ML19274E708

Forwards Discussion Based on Postulated Failure of Flow Splitter Plate & Postulated Evaluation of Reactor Coolant Pump Suction Elbow Splitter

ML19274E708
Person / Time
Site:	North Anna
Issue date:	04/15/1979
From:	Stallings C VIRGINIA POWER (VIRGINIA ELECTRIC & POWER CO.)
To:	Harold Denton Office of Nuclear Reactor Regulation
References
259, NUDOCS 7904170201
Download: ML19274E708 (51)

Text

.

VruuNIA ELECTHIC AND Pownu ContwNY H ICIIM ON D. VI HOINI A 23261 April 15, 1979 Mr. Harold R. Denton Serial flo. 259 Office of Nuclear Reactor Regulation LQA:EAB/pwc Attention: Mr. D. Vassallo U. S. Nuclear Regulatory Commission Docket No. 50-338 Washington, D.C.

20555

Dear Mr. Denton:

At the conclusion of our mee'.ing on April 12th where representatives of Virginia Electric and Power Co. and Westinghouse Electric Corporation made technical presentations which demonstrated that the flow splitter plates on North Anna 1 are structually sound, you requested that we provide a written discussion based on a postulated failure of a splitter plate and an accompanying Safety Evaluation assuming such a failure would occur.

This information is attached. Also enclosed is the written description of the analytical and ultrasonic examinations which we presented.

Based on the results of metallurgical and ultrasonic examinations, we have concluded that failures observed on Unit 2 were a result of high cycle fatigue which occurred early in life, and that this has not occurred, and will not occur on Unit 1.

This unit has operated for approximately one yegr, during which time the accumulated number of cycles is on the order of 5 x 10' cycles, well beyond the 106 cyc

- where fatigue failure would have occurred.

Neverthe-less, to confirm the cor. alued integrity of these plates the same ultrasonic examination will be conducted during the second refueling of North Anna 1.

We feel this information which has been reviewed and unanimously approved by the Station and System Nuclear Safety and Operating Committees is responsive to your requests, and it further supports our position that North Anna 1 will operate safely and that it should be returned to service promptly.

Your concurrence and prompt notification to Region II permitting power operation to resume will be very much appreciated.

Very truly yours, b. A'N. kdk2 M y C. M. Stallings Vice President Power Supply and Pr, duction Operations Attachment g

[/g cc: Mr. J. P. O'Reilly

\\

7 9 0417 0 Jo l -

SAF' fY EVALUATION OF REACTOR COOLANT PUMP SUCTION ELBOW SPLITTER NOR*ll ANNA UNIT 1 2

April 15, 1979

l

.INDEX

1) Description of Issue

2) Fragment Size Determination

3) Equipment Considerations

~

(a) Splitter Elbow (b) Reactor Coolant Pump (c) Reactor Vessel General u

. Internals

~

,. Fuel Assembly

. 4) Safety /.A.c,cident Analysis

5) Attachments - 4

, g.,....,_; _.

8 1e

/'

.N

~

h 4

i t

4 e

SAFETY EVALUATI0f!

0F-R 3CTOR C00LAllT PUMP SUCTIOli ELB0W SPLITTER x

1.

DESCRIPTIO!1 0F ISSUE The purpose of this report is to evaluate the impact on plant safety and performance resulting from the postulated failure of the reactor coolant pump suction elbow splitter. The t!RC has requested that this evaluation be performed as the result of an isolated splitter failure that occurred at florth Anna Unit 2 11uclear Plant some time during plant pre-start up

~

testing.

The splitter element for which this evaluation is being performed is in-stalled in a 31" x 90 elbow.

(See Figure 1). The splitter element is fabricated and installed in three sections. The splitter element is full penetration welded along t!.e axial length of the fitting and full penetra-tion welded between the three sections.

The splitter material is ASTM A-240 TP 304 cold rolled plate 1-1/16 inches in thickness.

The elbow

' material is SA351.CF8M. The splitter / elbow weld is located approximately

' l-1/2" from the the inside diameter of the pressure boundary.

a For this report various sizes of failed plate ' material'will be selected and evaluated,to determine their impact on plant and equipment performance.

The smallest size particle that could result from this postulated failure

. in quantities sufficient to affect plant / equipment performance will be

1. determined by reviewing the fracture characteristics of the failed splitter.

Intermediate size portions of a failed plate that could possibly pass through the inpeller will be detemined by a study of the.

pump geometry and interactions of the plate with the pump internals.

The effect of large portions of failed plate that could lodge at the inpeller inlet will also be considered.

Also included as attachments 2 and 3 are 'two additional analyses done by Wexinghouse as support to the arguments presented in this report.

t

.J.< FIGURE 1

... ~ _

SPLITTER PKP SIDE

/,

,s lf r,

s%

".. L' T.~. I!

%\\

t,

\\ \\

b I

\\\\

31" x 90 Eli

'\\

s

/

t l

e yi s s fl

~

~ -

I f,

~

./,

/

./

/

/-, e

/

/

's \\

lx

/

7

/

7 f/

's \\

\\

./.

SPLITTER g ->-

t

,=, x /

I

\\

s\\

//

/

ELDS

~

N

-,/

M

,'<y N N __.

N 1

y' meEsinE a

\\

9 S

G e

2-e

l. FRAGMENT SIZE DETERMINATION l

t i

3 For this evaluation the largest size fragment to be considered will be that which would not pass through the Reactor Coolant Pump but would lodge at the pump inlet.

i I

Intermediate size fragments will.be determined by 'considering the largest size particle that would pass through the reactor coolant pump and eventually find its way into the Reactor Vessel.

The possible size of this fragment i

will be discussed in a latter section of this report which describes the interaction of the pump internals with a failed splitter plate.

s The smallest particle to be considered was detcrained by performing a detailed examir;ation of the failed splitter plate.

The nature of the fracture surfaces indicate that the nost probable size of the smallest fragment could be defined by a 1-1/16" cube.

Smaller size fragments are possible but these would be limited in number and their ef f ects in the primar" system would be inconsequentin1.

O e

W g

s o

O W

mom 9

O e

e S

e 3

s EQ' IpMEllT C0ilSIDEPATI0g J

3.

... s (a)

Splitter Elbow

~

~

The pressure retaining integrity of the elbow is not af fected by the postulated splitter failure.

The splitter is welded to an integrally

. cast transition area which is 1-1/2" removed from the elbow wall. The elbow as-cast has a minimum calculated wall thickness tm = 2.88" based on the Af SI B31.1.0 minimum design stress S = 14,950 psi for m

SA 351 TP CF8M material.

The applicable code for the elbow is AliSI B31.7 - 1969 which pemits a design stress S = 18,700 psi resulting in a minimum calculated m

wall thickness t = 2.25" using the code equation:

~#

=

m

.~--

+

.Pd

~

.y n..., m

~

[i]

=

This conservatism results in an excess calculated t = 0.63 inches.

m Metallurgical studies performed on the cracked splitter samples conclusively show crack initiation in the weld area only.

Due to high cycle fatigue, the crack propagated along the weld line and then due to the character.istics of the geometry and vibratory mode of failure, propagated towards the center of the plate away from the fitting wall.

i

~

In order to obtain-an estimate of the integrity of the pipe wall impacted by a piece of the splitter plate,the gross assumption was made that the dislodged piece would be 31" x 19" x 1-1/16", weighing 175 pounds. For simplicity,'the piece was assumed to be a 31" cylinder, 5" in diameter. The calmaated impact force.is 5,000 lbs, significantly less than the punching shear f esistance of the elbow naterial, which is greater than 500,000 lbs.

e 9

0

. _ J.1'

-(b) P.eactor Coolant Pump Consideration of hydraulic passage cross secticnal areas and shapes in the pump impeller and diffuser lead to the conclusion that the largest piece of 1-1/16" plate which could conceivably pass through the pump could be no larger than about 9" by 9" square.

llowever, it's very unlikely that a piece anywhere near this size would pass cleanly from the impeller outlet to the diffuser inlet given the relatively high tangential velocity (2: 180 fps) of the impeller relative to the diffuser.

It is estimated that objects with dimensions greater than 3" in the radial flow direction are likely to be pinched or sheared between the impeller and The extent to which this may damage the pump is discussed diffuser vanes.

below.

.,. u.

Small Pieces m.

,,. ~n,

-~

T,here is a 1.5" radial clearance between the impeller outlet and the diffuser T"

tiilet, so smail objecL3 (less Liieri 1.5" 00 a side) nill iciid to p6ss through the pump hydraulics without pinching between the rotating and stationary parts and will at worst only locally dent the impeller and diffuser vanes as they

.bo'unce through. This would not significantly effect pump performance and is of negligible concern from the standpoint of the RCP.

Large Pieces Of greatest concern, with regard to the RCP, is the largest piece which could I

pass through the impeller and be pinched or sheared between the impeller and diffuser vanes. A piece too large to enter the impeller or which some how beccmes lodged in the impeller is of lesser concern since shaft vibration increase and a possible reduction in '. cop' flow will serve to alert the operators and allow a pump shutdown before pump damage of a more serious nature, such as shaft failure, can occur.

i s;

• .,3

. 7 *,g,

e G

'The largest piece (9" x 9" x l-1/6") which could reach the impeller exit would

, most likely cause very severe impeller and diffuser vane damage as it

' ' encountered the stationary diffuser vanes at. approximately 180 fps before it cleared the impeller. Because of the relatively large diameter of the impeller this would cause a very large transient retarding torque to be applied to the

. pump rotating assembly as the piece was sheared or otherwise deformed by the impeller. Under such a torsional 'rading condition the weak link in the rotating assembly, by a factor of more two, is the impeller key.

