ML19318C302

Forwards Addl Info Re Safety Evaluation of Facility Special Test Program.Provides Addl Assessment of Potential Upset Conditions Associated W/Postulated Rod Withdrawal Events. Fuel Clad Failure Is Not Expected

ML19318C302
Person / Time
Site:	Sequoyah
Issue date:	06/25/1980
From:	Mills L TENNESSEE VALLEY AUTHORITY
To:	Schwencer A Office of Nuclear Reactor Regulation
References
NUDOCS 8007010327
Download: ML19318C302 (11)

Text

,

- i 7

,-_;c -

r 27e -,-

400 Chestnut Street Tower II June 25, 1980 Director of Nuclear Reactor Regulation Attention:

Mr. A. Schwencer, Chief Light Water Reactors Branch No. 2 Division of Licensing U.S. Nuclear Regulatory Commission Washington, DC 20555

Dear Mr. Schwencer:

In the Matter of the Application of

)

Docket No. 50-327 Tennessee Valley Authority

)

References:

1.

Letter from L. M. Mills to L. S. Rubenstein dated May 9, 1980 2.

Letter from L. M. Mills to L. S. Rubenstein dated April 9, 1980 Enclosed for your review is additional information regarding the safety evaluation of the special test program for Sequoyah Nuclear Plant (SNP).

This supplement provides an additional assessment relative to the poten-tial upset conditions associated with postulated rod withdrawal events.

The conclusions of the assessment are that departure from nucleate boil-ing (DNB) is not expected for the potential events. Therefore, fuel clad failure is not expected.

The original safety evaluation and first supplement were transmitted by the previous submittals (see references). This supplement should be incorporated into Section 4.0 of the Master Plan for the Special Low-Power Test Program for the Sequoyah Nuclear Plant.

Very truly yours, TENNESSEE VALLEY AUTHORITY C3 %

L. M. Mills,,

ager Nuclear Regulation and Safety Enclosure (10)

Aco/

3 II

'""** O 39~l

_, n

v ENCLOSURE 7

,:2. -

'The safety evaluation report and supplement for the TVA special 'ow power tests provides an evaluatfon of each of the ten proposed tests for each FSAR accident initiating event. The results of that evaluation were surcarized by a categori-

.zation, for each test under each initiating event, of the baris, for acceptance.

In the great majority of cases it was concluded, either by reanalysis or by ccm-parison.with 'previcusly analyzed FSAR conditions, that fuel clad integrity would be maintained without need for hperator mitigating action. For the LOCA or steambreak events, it was concluded that the operator would have more than ample time (> 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />) to respond by manual action, e.g., manually initiate safety in-jection, to preclude fuel damage, Finally, in certain other cases, primarily associated with certain inadvertant RCCA withdrawal events, the postulated accident conditions were neither amenable to direc't analysis nor credit for operator intervention.

In particular, the pustulated accMent conditions were outsid.e the boun'ds of accepte'd analysis techniques so that fuel damage was not precluded either by analysis or identified operator action. For these cases,' the basis for acceptability was primarily as-sociated with the low probability of an inadvertant rod withdrawal event during the limited duration of the special tests.

' ~

This' supplement provides an additional assessment relative to the potential for.

.-s.

.:3 and consequences of fuel failure for these "unanalyzed" accident conditions as-sociated with certain rod withdrawal events. This assessment is partially

. based upon an attempt to bound certain effects which may exist for conditions removed from the range of direct model applicability. Additional informatien (attached) is provided for four areas:

.3..

i CG-1.

Thermai margin associated with normal test conditions. '

m 2.

The potential for DNS during accident conditions.

t

~

The clad temperature response assuming that D.1B cccurs.

3.

4.

Radiological consequences associated with presuc:ad gross fuel failure.

'.1 :~s..'.1'.::

. W........z....*:

'* ::.. ~. :' ' u.,. :. :

u j'- s =- dJ q % au.j.'..:-;..

~; '. *

.~.... ~ ~ : ~;

. --,.'. ~..y... < n;,: y - r n4,%l.W.':%.!-: x.i... -. M...; k... :.. t.

_. :-; g.:

a.* *rm.~pMm:e%ws:w,*r e: -.(.-y,..

