:

TN"T Reactor - I4~1I FW where:

TN = Normalized temperature T gy = Feedwater temperature TReactor = Reactor temperature l T Nodal = Normalized Analytical Prediction.

XTV-02-008 4.2 Revision 0 nutggb l

l e - - - - . . , - - _ _ . , _ _ , _ - . . _, . _ _ _ . . , , _ _ _ _ _ _ _ _ _ _ _ . _ _ , , _ . _ _ _ _ . . _ _

4 I

For these calculations, the values Tyg = 300*F TReactor = 550*F

~

were used for comparison and numerical convenience.

Although these values do not exactly correspond to actual operating conditions, the specific values used will have no actual effect on TN as long as the same values for Tpg and TReactor are used in each calculation.

The calculated values of TN are plotted against axial distance from the seal in Figure 4-1. These curves represent the analytical predictions for normalized external temperatures for each combination of azimuthal location (top, isttom) and leakage (0.0, 1.5 GPM).

4.3 Use of Field Leakace Monitorina Data TVA will be taking data with the Leakage Monitoring System following installation. This data will be used 4

by NUTECH to calibrate the system for leakage measurement.

XTV-02-008 4.3 Revision 0 nutagh m ,- - - _ , _ _ y-m_ y,_ , w___ g ,_,.... _ _ ,r ,-__m . . _ - ,_ _, _,--,y 5

From review of the analytical curves of Figure 4-1 it is

_ apparent that the temperature measured on the bottom of the nozzles is far more sensitive to leakage than is the top temperature. That is, the distance between the 0.0 GPM leakage and 1.5 GPM leakage curves for the nozzle

~

bottom is much greater than for the nozzle top. As a result, the bottom temperature is a more responsive

~

indication of leakage.

, To determine the actual value of leakage rate, the band

_ of field data for each nozzle is compared with the analytical predictions. This.is done by indicating the normalized top and bottom temperature bands on a curve such as Figure 4-1 at the axial location of the thermo-

~

couples. The mean bottom normalized temperature is interpreted approximately as seal bypass leakage rate by interpolation between the zero and 1.5 GPM curves as follows:

T o -T yg l

_ Leakage (GPM) = x 1.5 GPM (4-2) j To - T 1.5 where:

To = normalizcd bottom zero leakage temperature.

Tggs = mean normalized bottom temperature as measured by Leakage Monitoring System.

~

XTV-02-008 4.4 Revision 0 nutagh

. . _ = - . _.

W T1.5 = normalized bottom 1.5 GPM leakage tempera-ture.

- By the method described above, estimates of present seal leakage for each Browns Ferry feedwater nozzle may be

~

calculated.

~

j i

~

XTV-02-008 4.5 Revision 0 nutagh g--ow. .-- -- --, m.-r-W > -

w ---- r -- v~-

)Fl3h4NMyja g

.WiQ~ i s

%:s* a.

'~

gem. vsp w w as%

as4 y

! 3 q'.M;{P$k 3 l . $ $ (.' h$N,$d%fk%4I? I.Nh y -3

'[y?3TEGREAA&*hkkd - hnny :l-y4 %rn;x.m.gr mss %q~va6pq m I. -

.' !IfyNN i  ? l s-' -

AwA [

kAM- ,; ,  :~. --khMfr {p;,m.gg%e3M333)$SdM h

< 3- W%%w&rgh .m wgp)$d % p g &> <

_. w ;nQ:2 M*QQawy{j 4 . ., g j

% dkW -:  %

c.1 i$ . . 44'...:. ,;w&.p>+x

g ."yg. :,qs gp>4p:

.f:se.: Csh. 4 EhdE5 .s h

a[.&, N$$$$$$[:.$hh!h Wsfag fMd3ddW -:

Uk "

A

. . . . . Mr. 4sMf'iYgskhEkOd@M%;# Y <

ih '&QWWW);hp-fh .:

gi@-> r4 aAs.;.y;g3es-g;<%.g.dh,g::s.,:SQ- p o - gg p

'gg.

• t Q $f pali $g?hh$$4 d - M EA

%y% _ <

Q. ;fj?G+};fQ.Q j _, S

">yt

'k -

J

'%.,%t?.6M;n6-sf:.;. :y^WL fW:tMp 4 Q

fuggyc g

%. .q-

' qu4--gM - :: q:d4M,g6 N S ,w.y u .A.n. >j;si qvg . q,g

..y g c w w; s.ax

. M m

~-

yn n: 3

%nW?nL<*% < a.::d f g K hy ;je.>&;;i^

r .

g i "Qgyg g $ R@p%9W

~

Mp g v

.'h> .zw; pw g N$,sh:.

wm k-

$hm e$$s$f[mo.. h .:

l p

8 3

o z

,.' '.  % x w <f.wsr CD

-~W$4f8 g g

,7 m.

.e m mv4 %wg M}*f n+a w:

g c.