A stainless steel piece with a shear area of 5 t'o 10 inches squared jamming at the impeller exit - diffuser entrance could generate enough torque to seriously deform and possibly fail the impeller key.

Key deformation would cause a significant-increase in pump vibration levels as the impeller shifted off its rotational center.

Key failure would, of course, cause an immediate loss of loop flow.

In either case, the broken or damaged pump parts would be expected to remain within the pump, and shaft seal failure would not be expected. Diffuser damage in all conceivable cases would be restricted to the locality of the inlet vane ~ edges, and, nest likely be limited to deformation rather than fracture.

No gross failure of the diffuser structure would be' expected..Likewisc, the impeller damage would most probably. be deformation, possibly severe, but not fracture.

In all cases lateral vibration or bending of the shaft due to impacting of objects or imbalance is not considered to be a short term problem and would not be expected to cause shaft, bearing, or seal failure so long as operating i

time under such cbnditions was limited.

This means that pump shaft vibration levels should be continuously monitored, and the pump should be shutdown immediately upon the detection of abnormally high shaft vibration levels.

9 I

+4

(c) Reactor Vessel General Calculations were perfonr.ed to estimate the velocity and kinetic energy of a 304SS object 9" x 9" x 1-1/6" in size at various positions in the reactor vessel.

The total vessel ficw rate assumed was 315,600 gpm. (This corres-

~

ponds to the mechanical design flow rate for North Anna.).

The velocity in the cold leg nozzle is approximately 57 ft/sec and the kinetic energy of the object would be 1276 ft-lb.

This corresponds to the energy imparted by the f

object to the core barrel.

The object would then pass down the annulus between the thermal shield and reactor vessel.

The velocity in this region is approx-imately 3G ft/sec aim the kinetic energy of the object would be 506 ft-lb.

f The above corresponds to the velocity and kinetic energy of the object as it enters the lower plenum of the reactor vessel.

In the lower plenum of the reactor vessel the flow velocity decreases. A velocity of 5.5 ft/sec would be sufficient to lift the object through the lower core support plate.

The

. object would stop at the underside of the lower core plate.

It should be noted that depending on the orientation of the object in the vessel lower p'lenum the object might come to rest at t'he bottom of the vessel. See Fig. 2, 3 & 4.

Hor'th Anna has a loose parts monitoring system.

This system is capable of sensing objects with a kinetic energy of 0.5 ft-lb.

Assuming that an energy f

of 1 ft-lb is impacted by a 3045S object traveling with a velocity between l

f

~

3 20 and 35 ft/sec the volume of the object could not exceed 0.55 in, or approximately.15 lb. in weight. This implies that the loose parts monitoring system would be more than adequate to detect any object of appreciable size.

A complete description of the loose parts monitoring system is provided in At tachn.ent 4.

Internals The following presents a summary of the analysis and results obtained by an analytical review of the effect on the reactor vessel and internals due to loose par'ts resulting from pieces of a flow splitter entering the reactor vessel at the cold leg.

The items ieviewed^ relative-to thd internals and vessel fall into two main categories.

- Loose parts impacting on the lower internals structure.

- Loose parts lodged in areas of clearance and becoming wedged during periods of relative thermal growth.

t W

M

FIGURE 2-

~

s

~

t

'. (

d q A

h,, 3 % ; ', F 1

n1 9,A t m.n. _vr..=.. :. a u 2

.... ~ -

(2D:iE di':6.93 i

g d c=CYC:-:

17.1,

n[m,Il,l 7. *.. g

~

s (e,, _

y

, c. -,.--

ri

' $,..s~':& x;w1; %l

~

-(

][ji'3s

~

[

s

.o, aoa u

1 MR p

Hj Q.

a

.m..l x

. A.l_._l1 I:!

N

.s.

< a.. m Ae;= J.. J. w_s, y').,,

imq o n

i x

t. u

.,g c

m-en n:

s;.y nye

gi g. gm>

a l..L.,q e

3 l:'.~

w q

o. n., c -

m,..,.

1.

u

/

w

..s.

.._,.s..

}

b, !. _d

,1, "i

.S j'

[ b lF h l ' F p i1,j 9 [

]

a y

i N

. p.. j _1 ;,.y

.(-.,

eesseo

..].

l*

l Md,.b"1)L, J.

t i.

e %s o

9 I.b.1j-fg,',+

e s t C'i O

e e

4 i

~l4 O

O d

c

~~G

.C 4.9.4 p

G:-;irEp~a. Cf/,-2-d. TN. /,{

G. f.),CS'.'r2. ' $ _._.yj,.'3,

O F

..,..t--

t--

' ': ~ 't SQj g

m:_

% n i M 5+L.! W ?-~ ;. - 2,

2.;.. c.f e. '.O

, i:. ' s = ;;. C ":: y,', '/d.

I

/

R{,'E)..' ~~' ~, L.

a p.

/

\\

c. -- - -. - % -- p+,w/,. '-h, s \\, g 1;. -

,% fi'

.p.i

~~.~..}.. =.,c l.

.s.

C-6

=:cx,.nl 1, >! %..' d f 3.' x,. p'*'g' *v:. ~~.,'

/

'O

<; c.:

\\1,1: n

.- a f,,' ji'.;c D..,g &,.c @.

I

/,/,

6-

.. a.

cJ n

,,,v,1 q,l.Rij; q. f-Q jr : W e

p 9

.-<.,----a-

. > =.p > !; p r

... h; 2d;fD.-9 '_Y..'df['f".'c.h..

i. ~c.

@;p ';;.9

,;'t.,-;fl,'

3.-

j.$

VF G

ni; @/ f pi[f

\\ z o'2.B':'&$9grsw'o

'A ': ^

O e

j a

e e$n%s. y,r.

y. y,,_,u.9 n.m.i

,

• o

/

.,..w..

o

• n eyw j*

O

.' 4

• s==%-. '

O

-- :. /

O 1

'S p !.. M' =,:!iur j

/

e

3.,, bn 4 O

O O g 6 O;O O-w.y;..r m

Ft. *=. p :=. 7... ;%q...

/

O a ' ! ' 's d,;j ^.}.I '.3.j k j ;,,

/

ip 5-". -g-m=q17 9?'ss!

4. f h.I. ']l L!

S ' 11 i il 5 d

i..

ea y,

e' E

.- {

%,,S s.lI c.4'+( l1 c

yu' '

s

[

]

f 1-

._e w wm.

t

_H ul g/

a.

..m-m n. na ue n,

tw c.

l f,

h

,8 REACTOR VESSEL GENERAL ASSEMBLY t

8 y

y

}1 CURE'3

,!....g

.::.l

,g.-

1

~

o, ;.

\\

.a w

l{EAD & VESSEL

. UPPER SUPPORT ASSFJABLY

. '. c> *m ALIGNMENT PIN

~

~

g

.: N :: q v g >. ::

9.qq3.. g,. l ~.a..':'q) y -

7.

s 1 I:

.:@G.pp::,@,,:j p.@q> :

h:.

INTERNALS 3

j

' y['G'gg) Y j

_ SUPPORT LEDGE

(,

C',

. ),r_-

CONTROL RC3

~

. 2s. m.c-s

,, [

K,

-... O ".

, ~//

GUIDE TUBE

~

I UPPER SUPPORT w b ~r5 ~--o 1 M i

UPPER CORE

/.

/

- }.,

c-COLUI,,u,l N

x

,. ~ x

. ARREL B

- x,

,3.-

- N[; g r

~

i.,

l p

8

.1,

r %

- Q 9 G G i3 R.,j % /-W 2

I

^

t e.g,Ilt= e src +;-J @Ja /

N t>

TilERMAL SillELD tc.sl: = -t

'Ed[nNE$.MS.!?lM h hg. 2.f..w y g, 4 g ig.w-n:-pre:r 3

['/- UPPER CORE

.l i

/

rt V

hiMif:g...

c.,re p n ;cjyd:

IRRADIATION

.d, SPECIMEN GUIDES m N'y]I Ij g j.jc(i g, A[.6

);

PMTE s

g

[H

-w m

.t iY 91.

S'W~ '

LOYlER CORE I

y-)' 4.! !3[i....j

.g- [

BAFFLE ^

.--'il D) /l Qj BARREL

' 'T QY (j p 1 -

1 SJfi I

A l-ly; El

l......M

}f]- FUEL ASSEMBLY FORMEP' o

M l

p[=?]

c i..

s i

)

' ji

.*i fg 4

[' ;

l hi g LOV/ER CORE

~ ~ " ~ ~

}...

/'

PEATE g g ki; 3

yu ju i...... p.=g! j'4m ACCESS PORT U

V q

p r CORE e?n cl l

SUPPORT COLUMN

.... %w

% -y y/

._,._~,m.~m.,

l M M M MP-Ull@..x:

LOWER SUPPORT EEMS.E;%.- : >'M'%

C 2 :::

,i 3.

