ML19224D742

Transcript of 790605 Public Meeting in Washington,Dc Re Mark I Containment.Pp 1-38.Viewgraphs & Photographs Encl

ML19224D742
Person / Time
Issue date:	06/05/1979
From:	Bradford P, Gilinsky V, Hendrie J NRC COMMISSION (OCM)
To:
References
REF-10CFR9.7 NUDOCS 7907160246
Download: ML19224D742 (42)

Text

..

c, %

r

,/

l NUCLE AR REGUL ATO RY COMMISSION r

IN THE MATTER OF:

PUBLIC MEETING-BRIEFING BY NRC ON MARK I CONTAINMENT

(

Piece -. Washington, D. C.

Date.

Tuesday, 5 June 1979 pages 1 - 38

-c.cnen.:

( ~

(2C2) 047 3700 ACE FEDERAL REPORTERS.DiC.

OffcialReponen

.t.LL Ner-h C :itel Street Weshington, D.C. OCC01

)7 t- (-Ji )ii-} I NATICNWIC E COVERACE - C ALLY l

,. i

'/.'

'?

V

V 9 0716 02I[

~

1

CR5211,

/

DISCLADER This is an uno,,IC4 a1 - anscriPt of a meeting o# the United States

-~

• he sdav, 5 June 1979 Nuclear Reculatory Con:nission held on

^

Commissions s o-*4ces at 1717 h S-ee.,

ashington, D. C.

The i

and observation.

mwis t anscr1P-meeting was open to public a**e.'.danca 0

edi*ed, and it may contain has not been reviewed, correcte,,

inaccuracies.

The. transcript is intended sole,

-j, {"ene-alinformational e -

g ee ormal e

purposes.

As provided by 10 C:R 9 ers disc ssed.

Expressions or informal record of decision o-nece,7sa.41y reflect final of opinion in this transcript do no determinations or beliefs.

No plea n

er paper may be filed

~

v

~

as the resu3.t of or addressed with the Co==ission in any r.oc to any statenent or a g'.rnen co_.,- --> -. erein, except as the Con nissicn may author:.ze-356 CT

i i

2 c e-, /Q i

w I

CR521 R UNITED STATES OF A> ERICA l

Madelen 2l NUCLEAR REGULATORY COM31 SSION l

i 3i i

l i

1. l PUBLIC MEETING I

5l 3RIEFING SY NRC ON MARK I CONTAIN>ENT I

6i 7l Rocm 1130 1717 II S treet,

N.W.

B Washington, D.C.

i 9

Tuesday, 5 June 1979 10 The Cc=.ission met, pursuant to notice, at 9:55 a.m.

i I

11 '

BEFORE:

12 DR. JOSEPH M.

HENDRIE, Chairman 13 VICTOR GILINS G, Comatissioner 14 PETER A.

BRADFORD, Ccmmissioner 15 JOFN F. AEIARNE, Ccrmissioner i

16 MESSRS. EISENHUT, GRIMES, AND GOSSICK i

17 18 1

19 1

20.

i etIi

• 1 i i

• a.

4-

. ;.e.r.i nemmn. inc. j 25

.i it i

• J

/

3. "

b. ' 'J

',, }

)

t

2a I

? ROCEEDINGS 5210 delon mpbl CEAIRMAN HENDRIE:

Okay.

We'll get s tarted.

2 1

i 1

3i The Commission meeting this morning is for a brier-j i

4 ing on the Mark I containment situation.

i 5

This is a matter of some longstanding interest to l

f i

6 the Commission and its staff.

It's been -- I'd hate to say I

t I

how long ago -- but, what, a couple of years that some of the l

7 8

testing on the Mark III containment concept indicated we'd I

9 better go back and look at some of the other cret sure suc. e. res-i i

sion containment designs with re* gard to dynamic forces.

And l

10 !

I 11 there has been a substantial program for each of the contain-12 pressure suppression containment types, the Mark I ment 13 short-term program and there's also a Mark I long-term program.

l 14 I believe what we'd like to discuss this morninc in.

i 1

particular are the aspects of some tests which I guess at this l 15 i

i 6

point go back a few months, three to s?.x months, out of the 17 three dimensional testing in which the three dimensional i

l *'

results appeared to lie a little higher in terms of loads, 19 l and the two dir.nsional results which were of particular i

20 '

since the intuitions had been that it would be the interest 21 other way arour.d.

pl Lee, please go ahead.

n3 :I MR. GOSSICK:

I'll ask Mr. Eisenhut to go ahead.

1 24 ;

v2

= 5 = yw m.

m u, <1

..o,,.

-. ~.

a-e.xers

,oor m. sne.

i 25,

'4 hat we'd 1;ke :: do is -- You sert of gave ;s a

,i 3r7

(,;,g'.

,i s.

a

2b l

l mpb2 gcod summary and went through what we'd like to exc. lain today.

I 2

This subject was the subject of a couple of memos 3

l we sent to the Commission, one last August and one in I

1. l I

Februarv of this year, where we pointed out the difference i

5 between the 2-D and 3-D test results coming out of the test I

6' I

program.

l 7

We would like to just go through a summary, to a

refresh everyone's memory we will go through a short summary i

9 of the background of where we've been on the Mar'

,. _ o gr am,

10 I trying to get enough background so we can see actually what i

11 we're discussing.

i 12 Chris Grimes, who is the test manager running this 13 task for the NRC, will be going through th,e briefing.

14 MR. GRIMES:

Thank you, Darryl.

i

'S Before I get into the discussion of -he testing 16 program I would like to refresh your memory since out last i

17 meeting on Mark I in May of 1978 and go through a brief i

1 *- i i,

discussion of the suppression pool hydrodynamic loads and i

19 the history of the Mark I.

1

^0 (Slide.)

y :l I will go en to discuss the Livermore testing program, the results from tha: testing program, the conclusions that we drew, and then give you a status reper: on the long-24 i

term trocram.

ics-ETCef at ReOCr*tr1,Inc.,

~

SC 'l Iq,4;n.s

~~

\\--amn 1

i t

i

, r, -

3 I

t i

mo.b3 The first slide shows the cut-away of the Mark I

~

1 t

l containment.

There are 22 operating Mark I units, they have l

2 t

I 3-three units under construction.

i The significant features of a Mark I are the 4

f drywell connected to a wetwell by vent system.

During the 5i i

6

'60s the Mark I was developed from testing in the early I

Humboldt-Sedega Bay test facilities where there was no sig-7 1

I nificant hydrodynamic loads observed.

And as Commissioner I

8 i.

i Hendrie noted, testing of the Mark III identified these loading, 9

i e

10 I I

considerations.

?

i And in early 1975 we requested the formal Mark I l

11 12 reassessment and the Owners Group was formed and they elected to go.into a short-term and long-term program.

In early 1976 preliminary analysis of the Vermont-Yankee Plant predteted uplift of the torus due to the pool l

15 l

16 swell loads.

It was at that time that we focused on the down-i 17 '!, ward and u ward loadinc. comc.onents on the torus.

i 1 ~: 1 j

(Slide.)

i 19 'l The next slide shows the --

20 CHAIFF.AN HINDRII:

I must say about the last one i

'1' that the quality of the illustrations has improved since long

,,4

,1 years ago when I used to --

6 1

73 C 4

MR. EISINEUT:

Well, if we do this enough tines

^

i 24 we'll get better.

4C2-e9Cef 31 Af00fMf t, !9C.

9C i MR. GRIMES.

We have since 1975 develored

== l n

l 4

k

(

e

.) J J L

d

4 l

i I

1 mpb4 good illustrations.

2 (Laughter.)

The major pos1 dynamic loads are the download, v'ich I

i occurs shortly after the vents clear in the torus, the upload 4-1 i

i 5

which follows that, pool impact on the rinc header and the 6

structures above the pool, and the LOCA steam condensation i

7 and safety relief valve steam condensation which cause 8

oscillatory loads on the structures submerged below the I

1 9

cool and the torus walls.

i l

i 10 6 The only load that we'll be discussing today is the 11 pool swell upload, since it is the one that was principally i

12 affected by these three dimensional effects conclusions.

(Slide.)

14 The upload occurs after the bubbles have begun i

1 ~*

to form and a ligament of water ahnve the bubbles begins to 16 compress the air space.

The downward component inside the 17 l bubble then droo.s as flow rate into the bubble droo.s, and the i,

i 18 result is a net cpload on the torus.

