ML20006D157

Responds to 891226 Request for Addl Info Re YAEC-1683 on MICBURN-3/CASMO-3/TABLES-3/SIMULATE-3 Benchmarking.Hot Eigenvalue Std Deviation on Table 5.7 of YAEC-1683 Reduced to 0.00098 w/SIMULATE-3

ML20006D157
Person / Time
Site:	Vermont Yankee
Issue date:	02/02/1990
From:	Tremblay L VERMONT YANKEE NUCLEAR POWER CORP.
To:	Rosalyn Jones NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
BVY-90-011, BVY-90-11, NUDOCS 9002120161
Download: ML20006D157 (9)

Text

. _ . - -

1 I

r VERMONT YANKEE  !

NUCLEAR POWER CORPORATION l l

j Ferry Road, Brattieboro, VT 053017002

, , , , , -]

ENGINEERING OFFICE

1. 4 580 MAIN STREET

' l BOLTON, MA 01740  ;

(5001779 6711 1 February 2,1990

f. - United States Nuclear Regulatory Commission nyy 90 011 -

Document Control Desk ,

Washington, DC 20555 Attention: Mr. Robert C. Jones, Chief  !

Reactor Systems Branch  ;

Division of Systems Technology

~

Office of Nuclear Reactor Regulation

References:

a. License No. DPR 28 (Docket No. 50-271)

b. letter, USNRC to VYNPC, NVY 89-253, dated December 26,1989, i

Subject:

Res mnse to Request for Additional Information Regarding the Topical Report -

YA)iC-1683 on MICBURN-3 / CASMO-3 / TABLES-3 / SIMULATE-3 -

Benchmarking

Dear Sir:

- Reference (b) stated that the Reactor Systems Branch had reviewed the subject topical report and concluded that additional information was required for NRC to complete its review. As part of Reference (b), NRC provided an enclosure with a total of fifteen (15) questions relative to the .

E subject report, and requested that Vermont Yankee submit responses to these questions within 30 L days of receipt of Reference (b) to enable the NRC staff to complete its review. Reference (b) was li received by Vermont Yankee on January 9,1990. r Enclosed please find our responses to your questions. We trust that this information will assist you in your efforts to complete your review of the subject topical repon; however, if you i have any further questions or trquire additional information, please contact this office.

Very truly yours, VERMONT YANKEE NUCLEAR POWER CORPORATION f(sfl\ . A O* ' s . i leonard A.Tremblay,Jr. V [

Senior Licensing Engineer Enclosure cc: USNRC Region I Administrator USNRC Resident Inspector- VYNPS ', 0 h

.USNRC Project Manager- VYNPS 9002120161 900202 PDR ADOCK 05000271

-p PDC 4r

. - - - - - ~ . - . - -- - _ - - - - - -.

l i

RESPONSES TO REQUEST FOR ADDITIONAL INFORMATION CONCERNING YAEC 1683,  !

SPURN 3/CASMO-3/TAllLES 3/S MULATE 3 BENCHMARKING OF '

VERMOhT YAh KEE CYCLES 9 THROUGH 13 QUESTIONS AND ANSWERS i

1) Q. New will the reduced K,, unoerielnty impect the YAEC Licensing Analyseet A. On Table 5.7 of YAEC 1683, the hot eigenvalue standard deviation is reduced '

from 100281 with SIMUt. ATE 2 to 100098 with SIMULATE 3. Most of this 4 improvement with SIMULATE 3 is caused by a reduction in eigenvalue drift with exposure. As Figure 1.1 illustrates, both SIMULATE 2 and SIMULATE 3 track along well defined paths. It is the consistent behavior of the hot model's eigenvalue path whloh provides confidence in the predictions made during the licensing (e.g., end of ovele exposure, critical rod pattoms for initiating transients, etc.). Sinoe both modeIs have well defined pathways, the hot portion of the  ;

licensing will not be affected by this reduction in hot N unoortainty.

The reduction in cold 4 uncertainty with SIMULA'IE 3 will also not affect the conservatism of the licensing analysis. YAEC does not intend to change the  !

statistical level of confidence or the Technical Specification limit for cold shutdown margin as a result of the reduction in cold 4 uncertainty.

2) Q. Deecrthe the procedune used to a4ust the CASMO 30 pellet deneltles when the calculeled and measured assembly we@ hts for a given Ivel type dIeWien.  !