COLUMN N{

E.7EEh ll f-DIFFUSER PMTE

'tq% q n:.e:c -A

.a m ei

/

Q W!m='..,V9 d.?:: IwW5?, '.,

LOWER

'[

,1

s. s.

r, m.

pg,Js.,7.;%_ q{:/ m.br,h..k.. v.

GUIDE TUBE TIE PMTES eo,n.

f M-4 INSTRUMENTATION

. ji

.i A.v, Pt. y.m

(

Pl@l @':qm%p~d ;py rife /

CORE SUPPORT 3A

.M

~

FORGING f,

RADIAL SUPPORT ms g q gg 1

k[j.ll{5hhkj j

B0lTOM SECONDARY l

[)

SUPPORT ASS'Y j,

INSTRUMENTATION GUIDE TUBES a t,q,

_g s

V z'

i

[,)P.T, u,.i q.h dIhh.1] hh -

i W J lfh.

/g,

.3

.s

)

[

r.

s

:..Q.~.. *.

(

z.9

. s s.

, ; ~,

i

~ *

. a.

. ~...

. s,.

y s.

. \\.

.......z.

. J,

=

..y S

j-,.

(FIGU.RE'4 1:

'. l'a

. 1...

.a -

~ '. ::..

1.*

'.12...' x.

.,...I.-.----*--.

L- - e l' -

~.~*

-.r

.~.

- : c.

J.,

, 2,.

o-

. -... ::.,+-.:,..:,s

. 2:1.. ',

l

,i.,

w.

l

.l, RADIALSUPPORTKEY(6)

- 1..- '.y 1

1-?

4

. Tl., -

7

!!AllWAY COVER CORE SUPPORT

\\

I P A

A A

y, u(?'

k\\ y,. l b' V:/

v:/ '

l WDIf: WI 9 C 1

• q

=w 1

s g

l n

a

=

]

e -

j i

I n

i s

i j

l j

i h

td ws T1

! 4

'l Vigoll}6T 1lOM s

rBmMMI l UN i NiW1 I

Mt II:

U

• .*'e' M e n

,,L,f

! Ei :l1

.r w

m mm m#1 oo h h. i.

ii sN I

1 i in

. E d.

ik

-]

p s _,_,I.p.r md aJ

\\

! e

~

et 5, g m._

n-u.-

1

_ypm, w--r _u.. -

y' pr n

a

.y

[ [

!I f

g Q

• ~

SUPPORT RING s.

L0hER CORE PLATE l

l j'

i u, l

,nus c 1 i

l u

t.

l 6

i

s..

l i

UPPER TIE PLATE L,-

l 1

l l

l

,..o u,

.s g.r 4

4 m

.. m q

,g m

j

~~

g' f CORE SUPPORT g

l

~~ COLUMN LOCKING e

4' LOWER TIE PLATE

(,

f 1

,}

NUT i

3 y4 e,

c.

j

\\l l

[* N[14' d

f

- !!!STRUMENTATION GUIDE,

{b EHERGY ABSORBER ASS'Y 0FF-SET INSTRUMENTATION COLUMN

. i

, COLUMN (CRUCIFORM)(25)

SECONDARY CORE SUPPORT j

BUTTTYPECOLUMH(30

.s a

.. ~ -

'~-

\\ %.

a:

.. L

(

. ~

s e

Specific areas judged to be critical are as follows:

(1)

Impact on core barrel by flow splitter piece.

(2)

Piece of flow splitter lodged into one radial suppor t key and clevis (3)

Piece of flow splitter impact on bottom mounted instrtruentation tubing.

(4) Piece of flow splitter lodges between the energy absorber of the internals and the reactor vessel.

P,esul ts :

The resulting conclu ions are as follows:

(i)

Impact on core barrel by a piece of the flow splitter.

A piece whose size 9" x 9" x 1.1/16" will not shear..through the core barrel, but causes the normal and upset stress limit in the ASliE Code Sub-section tiG to be exceeded,'.This would result in local deformation of the cohe' barrel.

(2)

Piece of flow splitter lodged into one radial support key and clevis.

A size of I 1/16" x.50 x 9.0 could lodge into the clearance area between the internals and the reactor vessel clevis, at cold conditions and then wedge when the plant is heated up.

The resulting load imposed on the core barrel and reactor vessel causes the pre-load at the flange pic e is,sedgcd, and could rcsult 6v vu.vou

.v.

nc us.usion unu onu in yielding of the internals hold down spring.

The interference load and pressure stresses. produced on the internals and vessel would be within code allowable limits.

(3)

Piece of flow splitter impacts on the vessel bottom counted instrumentation tubing.

The impact force caused by a piece whose size is 9" x 9" x 1.06" would severely damage the tubing and associcted instrument, and cause stresses

, in excess of the AS!'E Code allowable values for normal and upset operation.

Plastic analysis of the tube indicates gross deformation.

The pressure boundary would not be breeched.

(4) Piece of flow splitter lodges between the energy absorber of the internals and the reactor vessel.

A piece with a contact area of one square inch lodged between the absorber and vessel will result in the reactor vessel stresses to remain within the code allowable limits when a load of 400,000 lb or less is applied.

As shown in the study for the typical three loop plants with steam generator plugs in the reactor vessel the load required to yield the energy absorber is between 147,500 lb for one column to 590,000 lbs. for four columns.

f_uel Assembly Coolant flow blockage can occur with an assumed splitter plate piece entering the lower internals. The blockage can hypothetically occur by simultaneously covering all four lower core plate flow holes located directly below a fuel assembly with a piece approximately 9 inches square.

Blockage can also occur from a smaller piece entering one of the core plate flow holes.

In both cases, the flow blockage causes local reductions in coolant flow. The effects of the coolant flow blnckage in terms of maintaining rated core performance, have been determined. With the reactor operating at nominal full power conditions, and the fuel assembly inlet nozzle completely blocked, the effects of an increase i.. enthalpy and decrease in mass velocity in the logier portion of the fuel asse:rbly would not result in the reactor reaching a minimum D"BR of 1.30 (reference Attachment 1 from RESAR-3S).

In reality, a local flow blockage is expected to promote turbulance and this would not affect DNBR at all.

Such debris A piece of debris larger than the plate thickness could be considered.

would enter from the lower internals.

The debris would then permanently lodge in the fuel assembly bottom nozzle plenum.

The bottom nozzle flow holes, being con-siderably smaller than 1-1/16, would prevent the debris from further novement through the fuel assembly and reactivity control components.

Thus, proper functioning of these reactor components, which includes the cantrol rod assemblies, is maintained.

N t

. t

4.

SAFETY / ACCIDENT /WAt.YSIS The mechanical and themo-hydraulic effects of a foreign piece upon the reactor coolant system, reactor coolant pump and the reactor core and core structures have been discussed in the other sections of this paper.

These concluded that there viere no significant hazards to those components presented by the hypothesized / presence of a foreign piece.

From the safety / accident analysis vicwpoint, three analyses are discussed. These are:

partial loss of forced reactor coolant flow; complete

loss of forced reactor coolant flow; and single reactor coolant pump loched rotor.

Considering an assumption of a foreign piece of a size sufficiently large to not pass through the pump rotor, yet small enough to be lifted up through the elbow by hydraulic forces, a situation with potential effects on pump flow

, can be hypothesized.

An orientation, where a foreign object is held against the impeller inlet, would reduce pump efficiency and very likely result in unbelance and con-sequently an increased amplitude of vibration. This, in turn, would necessitate a shutdoicn of the pump and a subscquent examination of cause.

The effect on loop flow likely would not be of sufficient magnitude to generate a reactor trip on low flow; however, in any event, the flow reduction

.would be less than the partial loss of' flow event, the results of which are shown to be satisfactory in Section 15.2.5 of the North Anna FSAR.

The Unit 1 splitter plate evaluations concluded that all plates were structurally sound. The sudden non-mechanistic failure postulated by the HRC would be considered to constitute a single passive failure in one loop.

In that event, there would not be a related failure or loss of flow in the other loops. The total loss of forced reactor coolant flow would not result and therefore, this accident analys,is is not applicable.

A third situation was theorized by the NRC, that of a locked rotor.

It would require a foreign object of appropriate geometry to enter but not pass

through the impeller and to then extend beyond the outer diameter of the impeller to a sufficient degree to impact the pump diffuser.

Considering the values of rotation initeria, relative to the structural strength of the object configuration, it is expected that the object would be def orir.ed sufficiently to clear the rotational obstruction with lesser damage occurring on the pump components, and a sudden and immediate locked rotor is not anticipated.

In a worst case, an obbreviated coastdown would be anticipated.

An abbreviated coastdown occurred on Surry Unit 1 at approximately 80!' power as a result of a sheared pump shaft.

The later results and evaluation of that event showed the core did not approach an unsafe condition and the results were much less severe than the evaluation of a locked rotor event, as reported in Section 15.4.4 of the North Anna FSAR.

Ihe effect of a foreign object interacting between the pump impeller and dif fuser would be expected to be less than those shown by experience with the sheared shaft.

/

4 s-

^

AT1ACI W!;T 1

~

Fuel I:od Behavior Effects from Coolant Flw Blockaoe N

Coolant flou blockages can occur within the coolant channels of a fuel assembly or external to the reactor core.

