ML20245H313
| ML20245H313 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 06/23/1989 |
| From: | Papanic G YANKEE ATOMIC ELECTRIC CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20245H320 | List: |
| References | |
| BYR-89-108, NUDOCS 8906290458 | |
| Download: ML20245H313 (600) | |
Text
{{#Wiki_filter:- i l l YANKEEATOMICELECTRIC COMPANY
"'C"*,f.s?l7g;"l"
?Y 580 Main Street, Bolton, Massachusetts 01740-1398
.Yau@xme June 23, 1989 BYR 89-108 United States Nuclear Regulatory Commission
-Document Control Desk Washington, DC 20555
References:
(a) License No. DPR-3 (Docket No. 50-29) (b) Letter, USNRC to Yankee Atomic Electric Company, dated H October 14, 1988
Subject:
RELAP5YA
Dear Sirs:
In Referenc'e (b), the USNRC accepted, for referencing in licensing l applications, Yankee's computer program "RELAP5YA". Additionally, it was requested that Yankee publish an accepted version of-the program manual with review questions and answers. Attached to this. letter is the manual. We trust that you will find this satisfactory; however, if you have any questions, please contact us. k. A Very truly yours, j YANKEE ATOMIC ELECTRIC COMPANY l a .. George .apanic, r. ! Senior Project ngineer Licensing-GP/ pac /0411v Attachment ge/
'f/
0 l' F h__m___.__. _ _ _ _ _ __ _ . . _ . - _ _ _
l
.- l O
l l i Response to 197 Questions on RELAPSYA provided in Letter, USNRC to YAEC and MYAPCO, dated May 11, 1984 l s. b J
\
!I i
i
'h l-l l
i
.V O
RESPONSE TO 33 QUESTIONS ON RELAP5YA March 1, 1985 l l' l l-I O
-/N 1. CONDENSATION HEAT TRANSFER AND NONCONDENSABLE GASES 5 ). Clarify how the condensation-heat-transfer correlations given on (Q.I.2)
Page 45 of Reference 8 for laminar flow at an angle to the vertical and for turbulent flow were derived from the reference cited on Page 45 of Reference 8. (A.I.2) Reference 8 is: V. H. Ransom, et al. , "RELAPS/ MODI Code Manual Volume 1: System Models and Numerical Methods" EG&G Idaho Report NUREG/CR-1826, Volume I (March 1982). i The reference cited on Page 45 of Reference 8 is: J. C. Collier, " Convective Boiling and Condensation", McGraw-Hill Book Company, London (1972). The equations in question are: Laminar condensation ,
,1/3 fp(fp ->$3 \)k g Cos & m
%)k= i766 _
4 %e Y$ _
\
Where $ = angle to the vertical. Turbulent Condensation . . ig !. 'l QF_ 095 Y% " 25 (2) W *1 T3h____IV3 p> g In Equation (1), there is an error in documentation. The coding (subroutine CONDEN) shows that Equation (1) is computed as:
- 3 fp M hp Mo 203%(M/DL) (3) 4 t. k = 1 7 6 6 _
GDe My . where DL = length of RELAPS volume = AL and DZ = O *E: * $[2. with A Z = elevation change of RELAPS volume and g = gravitational acceleration
= 9.80665 m/s2 , in 52. units used by the code Equation (3), above, can be derived as follows:
From Collier Chapter 10, the laminar condensation heat transfer coefficient for film-wise condensation, designated as Tig, is given by Equation (10.63): f,M %&% 3-%
; o 925 it* m r, O _
p h /"'N , whereh = angle with horizontal axis {g=condensateflowatdistance2fromthetop'per. unit film width l Hence fg ~ G B D C D,
~~
~ (5) gp:e 4.
Also, sin O = 62I (6)
, AL l Where h2 = elevation change of volume AL = length of volume substitutingforfgandsinein(4), 1[3 kb = o Cl26 4 f3 M i fe} la*lati (7)
G De \A RELAPS uses D2 and DL where DL = hL-and D2 = h2 . L 2 Hence = g DL OL In SI units used by RELAPS, g = 9.80665 m/s2 0.2039432 (D2/DL) (8) Hence = Substituting (8) into (7), noting that 4 1/3 can be taken out of the bracketted term: , fp af $ kg o 203%S2.(,DZ/pt}' (9) k ' = 1 4 6 2 3 '1 5 GDe Ms. This is for pure laminar flow in smooth films. It is known (Collier, Page 331, Section 10.4.3) that even at low film l i Reynolds numbers, experiments show wavy films. Hence, it is suggested (Collier, Section 10.4.3) that the above heat transfer coefficient be corrected by a factor of 1.2 to account for turbulence enhancement. Multiplying Equation (9) by 1.2: ,, i fprg4,p.2om.u.cta/m.y Ep_- 1762 I (10) _ GDe 4p , If can be seen that Equation (10) derived above and Equation ; (3), which is coded in RELAP5, are virtually identical. h
r , . p_ For turbulent condensation, Collier gives in Equation (10.97): P/2 gz
$y4b o.o 65 ny p %.
pg
& 5 t where-T {is given by Equation'(10.98) as:
- 2.
3 (12) T .L' --- S
- 2. T3 4 wheref}is'.an." apparent"interfacialfriction' factor evaluated _as for single phase pipe flow at a mean vapor mass velocity G g.
Substituting Equation (12) in Equation (11), and using Gg=
'/2
}f Y k = o o65 P
- b f r n3)
For single phase vapor flow at mass velocity gG , the friction factor in turbulent flow can be written (McAdams) as: 1 00% "" hCi Re 2o a where Re g a b N Substituting Equation (14) in Equation (13) and using p ._ C pgg, j O GS)
= o.065 [ l M1
- N .t #
,(I lV3{
t 1)e/p3}02 Comparing this to Equation (2), it can be seen that there is a l small difference - the exponent on the Reynolds number is 0.25 instead of 0.2. The res*J1 ting difference in ht in Equation (2) will be: . 025 1/2 a {4 - Rg3 - 8,e .o2.5-
~
hp ~ , Re3 o 2. S The' impact of this is expected to be small, but this difference will be corrected by using Equation (15) instead of Equation (2) in the subroutine CONDEN in RELAPSYA. i
(l Vfj- 7~ 7 = -- - --- 4
-r y <
, c.;,; - -; m- ,
, I t
e i
?
,, , 2.
.1- 4, ,
!O f
,jQ110).. . . Deleted; 1.. . .
e l
Reference:
Memo, .Ed D. Throm, ~ Reactor System Branch,. DSI ,to '
. Brian W.~- Sheron, Chief, Reactor System Branch,. DSI, "Sutmary. of.
- 06/22/84 YAEC-NRC Meeting.on RELAP5YA PWR SBLOCA EM Analysis".
July.17, 1984. J.6 b 5 I
I E F l [.L.%
' - v n,.
p..; e 1_ ' '] d II. NONEOUILIBRIUM' EFFECTS- .
) i 4
(Q.II.4); Deleted. ,j.Q
Reference:
- Memo,' Ed 0. Throm. Reactor System Branch DSI to
- Brian W. . Sheron, Chit f, Reactor System Branch, DSI,' " Summary of :i 06/22/84 YAEC-NRC M.'d inr, on RELAPSYA PWR SBLOCA EM Analysis".
July'17..1984.
?
O l '. I l l l --_ _ - - __
~' t, . l l IV. STEAM GENERATOR HEAT TRANSFER
\s ,/ ~ Patas 12-13 of Reference 16 describes the process Sandia used to get an initial steady-state for the LOFT L3-6/L8-1 calculation. "The steam flow valve was controlled to match the steam dome pressure using an exponential relaxation scheme. The desired secondary pressure had to be reduced to its
' lowest possible experimental value to yield good primary side temperature agreement, and that was only after manipulation of the secondary side heated equivalent diameter (which helps control the temperature gradient across the U-tubes). Using the strict geometric definition of heated equivalent-diameter, the primary side temperature would be approximately 5 K too high for a given' secondary side pressure and saturation temperature. Lowering the equivalent diameter by about an order of magnitude (a number based on the U-tube wall-to-wall spacing) resulted in cold leg temperatures within the high side of the experimental uncertainty when the secondary pressure was specified at the low end of its uncertainty."
(Q.IV.5) Clarify what hydraulic diameter will be used for the secondary-side tube bundle for SB LOCA calculations. (A.IV.5) The actual hydraulic diameter will be used. Also see A.IV.7. (Q.IV.6) Clarify how the choice will affect the friction and heat transfer calculations. (A.IV.6) The actual hydraulic diameter is used. See Sections 2.1.3.3 and 2.1.3.6 of Reference (IV.1) for its impact on the friction and
/~'N heat transfer calculations. Also see A.IV.7.
(_,) (Q.IV 7) Clarify if the new nucleate boiling correlation discussed on Page 176 of Reference (IV.1) eliminates the need to modify the hydraulic diameter. (A.IV.7) The new nucleate boiling correlation discussed on Page 176 of Reference (IV.1) does eliminate the need to modify the hydraulic diameter. The purpose of incorporating the new nucleate boiling correlation in RELAPSYA was to provide a better estimate of the magnitude of the heat transfer coefficient for PWR steam generator conditions. With a more realistic heat transfer coefficient, a reliable primary-to-secondary heat balance can be achieved without artificially modifying the input parameters (hydraulic diameter, cold leg temperature, and secondary pressure). Reference (IV.1) R. T. Fernandez, R. K. Sundaram, J. Chaus, A. Husain, J. N. Loomis, L. Schor, R. C. Harvey, and R. Habert, "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description" Yankee Atomic Electric Company Report, YAEC-1300P,
; ) Volume I, October 1982 (proprietary).
\/
_4
e
.,-q. V. SYSTEM VERIFICATION AND OTHER EXPERIMENTAL VERIFICATION f \
h/ . Clarification about why the' secondary pressure for LOFT Test L3-6/L8-1 was i underpredicted by such a large amount was requested in Section IV. Additional requests for. clarification are given in this section. 1 (Q.V.5) Clarify why the HPIS flow was injected directly into the downcomer rather than into the cold les as shown in Figure 5.3-1 of Reference 12. (A.V.5) In LOFT Test L3-6/L8-1, the HPIS flow was injected directly into the downcomer as described on Page 7 of the Experiment Data
. Report [ Reference (V.1)]. The RELAPSYA model shown in Figure 5.3-2 of Reference (V.2) represents-the actual test configuration. "The LOFT facility has the capability of injecting the ECCS flow at various locations. In Figure 5.3-1 of Reference (V.2), the more common cold les ECCS injection location is shown; however, the figure was only provided to show the major. components of the LOFT facility.
'(Q.V.6) Clarify how the calculated result compared with the data for loop flow, HPIS flow, hot-leg temperature, and core decay power.
(A.V.6) A comparison of additional parameters for the LOFT Test L3-6 prediction are presented in Figures A.V.6-1 through A.V.6-3. Figure A.V.6-1 compares the measured and predicted hot leg temperatures. The comparison shows good agreement for the first' 7-sg 1,000 seconds. Beyond 1,000 seconds, the RELAPSYA calculation () drops.below the data following the same trends as the system pressure; shown in Figure 5.3-3 of Reference-(V.2). The effect ~ of this underprediction of the Primary System pressure is also chown by the slight overprediction of the HP13 flow beyond. 1,000 seconds illustrated in Figure A.V 6-2. The decay heat curves are compared in Figure A.V.6-3. It should be noted that the data represents a calculation of the decay heat using the proposed 1978 ANS decay heat standards. The results of this calculation are presented in curve form in the Experiment Data Report (V.1). A comparison of the two results reveals the RELAP5YA decay heat values to be slightly higher than the calculated values shown in Reference (V.1). No comparison is provided for the loop flows since data for this parameter was not contained in the Experiment Data Report (V.1).
RH1 0 _ O - - - - - - - 0 0 8 2 L 5 0 0 4 2 _ 0 0 5 3 2 0 0
. 5 0
2 n o 0 i T F t A 5 0 O la T . 8 A 1 Luc D n i l 6 0 a 0 w C 3 5 o A L . f8 i 9 1 l Y F 5 T Y3 S P F 5L A O P 0 I L AT LF 5G0) H P LE . EO R O 4E 1 f RL S T a ( oto N O 0. E no-6 S 0M s 3 OI R A P M
- 5 I 2
1T rL 0 i a p m s t e O o C b 0 CT
. 0 A 1 2
Y - 5 6 P 0 A 0 L . d V. E O A R e r 0 u g 0 i
. d F O
0 o
. d4 0
0
. d 2
F J b O. 3 e; n
. * .d .d 9 3 -
- o oooo nE3h hDS a"E 23t Mekc O 'P' illl1, l llll !l It
,)ii]l
!y O _ - - - _ _ - - _ - - - _
0 0 0 0
- 2
, 1 1
l l U 0 0 e 5 4 2
- 0. ~
0 b 3 2 ~ O. o T F
. b O 0
2 L - n i 0 R 0 e n T r io
. 0 .
A 8 _ tut 1 D a a l 6 r u
- 0 e c 3 0 p l
L . 0 m a 8 e T 1 T C F g A O e 5 Y L 0. 0 ) L P O . 0C T 9ES ot A 1 L ( l E N I R O O S I R A 0. 0ME 0I 2 1T f ot n 6 o-3 o P s i M rL O 0 a t C 0 p
. 0 0 m s e R o Y
1 C T 5 P A 0 1 L 0 6 E . 0 R 8 _ V. A6 A _ -
. _ Y3 5L P
AT 6 0 0 0 e r u g _ LF
= EO RL i
F
=
c 0 0 0 4 0 0 _ 0 2 _ 0 0 i- - - - - - - ~ - -
,g. a g o f o Y o N' ag "g oY o 5 h " k 0N o " o .k o" $ C_ hN s' M@ 8S p0E O a'
1 j!!] il 7 _. O _ _ 0 3 - _ _ _ _ _ _ _ 5 2 0 iI ' 5 2 2 0 iI : ' b 0 2 n R r_ To _ T D Fi R iI ' h Ot _ O 7 L a - 1 l 6 nu ic
- l 3 = _ t a L nC 0 e T eI ' b l l A 5 Y F 1 y5 O aP L ) cA -
eL O C DE T E R 0 S f
' 5 ( oot N iI 2
D n O S 1 E M o6 s-
~
I I i3 R T rL R a P 0 pt M b ms O iI ' 0 oe CT C 1 A 3 Y - S 6 P A 8 0 V. L eI 5 A E 7 R e r u g 6 i R - 0 F Y 3 iI 5 L 8 d P 5 R T L F E O R L 8 o ii d mX 2 L _ _ _ _ _ _ .
; n_ 4 . _ _ . _ . . O 5 0 5 0 5 0 5 0 5 0 5 0 7 5 2 0 7 5 2 0 7 5 2 0 2 2 2 2 1 1 1 1 0 O _a[$* Eog k'
7l!l! l -
r , (Q.V.7) Clarify how the value for the two-phase discharge' coefficient
/~'N was chosen.
l .; l' ' - Q,1 (A.V.7) In the RELAP5YA calculation, a two-phase discharge coefficient of 0.75 was chosen since it resulted in an adequate calculation of the break flow and at the same time a reasonable prediction of the system pressure. For licensing calculations, YAEC will use the Moody critical flow model as per 10CFR50, Appendix K Requirement I.C.1.b. Additionally, a break' size sensitivity study will be performed. Therefore, the effect of break flow, or the magnitude of break flow, on small break LOCAs will be addressed in the licensing calculations. For this reason, the objective of the RELAP5YA prediction of LOFT Test L3-6/L8-1 was not to assess the critical flow model. The critical flow model has been assessed and the results of this assessment are presented in Section 2.2 of the RELAP5YA Manual (Volume III). The objective of the LOFT Test L3-6/L8-1 prediction was to assess the system calculation independent of break flow. Therefore, any discharge coefficient that resulted in an adequate break flow calculation was considered acceptable. (Q.V.10) Deleted
Reference:
Memo Ed D. Throm, Reactor System Branch DSI to Brian W. Sheron, Chief, Reactor System Branch, DS1, " Summary of 06/22/84 YAEC-NRC Meeting on RELAPSYA PWR SBLOCA EM Analysis".
}
J July 17, 1984. (Q.V.16) Deleted
Reference:
Memo, Ed D. Throm, Reactor System Branch DSI to Brian W. Sheron, Chief, Reactor System Branch, DS1, " Summary of 06/22/84 YAEC-NRC Meeting on RELAPSYA PWR SBLOCA EM Analysis" July 17, 1984 References (V.1) Bayless P. D. , and J. M. Carpenter, Experiment Data Report for LOFT Wuclear Small Break Experiment L3-6 and Severe Core Transient Experiment L8-1 NUREG/CR-1868, January 1981. (V.2) R. T. Fernandez, R. K. Sundaram, J. Chaus, A. Husain, J. N. Loomis, L. Schor. R. C. Harvey, and R. Habert, "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description", Yankee Atomic Electric Company Report, YAEC-1300P, Volume I, October 1982 (proprietary). Ns .J _.r - ._ - - - - _ _ _ . . _ . . - - _ . - - _ _ . . _ - _ VI. FLOW DECIMES
\
U )' .(Q.VI.12) Deleted
Reference:
Memo, Ed D. Throm, Reactor System Branch DSI to Brian W. Sheron, Chief, Reactor System Branch, DSI, " Summary of 06/22/84 YAEC-NRC Meeting on RELAPSYA PWR SBLOCA EM Analysis". July 17, 1984. (Q.VI.35) Deleted thru (Q.VI.39)
Reference:
Memo, Ed D. Throm, Reactor System Branch DSI to Brian W. Sheron, Chief, Reactor System Branch, DSI, " Summary of 06/22/84 YAEC-NRC Meeting on RELAP5YA PWR SBLOCA EM Analysis".. July 17, 1984. ( ( I g V11. CORE STEAM COOLING 'l l
.The FRIGG loop tests and GE level swell experiments were used to assess the j 'b/
interphase drag models. Page 7 of Reference 12 states for the FRIGG tests: ~
.j "The discrepancies between the data and the predictions at low void fractions, seen in Figures 2.1-4 and 2.1-6, are due primarily to the lack of a subcooled i boiling model in RELAPSYA." f i
(Q.VII.1) Clarify how much difference the subcooled boiling model would make in the results. 1 (A.V11.1) The subcooled boiling model would allow for calculation of void fractions greater than zero even when the equilibrium quality is j less than zero. Currently, voids can form only when the equilibrium enthalpy of the fluid in a particular volume exceeds the saturation liquid enthalpy. Subcooled voids could be of interest under steady-state conditions. However, for LOCA i calculations where depressurization is accompanied with loss of subcooling in a short duration, the absence of a subcooled void formation model has a negligible effect on the analytical results of importance. (Q.V11.2) Clarify why the void fractions near the top of the bundle are the same for each test for the data and the RELAP5YA result (0.3 in Figure 2.1-4 and 0.7 in Figure 2.1-6) despite the much dif ferent void fractions in the lower parts of the bundle. p Doesn't the fact that these are steady-state cases force t identical outlet void fractions for the same energy removal and same inlet conditions no matter what the interphase drag may be?