19 In the short-term program the only basis that we 20 had to understand these pool swell loads were 1/12th scale 21 '1 ccs:s conducted by General Ilectric, plus scme analysis

,,.i

q they had done using simplified computer mcdels.

h that

~~ j

n addition tc that testing program for the 24 1 icng-tern program we now also have a 1/4 scale tes: -- :

sm 5?Cef an ReOOr*ffs. !nc.

,C j

forget tc mention this:

1/12th scale tes; was a 2-D slab,

~~

i

{ it l l

,),)cb L'

aa

I I

i i

t i

neb 5 one downcomer pair with an average unit cell width.

f I

o This is a picture of the General Electric 1/4 scale 3l f acility which is the primary basis for load definition test i

i 4'

We also have a testing program i

l in the long-term program.

i 5

from the Electric Power Research Institute in Palo Alto which I,

6'I is a 1/12th scale three dimensional test.

It is a 90 degree

-l taquivalent straight cylindar, that is the toratal shape was l

7 I

a straightened out.

And they also have developed some verv i

9 sophisticated computer modelinc at the Jaycore Company in t

10 :

j California.

i I

11 And the Staff.has the Lawrence Livermore Laboratory 12 1/5th scale test, which has both 2-D and 3-D.

Also Livermore p'

has developed some computer modeling of the bubble formation i

I 1/

j and pool s.* ell processes.

j 15 !

(Slide.)

i 16 i

COMMISSIONER AHEARNE:

Can we go back to that

{

'l d/

picture?

18 MR. FISENHUT:

The picture of the test?

19 )

COMMISSICMER AHEAFNE:

Yes.

20 MR. EISEUFUT:

Would you go back to the last slide?

A9s (Slide.)

c1 s

,-44 j COMMISSIONER AHEARNE:

~4 hat is that black Object

,1 sert Of off at an angld?

Il b' MR. GRIMEST I believe that was a transducer Cap scs Feceras A eoor+ers. Inc.

t SC.

-.a t was left On prior to O'ne test, and it was OOrn Off by ~Pe

-I i

6 l

l, I

?

m o. b 6 bubble formation.

That is not typical of the plant.

They 2

1 don't usually do that.

t 3li COMMISSIONER AHEARNE:

I wouldn't think so.

l l

l 4'l

( Sl;.de. )

l.

MR. GRIMES:

The short-term program load was derived, 5

i 61 from 1/12th scale tests using a most probable load as I said, i

4 7

An intent of the short-term program was to demon-I concept.

I I

i 8

strate sufficient capability in the structure that would stand iI 9

full hydrodynamic loads and so they used their most probable i,

10 1 load concept other than a bounding load concept.

i 11 To assure sufficient margin in the structure, a

.i f actor of safety of two was used with this loading tiansient.

j 12 l

i 13 Plant unique loads were developed from the 1/12th scale test j

l by using sensitivity parameters.

However, because this was t

15 a most probable load concept, we were concerned about the 16 i

uncertainty in the load and what effects it might have.

17 (Slide.)

18 So a sensitivity case was derived, and the i

i 19 sensitivity was developed by reviewing the data to deter-20 mine what amount of un ertainty there was.

We concluded that there was a 50 percent uncertainty in the load.

With an su increased load we correspondingly decrease the allowable i

23,i facecr of safety simply to prove that there was sufficien 24 i capability in the structure to take this bc

r. ding load.

3.,w,, n worms. mc.

u.

J So with a higher loading w:2 use a factor of safety i

)IbJ b, b] g J

7 I

I 1

of one to ensure continued integrity.

However there is a non-mob 7 i

i 2!

linear response of the torus to upload, and therefore the l

1 I

3'i Cwners felt compelled to reassess possible conservatisms in thei l

i load because they did have to make certain modifications as l'

4I t

5' a result of their short-term program analysis, and they didn't I

6:

to make any more modifications than they absolutely had want i

to because they felt tr : the long-term program would identify l, 8

absolutely necessary modifications.

anv i

9 COMMISSIONER GILINSKY:

What is a factor of safety 10 i l

of one?

MR. GRIMES :

A factor of safety of one is the e

12 1 i;

ultimate strength, I believe.

I 13 MR. GOSSICK:

Strength of failure.

14 MR. EISENHUT:

Turn it around again.

What was done in the short-term program, as you'll recall, is the kase case, 16 the most probable load was computed.

And then over and above 7,'

that there was a factor of safety of two that was required.

18 For the long-term program we wanted to define the load better i

19 and required a factor of safety of four, the standard way i

of doing it.

We did a sensitivity case, and this in how we 6

i n;.

really think it's going to ccme cut, with a factor of safe.:y

_ two.

0:

i 0

3, We did a sensitivity case where we cranked up the o

,,1 2

loads and said 'Even if you increase the loads by 50 percent, i

sc..:.cerai 4.wms, inc. r y

.. a? vou still -- it ccmes cut equivalent to what this structure can 1,

f

]

[) a b k.,

s

S i

i I

1 mo.b8 actually take.'

1 2l COMMISSICNER AHEARNE:

A factor of safety of one I

l 3'

1 is not a factor of safety.

i i

I i

4' i

MR. EISENHUT:

That's rigat.

It's a factor of 5

t safety of 1.001.

I i

l 61 (Laughter.)

MR. GRIMES:

I said it wrong' too.

We said that t

8, the factor of safety has to be less than or ecual to.

So you 9

j approach one.

But, you know, it wasn' t too desirable to try i

t t

10 I and crank it so that you get right to that limit.

4 MR. EISECHUT:

The other piece of this is in going through this when you' re talking about a f actor of safety,,

12 13 it's the limiting component.

So there may.oe one piece of the,

s 1

14 facility that can -- that has a f actor of safety of two, or i

15 one piece that would compute down to the load's ecual to the 16 capability.

But most of the f ac;1ity has a much larger factor 17 of safetv..

In most plants it's Just one piece of the plant.

d IS I I

MR. SNYDER:

Can you mr.ke a general statement as 19 to where the limiting compenent is?

Is it the downcc=er?

20 MR. GRIMES:

It's a c.lant sr.ecific kind of c.roblem.

3.,I On these loads most cf the limiting ccmponents are in the n,

support structures, columns, saddles, whatever.

i 9,

_.-.4nv....,, :

. :..ninx. you can,t sa' fcr certain, M.,..

_es i

i

, I, but v.cu can certainly make an inference generally that it's s-sc..s.cers a.ecrmi. mc.

i

..e the sur. e. cres, sc=ething to do with supper:s.

Scme plants have o

i s *

=j..).)

\\'

l

9

\\

\\

l I

l.

i I

mpb9 basically a saddle sitting under the torus.

Some plants I

I i.

i l

al

^!

needed to do things to the actual columns.

Or in some cases, I

l if it waa 3'ust sitting on a column, they had to strap it down 3

s i

1 so that it did not jump under an upload, which is basically 4

i l

5; what happened at the Maine Yankee -- the Vermont Yankee.

t I

6}1 COMMISSIONER BRADFORD:

I guess I'm still behind on this factor of safety of one.

Does that mean that when i

I t

j 0

you run the conservative bounding assumption that some i

9 set of su ports will be stressed right to the limit?

~

l Y

10 1 MR. EISENHUT:

Right to the limit, it would for this, i

i 11 upper conservative thing.

Now that was the basis for the I

12 l short-term program, and that's what we did on the short-term i

13 program where on the limiting c6mponent, what we really thought, I

14 would happen would be the limiting e.iece in a factor of safetv

- i 15 of two, most of it had like four.

For the sensitivity case, 16 where we cranked up the loads and said 'What even if' for this i

17

case where you boundcd everything, the limiting components 18 still had a factor of safety of -- to the point where it 19 reached -- the capabilities would be with the --

20 COMMISSIONER AHEARNE:

Yield or ultimate strength?

1

• 1 'l 1

MR. EISENHUT:

Ultimate.

1

-e j Now I guess I should say we're getting ahead of l Ourselves a little bit.

Everyt'.ing we're finding cut bears

.i 24 0 out that -- I mean, there's twc acnsiderations:

3 -eceral A tCor*trs, lec.

ec,

'~ l Everything we're still finding out ' rears out that 1

7

/

3*

i

)[b 1

b

10 i

I i

l mpbl0 that that censicivity case still bounds everything we've seed. l I

t i

i But secondly, the other piece of it is on the long-term t

l prog::am we're going to fix it up so that there is a f acto.

i I

i of safety above what our new Isud would be, whatever the code l

1. )

l t

would recuire, whl:h is up nearly a factor of safety of four.

c' I

COMMISSIONER AHEARNE:

Now let me get back to that 6

t other point.