The effects of fuel assembly blockage within the assembly on fuel rod behavior is core pronounced than external bicchages of the same magnitude.

In bo15 cases the flow blockages cause local reductions in coolant flow.

The amount of local

, flow reduction,. where it occurs in the reactor, and hmi far along the flow strean the reduction persists are consideration:, which will influ-ence the fuel rod behavior.

The effects of coolant flow blockcees in

- terms of maintaining rated core perfcnnance are determined both by ana-lytical and experimentel methods.

The experimental data are usually '-

)

' use'd to augment analytical tool.s such as computer programs similar to Inspection of the DI!B correlation (Section the TilINC-IV progrcm.

4.4.2.3 and Refercnce [44]) shows that the predicted D."Dit is dependent upon the local values of quality and mass velocity.

-+

The TillMC-IV Code is capable of predicting the effects of local flow blockages on DNCIt within the fuel assembly on subchannel basis, regard,

less of where the flow blockage occurs.

In P.eference [63], it is shown that for a' fuel assembly similar to the Westinghouse design, THIHC-IV i1 y

accurately predicts the flow distribution within the fuel assembly vhen the. inlet nozzle is completely blocked.

Full recovery of the flow was

~

found to occur about 30 inches downstream of the blockage.

With the n.

t

% )3

• .-l.

p~

3,s #

I

y reference reactor operating at the nominal full povier conditions speci-J/

ficd in Table 4./,-1, the effects of an increase in enthalpy and decrease in mass velocity in the lower portion of the fuel assembly would not result in the reactor reaching a minimum DMUR of 1.30.

Fro:a a revicu of the open litera ture it is concluded that flow blockage in "open lattice cores" similar to tiie 1.'estinghouse cores cause flow perturbations which are local to the blockage.

For instance, A. Oktsu-bo[82] et al., show that the mean' bundle velocity is approached asymp-totically about 4 inches downstreara from a ficw blgcLage in a single flc.t cell.

Similar results.nere also found for 2 and 3 cells completc -

~

ly blocked.

P. Basmer[C33, et al., tested an open lattice fuel assem-bly in which 41 percent of the subchannels were completely blocked in the center of the test bundle between spacer grids.

Their results show the stagnant zone behind the flow blockage essentially disappears after l.65 L/De or about 5 inches for their test bundle.

They also found i.

that leakage flow through'the blockage tended to shorten the stagnant zone or, in essence, the complete recovery length.

Thus, local flow blockages within a fuel assembly have little ef fect on subchannel en-thalpy rise.

The reduction in local mass velocity is then the main parameter which affects the DNBR.

If the standard plants were operating at full power and no:ainal steady state conditions as specified in Table 4.4-1,- a reduction in local mass velocity greater than SG pcrcent would be required to reduce the DNBR from 1.74 to 1.30.

The above mass veloc-effect on the DNB correlation was based on i'.a assumption of fully devel-oped flow along the full channel length.

In rea'ity a local flow b1'och-age is expected to promote turbulence and thus would likely not effect DNBR at all.

Coolant flow blockages induce local crossflows as well as promote turb.u-lence.

Fuel rod behavior is changed under the influence of a sufficiently A

• -2

, ;,..w,

\\

Allticil:Elli 2 Summary of flesults of Analyses Performed to Support an Evaluation of the Effect, of a Small Loose Part Lodged Between the Reactor Vessel and the Bottom Plate of the Secondary Core Support IMRCll,1978

~

. -1

/

00:lTEllTS 1.

Introduction 2.

Analytical Model Used for Co:r.putations 3.

Calc 01ation Technique 4.

Force and Deformation Results for Several Transients 5.

Reactor Vessel Stresses 6.

Discussion e- -2

/

" 1.

Introduction A loose part was detected in the bottom of a typical }] PUR reactor vessel

/

by a Hetal, Impact gonitor* late in 1976.

Subsequent measurements using the Mill detectors indicated that the part weight was less than '.5 pounds o'r approximately 2" x 2" x l-1/16" thick.

As part of their evaluction of possible effects of this loose part, Uestinghouse was requested to perform preliainary calculations to determine the effects of the loose part, assuming that it became lodged between the reactor vessel and the bottom

. plate of the energy absorber. Additional assumptions to be made were that

1) owing to the ucight estinated from the Milt data, the part could be assumed to be under one of the four posts of the energy absorber, and 2) since the identity of the part has not been established, preliminary calcu-lations should include those assuming that the part was rigid.

Other conservative assumptions such as the assumption that the part perfectly conformed to the surfaces of the vessel clad and louer plate, and that these surfaces remain clastic were included in the calculations.

The forces computed from the structural model were used to evaluate the net h'olddown force at the core barrel flange and the stresses in the reactor vessel and internels.

Vessel stresses were tamputed with handbook formulas and include pressure, thernal and loose part induced s. tresses.

The results were compared with allowable stresses for normal on ation.

Internalt stresses were f;und to be within allowable values by comparison of the ioads imposed by the lease part with loads considered in design calculations.

Initial calculations were done to scope the effects of the part betoning lodged during heat-up or power escalation of the plcnt.

Calculation; were also pcrformed to determine a preliminary basis for heat-up (i.e., heat-up rate and frequency of ascertaining that the part had not become lodged).

The later calculations were done with a more detailed structural model.

The analytical methods and results for the most important cases con;.idered are discussed in subsequent sections of this report.

The results were obtained using the analytical model described in the next section.

Comparable to Vepco Loose Parts !1onitor "7s -3

e 2.

Analytical IMdel lised for Force-D::flection Comutations A flexibility nodel of the reactor vessel and intecnals was developed to compute forces and displaccrents resulting frou differential thermal growth between the reactor vessel and internals,' hydraulic flow forces and weight with a loose part ucdgcd between the bottcm plate of the energy absorber and the bottom head of the reactor ve.sel.

In addition to cross-sectional stiffnesses of the various sections of the reactor vessel and internals, the model includes estimates of the flexibilities introduced by local loading of the vessel, local loading of the internals (through one post of the energy absorber) and contact stiffnesses above and below the laose part (assuming a 5.2 square inch cont::ct area).

Since the area of the part is small (based on netal impact monitor results) relative to the size of the botttom plate of the energy absorber, it was assumcd that the part was lodged under one post of the absorber.

On this basis, the stiffness of the energy absorber uscd in the nodel was comprised of the stiffness of one post (and the associated cylinders) and the stiffness resulting from bending the bottom plate of the energy ab',orber with the con-servative assumption that the other three legs of the energy absorber were rigid.

l.' hen indicated by the force levels,'an approximate clastic-plastic force strcin curve was used to calculate energy absorber stiffness.

3.

Calculation Technique The analytical model was used to determine the forces and deformations resulting from a loose part becoming lodged between the bottom plate of the energy absorber and the lower head of the reactor vessel during heat-up or power escalation.

For each case considered, a part exactly fitting the gap was assumed to become lodged just before the transient was begun. The re-duction in the gap resulting from the change in temperature of the structure was estimated and the resulting deformation applied to the analytical model to determine the forces acting between the vessel and the internals through the loose part and to determine the not fcrce between the lower surface of the core barrel flange and the adjacent ledge of the reactor vessel.

If -4

n yielding of the energy absorber was indicated, this process was done itera-tively.

The force introduced by the loose part taken together with the weight, spring (fuel assembly and core barrel holddo.in springs) and hydraulic forces was used to determine net holddo.-in force that would exist at the core barrel flange.

To compute structure temperatures during heat-up transients, the temperature distribution through the vessel wall was calculated from data in the literature for a linear ter.perature increase of one surface of a plate.

Since the core barrel wall was found to be close to the fluid tenperature for the transients considered, it was conservatively assumed to be at the fluid temperature.

_4.

Force and Deformation Pesults for Several Transients The following assumptions were used in the calculations leading to the results for all cases listed below:

a.

l.cose part contact surface area cf 5.2 square inches (each side) b.

A ninimum value of core barrel holddown spring force c.

!!o yielding of the loose part or adjacent surfaces d.

The hydraulic forces were those for four pump flow at 700F or 5509F.

The limiting criteria was that the minimum ret force between the core barrel flange and vessel ledge is not less than 100,000 pounds.

Case 1:

Part becomes lodged at the beginning of a heat-up transient that starts at 700F.

1.

The minimum holddown force of 100,000 pounds is reached when the differential thermal growth reaches 0.038 inches.

2.

A temperature increase of 319F will result in an 0.038 inch relative growth if the heat-up rate is sufficiently slow that all structures are at the coolant temperature. -5 4

/

3.

A teuperature increase of 200F at a rate of 200F/ hour will result in a growth of 0.035 inches (i.e., the minirum hold-down force would be reached after.1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> at 200f/ hour).

4.

The force acting across the loose part after the transient listed in (3) and after all four pu.Tps have been shut down, will be approxii.utely 320,000 pounds (361,00') pounds, if the hydraulic force acting at the core barrel flango is conser-vatively assumed to act at the lower support plate).

Case 2:

Part becomes lodged at the beginning of a heat-up transient that starts near the full ttmperature, for this case the energy absorber yields before the minimum holddown deflection is reached.

As an example of the approximate results, if.

a heat-up rate of 370F/ hour occurs for one hour, the maximum forces on the vessel will be 270,000 pounds and the energy absorber deflection will be approximately 0.030 inches.