~
(A.V11.2) No. For steady-state conditions with a given inlet fluid condition and bundle power, the outlet mixture enthalpy is fixed. For thermal equilibrium at the outlet, the exit dynamic quality is also fixed. However, the exit void fraction will depend on the slip between the phases which is primarily ' determined in RELAPSYA by the interphase drag model. This can be demonstrated very easily by considering the case of homogeneous flow (limiting slip of unity). The exit homogeneous void fraction would be significantly higher than seen in the tests and predicted values. The Bennetit single-tube tests were used to verify forced convective boiling ~ for the region.below the CHF location. For three tests presented on Pages f^ 97-99 of Reference 12 the wall temperatures predicted agreed well with data (
>in the lower end of the tube where the Thom correlation was used and were above the data in the upper end of the tube where the Chen correlation was used.
(Q.VII.10) Clarify what void-fraction range corresponds'to the results presented.
.(A.V11.10). The range of void-fraction for.the three Bennett single-tube tests presented are' listed below and also plotted as a function of elevation in the attached figures.
Test Void-Fraction 5251 0 - 0.95 5276 0 - 0.97
.5359 0 - 0.99
(~
%)
1 i l 1 .sm l
. _ - - - - _ _ _ _ _ _ - _ _ - _ _ . __ _ - _ . _ _ _ _ . --__-____---_-__-_________-__L
680 - j - l' - ' - j_- l *- t j ' l l l l l'j' ' . i cr N : : -
%}
670.:. - 6 - .
- 6601 .
= . . :
- .= ; - .
=
a . 5 : ' k : .. N-660:- * - d [ . i : = 64Db- e sauert anta
,- x acursva sinen-cmcas .
- e neurstasons . :
' ' ' ' ' ' ' ' ' ' ' ' ~
630" ' tb ' 2b ' 3b ' 4b ' 6b ' 6b ' 7b ' 8b ' Sb ' id0 ' 1f 0 ' 1do ' 130 ' 1d' ' 150 (t DISTANCE FROM TUBE ENTRANCE flN.) 10 ' ' l _ L l _ T:[s
.B_ _-
2 .6.. S U g s w O --
;;; .4._
> -)
1
.2__
l 1 00 (i 2b SD 7b 1d0 id5 150 DISTANCE FROM TUBE ENTRANCE .IN.) Figure A.V11.10-1 Comparison of Wall Temperatures and the Calculated Void Fraction in Bennett Run No. 5251
)
\ ~ h I-680 * * * ' ' ' ' ' ' ' ' i i i i i 'a . u . i i ? . li ,.. . ~ ;
-( v l ' 1 \f) .
57 0. .
~
a . .
- k. . e -
we: 680" _e . e
.m -
e . 5 : .. [
- a. . e c . e -
w n --
+- 560. -
a . -
- a. * -
z : - 545"- . e scantit anta x aurnsirwa-c'<n - .
- acursincoas ,
' ' ' ' ' ' ~
530" ' '
'1dO' ' fido ' ' '200
. g' . . . g4 . . . b 6' ' ' Bb ' ' ' 1 d O ' '140'
'1dO' 0! STANCE FROM TUBE ENTRANCE !!N.)
1.0 i i i e i i '.
\.v)
.8..
g .6__ s-u E 4: A
' .O. . l o- .4 -
.2__
1 i
- s ' ' ' ' '
0.0. . 2b - Eb 7b Id0 Id5 180 1*l5 2C 0 OISTANCE FROH TUBE ENTRANCE (IN.1 4 r"
. {) Figure A.VII.10-2 Comparison of Wall Temperatures and the Calculated Void Fraction in Bennett Run No. 5276 j
l 8'
- ' - '- * - ' ' = "' - ' - - - 'i -
a' .' . 590 , e i a i e i a i a i 4- :
~
,- . ~.
l(m.). %'. 5802 .
- ~-
1' , 6703. : s-
.e e e a . .
a: . w I- : . Y 500: .-
- s : e :
d-a
- e .
650; . a soorti sein e . I x neursta :tuon.orma
- . .suem orn, :
5403 (I ' lb ' 2b ' 3b ' 4b ' Sb ' 6b ' 7b ' Bb ' Bb ' id0 ' IfD ' 110 ' id0 ' 140 ' 1 O! STANCE FRon TUBE ENTRANCE IIN.) 10 i i e i i i I
%./
.B._
g .6._ M E O .. E .4..
.2._
0.0 ' 150 id5 Sb 7b 1d0 C , Rb 013TRNCE FROM TUBE ENTRANCE (IN.) r Figure A.VII.10-3 Comparison of Wall Temperatures and the
/ -
.( Calculated Void Fraction in Run No. 5359
,- Page 192 of Reference 10 states: "Once the node is quenched, the internal heat transfer lockout flags (if activated as discussed in Section 4.5) are .( ) reset to zero and the heat transfer lockout flags of all other nodes will no longer have any effect."
(Q.VII.33)- Clarify if the bottom of the core quenches then are the lockout flags turned off for all the hisher elevations? Pages 205-206 of Reference 10 states: "Once CHF is exceeded at a selected heat structure surface, the internal Nucleate Boiling Lockout flag, IEMI, is reset to 1 ("True"). If the heat transfer logic subsequently returns to a nucleate boiling mode (Mode Numbers 2, 3, and 29), the resulting heat transfer coefficient is nultiplied by the value of KMNB, thereby forcing a degraded heat transfer calculation." (Q.V11.34) Clarify how this satisfies the Appendix K criteria of not using the nucleate boiling correlations at that location during blowdown because this just uses a multiplier times the nucleate boiling correlation. (Q.V11.35) Clarify why using the value of 0.5 (Page 265 of Reference 10) for the multiplier would be conservative. Pages 206-207 of Reference 10 states that the nucleate boiling lockout option and the return to transition boiling lockout option can be deactivated: "If the logic contained in the rewet and quench model (Section 4.3) has calculated that the node has quenched." f-i (Q.V11.37) Clarify why, if this occurs during the blowdown phase, that it does not violate Appendix K. ANSWERS A.VII.33. A.VII.34 A.VII.35. AND A.VII.37 FOR SBLOCA Questions V11.33, V11.34, V11.35, and V11.37 deal with Appendix K Requirements I.C.4.e and I.C.5.b.3 (Return to Nucleate Boiling and Transition Boiling Lockouts). These requirements are applicable to the blowdown phase of large break LOCAs in LWRs. YAEC currently intends to use RELAPSYA to analyze all size breaks in BWRs but only rmall break LOCAs in PWRs. YAEC is not required to use the option of locking out return to nucleate or transition boiling for the analysis of small breaks in LWRs. The code will be able to scleet any heat transfer correlation as justified by local fluid and surface conditions. ANSWERS A.VII .33. A.VII .34. A.VII .35. AND A.VII .37 FOR LBLOCAs in BWRs Requirements I.C.4.e and I.C.S.b.3 will be complied with for the analysis of large breaks in BWRs. Answers to the above questions for large breaks in BWRs are provided below: (A.V11.33) The RtLAPSYA Code allows a node to be quenched only after the rewet and quench model has been activated and the calculated
/"
local fluid and surface conditions indicate quench has occurred ( (see Section 4.3 of Volume 1). The rewet and quench nodel will be activated only after ECC injection has been calculated to L' I. 4 u b \ The locks will be removed at all elevations for a given w' occur. L heat structure when either the top or bottom node has been ; p j' N l calculated to quench. This is consistent with the assumption- '! l 1') that the reflood phase has been started. l (A.VII.34) Appendix K requires the use of a degraded heat transfer coefficient during the lockout conditions. YAEC's logic provides this degradation by modifying the. correlation with a degradation multiplier. It is to be noted that with the use of a wultiplier, the nucleate boiling correlation does not still remain a nucleate _ boiling correlation. .'Therefore..this use does not conflict with the Appendix K Requirement. (A.VII.35) Appendix K requires the use of a degraded heat transfer coefficient during return to nucleate boiling lockout conditions. A 50% degradation to the calcula^e3 heat transfer value implies a conservatism of 50%. (A.VII.37) The rewet and quench model is not activated until the ECC has
-been calculated to' inject into the reactor vessel. Therefore, return to nucleate and transition boiling modes are locked out prior to this period. The initiation of safety injection in the reactor vessel and quenching of the top or bottom node indicates the onset of reflood phase in BWRs. Therefore, the deactivation of the lockouts during this period is not in violation of Appendix K, Criteria I.C.4.c and I.C.S.b.3.
~
\m /
r 0
- =
i .. - l' ', ' s i 1
; )
7 Pages 270-271 of Reference 10 give.the. Appendix K Requirement I.D.5 for refill
'%. and.reflood heat transfer as a function of reflood rates. The submittal'then i states: _"The.above requirements is applicable to.large break LOCAs in PWRs.
~
'L 'As such, itz is not needed for our current RELAPSYA licensing applications." {
. (Page 271 of Reference 10). ,
i g. J(Q.VII.36) Clarify why the requirement.is not applicable to SB LOCA calculations. (A.VII.36) See answers A.VII.42 and A.VII.43 below. l
.( Q.VII.39) . Clarify if the trailing 0.5. factor in Equation 5.0-34 on Page 227 of Reference 10'is a multiplier'or meant to be an exponent on the previous term.
(A.VII.39) The 0.5 factor is an exponent on the previous term.
' f}
V
-l c
'Page 271,of I.eference 10 states that the Appendix'K Requirement I.D.5 on "is applicable to large break LOCAs and PWRs. As such, f]
f(;,/ refill and reflood: it is 'not needed for our current'RELAPSYA licensing applications. However,'a reflood heat' transfer model has been incorporated ~into RELAPSYA as a user selected option.**
- (Q.VII.42)' Clarify why the.reflood model is not needed for SB LOCAs.
(Q.VII 43) Clarify why the reflood rate of less than 1 inch per second cannot occur for SB LOCAs. (A.VII.42) AND (A.VII.43)- A sewet and quench'model has been incorporated in RELAP5YA to reasonably sinulate the thermal-hydraulic behavior expected during refill /reflood situations. This model will be used during small break LOCA calculations. This model has not been assessed against FLECHT data as required by the Appendix K Requirement I.D.S. .It is our opinion that this requirement is applicable to large breaks in.PWRs. The FLECHT test conditions are representative of the conditions expected during large breaks. FLECHT tests typical of small break conditions are not'available for the benchmarking of our rewet and quench model. However, the rewet and quench model in RELAPSYA was successfully tested against results from TNTF and TLTA runs. These are described in YAEC-1300-P, Volume III, Section 5.0.
- In our opinion,-the RELAP5YA rewet and quench model meets.the intent of Requirement I.D.5 by providing an adequate simulation of the refill /reflood '
behavior during small break LOCAs. Additionally, YAEC'E currently licensed. O model' employs a' scheme similar to the one being proposed; it uses the standard heat transfer package to calculate heat transfer coefficients during refill /reflood phases of SBLOCAs.
4 s 4 Page 15 of Reference 8' describes- a ' boron transport. nodel used in Cycle 14.. Page 2 of. Reference 9 states: ** Reactivity feedback from hydrodynamics and
' ('~'(
' \s /-
heat' conduction is provided but the effects of boron on reactivity has not-been included."
- (Q.V11.45) Clarify how the effect of boron on reactivity is included in RELAPSYA.
(A.V11.45) For PWR LOCAs, boron injection occurs via the ECC water. For PWR small break LOCAs, we assume that control rod / blade insertion occurs;' therefore, the effect of boron injection on) negative reactivity is negligible for these cases. For PWR large breaks, control rod insertion is not credited. Therefore, the negative reactivity associated 'with boron injection can be , included as a reactivity' table that is used by the reactor kinetics model.
~ For BWR LOCAs, boron injection from the Standby Liquid Control System does not occur.
/'%
i 1
L
-m l
L
'l 1
i i [ (Q.V11.53) Clarify how the Appendix K Requirement I.A.5 on calculating the Cs . hydrogen generation is satisfied. 1
-Q/ The. Baker-Just equation has been incorporated in RELAPSYA to (A.VII.53) ca!culate the energy release and cladding oxidation caused by the Zircaloy-water reaction. The reaction is assumed not to be steam limited. The Zircaloy-water reaction at the inner surface !
is initiated at the onset of clad rupture. The reaction at the l inner surface is calcriated for the entire surface area of the { ruptured node. In our licensing calculations, the ruptured node. will not be allowed to be less tha,n 3 inches in length. The above model is in compliance with Appendix K Requirement I.A.5. ( 4 Energy generation rate and oxide layer thickness are the two parameters used directly in RELAP5YA calculations. Hydrogen < generation values are inferred from the oxide layer thickness i values. .The hydrogen generation value is used in conjunction with 10CFR50.46 to assure that 50.46 b(2) and b(3) requirements 3 are satisfied. This will be assured in the following manner: 1 1 (b) The sum of inside and outside clad thickness will be calculated for.high temperature locations in the hot rod. q It will be ascertained that at the end of the transient the , sum does not exceed 17% of the original clad thickness at any given location in this rod. (b) At the end of the transient, the hydrogen generation in the , O core obtained.from the calculated cladding oxidation thicknesses, will be ascertained to be less than 1% of l theoretical maximum. (Q.V11.54) Clarify how the Appendix K Requirement I.B to take account of clad swelling and rupture on hydrogen generation is satisfied. (A.VII.54) Clad swelling and rupture models recommended in NUREG-0630 have been incorporated in RELAPSYA. The degree of swelling and j rupture is taken into account in P"LAPSYA for the calculation of ; gap conductance, clad thinning, clad oxidation thickness, and metal-water (M-W) heat generation rates. Thus, the RELAP5YA calculation is based on applicable data and is done in a manner not to underestimate swelling, incidence of rupture, and the ; accompanied M-W reaction as per Appendix K Requirement I.B. I l The requirement on hydrogen generation is satisfied in an i indirect manner (please see (A.V11.53)). The hydrogen generation is a consequence of M-W reaction. Therefore, compliance with Requirements I.B on M-W reaction as described above simultaneously complies with Requirement I.B on hydrogen generation. 1 Clarify how'the hydrogen generated is modeled to be transported j (Q.VII.55) in the system. !
L , l T(A.V11.55) Licensing calculations submitted to the NRC will be performed in accordance with Appendix K criteria and will meet the l I.{w_)'Y requirements of 10CFR50.46. These. criteria allow for the l availability of ECCS after accounting for.the' worst single j failure. Also, a successful licensing calculation will be within 2200 0F,'S 17% maximum clad oxidation and 1 1% core-wide oxidation. It is expected that'for these calculations where the amount of hydrogen generation'is's'1%', the' impact of hydrogen on hydraulics and heat transfer characteristics of the system is I ll going to be negligible. Therefore, in our licensing calculations, the hydrogen transport is not modelled. f vn
.O
,v . .
.p .
IX. -BREAK FLOW tt (Q.IX.13) Deleted
Reference:
Memo, Ed'D. Throm, Reactor System Branch DSI to. ' Brian W. Sheron,' Chief, Reactor. System Branch, DSI, " Summary of 06/22/84 YAEC-NRC Meeting on RELAPSYA PWR SBLOCA EM Analysis". ; July 17. 1984.
)
1 l
')
1 O l RESPONSE TO 26 QUESTIONS ON RELAPSYA April 30, 1985 4 0 . 1
)
L
- 1. C' CONDENSATION HEAT TRANSFER AND NONCONDENSABLE CASES U; Clarify the procedures used to account.for the effect's of (Q.I.21) r noncondensables on the secondary side'of the steam generator.
(A.I.21) - Two potential noncondensable gas sources exist for the secondary side of steam generators when a pipe breaks.in the Primary Reactor Coolant System:
- a. Noncondensable gas in the secoridary' side coolant during .
normal operation.
- b. Noncondensable gas introduced with auxiliary feedwater.
During normal operation, the air ejector limits the amount of noncondensable gas in the secondary side to'a very small value The Auxiliary (e.g., less than 1 sefm for the Yankee plant). Feedwater System' draws suction from the demineralized water i, storage tank. Although this tank has an atmospheric gas cover, the pressure increase due to the auxiliary feedwater pumps will tend to retain the very small amount of dissolved gas in solution. Therefore, we conclude there are no significant noncondensable gas sources on the secondary side, and the small amount present will have no significant impact on the secondary side behavior.