It sounds like what you'ra saying is that your 7

sensivitity case, v.ou increase he loads by 50 percent, which 8

i, i

is what you s av.

v.ou think is the boundin case of the oro-9 i

10 I gram brings some segment of material to ultimate strength.

f 11 MR. EISENHUT:

Yes.

l 12 COMMISSIONER GILINSKY: Now how well can we calculateii l

13 that?

l i

14 MR. EISENHUT:

This has actually cot conservatism t

1 ~5 because this is the code allowables which have the margin in it, 16 built into it also.

17 I remember we went through this in some discussion 18 !

last May because -- even when you get to that point that you 1c,

say that the loading through the component, whatever it is, 20 a the structures, materials, is up to the code allowables for

'l You still have conservatisms built into it.

ultimate.

9a 1

-' g COMMISSIINER AHEARNE:

Well, are _,:u saying that i

23 h ycu,re reaching in this case the code allowable, which is scme

.i 24 1 racecr _ess : nan ne actual strength of the material, cr are sco-dedef al A tCor*ers, t r*C. j

• C l

vcu sav.inc. that you're reachin? --

~~

1 1

.[)h b U,' i*

n' n i

)

I

11 I

i t

i i

i 1

mobil MR. EISENHUT:

No, even on the definition of I

f 1

i 2

ultimate within the material.

i, i

3 j

MR. GRIMES:

I'd also like to point out that we're I

e talkinc about one component, like one column would reach this.

4 I

5i It's not that each column would be stressed to this particular '.

L i

4 4

6l it's normally one or maybe a number of columns together.I l

limit, I

7!

I COMMISSIONER AHEARNE:

I was just trying to under-l I

I I

stand this first issue.

I 9

DR. LIAUN:

When we talk about capacity we do mean i

10 I t'e ultimate strength of code specified minimum ultimate strength 11 of the material.

i COMMISSIONER AMLUNE:

Now does that mean, when 12 13

(

you use,the phrase " code specified ultimate strength", is that I

i i

11 the ultimate strength of the material, or is it something less?

DR. LIAUN:

No, minimum, minimum value.

16 MR. EISENHUT:

That's why I meant that even when i

17 you reach the code specified minimum --

18 DR. LIAUN:

Minimum ultimate strength.

19 '

MR. EISENHUT:

-- that's got conservatism in it.

f 20 i CHAIFF.AN HENDRIE:

It means the following:

91 '!!

If you have a batch of steel and you want Oc class

^

,, 1

.. s it the code class, say, 42, you do some testing on it and all

.i i

s n.

I c:

sa <

_ One tens;_e specimens have to snow u_timate streng:n d

's,

tc..

cer.i a.conen,nc.

strecctn er viel:.

as the case above that code sreciriec viel:,

9C may be.

It doesn't mean that there is an engineering safety r

y Li 7 I' f f

JJJ u i i

19 I

1 I

mobl2 factor bui?t in.

It's a material croperties qualification I

l 2'I which says that a lot of --

i I

l 3l COMMISSIONER AHEARNE:

So what it says is that i

I v.ou are guaranteed that any material that you use has an 4

Si ultimate strength no less than that.

',i 6

DR. LIAUN:

That's correct.

i I

7' COMMISSIONER AHIARNE:

But it does mean that some j

i i.

8 mav have an ultimate strength --

l

^

9 MR. EISENHUT:

Frcm the material's conservatism.

10 l i

DR. LIAUN:

Let me quote you a number for the type 11 of material that we're talking about.

1 12 The ultimate strength ranges from 70 to 90.

f, MR. EISENHUT:

But within the code allowables there !

13 is material conservatisms and there are -- Are there any other !

14

.c conservatisms?

16 DR. LIAUN:

Yes.

Most of the coding we were talk-17 is based on static or pseudostatic type loading ing about la i i,

tested at extremely low strain rates.

But the type of loading f

19 :

phenomenon that we're talking about is of verv short duration.

l i

So based on scme test of material on yield strength and ulti-Of

~

'l mate strength at high strain rate you can expect a normal-

,, i l

ly up to about 20 or 25 percent in additica te what's 1

m specified in the code.

^~

e.

1 So the firs: effect we're tal.cing about cn the s..,:.r i a comri, v ne.

c 4,e i

minimum property to what mean value of the test specime. you a

)

U i

e

13 i

I 1

or so.

mpbl3 can expect reasonably an additional ten percent l

2 CCMMISSIONER AHEARNE:

Is the time duration i

3 i

escecially short?

Do you actually get a c.ressure pulse?

f i

4 DR. LIAUN:

I beg your pardon?

l 4

l 51 l

COMMISSIONER AHEARNE:

Do you actually get --- !s I

I 61 the time duration sufficiently short that vou actually get a i,

i I

i 7

shock pulse going through that?

j l

DR. LUAUN:

That's what test data says, yes.

l 8

i i

9 MR. EISENHUT:

So the point I was making is there i

1 10 I is the actual materials consideration's conservatisms, the 11 some of the same considera i static, the dynamic considerations, 12 tions we 've been discussing previously with various aspects,

I 13

(

that when you cet.to the minimum soecified code allowables 14 for even ultimate, it's not necessarily ultimate, there are 15 several factors on top of that yet.

So that you' re confident 16 1 l

that you're really, you know, bounding the problem with some i

i 17 '

other conservatisms.

18 COMMISSIONER GILINSKY:

Now what happens with 19 ultimate strength, say, with a column?

Is it going to buckle, 20 '

tear, or what?

21 9]

MR. EISENHUT:

The simple line is that it fails.

.. o It could buckle, it could -- and what we're really talking

m.,

4 i abcut here is an upload.

1 24 4 COMMISSIINER AEEARNE:

Your maximum '. cads are tc. r.c.r.i 9.xmn. inc.

,c 6.

vastly greater in the dcwn, as far as pressure.

3 d

)Jb b 'i U h

- /

7

14 i

i, 1

mo.bl4 MR. EISENHUT:

Yes.

2 MR. GRIMES:

There is a greater capacity to take the:

i, 3:

load in compression in the columns than it is in the upload

,t, where you tend to try to pull a joint out of the torus.

l 1

MR. EISENHUT:

If you have a big column bolted down 5l i

i 3

on the floor, if you tried pushing the column, it's got a 6

7 lot more carability than if you tried lifting the column.

i 8

COMMISSIONER AHEARNE:

So v.ou are talking about I

9l

-- Well, are vou talking about the failure of the bolts or the,

t, 10 !

i i

structure?

11 DR. LIAUN:

A sheer pin.

Most of them are dic'c.ated i

1 by the sheer capacity or the pin connection at the bottom of I

12 i.

i 13 i

the columns.

t COMMISSIONER GILINSKY: So the pins would fail?

15 MR. EISENHUT:

To me, if you hit the actual 16 ultimate, you fail the pin.

17 DR. LIAUM:

Actual ordinate, you fail the pin.

13,

MR. EISINEUT:

As Dr. Liaun said, there's these i

19 other factors in there, though, why we think that even if --

20 COMMISSIONER GILINSKYr I'm trying to understand i

21 '1I what failure means.

,,I, DR. LIAUN:

And those factors I mentioned were nct

.. i

.t

. sa ety O _ two, it'S On OOp Of that, 10cluCeO in One CaCOOr 02

.J t

l 24 '

A eCo rMrt, j fic, ' on top of the two.

.cr.e 9C er s t 94 i

COMMISSICNER AHEARSE:

Ncw you said that ycu were

~~

1 a

9 7

/

6

<L J

s

I mpL15 worried that your experimen s or your analysis had a non-linear; I

I 2

response.

i t

3 I MR. GRIMES:

That's right.

If you increase the 1

load'by 30 percent, for example, the stresses go up by almost l

i a factor of three.

That is the way the structure responds.

5 i

f i

6

Once you overcome the weight of the facility, the weight of 7

the water and the mass of the steel, then the rest of your i

8' load goes simply into stress at these local points.

9 So the Owners Group and General Electric went back I

i i

10 !

to review their most probable load specifications to identify a n v. what thev. felt might be conservatisms in those so.ecifica-11 i

l 12 And at that time they concluded that three dimensienal l tions.