Case 3:

Part becomes lodged just after the escalation from zero power to full power is begun (assuming original design conditions);

1.

Relative thermal growth (assuming no thermal lags) of 0.053 inches occurs.

2.

The system forces across the loose rt are calculated to be just above the yield point of the energy absorber so that a force of approximately 220,000 pounds will exist across the loose part with four pumps in operation.

_S.

Reactor Vessel Stresses Using handbook formulas, it was determined that the combined primary, secondary and shear stress allowabic values for the reactor vessel are net when a load of 400,000 pounds is applied by a loose part over a one square inch area (larger allowable loads result if a larger area is used).

The analysis is based

. -6'

9

'on the load over this small area causing a local membrane stress and the membrane and shear stresses across the vessel wil being secondary or self-relieving.

Therefore, allouable stres.s limits of 40,000 psi (1.5 Sm) were used for the rembrane stress intensity and 80,000 psi (3.0 Sm) were used for the stress intensity developed from the con.bined primary, secondary and' shear stresses.

The initial high hearing stress from this displacenent (deformation) controlled loed over this small area was considered a local hertz contact stress and would be reduced to allowable code limits after a local deformation of the vessel clad on the order of 0.040 inches.

If the code allowable bearing stress limit of 1.55y (51,000 psi) must be met, the required contact for the load must be approximately G.5 square inches for a 400,000 pound load.

6.

Discussion With the assumptions used in the calculations, a heat-up rate of 200F/ hour with checks for loose part freedom once an hour can be used without reducing the core barrel holddown force to less than 100,000 pounds.

The results also indicate that higher heat-up rates and/or longer intervals beti.cen checks for looseness are possible if material properties and hydraulic

-forces at intermediate temperatures are used, if part of the heat-up is done with less than four pumps in operation, or if flexibility of the loose part is included.

The higher heating rates will cause increased plastic deformation of the energy absorber and somewhat higher forces on the vessel.

The differential thermal expansion that occurs during power escalation results in forces at the loose part that are slightly higher than the force required to yield the energy absorber.

Oscillatory stresses in the vessel and internals during normal operatior, with the part lodged and changes in stresses due to seismC' and loss of coolant events, have not been evaluated.

. -7

/

AlTACIW.Ef;T 3 S. G. TU3E Pl.UGS 111 TYPICAL 3-1.00P VESSEL

==

Introduction:==

There are t.;o possible consequences that could arise from the cxistence of steam generator tube plugs in the reactor vessel:

1.

Loose pieces irapacting upon the lower internals components.

2.

The plugs becoming ucdged between clearances and causing high forces during periods of relative thermal growth.

These plugs are approximately 3/4 inch in diameter x 6 inches long.

The first consequence was excmined using known or estimated flow velocities in the lower vessel plenum and was judged to not be a serious problem, provided the plugs are not left in indefinitely.

The plugs are too large to enter either the core region or the drive line area, and the chance of small pieces breaking off and nigrating upward is judged to be remote.

For the second consequence, several areas were identified where the possibility of tube plugs becoming ucdged was considered.

Of those studied, only one area was subsequently judged to have any real probability of occurring - and that is the area at the bottom of the vessel, where a clearance exists between the vessel and the_ secor dary core support base plate (see Figure 1).

A close clearance exists between the underside of the base plate, at its per-iphery, and the reacter -vessel.beite*a head...Thir.clearanceis1.00 inch (nom.)

cold. At the end of normal heatup, this gap has closed to 0.375 inch, and eventually stabilizes at 0.500 inch during steady state operatiu..

Thereforc, if stcm generator tube plugs (approximately 0.72 inch at solid end) were to become wedged between the base plate and vessel before or during heatup, the constriction against thermal growth wot.ld cause high forces to cr.ist.

Results of Analysis:

A study was performed to determine:

. -1

'l.

The nagnitude of forces produced by a number of wedued tube plugs, 2.

The possible conscr[uences of these forces upon the reactor vessel and internals.

A test was performed to deternine the forces that would exist at given deficctions.

For this test, the solid end of a tube plug was compressed between two flat 304 SS plates in a load nachine.

2 attached figure 2 indicates that forces of 49,000 lbs and 72,000 lbs. vould exist at deflections of 0.250 inch and 0.375 inch, respectivelv.

The 0.250 inch deflection repre-sents the remaining vertical groveth of the internals (relative to the vessel) once contact has been nade with the tube plug, while 0.375 inch is for the end of heatup condition.

Correcting for operating temperature reduces the above loads to 42,000 lbs. and 62,000 lbs., per wedged plug.

If nore than 6ne plug vere tiedged between the vessel and base plate, the load generated would increase accordingly.

Thus, i f in the worst case, all eleven tube plugs were wedged bencath the base nlate, the maximum theoretical load that could be generated would be 602,000 lbs., based upon the results of the i.es t.

To assess the possible consequences of these forces, the following areas were studied:

1.

Stresses in vessel bottom head 2.

l.oad capaci'y of secondary core support energy absorber 3.

Load capacity of internals hold down spring 4.

Longitudinal stress in vessel shell 5.

Stresses in internals core support, core barrel and core support columr.s Of these five areas, only the first three were found to be significantly affected. -2

Analysis of the secondary core support enercy absorber assanbly, indicates that assably co:nprising four energy absorber colinans will yield at forces in the range of 430,000 to 590,000 lbs., depending on the actual yield strength properties of the naterial used.

Any one column would yield at between 107,500 and 147,500 lbs.

Yielding of the energy absorber assembly would ii.: pair its ability to liuit the force produccd by a postulated core drop accident (core barrel failure) to within prescribed values.

Yielding of a single energy absorber colu:an (as a result of eccentric loading under

.aed to be as serious as the the base plate), while not desireble,

• v3 general yielding case.

The load capacity of the internals hold doun spring was emnined for tuo con-ditions - steady state operation (mechanical design flow) and the hot pump overspeed condition.

During steady state operation, a contact force of 512,000 lbs. exista, veen the core barrel flange and the vessel ledge (Figure 3), while a contact force of 396,000 lbs. exists during hot pur:p overspeed (Figure 4).

Current design practice is to consider 100,000 lbs.

of the contact force as margin against uncertainties, which leaves 412,000 lbs. and 20G,000 lbs. as reserve contact force during steady state operation and hot pump overspeed, respectively.

Any force acting upward through the base plate (due to vedged tube plugs) would act to reduce the reserve contact force describcd above.

If this con act force were overcome by the upward force of the wedged steam generator tube plugs, the consequences to the internals could be serious.

As contact is lost at the vessel core support ledge, flow through the resulting gap would tend to equalize the pressure above and belou the core barrel flange.

This in turn might cause the lower internals to slam doun upon the ledge, where the process could repeat.

==

Conclusions:==

The loads discussed above are sumarized in the following table.

. -3

/

.5 Co::Ponent Allowable 1.ond 1.

Vessel bottom.iead 450,000 lb. to 500,000 lb.

2.

Energy absorber assembly 430,000 lb. to 590,000 lb.

3.

Iloid down contact force 296,000 lb. to 412,000 lb.

Thus, during steady state operation, the limiting load that can be tolerated is 412,000 lbs.

If the hot pump overspeed condition is considered, the liraiting load reduces to 296,000 lbs.

The forces generated by wedged tube plugs are sur,tnarized below.

Allouable !!un.ber Condi tion Allouable Force of Plugs 9.80 plugs Steady state operation:

412,000 lb + 42,000 lb.

=

7.04 plugs llot pump overspeed:

2.96,000 lb. i 42,000 lb.

=

6.64 plugs '

End oa normal heatup:

412,000 lb. ?

62,000 lb.

=

lleatup + pump overspeed:

296,000 lb. + 62,000 lb.'

4.77 plugs

=

From the above, it can be seen that if a hot pump overspeed condition is con-sidered to occur at the end of a nc~nl heatep, T.9 r. ore than four (4) wedged steam generator tube plugs cun bc tolerated-beneath the energy absorber base plate.

If the above-postulated transient is not considered viable, then the number of wedged tube plugs that can be tolerated increases to six (6).

. -4 9

g

[~

%/

/

P-,,4 N

M..v '

m. -- _a._

x a,

n-

~

-- --- _.J..,: ; r-. -.

s pgj CN,_._.____._..._..

_ =_ _, _,; \\_.=, _... _ ; _

a _

{NN \\j.[

/

s

\\/

--. 5,. n,..~

\\

_u._'. = ]> \\ N {

. h_. _....

l

-E.

- c.1 -,.+3, g.

/

},sl wm) jf

-t w i

ria

- = m,=1

-J~

+4n 4;-

- - - ---- = >

=

\\ \\ \\ \\ \\. 'E I

,'h ggg

'~

lI l

n.yh--l

j. te g

_ ]

N x N s C "'

_ w -._..._.

. _a _

\\

'J

~

Y y

_.. _, _ _ =

nuenc v

'i,7~/

hfl.U n]ERGY h C S.

NN NT, 7

/)

e^se et^7 c

• h\\\\ b I

/

f_ _

._1

/,/

~

• M w k.....I _

.. C

/

f

,. _ ~.

l

.~

-_. ;.=

A.