M a:

%w?; w>9 ,

g H

w O g

d *e

h na%rwu .>.y'w -. g

%7 ya~g:. l n.y:M - u)

vs% 3 2 9?d32 o mn 9p m

e

., u

94. a T .:- $

m

. . . g

. 4

o M

e

>=

. a z

. s.

e e

e e

e e .'

sauvute4A assmwuca XTV-02-008 Revision 0 4.6 nutgrJ)

5.0 FEEDWATER NOZ2LE STRUCTURAL INTEGRITY ASSESSMENT 5.1 Introduction Crack initiation in BWR feedwater nozzles has been demonstrated (Reference 2) to be primarily a f atigue process. This section describes the evaluation of the fatigue ef fects on the feedwater nozzle due to rapid thermal cycling, system cycling, and unsteady flow cycling. This evaluation is based upon a review of Browns Ferry operational data (Reference 4) and the revised stress reports which describe the new thermal sleeve configuration (Reference 5).

5.2 Plant Specific Thermal Duty Map A plant specific Thermal Duty Map (TDM) for Browns Ferry was developed based upon a review of plant operating records contained in Reference 4. Browns Ferry Unit 1

__ Cycle 4 was selected for study as a particularly severe cycle, based upon recommendations of plant pe rsonnel .

This map describes the average hours per year spent in various operating regions, defined by power level and corresponding reactor and feedwater temperatures. The data were divided into 27 regions. Table 5-1 prest ts the resulting Plant Specific Feedwater Duty Map.

XTV-02-008 5.1 Revision 0 ,

~

nutggb

, - - - , w , - - , . - - - , , - , - , . . , ,

- ,.g , ,.w~ , ng. - ,

s Regions 1 through 19 were developed based solely upon

,, Browns Ferry data. Regions 20 through 26 have been added to Regions 1 through 19 in accordance with the General Electric Generic Feedwater Duty Map (Reference

5) to account for periods of reactor power maneuver-

.ing. Region 27 represents periods of cold shutdown.

5.3 Rapid Temperature Cycling 5.3.1 Description of Rapid Temperature Cycling Phenomena Rapid temperature cycling (on the order of 0.1 Hz to 1.0 Hz) occurs at the nozzle inside surface due to mixing of hot reactor water and cold feedwater. The most dominant

~

cause of this cycling is bypass leakage of feedwater past the thermal sleeve seals. However, rapid cycling can also occur in the absence of such leakage due to the mixing of hot reactor water and a cold water boundary layer which builds up on the outside surf ace of the

__ thermal sleeve.

The magnitude of the nozzle surface thermal cycling which occurs due to this rapid thermal mixing phenomenon

~~

is given by the following expression, which is derived from experimental data presented in Reference 2.

XTV-02-008 5.2 Refision 0 nutggb

ATp _p = X X C 3 XC 4X (T R -TFW) (5-1) where ATp_p = Metal surf ace peak. to peak temperature range.

K = Amplitude coefficient for a given frequency of cycling, from Table 5-2.

C3

= Coefficient from Table 5-3.

C4

= Leakage coefficient from Table 5-4.

TR

= Reactor water temperature.

Twp = Feedwater temperature.

In order to determine the number of cycles at various peak-to-peak temperature ranges from equation 5-1, it is also necessary to know the amount of time which is spent at various feedwater temperatures, reactor temperatures and feedwater flow rates. This information is contained in the Browns Ferry thermal duty map described above.

5.3.2 Fatigue Usage Calculation l A computer program (DAMSUM) has been developed to

' ~'

evaluate the rapid cycle fatigue usage factors for BWR feedwater nozzles.

i __

XTV-02-008 5.3 Revision 0 nutgrb

c. >-- -- + + - - . . - . - - ---.,,--m

A series of constant leakage rates is input to the program.

Using this data the program determines ATp_p (Equation 5-1) for each flow map region using the appropriate feedwater flow rate for the region.

~

Alternating stress is then determined for each value of ATp _p using the following equation:

EsaT 0

a"2 (1 v (5-2) where:

ATp_p = Peak-to-Peak Temperature Range from Equation 5-1 E = Young's Modulus a = Coefficient of Thermal Expansion v = Poisson's Ratio The properties E and a are determined in the program from tabulated ASME Code values at the average tempera-ture about which the cycling occurs.

For incorporation into the DAMSUM program, the fatigue curve was linearized into straight line log-log segments as follows:

XTV-02-008 5.4 Revision 0 nutggb y - - - - . - -

---

e -

g--.-.---___..,_,e . - - . - - . - , , , - , - - . - - - - - - - . -

log 10N = C - B log 10 S a (5-3) where log 10 Nn ~ 109 10 N n+1 8 = 109 10 8n+1 ~1 910 8 n C = loggo Nn+B log 10 S n where N n, N n +1' S n and Sn +1 are values of S a and N at the beginning and end of the log decade surrounding the

-- calculated value of S, (See Table in Figure 3-2).

The allowable number of cycles (Nallow) can thus be determined for each calculated value of ao lt fr m the fatigue curve, and the cumulative fatigue usage factor due to rapid cycling is computed versus time using the following multiple summation:

I J K l

U={i=1 ([j=1 k=1 [ Njk! Allow jk'i ( ~

where:

1 N jk = Applied number of cycles for the 4th amplitude and frequency in Table -2 XTV-02-008 5.5

~

Revision 0

nutggh

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

and for the kth flow map region in

_, Table 5-1.

N allow, jk = Allowable number of cycles for Salt for the jth amplitude and frequency in Table 5-2 and for the kth flow map region in Table 5-1.