1 13

(

effects for conrervatism because it was felt -- and the 1

Staff agreed -- we felt that a 3-D reduction should be expected,-

14 i

I ~t because the two dimensional cell confines pool swell and leads 16 to, you know, trying to move everything in the upward direction; and also the irregular downcomer spacing should lead to a 18 staggered swell process.

19 i

And so General Electric performed an analysis where 20 1

they did some cell calculations to determine what that conserva-r 1

21 'll tism would be, and they calculated to 20 percent conservatism.

t They calculated some other conservatisms too, like break b

4 opening time and ccndensation in the drywell and things like i

a, that.

But we igncred these and we concentrated on the btg sc.;,cesi =ecores,inc.

• C one, beinc the 20 percent 3-D conservatisn.

And tha led us 4.

a

16 l

i I

mpbl6 to our sensitivity bounding load.

I I

2 (Slide.)

j t

3' l

The downward load was simo.lv. the 50 cercent.

i I

i 4,

Multiply the base case by 1.5.

And on the upward load it i

I 5

so the 50 percent with the 20 percent conservatisms, was I

6l that's 1.5 times

.8.

It ends up being a 20 percent increase I

7 l

in the load.

i i

I 8

The two 20 percents sometimes get to be confusing, j

9l but the 20 cercent reduction led to the 20 percent increase i

I l

i 10 1 in the load.

(Slide.)

12 COMMISSIONER AHEARNE:

Now the downward p'ak that I

13

(

vou end up with is.what?

f, 1

14 MR. GRIMES:

Scmewhere around 24.

15 COMMISSIONER AHEARNE:

With that rapid a rise 16 with that peak, you still don't have any problems of a down-17 ward load?

18 i

MR. GRIMES:

It's not that we don't have problems.

I

9 (

There are scme parts of the strr-'"re that are sensitive to 20 i that load.

But they either beefed up the columns to take it, 21' if they had close stresses, er they had saddles, and the

.m.

saddles could usually take it without any trouble at all.

23,

- -, m.-

.w: -

.w,-

t,.,. o--

c.~-w c Jm.,.

e-w-gs,. e_a.r <.e <:

-c a

..._o

.v 2 :

2d i l

em, a ecomn. inc. l here, :s this T.easured or calculated?

j!

MR. GRIMES-This is calculated.

This is a lead i

[)

b.(

7 I

17 l

i mo.bl7 scecification that was used.

2 COMMISSIONT.R AHEARNE:

I wanted to ask:

Can you i

3i l

get any kind of a shock wave formation?

4 MR. GRIMES:

No.

i S

COMMISSIONER AHEARNE:

Even with that --

j i

l 6

MR. GRIMES:

I'm sorry the time scale isn't marked l

7 on here.

It's from zero to 1.4 seconds.

So the rise time is 8

like

.2 seconds, which is sufficiently slow to avoid shock 9

waves, but it's fast enough to move the structure.

t i

10 1 I

MR. EISENHUT:

You know, you can see up there on the :pload where we put in a factor of eight for 3D and 2D 12 effects, and in the bottom we only used a factor of one.

13

(

So this is why we felt it was necessary to notify l

l 14 the Commission last August, when we laarned that the factor i

i 15 of.8 may in fact be

.1, or 1.0, particularly in light of i

16 this question that for the sensitivity case the factor of l'7 safety would be greater than or approaches one.

So cf course 18 it would mean that if the sensitivity case went up it would 19 get you into problems where you theoretically have a facter of 20 safety of less than une.

i e 1,J sut since then we've gone back, as we're coing to d

l 99 9

l!

arter we C.et t.nat verv. O. r e l i m i n a r v. in!crmatien, J

Co OnrcuCn CCW,

'l eq !

l mw J wnere we came cut and what are the actual numbers we're ccn.ing

f 2d o b..-

.d, u ses ;,ceral A eoorwrs. Inc.,

I

e..:

,j MR. GRIMES' Just 20 Summarite On the shcrt-terr l<

I L p

/

h b i '-

i

l 1

I most probable mpb18 program, we had the two cases, a base case, 2 l factor of safety had to be greater than I

load, limiting component, i

l 3i I

l or equal to two.

l i

The sensitivity case used a bounding lead based on 4

I i

5 1/12th scale data and the factor of safety had to be greater j

I 6

than or equal to one on the limiting component.

7 (Slide.)

{

i 8

Our short-term croc. ram was summariced and our 9

safety evaluation which was issued in December, 1977, that i

I i

10 '

was followed by the exemptions which we issued to the Mark I j

11 plants in March, 1978.

The outstanding concerns that we l

12 identified at that time that we would have to resolve in the

(

long-term program were to confirm scaling, which we nave i

i 14 completed.

We have both industry and NRC courses to confirm 15 the scaling techniques.

It was a download anomaly in the i

16 short-term program that we discovered to be a window response i

in the test facility, which was atypical and led to an 18 exaggerated loading condition.

We have resolved that with 19 l

all of our testing programs.

I 20 And we have a full-scale test facilitv for the 21

.I condensation loads.

Testinc is cc=pleted and the data is d

v" "j

currently under review.

The only thing that we haven't 4

2' 'l

• i ccmpleted is this 22 load reduction.

We have found that 1

24 i ics.eocersi AecorMrs, Inc. j there is no lead reduction for 30.

l New that conpletes the discussion of the shcr -term i

J.; b' L ".o j

~j f 1

c i 14

19 i

I, I

and now I'll go to a brief discussion of the mpbl9

program, I

2 Livermore test facility.

I i

3i The Lawrence Livermore Laboratory was reques:ed j

I I

1. l to perform a large-scale three dimensional pool swell test l

5!

and the Office of Nuclear Reactor Research contracted the I

I 6

Lawrence Livermore Laboratory in 1977 to perform those tests.

l t,

7 Through bench tests and analysi., the Livermore i

8 selected 1/5th scale as being large enough in scale and a l

1 9

90 degree sector as being sufficiently 3D.

10 1 (Slide.)

i 11 The components of the test facility included a i

12 drywell, which included a three dimensional 90 degree sector l

13 and a two dimensional sector was added to provide direct 2D

\\

I and 3D comparisons.

15 COMMISSIONER AHEARNE:

These are the same facilities?

16 MR. EISENHUT:

Yes, one is looking down into the l

facility.

18 l

(Slide.)

i i

19 1 This facility on the first photo is actually under 20 1 the -- below grade.

i 21 l rrMy. - - 0h r R

.,-u.

,w,. :

_ see.

%... : 2 (l

(Slide.)

• 3, MR. GRIMES.

In the 90 degree sec cr we have two 1

main vents and d:wncemer pairs.

The load was derived from 2-AC2 etcef 31 R fCOf*ff t. INC.

,e o

^~ l local pressures measured along the head of the instrument at 1

n U

7 [. b u

s

.). b N2

20 I

l' mpb20 one-half of this sector and the individual pressures were l

2 analytically combined to come up with a net result of load.

l 3

When Livermtre first identified their load calcula-l tions'we felt that there was a potential error in the calcula-I i

5 tion technique, the way that the individual pressures were i6' analytically combined.

And so we requested a reanalysis of j

i Il the downward an-pward load.

And we also requested an un-

{

l 8

certain calculation be made in both the 2D and the 3D because we felt that'there might be a chance we were seeing a calcula-Y t

10 I l

tional error in the loads.

11 (Slide.)

1 I

12 So Livermore -- Well, the Livermore original loads I3 led us to send you the August.4th, 1978 memo identifying a

(

j ',

The original loads that Livermore calculat-potential problem.

1 **

ed were essentially equivalent to the short-term program or i

16 a little bit lower.

I 17 :

(Ove rlay. )

I 18 '

3ecause the loads in the -- the two dimensional 19 loads were below both --

1 20 COMMISSIONER AHEARNE:

That was the short-term base 21 1

case?

M *). ]

MR. GRIMES:

That's the short term base case load, not the sensitivity case.

24 The Livermere loads were then recalculated using 3 E sy**f a? RtOOr*tft,Irtc, '

SC l a recalculational technique.

They revised the analysis.

The a

[

4 [

N i

]

I

21 mpb2' "evised analysis showed that the three dimensional load was 2

still higher than the two dimensional load.

3 l

(Slide.)

t i

4 But both were below the short-term program.

j i

5 These numbers were based on a hand-calculation I

6'i that was performed bv. Livermore personnel and some or our i

o I

i 7

from the Brookhaven National Laboratory, myself, j

i consultants l

8 certainly after that August 4th memo.