- ; =a =,p.. - - -.-

y/

O

c.

am xx

.1-4

' j ',_

.--.- _J,

+ MT up.

i LN

.m

..; =3 _._ _.-

-:-ty.--d..-g,= -

~

anx.

l

E f

e

,h\\ x x.}g--//

j x

.u

=

D'-- i.I.,Ns NN,

"-+ m i a.

c. _.i ~._ s m

. = r-

,-m..---

a_

=

m._

~

z s 1

.;.. e.,>.'l f

-s

's'

/ /. V / )\\ /

t) 9

^

s

['[

'l)(f)

.{'

(f

{

O! (1 p) 8 5

(.

r co g

OO (p

i 2

(1 t'l q

t)

(

.{-

i s

i,.

t.

t (d

1 d t,

h d[

<7 y-

'.9 1

t O

8 s

N

..).

.(

.t

,>. l

, o.

e 1

0 s.

.)

2>

s

(

7l pQ f.,

x 3 o,

, -r tg

'\\'s 7

Q

'.1 U.)

JlE.

t

c.,5 e-

- }

\\

<t

p. s -

3 c

0 4..

te J

tj.'

[l (g)

. t..

a

-..s,.

s n

! sa e)

,N j.

d rz

(

q O o

31 i<

i

-(

(

t.

CJ g>

I g, 3

.?

'I o

,4

......... \\

.. _.. })

yo

-y-x i

w4 s

s

'n..

t

\\

. r, i

\\

a p

2

)

,1

\\.

.o 9

1 N

\\

~

.\\

~

..g

\\

q

.y \\

.3 o

\\

fil

\\

k

)

.N,

,.. \\ -

c-1

~-

1 --

1

=

O.

s

-t -

--o n F-O, Q

o n

9

%\\

e-(N p\\

,~

... ~.

'51'ETCll LllC ET

.t c,, i1.,i s U EST!!!G!!OUL E El E CT RIC CO RI'O R ATi o tt 6-9.7- / 7 D FI G VR E

'3'.

'.. ~.

/

l y

13 F?., co o 4f:

c c

,i o

it

,s:'

-sl.-m~~,,-..%,.,_~-,,%.-.__

f..

f.

f'xx'Q'd'N R'!M'$'s.s'N'x'4'%.s.SS-ik's'e'.'

'S x'

N i

q x' i g&M d,. N Nx '.

i,$N:ss$s 7ie:f';s t:q.@.S'..

Sy

e. 2< s.t.e ::':.N

. Lsc hww.~3 L,_____w Mff:;gc&- 18\\.

tw

/

i Nd d d 1(N'}

_._M..

f:' ;.z.$' 'N.sW.s.)!b.s)~

[slf.:, ca.

De sic N -

/ll'N s.

s

.sss s s ns

.s.x.

. s; w M,m.w.ms\\s.s'\\lSN'9sY.. '.! !,sdi os

.x

,d,' 3 s sn.. m 1

x

/.

4%00$

'N s\\'

5}2.,000'"

N'N jg.

s h'

's

\\s t

\\\\

k..\\ \\ N

\\

\\'

s s.-

lsss. N \\ N NN,N'.,\\ sN

'N N

v.,

s.

s

\\\\xs Nl,xN xs,N-t

o \\\\y t

\\\\

kD-

.,,,,, bn'g'_x'i W,'6 s -?

\\(',.\\>s\\x -

.\\

'?

f 1:.OV 'ER y=f;. L u, h

.\\

.:s 1in enum.s y-

-y

'KON.(

'N.'

.,3

' 5%g y,.. s x, y

e

'Ns s

N s.'s Ns iss'Oh'N'9 x

ss s 7

xx

.L..::..'.....:'...,.-

...7%ooo0.0.'eQ s.

[-

g..'TCil L!!E ET-SKt n.,,,

V E S T illG!! O U L E El. E C T I;! C C o i!!' Ol: A T I O tt

~

6-yy

7

.a..FI G U R.E.4 l

1. 3,nas, coo s

t.

\\

/,:!,;

'/, // {\\,NN's:s'-(s NlI.; ?s,g'N J,N's,Q'NN'.N x :

s'-

?

,/

t.....___,........_,_,..,-h_._

x

/

'2 N

/ cs q.NN'A's'M cy.Q -

f f~. $

, 'Y

~'N ; N'

.i ~

/

/.

=.~x

(/

I' w'W n n1,Ni, i

m N.

t l/a kD's..g w,y/.

tu t

)

'., s *

/,os,N's K,N s s',N.s...N.s N

Ns

\\-..

. \\

//

t

,$sNs s

sNs s

/ ! s NN g \\*ys.

':Nxx s A'. N NN'\\\\s. i \\ l LN -

. HOT P D Ag Ig I

//

\\ N's sh s s s g

,< w NN's'ce:q.Nx: :;.si rg

"/k ['~',8 ~'N'xb'$'$'s) Ib OVER S P.E E D 1

iQ's'l

'"$\\\\K's\\'g'"Jsl'i'91-4 3 ?.,o oo #

fj/ /j'Le

'/ \\/

3 Nss 4

'xxxi xx'x sx s N'i ce 3 cib 000 ~"*

\\%x., r%s i

s.

Ie.,

a 9 331 g A s 3~ ~

e. - )- -8'5'dN!'NkN i.

'\\

~

s W'&'%'%%%'Q'N

..um.)'$_bd$$['SS$.$

k 4z ovo # cuer)

Lowen zunwum.s n

s

\\'s's@A'$ 'Nd gQx N' 7

3 s, s.

/

\\

N s ss... s

.h.

Ns N gh'sh s

s s

~..

m

,m

N11ACll:!!1T 4 LOOjLI'AliTS MONITopi' G The Loose Parts Monitoring System, installed in North Anna Units 1 & 2, consists of a total of 10 t ransducers per unit, five active and five passive (in-stalled spares). There are 2 transducers on each Steam Generator, one on each nanway on the lower section of the generator, wit h one active and one passive.

There are 2 transducers in the reactor vessel flange area and 2 on the lower reactor vessel hemisphere.

One transducer in each location is active.

The basic uensitivity of the transducers is.05 f t.-lb of impact energy at the transducer location.

Of course, if the impact of a mass is not exactly at the transducer the signal vill be at tenuated as it travels through the materials in the coolant system and finally reaches the transducer.

After the impact noise le detected, it is then transmitted to the control room where it is amplified for readout and alarming functions.

During initial insca11ation and checkout of the system, the attenuated factors at various locations around t.he transducers is checked using a tool which imparts a known impact to the materials being tested.

The data from the initial test makes it possible for the system to be used to evaluate any future signals as to size and location of a loose piece of material in the Reactor Coolant System.

The impact energy of an object is basically dependent on its velocity and mass.

Therefore, calcuations vere nade to determine the minimum particle size that the Loose Parts Syston can detect in its present configuration. The following flow velocitics were calculated using the minimum allowable Reactor Coolant flow rate of 92,800 C.P.M.

North Anna Unit: 1 & 2 Flow Velocities Coolant Temp. - 5470F f

Coolant Pressure - 2235 PSIG i

Coolant Flow (GPM) - 92,800 CPM R. C. Pump Discharge - 27 1/2" I.D. - 50.18 ft/sec Reactor Outlet

- 29" I.D. - 45.14 ft/sec R._ C. Pump Suction

- 31" I.D. - 39.46 ft/sec -1

Calculations Done la Roch? ell. Loose Part: Monit orinn Syst en 'tanuf ac turer.

Coolant Tenp. - 547*F Coolant Pressure - 2235 PSIG Coolant flow (GPM) - 93,800 CPM Noninal Pipe Size - 30" 1.D. - 42.5 ft/sec The calculation of impact energy was based on a weight of 1 lb. t ravel ing at 42.5 ft/sec hitting the core barrel.

An object of this size will generate an impact equivalent to 28.3 ft-lb.

Initial test data indicates that, based on trans-ducer location and a t tenua t ed. f actors, the signal would he at tenuated by a maximum of 10 lb. or by a factor of 3.16.

The resultant si;;nal at the transducer would therefore be 28.3 fr-lb or 8.95 ft-lb. which is well above the alarm setting 3.16 of 0.5ft-lb.

Since the flow velocity in the coolant system is const. ant a correlation can be made to show the smallest mass that will cause an alarm.

1 lb.

x lb.

x lb. =.055 lb. which 8.95 ft-lb

.5 ft-lb is < 102.

1 oz. will impart.56 ft-lb. of energy; therefore, we feel that a particle of this weight could be detected and that it is feasible that it would travel at the same velocity as the coolant.

Since the start-ul test data for North Anna shows that a flow rate of approximately 105,000 CPM exists in the coolant loops, we feel even more confident in naking this statement.

Of course it is not realistic to assume that. a mass of 1 lb. would travel at coolant flou velocities; however, it would only have to travel at 1/4 of those velocit les to generate enough impact energy to cause alarms.

The Loose Parts Manitoring System functional capability can La verified during operation in two ways.

First of all, transducer operation may be checked by comparing initial background noise profiles with present profiles using the vibration made of operation of the system.

In general, background noise in a mechanical system tends to increase with time; therefore, a noticeabic decrease in background noise vould warrant further investigation into the system functional capaollity. On the other hand, a rapid increase in background noise level, even below alarm settings, would also be indicative of a monitoring system problem or possibly a mechanical component problem.

