ML20137J402

External Circumferential Crack Growth Analysis for B&W- Design Rv Head CR Drive Mechanism Nozzles

ML20137J402
Person / Time
Site:	Three Mile Island
Issue date:	12/31/1993
From:	Yoon K BABCOCK & WILCOX CO., BABCOCK & WILCOX OPERATING PLANTS OWNERS GROUP
To:
Shared Package
ML20007F901	List: ML20137H777 ML20137J402
References
BAW-10190-ADD01, BAW-10190-ADD1, NUDOCS 9704040013
Download: ML20137J402 (18)

Text

-.-- _ _ - - - _ - _ _ . _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

). _ _ . . - - _ _ - .

j BAW-10190P, ADDENDUM i DECEMBER 1993 i

4

' B&W OWNERS GROUP PROPRlETARY #'"- #Y""  % d ?c7/.

.l 1 \

i -

."LOWNERS GROUP

$4 .

) MATERIALS COMMITTEE ,

_ * . -r

- - m 1; ,, , ; , > -_  ;;grn~F~

j

?

. . L, -

..".~-=.-* -

i j External Circumferential Crack l Growth Analysis For B&W l Design Reactor Vessel Head i Control Rod Drive Mechanism Nozzles j

l i

! B&WNUCLEAR TECHNOLOGIES

! 9704040013 970331 PDR ADOCK 05000209 P PDR

(

u . - .- . - - - - - -- . _ _ - _ __- -___--__ __

you-psymekny &

B&W OWNERS GROUP PRO I RY tsc-94-574 BAW-10190P. Addendum 1 December 1993 EXTERNAL CIRCUMFERENTIAL CRACK GROWTH ANALYSIS FOR B&W-DESIGN REACTOR VESSEL HEAD CONTROL R00 DRIVE MECHANISM N0ZZLES 43-10190P-01 l

by i

K.K. Yoon (See Section 7 for Document Signatures) l Prepared for

, Duke Power Company Entergy Operations, Inc.

Florida Power Corporation GPU Nuclear Corporation ,

-Tennessee Valley Authority Toledo Edison Company '

l l

Prepared by B&W Nuclear Technologies Engineering and Plant Services Division P. O. Box 10935 Lynchburg, VA 24506-0935 i

/ /

This docume t is the property f the B&W ners Group.

i Distribution or reproduction o this document organizatio s not in the B&W O ers Group is prp lbited Individuals or  ;

without the written c sent of the B&W O ers Group. /

. /

4 Ap,A<y pu fcc-n -s74

~ EXTERNAL CIRCUMFERENTIAL CRACK GROWTH ANALYSIS FOR B&W-DESIGN REACTOR VESSEL HEAD CONTROL ROD DRIVE MECHANISM N0ZZLES

1. Introduction .i' only one out of approximately 2600 CRDM nozzles inspected to date revealed a leaking through-wall crack. This same nozzle also exhibited an indication of circumferential cracking on its outside surface. It is noted that development of circumferential cracking on the outside surface is only possible after establishment of a leak path for the primary water environment. Primary water  !

can only reach the outside nozzle surface (during operation) when a through-wall.  ;

crack is present. The predicted time, therefore, for e'xternal crack initiation  ;

and growth is in addition to the time for the internal surface crack growth to -

-become a leaking through-wall crack.  !

Unmitigated circumferential crack growth from the outside surface of the nozzle prtentially implies a completely detached upper portion of the nozzle i thereby ieading to the scenario of nozzle ejection. This addendum to the Safety l Evaluation of B&W Design Reactor Ve'ssel CRDM Nozzles provides an evaluation of this concern specifically for the stresses presented in reference 1.