II. ' NONEOUILIBRIUM EFFECTS 'fi 4} 4 i { '
,' Temperature gradients in the pressurizer liquid could affect the condensation rate as the pressurizer refills. As the pressurizer refills, if
.the liquid surface becomes saturated it coulo insulate the steam above from the subcooled liquid below. Combustion Engineering did a bounding calculation
~
with no condensation so the liquid acted like a piston. "The results showed that for an equilibrium model, the pressurizer refilled to 34 percent of full. whereas for the' piston model, the pressurizer refilled to only seven percent of full" (Page VIII-32 of Reference 2). Clarify how the pressurizer will be modeled in SB LOCA analysts including: (Q.II.12) What noding will be recommended? (A.II.12) Two nodes will generally be used in the pressurizer for LOCA licensing analyses. One node will represent the normal liquid portion, the other will represent the steam space. Any deviations from this noding will be documented and reported. b . Pressurizer plays an important role for non-LOCA transients, and its model is' crucial to accurately predict the outcome of such events. The pressurizer does not play an important role during l blowdown and core refill phases of LOCAs which are important for licensing analyses. The noding may be important during pressurizer refill phase. However, the core is covered with several feet of water during this phase and has no impact on l- licensing results. (Q.II.13) How will the wall energy storage be modeled? (A.II.13) The pressurizer wall will be represented by two heat structures; one in each of the two pressurizer volumes. To maximize the stored energy in the wall, the outside surface will be modeled as insulated and the inside surface will be initialized to the
, pressurizer saturation temperature.
t i . - . _ _ . . _ _ _ _ _ _ _ _ _ - _-__ - __ _ _ __ __________ _
(Q.II.14) What wall heat transfer will be modeled? ( } (A.II.14) RELAp5YA standard heat transfer package will be used-to calculate the appropriate heat transfer rate. The heat transfer package is described in Section 4.0 and Section 2.1.3.4.2 of Apprndix A in Reference 11.1. (Q.II.15) How will the liquid temperature distribution be modeled? (A.II.15) Liquid temperature distribution in the pressurizer will not be modeled. The liquid will be initialized at saturation conditions. It is true that a temperature gradient in the pressurizer could affect the condensation and the pressurizer refill rate. However, the core is already covered with several feet of water at the initiation of the pressurizer refill. The pressurizer refill will not influence the results of SB LOCA calculations, and the liquid temperature distribution is not needed. p s/ (Q.II.16) Clarify how accurate the condensation model discussed in Section I is for a horizontal interface between the liquid and steam. (A.II.16) As discussed in A.II.15, the condensation rate during the pressurizer refill does not impact SB LOCA licensing results. Therefore, the accuracy of the condensation model used for the pressurizer refill process is not important for LOCA licensing calculations. (Q.II.17) Clarify how the pressurizer refill process has been verified against experimental data and what range of refill conditions has been considered. (A.II.17) As discussed in A.II.15, the pressurizer refill process does not impact LOCA licensing results. Benchmarking for pressurizer refill is considered unwarranted. L l
)
<g
'L
,j 11 References JW.
(II.1) Fernandez,'R. T., R. K. Sundaram, J..Ghaus, A. Husain, J.'N. l Loomis. L. Schor, R. C. Harvey, and R. Habert, "RELAP5YA - A Computer Pro 5 ram for Li5ht-Water Reactor System Thermal-Hydraulic
> Analysis, Volume I: Code Description," Yankee Atomic Electric Corepany Report YAEC-1300P, Volume I (Octhber 1982)(proprietary).
i . e j z- _
VI. FLOW REGIMES
# In some small-break LOCAs calculated with TRAC for B&W plants, oscillations have occurred between the two cold legs of the same loop.
This phenomenon has been calculated in both the loop with a small cold leg break and the other loop. These oscillations occurred after the candy-cane voiding caused the hot leg flow to stop. First, the flow moves from the vessel into one cold leg, through the pump, loop seal, steam generator lower plenum, other loop seal, other pung, and other cold leg back to the vessel. ! Next, the flow reverses and goes the other way through this circuit. The i oscillations are driven by the loop-seal height and the density changes caused by the HPI water injected into the cold leg as the cold les flow moves toward the loop seal. Both B&W and CE plants have the potential for these oscillations because of the two cold legs per loop. If these oscillations { occur, they can cause cold leg temperature oscillations with resulting break l flow variations. Unless the models include the two cold legs per loop plus the actual HPI injection locations, the codes cannot model this phenomenon. (Q.VI.14) Clarify how this potential SB LOCA phenomenon will be modeled. 7s i \ wJ (A.VI.14) Yankee will model the actual configuration of the plant. For those plants with two cold legs per loop, the two loops will be modeled separately and the HPI will be injected at its correct location. Any deviations from this approach will be justified. (Q.VI.15) Clarify how bypass is modeled or justify why it is not necessary between the downcomer and the upper head and between the core and the upper head. Bypass could change the temperature of the upper head and thus change the time when flashing begins. Bypass could also possibly add water from the core to the upper head during a period when the upper head was partially voided and, therefore, condense steam and drop the pressure. (A.VI.15) Figures VI.15.1 and VI.15.2 represent the nodalization used in the pump trip studies performed for the Maine Yankee plant and Yankee plant at Rowe, respect.vely. Similar noding schemes will be 8 () employed for the small break calculations. Yankee uses a detailed m 1 1 nodalization of the, plant. The nodalization used for the Maine
'~ ' Yankee plant and Yankee plant are not identical. These
) differences were necessitated to model the actual configuration of the plants. The b'ypass model between the downcomer and upper head and between the core and upper head, intended for MY and Y, are _..
discussed below: Maine Yankee (NY) i Bypass flow between the downcomer.and the' upper head is modeled. It is represented by the junction connecting Nodes 3 and 50. Bypass between the core and the upper head is also modeled. The bypass is represented by Nodes 61-1 through 61-4. Yankee (Y) Bypass between the downcomer and the upper head is modeled. It is (vh represented by the junction connecting Nodes 5 and 85. t v Bypass between the core and the upper head is indirectly modeled. This is due to the Y upper head region geometry. Y has control blade shrouds which are open at'the higher elevations of the upper plenum. Bypass flow from the upper head to upper plenum is modeled by Node 75 and its associated junctions. The bypass from the core to upper portions of the upper plenum is lumped with other upper plenum flow since only about 5% of the total core flow is associated with the control blade shrouds. (Q.VI.16) Clarify how the recirculation flow from the upper plenum up into the upper head and back down into the upper plenum will be > modeled, or justify why it is not necessary to model this flow. This recirculation flow could change the upper head temperature and with it the time when the upper head begins to flash. It might also cause some of the bubbles formed in the upper head during depressurization to be swept out and accumulate in the high l , points of the loops. l
,es V
l
- n. .
J (A.VI.16)- Figures VI.15.1 and VI.15.2 provide representative noding schemes l 1
')
.which will be used for SB LOCA calculations. The upper plenum and
.the upper head noding intended for NY and Y analyses are discussed below:
Maine Yankee The upper head.and upper plenum regions are modeled by nine nodes. These are Nodes 46, 45, 50, 85, 95 and 61-1 through 61-4 in Figure VI.15.1. The recirculation path is established via the control rod guide tubes situated - in the upper portion of the core, just below the upper plenum region.
- Yankee The recirculation flow from the upper plenum to the upper head and back is'modeled for Yankee. Nodes 62, 85, 75, 90 and the associated junctions represent this flow.
Clarify if the left side of Equation 2.1-29 on page 38 of (
\s '
(Q.VI.22) Reference 10 should be the reciprocal of the square root of h T rather than the h Tshown. (A.VI.22) There is an error in the documentation. The left side of Jguation 2.1-29 is the reciprocal of the square root of >T and not ATas shown. 9 3
\m
-7 .
_m___________.______________...-._____m.__
f
?> <
a b' . j.- , f'
~
if. E y ., Ta :
~
4 3: S i a e., t , . v T w m. :
'C ' * 't~5 ja 1.&l3-I . _ .
6" . ', ' "4 IN ii"~ [Dil : -: ::-F*-0[} 8--i
-(
; i,
*a ii i :, I 3 2 .g 3
M l +
> C 5 o .E.
j 'l.L. ? - 0 g - e 7 e
- t t
= 'I l * -o Ia g I
~[ .~
., c.
1 .y co' c _- cs e gm. CU g, C C
- s -
g g[, ( ~d*ii'"*-) oo Yl ? ? ?
# D C r ! l.fl Il t!,;!
- :. EU
_E:: _ cn Ka- . ee
. *g
.t: W ii !: IF m _
OO C. O . u e -r OA
- WA n
'I
~k
- lfI
.l:- _
b, :1 1 S $ [ M. m Wa #N
,e , ; , ,, g 0
k g. bi - 9 0 . Q (? "'; $ e L I'y t w 1*? Wr: ? ;N , l 7 5 :) i( . . g.tD(tp 7vr
- f. J n
~
e.
, . i g
- y E
P 8 n=u.n 4 5 d I
\
l 1 TI 1 e I
"L
/
r r e ) o t t a0 a w2 r y d9 er P w( na ed C n o mc a e m eS a t e~ S t I S7I s 0 9 M6ni3Y9 g 4
= l 3 2 t 0i 0
- I I I (
9 9 e p
~ _, _ l o
o r eo _ s w
~ t t r
0 os r o
\
e )U bI 7
*p 1
4 t n sasn kst P eee 0 r r s 2b pe 1 ssp r 0 e u3 e 0 2 e E aae TI n 2 , 4 uv Ehh r
, i R fQ 2
i 3 e H ppp L ) t T ooo 0 ooo e n Rt t ( 7 C o Sll l A u 'M eccntt S N 0 t
- r. kot aa
- 0 6 - .
P_ O or nn obii t 0 0 . O L M eee 3 7 2 7 E L A Al l l vggb 5 nnu o i dr C Pii o s I A ssd 0 3 P L 0 Y E - - a r e 2 H T D t 0 E i 2 T r 0e u o s e P V3J l r P e h j p l l l I , 5 74 - .. g 2 ( 2 6 . 5 s-2
+D*I#+5 3 1 2 3
95 2 g
+ I-- !. I 5 l I
. l g
4 R . g l 0 1 C 3
. +[ k, r
i VII. CORE STEAM COOLING
'd; (Q.VII.5)~ The FRIGG test restilts shown on pages 14-15 of Reference 12 are for mass fluxes cf 729-1464'kg/s-m . Clarify what range of mass flux occurred in the GE level swell test and how the new vertical interphase drag models were assessed for the flow range of less than 729 kg/s-m .
(A.VII.5) The mass flux in the GE vessel ranged from about 0 to 100 kg/s-m , with the higher fluxes occurring immediately after blowdown initiation. Steady-state void profile data at these low mass fluxes are generally not available. Hence, the only assessment of the drag models against steady-state data were in the FRIGG tests, with the lowest mass flux at 729 kg/s-m . Other assessments of the drag models at low mass fluxes were in the TLTA boiloff test (Section 5.2.6, Reference VII.1). In this test, the mass fluxes were generally in the range of 0 to 20 kg/s-m . (Q.VII.11) Page 100 of Reference 12 states: "The original Biasi and Griffith-Zuber CHF correlations were developed from data taken in simple test geometries. These included single heated tubes and, for the latter correlations, heated annuli. However, most applications of the new CHF option in RELAPSYA will primarily_ involve bundle type geometries. To extend the use of the correlations to bundle geometries, the Biasi correlation has been slightly modified. The modification involves the use of the heated to wetted perimeter ratio. The new CHF option has been assessed against more prototypical conditions." Three facilities were used to verify the model. "The five Columbia tests were selected primarily because their test sections had representative, nonuniform power distribution and full-length, prototypical fuel rod simulators. The four GE nine rod tests were selected primarily because they had larger rod diameters and a fairly broad mass flux range. The three ORNL THTF tests were selected primarily because they involve a fairly large rod cluster, a fairly
, broad pressure range, and intermediate to high
l qualities at CHF" (page 101 of Reference 12). The modified Biasi
/"'*
CHF correlation was assessed by these tests, and the systems tests s_- were used for other regions. _For all the ORNL tests,. the calculated CHF location was below the data. For the three tests in Columbia test Section 68 (page 108 of Reference 12), the-RELAPSYA calculated CHF location is as high or higher than the data. Clarify why this is conservative when the calculation used the average power whereas "of ten CHF was experimentally detected on rods with the highest normalized radial power factor" (page 103 of Reference 12). Also, for three of the four GE CHF tests (page 109 of Reference 12), the calculated CHF location was above the data. (A.VII.11) The RELAP5YA CHF' calculation was for an avetsge rod, while the Columbia test bundle included higher than average power rods. CHF was generally detected on the higher. power rods in the tests. RELAPSYA CHF calculations for a higher power rod would predict a lower CHF location. A s_, k The RELAPSYA CHF correlation is a best-estimate correlation.
~Therefore, it was. assessed for accuracy rather than conservatism. ,
i This correlation tends to be conservative, particularly at medium I to high qualities, Which are of greatest interest for LOCA analysis. (Q.VII.12) Clarify why this is conservative. (A.VII.12) CHF predictions for the GE CHF tests are shown in Figure 3.2-9 (Reference VII.1), attached herewith for convenience,~which also includes the results of other CHF test cases. The CHF quality is accurately predicted for two of the four tests. Of the remaining two predictions, one is high by 16% and one is low by 5%. It should be noted that the RELAP5YA CHF correlation tends to be conservative at medium to high qualities, which are of greatest interest to LOCA analyses. s
i Additional test cases performed in the RELAPSYA assessment consistently confirmed the adequacy of the CHF calculation for L - LOCA conditions. These test cases and results_are given in the follow 2ng sections of Reference VII.1: THTF - Section 5.1 TLTA - Section 5.2 LOFT - Section 5.3 It should also be noted that in RELAPSYA licensing analyses, uncertainties in power level are added to insure conservatism. (Q.VII.24) Page 180 of Reference 10 states: **This update also contains a modification to the heat transfer logic in RELAPS MOD-1 such that the heat transfer coef ficient used in estimating wall temperature at CHF is a pre-CHF heat transfer coefficient. The original logic was not able to predict this heat transfer coefficient [ consistently. Clarify how this change has been assessed. k. (A.VII.24) The change referred to as a modification on page 180 is actually part of the new CHF calculation contained in subroutine CHFCAL. The old CHF calculation not only calculated a critical heat flux value, but would also sometimes set the variable HTCOEF to a non-zero value to allow for variance of critical heat flux with respect to wall temperature. A non-zero HTCOEF affects the calculation of wall temperature at CHF and causes CHF to be modified in subroutine PREDNB when CHF is exceeded. Wall temperature at CHF affects the transition boiling heat transfer and the modified value of CHF is printed in RELAPS MOD-1 major edits. Therefore, the effects of HTCOEF are apparent in RELAP5 { j MOD-1 results. The CHp correlation implemented in RELAPSYA is not a function of wall temperature, so the variable HTCOEF is simply set to zero. } j O l l l
The CHF calculation assessment included this part of the f~y calculation.
,\,,)
(Q.VII.44) Figure 3.2.10 on page 34 of Reference 16 shows the decay. power for LOFT Test L3-6 was significantly underpredicted by Cycle 14. Clarify if there is a problem with the Cycle 14 and also RELAPSYA decay heat models that causes the power to be underpredicted. I (A.VII.44) RELAP5YA has been developed from the-RELAP5 MOD-1 Cycle 18 code. Therefore, the decay heat models contained in RELAPSYA are consistent with Cycle 18 and not Cycle 14. The RELAP5YA calculated decay heat has been compared to LOFT Test L3-6 in Figure A.V.6-3 of Reference VII.2. This comparison shows that the RELAP5YA calculated decay heat does not underpredict the LOFT results as did Cycle 14. (Q.V11.46) Page 53 of Reference 10 states:
"In addition, a small fraction of the core power is directly deposited in the moderator due to gamma f- s and neutron interactions with the fluid. The direct moderator
,_) heating effect Q ,d is n t stated explicitly in the RELAPS MOD-1 documents, but the computation is contained in the point kinetics subroutine, RKIN." Clarify What fraction of the core power is modeled as going directly to the moderator.
(A.VII.46) Applications of RELAPSYA to date have not assumed any core power going directly to the moderator. However, future applications of the code may assume a non-zero value for direct moderator heating. Justification of any non-zero value for the direct moderator heating fraction will be provided upon application. (Q.VII.47) Page 13 of Reference 10 states that algorithns for heat transfer processes account for: " Radiative heat transfer with a participating fluid medium." This does not seem to agree with
- Clarify statements made elseWhere in Section 4.4 of Reference 10.
how/if the participating fluid medium is modeled. O
. ;y a e ,
w r
~
(A.VII.47)- The' cited statement has been misquoted and taken out:of context. The original statement occurs-in Table 1.1-2: " selection Criteria 9-s 4 Y for LOCA Analysis Methods". This table lists criteria to evaluate' Q.- thermal-hydraulic codes and to select one as s suitable starting-
~
point for YAEC's LOCA Methods Development Program. These criteria i n also reflect-features YAEC seeks in order to meet both the near-- and_long-term objectives-of this. development program. These objectives include both best-estimate-and improved evaluation.
-model analysis capabilities. This particular feature (Item 2.2.6.e) is not' required nor considered crucial for ECCS evaluation model capabilities defined in 10CFR10.46, Appendix K.
p Section 4.4. of Reference VII.3 same as-Reference 10 of NRC-questions, clearly states that.the multiple surface radiation model is for. solid-solid radiation heat transfer only, and the surrounding fluid is assumed to be transparent (i.e., nonparticipating) to radiation heat transfer. Table 1 in Appendix A of VII.3 points out ttnt RELAPSYA (and i : RELAPS MOD-1) does contain a radiation heat transfer term, hp , within the vapor correlation used for low flow post-CHF transition-and film boiling resimes. g
-(Q.VII.48) Page 203 of Reference 10 lists the variables in the RELAP5YA thermal radiation model including the steam absorbency and emissivity. ' Clarify how these two properties are used.