-2

~..

...n.

=s m es.~~ aL..nuw w..-.n.

~ :.>.> :vn.m iMT k..-s2CT:*rs'~'.n%. tit ':- ':C 2x:v *.'* =.-M: 2 ?.* ' : *= * : R *-" -

~ 4* -

giy :4:ya+:

r4:.i..

4.../. n := ^

-.~

The conclusions of this assessment are as follows:

l.

DNB is not expected for the limiting thermal condition cssociated with any RCCA withdrawal event..

w 2.

Even assuming DNB, there should be adequate heat transfer to prevent

~"

clad overheating.

3.. Fuel. clad failure is not expected.

4.

Even assuming 100% clad failure and other extreme conservatisms, the re.suli;ing offsite dose would be small.

..s DESIGN CONSIDEP.ATIONS Margin to hot channel boiling has been incorporated with all normal test cor.di-tions by establishing a lower bound requirement on the degree of reactor coolant subecoling. For example, the rate of planned oower (1 crease for test ne=ber 8 (stagnant f1'ow startup) is restricted by test' procedure to maintain margin to hot channel boiling. This test requirement assures that pcstulated accidents are initiated from a condition of excess thermal margin.

y.

ONB CCNSIDERATIONS For certain cooldown transients, the conclusion that' DN' B is precluded was drawn based on use of the W-3 critical heat flux correlation. Althcugh the analyses for the cooldown events discussed in section 4.2.3.2. result in mass velocity be-low the range of direct applicability of the correlation, the reactor heat flux was so low relative to the predicted critical heat flux that even a factor of 2 k

would not result in serious concern for DNB for this event.

For the non-cooldown transients the limiting conditions, 'with respect to DNB, are farther away from the W-3 range of applicability because the coolant temper ature is higher and the power-to-flow ratio is larger.

eV-+

F

}

y Cc=parison of the W-3 DNB correlation to low flow DNB test data and correlations (references 1 and 2) indicate that it will conservatively predict critical heat flux at low pressure (E 1000 psi) conditions with low coolant flow, pcol boil-ing critical heat flux values (reference 3) at these pressures are higher than those predicted by'the low flow correlations.

Further review of the data in reference 1 indicates that the critical heat flux at higher pressure is signif-

~

icantly lower than the above data at 1000 psi. The minimum critical heat flux 6

2 of the data set is.16 x 10 BTU /hr-ft for a data point at 2200 psia at a mass 6

2 velocity of.2'x 10 lbm/hr-ft,

Since the exit quality for this data point was 64%, it is unlikely that the

~

reacto'r would be able to maintain a heat flux of that level due to the nuclear f'eedback frcm voiding. The power distribution would tend ^o be peak to the bottom thus further reducing the local quality at the peak heat flux locations.

Also the pool boiling correlations in reference 3 shew seme werease in critical heat flux.above 1000 psia to the maximum pressure of applicability of 2000 psia.

However extrapolation of the pool boiling correlations to a value of zero critical heat flux at the critical pressure (3206.2 psia) would not result in icwcr critical heat fluxes at any pressures above 2000 psia expected during these tests, than showninthedata:etfredreference1.

Since the core average heat flux at 10';

of nominal power (hichest expected power for heatup events) is only on the order of.02 x 106 BTU /hr-ft2 a large peaking factor would be required to put the reactor heat flux as high as the critical heat flux.

For the transients considered the only ones that lead to significant off normal the peaking factors are the rud motion transients. The rod withdrawal from sub-critical is a pcwer burst concern. As such, it is expected that even if DNB oc-curred, the red surface would rewet.

For the rod bank withdrawal, the combina-tion of maximum power and peaking factor would result in a peak power lower than th'e data referenced above. ~Given the lack of data, it is difficult to completely preclude DNB, although a prudent judgement indicates that it is indeed remote.

.)

[:, : :. - :...;.;. -. :q.;. '... - & :..... Q

........i=v I

." 5 :1{-.5.

.7 n..w....~.......,.....a..

.r n..-

.n.