( )i = Usage factor for the ith year of reactor operation.

Results of representative rapid thermal cycle-induced fatigue calculations as described above are contained in Tables 5-5, 5-6, and 5-7. Figure 5-1 shows the relationship between rapid cycle fatigue usage and leakage rate predicted for Browns Ferry.

The above fatigue usage calculations assumed a series of constant leakage rates. In reality, leakage can be

_. expected to increase gradually with time due to corrosion of seal and pipe materials.

A series of calculations was performed in which an

~

initial interference fit and corrosion rates were assumed, and leakage was allowed to increase based upon mean gap calculations performed by program DAMSUM.

XTV-02-008 5.6 Revision 0 nutagh

Fatigue usage per year was calculated based upon predicted leakage for that year. Cumulative usage was also calculated.

~

5.3.3 Rapid Cycle Evaluation for Calculated Leakage The rate of accumulation of rapid cycle fatigue usage for increasing leakage rates is chown in Table 5-8.

This information, together with the system cycling values determined in Section 5.5 predicts the current fatigue accumulation situation. If the leakage rate changes with time, the revised fatigue usage can be determined by entering the curve of Figure 5-1 at the

~

calculated leakage rate to determine the associated rapid thermal cycling-induced fatigue usage. This value can be used with the system cycling value as described above.

-- Contribution of each thermal duty region to total fatigue usage f actor is shown in Table 5-9 for a repre-

~

sentative case. For different leakage ratas, the magnitude of the individual contributions will vary, but the relative importance of each region will remain substantially the same, as long as the calculated stresses are above the endurance limit. When stresses XTV-02-008 5.7 Revision 0 nutgLqh

m. *- . , ,. - - - - - ,w ,---

1 .

fall below the endurance limit, no fatigue usage

-- contribution is expected. By inspection of tables of this type, those operating regimes which contribute most strongly to rapid cycle fatigue damage can be identified.

It should be noted that in Figure 5-1 the slope of the fatigue usage accumulation curve is shallow at low leakage rates. A small increase in leakage will produce only a small increase in fatigue usage for low leakage rates. In the vicinity of 1 GPM leakage however, the local fatigue curve is much steeper'. A small leakage increase in this region can produce an appreciable increase in fatigue usage. This comment is presented for use in planning repairs, etc.

Leakage past triple thermal sleeve-type seals has

- historically been very low. Leakage is expected to increase gradually with time (over a period of years)

~

due to seal degradation and corrosion. In Figure 5-1, i

it may be seen that below 0.5 GPM leakage, the 40 year fatigue usage due to rapid thermal cycling is minimal.

Up to about 1 GPM, the fatigue usage is still very low.

XTV-02-008 5.8

_ Revision 0 nutggb

By continually monitoring leakage using Browns Ferry's monitoring equipment, the necessity of repairs to thermal sleeve seals can be anticipated well in

_ advance. Properly timed repairs should prevent accumulated usage from exceeding allowable values.

Since thermal sleeve seals degrade only gradually, sudden changes in data from the leakage monitoring system (e.g., large step increases indicated leakage) are indicative of monitoring system component problems such as thermocouple detachment. A thorough LMS component check should be made at the next convenient time.

Unstable Flow / Low Flow Evaluation

~

5.4 TVA provided data (Reference 6) which describes varia-tion of feedwater temperature within the horizontal portion of the pipe upstream of the feedwater nozzle.

_ Resistance Temperature Detectors (RTDs) located at 12, 3, 6, and 9 o' clock azimuthally were installed to monitor the thermal effects of feedwater low flow l operation.

The data, which were taken at 1 minute intervals, l

suggest that the feedwater flow at Browns Ferry is very XTV-02-008 5.9

, Revision 0 l

nutggh

accurately controllable. During typical start-ups and

__ scrams, few thermal transients were observed (4-10 generally). Transients had a period of 1-2 hours.

Thermal cycling of this character is very mild compared

~

to most of the cases previously analyzed. A comparison of the crack growth cases considered in Reference 2 with the cycling experienced at Browns Ferry shows that a 1/4 inch crack would require roughly 20 years to propagate to a 1 inch depth under the impetus of such cycling.

This is a worst case estimate.

! A review of plant piping and operational procedures revealed that the feedwater system itself is rarely used when extremely low water flow rates are required to maintain vessel water level. Instead, the Control Rod Drive Return line, which is tied into one feedwater loop and which has a capacity of 100 GPM, is used to supply vessel make up flow. Since the flow can be maintained steadily, thermal transients due to intermittent opera-tion of the feedwater pumps are minimized. Feedwat'er

~~

flow is stabilized by pumping to the condenser until the greater capacity of the feedwater system is required.

i Consequently, the crack growth concerns related to cyclic low flow feedwater system operation are XTV-02-006 5.10

~

Revision 0 l nutaq,b m m- s- - - - - -r--- e. w- - - - - -- -e. - mw- - - - ey-- ne y - ---%,- ,.--g--yi---9.---9--mm,y9 y y, - - - - - - . y,---, z.-- ------,-.wrww-w-m--- --

-.--av---

.. y

, s.

. g

~ #

s s a

5 i ;,~

3

,g,,.y N.s.