A. A zero depletion stop CASMO-3G case is run, for the majority lattice of the fuel '

bundle, using the nominal pellet stack density as an initial guess. The output i- from this zero de etion case provides the lattice cell area (cm' and heavy metal _

density (gmU/cm. Multiplying these values by the active fuel noth (cm) produoos an overall HM loading for the assembly (gmU). The latter is compared to the batch average of the as loaded weights. The ratio of weights (as loaded to CASMO 3G) provides an adjustinent factor to correct the next CASMO 3G ,

i input density. ,

In general, one iteration is sufficient. For the fuel benchmarked in YAEC 1683, this procedure resulted in a 0.2-0.3% increase in stack density to achieve the batch averaged as loaded weights.

1 l 3) Q. Will the Vermont Yankee SIMULATE 3 model be applied to fuel types that

are not included in the Cycle 913 data bene and how does the YAEC 1683

benchmeriting justify this application? What are these new fuel typee?

l A. Yes, the SIMULATE 3 (and CASMO-3G) model benchmarked in YAEC 1683 will I L be applied to fuel types not included in the Vermont Yankee (VY) Cycles 913 o data base. Justification is based upon the concept that: This benchmark, along with others performed by YAEC and the code vendor, are consistent over a wide range of fuel parameters; therefore, small extrapolations beyond these ranges should exhibit similar behavior when modelled. 1 l

I i

i

, o. .~

lt is not possble, at this time, to define the future fuel types that W might

. However, the fuel vendors generally introduce design changes in evolu ary steps, which are usually tested by means of lead test assemblies (LTAs). This provides a long lead time In which to test SIMUMTE 3 against vendor results for the LTAs. This should allow YAEC sufficient time to respond L

to any licensing or operational impilcations in an orderly manner.

4) Q. New de the CASN040 and Sn0(JLA1E4 lattlee parametere con 1 pere in the Sn0ULATE 3 "AugNt" calculationt A. The ' Audit" calculation of SIMULATE 3 shows that the CASMO 3G oross-sections, etc., are accurately transmitted by TABLES 3. For QA wrposes, YAEC has found that the best overall figures of merit are the SIMULATIE 3 generated and M' listed in Audit. Typically, the absolute differences observed values between CA of K,SMO 3G and SIMUMTE 3 are less than .02% for K,, and less th

.07% for M'.

3) Q. Since some "Seewey" exists in the specifloation of the thermelhyttraulle parameters, how were theen pernmetere sedected and how sensitive are the cefoudetions to this sofection?

A. The " leeway" referred to on page 24 of YAEC 1683 consists of the following choloes:

a.) Whether to use plant process computer generated subcooling as an input, or to use the internal SIMUMTE 3 heat balance.

b.) Whether to turn on the spacer correction in SIMULATE 3, or leave it off.

For the benchmark, the plant process computer values of subcooling were used,

, because they were readily available. A subsequent sensitivity study, using the heat balance, gave virtually identical results to those shown in YAEC 1683.

With regard to the spacer correction, Cycles 913 were initially depleted without the spacer oorrection. Turning on the spacer correction resulted in minor

, changes: The spacer correction brought the hot eigenvalue closer to 1.0 and l slightly improved the comparisons to plant instrument (TIP) readings as illustrated in Figures 4.10 4.12 of YAEC-1683.

6) Q. Deecribe the thermel hydraulic adjustment made to account for the spacere.

• Why wasn't this ef0nct accounted for by an adjustment of the local nucient l Pe!9Melefei l A. The EPRI Vold model, used in SIMULATE 3, does not account for the highly localized pressure drop and voiding effects in the immediate vicinity of the speoers. The spacer correction is an adjustment which aooounts for the local l

accumulation of steam above the fuel spacers by increasing the local vold fraction. The local void adjustment decreases to zero a small distance I downstream of the spacer. Since the local accumulation of volds above the j spacer is a physical thermal hydreon c phenomenon it would not be appropriate to model its efisct with adjustments to the nuclear parameters.

I

• I 5

- 1) Q. New nos the reentor period converted to reecdvlty in the coldcritical measurementet l A. The reactivity correction ( p) for reactor period (T) was calculated using the .

in-hour equation. For large periods, the In hour equation reduoes to:

l 0 he p=E 1 +AT ,

The six Oroup 5, and A point kinetics parameters are generated during the  !

licensing of each cycle. The values of 5, change slightly from cycle to cycle, and -

chance with exposure within a given cycle.' The use of cycle and exposure i spoolfic values for 5, causes some variation in the period corrections used in the -t benchmark. These are plotted versus stable reactor period in Figure 7.1. .

3) Q. What Irlput does FWWR providet .,

A. FIBWR is an approved licensing code which provides SIMULATE 3 with the split l

between active channel flow (moderator in contact with heated fuel rods) and i byMiss flow (moderator outside the channel #us flow up the water tubes). The ,

spit between the two is generated at severa different values of total core flow.

The bypass flow is input to SIMULATE 3 as a table versus total oore flow.