• ~~~~l".29"2. & N O$ 6.~T T l$ W $1' M 4T-[.d D ?"f f i?.W 2 2 ?:

O N "-

~ =~ ?W'%.$b

.;:. ;. v y ;7 ~ v ~ -:.- - - 7. 7

1. .-~

-. 2;

't C1.AD TEMPERATURE C0l!SIDERATIONS

.Should DitB occur, the peak clad temperature reached would depend primarily on the local nuclear power transient following O!!B and on the behavior of the post-DNB heat transfer coefficient. '

~

For a rapid power transient, as is illustrated by the SER analysis for RCCA bank withdrawal from a subcritical condition, the fuel temperature reactivity feedback and reactor trip on a nuclear flux signal would. shut down the reactor before suf-ficient energy could be generated to cause a damaging rise in clad temperature.

In that case, the maximum clad temperature calculated was under 1200*F even as-2 suming an extremely low heat transfer coefficient (E 2 BTU /hr-ft _.7),

~

A'possibly more limiting condition for RCCA withdrawal would be the case in which a power increase causes DNB but would either not result in reactor trip on high nuclear. flux or the trip is delayed.

In the foimer case, a steady state condition with' h'ot spot DNS could be postulated.

In this state the cTad temperature could be calculated given only the total core power, local heat. flux.. channel factor,,

heat transfer coefficient and saturation temperature.

The core power is postulated to be essentially at the power which would cause a r'eactor' trip on high power Range Neutron Flux low setpoint. The trip setpoint

~

~

is at 7% for these tests. To allow for calorimetric errors and normal system errors' trip is assumed to occur at 10% of ' rated thermal power (RTp), unless a large decrease in downcemer coolant temperature occurs during the test. For the cooldown phase of test 98, a temperature reduction of up to ICO*F could occur.

At the minimum ~ temperature of 450'F, using an assumed upper bound deviation rate of 1% of reading per 1*F for the excore nuclear instrumenta-ion, a maximum true core power at trip would be 20%. No other tests with natural circulation are '

expected to result in significant downcomer temperature reduction.

In tests 3

~.

and 5, depressurization to less than approximately 1450 psia could recuire temp-F erature reduction, as is indicated in Figure 1; however, such low pressures are 7

not expected.

e..

Figure 2 shows the relationship of peak clad temperature, local heat transfer co-j l

7 efficient, and the product of heat flux hot channel factor (F ) times core power q

(

=

s..e..

f _4 05..fr..,t,

. 3 4 ;'.,..

C- =.

' -?.,

.. ;;.5,E~,j. ;, E.?,1.. :i;; ;.- -S

- =ga-

- =_u +

= _,

~

.. ;.k'. > 4. c.;.p.*.d u..M n:;...:3.*:

..r

~.g y..,,

• g.. y.

r

' %'.A'9.ijf/h'!."M.7x.WP.r.v

.. hefN. Ff F.5M,;,,.9.a HT..,.M.w.C.g.,2..MP. ? **V..

8 $...,.9..6 9

7

. g,-p._ - -

.(

3-

~,

y.

~

y

..:. v..y

... -. ~

1 (fraction of RTP).

For the event of an uncontro11ed RCCA bank withdrawal, and

. for all tests except the cooldown phase of test 98, the upper bound of this heat flux product is approximately 0.34.

For the single case of the cooldown phase of test 98, the 'mper bound of this product is approximately 0.7.

For the event of uncontrolled uthdrawal of a single RCCA, the value of 0.34 is also bcunding foi all conditions except for ' test 9B beration phase (high' bank D insertion) and l

cooldown phase (high trip error at low temperature), and for these isolated events the bound is well under 0.7.

Using these values, the heat transfer coef-ficient required to keep the peak clad temperature below 1800'F, the threshold of significant heat flux increases due to zirconium-water reaction, can be found from Figure 2.

