ML19224D163

Notifies of Status of NRC Knowledge,As of 790523,re C Michelson Jan 1978 Rept,Decay Heat Removal During Very Small Break LOCA for B&W 205 - Fuel Assembly Pwr

ML19224D163
Person / Time
Site:	Crane
Issue date:	05/23/1979
From:	Advisory Committee on Reactor Safeguards
To:
Shared Package
ML19224D160	List: ML19224D159 ML19224D163 ML19224D165
References
ACRS-SM-0131, ACRS-SM-131, PP-790523, NUDOCS 7907110014
Download: ML19224D163 (6)

Text

.

m g-v.

c.

e y r,2v.. +

c.+3:n 3 ~ u o.b,v =: ;.y Q. T,; C ".' ' '.-

. v: 7.

~._

f.~

. m --

yG ?ly ',1'

).:;"'s' V ry.

(. ;: '

)h'.,_.l(,b.i.F

- i

h., @. '-

$ : '[.-

z STATUS OF Kt0'lLEDGE Ill THE OFFICE OF fiUCLEAR REACTOR REGULATIU1 AS OF MAY 23, 1979 C0tlCERillf;G THE JAtlUARY 1978 REPORT BY C. MICHELS0ft EtiTITLED " DECAY HEAT REMOVAL DURIrlG A VERI SMALL BREAK LOCA FOR A B&W 205 - FUEL ASSEMBLY PWR" As far as-we can ascertain at this time, the f!RC staff first received a

1. s gi

\\

'S copy of the January 1978 Micheison report early in April,1979 (enclosures 5

\\{ g 4 s.

h 4 % y.,.g 2,3,4,5).

The report was not formally transmitted from TVA but informally 3

T F

provided by Dr. Michelson upon request of the staff. More recently, it tkT A { TN

has been learned that a copy of the Michelson report was formally transmitted S

$J to B&W by TVA in April 1978, and copies apparently were available to the I

ACRS. This action is presently being investigated by the tiRC's Office k

of Inspector and Auditor.

It is clear however, that the January 1978 u

4 N[b report had not received formal flRC staff review prior to the TMI-2 accident.

9.na% We have also ascertained within the past few days that two handwritten documents which were apparently drafts of the material which eventually became the January 1978 Michelson report were informally provided to a member of the f1RC staff in late 1977 or early 1978 by Jesse Ebersole, Mr. Michelson's supervisor at TVA and a member of the ACRS (enclosure 6).

The staff member recalls discussing the general areas of natural circula-tion and the effects of noncondensible gases in about that time frame with Mr. Ebersole. He does not recall responding formally to the handwritten naterial provided by Mr. Ebersole.

In January 1.978, that same staff member originated a Reactor Systems Branch review reminder (enclosure 7) which in part treats the concerns raised in Mr. Michelson's report of January 1978.

From our review of the letters between TVA and B&W now available to us (enclosure 8), the Michelson report apparently was not considered by TVA 260 Og 7907110M EliCLOSURE 1

ymvmw...m n..=

t*

y, -n -

~,&.% x ;, 1 '..':h;," M '

-ri J V -.;P n- % ies. V. x m c'

?"'~

S,

~1w + + -,

' ~. V T.i.i m.

S..

$. z.N.& '.'. '

f *_*.

~ - 7'~-

~~. N

~, - -

.c g.

'; ^

2 to identify any specific safety problems, but rather to identify a number of concerns regarding core cooling during very small break LOCAs.

Exchange of technical information,' including concerns such ac in the Michelson report, is frequently carried cut between the vendors and the custcmers without flRC involvement.

If the concerns identified in the January 1978 Michelson report were subsequently ~ determined by B&W or TVA to involve defects which could create a substantial hazard, then they should have reported them to MRC.

Since TVA apparently did not know initially if any safety problem existed, and B&M, in a letter' to TVA on January 27, 1979, subsequently asserted that none existed *, neither organization apparently believed it was necessary to report these findings to fiRC since such reporting did not occur.

However, flRC's Office of Inspecto'r and Auditor is conducting an independent investigation in order to determine if failure to report the flichelscn conclusions to ilRC by either organization constituted a violation of the requirements of 10 CFR 21 (Reporting of Defects and Noncompliance).

The January 1978 report by C. Michelson documented' concerns regarding the ability of the core to remain covered for breaks in the B&U primary coolant system smaller than those breaks normally analyzed for licensing applications.

The concerns primarily focused on the lack of documented information which confirmed that the consequences of breaks presently considered for licensing applications conservatively bound the consequences of smaller breaks.