In either case a periodic check of each channel should reveal such occurrences.

The second method would be used to determine that the electronics were working properly using built in test signals.

By using frequent surveillance intervals, we believe that the Loose ' arts Monitoring Systen can provide us with an early indication of metallic particles i<

the coolant system that are as small as 1 oz. in weight. -2

AfiALYTICAL EVALUATIO:lS_

1.

IllTI:0DUCl10:1 In ticis section will be presented results of analytical evaluations conducted to identify the cause of cracting observed in Unit 2 clbow C and analytical predictions of subsequent crack behavior had operation been continued.

A brief sumary of pertinent facts and analytical results follows Apprcpriate details are presented in Figures and Tables.

A.

Observations 1.

Pattern The cracking observed in the splitter plate is characterized dominantly by two large cracks located at opposite ends of the plate (leading and trailing edges) extending approximately 19 inches along the plate (leading edge) and 15 inches along the plate (trailing edge). The two cracts are diagonally opposite one ano'ther in the plate.

Both cracks started in the welds, follow the weld for about 10" and then hoch in towards the centet; of the plate and towards each other.

2.

liodes of failure Investicated The following causes of craciling wel e investigated, Crack extension of an initial flaw by application of a large load a.

!{o large, abnormal (non-cyclic) load source could be identified nor could any evidence of a pre-existing flaw be found.

b.

Stress corrosion cra ding, fio evidence of stress corrosion cracting was found.

c.

Material properties.

Results of material investigations revealed no abnormal or unacceptable material properties.

d.

L'elding practice.

fio abnormalities in welding practice'were uncovered.

AE-1

c.

aticue.

fractographic analysis of the leading edge crack at.

r several locations clearly established that cracking occurred by fatigue.

St.riatica i:.easurements were in the range of 1 - 5 micro-inches per cycle and yield an estinated range of cycles to failure of 110,000 to 530,003 cycles.

TMse results support ruling out causes a, b, c and d, above and suggest a vibration source of loading stenaing f rou either structural induced or flou-induc ed vibrations.

Striation measurements indicate high stress levels ranging from 20 to 60 ksi.

(See Table 1).

D.

Sources of Fatique toadinq 1.

Structural Induced Vibrations Experimental vibrations measurements were assumed to apply to the

!! orth AnnD Unit //2 loop C and vibration and stress analysis results on the splitter plate show that the cracking was not due to structural induced vibrations because of very low vibration stress mplitudes (1 10 psi).

2.

flow Induced Vibrations a.

Itatural frecuency of flow splitter plate. The first two nate al frequencies of the flow splitter plate were determined by finite element analysi, to be 149.15 and 148.18 cycles /sec. resulting from the first two closely spaced modes, l'iode shapes are shown in figures 1 and 2.

b.

t!alur_al frecuency of elbow.

The natural frequency of the elbow is deterr ~,ed to be much less than that of the splitter plate.

Estimates are 57 cycles /sec, first mode, c.

forcing f:ecu"ncy due to [3mp induced vibrations.

Experimental vibration measurements made show that the predominant frequencies related to pump operation aro 20 Itz,140 llz and 273 112.

The first frequency is associated with the shaft rotation at 1200 rpm.

The second and third frequencies are associated with the blade (7) passing frequencies.

The second frequency (140 liz) suggests a possible resonance of the plate (148 liz), however as mentioned AE-2

above stress amplitudes were insufficient to cause the observed cracking.

d.

Excitation f re3uency due to vortex sheddino.

Considerations of elbow and plate geometry and flow and fluid conditions result in excitation frequencies of 98 Itz, (lower bound) and 136 Hz, (upper bound), again indicating possible resonance of the splitter plate.

pesults of the modal analysis (Figures 3 and 4) show that (1) sufficient stress amplification (4 32 to 4 72 ksi) could occur to cause the observed cracking (See Table 2); (2) the stress distribution in the plate is consistent with the cracking pat. tern observed in elbow C, namely, the cracks start at either leading or trailing edges, follow the. weld and t. hen turn into the plate center-(Figures 3 and 4);

(3) af t'er. progression of cracking the natural frequency o, the plate would drop to 64 Itz and a reduction in stress amplification, clininat.ing further crack extension (See figures 5 and 6).

Figures 5 and 6 also show that the c' rack would turn towards the 1

center of the plate.

t C.

Sur:na ry_

The information presented herein shows that a postulated resonance in the flow splitter plate induced by fluid flow is a plausible explanation of the observed cracking in the splitter of Unit 2, f

8 and that this cracking would not be expected to progress.

Further-rore, because of the low natural frequency of the elbow, and the geometry of the cracks (hooking into the plate), this mechanism is not expected to have any effect on the pipe wall.

This mechanism also indicates that the cracking observed occurred over a period of 12 to 60 minutes, significantly less than the time of operation in Unit #2 (17 days) and considerably less th.n the operating period of Unit !),1 year.

liad cracking occurred in Unit fl, (1) it would have been detected, and (2) it would have quickly progressed and stopped due to the drop in natural frequency of the plate and thus eliminating the assumed source of loading.

AE-3

Irrespective of the loading source, given the plate geornetry and constraints, the maximum stresses can only occur in the diametral direction.

Therefore the fracture rnode cannot be other than tiie one observed in North Anna Unit. i"2, l. cop C.

t

• e S

S t

i 1

l e

4 AE-4

WESTINGilOUSE I",*.'It SYSTET.1S DIVISION

"^"

I Tr\\BLe i Frm G ut Di ric es E m ur, m ts E>i Ei (\\gaA Vad [2.

".. ". ('.%c rv.// 4lnle g

<u nnin onii

< n,. o. n y unii e n,m n v g-(;['t'rRC[4 O/pf (-['f

<VM (hrn a y,

eh

.. s i o n.-

io.

catc.ru).

V G.Lt h \\ " 2 At/;)

,$i 49_

(4 7 Ohf,,fjl5f(,~.

i.

o 5

b-IN y# e a

~

s D

0 O

f a

n q

,)

y 1 ('

i s b c,n.

aC. v c-1

.N I

10 1A t

o b

$s 9,

9 9

f f.

f; h(

M d

d a

m G~

d ti ti r)

+1 1-(

+{

E It v n

^

E

-y es a

L' E hl 5

6 o

ty c

3~

F) to 4

- l-(f q

~

.. n op 5

n o

o r

i

<<j

4

+

in tn

+

v Q

't in Ee n

to t)s

.n D

S

.o o

.o_

q o

o u

'o s

0-U o

o 0

o o

d8

[1.-

o o

o o

s-N s

N o

o c

o

.o d

O N

?

IO

$n 7

d M

._l

'9 s

o

/

FM 5

If)

~ ' ;i

~

,a J

O t ')

H-I tn tn.

b O

Nd "I

O 15 0

tD

.J 8

d 00

_o m,

g-g to s o.

. o' l o.

lo n

g 7

o

~.

y.

y.

y

%<G X

y y

to to g

u s

^uinon oni cuco uv unu cueco.nv onit nt v.

nr v.

NO.

oATE WL f.l l N G H O U Si: ( Of LM 1,5? DC

WESTINGilOUSE IT!!! SYST Ef.13 DIVISIOTJ

" '2'~

~

TiI((

'i'ii BL C.

'2-fm p.tJ 6,JT Sr R C. '.> s Esitrm1E 5

(

Kiaic;6

,,m r,7; c. n,-

t k(nh (*atits Urs LT 'lt g, 3, (>,.g,g,,. f 9hl 3<3

y. 4 9% j _,,, <c, c,- ;, n,afp o

O g,3 mon u pstirag.

~

UltOUs'

~

V G est<\\- \\ W 1 V g/j;

,, g _

,,77.Q,m jgm g

'(#-

C/ LL. f 40 YoRTex

ShebDwc, Irvau ceD Re w nuco 0 W%Id ActT b'PGDcem GMT 3

i

- pa2 i

2[

7

-- FTT '

i mi s,

_i M\\cG.E

\\b u[!!E ff

, oo oo Wj

[' n F Lo to Dcru ny f6 G cre ntt rs Ltre D mr3.s 3 1 bod /A qt5 i

cd :

Tte T mcK u c,1. r, t N

=

i, ou

(.

d:

L &J Gvth ce Tte ts y

3t 7

'S

=

STgc7) R L nt u 4 o.2 n o ttw 7

v. c..

o - o /, o. 0 7-Fr-o c r< ON cp cc.vtict),t.

O Y

=-

32 4 m PirJ 6 3

l o o ovig xC c63) Y Si Cg

- l

.. o, o g y

q.

7%*D l ~,)

f i]I

. A LS X (. i - ) g

__J -

f i - 11 9 F s sa s'M E Lemo eJT Cnt. coLrrtt aN S C, toc 0.o r4 UN \\T DAO A<<AL\\pl5 Gtv6

- 6 o 2.

C, k f.T

13 C lC SI

.6

% U#1 ref.T_,1I 9

.o t

{;gij h hb7.(d b llI w

nry s TcM 5 cia merc n re cv t>

(, /-)

OtW mc0

• l b R

AU llf ull (DAli Cit P;'O. U Y DATL Cte r;*D, U Y DAll (q [ y, og y, NO.