(A.VII.48) These two variables are not used in the multiple surface radiation U model at the current time. They have been defined in the RADHTC block to allow for incorporation of the fluid as a participating medium in the radiation heat transfer process at some future time. (Q.VII.49) Page 198 of Reference 10 states: "For radiation heat transfer to be calculated, the temperature of one of the' surfaces must exceed a cert'ain specified temperature (default is 900 degrees K) and the I void fraction of the volume adjacent to the heat structure must exceed a specified critical value (default is 0.75)." Clarify [, _ - - - _ _ --__2____- _ _ -
t l what is done if some surf aces in the network meet the requirement f and others do not. q) } i (A.VII.49) The radiation calculation will proceed for all solid surfaces in f
.i the radiation network when user-specified temperature and void ]
fraction criteria are satisfied. This feature allows the user to avoid radiation calculations when conditions are such that the radiant heat fluxes will be relatively small. This, in turn, will avoid unnecessary computational costs. ] l i l (Q.VII.50) Clarify how accurate the assumption of neglecting the water is for ; void fractions of 0.75. l l (A.VII.50) For void fractions of 0.75 and below, the liquid mass fraction is j 1 relatively large, ranging from 0.999 at 15 psia to 0.710 at 2000 psia. Based upon our experience, heat structures will generally be well cooled by convective heat transfer to the two-phase fluid for this void fraction range. This includes rewet and quench phenomena encountered during reflooding conditions in hot f]t t
'w/ bundles. For example, this is evident by examining the void fraction and heater rod temperature histories presented in Reference VII.1 for THTF Reflood Tests 3.09.100 and 3.09.10Q (Figures 5.1-17 to 5.1-28), and in the TLTA Boiloff Test 6441/6 (Figures 5.2-49 to 5.2-62). Thus, the radiant heat fluxes will be relatively small compared to the convective heat flux to the fluid for void fractions of 0.75 and below. Therefore, we believe this assumption tends to be slightly conservative, yet close to reality.
(Q.VII.51) Justify that the clad geometry changes in SB LOCA calculations do not change the view factors used. (A.VII.51) This question will be addressed in the documentation for j plant-specific LOCA-EM sample problems. ] f 1 (Q.VII.52) Clarify how the thennal radiation model will be used in EM analyses, ig) LJ l I I
l ?. .) (A.VII.52) The Multiple Surface Radiation model has not been used in plant j f's applications of RELAPSYA to date. Clarification of how this model will be used in EH analyses will be provided in the documentation for~ plant-specific sample problems when this model is u'ad. , I References !
-l
.(VII.1) Fernandez, R. T., R. K. Sundaram, J. Chaus, A. Husain, J. N.
.Loomis, L. Schor, R. C. Harvey, and R. Habert, "RELAP5YA - A q Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume III: Code Assessment," Yankee Atomic Electric Company Report YAEC-1300P, Volume III (October 1982)(proprietary) .
(VII.2)- YAEC Letter to USNRC, dated March 1, 1985, " Response to the NRC Questions on RELAPSYA". (VII.3) Fernandez, R. T., R. K.-Sundaram,.J. Chaus, A. Husain, J. N. Loomis. L. Schor, R. C. Harvey, and R. Habert, "PELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic d Antlysis, Volume I: Code Description," Yankee Atomic Electric Company Report YAEC-1300F, Volume I (October 1982)(proprietary).
/^%
U
-- -_-_-_.-__--_-__J
t
;+: ,
. uO ..
1.0- + i i i i l l l 0.8 - Q . .
',U O;6 ,
o _,, e 3 . . 3 w v yO O'. 4 - 3 . i a u . O l
. K,e(
0.2 - A Columbia - o 3 O GE O ORNL - l l: L 1r 0.0 L_g ,
; , ; , l ,
0.2 0.4 0.6 0.8 1.0 Experimental Critical' Quality Figure 3.2-9: Comparison of Calculated and Measured Critical Qualities . G 4
-O l
j l e sj s IK. BPEAK FLOW
- ,h .
-( . .
.Page'21 of Reference 12 states: " Based on these calculations,.it.
(Q.11.15) 1 is recommended that the' option using static pressure.and enthalpy from the donor volume be selected." Clarify if there are any.SB LOCA applications Where the area of the break would be large enough to cause the donor volume static and stagnation pressure and enthalpy to differ by more than a few percent. . (A.II.15)2 0.5 ft ' is the largest break size customarily analyzed _for SB LOCA' licensing calculations. The primary loop flow area is 6.12 ft 'for.the Maine Yankee plant and 1.42 ft for the Yan'kee plant at Rowe. Therefore, the largest area' ratios for these plants are about 0.08 for NY ' and about 0.35. for the Yankee plant. :For these area' ratios, the critical flows predicted by the static.or stagnation pressure and enthalpy in the donor volume are not expected-to differ more than a few percent. (Q.II.16) Pages 27-28 of Reference.12 states: "The various checkout j
\/ problems. described in Section 2.2.1 indicate that a stable solution'is obtainad when the Moody critical flow model is used d
with the static pressure and enthalpy taken from the donor volume. It also indicates that the stagnation pressure and enthalpy option may be used when the break-to-volume area ratio is less than about 0.5." Clarify whether the static or stagnation option is recommended then for SB LOCA calculations. (A.II.16) The static option is recommended for SB LOCA calculations. This recommendation is based on our experience with the use of the stagnation option. Sometimes the stagnation option resulted in computational problems for open conduit flows. The problem was caused by feedback between the mass flux and stagnation condition calculations. No such calculational problems arose when the static option was utilized. (Q.II.17)
- Page 122 of Reference 10 states: " Application of the Moody model involves no explicit assumption about the slip ratio (v /Vg ).
n- - _ _ - - _ - . 4 Experience has shown that with~the Moody model, RELAPSYA yields fN slip ratio values in the range of 1.0 to 1.3 for a typical
\~- blowdown calculation. These values are considerably lower than the theoretical values derived by Moody, but unlike Moody's values, they are compatible with the RELAPSYA equation set.**
Clarify what these slip changes do to relative amounts of liquid and vapor that to out the break and how much they change the depressurization characteristics of SB LOCA calculations. (A.IK.17) The RELAP5YA calculated slip ratios'were selected to provide consistency with the code. The selection resulted in the reduction of mass-error and a substantial decrease in computational time. Additionally, the use of the RELAPSYA slip ratio'is consistent with the implementation of the Moody model in current homogeneous codes used for licensing analyses. RELAPSYA calculates lower slip ratios than Moody's theoretical model. -RELAPSYA will calculate relatively larger amounts of e- liquid and lower amounts of vapor leaving the break. This will
-f A. lead to a slower depressurization of the vessel than would have been calculated had Moody's theoretical slip ratio been utilized.
Therefore, the RELAPSYA slip ratio depletes the RV inventory
' faster and refills it slower due to decreased ECC flow rates caused by higher pressures in the reactor vessel. Thus, the Moody model employed in RELAP5YA is expected to yield conservative SB LOCA results.
4
l. k i, j . O . A*. . RESPONSE TO B0 QUESTIONS ON RELAP5YA July 1, 1985
~
O - O s
l L I'
/ I. CONDENSATION HEAT TRANSFER AND WONCONDENSIBLE CASES
[& Nine potential sources of noncondensible gases are listed on Page
(1) dissolved hydrogen in the pf ma,ry i
VIII-49 of Reference 4 for a PWR: coolant, (2) dissolved nitrogen in the accumulator water,4(3) dissolved air in ' the refueling water storage tank, (4) hydrogen released from the " sirconium-water reaction, (5) f ree nitrogen used to pressurise accumulators. (6) hydrogen released from radiolytic decomposition of injected water, (7) fission and fill gas in reactor fuel, (8) hydrogen gas (free and dissolved) in makeup tank, and (9) pressurizer steam space gas." O.7.3. Clarify which of these are treated by RELAPSYA and what rates are used to predict the amounts entering the system from each source. A.I.3 Non!.nensible gases are expected to have a negligible effect on the ( outcome of small break LOCAs. Hence, the treatment of noncondensigles in our calculations is not warranted. Appendix 1.3-1 provides the justification for excluding the treatment of noncondensibles in our-analyses. J I k
6. APPENDIX A.I-3 (" BOUNDING ESTIMATES OF NONCONDENSIBLE CAS SCURCES 7 AFD THEIR INPACT ON SBLOCAS, ,, . , ,
~
1.0 .INTRODUCTIDW During normal operation, a concentration between 25-50 cc of hydrogen /kg of water is dissolved in the Reactor Coolant System with an associated buildup in the pressurizer vapor space. Some concerns have been raised that this noncondensible gas can be released during a SBLOCA and inhibit adequate core cooling. The major concern is centered around the belief that the noncondensibles generated may mig 7 ate to the steam generator U-tubes and inhibit flow or sufficiently degrade heat transfer to inhibit' adequate decay heat removal. The steam generators are the preferred mode of decay heat removal for a significant pericd of time in only very small breaks. For example, licensing analysez for breaks larger than 0.02 f t in the Maine Yankee plant show that () most af the decay heat is removed through the break (Reference I.,3-1). For breaks rmaller than 0.02 ft , the steam generators are relied upon to remove some decay heat for a significant period of time. ! section 2 describes the method used to estimate the nine potential noncondensible gas sources. 11 ant-specific calculations for the Maine Yankee plant and Yankee plant at Rowe are provided. Section 3 provides an estimation of the noncondensibles' impact on the heat transfer capability of the two plants. 2.0 ESTIMATION OF NONCONDENSIBLE CAS SOURCES The release of noncondensibles from the nine potential sources are calculated below. These calculations are performed for both the Maine Yankee plant and Yankee plart at Rowe. O \
L MAINE YANKEE PLANT _
\
- 1. Dissolved hydrogen in the primary coolant: The calculation asse.aed the maximum RCS concentration of 50 cc of 2H /kg of water ,
It is conservatively assumef that all
'(Reference I.3-2) at STP.
the hydrogen is released and remains in the primary system.
- 2. Dissolved nitrogen in the accumulator water: The concentration of dissolved nitrogen is determined from Henry's Law (Reference I.3-3)
~
at 104 F and 230 psig to be 2.04 x 10 lbm N2/1bm H20. This concentration is used to calculate the nitrogen introduced during accumulator water injection. The pumped safety
- 3. Dissolved air in refueling water storage tank:
injection water from the RWST was calculated to contain 19.9 cc of air /kg of water at STP based upon Henry's Law (Reference I.3-3). The entire RWST inventory is' assumed to be injected. The calculation
- 4. Hydrogen released from zirconium-water reaction:
is based on 1% zirconium reacted and 37840 fuel rods. V The nitrogen
- 5. Free nitrogen used to pressurire accumuistors:
contained in the accumulator gas space is assumed to be released when the accumulator liquid empties.
- 6. Hydrogen released from radiolytic decomposition of injected water:
concentration in the primary system falls The time when the H y below 5 cc of Hy /kg of water is determined to be the time to For start the calculation of radiolytic decomposition of water. the 0.02 ft Maine Yankee SBLOCA, a concentration of 5 cc H2 /kg At this time, the H O is reached at approximate?y 4700 seconds. 2 steam generators are not relied on to remove decay heat. The calculation is performed
- 7. Fission and fill gas in reactor fuel:
assuming all fuel rods are at end-of-life conditions.
8.-Hydrogen in the make-up tank: For the Maine Yankee plant, the 1 make-up tank does not release water into the primary system during (j a SBI,0CA.
- u. -,
pressurizer. vapor space: 630 cubic feet of hydrogen at STP is ' 9. assumed to be contained in the pressuritor vapor space! This represents 90% of.the available gas volume. For Maine Yankee, Table I.3-1 provides the volume at (STP) and mass of the noncondensibles from all the nine sources identified above. It is to be recognized that the core would have to undergo transients that are not predicted to occur for applicable small breaks, in order for many Present Appendix K small of the sources identified above to be important. break plant licensing analyses show that clad temperatures remain low and no cladding rupture is predicted. In addition for breaks of 0.02 ft (Reference I.3-1) in area and lower, the system pressure does not reduce to the safety injection tank pressure. The only sources that may be introduced are: (a) hydrogen dissolved in the primary coolant, (b) air dissolved in the (/ refueling water and (c) hydrogen contained in the pressurizer vapor space. at STP. The noncondensible gas volume from these sources amounts to 2011 ft 688 ft For the Maine Yankee plant, the total primary tube volume it per steam generator. The noncondensible gas volume at the steam generator The reactor safety relief valve setpoint (1000 psia) is about 60 ft . vessel upper head volume is 1476 ft . Most of the noncondensibles are impact on decay expected to migrate to the upper head without causing any However, if all the gas heat removal capability of the. steam generators. generated in this conservative calculation were to migrate to the steam generator U-tubes and reside there, the gas would occupy about 3% of the U-tube volume at stable conditions. YANKEE Pt. ANT AT ROWE
- 1. Dissolved hydrogen in the primary coolant: The calculation assumed a maximum concentration of 40 cc of H 2 /kg of water t
SL , y j (f < is l i l.: (Reference I.3-4) at STP. !!t is conservatively. assumed that"all.
/ the hydro' gen is released and remains in the primary system.-
v < I
- 2. Dissolved nitrogen in the accumulator water: The dissolved- ..w. ..
concentration of, nitrogen is determined from Henry's Law (Reference
~
.I.3 3) st'14.7 and 70 F;to be 1.9027 x 10 lba N2 /lba
- I
'M yo. This concentration is used to' calculate the introduction of j nitrogen during accumulator water injection.
l 3. Dissolvod air in refueling water storage tank: The concentration of dissolved air is determined from Henry's Law (Reference I.3-3)
~
to be 1.6364 x 10 lba air /lbe H2 O at 14.7 psia and 130 F.
.The entire RWST inventory is assumed to be injected.
- 4. Hydrogen released from zirconium water-reaction: The calculation is based on 1% zirconium reacted and 17336 fuel rods.
- 5. Free nitrogen used to pressurize accumulators: The accumulator at the Yankee plant is isolated just before the liquid is completely depleted. This is done to avoid accumulator nitrogen injection into the reactor vessel.
- 6. Hydrogen released from ra.a8,olytic decomposition of injected water:
The radiolytic decomposition of water is initiated when the H2 concentration in the primary system falls below 5 cc of H2 /kg'of water. For a 1" break in Yankee Rowe, this concentration is reached at about 3000 seconds. At this time, the steam generators are not needed to remove decay heat.
- 7. Fission and fill gas in reactor fuel: The calculation is performed assuming all fuel rods are at end-of-life conditions.
- 8. Hydrogen in the make-up tank: For the Yankee plant, the make-up tank does not release water into the primary system during a SBLOCA.
- 9. Pressurizer vapor space: 185 cubic feet of hydrogen at STP are assumed to be contained in the pressurizer vapor space. This i O represents 90% of the available gas space.
i q
-s- 1 l
t !
- _ _ ___ ___-_- _ h
i 1 l - For the Yankee plant at Rowe, Table I.3-2 provides the volume (at STP)
,C and mass of the noncondensibles from all the nine sources identified above.
V . For breaks of 1" diameter and lower (for which the steam generator is desired to remove decay,, heat), the system pressure does not reduce tY the accumulator actuation pressure (Reference I.3-5). Also, the clad temperatures remain low and no cladding rupture is predicted. The only sources that may be introduced are: (a) hydrogen dissolved in the primary coolant, (b) air dissolved in the refueling water and (c) hydrogen contained in the pressurizer vapor space. The conservative calculations performed indicate that for the Yankee plant 440 ft at STP of noncondensible are available during the period the steam generators are desired for some decay heat removal. For the Yankee plant, the total primary tube volume (4 steam generators) is 538 ft . The noncondensible gas volume at the steam The generator safety relief valve setpoint (900 psia) is about 13 ft . reactor vessel upper head volume is about 378 ft . Most of the noncondensibles are expected to pigrate to the upper head region without Q causing any impact on the decay heat removal capability of the steam generators. However, if all the gas generated in this conservative calculation were to migrate to the steam generator U-tubes and stay there, the gas would occupy less than 3% of the U-tube volume at stable conditions. 4.0 EFFECT OF NONCONDENSIBLES ON CONDENSATION HEAT TRANSFER The impact of the noncondensibles on heat transfer and thermal hydraulics of the two plants is assessed in.this section. It is concluded that the noncondensible gases released at the Maine Yankee and the Yankee plants during small break LOCAs will have a negligible effect on the hydraulic and therinal aspects of the two plants. The two effects of noncondensible gases are:
- 1. Reduction in the available temperature difference for given primary i
and secondary system pressures. rm i j V) { l I s i
- 2. Reduction in condensation heat transfer due to the resistance of the noncondensible gas boundary layer.
'i n V The reduction in the available temperature difference is negligible as illustrated in Tables I.3-3 and I.3-4 for the. Maine Yankee plant, and Tables I.3-5 and I.3-6 for the Tankee plant at Rowe. These tables assume th't 20N ft (STP) at Maine Yankee and 440 ft (STP)atYankeePhantof' noncondensibles are evenly distributed in either the primary system or the steam generators. The presence of noncondensibles affects the saturation temperature due to the partial pressure of the noncondensible gas. These temperature differences of less than 3 F will not significantly affect the small break calculations.