9) Q. What exief and redief distribution of channelbow aes assumed in the  !

Vermont Yankee oniculationst What was the bests for this distribution and did it depend on lluel burnupt .

A. This is described in the answer to Question 14.

10) Q. Shouedn't the TF uncertelnty be 1d5 rather then 1/2 of the rms dINorence between symmetric TFst l

A. Yes it should be 1AE The description of total TIP uncertainty provided on .

l pa0e 44 of YAEC 1683 is in error. However, the values for total TIP uncertainty  ;

shown in Table 5.6 of YAEC 1683 are calculated properly.

Q. New were the non equilibrium state points which were excluded from the I 11) anicuention/tneesurement comparisons identitledt Were other Cycle 013 statepoints exciuded from the benchmetic comparisons and, if so, why were they excluded? Why ston't Cycle 11 and 12 EOFPL cold critical comparisons included?

A. As part of standard oore follow guidelines, YAEC receives TIP data several days to a week following control rod maneuvers at the plant. This data is taken at, or near, xenon equilibrium conditions. The deelslon regarding the xenon equilibrium '

status of these transmitted TIP sets is made by Vermont Yankee (VY), based upon the steady behavior of power and flow versus time. All data provided to

. YAEC by VY, as part of standard core follow, was used in the benchmark. YAEC L did not exercise any discretion in eliminating data from the benchmark.

l Regarding the cold criticals: Cold criticals were not performed near EOFPL for Cycles 11 and 12 because VY did not scram near EOFPL in these cycles. End of cycle (EOC) cold criticals have not been performed at VY since Cycle 10.

i

~ , . .. . _ _ _ _ _ _ _ _ _ _ _ -

i

~

\

12) ~ Q. New ed he ee60uinnonNnensurement dItnerence W the sepecenHod crtNoel compere wish she typical dlNorences obtelned for the in-sequence crtNcel i maneuremerWet '

A. The agaoent rod critical (local critical) is the first critical shown on Table 5.4 of  !

YAEC 1683. The local critical eigenvalue is .99258. This compares to an  !

'3 '

avera0e for all oold criticals of .99680+.00168. Thus, the local critical falls 1

outside the range of the standard devTation.

As a solitary data wint, the deviation of the local critical from the average does not indicate a weamess in the methods, per se. As proof that the overall BWR  ;

methods are working, refer to the SIMULATE 3 topical YAEC 1659. The  !

' benchmark of Quad Cities, shown in Table 4.2 of YAEC 1659, contained 10 local t ortlicals and 8 in- uence criticals. The local criticals avera0ed .998310005. t The in sequence s averaged .9967+.0026. These statistically overlap, demonstrating that SIMUL. ATE 3 local critical results are consistent with in-sequence critical results. l i

13) Q. The SnM. ATE 4 model has essentially no factore M adjusting  ;

. the ce6culenons in order to improve agreement with cycle specific  :

maneuromonte. N the calculation /meneurement difnerenone for a future .:;

cyone incronen above the values glVen in YAEC 1603, how will this inerneeed uncertainty be accommodated? Will the model be adjusted or ,

wit the uncerteenty merpin in the calcuandons be incroceed? Now will this change be documentedT i A. If the odculation/ measurement differences for a future cycle increase above the ,

values given in YAEC 1683, the increased uncertainty would be statistically factored into the data base of those important parametere that affect lloensing. l The model would not be adjusted under these circumstances. Changes in the 7 uncertainty margins, and how they are implemented to preserve conservatism in licensing, would be documented in the licensin0 process.

If, for any unforseen reason, chances to approved methods would be required, these would be reported to the USNRC.

14) Q. Provide a doenited description of the technique used to incorporate the results of the CASn004G calculations for the bowed chennel bundbee In the CASn004G/Sn0ULATE4 model. In addition to the doNection of the channel nells, are the fuel pins in the lattloe also assumed to be displeoed? How i are the eNects of channel-bowing on fuel rod power penking accounted fort

' A. A conscious decision was made, when building the model, to keep it as simple as possible, while including the majority of known physical effects. With regard to channel bowing these physical effects are:

E a.) Axially, the channel provides structural rl0l dity to the fuel. The spacers are in virtual contact with the channel and move with it, b.) New channels have some bow which averages approximately 30 mils. The manufacturer pre-orients the bow away from the wide wide comer to preclude any control rod interference.

c.) The fast flux gradient in D lattices creates a positive feedback in differential Zr growth that causes bowing away from the blade to increase with exposure.

l l

I u .w _ . _ _ _~, __ __ _ _ _ _ _ _

h I

I d.) The bowing in exposed channels has an axial shape that seems to follow the I axial power distribution history.  ;