2. External Circumferential Crack Growth Upon examination of,all axial stresses of the hillside CRDM nozzles in the B&W Owners Group plants, the maximum tensile outside stress was found at the top of the weld zone on the downhill side of the nozzle as shown in Figure 1 (Figure 2-8 in reference 1). This is also the location of the maximum stress gradient across the nozzle wall which ranges frbm a high tensile stress on the outside of the wall to a compressive stress on the inside. On the uphill side of the nozzle, there is' a reduced tensile axial stress on the outside surface. The stress gradient at this location is significantly less as can be seen in Figure 2 (Figure 2-9 of reference 1). The average stress around the nozzle is relatively constant at approximately 20 ksi (Eigtre ?). The axial stresses are primarily comprised of secondary stresses resulting fra the. welding operation and thermal. loading. The remaining portion of the total axial stress is comprised of pressure induced primary stresses. For an internal pressure of 2200 psi, the pressure. induced stress in the axial direction is approximately 2 ksi'

. Al - 1 e

e

and is insignificant compared to the residual welding stresses and thermal

~

stresses. A rigorous three-dimensional finite element analysis, including crack gru nh with time, would demonstrate the effect of stress relief as a circumferential crack propagates through the nozzle wall. However,'this would

- require an inor.dinate amount of computational effort. Instead, a simple but conservative crack growth analysis is presented here that treats the residual and thermal stresses as primary stresses and uses the crack growth rate eg'uation i presented in reference 1.

The worst-case stress profile is located on the downhill side and is shown in Figure 4. A cubic polynomial fit of the stress data resulted in a set of coefficients to be used for a stress intensity factor calculation using the Raju-ca) solution for circumferential surface cracks (Figure 5). The outside Newman surface of the nozzle was machined and most likely contains a small cold-worked layer. The welatng operat. ion recrystallizes this cold-worked layer in the heat affected zone and no additional machining is performed after welding. With this 4 information, it cir be argued that the crack initiation for the outer surface would take a longer time than the inside surface where the machining process, which was performed prior to welding and in some cases after welding, introduced a layer of cold-worked material. The.refore, a very small initial crack size was  !

assumed (i.e.,1.0 mil) on the external surface of the nozzle. The resulting  ;

crack growth is (fiven in Figure 6. This shows that a minimum of 6 years is required for a 1.0 mil crack to become a through-wall crack if all the applied  !

stresses are primary stresses. In reality, considering the self-relieving nature of the residual stress, the crack would have arrested, probably 60% through  !

the thickness. As tated previously, the pressure stress is only 2 ksi. This translates into a stress intensity factor much lower than the threshold value for crack growth (9MPa/tii). This pressure stress is very low compared to the 90 ksi starting axial total stress at the outside of the nozzle. .

Even if a through-wall circumferential flaw gia." is possible, it would take more than 40 years for this flaw to grow circe Jerentially along the elliptical weld zone toward the uphill elevation. Here again, relief of the ,

residual stress will prevent complete growth of the circumferential fl aw.

Therefore, it is concluded that there is no possibility for an external ,

circumferential flaw indication to grow circumferential1y to the point of becoming a safety concern. .

i Al - 2 s

3.

' Gross Leak-Before-Break: The Ultimate Safetv Feature If it is postulated that a circumferential crack propagates through-wall and grows circumferential1y along the weld-nozzle interface region, the potential -

safety concern is detachment of the upper nozzle from the lower nozzle section and its ejection from the closure head.

This event is not likely to happen due to the following retsons.

. The first reason is as described above: more than 90% of the crack drivi force is self-relieving residual stress. ,

Upon opening of a crack, a significant '

part of the crack driving stress will be relieved so that either the crack growth rate is drastically reduced or the crack growth is terminated.

The second reason is due to a gross leak-before-break mechanism.The net section limit ligament is less than 10L Postulating that a large portion of the-nozzle cross-section contains a through-wall crack, there is ample room for leakage in Figure 7.

to occur before approaching the net section limit ligament as depicted This'will allor 'a detectable leakage of steam through this large i

crack, thereby providing ample warning to prevent the failure of the nozzle.

Furthermore, when the limit load causes the remaining ligament to start ido )

stretch, it would do so gradually and not be an instantaneous catastrophic failure . since Alloy 600 is a very ductile material. In addition, evidence indicates that th_e nozzles are in an oval shape due to interaction with the closure head deformation.