The presence of noncondensible gas in a condensing vapor influences the resistance to heat transfer in the region of the vapor liquid interface. This increased resistance results in reduced heat transfer. The increased resistance is due to the formation of a noncondensible boundary layer. The noncondensible gas is carried with the vapor towards the liquid vapor interface where it accumulates. Therefore, the partial pressure of gas at the fQ liquid vapor interface increases above that in the bulk of the mixture. This situation is shown in Figure I.3-I. The effect of noncondensibles on the condensation heat transfer is calculated using the method out.ined in Reference I.3-6. The overall heat conduction through the liquid film when equated to the sensible heat transfer in the diffusion layer plus the latent heat liberated at the film interface results in the equation: h (T g - Tg) = h,(Tg -T CI
+K hg (PCO
~
GI am where: he = heat transfer coefficient in liquid film, Btu /hr ft 0F 2 hs = sensible heat transfer coefficient, Btu /hr ft 2or Kg = mass transfer coefficient, ft/hr l-I i L _ ___
m 1
'hgj= ' intent heat of vaporization. Btu /lb.=
P, = :los mean partial pressure of noncondensible.
~
AI A0 Iram ." I 1n Pgg/PA0 ' *"~ + I: , Oi.her' terms are defined in Figure.I.3-1.-
'For the case d ere the gas mixture is stagnant, the sensible heat a
transferred through the diffusion layer need not be considered. Kg,-the mass
' transfer coefficient, is defined in Reference I.3-6 as:
a0.373 1.02D gH 3f(y{,i-
- 1)
-(2) g" M MD d
where: D = volumetric diffusivity, ft2 /hr H = tube height..ft acceleration of gravity, ft/hr2
.g =
= mixture density, Ib/ft 3
$a = vapor density in bulk mixture,.ib/f't 3
$1 = vapor density et interface, Ib/ft 3 h = mixture viscosity. Ib/ft-hr The film heat transfer coefficient, h , was evaluated using the Nusselt film condensation model.
*0.25
= 0.H 3 g( 3 -f)gh,k y g e g M(Tg - T,)
O where: I 1
= liquid density, Ib/ft 3 fy = vapor (ensity, ib/ft 3
e ki = liquid thermal conductivity, Btu /hr ft 0F bl = liquid viscosity, Ib/ft-hr . All liquid properties are evaluated at: 3:
. T g
=T + 0.31 (TCI ~
w The solution can be arrived at by iterating on Tg7, unW Equation (1) is satisfied. t Two sets of calculations have been perfor1ned to illustrate the effect of noncondensibles on heat transfer from the primary side to the secondary side of the steam generators. These calculations assumed! For Maine Yankee: a) A secondary side pressure of 1000 psia. b) 2011 ft (at STP) of noncondensible (hydrogen) are released. I ,
\
c) The noncondensible was evenly dispersed throughout either the entire primary system or in the four steam generators. For the Yankee plant at Rowe: a) A secondary side pressure of 900 psis. b) 440 ft (at STP) of noncondensible (hydrogen) are released, c) The noncondensible was evenly dispersed throughout either the entire primary system or in the four steam generators. Figure 1.3-2 and I.3-3 give the results of these calculations as a function of primary side pressure for the Maine Yankee and Yankee plants, respectively. The top curve corresponds to the case Where the noncondensible was evenly thixed throughout the entire primary system. The bottom curve in each graph corresponds to the case where the noncondensible was uniformly i
U 4 2 distributed in the steam tenerators ottiy. The graph's ordinate represents the
'D heat flux with noncondensibles divided by the heat flux without noncondensibles for the same primary side pressure. As the curves show, for a primary side pressure which is approximately 200 psia greater than the secondary side pressura, the reduction in heat transfer is less than 10% for J
Maine Yankee and about 13% for the Yankee plant. Due to the large heat transfer area (151403 ft for Maine Yankee and 42575 ft for Yankee) available for condensation, this type of reduction presents no problem. If the primary side begins to approach the secondary side pressure and the effect of noncondensible becomes more pronounced, the net effect will only be to stebilize at a slightly higher primary pressure (about 25 psla) than that for the case without noncondensible. The significant margin in overall-heat transfer is illustrated in Figures I.3-4 and I.3-5, which gives the heat removal rate from the secondary side foe- various primary side pressures. For reference, the core decay heat at 500 seconds has been noted on the ordinate. The top curve in Figures I.3-4 and I.3-5 corresponds to the case without any noncondensible present. The middle curve represents the situation Where the noncondensibles are mixed evs,nly throughout the primary system. The bottom [
' curve corresponds to a case where the noncond 9 nsible gas is only present in the steam generators. Comparison between the steam generator heat removal rates and the core decay heat rate at 500 seconds illustrates the significant available margin in steam generator heat removal capability even when the maximum amount of noncondensible is assumed to be present at the most detrimental location.
In summary, the following conclusions can be drawn. Woncondensible Cases in Steam Generators - St'ar.nant Conditions
- 1. The effect of the presence of noncondensible gas source on steam generator heat' transfer rate is small (10% for NY and 13% for YR),
for primary pressures that are 200 psia or higher than the steam generator secondary side pressure. This small heat transfer degradation has no significant impact on the primary side pressure. > O t .
2 .~ For primary pressures that approach 'the secondary side pressure, the steam generator heat transfer degradation due to the presence ['- \" ' of noncondensible gas is higher. 'swever, this will only lead to increasing the primary system pre ,sure by about 25 psia for the same decay heat rate. .
- 3. The core heat transfer is not affected for these cases since all the noncondensible gas is assumed to reside in the steam generators.
Eoneondensidie cases Eveniv Distributed - Sternant Conditions
- 1. For these cases, the heat transfer rates in the steam generators and the core are not significantly affected.
References (1.3-1) Liliane Schor, et al., " Justification of Reactor Coolant Pung Operation During Small Break LOCA Transients for Haine Yankee," YAEC-1423. April 1984. \,/ (1.3-2) Maine Yankee Plant, Reactor Coolant System, System Description No. 4467-010 Revision 0. Combustion Engineering. Inc. (1.3-3) Perry, Robert H and Chilton, Cecil H. , " Chemical Engineers' Handbook," Fifth Edition, McGraw Hill Book Company. (1.3-4) Yankee Plant at Rowe Technical Specifications. (1.3-5) J. N. Loomis, et al., " Reactor Coolant Pump Operation During Small Break LOCA Transients at the Yankee Nuclear Power Station," YAEC-1437, July 1984.
" Convective Boiling and Condensation," Second (I.3-6) Collier, John C.,
Edition, McGraw Hill Book Company. C
\
TABI.E I.3-1
~
.L/ Sources of Woncondensibles for the Maine Yankee Plant _
Volume ~ (STP)_. Hass Source.- 515.0 ft3 ~ 2'.69 lb l'. Dissolved hydrogen in the primary coolant 768.0 ft3 56.0 lb
- 2. Dissolved nitrogen in the accumulator water 866.0 ft 3 65.09 lb
- 3. Dissolved air in RWST 4127.0 ft3 21.57 lb 4 Hydrogen released from zirconium water reaction 96780.0 ft 3 7057.0 lb
- 5. Free nitrogen used to pressure secumulators !
N/A N/A
- 6. Hydrogen released from radiolytic decomposition and injected water 1236.9 ft 3 He 10.08 lbm
- 7. Fission and fill gas in reactor fuel Ar 0.30 lbm N 0.31 lbm Kr 5.52 lbm Xe 61.42 lbm N/A N/A
- 8. H2 gas in makeup tank 630.0 ft3 3.292 lb
- 9. Pressurizer steam space gas Note:
a) For breaks requiring the return to natural circulation no fuel rod rupture or oxidation is predicted. b) For breaks requiring the return to natural circulation, the SIT's do not inject water. O \
TAIAE I.3-2
' Sources of Woncondensibles for the Yankee Plant at Rowe '
Volume ' (STP)_ Mass Source Dissolved hydrogen in the primary '117.3 ft3 0.613 lb 1.. coolant Dissolved nitrogen in the accumulator 11.38 fL3 0.8298 lb' 2. water
.137.89 ft3 10.365 lb
- 3. Dissolved air in RWST 813.74 ft3 4.2526 lb
- 4. Hydrogen released'from zirconium water reaction N/A N/A
- 5. Free nitrogen used to pressure accumulators N/A N/A
- 6. Hydrogen released from radiolytic decomposition and injected water Fission and fill gas in reactor fuel 103.37 ft3 He 1.05 lb
- 7. Ar 0.05 lb W 5. 85 x 10-4 Ib Kr 0.044 lb Xe 0.49'1b l H 9.0 x 10-5 lb N/A N/A
- 8. H2 gas in makeup tank .,
184.5 ft3 0.9642 lb
- 9. Pressurizer steam space gas Note:
a). For breaks requiring'the return to natural circulation no fuel rod rupture or oxidation is predicted. b) For breaks requiring the return to natural circulation, the SIT's do not inject water. O I t
TABLE I.3-3
. []; .
Maine Yankee
;(j Primary Temperatu,res With and Without Woneondensibles ,
Assumes Woncondensibles Evenly Distribute $ in Primary System 4*> . an Primary f(OF) w/o T(DF) w/ d T(OF) Pressure (psia) Noncondensibles Noncondensibles 1050.0 550.53- 549.89 0.64 1200.0 567.19 566.59 0.60 1400.0 587.07 586.53 0.54
'1600.0 604.87 604.38 0.49 1800.0 621.02 620.57 0.45 2000.0 635.80 635.39 0.4 i
O .
i 1 i TABLE I .3-4 t
' Maine Yankee ,-
Trlmary Temperatures With and Without Woncondensibles 1 Assumes Woncondensibles only in Steam Cenerators_ , J Primary ' ) T(OF) w/o T(OF) w/ I Pressure Woncondensibles (psia)_ Noncondensibles .$ 547.26 3.27 1050.0 550.53 564.19 3.00 1200.0 567.19 584.36 2.71 1400.0 587.07 602.38 2.49 1600.0 608.87 618.72 2.30' 1800.0 621.02 633.66 2.14 2000.0 635.80 O . I 1 l l l 0 - -. _ _ _ - _ _ - _ _ - - - _ _ - _ - _ _ - _ _ - _ _
TABI.E 1.3-5
.N Yankee Rowe Primary Temperatures With and Without Woncondensibles
' Assumes Woncondensibles Evenly Distributed in Primary System nr Primary T(OT) w/
d T('F) Pressure T(OF) w/o (psis)__ ~Woncondensibles Woncondensibles 544.58 544.05 0.53 1000.0 550.53 550.01 0.52. 1050.0 567.19 566.72 0.47 1200.0 587.07 586.65 0.42 1400.0 604.87 604.49 0.38 1600.0 621.02 620.67 0.35 1800.0 635.80 635.48 0.32 2000.0 4 l
~
> t
,,q . .
TABLE I.3-6
\ Yankee Rowe
- Primary Temperatures With and Without Woncondensibles Assumes Woncondensibles Only in Steam Cenerators Primary T(OF) w/ / T(#F)
Pressure T(OF) w/o (psia) Woncondensibles Woncondensibles- . . . 544.58 541.74 2.84 1000.0 550.53 547.78 2.75 1050.0 1200.0 567.19 564.66 2.53 1400.0 587.07 '584,79 2.28 604.8,' 602.78 2.09 1600.0 621.02 619.09 1.93 1800.0 635.80 634.01 1.79 2000.0 O . O t
M THE INFLUENCE Or NONCONDENSIBLES ON
- INTERRACIAL RESISTANCE q
- 't k[ ,
1 DIFFULION g .;, LA,YER , BULK MIXTURE l
- QP l
% L .
I O PGO s Q' . s I N D( PGI
% F
% I s L Q TGO M
.k
- TGI W s
/~ ( pgg O]
1
' C PAO
. i P = ~ TOTAL SYSTEM PRESSURE PGO = STEAM PARTIAL PRESSURE IN BULK, MIXTURE PGI = STEAM PARTIAL PRESSURE AT INTERFACE PAO .. NONCONDENSIBLES PARTIAL PRESSURE IN BULK MIXTURE PAI = NONCONDENSIBLES PARTIAL PRESSURE AT INTERFACE TW = WALL TEMPERATURE TGO = BULK MIXTURE TEMPERATURE TGI = INTERTACE TEMPERATURE Figure 1.3-1 The Influence of Noncondensibles on Interfacial Resistance O l.
t
9 i N.) m
- o. . .
I C' I - - 1 I -O N.
- N Po ,
55 a-
- o. e 8
-c 88 5 58 N
== g is i- ;
<i<, o e 3
e O -" di ' -Om- a
*>) CD M { $
=
g
@ 5 Q- =
o. o v ~3
.o 1
- o L.t3 a
- Sl CD y ,$
o oy
=
e (n
.s.
E. o C$ % b
- g- O Q_ u
- l) < y
.* w in3
- o. - -
-O u o
( N
=A 3 m
O. O O O l ! ! l l 9'O VD E*0 0'O l 0'I 8'O 3198 IU3H O t
- ~-- - _ - _ - - - -.
y , T
'4.'
's
. 4 s# .
- o. " ' . -
l- 1, g g o . h 8E-
** 9 o
wp' Eoz -o o 88 zz N EE-11 o.. =
- o a.
e s9 o ,, -o C_D E' o, m C C
- o. g~ '
e. M ) -o 1 g LO w .
- eo.
C o z. 8-c5 D *E fl ~ C$ (f) %$ AJ - > i tn ee LJ EI og z 05 w
-o 7.
~ "
( N be
- o. -
a
-o m
- o
~
o. b o CD I I I i 0*I 90 9'O &~0 20 0~0
'3198 IU3H
\
~
qv) . o..
. .. i i i g .
; N N
Br j 55 o . o , gO _a .
= 0 :
ge m x b
- o. .2
"" _8 - i i
s a- & u
. in 8 Q_
.o w y
- o o c
_ -o W d s%O *
.O I .
(O i CD 5 s 9 o w% x _ 0 0_ a e w e u g- = 8
- 9 :
a a i
< -o - :
. m
, j u :
E E
, o. w
- < 0 <
sa o 1 o 1 o ooos ao$ot oo$oc.0obot 0obot oo (8H/n18) 930'I x XnlJ 193H m (J 21- l t ;
)
,-Q . -
3 p u ..
- o. -
, i e i i i r
- g. .
N *
, N o in n
zz c. o
.. _a all n i Mw e
=Us o :
iss n., _a d i e CD ' :: uoo - E cc t o- .g g cn r _ - o o_ "
=;
E. E o e e 9 ocD st i, ' o ,= _ -o m == I cn W C: ! tt Oc se o a_ - o T- _ u -o n
= 0 a 8 8
- E a
$ . E
, o
._ u o w
- c a.
E u o . o o
.# CD I I I I I 0'00k10'00E10'00010 009 0 009 0*00F 0 002 0'O (8H/nl8) 930'IxXnlJ 193H
(' 2 2.-
\
.q i
L 0.7.4 h/ Justify why any sources are omitted. , i
)
1 4 4.~ .
'I A.T.4 .. -
Please see Response A.I.3. i 0.T.5 Clarify how mixtures of more than one noncondensible gas such as nitrogen and hydrogen are modeled, especially if fractions.of each are i changing in space and time. A.I.5 Please see Response A.I.3. 0.I.6 i' clarify under what conditions natural circulation flow can be interrupted by the accumulation of noncondensible gas plus steam in the U-bends,andwhatconditionsarenecessarytoke-establishnatural circulation flow. A.I.6 Please see Response A.III.1. . t
.r 2.d.:.1
\ Clarify.how the pump curves are changed if noncondensibles are present.
.1 #'
'F '
A.T.7 .- . Input data that represent the pump homologous curves are not changed for cases where some noncondensible gas might be'present for the following reasons:
- a. The most prevalent constituents in pump components will generally be water and/or steam. Large concentrations of noncondensible gases are not expected in pump components for LOCA cases that' comply with the ECCS criteria in 10CFRSO.46. For example, see the answer to Question Q.I.3. .I i
- b. Since large concentrations of noncondensible gases are.not For example, expected, they are excluded from our LOCA analyses.
see the answer to Question Q.I.3.
- c. Perhaps the most systematic comparison of steam water versus air-water mixtures on pump two-phase performance is presented in Reference 1.7-1. This study indicates that both two-phase mixtures show similar trends of head and torque degradation as a function of void fraction. There is some indication that the air-water data may involve somewhat scre head and torque degradation than the steam water data (see Figures 6-2, 6-4, 6-5 and 6-7 from Reference 1.7-1 attached). However, the dat,a scatter (particularly in the air-water data) preclude drawing quantitative conclusions with a i high degree of confidence.
l Therefore, we use pump homologous data derived from steam water tests for all LOCA analyses. i Reference j (1.7-1) Kamath. P. S., Swift, W. L., "Two-Phase Performance of Scale Models 1
)
l of a Primary Coolant Pump," EPRI NP-2578. Electric Power Research s Institute Palo Alto, CA. September 1982. I
~24-
\
s m_________.__._.__ _ _ _ _ _ _ _ _ _ _ . . _ . _ _ _ ._ _
y, -. .__ _ _ L3 Creare Air / Water REFERENCE I.7-1
- ,f' i \
O asu . 1.5 --' . j a i. , g s* 4 e k Single e 0 2 e !' Phase 1.8 -- e sp e "z ee De
=
- e. e 6- o c'
- j. .. ..
,ee e
D
'e e
. .o
- 8 e e.
De
=
e g e e j, a- _me I e , e e D D0 m." U b a O A z s.s :
/% ,
e.: a s. e .'s 8Mee %75 Is ._ sg-- Ese I g,,.
. Yohd fr action. gr O
e Eb 4 4 e
-0. 5 --
4 4
- 4. s - -
Figure 6-2. Comparison of Creare Air / Water Head Correlation wit.h Creare and B&W Data For V/a, hMen O.80 and 1.20 (
\
\
1 a l I' C
.: ; [+ -
i' \ V) D Creare Air / Water REFERENCE I.7-1 e B&W s.s -
. D e ,
, m to m a C m 5
- .e - a, eg 8 e E ,- e ge o e
% e e .