.4 V'arious film boiling heat transfer correlations have been reviewed to evaluate '

the heat transfer coefficient for post-DNB conditions. Although no correlations were found which cover the complete range of conditions being tested, some data exist which can be extrapolated to obtain representative heat transfer co-efficients. The Westinghouse UHI film boiling correlation '(reference 4), was developed at low flow cor.f.'.tions,similar to those postulated for incidents oc-curring during the TVA tests. This correlation was. extrapolated to the higher

~

pressure conditions of the tests to obtain representative film b,,iling coeffic-ients'. This resulted in a heali transfer coefficient in excess' of [

3"

at 2200 psia and 5% flow with quality between,10-50%. Other film boiling heat transfer correlation's,' developed at higher pressures, were also examined. These correlat' ions were extrapolated down to the.1ower flow conditions of the TVA tests as another approach to obtain represantative fijm boiling.c~oefficients.. Using both l"-

the Mattson et al-(reference' 5) and the Tong (reference 6) film boiling correla-tions resulted in post-DNB heat transfer coefficients in excess of 150 STU/hr-ft,.7 -

2 i

at the conditions given above.

~

~

These results indicate that a clad temperature excursion resulting in fuel darage-i-

is not likely to occur even if DNB is assumed.

D 8

e d.

i

'ee.

4 6.*Ia 4

i'

.._em

DOSE AtlALYSIS-C0ftSIDERATI0tlS

' '" ~ ' ' ' '

^ ' " ~ ~ -

The dose analyses were performed for a hypothet'ical accident scenario using conser-vative assumptions so as to determine an extreme upper bound on postulated acci-dent consequences. The analysis assumed a reactor accident involving no pipe-break with a coincident loss of condenser. vacuum..This accident scenario is representative of the Condition II type events analyzed in the FSAR. The assumptions made in the

~

analysis include:

~178Mwt(5% power) 1.0 dose-equival.ent I-131 RCS activity (tech spec limit) 500 gpd steam generator leak in each SG (tiech spec limit) 100% clad damage and gap activity release 10% iodine /;.oble gas in gap space

,100 DF in steam generators 500 iodine spike factor over steady state 509,000 lb. atmospheric steam dump over 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 1.7 x 10-3 sec/m3 x/Q percentile value The results of the analysis show that the two hour site boundary doses would be 5 rem thyroid, 0.9 rem total body and 0.4 rem to the skin.

The analysis of the accidents has incorporated some very conservative assumptions which goes beyond the normal degree of conservatism used in FSAR analyses. The most k

prominent of these assumptions and a brief de'scription of the extreme conservatism includes:

h 1)

Equilibriumradionuclideinventoriesestablishedat,5dpower.

For Iodines, this requires s 1 month of steady state operation at 5% unintergeoted.

t i

2)

Fuel clad gap inventories at 10% of core inventory, this is a time' dependent, temperature dependent phenomona. At 5% power, very little diffusion to gap space is expected for the short test pericd.

3) 100% fuel rod clad damage.

~'

i l

.-~..

_7_

....2....

4)

Primary to secor.dary leakage to tech spec values. 'Since.Sequoyah is a new

--5;,

plant, no primary to secondary leakage is expected.

If leakage were present,._

it would most likely slowly increase in stepsy +o,,t,ech, spec leveTs. 2,.J_i._

s 5)

Percentile meteorology, there is a 95% probab..ity of better diffusion character-istics and thus lower offsite doses.

For these reasons, in the unlikely event of a potential accident during the tests, the resulting dose,is small, even assuming 100% clad damage and other extreme conservatisms.

OTHER CONCERflS

.The LOCA analyses presented indicate that there is over 6,000 seconds for the operator to take action. This is more than sufficient time for the operator to take corrective action.

Some transients were not analyzed or discussed in this supplement due to the combination of the low probability of the trc7sient cccur-ring and the very short time period of the special tests. This is true for the

~

rod ejection accident. The combination of the low probability of occurring and the bounding dose evaluation for a condition II transient given here indicate that these events do not need to be analyzed. Similar dose calculations have been done for the steamline break accidents which results in scmewhat nigher doses th.an the condition II analysis. These dose results indicate that the fact that the !{IS channels are not completely qualified does not alter the conclusion that the results are bounded.

O

(

i

,, '. 5g

.z..>..w l

.5:.44.-A'.IA.w Ms.j.-@.;..v.4 h.b.2

.. :~

0.g. ~

..;.,:.. :S, x.

..n..e? M.u.;.. :,.,..s ~.m.2 y.~M~ g.:u m ws.s u w c& a.:.m.~.v.+

..;>.ix. M e. nim -.