,' z minimized. The preserat method of system operahion is'

=' '

', \ ,;  ? , ,

'a considered adequate. Modifica'.; ions to feedwa .er low r . = .

.. fl,ow control equipment solely for the purpose ,of.

- mitigating low ' flow / unsteady, flow . induced thermal cycling is not deemeil to be necee,sary;, ,

? ,

s

=

q, ., <

9 w 5.5 System Cyclinc -

System cycling'inclu. des the ef fects of major operational ,

transients on the reactor pressure boundary. 5,Such s.

., s ,,

transients include. the pressure and thermal transient ,,

2 '

portions of start-ups, shutdowns, scrams lli.urbine , rolls, *

, 2

~

etc. Fatigue due to cystem cycling contributes, "'

to both

• _p crack initiation and crack growth. fonsequently, a fatigue usage factor due to system cy' cling must be added \

~

\ ', , .

to that due to rapid therme.1 ~*

cycling to determine ,-

the,

__ u tdtal fatigue usage factor.

1 .,

~

s. , , ,

~ .,,.

Values for' the f atigue usage facter due to system -

,; c

- cycling Were taken from Reference 5 for the Browns Ferry feedwater nozzles. The' system cycle fatigue usage factors for, 40 years are approximately 0.40 bet.Jeen the

, thermal sleeve : teals,g and 0 [3 at the nozzle blend

~ ~

radius, based upon the referenced Stress Reports. TVA maintains an on going record of System Cycling base. i upon transients actually experienced at each unit. ,'

• s.

XTV-02-008 5.11 -

s Revision 0 ..

J.s nute_cb s

, r o',

i

[ . i

_ .D

~

5.6 Total Fatique Evaluation The system cycling fatigue values were added to the

~

values determined for rapid thermal cycling to generate a total usage factor. These factors are compiled in Table 5-10 for various leakage rates.

5.7 Reactor Water Clean Up Cross-tie Evaluation NUREG 0619 (Reference 7) proposes several methods for reducing the fatigue usage of the feedwater nozzles.

One of these recommendations is the installation of a cross-tie line to allow Reactor Water Clean Up (RWCU) return flow to be routed to both feedwater loops, (RWCU flow presently returns via loop B only except on Unit 3). The mixing of high temperature RWCU water with lower temperature feedwater would tend to mitigate the

-. thermal fatigue effects of feedwater flow on the feedwater nozzles by decreasing the temperature l

~~

l difference between reactor and feedwater. To evaluate the effectiveness of such a modification, the thermal duty map developed above was revised to include the effects of RWCU flow. The following characteristics of the RWCU system were assumed:

[

l i

XTV-02-008 5.12 Revision 0 nutggb l

1. RWCU return temperature at normal operating reactor temperature (539'F) = 472*F.

2. RWCU flow (1004) = 300 GPM These numbers were taken from discussions with TVA personnel, and from Browns Ferry Drawing 47W810 (Reference 8).

It should be noted that since RWCU return heating is accomplished by a regenerative /non-regenerative self heat exchanging system (i.e., there is no source of heating other than RWCU inlet flow), the achievable RWCU return temperature is governed by actual reactor temper-

~

ature. In other words, a drop in reactor temperature must necessarily produce a comparable drop in achievable RWCU temperature. Furthermore, the clean up system at Browns Ferry is designed as a high pressure system.

This implies that as reactor pressure drops, maximum

-- achievable RWCU flow will also drop. The assumption that the above temperature and flow are available throughout the thermal duty map produces results which are an upper bound on the improvement in fatigue usage

~

which may be expected through the use of RWCU. A lo<er limit on expected improvement can be obtained by s

assuming that RWCU is only available when the reactor is XTV-02-008 5.13 Revision 0 nutagh

,y- ,.- .c -. ,,-----,e-- , . - . ,,w e- y-y- - - -- - -

~

at operating temperature (539'F). Analysis of these two

_ cases provides an indication of the range of f atigue usage improvemant which may be expected.

Comparison of the results of rapid cycle evaluation with

~

and without RWCU shows that:

l. Beyond the second thermal sleeve seal an improve-ment in fatigue usage of less than 1% would result from incorporation of the mitigating ef fect of RWCU.

2. Between the seals, and improvement of about 6%

could be expected.

To corroborate these analytical conclusions, Browns Ferry directly monitored the temperature ef fects of operation with and without RWCU cross tie on Unit 3, using the system described in Section 5.4. This data

_ showed no detectable difference in temperature measured upstream of the seals (Reference 6).

These improvements are insignificant. Consequently, the

~

installation of a RWCU cross-tie modification at Browns XTV-02-008 5.14 Revision 0 nutg,qh

- - - , - - ~ , . . . , -- -- e - +e -,n.- - -

Ferry Units 1, 2, and 3 is not considered to be justified unless additional operational flexibility is gained thereby.

5.8 In-Service Inspection The analysis presented herein shows that the total fatigue usage f actor remains substantially below the ASME allowable value of 1.0 during the lifetime of the

~.

Browns Ferry Units, unless secondary thermal sleeve seal

__ leakage exceeds 1.0 GPM. With continual leakage monitoring, the potential for crack initiation can be accurately assessed without inspection.