2
~g e ee a 0 1 8 e c -
e c e O D D e e a e oG
- ae c es "
- e B C,9 p r 8 e e
.t e
i . o *e *D 8 e ee , e r T i . . .. ..e8 .. . (' . , , _ . e:s e:t n:s ~ e:s n:s n:s s:i ~ e.'s n:o s.'s Vold Froetton. or
-e . 5 --
-1. e -
. Figure 6-4. Counparison of Creare Air / Water Torque Correlation with Crearv Air / Water vs B&W Data For v/g Between 0.80 and 1.20 O --
.y-P l
,( ' o C-E . I REFERENCE I.7-1
~
o cr. ore steam / Water l s.s -- " j e
- t. s ->g e e
- n, s
e- e
.c e
. e t , o E-s.s--
t B e e ig. .
. m E m E a
~
2 ' i 4 g , , , . . - -
- = : :
s.s
= 8.8 ,;, s;c s;s el. s:s s:6 s:t v.s s.,
Vold Frection,of 1
- m
-s . 5 a' l
1 l l
-s.e -
Figure 6-5. comparison of creare steam / water need correlation with creare and c-E Data For p ,e500pelaForv/cgSetween 0.90 and 1.10
\
27-
l \. l .i O c CE Pupe < 568 REFERENCE 1.7-1 e' Croare Stamm/ Water
._ u s.s -- .-
. ?
4' 4 e e- D s.a - Do o # e . e e
% e s
e
- ee s
- e
~
eao eg fo . q e c h e.s -- Om m. I .
.b.
E - h e O i ... 4 ' .:. , a aa a aa a a ... Vold Frec' ion,or 4
-0. 5 --
1 4
.s.D -
Figure 6-7. Comparison of Creare Steam / Water Torque Correlation wit.h
. creare steam W ater and C-E Data For v/ g Setween 0.90 and 1 10 and p < 500 psia k
s
i l Q.I.8
'~
; I s~ ' The condensation model is described as follows on p. 35 of Ref. 10:
"['g = K (1-X+Xc ),(X,-X), where: .
K = 1.0x10 (Kg/m -sec), empirical constant
~
X = 1.0x10 , empirical constant." c "From the code assessment, we have inferred that this model tends to overpredict the condensation rate when subcooled ECC is injected into a stean environment. This causes the system pressure to be lou j and results in degraded heat transfer in the fuel bundle region as seen 1 in the TLTA end LOFT code assessment cases discussed in Section 5.0 of Volume III. An improved model in this area is' desirable for best-estimate analyses. For licensing analyses, the degraded' heat transfer effect l and the conservative assumptions imposed by Appendix K yield conservative results." y\ i '% -)
"The presence of a noncondensible gas is taken into account in !
the vaporization and condensation models described above." Clarify l how the presence of a noncondensible gas is taken into account in this equation. A.I.8 Significant quantities of noncondensible gas are not expected within the primary or secondary coolant systems as discussed in the answers to questions 1.3 and 1.21. However, the current RELAPSYA condensation model described above does account for the presence of noncondensible gas. For the same total pressure and gas mixture quality, increasing amounts of noncondensible gas will decrease the static quality difference, This is further X,-X, and therefore decrease the condensation rate. explained below.
, -~3 r
,7 . . , ;. m o , !- The static; quality of a gas. mixture is defined by equation (77) on page 21 of Appendix'A in. Reference 1.8-1 as follows:
(1) X = X, + Xy-
, where:
X = static quality of the gas mixture X, =- static quality of the noncondensible gas
= . static quality of the vapor
-X, t
Note . that for a' fixed value of the mixture quality, X, the vapor quality will decrease ~as the noncondensible quality increases toward X. :This causes a corresponding decrease in the vapor partial pressure for a fixed total pressure. The two-phase, multicomponent mixture specific internal energy is' defined - by equation (80) in Reference I.8-1 as follows: lt (2)
;[ .,
U =.(1-X)Uf +.X,U ,+ X yy U where: U = mixture specific internal energy U = liquid specific internal energy f U,
= noncondensible specific internal energy Uy = vapor specific internal energy Equation (1) can be used to eliminate X, in equation (2) to obtaint U = (1-X)Uf + X ,U ,+ (X-X,)U y W
i _ - _ _ _ _ _ _ _ _ _ _ = _ _ _ - _ - _ _ _ _ _ - _ _ _ _ _ _ _ _ . _ _-- _- .____ _ ~ . _ . _ ___
k ; i ., - ( ,g..-- , is obtained Tum equation
- The equilibrium quality for the gas mixture, X,, .
gr. ~(3) by setting / f .I Uf = Uf ,, =Uy,- Uy X.= X,, .
~'
- where,U f , and Uy , 'are the liquid and vapor specific internal . energies l
at the saturation state corresponding to the vapor partial pressure, (T,). We then obtain: P y
.(4)
X,,= (U-Uf ,'+ (Uy,-Uy)X,]/(U y ,-U f ,) o.
'l The static Lquality .dif ference, (X,-X)', is obtained by subtracting X from equation (4) and then substituting equation (3) in to eliminate U. This yields:
(X,-X) =.I(1-X)(Uf-Uf ,) + (x-X ,) (U y-Uy ,) D(Uy ,-U f ,).<(5) We now examine the limiting condition where the noncondensible quality, When X > 0.5, the X,, approaches the fixed gas mixture quality, X. _ least massive phase is the liquid. Therefore the 5 equation model sets-the . liquid phase to saturation conditions, Uf = Uf , (Py ). Then equation-(5) reduces to:
~
(X,-O = (X-Xn}( v~ vs}!(U vs is When X < 0.5, the less t which tends towards zero as X, tends toward X. massive phase is the gas. Therefore the 5 equation model sets the vapor phase to saturation conditions, U = U , (Py ). Then equation (5) reduces y to: , (7) (X,-X) = (1-X)(Uf-Uf ,)/(U y ,-U f ,)
~
1 O
From pages 21 and 22 of Appendix A, Reference 1.8-1: [ A)m .' ?
=U
' L/.
U f
'f (P, T )f.
U,'=U,[P),T,M g g y ,
~
As the noncondensible quality increases toward X, the vapor pressure decreases and the equilibrium temperature approaches. the liquid temperature. In the limit, U,dUf f and the static: quality difference tends to zero. Therefore, the condensation rate approaches zero as the noncondensible quality approaches the gas quality. References
'(1.8-1) R. T. Fernandez, R. K. Sundaram, J. Ghaus, A. Husain, J. N. Loomis, .
L. Schor, R. C. Harvey and R. Habert, "RELAP5YA - A Computer Pro, gram for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description," Yankee Atomic Electric Company Report YAEC-1300P, Volume I (October 1982). (Proprietary) 0.I.22 Clarify how the effect of any buildup of noncondensible gas in the volume attached to the HPI or accumulator is included in the condensation rate during injection. A.I.22 See the answer to Question Q.I.8.
' f 32-k 4
1
\ ' ) O.T.23 V) Pages 166-168 of Reference 10 describe the energy transfer to the accumulator nitrogen as the_ sum of three terms: (1) the heat convected' ' j from the wall to,the nitrogen, (2) the heat convected from thi liquid
- surface to the nitrog'en, and (3) the additional energy trans,fer caused by condensation in the gaseous phase of the liquid that is vaporised at the interface. In the presence of mass transfer, the driving potential' j i
for heat transfer (Lewis number assumed equal to 1) is the enthalpy dif ference rather than the temperature difference (Reference 14), clarify W y the heat conveeted from the liquid water is in terms of temperature rather than enthalpy. I A.I.23 The accumulator model presently incorporated into RELAp5YA corresponds i to the RELAP5/M001 Cycle 18 version described in Appendix A of Reference I.23-1. The present model supersedes the model described in
. Chapter 3 of Reference I.23-1.
The heat convected from the liquid water is given'in terms of an enthalpy difference as indicated in Equation (208) of Appendix A to Reference I.23-1; shown below: Qcond " vap I w' ~ f D The temperature difference Wich appears in Equation (206) is one of the terns used to assess the water vapor concentration gradient at the liquid-gas interface. The concentration gradient is used in the context of Fick's law with the objective of assessing the evaporation as described in Appendix I.23-1. rate My ,p. O \
APPENDIK I.23-1 Assessmentof the Evaporation Rate at the Liquid-Gas Interface C' . . Equation (206) of Appendix A to Ref'erence I.23-1 can be derived by assuming that the vapor,1, ration at the liquid-gas interface,in the accumulator can be approximated by a quasi-steady diffusion formulation. Then, applying Fick's law, which is given by Equation (3-12) of Reference I.23-2 in the form: (1) h,p=-A,e y ed* my ,p where: hvap = vaporization mass transfer rate,(Kg/s) Aw = liquid-gas interfacial ares,(m2 ) d = diffusivity of water vapor in air,(m2 /s), equal to that in nitrogen These quantities are defined consistently with the nomenclature of Appendix A to Reference I.23-1. Furthermore,fisthemixture(Nitrogenand ( vapor) density at the interface (which is assumed to be equal to the saturated water vapor density, , at the liquid surface). Finally, my is the concentration of vapor at the liquid-gas interface (kg of vapor per kg of mixture), i.e.: van ' van (2)
,vap , y, ,
Assuming that the mixture density is approximately equal to the vapor density, applying the perfect gas assumption to the volume ratio in Equation (2) and substituting the result into Equation (1) gives: h,p=-A,e y g ed e (h(Py /PD' where P is the vapor partial pressure at the interface and P D p mixture or dome pressure, which is not dependent on x. t
L The perfect gas law'can be once s' gain'used to estimate: p , b P dx vap E T1bdx T g.
-(4) p
.- Substituting Equation (4) into Equation (3) and integrating;over a diffusion length, L ,p leads to Equation'(203) of Appendix A to Reference I.23-1 in the form:
(5) yp" g e A, e h3 e h, e (T, - Tp)
~
whereh,=TD represents the equilibrium compressibility for a perfect gas and h =M p in (mN h a man transfer coemeied. 3 The mass transfer coefficient can be written in terms of the heat transfer coefficient by applying Reynolds' analogy.as in (Equation 11-31) of Reference I.23 M, which is easily rewritten in the form of Equation (204)'of Appendix A to Reference I.23-1 after minor algebraic transformations.
- O ,e,.r.nc.s (1.23-1) R. T. Fernandez, R. K. Sundaram, J. Chaus .A. Husain, J. W. Loomis, L. Schor, R. C. Harvey and R. Habert, "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume 1: Code Description " Yankee Atomic Electric Company Report
.YAEC-1300P, Volume I (October 1982). (Proprietary)
(I.23-2) W. M. Kays, " Convective Heat and Mass Transfer," McGraw-Hill, 1966. l
-(1.23-3) J. P. Holman, " Heat Transfer." McGraw-Hill, Fourth Edition,1976, Chapter 11.
I 1 O -3s-1 s
I 9 1.24 ,
'6 f, . )
'u Clarify how the energy transfer from evaporation of water at the liquid surface is included. Assuming the heat- and mass-transfer analogy, the mass flow of vapo.r leaving the liquid surface is equal to g ,(( - w i )/(w
- 1)
- (Reference 14), where g ,is the mass transfer coefficient and w ,is l
the mass fraction of water at i the interface or at twin the main l nitrogen volume. A.I.24 The primary objective of the accumulator component model is to calculate the injection rates into the cold leg. An important
. parameter in the. calculation of these flow rates is the accumulator gas dome pressure history. The gas dome pressure is somewhat sensitive to Therefore, the energy the condensation rate of steam in the gas space.
transfer to the gas space from the evaporation of water at the liquid surface is estimated by assuming that all the evaporated water (Equation (203) of Appendix A to Reference I.23-1 W ich is derived in Appendix 1.23-1) recondenses (Equation (207)) and that the heat of V . . The impact on condensation (Equation (208)) is added to the gas space. the injection rates due to variations of the energy (temperature) of the liquid slug (caused by heat removal from the liquid slug due to evaporation) is negligible. Therefore, the evaporation energy, which is supplied by the liquid slug ic neglected and the liquid is assumed to remain at constant temperature. In summary, the condensation energy is e. counted for in the gas opace. In the liquid slug, the evaporation energy is neglected because it has no effect on the injection rates.
/ \
b l i
0.I.25-Justify Equation (3.4-32) of Reference 10. Please supply a copy of . Reference 3.4-1 of Reference 10. , , , A,I.25 This equation corresponds to Equation (206) of Appendix A to Reference I.23-1 in the current acetniulator model. The equation is justified in Appendix I.23-1. Requested reference is supplied as Appendix I.25-1. O O . i i
)
y ... o .] 'f . i
~
1 APPENDIX I.25-1 l
.I A MECHANIST 8C ACCUMULATOR MODEL FOR LIGHT I WATER REACTOR TRANSIENT ANALYSIS' )
/^)
( b/ K. E. Carteen, D. L. 88.gol. V. H. Mennom and J. A. Tesep rcsc ideso. iac. edeha Natsonal (apensering Laboratorv idaho Fans, Idaho I l _
'AB57RACT V Vo lume A lumped parameter accumulator model has been
~
rieveloped and installed into the computer prograr. 2 Elevation RC L AP $. The model includes heat transfer from the creek wall of tne tank, f<om the water surface to the dome vaporization from the water surf ace to the e Thermal dif fusivity dome, and condensation of vapor in the dome. The a Coefficient of isobaric therinal expans ton Int ftature also includes heat addition to the a Change of j nitrogen dome due to the condensing process. 6esults from data comparisons are presented NOMENCLATURE , Kinematic viscosity ,! Cross-sectional flow area Subscripts C N1tiplier for heat transfer correlation c Condense Cv $pecific heat at constar;t volume f Liquid D Diameter g Vapor Diffusivity K Upstream volume indes (Accumulator) f] d f Coefficient of friction L tbwnstrtaa volume indes (System) g Gravitational constant 1 Surge line Gr Grashof number n Nitrogen h Enthaley 5 Saturated hg. hl Wat transfer coefficients t Tank h3 mss transfer coefficient v vaportration k Thermal conductivity INTRODUCTION L Le ng th This p' aper describes an accumulator model that n g,, has been developed for installation into a systems 8 b . Nss transfer rate analysis computer code (ItCLAP5 ). for analysis of P Pressure light water reactor (LWR) behavior. This model features mechanistic relationships for heat transfer ir Prandit numbe, from the tank wall and water surface, condensation 0 Total Mas transfer in the vapor done, and vaporlastion from the water j' A Nitrogen gas genstant T Temperature a. The Itt;LF5 computer code and its documentation. IELAPS/ MODI. Code Description Vol. I and !! by t Time- t'. H. Ranson et .al.. are available from the National Energy Sof tware Center. Bldg. 208-acoe C-230, 9700 l H Internal energy Soutn Cass Avenue. Argonne, Illinois. Nort suppened by uw U. 5. %ense Re,Asewv.'M N d Ad**' AW A* marsh e DOE Ceavset No. DE4C07-7a100 t570. p I i
i r , Sir f ace to the dome. 1%e a:c mslitor model'Ionsists' %,,,,,,,
* * {
of a hydrodynamic model and a heat' transfer model. f] . **= v.
"I
) 'The latter includes the vaoortration and "*""** ~ -
condensation effects. , 1
. - 1 The hydrodynamic mortel is descr,1)ed along with
)
the heat transfer models for the accumulator, t Perfomhnce of the model is verified by data, comparisons between the model results and the U,l D -( LOFT L31 test and also a Westinghouse Electric Corporation LDper Heat Injection (LHI) accumulator perfomance test. .. NYDR0 DYNAMIC MODEL c . u ~, . l I i The accumulator is modeled as a lumped-parameter component. This model was chosen 3 over a distributed model Decease the spatial m gradients in 'the accumulator are expected to be . small. . rs .r ~ v *. The accumulator model and associated notations Figure 1. Typical accumulator, are shown in Figure 1. The basic modeling assumptions are: Using these assumptions. the basic equations
.j/N j '
governing thermal. hydraulics of the tank and b/ 1, Heat transfer from the accumulator walls.
'l Heat and mass. transfer from the liquid are discharge line can be written as l
modeled using natural convection
- 1. Conservation of mass (nitrogen)
. correlations assuming similarity between heat and mass transfer from the 11guld Mn . tonstant . e, V, (1) surface.
The nitronen is modeled as an ideal gas 2. Conservation of energy 2. with constant specific heat. Thesgmin the dome entsts at a very low partial Nitrogen pressure and is not modeled dirtCllv. The . du dV energy released as a result of veoor W condensation is transferred to the M n
. -P gg+0 ni troge n.
tan 11
- 3. Because of the high heat capacity and large mass of water below the interf ace. di W the water is assumed to remain at its M, C,t M.-O g initial temperature.
- 3. pbmentum eavation
- a. The model for liquid flow includes inertta. well friction. fore loss and 9ravity ef fects. (t .