Effect (a) is implemented in the model by having the entire lattice, pins and all,  ;

deflected in CASMO 3G by the amount of bow. This is achieved in CASMO 3G .

by simply adding to the dimensions of the wide water gap, while subtracting an  ;

equal amount from the dimensions of the narrow gap. Since the pino keep the ,

same geometry relative to ths channel walls, they have effectively moved with the ]

channel-bow.

l Changing the width of the water gaps affects the magnitude of local peaking as a function of bumup. The local peaking factors that CASMO 3G produoes for the  ;

deflected lattios are input, via TABLES 3, directly into SIMULATE 3. Therefore, .

the effects of channel-bowing on fuel rod power peaking are accounted for.

Effect (b) is implemented by modelling the bow In CASMO-3G starting at zero )

exposure.  ;

Effect (c) is partially implemented by using an amount of channel bow which is  ;

greater than the as menufactured average bow (approximately 30 mils). As  ;

stated in Appendix C of YAEC 1683, a value of 40 mils was arbitrarily selected by the investigators to approximate the core average channel bow. Thus, some amount of increase due to bumup is included in the model. However, further change in bow during the cycle is not modelled. The amount of bow is kept 1 fixed for the lifetime of the lattice.

Effect (d) is not included: A detailed axial shape was omitted to keep the model ,

simple. The middle lattices of each fuel type are modelled in CASMO 3G with ,

the same deflection for the 0%,40% and 70% void depletion cases and branches. Figure 14.1 shows what the effective axial representation of the bowing is in the model. For axially zoned fuel, the top and bottom fuel zones are assumed to have no deflection. All middle fuel zones are run in the .

CASMO-3G model as if they were uniformly deflected by the full amount of the assumed bow.

l To summarize the answer to Question 9, based on the details provided above:

The axlal and radial distribution of channel-bow in the model is uniform. The amount of bow (uniform deflection) modelled is arbitrarily set at 40 mils. This amount of bow exceeds the as manufactured amount; therefore, some of the increase in channel-bow with bumup is included in the model. However, the '

assumed amount of deflection in the model remains fixed during the cycle.

16) Q. in view of the wide tenge in coolant temperatures during the cold critical e L tests (Table 5.4), is a temperature correction applied to the cold critical s@envalues of Figure 5.2?

A. The temperature of each cold critical statepoint is explicitly input into the given '

SIMULATE 3 case. This is sufficient to al;ow SIMULATE 3 to interpolate between l cross section sets at 68'F and 300'F. No correction is made to the final SIMULATE 3 answer except for the reactivity adjustment for reactor period (see answer to Question 7).

y n ,. y ,,--,,v, -~r- .,-<+e -- ,-----,--w-v-- -- -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - -

T; \

.~-

Pfguro 1,1 i

N-t

• U T AWs 1.016 1.0 14 - 5A x CYCn 33 44f 4 y 4

+ CYCW 12 gg ,.

'2 , cr,w ,,

1. ~+ B &x O CYC2 to +

ana c 1.010-O CYC W g

++ d;

,nane c

,,, .ejp.g o stuvurE--2 p a n- -

1.00a =

+p ,g' Y

+

fa h n

@ 1.00e=

c a %gga g

g*x xN mT "(*x # **,

j-1.004 .

%<w >W _

f a 1.002 =

0

,NN' gg.).t 1,000- 3 d O

-- ++ l CDO c

q: 4. xqbs *e.%

0.998- k h} o xs .

xx '% SlMUMTE-3 x y A

0.996- -

0.994-9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 N.0 18.0 19.0 20.0 21.0 CORE AVERAGE EXPOSURE (GWD/MT)

I

• Figurs 7.1  !

{

l REACTOR PERIOD VERSUS THE PERIOD WORTH USED IN THE VY SPECIFIC BENCHMARK OF THE SIMULATE 3/CASMO3 MODEL 1 0.001 e <

1 i 0.0009 0.0008-e e

0.0007 e ,;

e e I

{0.0006 l I o

_ 1 e

0.0005 -

oe g I 8 -

ot '

@ 0.0004- ,

e  ;

e, 0.0003 g

l 0.0002 o

0.0001 0.0 .... .... .... . . . e. .... .... .... ....

0 50 100 150 200 250 300 350 400 450 500 REACTOR PERIOD (soc)

{ s s . .* .

Figuro 14.1 l

)

i UNIFORM DEFLECTION OF THE MIDDLE ZONE l AWAY FROM THE WIDE-WIDE CORNER i l

no bow actual model l I '

!' Top l I l l I

\

~

Middle i l

1 L Bottom I l l l l

v Simple Model For Channel Rowing

,