9 Since this slide was made Livermore has completely i

10 '

They have developed a better calculation-i reanalyzed the data.

i 11 al technique.

3D loads have come back up to be about equivalent I

12 to the short term program, the short term proc *am most 13

(

probable load.

i i

Id COMMISSIGNER 3RADFORD:

Is it surprising to see 15 thatsortofdifferencebetweenthe3Dandthe2Dcalculations?l 4

16 MR. GRIMES:

Well, I'll get into that.

We've reached the point where we have identified a l

17 18 1 i

ncaber of differences between all of these facilities, including i

19 the 7.5 90 degree at Livermore.

They weren't perfectly matched, i

20 i

because, if you will recall, there was a long vent line lead-l 21 :

J ing from the drywell to the 7.5 degree sector which led to

.e.

n i

some cacacitance effects in the vent #1cw.

a o

2 ~'

MR. EISENHUT:

A simple a swc_ to that is nc, it's 24 s c. _.e w, a ex m n. i nc. ',

not sur risine.

It is at firs: blush, but when you icek at SC.

the actual fine tuninc differences of the medel it's not.

~~

4 l

,c-v.

a

)sU U/r

[

I

22 l

l l

\\

mpb22 COMMISSIONER AHEARNE:

It is surprising but l

2i i,

exc.lainable.

l 3i l

MR. GRIMES:

It's surprising and becoming more i

i I

i 4

explainable.

I i

5' MR. EISENHUT:

We really thought the -

'3e thought I

6 we should see reductions in going to 3D effects.

This chart l

shows what we had in mind last fall when we told you that we l

7 I

8I received 3D.

We did the 2D, but it wasn't really alarming at I

i 9

that time because 2D was so much less than the most probable i

t and even the higher 3D was less than most probable.

i i

i 11 COMMISSIONER AHEARNE:

Now you're saying that the l

reanalyzed Livermore, that is the 3D back up to near that.

I 12 i

(

MR. EISENHUT:

Could we go back to the last slide?

I 14 (Slide.)

15 COMMISSIONER AHEARNE:

I think what you're saying i

l i

16 is the reanalyzed 3D, it's closer to the original.

MR. EISENHUT:

Right.

What you've seen is there's 18 a function o' time.

The 3D Livermore loads were ccmparable 19 ;

to the Mark I short term probable.

They decreased, and they 20 '

have gone back up as they keep being refined and refined.

i l

21 d' V

CEAIRMAN HENDRIE:

You see, John, not only are the

,, !I leads themselves Oscillatory in nature, but the calculation of

,a 1j any given load at any given time is itself, as a functicn of l

24 time, Oscilla: cry.

sc.. cero a eoomn. ix..,

\\

,C (Laughter.)

I, b b.i b b2b

,1 o

23 i

I.

1 m o. b 2 3 CEAIRMAN HENDRIE:

Presumably it's a damped curve 2

and eventually, you know something will become smaller --

l 3

l COMMISSIONER AHEARNE:

It has something to do 'vith i

l 4lI i

mox constant for gravitation.

j i

5 (Laughter.)

t 6

CHAIRMAN HENDRII:

I wouldn't be a bit surprised.

7 MR. GRIMES:

We are oscillating in, we aren't 8

oscillati~g out.

I i

9l CEAIRMAN HENDRIE:

As long as it's converging.

i i

i 10 I MR. GRIMES:

It is converging.

That is the reason 11 i

for our conclusion, by the way, that we feel that the short l

12 term program is still valid, because even though our conclusioni about three dimensional effects was wrong -- Could we have-14 the slide reversed?

Not upside down, sideways, i

l 15 l (Slide.)

i 16 Thank you.

17,

I We know that our conclusion regarding three dimensional effects was wrong.

However, because o f our 19 increased kncwledge about the magnitude of the loads, we still

+

20 l

iudce than the short term program mosu probable load is indeed 21 most probable.

And the sensitivity case is bounding of al' of i

22 :l3 the data that we have today.

And therefere the conclusi:ns

'3 'ij regarding the structural calculations in the shcrt tern program

^

1 24 i

--- l va" ='sc-e di a-a

> ;.e.i n.cory n inc.

,e New for che long term program what we are dc ng g_

e I

l s.)

. /

L.

o,<

mpb24 right now is going bcck and comparing these data, trying to I

make them comcarable at the same test conditions and to wash 2

l 3

out the facility dirrerences in order to understand exactly l

i what.the loading conditions are going to be.

l 4

5 i

(Slide.)

l i

6i We collapse the data to equivalent conditions as i

i t

I best we can with analysis or sensitivity tests or whatever is f

I available.

And we consider all cf the data that's available l

B 9

because each of the facilities has its certain quirks or i

r 10 l i

certain difficulties.

What we've got left is an uncertainty I

f 11 band, and then we will consider.

.t uncertainty band when we i

1 12 select the load that shall be used for the long term program 13

(

analvs:.s.

i l

14 COMMISSIONER AHEARNE:

The GE and EPRI, are those 1

B 1 ~*

the only data points that you have?

16 MR. GRIMES :

That's only a few data points we have I

17 because the bulk of their tests were done at diffarent ncminal 18 l conditions than the bulk of the Livermere tests.

That's one i

t e

i 19 !

of the difficulties we've had in making the data ccmparable.

20 I' CCMMISSIONER AHEARNE:

So if I look at that, are GE and EPRI making extrapolaticns from ancther regica of the

~2 chart?

3

+1 m3 -l MR. GRIMES:

That's ccrrect.

.j a

24 But that is definitelv the trend that thev teu,

' cs-e_ecerai A ccomrs, Inc. ;i ec

",i that 2D is larger than 3D based :n the regi n that tOey 4

l J

j f~ 0 p.~ ;

f d

) 1 i,

25 I

1 I.

I I

mpb25 investicate.

i l

2 COMMISSIONER ASEARNE:

I'm just our led at how to i

1 t

i i

j interpret what significance I should give to the EPRI and G",

3 i

i I

4 'i data lines.

In the first place, they've obviously not l

t i

i 51 comparable in the sense that this says 3D is less than 2D.

~

l 6l MR. GRIMES:

That's correct.

We believe that that i

i, I

l is because they tended to do all of their tests at nomir.a '

7' i

I 8!

lonc term p 0 ran conditions.

1 i

I 9

MR. EISENHUT:

Excuse me.

i i

1 1

10 1 There are really four sets of data on the slide.

1 i

CCMMISSIONER AHEARNE:

I understand that.

11 12 MR, EISENHJT:

If you look at the bullets at the

/

13 tcp of the 90 degree sector, Livermore is in 3D.

So it shows

(

i I

14 at Livermore the 3D is greater than 2D.

It shows -- The 15 previous piece of data we had was 1/12th scale data, that l

16 earlier we showed, but it's the first part.

It doesn't i

17 reallv show that 3D is less than 2D.

1 18 :

COMMISSIOJER A.9EARh ':

What I'm saying is you had i

i 19 the other two sets of data which were represented by extrapolat, 20 \\

ing lines.

If they are to be reviewed at all as comparabla 21 h then thev woulf rav 3D is less than 2D.

a

4 99 !

v_re.

. u.a. i s - -._.._,.

.u.a.

a._

,,I,

.u.

. _ _ _c re...c..-.

m.w. _, i _e _._4_...

-i n

i I

24 {

COMM SSIONER AEZA.5E:

And sc therefore I'n pur: led o

,e o mo r n. inc.

,e

^~ l by what conc'_usien I'm supposed tc draw.

L3 n : '

7RlJg 4. 0 1,

26 i

l I

I 1

apb26 MR. EICENHUT:

The conclusion you're supposed to 1

2 I I

draw, tnen, is we're going to look at that and notice that I

I 3'

our own independent study that we feel we understand is l

i 4l1 i

the Livermore test, and it still gives us this anomaly where j

5 the 3D is greater than 2D.

So there is going to remaan 6

uncertainties, as Chris mentioned, that when we're done with 7

the program for the long term program so it's very likely i

1 8

that what we will do is therefore take the load and add some 9

factor on top of that to be sure we cover this all up.

We'll 10 add that factor on the top, then add on the factor of safety 11 to get to the code allowables.

i We're not going to address the code allowables, j

12 13

(

the extra allowables in the code.

That's just there for l

14 additional comfort.