,[)ef,,.'.a,vg I 22
- . L :
t 1
- ________-____ _- _ a
-__ ~ __ _-- - - _- __. _
. . (P . P )Ag g
- egg (g *gg)A, g (a ) Og. ng (Tg - T, ) . (g)
.m .
f . where f is the coef fiCient of f riCtion. The wall Convective film Coef fICient has been defined u i g a correlation for turbulent natural 4 ,$ tate Relationships -. ConveCtle for verticai plates as v PVn * "n"n , f (5) . h 3 0.1 [ (GrPr)g$3 (10) Un . C, In . (6) n where L is the sum of the lengths of the tank wall above the liguld surface and the radius of tne tank Using [quations (1) and (6), the nitrogen too. The Grashof number is given by energy equation (Equation (2)). can be rewritten as 3
&.ga(T-T,)h g (11)
M, C [di .-P g vA gy + Q. (M where L is as above plus the radius of the liculd Olfferentiating Equation (5), and solving for the surf ac e. pressure derivative yields In addl i n to wall heat transfer, convective heat dPg Pg di transf e from the accumulator liguld is assumed n Ex d #n (0) to be W
- KW ~ QW O*A f Kh 2 (Tg - T,) (12)
[quations (4), (7), and (8) comprise the system (- of three dif f erential equations used in the ( accumulator model. They are us'ed numerically to where advance T,, Yn , and Pg in tier. HEAT TRANSFER 2*I*" ~
- The variable D is the diameter of the water surface The heat transfer to the nitrogen in the dome and the Grashof number is the same as above.
has been modeled as being due to two dietest processes. The first mode is convective heat transfer liquid is also assumed to evaporate from the due to the temperature difference between the water surface due to the density difference between nitrogen and the accumulator walls.and ,11ould. The the vapor at the water surf ace and in the nitrogen second mode of heat transfer is postulated to be due 'above. The evaporation rate is assumed to be ta condensing water vapor in the accumulator dome. The steam in the dome is assumed to be saturated at . v
.h (e,g - og )
the gas temperature. Steam is assumed to be 3 cinvected from the liquid surface due to the motion af the nitrogen. A mass and energy balance for the . e - Ty (13) steam is then used to calculate the heat transfer to Il the nitrogen due to convection of vapor from the
' liquid surface and subsequent condensation in the where accumulator done.
'g g . a function of the liguld The heat transfer from the tant wall has been teaoerature.
modeled with a lumped-parameter model The mass transfer film coefficient, h3 , is u ,, I i
h j l i determined assuming siellerity between heat and mass CALCULATED RCSULTS l 6 transf er from the water surf ace (both are conserved f
/ <
- properties modeled by similar equations), and is The accumulator model described previously has
\ been tested in REL AP$ and gives good agreement with
. given by experimental data from LOFT Test L31. and a W'5II"9"0"I' I *CC"*"I*I'# "'I"# AC ' ' -
(g)U3 (14 ) h3.hg\n " test (WCAP 8471 . The' comparison to the q experimental data in both tests is of the form of where d is the di.ffusivity of water vapor in at pressure versus nitrogen volume. This type of defined as comparison makes the effects of heat transfer in the accumulator more apparent than either pressure or 2.3 temperature versus time. comparisons, and also P, 7 d . 0.239 y- (15) obscures most of the ef fects of uncertainties of the f K boundary conditions. The LOFT data used for where P, is atmospheric pressore and 0.239 is a comparison are liquid level and pressure history; dimensional constant that gives *d* the units of I the Westinpouse data used for comparison are
,'m /s). integrated mass flow and pressure history. ,
The mass transf er rate is then The LOFT accumulator was simulated using the : f0110 wing measured data. Nitrogen volume is p3 . " (161 1.39 m3 water volume was 1.97 m3 (0.45 m3was m, . h2 (hI b '9 'g (Tg - T,). Jrge line and 1.52 m , in tank), system pressure was 4.37 MPa. and the water and rogen temperature The steam in the accumulator dome will condense as was 304.7 K. A form loss of 24. was included at the temperature decreases and the additional heat of the surge line exit to model wall form loss as well
.O vaporization will be transf erred to the nitrogen in as wall friction and form loss in the surge line.
i 1
'v/ the dome. The condensation rate is calculated Time step site was 1 second.
assuming that steam is saturated at the temperature of the dome The pressure drop between the accumulator and the system was approximately 0.15 MPa. providiag a [nc. + h h,eg). p 6 MW W W acWh W WCW. Figure 2 shows a pressure versus vapor volume The heat transf er due to this enndensation is comparison plot between the LOFT measured data and calculated results. Note that the plot includes a
~
calculated isothermal Curve (uppermost Curve) and a Oc. h q(T,)a h (T,) " h(Ve) 99 calculated 15entropic curve (lowermost curve). In the early part of the blowdown (high pressure and low vapor volume) the espansion of the nitrogen gas
- m, (77 ) - hg (T,). . (18) h,5 5 .
15 essentially adiabatic. As the gas expands into the accumulator dome. its temperature decreases. The total heat transfer to th,e nitrogen is then causing a temperature difference between the gas and calculated as its surroundings. The surroundings then loose heat to the gas, causing the gas pressert to decrease dU less rapidly. The calculated and taperleental 9 0.O g
- Of
- 0,
- e, ,f. m) pressure versus nitrogen volume curves are seen to be in good agreement, indicating that the heat transfer was accurately predicted. Appres taate ly sisty perr.ent of the heat transfer was calculated to
[\ be due to condensation. 2a 41-
\
u
- l. ' S i i i 1.28 g i g 1 . . . . I ce vistea Isotheres) _
- - I altvistea t Isothermal I? f . k(taps O P5 calculation
! ?
~
- - 5 cattv14 tea inentropic , g ,
"-""$ culatea Isentropic
.J ..
s
% o
_I s ; \ . t \,' s 2 \ G. i 3 3 - \,'s ,, t .64 - V - p A.,*s s l a ' I = . s ' . . ' ,. t s N 2 - T A N, N -
.32 -
2 - s g,
*' $ %q, "'"..,,I
- : ;yy 1 I i i I 1
',, ,3 g ;3 y, .25 .52 .77 1.03-Nitrogen Volume (m ) 9e u atio D,U t fleure 7. L0ri L b1 Accowlat. 'i..e3 Wes tinghouse .m. A. .nula tor.
The Westinghouse LH1 accumulator was the secon SUMMRY emperiment simulated. This accumulator was tested (in place) by allowing it to discharge into the The results of these comparisons to data atmosphere (pressure vessel with cover of f). [be to indicate that heat transfer, especially from the the large pressure drop this performance test condensing vapor in the dome, is important for resembles a large break type of experiment providing accurate representation of accumulator simulations. a fast accumulator transient for verifyinc the The RELAPS accumulator model gives good results for accumulator model. This accumulator was smulated both slow and fast transients, isentropic or by using the measured data and performance polytropic. This techntout gives the user a
\ information contained in WCAP-847 as input. The mechanistic accumulator model to use in the uncertainties for the data were not given in this aulation of light water reactors, proprietary report REFERENCES The form of the converi.es between the experi-mental data and calculated results are of the form of 3. J. p. misan, wat Transfer, ath edition..New pressure ratios versus volume ratios in order to York: McGraw-Hill Book Company. Inc.,1976, preserve the integrity of the proprietary data, pp. 280, 24 4-245.
figure 3 shows the results of the comparisons between the calculated results and the experimental data. 2. R. J. Hanks,1958 Water vapor Transfer in Dry The caleviated isothermal (uppermost curve) and soit, soll Selence Society Procedin95 1956, isentropic (lowernost curve) curves are included 22:3'72-374. for reference. The experimental data shows a similar behavior seen in the LOFT test, that is, the gas 3. WCAP 8479, Westinghouse Electric Corpora-expension is initially an isentropic process. Then tion Nuclear Energy Systems, p. D. Box as the teaserature difference between the nitrogen 355, Pittsburgh, PA 15230 and its surroundings increases, the gas pressure decreases less rapidly. The calculated results ., e, p. nayless, J. s. surlow, R. H. Averill. att again in good agreement with the empeeteental Experimental Data Report for LOFT laselear wall. results. Break Emperiment, letREGICR-ll45, EG44-2007, January 1980, pp. 2A,127,140.
.t' t
25 t
l
. 5. M. J. Welland, tof f f5perfment operattaa sates acy warranty, empressed or imp 1ted, or assumes Specification Small areak Test Sectes (3 any legal liability or responsibility for any third -
8tvember 1979. ett L3 Sortes (05. Rev A. pp.15. Party's use. or the results'of such' use, of any 1 information ' apparatus, product or process disclosed NOT1ct in this report, or represents th9t its use by such third party would not infringe pelvately owned This rtport was prepared as an accpunt of work rights. The views,espressed in, tnis paper are not; sponsored by an agency of the thited States necessarily those of the U.S. Nclear Regulatory fevernment. helther the thited States Govertunent Camruission. noe any agency thereof, or any of their employees. O - 0 P
\
M t
L 1. L . o.I.26 1;; r Clarify how the analogy between heat and mass transfer was used. A.I.26 4 The analogy between heat and mass transfer was used as described in -4 c Appendix I.23-1. The expression for the relation between the mass 'l transfer coefficient and the heat transfer coefficient (Equation (204) of Appendix A to Reference I.23-1) is simply an algebraic manipulation of Equation (11-31) of Reference I.23-2. 0.I.27 Clarify how much effect the evaporation and condensation in the accumulator will have on the calculated pressures and injection rates for SBLOCAs. f A.I.27 6 . The effect of evaporation and condensation in the accumulator on the calculated pressures and injection rates are expected to be negligible. For example, the Maine Yankee test described in Section 2.4.1 of Reference I.23-1 was calculated twice: once with Q cond given by Equation (208) of Appendix A to Reference I.1 and The ca.culated accumulator pressures and flow again with Qcond = 0. rates are essentially identical as indicated in Table I.27-1 and Figures I.27-1 and I.27-2. , O _ _ _ _ _ _ _ _ _ _ . ,
9 ; g lI
'6666666666655555555 0000000000000000000
***+**+++ * + * + *+ ** +*
E EEEEEEEEEEEEEEtEEEE R U) 00Ce377885254876336 0003815973782957500 se eT D
. S A S P 2225721871233809633 8887652727148597190 r
u s N O E( R P 2222222110059495286 . sf es C
/ 111111111 119S377746 rn P a P
A V rT E o t t T aa l e U 001753592450 9426 uH O E H T T 452956.044380 . 5125 m u I A R )C 6 1363879968 49517 cD W E S 0 0344101233 32221 cN A O W / . . 1 1 357911111 11111 OG L 3O0 e/ eP F (K kA nV aE Y 6666 6666666555 0000 0000000000 55555 00000
+++++
eu no
++*+
EEEE ++ + EEEEEEEEEE+ +++++* EEEEE 32932 ih at E R 0008 0003 6544571300 9565941664 40896 Mi U ) 2225 7221207263 05366 d S 41 386 ed 6528382562 o S A 8887 41853 t D N E P R( 2222 2221100728 aa l O P 1111111998
~ 88777 uh C 111 1 '
ct
- / li P
A aW C l V E f ot H T E 1 1 743453166' . 43077
. nR I
W T A )CE 4
. 52249.56042
. . . 82 37558 84185 o
sw R S 6 13740701 33322 i W G/ 0 0344 202334 357911111 1111 1 rl aF OK . . 1 1 p L F( O00 m o C 1 E C
) 330 000 0003000'000 000000000000000 00000 7 2
M E I S 000050505050505050 505 . . . . . I e T( . 25711 025702570257025 1122223333644 l b 3 . a T 1T O
!l
l ji L. -.i i C 0 5 .
- 4 .
- L
~ -
~ R -
E - '
- F
- S N
s A d'
~
R n a
' T 0
. 6 h~ -
13 t N T i W s A E s
' e s H r u
' s -
- D s -
et s
- N r P s e s
O rT C o e
/ 0 t
a ke x's P i 7 l u n A 2 m a V u Y
) c A's E C E
c e A n i R E N T U S ( d M e t r a F s' O E l a o H u f O S. N. A T I W M I T l c a d Con R s 0 f C T s 01 o/p n va T N o s E A i r t E a u p o h H_ mt o i D CW N 1 O 7 C 2 L 0
. 9 1 P.
A e r V u g E i F H T I. W 0 Os o PJ 8- oO oc o'. 0 o* .. ,
<Cv aeacmweC O ,g.
Lll6 'l ll l -
fl 4 I f o (; ' hw*N , , L.! . o . e v
,a .s es
,_ 4 m.
i- , 4 *. i-2 ) l
/ O i i
e O e
/. N l C E ;
k
-o c
.c
. e
_,0- 'M 'n y^ 0 " t 'O a:
/
O Z > s- Y ?Y w e-MI . \' '* h .k O & d' g. c' N)
~c ts_l. %
~1
'J:i EO Z!
m g ..
< U ue.
M W: 7/; e z. gi " 0u i W 0&
= . r' H<'. .- \
ta i 34
% F 08 o R C- e
~
o c. Zi8 c2 0 8 "'
'U! tY M ~
E. .E EG 4A OE M- n. N
, o -
s.'
- c3 h
a
.k O
O O0041 00$71 00b6 0Ob9 0 Obc 0d (s/o>0 salva mou !
.r
\,
\
0 7 28 (
'V The following questions are concerned with the calculation of the Maine ,,
Yankee accumulator test presented in Section 2.4.1 of Reference 12. page 87 of Reference 12 says that the " valve took approximately 23 seconds to reach the fully open position." Table 2.4-1 of Reference 12. shows a fixed friction factor and form loss. coefficient. Clarify how the valve' opening process was modeled, especially because almost half of the 48s transient involved the valve opening time. Figures 2.4-1 and 2.4-2 of Reference 12 do not indicate which curve is the data and which is the RELAp5YA result. A.I.28 The friction f actor and fom loss factor given in Table 2.4-1 apply to the surge line and its connection to the accumulator tank. The valve opening process is modeled utilizing the motor valve (MTRVLV) component as described in Section 7.10 of Reference I.23-l'. A trip logic is combined with an appropriate valve opening speed resulting in valve throat ratios as shown in Figure I.28-1. This valve area history is ['] V believed to simulate the actual valve opening process. The variable fom loss coefficient for the valve opening is calculated by the abrupt area change model associated with this component. l t l l (
f C. I I I f y-o
.s+
e o ,.
-N lj o ;;
N' *
-N .
. \xN e
r N- 9a _ Noen - t 3 O N'N ou i - N _d m-r . e N . N o i-N. _d - y N 9
~
\.s _a 9 .
m _ - g C
\
\ 9 w
Q d 90 t'0 E*O O'O 01 B'0-(XUWU/H19) 01198 1908H1 3AlWA
\
0.I.29 7]. Clarify which curve is which and why the difference in pressures is . increasing af ter 32 s even though the liquid levels are virtually
. identical. .
).I.29 As described in A.I.23, the accumulator model described in Chapter 3 of Reference I.23-1 has been replaced by the RELAPS/ MOD 1 Cycle 18 model.
The Maine Yankee accumulator test has been recalculated with the revised model. ' Figures I.29-1 and I.29-2 compare calculated and measured accumulator pressure and liquid level, respectively. The accumulator pressure and liquid level are well predicted. 0.I.30 Clarify if any temperature measurements were made in the liquid and nitrogen regions that would indicate the temperature difference that
. f.
existed during the transient. , p.I.30 No records of f.emperature measurements in either region have been found. O t
.)
i I O I i 1- f a m i g f
,s .
i q
- o a
-d a
+- :
4"
-88 9 u
-m n u .
..s 8E 5m
-a ts.
n ,. e am 55 C3 5
- gum -m em
~ es 9o $$
O U -
. -o N b -e r :s uo H .
-t l e
-m
- gx:
- ~
B'8 z u. 80 c. . _ s@
& -S DE i 95 o .
- <i s$g- -; c 822at-o
<t-I d
I ' i I I 0*91 0*ti 0 *El 0*01 0*8 0*9 0*t
,01x (WJ) 380SS38d 801970W0008 O ..
Il ! ! l
)lii{ -
O - - - - 0 5 4-0 _ t
. s I i0 e
- 4 .T e_
) e
.k d n 0 n a oY fC I i5 / e pn 3 ai va
. EM t r 0 uo of h
I I 0 .t s, =- 3 i l wev a
.d e nL _
a a d - 0 )C hi t u I I5 2 E i q wi S ( L ( - d r eo t t aa _ 0. 0l 1 E l l uu O cm I I -
. 2I l u ac T Cc A
f - od 0 e nr
. ou ss I
I 5 i a 1 re aM p N md O on 0 Ca C
/N a
I PO I 0 2 AC 1 9 . V/ 2 _ EP A 1 TV e UE 0 r O . u 1 A RHH i 5 g TTT i F AII DHH 0 0 T. - _ .
.O +d od t 9 g o
~3xN g1.>33h_
G 3aasJ - O E' l'{:tl l
III. HOT-LEC OR U-BENDJPHASE SEPARATION L l kf 0.III.I. 0.III.2. 0.III.3 and 0.III.4 ,
- v. ,.
The procedures uged to interrupt and re-establish natural circulation are not clear. (Q.III.1) Clarify how natural circulation will stop and be restarted as vapor and/or noncondensible gas builds up in the top of. the steam generator U-bends. (Q.III.2) Clarify what noding detail is needed to model this process accurately. (Q.III.3) Clarify how the models have been assessed against data. (Q.III.4) Justify that the models for the interruption and re-establishment of natural circulation. are mechanistically realistic. A.III.~1. A.III.2. A.III.3 and A.III.4 Natural circulation is the preferred heat rejection mode for small breaks in which the break flow does not remove the decay heat. From References III.1-1 and III.1-2, the maximum break size for which natural circulation is important.is a 3" break for Maine Yankee and a ( 1" break for ,the Yankee plant. Break sizes larger than these would not be affected by the loss of natural circulation since sufficient energy would be removed through the break. From Reference III.1-3, there are two mechanisms which could result in an interruption of natural circulation. One is the collection of noncondensible gas in the top of the steam generator U-bend. The other-is an imbalance in the secondary steam generator inventories or heat removal capabilities. l The imbalanced secondary condition does not preclude adequate heat removal since the unaffected steam generators adequately remove decay heat and provide natural circulation flow paths. The collection of noncondensible gas in the U-bends does not appear to be applicable since the Semiscale tests (Reference 111.1-3) showed adequate heat removal and no flow stalling with up to 13% of the primary volume filled with noncondensible gas. This is due to the fact i V 1
u 1, that the noncondensible gas disperses and does not collect in the U bends to stall the flow. Only when the noncondensible gas was
!/~'
l p ." injected.into the steam generator at a rapid. rate did the flow stalling occur. . 4
. For Maine Yankee and.the Yankee plant, small break LOCA con,ditions.for .
which natural circulation is desired, the amount of noncondensible gas L calculated to exist is about 0.5% of the primary system volume. Therefore, interruption of natural circulation.is not expected for email break LOCAs for Which natural circulation is required for decay
-heat" removal.