.+s

m

_g_

~

. - t 4..,. m c'.s *

. : c:J;t a.4,:. :n-y 1.s.;.-

.- s.

e

. ;g -= i-. ~

?-

s

..a

~* -

.. \\.x;-

.p

~~"

REFERENCESI

[..,"

~

1.

J. S. Gellerstedt, R. A. Lee, W. J. 'Oberjohn, R. H. Wilson, L. J. Stanek,

" Correlation of Critical Heat Flux in a Bundle Cooled by Pressurized Water", Symposium on Two-Phase Flow in Rod Bundles, Code H27, ASME Winter Annual Meeting, November,1969.

s 2.

Hao, B. R., Zielke, L. A... Parker, M. B., " Low Flow Critical Heat Flux",

~

ANS 22, 1975.,

3.

Lahey, R. T., Moody, F. J., "The Thermal-Hydraulics of Boiling uatar Nuclear Reactor", knerican Nuclear Society,1977.

4.

WCAP-8582-P, Vol. II,." Blowdown Experiments With Upper. Head Injection in G2 17x17 Rod Array", McIntyre, B. A., August, 1976, (Westinghouse Proprietary) 5.

Mattson,. R. J., Condie, K.' G., Bengston, S. J. and Obench'ain, C. F., "Re-gression Analysis of Post-CHF Flow Boiling Data", paper B3.8, Vol. 4, Proc.

of 5th Int. Heat Transfer Conference, Tokyo, September. (1974).

6.

Tong, L. S. " Heat fransfer in Water-Cooled Nuclear Reactors", Nuc. Engng.

and Design 6, 301 (1967).

.~

s I

Q

1 i

y., - ; p.

.c1:; 4 i:

i.1

. \\ 'O

)

1-

. !..v.,(g.:q;.r-

.h e-

. 'tp.;;] l;.;

i

.1

,'{

• sj.* ;. 3 :'

,3

>:..39 :.

,

:y y

3 i.

c-3 ;q. e:g.

M 100 i:

pp l;, :d h 1

3 e.

~

...n

?

, j p

Peak Clad Temperature a 1800*F L ?,f*;..f.')

i

• 1

' u e

ll

*s -

to s,

/ 1:.

!.t m

,O 80

.g,. t e! 5

. 2r -

a

'f C. e f'; I.,g f

y: Heat Transfer

. ' 2:. :

',.'. % 'ti

.t

. Co:fficient

!P,...

Btu /Hr*F Ft.2 Pressure si 2600 psia r>-

c h... i il '

60 1600 psia

'.(g,3,

7. 2,

W; r

'jh.

ts

+

40

( r,,

. : l

...t.s 20

,+

\\

i N

0

~

0 0.1 0.2 0.3 0.4

0. 5 0.6 0.7 F x Power Fraction F.,

o-Q

• • g h

' FIGtlltE 2 lil Af ll:Af!". flit Col'lI'ICII'llT VS. lil:AT fl.IlX FOR CLAD li TEMPER 6TURE OF (800*F i

j 1

.._..__.__...m.____

___m

. ' Jil df' ; *; *. :.

8 2500 i-

=

=

• .;+

2400 - Limited by

~

.e -

1-PPrimary - P a-

2300,

- Secondary = 1600 psi e

' ~ ~ 1 ':..' g.

98 Cooldown h 1, 2, 7 j

r7 l'l i

9B Boration 2200 p

(2 or 3 )

8 C (1 & 3%)

~

I l-c L'

.(0 - 3%)

. ;p' l o 1.

t Cs ',i '

2100

/

3. - l:~

/

j.

l 2000-0

. i;. I ;

Test No.

4 RCS Power Level (1%)

' ]'j 1900-Pressure.

' l.

l Psig 3, 5 J(3 + 1 1/2%)

~

~,

1800-i a

1700

.n l 1600 I

1500 I

1400

/

i.

/

4 Limited by Hot Leg l

t

/

Subcooling = 20'F at 1.5%

130(,

l

/

5.

./

,r

.s t.

l 120fl t

l 460 480 500 520 540 560 580 l :

. Active Cold Leg Temperature, F

I l

l

1.,

lIGURE 1 flAIURAl. Cll!Cill.Afl0ll lESI C0!!DIT10!!S T

l

.