If the thermal sleeve seals are shown to allow only minimal leakage, NUREG-0619 suggests that the NRC may be receptive to delaying required inspections on a case-by-case basis. The justification for such delays would be that by continually monitoring leakage and evaluating fatigue usage resulting from the measured leakages, repairs to thermal sleeve seals could be rianned to prevent the fatigue usage from becoming excessive. If l leakage remains los, minimal fatigue usage may be 1

~

expected. If leakage increased with time, repair can be planned suf ficiently in advance to prevent signific it

~

XTV-02-008 5.15 Revision 0

4 rapid thermal cycle fatigue usage accumulation. Note i I

_ that leakage is not expected to increase suddenly, but rather gradually with time.

t

~

i 1

l ,

I i

1 l

l I

I l

1 -

t -

I

, XTV-02-008 5.16 Revision 0 l

m--- , - . , , , , < - , - _.---. , _ ,, ,,..ww_mm._,,

i-7.- ,..-,_,_,,_,,,,_,m.m,-.,.,y.r,,_,..,_-_m,,,.r,..w.w,._, _ , . , . _ , ,,..n-,w, _ . . -n,e--.r..m.w- ..nc,w,-.m.,

Table 5-1

~

BROWNS PERRY PLANT SPECIFIC THERMAL DUTY MAP Region  % Power  !!ours/ year Rx Temp FW Temp 1 100.00 5528.00 539.00 372.00 2 95.00 1195.00 539.00 360.00 3 80.00 373.00 539.00 350.00 4 70.00 302.00 539.00 337.00 5 50.00 94.00 539.00 315.00 6 50.00 3.00 539.00 285.00 7 40.00 42.00 539.00 305.00 8 40.00 14.00 539.00 240.00

~

9 32.00 14.00 539.00 185.00 10 23.00 12.00 539.00 175.00 11 23.00 8.00 539.00 120.00 12 15.00 16.00 539.00 120.00 13 5.00 64.00 530.00 120.00 14 5.00 14.00 470.00 120.00 15 5.00 10.00 360.00 120.00 16 5.00 40.00 240.00 120.00 17 5.00 48.00 512.00 210.00 18 5.00 29.00 298.00 210.00 19 5.00 13.00 536.00 308.00 20 0.00 43.00 340.00 300.00 21 1.00 .40 360.00 350.00 22 2.00 1.78 350.00 190.00 23 2.00 1.38 340.00 125.00 24 2.00 .25 330.00 70.00

_ 25 2.00 1.60 400.00 190.00 26 3.00 .38 340.00 160.C^

27 0.00 900.00 70.00 70. >

emp XTV-02-008 5.17 Revision 0 nutgrJ) m - - ---

Table 5-2 AMPLITUDE / FREQUENCY DATA FOR RAPID CYCLING

_ BETWEEN SEALS DOWNSTREAM OF @EALS Index Amplitude Frequency Amplitude Frequency I

K Cycles /Hr K Cycles /Hr 1 1.00 45 1.00 15 2 0.95 35 0.98 15 3 0.90 20 0.955 15 4 0.85 120 0.91 30

_ 5 0.77 100 0.84 75 6 0.66 105 0.75 120 7 0.56 100 0.65 150

~

8 0.46 100 0.55 180 9 0.36 100 0.45 450 10 0.26 1200 0.35 1200 11 0.15 7500 0.20 7500 XTV-02-008 5.18 Revision 0 nutggh

, - , . - , - - , , . - - - r-,--- ..w.-- - - - - - - - . .

l f

i .

l Table 5-3 COEFFICIENT C 3 1  ;

. t i

100% Rated 20% Rated 0% Rated Pt. Feedwater Flow Feedwater Flow Feedwater Flow Nozzle

- Bore 1.0 1.0 1.0 s

i .

I l

l -

l

~

l I

l l

l i XTV-02-008 5.19 Revision 0 i

f

4 Table 5-4

' l COEFFICIENT C 4 Thermal Sleeve O GPM .35 GPM 0.5 GPM 1.5 GPM Bypass Leakage Coefficient 0.1 0.1 0.24 0.3 C4 i If 3

i -

l l ,

1-l l

p 1

(

l XTV-02-008 5.20 '

Revision 0 1

f

- . , - , - , _ . -e,,,, ,-.-----,,,,~,,--mw,,,,.--._,,,,w.,,, .-n.,._,,--n.,, ..-,,_,n-.,,---,,n, , ,., ,_ -_..,,.,,, ,--

Table 5-5 RAPID CYCLE FATIGUE USAGE (Nozzle Blend Radius):

1.0 GPM LEAKAGE Cumulative Fatigue Usage

_ Year Factor Leakage 1 .003 1.000 2 .005 1.000

_ 3 .008 1.000 4 .011 1.000 5 .014 1.000 6 .016 1.000 7 .019 1.000 8 .022 1.000 9 .024 1.000 10 .027 1.000 11 .030 1.000 12 .033 1.000 13 .035 1.000 14 .038 1.000 15 .041 1.000 16 .044 1.000 17 .046 1.000 18 .049 1.000 19 .052 1.000 20 .054 1.000