DATE WELilNGitOur., I Otat LL:'13C

WI.ST ING!!OUSE l'V.'M SYS~l EMS DIVISION

"^"

'T 0%BL C 2

(Croidtt,1v'LD) or t n y, D, it Y L'All

. H A 'l t ur g;A g e

.s.Ct.ter,<~y'/al71 <f.3,z'_,,,'T[3/pf fy,$)

At>18 40f t tsA t t

• • lit 3 Jf L 1 gj Nenn [ht, Irs U./n /1 f it[f~f.o.

., s u i ts e 4.0 C A L C. f t).

b bib" kU I

//

A[

((7p

/[ '{

,' {r.

bbbbb-OE b

(') L'h hi b i (I'!b

'l = a, a 't cow cae swu s nn

[Le To 4

36 h.5I c~

U f

~

in A Ti~

AtilHOH 13 Alt C H F.'D. ti Y DAIL Cit MD. LA Y Rt V.

f1C V.

rJO.

DATC WL SlaNGitOUSE. f OHM %213C

WESTINGilOUCE ITill SYSTEf.1", D! VISION Ttil D Y$0 YT (() D O R',

00(h f 0 } *

• E' l.h l$ l.

J

' 4803t Cl

/ tJ t e nsli smye g,, i, g 3 gi y g,,g i g g,,. g, g y tJUf*1 h fb'r'/5 l%>IO l

,rg }, {ig g g g'/ Qh d )'f C A L L. f J O.

I 11,1 f JU.

,pt(su, ~ ~

• I8-Y Yb {bilA Apl

.pj g.(

U t) r 9

T%

/

N

\\

6

/

l l

/

\\

\\

/

/

/

/

Q' (

/

,l/

l

/

,r/

\\

/

/

/

c.

I

/

j

/f I

/

/

/

/

/

/

/

/

h v:nE l

(4

\\,.

)

/

s

/

/

(

/

/

q\\

/

>/

w AU1 Hof n OdTt' C e t t,'ti. It Y DAll

(. i t K

• D Il Y

[> A l l g t g' V, rig v, NO, DAl(

WLSTlFK.140VSt. I oltu M713c

'. -)

WESTifJGilOUSE PWH SYSTEMS DIVISION Ibe d Q(

2

{ b @ #3/l h (O d)(!,

bI\\[\\hb ' T l h O '.#.

j Of

• 'If D >l t i Au l 6t(fit f aAll 6 s i t.* [i, es Y t >4I L C s t r'.'!)

li v t s /, g t (Q Q, Ur.l Il O h l'l

, }; ;

y,q l Al h

$ $ls1 \\ \\\\

%IISI.

c o(l 1.,

,.$e R

iO

(. /i t C, f 4 0, 6Ill id O,

. it O Ul*

s(

• fyl }

f/'['{}ll(ff~f', / l f, e

t

/

N

/

x

'/

\\

/

'N

/

N,

/

/

N

/

.i. -

.g[-. 3. w'?

-'/

/

/

l a-.

/

y'

/

v

/

/

/

/

/

/

/

/

j gN,\\

/

/

97

/

/

/

\\

j ix

\\

k

/

sy

/

\\

/

(

/

/

\\

,.*g AulttOf t DAll Cit h* D. Ll Y t>All Celt.*D.46Y DAll g g g, y,

,gg y, NO.

11 A l l~

Wi hllf4(ittOUSL f Of tM L';71;sC 6 '

b b

3

y 0

8 0

0 0

9 0

(.

1..

2 li'.

• 5 4

il ea i

(v.v )

ToF T' I E ll.

I2' Co#D '#I"

' 'tC S s Di * " ' 'a g.,

yg,y to,; G iqu o t e m' l

m - - e n, e <, 4' mu, ~ - -

<, ga l

3,ac o,i a e

3, s,

4*

' j)3 [

(

' >o

'yp Q,

/r.

43

• i>.,,,g;Tn

'jp h,t A

[<. o s

e 3

Q> f(5) l I '

(6)

I

\\(,

('

(

ae i,t',s O,n.

(%

,a

.,,s i

(1)

" ',Q (8 3) gg d

(O' )

l'f;Il

'i o

-nf

(" )

( il (s s )

3 j

h

([,3 l' N I

d f i th y jr (s0 1

gil i

I s

w

\\

q eqont. nuo cepas tTE FiGucm 3

Fi6,sr cn oi3nt sr o.c.ss e.r

(-

1

~

3

.ee a.am e p W998 9

Gr*'

A us gy

.O e

e 5

e e

i'-

p, n;

T

.s a 6nuot.mL PP GCdcW6 h""

p s ) _. L o ou racTrf L T6 P BC"ba

li p

n

.,' g

_m

/

I, E

i, i 'a,% p W

r-3:

s

%p n,T9-)

A?j%,s s

g e

o > #,,,h 'h,xt ' @

l g W,,,, h, %

%, 4 ) y y #

L ss, 4,

e

'tdt

,7.atc3

,o' l

i W"\\

E

_e_s____

Rcone 4

sco a s m oont sto.e ue s

- ()

O

(.

E-t!

..y

-e-8 5j

( y. y ) - L ord C, r g ()gg r 4.p p gg g

,J j

K'X Din rod TG A L 10P 6)CrJCtert. SW C S 5

[

ll'

/.>'3 l/M i

e l

o n'>%

I I

('

l I

^ -)

'* -d g

1 T

L ')

(S k

t 4

Crscn t

t

,o ol

( 'd' L'

I\\

k' 6

/\\

\\

g\\

L' 3

(

! unt1L91-s e) l'a

-b 4

/

( >,,3 '/ d (24

(,

l

-r 24 g

~lh W"l)

Vri W

0 8,l 7,

s u.

t ot1

.s f

(s0 e,)

o V31 6

(<>l

/

I

$3) g'TIv ['O

( ol 9

' 2-I)

/ f /

(3 )

j/

CR(\\Q r

~~

4)

L'O

/////I 17f

?

une. e e m suececs n n.c u.uw &~o D PPoS !TE i

1 Ft Goe E 6 SC CoNo in o DAL.

Sm s s e S ; -f= G4 -3 \\ nz e

f e

0

(.. i..

h.,

• I l

J k

D'Y)

Lo N G11b Dins AL ro e Ecuoirrc Srrto 5

.4 7t? ir4 Der 6 STc0 b5 P

r Y,y.

9s hrn CTA ht lI r

I'lhD

-^-

p l

I,I O

I l

U g

i a

A,,

t s

f f I (is c(M'

- u '", \\f(' (A-

.)

9 1f wNe)4 a, N c ',p J

i N

,,,, a,,,, o e ', #g e

l.

yr

. ;,3 g,e a

e, ve t.m - {i)[, /L >

1 I,

Gh CO, g

U

//

g,, 's)

US)

/

v 2.)

( 'i' (o )

(.'> )

j //

C,(LD

}

T>

co

z

,,p7 tm

</77F '

1 i?

I i

Ncat : b rrnm sac-c tea ra c enoAt (wo me,ru FiGoce 6

Rasr moont sm e s.s e.c, -f = c.4 a s in.

\\

a Os

\\

I e

4 UL1pMD: llc f.Xf.".l!lAT 10:1 As a result of the cracks observed on Unit 2, an ultrasonic examination (UT) procedure was developed to enable a voluruetric examination to be made.

The objective of this procedure was to exaiaine the splitter plate to elbow neid from the cutside dic.e tcr (0.D.) of the pipc.

A straight beam technicp:e utilizing a 1" transducer and a 1 negahert? signal was employed.

The piccedure was used on loop C of Unit 2 which experienced the cracks.

This test provided excellent results and clearly de:nonstratmd that cracks of the nature found on Unit 2, would he detected by the procedure.

Af ter the procedure was developvo, a UT examinatien of each splitter elbo.

on Unit 1 tras perforraed.

As a result of this examination it was clear that the flow splitter was structurally sound and could perfora its intended function.

!!o defects as found on Unit 2 existed.

• It should be noted that tha procedure employed would detect any crach which occtnied in the ueld area cod apbroxinately 4 inches deeper into

.the splitter plate (10 iw hes f ron 0.D. ).

The only stipulation is that the crack surface rc.ust be perpendicular to the transducer. As shown in the analytical evaluations the only cracks that, would be anticipated to occur in the flow splitter would be identical to those observed in Unit 2, and would be detectable by the UT r;rocedure.

It should be noted that on Loop B of Unit.1, two reflectors were found as a result of the ultrasonic examination.

The first reflector was approximately 22" from the leading edge from the splitter plate and was approximately three inches in length.

The second reflector was approximately two inches fro:n the previous reflector and was 1/4 inch in length.

These reflectors occurred in the weld icetal at the junction of the longitudinal and lateral ueld of the splitter plates.

It was obvious from our inspection that these reflectors were not cf the type found in Unit 2.

Based on their location it could be expected that these reflectors UE-1

\\.

would be slag inclusion, lack of fusion, or some other type of weld imperfection due to the high amount of 1: eld material depor,ited in this particular area.

It should be noted that if a crack occurred in this particular area it would r t cause the type of failure which resulted in Unit 2 because of its location (approximately 22 inches from the leading edge of the splitter plate).

4 e

t

\\

e O

e D

0 e

o O

e UE-2

.