Therefore, there are gaps between'the nozzle and the head that will provide sufficient leak paths for steam to escape with fairly large volume thereby providing leak detection.

. Sufficient contact area would also remain to resist any slippage due to the ovalized nozzle.

The finite' element analysis results indicate that the maximum gap is -3.0 mils with an average of -1.0 mil during a normal operation.

This is an inherent safety feature that keeps the outside surface cracking from becoming a safety issue .

4.

CRDM Nozzle Straiohtenino CRDM nozzle straightening was performed by B&W during the manufa process for various nozzles to fulfill a straightness tolerance of 0.0055 inch / foot.

Manufacturing records only indicate the total number of CROM nozzles that required straightening.

There is no information available that indicates how far out of tolerance the nozzles were nor what means were used to Al - 3

l

l f

them.

The straightening process typically involves permanent bending. It is believed that in the case of a CRDM nozzle, straightening was performed after welding to the closure head by pulling the top end of the nozzle to one side.

This process would impart a very small permanent deformation to the CRDM nozzle, on the order of a mil, near the OD of the closure head. The shrink-fit portion '!

of the CRDM nozzle at the top of the closure head serves as the fixed end of a cantilever beam. Therefore, it is concluded that the straightening procedures i

utilized during manufacturing would not affect the stress nor the deformation near the high stress weld zone.

5. Conclusions Based on the above evaluation, it is concluded that the occurrence of '

nozzle detachment is physically impossible during the design . life of the B&W-plants considered in this study.

6. References .

1 l

1). BAW-10190P, Safety Evaluation for B&W Design Reactor Vessel Head Control

-Rod Drive Mechanism Nozzle Cracking, May 1993.

2). Raju, I. S. and Newman, J. C., " Stress Intensity Factor for Circumferential surface Cracks in Pipes and Rods under Tension and Bending Loads," ASTM STP 905, 1986, pp.789-805.

l s

Al - 4 i

I

)

7. Certification 1 This report is an accurate evaluation of the safety and c'onsequences associated with the possibility of external circumferential primary water stress corrosion crack growth analysis for B&W-design CRDM nozzles.

-1 l

4 K.K. Yo ' Date b

I Fractur echanics i i,

This report was reviewed and was found to be an accurate description of the work reported.

1 0% <eJ IE Y93 A.D. Kata Date Fracture Mechanics Verification of independent review.

Y ~

j K.E. Moore /flc'kf3 Dat'e Materials and Structural Analysis The document has been approved for release.

A.W. Robinson

/kVb5 Date 4

. Program Manager e

Al - 5 t-e e

Figure 1.

C' Bored CRDM Nozzle - Downhill Opr Axial Stress 64.4 KSI Yield Strength - Single Pass Weld 100000 - - - -

Wold

-- Inside t

Shrink Area' 't  !

80000 ---- --

• h 't" -

Fit Outside

-- - - -~f- - g{gg gg -

60000 - - - - - - - I ~4- -

I \

"--+"~^- - - -

t -

t -

- 40000 - - - - - - - - - - - - - -

1

-g- -

m + -

- -h- - - - - - -

a. .

t  ;

-u

s  :-

m m 20000 - -- - --- -l -

t- i-e )

-- - - * -i - - - - - - - -

x -

e s

r

-

M +-

^

0 -

,. . # ;__ _ =

m p s -

.~ ..

\

\

-20000 - - - -- - -- 3-h 1- -

p' '

i

+ .- - "- i --

- \ / '

p E s , i

-40000 - - - - - - - - - - - -

5 -

4 S-'J q -

- - - - - , 5 -~ "-- - - --

.. U r k 3

. . . . . . . . .,iy e s . e... .........,...,....e, ,

e i

e

• t 0 2 4 3

6 8 10 Distance from Bottom of Nozzle (inches)

Al - 6

Figure 2. -

C' Bored CRDM Nozzle - Uphill Opr Axial Stress 64.4 KSI Yield Strength - Single Pass Weld 3o0000 .