_. 21 .057 1.000 22 .060 1.000 23 .063 1.000 24 .065 1.000 25 .068 1.000 26 .071 1.000 27 .073 1.000

_, 28 .076 1.000 29 .079 1.000 30 .082 1.000 31 .084 1.000 32 .087 1.000 33 .090 1.000 34 .092 1.000

~

35 .095 1.000 36 .098 1.000 37 .101 1.000 38 .103 1.000 39 .106 1.000 40 .109 1.000 XTV-02-008 5.21 Revision 0 nutggb

Table 5-6 RAPID CYCLE FATIGUE USAGE (Nozzle Blend Radius):

1.25 GPM LEAKAGE Cumulative Fatigue Usage

_ Year Factor Leakage 1 .007 1.250 2 .013 1.250

_ 3 .020 1.250 4 .026 1.250 5 .033 1.250 6 .039 1.250 7 .046 1.250 8 .052 1.250 9 .059 1.250 10 .066 1.250 11 .072 1.250 12 .079 1.250 13 .085 1.250 14 .092 1.250 15 .098 1.250 16 .105 1.250 17 .111 1.250 18 .118 1.250 19 .125 1.250 20 .131 1.250

_ 21 .138 1.250 22 .144 1.250 23 .151 1.250 24 .157 1.250 25

.164 1.250 26 .170 1.250 27 .177 1.250

_ 28 .184 1.250 29 .190 1.250 30 .197 1.250 31 .203 1.250 32 .210 1.250 33 .216 1.250 34 .223 1.250 35 .229 1.250 36 .236 1.250 37 .243 1.250 38 .249 1.250 39 .256 1.250 40 .262 1.250 XTV-02-008 5.22 Revision 0 nutggb

. . _ _ _ . . - - . . , . - , _ . , _ , , - , --_.._w-, ,- . . , , . m _ _ _ _ . ,_ - . , , , , , _ . _ - . , . _ f

-- _ _ _y.,

Table 5-7 RAPID CYCLE FATIGUE USAGE (Nozzle Blend Radius):

1.5 GPM LEAKAGE Cumulative Fatigue Usage

_ Year Factor Leakage 1 .014 1.500 2 .027 1.500 3 .041 1.500 4 .054 1.500 5 .068 1.500 6 .092 1.500 7 .095 1.500 8 .109 1.500 9 .122 1.500 10 .136 1.500 11 .149 1.500 12 .163 1.500 13 .177 1.500 14 .190 1.500 15 .204 1.500 16 .217 1.500 17 .231 1.500 18 .245 1.500 19 .258 1.500 20 .272 1.500 21 .285 1.500 22 .299 1.500 23 .312 1.500 24 .326 1.500 25 .340 1.500 26 .353 1.500 27 .367 1.500

_. 28 .380 1.500 29 .394 1.500 30 .408 1.500 31 .421 1.500 32 .435 1.500 33 .448 1.500 34 .462 1.500

_ '35 .476 1.500 36 .489 1.500 37 .503 1.500 38 .516 1.500 39 .530 1.500 40 .543 1.500 XTV-02-008 5.23 Revision 0 nutagh

- - , - ,----e e-, _ _ . , - -, , - - - -- - - - --, -e - - - - - - ,--m-- ,-,-,-rv ~ - - , - - -

- - ~

Table 5-8 RAPID CYCLE FATIGUE USAGE (Nozzle Blend Radius):

LEAKAGE INCREASING WITH TIME Cumulative Fatigue Usage  !

__ Year FIT Factor Leakaae 1 .0046 .000 0.000 2 .0037 .000 0.000

_ 3 .0029 .000 0.000 4 .0021 .000 .035 5 .0014 .000 .103 6 .0008 .000 .171 7 .0001 .000 .239 8 .0006 .000 .307 9 .0013 .000 .375

^

10 .0020 .000 .465

11 .0026 .000 .567 12 .0033 .001 .669 13 .0040 .002 .771 14 .0047 .003 .873 15 .0054 .005 .975 16 .0060 .009 1.051 17 .0067 .013 1.119 18 .0074 .018 1.187 19 .0081 .025 1.283 20 .0088 .034 1.385

__ 21 .0094 .047 1.487 22 .0101 .061 1.500 23 .0108 .074 1.500 24 .0115 .088 1.500

.0122 25 .102 1.500 26 .0128 .115 1.500 27 .0135 .129 1.500

__ 28 .0142 .142 1.500 29 .0149 .156 1.500 30 .0156 .170 1.500 31 .0162 .183 1.500 32 .0169 .197 1.500 i 33 -

.0176 .210 1.500 34 -

.0183 .224 1.500

_ 35 -

.0190 .237 1.500 36 -

.0196 .251 1.500 37 -

.0203 .265 1.500 38 -

.0210 .278 1.500 39 -

.0217 .292 1.500 40 -

.0224 .305 1.500 XTV-02-008 5.24 Revision 0 nutagh

,.c -- - - - - -

Table 5-9

~~

RAPID CYCLE FATIGUE USAGE CONTRIBUTION BY MAP REGION LEAKAGE = 1.5 GPM Fatigue Cumulative Usage Per Fatigue Usage