• In a more recent submittal; letter, J. H. Taylor to R. J. Mattson, dated May 7,1979, transmitting " Evaluation of Transient Behavior and Small Reactor Coolant System Breaks in the 177 Fuel Assembly Pian *" (Volumes I and II),

B&W confirmed the conclusions of their January 27, 479 letter with more detailed evaluations and analyses and updated the bu,. c.aluation of the Michelson report (Volume II, Appendix 5; enclosure 9).

n

,2 6 0 u._

p 6,. m,,; m ' : '

r~ceratr.m ~.e, ~n-me-, m.... m,,.,,..;..,

Tm re s

' 5-;

h,C?C3 .

.Q%jiW W

• q$*'@i.>. &

• r - C "' y ' x.~

w

\\

d G.S'Q? j"' e

.D

& Q:X K

• yg; *? '

c. ;:n. =

~. 2.- ?".

r ct.w w...;

• = -

r ~.-

G,_.,

%"^'

w

t

,.'}_{.5:~

3

.e.

The basis for Michelson's concerns were hand-calculated steady-state mass It did not account for detailed effects (e.g.,

and energy balances.

transient terms and geometry) usually modelled in the more so,histicated computer codes used for licensing.

The four most significant items of the report which had a direct bearing _

on the events at TMI-2 as well as the behavior of B&W plants to small breaks are as follows:

1. ) Very small break plant response differs significantly from plant response to small breaks described in Safety Analysis Reports.

2.) Natural circulation plays an important role in core cooling following very small breaks and could be interrupted because of steam formation in the hot leg piping.

3.) Pressurizer level indicatio'n*is not a correct indication of syste:n

~

water inventory, and 4.) Small break isolation by operator action causing system repressuri-zation with subseugent relief and/or safety valve failure.

The report brought attention to the fact that very small breaks in the primary coolant system behave differently than small breaks previously ar.alyzed and therefore provided an indication that different emergency procedures might be needed for very small breaks. The January response by Sri to the Michelson report did not address this question and no changes were made in the emergency operating procedures. The May 1979 submittal confirmed the behavior of the plants to very small breaks as described by Michelson and provided guidelines for the preparation of emergency proce-dures in the event of very small breaks. These guidelines are presently being adapted as emergency procedures for the various operating B&'./ plants.

260 052

7;ys'yqqqyem?.=wq.ww;g5.n - 2 c~c nug, wn-v-~,m.= r.:m ::- 3 q ;;. X. -

.. cf 3 e

.--5 t~

~

.c.

~

y a, r.

3.

y

.r.I -

h, l N'I["

5.,;

~

'%'#*i

.) 'e '

4 The report also brough attention to the importance of natural circulation for very small breaks. From this, coupled with the thermal-hydraulic behavior of the Three Mile Island plant, it was learned that previous modeling representations were not sufficient, and that additional nom.iza-

~

tion in the pressurizer and steam generator models was needed to more.

accurately represent the expected system behavior.

As a result of these model changes, analyses have confirmed Michelson's prediction chat natural circulation could be interrupted. Howei r, these analyses also showed that core uncovery does not occur for any of these very small breaks.

In the TMI-2 accident, the pressurizer leve' indicated the pressurizer was full of liquid. The operators mistakenly interpreted that to mean the system was full of water, and shut off the high pressure ECC injection. The indication of a full pressurizer while other parts of the primary system may be voided could also occur for small breaks analyzed for licensing. As stated in the May 1979 submittal, additional operating procedures will be given to all plant operators such that system pressure will be a main meaiurement system for inventory determination.

In addition, hot and cold coolant loop temparatures would also be used.

In the TMI-2 accident the power-operated relief valve on the pressurizer failed opsn during overpressure. Subsequent isolation of this failed valve with an upstream block valve resulted in break isolation.

While the events in the TMI-2 accident did not follow the sequence postulated by Michelson, both valve failure and " break" isolation did occur.

260 053

' =

iDf 7;.'CQm M.:' D'8?-? 4.a

^..Wjac f*.

~-- Q'i~l.i.

kW'4 1:

'~'

TM P ',

%V,^. * -

~.%.

XM

^

l Q, h._

'. ;..Y. * '

!(.;, _ y m.

t.7.

j i..T r }'q,.3 5

The isolation of a break

• is not specifically considered in safety analyses.

B&W stated that the isolation of a small break and subsequent repressuri-zation does not prodcca a less safe condition than not isolating a break.