.... .. .. ._.. . . . . . . . . . . Inside

'  :  :  : =

. .sh. ri.nk 80000 - - - - - - - - --i-- -

Fit Outside i-

- 3 t - Area

.  : -.-4--

QQQQQ . . . . . . . . .

. . . ' .. . . . . i

...:....... . . ...'.........ue....

~

' it  : -

l

\

40000 --'--*'-

A-"

m - -

\-

ih - i g ..-/

i w

C1

, -g  :

m 20000 en I  ; i  :

-J' I

i-r1

-I' a-

-i' - - -

s, W

1

/ --

t. .-

0 , ,

/ .

g ~ i)

~ A: / .:

y y. .. j .- .-- .-

fs #W-

-20000 - - - - - - - - -

-- ---b N *l / 5l. I[

- - - -" i -- - - - - i!-Fi-W;8 Y .

. E 4QQQQ . . . . .. . .

.. . : ' . . . . ... :... . ... .. 4.h...,... .ay;;u.... a1. ..

~

W61d - .

-60000 - --- r!ea;!  !

-+- -

-- - + +" -

i t- c.

-" ;ab = =+ : -

1 0 2 4 6 8 10 Distance from Bottom of Nozzle (inct es)

Al - 7

Figure 3. .

C' Bored Nozzle - Avg. Axial Stresses Above Weld 64.4 KSi Yield Strength - Single Pass Weld -

-100000 . -

Steady. State Operating Stresses . -

. 80000 -

=: -- - - .

--+-----. --

60000 ----------i----------------. - -

+----- .--

~

4o000 ...........i.. . . . . .i . . . . . ..;.. . .; . .

m - i .

CL  :

g 20000

+--

e  % =

=

.m -

en  : - .

0 . -

i

-20000 ------i---- - - - i- -

-i.-

~

.. i-- . -

40000 . . . , .

3...

3 ..;. . . . .z., ..;... ..

-60000 ' ' ' ' ' ' ' '

0 20 40 60 80 100 120 140 160 180 Angle from Downhill Node 892 to Uphill Node 812 '

Al - 8 .

w W"' *-v m enee a en er man am='M  %=- .-em 4 * - - - -

Figure 4.

~ AXIAL STRESS (MAX LEVEL)

TEMP = 600 F 120

- FEM 100[ ~

O Cubic Fd gg -

O

• y _

_ 40 -

(n ~

(O 20 -

0 0 -

-20 -

m

-40 -

I

-60 I I I I O 0.2 0.4 0.6 0.8 1 1.2 a/t Al - 9

Figure 5.

External Circumferential Crack on a Cylinder (Ref. 2) '

N s l

/ -

- i

\

\

j' /

/m , /

/g

' : 2c

\

\ j l'yjY

\

/ ,j'

/' w l

~

e 9

9 Al . yo

6 i

I 5

i I

4 i

s l

sk yc ,

aa nr AC . .

h t

cr y wil oCn r l o )

s a n 6

. Gc ki t i o

r a

e cpt e as a c

I 3 (y 1 r r u Csl o e 1 g l amg e i i

m -

i it T

F nSed l

e O ,.

A e

r an l

mr f

ut e .

cx r E _

i C

I 2

s s

e n

k i

c t 1

h T_

l e

zz g _

o e -

N

/

dc .

O h

t p _

e D -

k - - -

c _ -

a r 1 0

8 6 C O 4 2 0 0 0 0 A,- .

Figure 7. CRDM Gross Leak-Before-Break Georetry np N

v i

I i

l Mll. Average

\ 5 Gap (operating) '

Counter-Bore '

i

// N

\ R emaining Ns Ligament s

i

\ Circumferential Crack Al - 12 2

l ATTACHMENT 3 Requested License Amendment Revisions to FSAR Design Basis 1

i i

l i

d f

J

-, , n

TMI-1/FSAR 5.2.3.2.6 Missiles Missile protection for the Reactor Building liners a) The building'and liner are protected from loss of function due to damage from such missiles as might be generated in a loss-of-coolant accident for break sizes up to and including the double-ended severance of a main coolant pipe.

b) The engineered safeguard components required to maintain containment integrity and to meet the site criteria of 10CFR100 are protected against loss of function due to damage by the missiles defined below.