.. MAP AT , Em Sa Region Factor 1 50.100 239.76 8.58 .0059 .0059 2 52.895 239.58 9.05 .0024 .0083 3 53.298 239.56 9.12 .0008 .0091 4 55.146 239.44 9.43 .0010 .0101 5 57.120 239.32 9.76 .0005 .0106 6 64.770 238.83 11.05 .0001 .0106 7 57.564 239.29 9.84 .0002 .0108 8 73.554 238.28 12.52 .0010 .0118 9 78.352 237.98 13.32 .0018 .0136

~~

10 36.400 240.62 6.26 0.0000 .0136 11 41.900 240.27 7.19 .0000 .0136

~

12 41.900 240.27 7.19 .0000 .0136 13 41.000 239.20 7.01 .0000 .0136 14 35.000 232.02 5.80 0.0000 .0136 15 24.000 218.85 3.75 0.0000 .0136 16 12.000 204.48 1.75 0.0000 .0136 30.200 17 237.61 5.13 0.0000 .0136 18 8.800 211.99 1.33 0.0000 .0136

__ 19 22.800 241.10 3.93 0.0000 .0136 i

l 20 4.000 217.59 .62 0.0000 .0136 l

i 21 1.000 220.30 .16 0.0000 .0136 22 16.000 218.09 2.49 0.0000 .0136 23 21.500 216.49 3.32 0.0000 .0136 24 26.000 214.94 3.99 0.0000 .0136 25 21.000 224.80 3.36 0.0000 .013' 26 18.000 216.71 2.79 0.0000 .013u

! 27 0.000 183.82 0.00 0.0000 .0136 l

XTV-02-008 5.25 Revision 0 nutagh I

_ __ _ - _ = ._ _

o Table 5-10 i

TOTAL FATIGUE USAGE - 40 YEARS (ASSUMING CONTINUOUSLY INCREASING LEAKAGE) 40 Year

• 40 Year System Cycling Rapid Cycling Fatigue Usage Fatigue Usage Unit Location Factor Factor Total

~

l Nozzle .2575 .3056 .5631 radius

-- 2-3 Nozzle .2205 .3056 .5261 radius w

W

• From Reference 5.

l l

t 1 -.

um XTV-02-008 5.26 Revision 0 i

nutagh

---+.,,----,-,,-c m. -. - . , , -. -. ,------ ,-.-. .---- -- -_ ,.me.- ,--.--,y .-, .,- - - .

0. 6 -

~

0. 5 -

'~~

E o

o g 0. 4 -

8 4

en

\'

D 63 g 0. 3 -

C f

5 o .

_ g 0. 2 -

Q E

f 0.1 -

0'O . s s 0.0 0.25 0.50 0.75 1.00 1.25 1.50 LEAKAGE (gpm)

Figure 5-1 FATIGUE USAGE FACTOR VS LEAKAGE RATE

~

XTV-02-008 Revision 0 5.27 nutggh l

l

6.0 CONCLUSION

S 6.1 Results The NUTECH evaluation of the Browns Ferry feedwater nozzles has produced the following conclusions:

1. If leakage past the secondary seals is low, fatigue-induced cracks of the feedwater nozzle bore and blend radius are not expected.

2. By continually monitoring seal leakage, increasing f atigue usage can be anticipated suf ficiently in advance to prevent crack initiation. Seal repairs could be planned as necessary. Such a program may be acceptable to the Nuclear Regulatory Commission as an alternate to in-vessel dye penetrant inspection of the feedwater nozzles. (Reference 7).

3. Browns Ferry's present method of operation at low feedwater flows does not significantly contribute i

to fatigue usage or to growth of existing cracks.

System flows at very low power levels appear to be accurately controllable with existing equipment.

Therefore, it does not appear to be necessary to

~

upgrade the present system if the concerns of NUREG-0619 are the only reasons for doing so.

XTV-02-008 6.1

- Revision 0

s

4. The potential improvement to nozzle fatigue service life which would result from a proposed Reactor

_ Water Clean Up System cross-tie modification is considered to be insignificant.

6.2 Recommendations NUTECH offers the following recommendations for further actions by TVA.

l. Leakage monitoring using the new system should be

- implemented. Data should be taken waekly to detect leakage increases with time. This would be indicative of seal degradation due to corrosion.

2. If leakage rates appear to be increasing with time, l repairs to the thermal sleeve seals should be l

~

planned suf ficiently in advance to prevent total

_ fatigue usage f actor from exceeding 1.0 during the remaining plant life.

W l

XTV-02-008 6.2 l

Revision 0

~

i nutagh

l. _ . ,. . - . - - - - - - - - - . - - - - -

i .

7.0 REFERENCES

1. Browns Ferry operational information. Transmitted by letter from C. R. Favreau (TVA) to H. L. Gustin (NUTECH), dated March 19, 1982.

2. Watanabe, H., Boil'.nq Water Reactor Feedwater Nozzle /Sparcer Final Report, NEDE-21821. General Electric Company, March, 1978.

t

3. American Society of Mechanical Engineers, Boiler and Pressure vessel Code,Section III, 1980 Edition.

4. Browns Ferry system operational data and computer logs for Browns Ferry Unit 1, Cycle 4 provided by TVA personnel to H. L. Gustin (NUTECH) during site visit March 15-17, 1982. NUTECH File Nos.