15 But because of the question you asked we are going 16 to add some additional comfort on top of those.

17 MR. GRIMIS:

As I said before, what we're trying 18 to do is collapse the data down to equivalent conditions.

I 19 !

The band that's left is an uncertainty, the degree of which 20 we know.

21 d

CHAIRMAN HINDRII:

What is the approximate fractional

.I aa q value?

Ten percent, five percent?

y I

MR. GRIMIS:

This one is on the Order o f five to ew 2d

• an o p

• g "- ~*

• Ace 'eceral Aeoorters. Inc.

~c sw a

d==

a' sh e

w4e w

e e

que as w

a w

i k

UJ[

(ll<.

V

.? /

I I

i i

i mpb27 you think you'll have as an uncertainty?

l I

i MR. GRIMES:

On that order of magnitude, I think, i

2 I

l 3 I because we're trying to look at the spectrum of conditions.

i I

I And most of the test data, it appears that the nominal l

3 l

5 conditions are around the middle and we're trying to show i

6 that we have both of the ends covered.

7 COMMISSIONER AHEARNE:

To collapse data from a 1

5 i

e I

8 varietv of experimental devices, 2D and JD, and then end up I

'I with five to ten percent error bounds, you would have to have 10 i a very firm understanding of all the phenomena that occurs.

l i

MR. GRIMES:

We hope to be able to convince anybody 8 11 I

i l

12 {

that we do.

j f

1 13

(

COMMISSIONER AHEARNE:

Fine.

l, i

MR. GRIMES:

We've gone through all of these facilit-14 1 ~5 I ies with a fine tooth comb, and I think that my overworked

\\

16 '! consultants would feel that we know as much about these I

17 !

facilities --

COMMISSICNER AHEARNE:

Fine.

Then there's a very

[d 19 <

si=ple way of showing it.

You would have them predict the results of another experiment, do the experiments and see if lo 21 l thev. track exactiv. the c.redictions, and then you would have g

a

$9 i

established that.

MR. IISENEUT:

That wculd be one way, certainly.

CHAIRMAN HENIRIE:

Well, we're not in a very goed

..s.c.,si neocre m.inc.,

^4 cosition to build a whole new facility.

^~

3 ;b bb

28 t

i l

1 m=b28 COMMISSIONER AHEARNE:

No, no.

I'm just dubi-us I

~

2 that when you have that wide data scatter and you've got.0 I'

3-and 3D, that the level of understandinc. of the methodoloc.v.

i to get to five to ten percent error band collapse and all that l 4

l i

5; j

is very....

6 CEAIRMAN HENDRIE:

I think what you're looking at, I,

7 I think your best data is the Livermore, and I think all you're doing is sav.inc. 'Well, there's some indication just l

E 9

to cover residual possibilities we'll add some small incre-i I

i 10 '

ment onto that to account for the fact tha't some of these i

I 11 '

other experiments which we think were less representative... '

I 1

COMMISSIONER AHEARNE:

Now that I could understand. I 12 i

t, CHAIRMAN HENDRIE:

I don't think there is an i

I 14 !

implication that you understand the hydrolics, the two-phase 15 hydrolics.

16 COMMISSIONER AHEARNE:

Well, that's what I thought 17 that collapsing the data meant.

i l

1 *: 1 CEAIRMAN HENDRIE:

I think what s t.'.ppening here is i

i 19,

that you're anchoring on that Livermore 90 decree sector stuff 20 and just saying 'Well, we've got a batch of other experiments 21 c her f acilities and we'll try to ccrrelate those with i,

at 1,

me

,1 what we think is the correct answer, and we're jus: trying a i

as i i

little ten percen cr whatever en the requirenents just

~

1 24 because we want to be censervative.'

And it's in that sense 339Cef te R fCOr*Tes, I nc,,

S4 l rather than ecliapsing and correlating.

[)b b b

i

.i

29 i

i l

i 1

=pb29 MR. GRIMES:

If I might go on --

i CHAIRMAN HENDRIE:

Before you get off this thing, i

i I

3' let me take you off on another line.

i 4l The abscissa is drvwell cressurization rate which l

5:

goes all the way frou quite small breaks so that the pressuriza-l 6 !

tion rate is small --

i 7

MR. GRIMES:

This range of pressurization rate is simply over the plants that we have, what is the design 9

basis accident pressurization rate.

I 10 I I

CEAIRMAN HENDRIE:

This is the double-ended full-l 11 1 j

scale banger, and the variation is plant by plant, rather than '

J I

i 12 o' Ah, so.

In that case, I don't object to the lack of i

i l

l 13 slope or gradual slope in pressurization.

i 14 MR. GRIMES:

As you go down toward the zero end of i

15 this curve, the thing takes off, you knew, exponentially, and 16 1 then out in the high pressurization ranges it's relatively 17 I

flat until you reach a o.cin: where v.ou get instantaneous t

18 :

l bubbles.

19 i CHAIRMAN HINDRII:

And down at the low end what 20 you're doing is evenuually pressurizing the drywell enough to l

31 al clear the vents,put in the bubbles and se on?

^

,, 1 "qi

.v.2..... a. s.,..n.

4.

.y....

,. i 4E il IHAIRMAN EINDRII:

A:.d you don't get any of this 2s '

seuff.

...;,c.,,,

a o,,,,,,,nc.

,c Now le: me ask you another cuestion, and then I'll

. 7.

i,

..J i

30 i

I l

1 mpb30 quit.

1 2

Recent events suggest that one ought to ask oneself ~

4 3 i whether there would be a possible occasion and what happens I

I g;

1 if the relief and safety valves stay open?

So in terms 5

of the forces you're looking at here, which are for the i,

6i full design basis loss of coolant accident, what do the i

1 7'

I relief and safety valve blowdowns look like against this scale?;

i I

8 If you blow a wall up and let them run, do you get i

9l ccmparable forces?

Is it substantially smsller or comparable? :

10 I Is there any subcombination of orie, three, five and seven that 11 does sometning?

I 12 l MR. GRIMES:

The safety relief valve discharges l

13 tend to be more localized anr even is h a number of them going,

14 off there is attenuation between the LOCA conditions.

Thir 15 loading condition that we're looking at here is on the whole 16 structure at the same time, and I'd say that there are certain 17 ;

cceponents in the system that are sensitive to safety relief 13 i

valves, but I'm not sure about components.

19 We use a different kind of analysis to determine 20 what the response of the structure is.

21 MR. EISENEUT:

We didn't specifically include that n.n.

in here, but we are lookirg at all the different phenomena I

eq i

associated with this.

24.

merms. inc.1 We are gcing through each of these possibilities a.; c y.i m 25 i in great detail as part of the long term program, and there J. l. (

(3, j ji vU

31 I

l I

1 mpb31 may well be some localiced modification.

1 2

CHAIRMAN HENDRIE:

Well, I can remember at various i

I i

31 times baffles have been torn up and there has been some I

4 think there has been damage to s*-"-*"

al damage to them.

I

'I I

supports and so on, internal succ. orts.

I 6

MR. EISENHUT:

The solution is probably changing l

I 7

out the quenchers in the bottom of thei discharge.

i 8

MR. GRIMES:

Replacing ramshead quenchers.

They are greatly reducing those loads l.

9 MRc EISENEUT:

10 !

l It's just that we tried to concentrate here on i

11,

strictly the main load and the LOCA.

12 i

Okay.

Onward.

\\

13 I

(Slide.)

i MR. GRIMES:

The long term program definition will 15 cons:.st of three dimer.sional testing in a 1/4 scale facility.

16 Each plant's geometry, specific geometry will be tested to 17 l provide them with a base load and condition.

There will be a 18 conservative load definition in that all of the things that i

19 influence the loading condition, thev. alwav.s co to the worst i

20 case loading condition.

71 4

Also new we do have a larger data base, so uncertainty 9, il

^^ ]

in the magnitude of the loads is diminishing substantially.

1

?

07'~ l We have the three dimensional effects under review by both and the EPRI 1/12th scale test.

And / hen W

,c..;-cer., acoomn. me. l the 1/ 5 th scale test

m. d this slide was developed we had targeted cc=pletion dates ic r p

a l

7Uf m-q

32

==b32 for development of our position as May 31st.

Consultants at I

~

l 3rookhaven have just submitted to us a number of options that 2

i they would 12xe us to consider to resolve this issue, and we 3

i 1

1. l received those vesterday.