During the detailed plant design, the missile protection necessary to meet the above criteria has been developed and unplemented using the following considerations:

a) The reactor and reactor coolant system are each surroundec by reinforced concrete and steel structures designed to withstand forces associated with double-ended rupture of a main coolant pipe and any missiles that may be generated, b) The structural design of the missile shielding has taken into account both static and impact loads and is based upon a barrier cross section with energy absorption capacity at least 25 percent greater than that required when considering a potential missile.

c) Components of the reactor coolant system have been examined to identify and classify potential missiles according to size, shape, kinetic energy, and driving force for purp ses of analyttng their effecto.

d) The types of missiles for which missile protection is provided are:

Valve stems up to and including the largest size used Valve bonnets Instrument thimbles Various sizes of nuts and bolts Reactor vessel head bolts centsel red dePee :: henir r e) The list of credible missiles on the reactor vessel head and the methods for identification of potential sources ofinttmal missiles are found in Reference "TBD" {

i Reference "TBD" = Framatome Technologies, Inc. (FTI), Document 51-1240140-00,

" Reactor Vessel Missile Shield Removal Repon", dated December, 1995.

t

From TMI-l FSAR Para. 3.2.4.3.2.1 I TMI-1/FSAR

j. Position Indications Two methods of position indication are provided: one, an absolute position indicator and the other, a relative position indicator.

The absolute position transducer consists of a series of magnetically operated reed switches mounted in a tube parallelto j the motor tube extension. Each switch is hermetically sealed.

Switch contacts close when a permanent magnet mounted on the upper l end of the lead screw extension comes in close proximity. As the l lead screw (and the CRA) moves, switches operate sequentially, l producing an analog voltage proportional to position. The accuracy of the analog signal is 2.0 inches (full scale is 139 inches), and the accuracy of the readout is 3.4 inches.

Additional reed switches are included in the same tube with the 1 absolute position transducer to provide full withdrawal and  !

insertion signals. The relative position transducer is a small 4 pulse-stepping motor driven from the power supply for the rod drive motor. This small motor is coupled to a potentiometer with ,

an output signal accuracy of 0.97 inch producing a readout with l an accuracy of 2.4 inches. I

k. Motor Tube Design Criteria The motor tube design complies with Section III of the ASME Boiler and Pressure Vessel Code for a Class A vessel. The operating transient cycles, which are considered for the stress analysis of the reactor pressure vessel, are also considered in the motor tube design.

Quality standards relative to material selection, fabrication, and inspection are specified to ensure safety functions of the housings essential to accident prevention. Materials conform to ASTM or ASME,Section II, Material Specifications. All welding shall be performed by personnel qualified under ASME Code,Section IX, Welding Qualifications. These design and fabrication procedures establish quality assurance of the assemblies to contain the reactor coolant safely at operating temperature and pressure.

In the ghly unlikely ent that a pre sure barrier c ponent or .

the co rol rod drive ssembly does f I catastrophic lly, i.e.,  !

rupt e completely, e following re uits would ens  :

Control Rod rive Nozzle i

1)

The asse ly would be ej ted upward as missile until t was sto ed by the miss e shield over he reactor. T s upward motion would h e no adverse e ect on adjace asse. lies.

l

a THI-1/FSAR 2 Motor Tube The failjlre of this comp nent anywhere ovethelower/

flange ould result in missile-like ection into t missi e shielding ov- the reactor. his upward mot n wou have no adver effect on adj cent mechanism

( ference 72).

3.2.4.3.2.2 Axial Power Shaoina Rod Drive Mechanisms For actuating the partial length control rods, which maintain their set position during a reactor trip of the shim safety drive, the CRDM is modified so that the roller nut assembly will not disengage from the lead screw on a loss of power to the stator. Except for this modification, the shim drives and the axial power shaping rod drives are identical.

i i

I i

l l

l