This is because any repressurization that results in relief and/or safety valve opening or failure is bounded by small break analyses sizes slightly larger than the valve opening size. The staff agrees in p inciple with

.~

this explanation. However, we will require all applicants and licensees to analyze very small breaks which exhibit repressurization with subse-quent pressurizer valve failure as part of their evaluation of plant response t.o small breaks.

The significance of the Michelson report findings to conclusions regarding the acceptability of consequences due to small breaks in B&W plants will be addressed in detail in a stafbeport to be issued shortly. This is one of the considerations requiring resolution before restart of the presently shutdown B&W plants.

The key conclusions of the staff evaluation of the January 1978 Michelson report and the B&U response to the report are as follows:

1.) The overall behavior of the plants to small breaks was shown to be, for the most part, consistent with the beba'ior as predicted by

!!ichelson and, within the expected accuracy the B&M analyses sub-stantiate Michelscn's hand calculation results; 2.) This behavior did not result in unacceptable consequences and the core is not calculatsd to uncover for the small break accident scenarios postulated by TVA (Michelson);

260 05?

.a :v...

e,,,,.a.~....,.

,,,.,J (p

.,, ~. - -

~ +

,..,Yi.J-U-

,s

.. * **'^

.e g. ;,.,),

.s.

_1,--

.r...

.~-

~.

Y.'-

-;!+ - l,.

.,s 6

3.)* Applicants and licensees will be required to include, as part of their ongeing re-evaluation of plant response to small breaks a) analyses of breaks which exhibit repressurization with subsequent pressurizer valve failure, and

')

documentation of analyses and data which support the conclusion-o that steam condensation-induced structural loadings are bounded by the large break LOCA structural loadings.

260 On

D * '"

, ' [- _- R.. r.2@ ';. '; lj ~

T '.'. %,... -.~'.c

[@,

, uy - ;},.O.'2..'

. N ~r.

+m.7., -

',3 a.: -

le's.T F 0 STA

~.-

t &c

..a;

a..

x.

~

g 3.g. y.:.J "

• e,

] M.'

i 7 2,.".

,.,y fJUCf. EAR REGUI.ATC::. J1.7.;isslofJ

/*.*

.[!

]

1*/ ASH 1.'iGTON, D.

• 1551-

.fi.f. #

1

~

• ,e

,,, g

. jai! 10 tid I:0TE T0:

RS3 ltembers

'(Oit:

T. ii. trovak

.UBJECT: LOOP SEALS lit PRESSURIZER SURGE LIrlE Loop seals ir. the pressurizer surge line are used in some plant designs (noted in B2.1).

Under ordinary circumstances, these configurations are inconsequential because the saturation temperaf.ure in the pressur. zed

1. -650 F) is the highest temperature in the primary system.

However, 0

.inder upset conditions (such as prolonged relief valve opening) and "ccidents where significant voids are formed in the primary system, it

.ay be possible to end up with a two-phrse mixture in the pressurizer that is not a' the highest temperature 1.. the primary system.

Under-these circumstances, additinnal loss of primary system inventory or shrinkage in t';e primary system may not be ' indicated by pressurize:-

~

evel.

This situation has already occurred at Davis Besse I when a

.elicf valvi.

tuck open.

he loup sea! results in a nanometer.effe.ct as s::own in Fig.1.

If there is sattrated itearn at 1200 psi in the hot leg pipe and the level in the rcssurizer is 60 feet, the pressuri er pressure, P, would only have to i

z t e about IDd psi, which corresponds to sat eation temperature about 50F

' clow that in the hot leg.

Thus, the pressurizer temperature does not l

nave to be sirfr.ificantly lo.ver than the temperature in the primary tystem.

A1:. hough the s3fety andyaes do not require termination of the makeup

,ystem, nperators would control makeup flow based on the pressurizer

, level as par t.f their normal procedures..As a result, under certain onditions w: cre the pressurizer could behave as a mancmeter, the operator could trreneously s. it off makeup flow when significant void occurs else-i there in the system or loss of inventory is continuing.

It is recommended that the bases for the design requirement be studied

^

arefully for all CP reviews with the object of Jetermir-g if the loop meal.an be eliainated.

For OL reviecm, procedures shoo.d be reviewed

.0 casure adequate information before the operator terminates mkeup f!cu.

O DUPLICATE DOCUMENT Thoma a

React' } Entire docuraent previously entered rnclosure:

into system under:

ANO ] h b l} b b I d

-runtact:

Sandy Israel,f;RR No. of pages:

2 2 b rd r rdJ 49-27591

_