We're hoping to have those in the l,

t 5

mail this week and start -working on developing our po ision.

I 6

CCK'ISSIONER AHEAPSE :

So what is your target date?

7 MR. GRIMES:

The end of this week, I hope, or how-8 ever long it takes to get all of these options.

i MR. EISENHUT:

Well, this is a little misleading 9

10 1~

because we won't have the program complete.

The completion 1

11 for the long term program is going to be a couple of months 12 later.

i l

MR. GRIMES:

Right now in the status of the lonc l

'3 1

I#

term program we have been reviewing load definition for about 15 We have 46 reports submitted to support the load a year.

16 definition of 24 loads, and we've worked through I believe 17 about half of those loads.

i I8 l Right now we hope to have positions on each of the i

19 '

loads developed by July so that we can begin the analysis for i

'C '

the long term program and also the Mark I Cwners have started 21 ]

submitting their implemenuation schedule.

They have in the a

,na past, and they continue to make modifications where it was il

'3 convenient for them Oc try and shorten the implementation, the time it will take to completely resolve the Mark long terr

,aJecef al A eoo rters, I FC.

SC i

p!cgram.

ng 9-Jd bs) 4

33 i

i i

i

=pb33 COMMISSICNER GILINSKY:

And how long does that take?l I

2f MR. GRIMES:

The implementation?

We're just getting I

k i

our schedules now, and I'd hate to give you a number because in 1

I i

i a number of cases we hoce to be able to negotiate with the i

5 utilities --

6 1

MR. EISENHUT:

Well, let me put it this way:

4 i

7 We're going to lean awfully hard to have all of 0

the modifications in place by December,

'80.

That's what's been said for a long time, but we'd like to stick to that if 9'l 10 1 at all possible.

i 11 We've recently sent out letters asking all licensees to propose to us their schedule, recognizing that a lot of j

12 l

I 13 these modifications take weeks.

If they can't get all of the i

I4 modification s done by December

'80, we're going to try to get 15 a priority scheme.

16 Part of this, as Chris mentioned, is we've b.en i

17 !

I sort of implementing as we go along.

The minute we can pin 18 >l down the pool swell loads, for example, we're going to sit 19 <

down and tell people 'Here's how we think you ought to go

'0 about cranking it up for your own plant, and get on with it.

i

' l "'

i Send us your plant analysis.'

.i

$"9 l

What we're still shooting at is December,

'80.

33l' There are going to be scme hardship cases probably where

..i utilities which have mavbe three or four units, three or four i.a

1. ,,eer o a.oenen inc.

,c,

3WRs are nct going to be able to bring them down all a: the

~~

1

~ c

.s

,s ul0

34 i

l l

I mob 34 same time without having a major impact to make these mod:. fica i

2 tions which may run twc or three months.

What they would hope '

i 3

to do would be to extend the refueling outage which they 1,

l normally stagger the refueling outages.

For example, the l

4+

I i

5' Dresden 2, 3, the Quad-Cities 1, 2,

they're sitting on the 1

6 If you require them to make the modifications same system.

by December '80, they're probably not going to get approval l

7 i

to do It until some time, maybe early

'80.

Therefore if 8

they have to shut down their plants to fuel them, they're 10 !

1 going to overlap.

11 So from a practical standpc!nt we're just going to 4

12 get them all in place.

That's why we're looking at getting i

\\

i 13 as many done as possible, and secondly, trying to get all the I

l major modifications done.

So we've sent out letters and 1 ~5 asked everyone for firm schedules to we can really try to 16 l orchestrate what we're looking at.

17 Our reviews are also finite about manpower to i

18 '

program our review.

So we have to program to their needs.

1 19 COMMISSICNER GILINSKY:

Could you just rem nd us i

20 what is important?

What happens if the supports fail?

71 l

MR. IISINHUT:

If the support fails the torus would

~

l move around too much.

Ycu could have an 2005 pipe that is N

9, d

^~ l attached to the tcrus that could be torn loose.

24 LI The tcrus an a EWR Mark I is the poc1 cf water that a :.we a.w.m. inc.

,e vou ase.

You draw ICCS water from it.

If you would, 4

1

~

/

m

)I)O U /

5

35 i

t i

l 1

mob 35

'herefore, icse that piping say following an accident -- and are the only reasons we're really icoking at te cccide'..

I 3 I this -- you could have yourself in a situation where you're i

.I

'l concerned about water supply.

i 5l The second concern is even if you have an uplift t

+i of the torus, as Maine Yankee -- Vermont Yankee, Vermont I

Yankee was concerned about in early

'76, if the torus actually 7

i I,

8 literally jumps and comes back, you don't knew what structurally 9

the integrity of the containment, the torus as part of the 10 '

containment is.

So it's a major concern.

11 Each one in each plant may or may not require modification.

On the download there are a few plants that i

12 f

I, l'

have to install saddles to increase their capability to take

~

f I

the downloads.

15 COMMISSIONER AHEARNE:

About the columns?

16 I MR. GRIMES:

The columns can handle it in some 17 !

plants and can't in others.

It's a plant specific model.

I 18 COMMISSIONER 3RADFORD:

When you talk, though, it i

19 would be having maybe a factor of one with regard to the pool J

1 20.1 Your item two is also an upward pressure.

swell impact.

1

,1 1 1,

Would that add to --

Was that added to the uplift

~

e, o

^^ /

Of the pcci swell --

a 77 MR. GRIMES' It's added to uplif t in ';he calcula-i

,1 4-cion.

a.s.cero a.cor-,rs. inc.

,c j COMMISSIONER 3RADFORD:

I see.

N'.) (f)

I

<m

36 I

t I

i t

mpb36 ll MR. GRIMES:

It's included.

I T

I 2

Now the cool swell i=cact is crimarily concerned l

about damaging the header or damaging things inside the contain-3 4'

ment or ge.lerating missiles.

S CCMMISSIONER BRADFORD:

Pool swell impact along, 6 !

then, isn't what you're worried about; it's the supports?

MR. EISENEUT:

Well, it's all tied together.

l 7

t i

8I MR. GRIMES:

The pool swell pressure load.

By

' impact" I was referring to tha slap of water against the 9

i 10 i structure inside the torus.

MR. EISENHUT:

Inside the big header.

And there's 11 i

12 !

cap walls and all kinds of things where, if this water comes I

up and hits it, if you have a solid cap well, for example,

/

I3 a'

it's going to want to move the cap well around.

15 So there's a lot of related issues in addition to 16 this compression, the lifting of the torus.

And again, since 17 term program only required a f actor of safety of this sort 1 ~:

two in the most problem case, here we're getting back to the 19 i l

criginal facility, it means increasing the factor of safety, 20 therefore it means generally scne modifications on the weakes 1

21 "

i links of the system.

,* n 1

CCMMISSIONER AEIAR:!E:

Was the design basis release

,",dl the strongest pressure pulse that you see?

2s !

MR. GRIMES:

Yes.

It's the fastes: pressurization.

.ce-Fsceral Aecor ers, Inc. j

't

^~ j MR. EISENHUT:

It 's an instantaneous break.

1, O

V f

1

)

s 4

37 1

mpb37 COMMISSIONER BRADFORD:

How are these five loads 2

combined?

3 MR. EISENHUT:

Which five are you referring to?

4 MR. GRIMES:

If you want a structural analysis --

5 COMMISSIONER BRADFORD:

I assume these five loads 1

6 occur close together, 7

. iR. GRIMES:

The pool swell pressure load combines 8l l

with the header impact to give the net loads on the torus, 91 l

SRV :-condensation and steam condensation can be a load 10 combination for a smaller size break.

11 We've looked -- in the short term we've looked at 12 condensation loads and felt that they would primarily be of petite concern.

And.so we left that to the large scale 14 program.

15 CHAIRMAN HENDRIE:

I think these loads occur in 16

-- one knows something about their timing.

So unlike the 17 case of the seismic loads, dif ferent components of the 18 seismic load, where you don't have a basis for establishing 19.

I and phasing the way the analysis is done most of the time, 20' and then have to account for not having that when you cc=bine 21 them, you know pretty well which is going to occur.

22, MR. EISINHUT:

I don't really think they're related.

23 If they would interact, we'd look for some approach either 24,

"so-*--

=RS=*

<~

'"~a

-+

ce c ascerw g Oorcarv 25:

It will either be absolute or SRSS.

)bb b.b

1 38 1l mpb38 CHAIRV.AN HENDRIE:

All right.