165.1202.0012, and 165.1202.0013.

5. General Electric Company feedwater nozzle stress

~

reports and analyses.

a. General Electric Stress Report 22A5594,

~

"Feedwater Nozzle," Browns Ferry 1, 2 and 3, Revision 2, June 1, 1979.

XTV-02-008 7.1 Revision 0 nutagh

= _. _ _

i w

b. General Electric Design Specification (Repair)

-- 22A5584, " Reactor Vessel," Revision 1, March I

21, 1978.

c. General Electric Design Certification 22A5S39,

" Reactor Vessel," Revision 1, October 7, 1977.

d. General Electric Design Specification 22A5593,

" Reactor Vessel" Browns Ferry 1, 2 and 3, Revision 2, May 22, 1979.

e. General Electric Stress Report 22A5562,

~

" Reactor Vessel," Revision 0, October 7, 1977.

6. Browns Ferry feedwater low flow control data.

Transmitted by letter from T. F. Ziegler (TVA) to L. C. Hsu (NUTECH), dated June 28, 1982.

7. NUREG-0619,"BWR Feedwater Nozzle and Control Rod

~ Drive Return Line Nozzle Cracking," U.S. Nuclear Regulatory Commission, November 1980.

8. Browns Ferry feedwater and reactor water cleanup systems design drawings. Transmitted by letter

., from T. F. Ziegler (TVA) to H. L. Gustin (NUTE d),

dated May 3, 1982.

XTV-02-008 7.2 Revision 0 nutagh

\ ---- _ .. - -.

+

ENCLOSURE 2 PLAN FOR RESOLUTION OF CRACKING PROBLEM AND CONTINUED MONITORING OF FEEDWATER N0ZZLE PURSUANT TO NUREG-0619 BROWNS FERRY NUCLEAR PLANT

1. Sparger and Thermal-Sleeve Design Modifications (NUREG-0619, Section 4.1)

Nozzle cladding has been removed and GE triple-sleeve spargers have been installed on all Browns Ferry units.

2. Low-flow Controller (NUREG-0619, Section 4.2, and NRC Generic Letter 81-11)

Based on the NUTECH anaysis, installation of a low-flow controller having the characteristics described in section 3.4.4.3 of NEDE-21821-A has been found to be unnecessary. First, the high-cycle fatigue analysis indicated that no cracks will be initiated in the feedwater nozzle blend radius over the 40-year plant life assuming a thermal sleeve leakage of less than 1.5 gpm. Second, even it a 1/4-inch crack existed, it would take 20 years for the crack to grow to one inch (worst case estimate).

TVA intends to take appropriate action when a crack is indicated by UT examination or when thermal sleeve leakage greater than 1.5 spm is measured. The NRC will be notified in either case.

3 Reactor Water Cleanup Reroute (NUREG-0619, section 4.2, and NRC Generic Letter 81-11)

Again based on the NUTECH analysis, rerouting of the reactor water cleanup system at Browns Ferry has been found to be unnecessary. The analysis has concluded that the benefit of such rerouting is insignificant and this fact has been supported by experimental data at Browns Ferry. This modification had previously been done on unit 3

4. Inspections and Leak Detection (NUREG-0619, section 4.3)

TVA will continue to perform the UT and visual inspections at the intervals required by Table 2 (page 18) of NUREG-0619. With regard to the routine PT requirements, TVA proposes an alternative to the NUREG-0619 inspection interval of nine refueling cycles (or 135 startup/ shutdown cycles). NUTECH's worst case crack growth estimate predicts that it will take 20 years for a 1/4-inch crack (undetected) to grow to one inch. Therefore, TVA proposes to perform the PT examinations on each unit after 20 years of operation, starting from when the nozzle cladding was removed. Cladding has been removed from the Browns Ferry units, and startup of the unit from the outage when that was done is as follows:

Unit 1 - January 1978 Unit 2 - June 1978 Unit 3 - December 1979

\

. 1 l

These dates are proposed as the start of the 20-year interval for each unit as discussed above. If the end of the 20-year period occurs during a cycle of operation, the PT examinations will be performed at the following scheduled refueling outage. Performance of the PT examinations on this proposed alternative inspection interval will result in some reduction of cumulative exposure to personnel over that to be incurred from the intervals currently in NUREG-0619 We are hopeful that advances in ultrasonic testing (UT) technology will totally eliminate the need for PT examination of the feedwater nozzles.

However, until approval by NRC of improved UT techniques TVA will perform the PT examinations in accordance with the alternative outlined above.

By letter from H. R. Wisenburg to H. R. Denton dated October 13, 1981, TVA committed to install a leak detection system to monitor thermal-sleeve bypass leakage on each Browns Ferry unit. In accordance with NUREG-0619 (section 4 3 2.4) we will keep the NRC staff informed as to its performance and assessment of leakage measurements. The NUTECH analysis has shown that leakage values less than 1.5 gpm will not result in a 40-year fatigue usage factor greater than 1.0. Therefore, we commit to remedial measures or additional analysis only if the leakage exceeds 1.5 gpm.

A further benefit of the leakage detection system will be to define the time when seal refurbishment becomes necessary. GE seal refurbishment schedules based on estimated corrosion rates no longer apply.

k