2 Other questions?

3'l (No response.)

4 CHAIR'G.N HENDRIE:

Thank you very much.

S' (Whereupon, at 10:55 a.m.,

the conference 6l was adjourned.)

71 8

9 Bi e

12 13 7

l 14 15 16 17 18!

l t

19 20:

i i

21 22 23 "

21:

.c=a.c e scom=ev' t.

25!

4 g

i t

f%RK I RE'IBl

• EASIS FOR SHORT H i PROGPR1 LDADS

• SiCRT Hi PROGPRI C0fiCLUSI0f6

• LLL C0tFIPJ% TORY PESEARCH PROGPRi

• EASIS R)R L0f'G tem PROGPRi LOADS

• PROGPR1 STATUS I

E

./

f b

h p

+.n m

w.

x

.s

~/

r,

,7 5-4 gi ce

. :.w... -

1 T's -

a.=

' A %%. gggp,5 s

w s

• d

T%j' in ht=UQ. sty %pf? Q,S:j'v~a= f iS R+ Q alCW.

t w ?~'-T"?i4~2%- N:':&s- =.,Wca.ws?

. 2 r.; w a

. w$.r:=y:2:.. ' _, p.s:2:; =l~:5%[

a-n:si

.~. 'E ' i 2*

'C:.YYQ

+.:;; %.:.: : &; : # 4-- 9 = ~

x1.yfy""M:..a v9 DRYWELL TORLIS 6

g m n.

MAJOR POOL DYNAMIC LOADS

1. DOWNWARD PRESSURE

2. UPWARD PRESSURE

3. POOL SWELL IMPACT

4. LOCA STEAM CONDEttSATION

5. SAFETY-RELIEF VALVE STEAM CONDENSATION

,I

Djm ~ e, p. m,t nt ;

y J v' t y j,

a f-;'. x : 7 :

s_.

,, yy, >-

' ? W ? f-f ;

h 3 4

_ 'QQ ;;.

. :)c.~b.i.?l0

, ?' ? f., &

8

gg),;$g

. ]T'

$)s" 14**f"~"' ~

%th,

~

g

=~:

7

..a::asswr s

4

'y up l

Ilil,

)

(u

]flV

_-l n

3 e' g,

y

.t..

. 6y 4*q F,

6 h

g

.g w

5

~;

mzw GCi=,,

a)rmsp;g:\\i-!f. '.

5^.y ;., -

n

.,, ~

d'

/pR-

.eg%Q$, Q; u < 3 # E t y;.:

G7: 1.t... ~. :

c x.ss.N:::.5,..G,? ~

v -

p-'

_ y

~.(6',.

I v } }f( ~Q,; j4 g.

J 4

s s

i 9,a.

- y).f.L t *,(,

)

M y. ;L.i J.

P 1

a u

n 4.;.gy i

i<

1 i

1

...4g, g

y-s>

g. z,,

• p,%{'

a

,f l

e

_ ~<

s

[. j(.

~

+

g 3

h';!7-ag,

'. t.,,

I

'e

,:+

~

s

. - (a

)s (t 2 n

D n=

e A

oy pt e O

Emfa L

sos L

A C f

Cgo

)

A s

nr d

Ei o t

C n

Sit o

Amc PI c

a i

T T e

BLF s

S R

(

E

.E V

M IT S

U i

R OT i

N W

P O

U D

i i

O3 po $w$goF s

sp p

6 8

1 m

,nB

,, e C

STP SENSITIVITY CASE UP.c. = UP.c. X (1.5) X (0.8) s b

6 ps.i-l\\

where 0.8cc3-D/2-D l

./

~ '

-UP

' ~~.,

o i

i i

i

- s\\

TIME (seconds) o d

- DOWN 'i l

i l

O 8

BASE CASE FOS 12 P

s 6

8 SENSITIVITY CASE FOS 21 l

u>

D i

E i

s o

F i

s DOWN.c. = DOWN.c. X (1.5) X (1.0) l i

s b

8 where 1.0cc3-D/2-D i

s t s3 18 psi

' [ s' n

u si 8 cs i si e iil8

'lI

'Q 2-s.I I

.l s

l

a ta d

2 e

1 l

f a

ocs S

S e-S t2 I

S O

a1 O

t F

/

F S1 A

st tn B

d n

m a

e e

on o

n sr R

o Edf o

l E

ep S

ae p

lb m

Aog m

S L

a o C

d o

b c

gle c

F o

Ynw T

E g

g r

Ti e

n pn do si S

S I

i Asi Vnn ai t

t t

uKb m

Com To I

EMiL S B L

I i

S No e

Ao o

E B

S

=[2

SHORT TERtl Pf0GPRi CONCLUSI0tS C0tFIPM POOL SWELL SCALING C0 FIR 1 THREE-DItBEI0tFL POOL SWEli EFFECTS RESOLVE IB&ARD LOAD NTfALY RJLL-SCAll FF0T0 TYPICAL STEN 1 CCfDSEATI0tl TESTS ARE tECESSARY 6

4

-4 m-m-

e m

.e

+

y p:s~,1+ S (F

.,a

,. g i,

g i

\\

f s'

}

j

(

idi i

"P l l) y

,s s.

i m

(

\\

d i

i

)

f e

5 d W

J.L

[l r

, 4 m

_ ~

' ^

~.

-af,y f' g g j ; L i T

~ ' ~ ~ ~

' ' ~ '

L. ;z;., gg

. -(

s 1

g,.

w.t1..

1

\\-

a w

=

%m

'l 'kll"//

- ),

l r-g 4:-~[.//*

7,,

y'id' #

.?

,, ;~

.~

I. R p'sy i.

l

~,-

T

~

lf,. Q, s

a.

i

. y s 's-,

4 7'3 :

w

/

d

'O i

W s

^-

m e

.s e

1 a-g3; i

y w;

%~,

4 *.;

y N

g' g.,

s,

x&g; 37

-$;ksfi g

6 ji c.

^

I s

N I

P90R0 gg

LAWRENCE LIVERMORE LABORATORY 1/5 Scale 90 Torus Sector Main Vent s

I

\\

(j ooa

/

oaaa c,

i' Downcomer Pair Ring Header

L D

L M

k"N E

DL A

S RE R

V D D

I AH G

E 3 2 b

O R

A WT R

L PM P

L U OT M

Y L

DRI R

s L

E EFI T

.N HD C

-T

/

TEA R /

,uNQ. N'

\\

OV F O

OI T H

E S

s 5'

M R S ME E I

I

\\, ' '-

T K

SDT R

\\

h s FSE A

CA

~

/. /, /

OEL M

N RC O

'7 OS L

/

/

/

/

S A N/

F 5 N

I I

RE /

I

/

1 G

/

f AR I

PU R D D O 3 2 MS L

t O S L

i E

L C R P

m u mm]

k @<3n.o u :_

?,,,

n

4

N D

3 OL E

r r

o SA L

o t

A t

c I

c e

RN C

e S

S S

AIM

-2 o

1 0 5 e

A T

P

/

9 7 E

M O 1

N R

A I

O A

R S

P E

T CE x,-

N-A N

N M

O D

I D

O L

T A AI L

A I

D S T 2

L Z

O I

R E

LED L

U FHN A

S C

S OTO S

E R

RC 4

/

P N

O OT 1

L E

L FS G

TE E J_f -

A E

I W

A U T Y

R T Q D

NI

~

S H

~ -

EN s

E C R

~

E P T E

w g m m Q a. O < 3 S y E R

1 d

n:.

'75

.t s

s L

t e

9 st 7

O e

9 t

e O N 1

s Da PO 2

c y

3 1

t I

et i

MIT s

n y

lar a

i c o a

M AN t

s w

re I

(

RF

%n cn GE i

o u

a it s

s

.t OD wa n

t ei s s s ss e d sf et set t RD a

e st t ss v

n l

i P A es b d t eDteeeo t

r i

t t

ad eD 3

DDr e

MO t

a a

o s2De23 el p L

dl a

2l d

R e

a eenm E L e e bl ecl al saa uo l

uiv ). a ca ccs c

T L qt st ana sc2sst E

nvdd2 cd i

s1

- / 55 e

e GW uri 1

41 //f t t

r eie//

1 1 f e

sd g1 t

N S nn nr RLL eg I

1 r

EE PLLDa O

ao oa lPCcLGGELL3T L

e o

o o

e S

,s

.