Since the interruption and re-establishment of natural circulation is not expected for small break LOCAs, Questions III.2 to III.4 are not applicable. References (III.1-1) L. Schor, et al., " Justification of Reactor Coolant Pung operation During Small Break LOCA Transients Maine Yankee," YAEC-1423 April 1984. (111.1-2) J. N. Loomis, et al. , " Reactor Coolant Pump Operation During Small Break LOCA Transients at the Yankee Nuclear Power Station " YAEC-1437, July 1984. (111.1-3) D. J. Shimeck and C. W. Johnsen, " Natural Circulation Cooling in a Pressurized Water teactor Geometry Under Accident-Induced Conditions," Nuclear Science and Engineering 88, 311-320 (1984). 0.111.5 Clarify how the heat transfer at low flows in the top of the U-bends is modeled for two-phase flow and also if noncondensible gases are present. O i
l-l \- 1 A.III.5L
? s 5 No special treatment is given to the heat transfer in the top of the U-bends. The standard RELAPSYA heat transfer models described in sections 2.1.3.6 and Section 2.1.3.4, Appendix A of. Reference'III.5-1 i ',
1
. are used. The effect of noncondensible gases is not considered and has no significant impact on the heat transfer coefficient.
1 References' (111.5-1) R. T. Fernander, et al., **RELAP5YA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis Volume I: Code Description," YAEC-1300P, October 1982. (Proprietary) v I L L
f IV. STEAM CENERATOR HEAT TRANSFER L ;
't%
)
l V ' O.IV.1 w Page lii of Reference 15 (Which modeled TLECHT SEASET Test'23402'Where *
- the secondary was hotter than the primary) states that: implementation of the code's time step control algorithm occasionally allowed the primary side temperature to become hotter than the secondary fluid temperature, leading to unphysical oscillations which could be eh ninated by the user reducing the time step used." Clarify how this oscillation problem will be avoided in SBLOCA calculations.
A.IV.1 In the analysis performed so far with RELAP5YA, i.e., Pump Trip Study for the Haine Yankee plant (Reference IV.1-1) and the Yankee plant at Rowe (Reference IV.1-2), and the simulation of the LOFT L3-6 experiment However, if (Reference IV.1-3), no such oscillations were encountered. such oscillations are encountered, the time step will be reduced. m References (IV.1-1) L. Schor, et al., " Justification of Reactor ("colant Pump Operation During Small Break LOCA Transients, Haine Yankee," YAEC-1423 April 1984. (IV.1-2) J. N. Loomis, et al., " Reactor Coolant Pump Operation During Small Break LOCA Transients at the Yankee Nuclear Power Station," YAEC-1437, July 1984. R. K. Sundaram, J. Ghaus A. Husain, J. M. Loomis. (IV.1-3) Fernandez, R. T., L. Schor, R. C. Harvey and R. Habert "RELAPSYA - A computer Program for Light-Water Reactor System Therma)-Hydraulic Analysis, Volume III: Code Assessment," Yankee Atomic Electric Company Report YAEC-1300P, Volume III (October 1982). (Proprietary)
\
0.TV.2 IO) 1\- / Clarify what steam generator noding will be used for SBLOCA calculations. .w. . J A.IV.2 f Figures TV.2-1 and IV.2-2 represent the nodalization used in the Pump Trip Studies for the Maine Yankee plant and Yankee plant at Rowe, respectively. Similar noding schemes will be employed for the small break calculations. 1 As described below, a detailed nodalization of the steam generator secondary was used for both the plants to ensure realistic heat transfer behavior across the steam generator tubes. i l BAINE YANKEE All the U-tubes in the steam generators are lumped into a single flow path. The primary side of the steam generator is modeled with 1,0 volumes and (m} the secondary side consists of 14 volumes. All area changes, baf fle plates and dome structures (i.e., separators, dryers) are accounted for in the nodalization through the loss coefficients and geometries that e e entered as input. Besides the U-tubes, heat slabs representing the external walls are included in the model. JANKEE PLANT AT ROWE i For the Yankee plant, a similar nodalization is used for the steam generator as for the Maine Yankee plant. The primary side is modeled with 8-volumes (6 in the tubes) and the secondary side consists of 11 volumes. All the flow path changes are accounted for in the calculations of the geometry and loss coefficients which are entered as input. The walls of the steam generator U-tubes and veksel are modeled as heat i structures. i
.J il
[
'. f*
t t IYe
.t :-
e O O mm il~" _ . 4 8 In 'iet-:- 3 . k* I
- . s it
'e II $,
]T g g tl
!E r
f-h -
<=
Ey 1- L -
,4 . . i . . . .- -
EU r} i" #r * ~ ' t;le:'- : - f" W' A - I-o
- a l.J I;I! I!. ) 'u :
)g<
O) ' . _ j
'%2 h - : - 0 2, ! }-- ,,,,
EM es y y g. c v:
.g t
j (iniglijliiP - C ( q ,E D - 31 i e p. s- 8 7: I .t n'- I.E
. }g f, *,
* ,, . .* . , g j
.{ f
~
if tr
.i .
g lp;;
. i a,[
f- :. - r
~
. { N G
O -5B-
\
e
-l n .. I u ,
i s.c g 4
- 1 t * .
j_E. e g- . ar- Tjh i+ .! t " 19 d' L _
" s- E 3,
I I --- e I i& h 5! t !-
/ 5 l't E
El!i*;E 25
.: 5 lz- S 5 22 mj$r5a g g 5 13 j
- - - - E e -
tti
--- ,E
" !is C)
V
, I - --/ E6 <
,rs::
E U
' 'Ij f zma
!!! PS
{T T tg 4-I l
~
' ) E e
a g a rr$ n j g 5 'f' 1 % 4 A I e l S l { q)L* - e i E 5" l I
- ;- \
l r s, : 4
-: <r i b 2
,. :: + : - j!ill + :- : E e . ,.
-l :- ' -
+ ::
L
** s, <-
N.jl m # f _ I i 2
# l
\
d
,t i
i 0.TV.3 clarify W at noding is needed to model steam generators for low flow , and' reflux cooling conditions in SBLOCA. t , l
~
.. , j a . TV . s.
Figures IV.2-1 and IV.2-2 provide representative noding schemes which
'will be used for all the phases of a SBLOCA accident, including low flow and reflux cooling conditions.
O - d 0 .,,. i _ _ ___=____ _ _ _ _ _
V. SYSTEMS VERIFICATION AND OTHER EXPERIMENTAL VERIFICATION s" .
- - 4
, (_ o.V.12
- m. . 1 1
Clarify how single-phase and two-phase reverse flow;through the pumps . is modeled for locked or unlocked rotors.
).V.17 Single-phase and two-phase reverse flow through pumps _with locked or unlocked rotors would generally be modeled using the Centrifugal Pump Model.
This model is described in detail in Section 2.2.6 of Volume I, Section 2.3.1 of Appendix A, Section 3.2.9 of Appendix B and Section 7.11 of Volume II Appropriate input data would be used as contained in Reference V.12-1. For the unlocked rotor discussed in'the answers to Questions V.13 and V.15. case with loss of electrical power, the pump angular momentum equation would be solved to determine the time-dependent pump speed. For the locked rotor case, either the pump speed can be entered as zero in a Pump Velocity Table or the Pump Stop Data Card can be used. Note that the Yankee plant has check < O valves in each cold leg that prevents significant reverse flow from occurring. O Reference (V.12) R. T. Fernandez, R. K. Sundaram, J. Chaus, A. Husain, J. N. Loomis, L. Schor. R. C. Harvey and R. Habert, RELAP5YA - A Computer Program for Light-Water Reactor System Thermal-tfydraulie Analysis.** Yankee Atomic Electric Company Report YAEC-1300P (October 1982) (Proprietary) . I
1 i i I 0.V.13 4
/' % Clarify how prototypical the data used is for full-scale pumps. '( ,) .
A.V.13 ,.
- l Table V.13 summarizes the single- and two-phase steady-state data l sources used to assemble the pump performance curves for the Maine Yankee, Yankee and Vermont Yankee reactor coolant pumps.
Single-phase homologous data have been selected to model full scale j pump behavior in the following order of availability and preference:
- 1. Plant-specific full scale pump data.
- 2. CE-EPRI 1/5 scale pump data (Reference V.13-1).
I
- 3. Maine Yankee full scale pump data.
The first preference is to use full scale prototypical data when it is (
' - available from the pump manufacturer. The second preference is to use CE-EPRI 1/5 scale pump data for three reasons:
- a. The geometry of the CE-EPRI 1/5 scale pump is more similar to that of the Yankee and Vermont 7ankee pumps than the Semiscale pump.
l
- b. The specific speed (an important similitude parameter) of the CE-EPRI pump is closer to that of the Yankee and Vermont Yankee
~
pumps than either the Maine Yankee or Semiscale pumps. Pump specific Speed Vermont Yankee 3802 Yankee 4687 Maine Yankee 5556 CE-EPRI 4209 Semiscale 926 1' l
~62-t Li__________._.______ _ _ .i
I
- c. This choice is' consistent with the two-phase homologous data base in the piincipal regions of interest (Curve Types'l-6) discussed A.
, jd below.
The third preference is to use the Maine Yankee pump data'since the
,. geometry, scale and specific speed are closer to those for the Yankee and Vennont 7ankee pumps than the Semiscale pump parameters and desigh. These data are only used for curve Types 7 and 8 in the Yankea t
and Vermont Yankee pump models. At the current time, no full scale two-phase homologous pump data are available. Therefore, the choices are limited to the CE-EPRI 1/5 scale steam water-data, the Semiscale steam water data and the B&W 1/3 scale air-water data. The CE-EPRI data are used for Curve Types 1 through 6 for the following reasons;
- m. Steam water is the two-phase fluid medium of principal interest.
- b. The geometry and specific speed for this pump are closer to the A full scale pump parameters than those of the Semiscale pump.
- c. These data are more consistent with the single-phase homologous data celected for the full scale pumps.
The Semiscale steam water data base is used for Curvs Types 7 and 8 Data are (positive flow negative speed) in the two-phase region. However, thest antered for these two curve types for. completeness. regions are not imoortant since it is unlikely that these conditions
-will be encountered in IMk LOCAs.- In fact, this consideration is the principal reason why these regions were gscerally excluded from the f
CE-EPRI 1/5 Scale Test Matrix (Reference V.13-1). Reference (V.13-1) W. C. Kennedy et al. , ** Pump Two-Phase Performance Program," Volume I through 8. EPRI NP-1556. Electric Power Research Institute, September 1980. O -63 I
-1
]
TABLE V.13
,em Data Sources For Reactor Coolant Pump Models
(
\_ -
4 5 6 7(*) .8(a) ) 1 2 3 Curve Number MVD MAT MVT, MAR "MVR MAN ,MVW MAD Head Curvas BVD BAT BVT' BAR BVR_ __. Torque Curves BAN BVN BAD
'l Maine Yankee MY MY MY(b) gy(b) MY MY Single-Phase NY MY 33 EP(b) 33 EP EP EP EP EPID)
Two-Phase ! o Yankee Rowe EP(c) gy gy YR YR(c) YRCC) EP(c) Single-Phase YR EP(c) 33 33 EP EP EP(c) EP(c) EP(c)
.Two-Phase Vemont Yankee EP EP MY MY VY EP EP Single-Phase VY EP EP SS SS EP EP EP EP Two-Phase Rey: NY = Maine Yankee Pump Data j EP = CE-EPRI Pump Data '
YR = Yankee Rowe Pump Data YY = Vermont Yankee Pump Data SS = Semiscale Wo,es: ot (a) These regions are very unlikely to occur in LWR LOCAs. (b) Maine Yankee has anti-reverse rotation devices; therefore, reverse speed (this region) will not occur. y (c) Lenkee has ronreturn valves in each cold leg; therefore, reverse flow > (this region) will not occur. l l D d l t o -- - __ -- - - - - -
)
- 0.v.14 -
p/ '\s - Clarify what range of conditions have been used to test the pump model. , s
- m. .
s A.V.14 ,, , RELAP5YA calculations with pump models have been compared to data from three tests that cover a wide range of conditions. These comparisons j are described below:
- a. Yankee Pump Coastdown - Single-Phase Conditions The initial startup test program at the Yankee plant included a test where two pumps were tripped and the remaining two were left running (Reference V.14-1). This caused a normal reactor scram on low pump current, and a rapid flow coastdown in Loop 1 and Loop 4 that contained the canned rotor pumps that were tripped. The primary system pressure was relatively constant at about 2100 psia and maintained a large liquid subcooling of about 120 F in the pumps during this transient. Since each loop contains nonreturn valves, there was no significant backflow in these two loops from continued operation of the remaining two pumps. Plant test records contain the square root of the pressura difference between the inlet and outlet plena, normalized by the initial steady-state value. This war taken as an indication of the " normalized flow" in each trirped loop.
These two curves are shown on Figure V.14-1 which is attached.
~
The Yankee SBLOCA model (Rtference V.%4-2) was modified to simulate this plant transient. The square root of the pressure difference across the steam generator plena, normalized by the initial value, was monitored by control variables. These parameters yielded similar results for both tripped loops, and are also shown on Figure V.14-1. The comparison shows that the code calculated the coastdown characteristics well. The short coastdown time results from the relatively small inertia of the canned rotor pumps in the Yankee plant. This comparison f-t
\
provides etnfidenes in tha homologous mthod fer predicting single-phase, transient pump performance.
//
l
b. TLTA Pump Coactdown - Sinr.le-phase Conditions 76-TLTA Test' 6425/R2 was a large break blowdo b test with ECC injection perfomed in the Two-Loop Test Apparatus b'y General Electric. This was one of several TLTA tests selected for RELAp5YA code assessment (see Section 5.2 of Reference V.14-3).
A review of reported test data indicates that the centrifugal pump in the broken loop was tripped and isolated by two quick closing valves very rapidly to prevent potential damage to that. < pump. However, the coastdown flow rate for the intact loop centrifugk1 pump, tripped at time zero, was measured and reported in Figure J-74 of Reference V.14-4. This test data is reproduced in Figure V.14-2 and compared to the transient mass i flow rate predicted by the RELAp5YA model of TLTA. This comparison shows that the RELAp5YA model of the TLTA centrifugal pump simulated the coastdown flow rate reasonably well. The corresponding decrease in pump head is shown in Figure V.14-3.
*No test data were reported for comparing this parameter to, but the calculation shows a well behaved coastdown characteristic.
During this 18-second coastdown, the calculated results indicate the pump outlet pressure decreased from 1234 to 1060 psia, the liquid subcooling in the pump decreased from 26 F to 17 F, and the fluid remained as single-phase liquid.
- e. LOFT Pantoperation ene" Coast (cwt;_- Single- and No-Phsse Cond i t ic,n_s, LOFT Tekt L3-6/LG-1 was a PWR raall heeak test that cambined a long period of pump operation under single- and two-phase conditions, a subsequent trip of both pumps at 2371 seconds, and l continued testing until 7469 seconds. This test has been simulated by the RELAPSYA model of the LOFT facility from 0 to 2460 seconds for code assessment purposes. The pump models used LOFT test data and a fixed inertia rather than the variable b
t
p 31 f.. inertia that exists.in these components.=Many calculated parameters are compared to t'est. data in Section 5.3 of Reference LW
.Y.14-3. -Figures V.14-4 through V.14-6, attached, show additional comparisons of predicted pump performance p.arameters to available. test data, j Figure V.14-4 compares the calculated pump head to the differential. pressure across the two pumps (measurement PdE-pC-001). Both curves show essentially the same head degradation as the fluid changes from initially subcooled liquid?
to predominantly steam at about 1000 seconds and beyond. .This
~
comparison provides confidence in the homologous method for f I I modeling pump two-Ph ase perfo:sance. Figure .V.14-5 compares the calculat'ed and measured pump speeds. Both curves.show essentially a constant speed during the period when electrical power was supplied to the pumps. This was followed by a rapid coastdown when the pumps were tripped at-about 2370 seconds. The spikes in the test' data from 0 to 300 e seconds are attributed to. unexplained noise *'. by the EG&G test analysts. A slightly lower initial speed (3082 rpm versus the test value of 3200 rpm) was used to match'the flow rates and pressure drops around the primary system. Figure V.14-6 compares the calculated and measured pump speeds durint, the coastdown period from 2370 seconds to 2450 seconds. The two t.urves show similar coastdown trends, but with two nr. table dif ferences. First, the two curvas are oUset by akut. 4.4 seconds due to the lower in'itial speed-and the' plotting edit J i frequency for the calculation versus the measured speed i' history. Second, the slope of the calculated sepeed history is more negative than the test data for speeds above 1000 rpm, about the same for speeds between 700 and 1000 rpm, and less i tiegative for speeds below 700 rpm. These differences are primarily due to the fixed value for the pump moment of inertia used in the. calculation versus the variable moment of inertia in the LOFT pumps. Table V.14-1 compares the approximate LOFT O %
inertia values to the value of 274 lbg -ft used for our simulation. The slope of the pump speed equals the shaft torque
' divided by the moment of inertia for a free-wheeling pump: .
dw/dt.= - /I. ;. . 4 For speeds above 1000 rpm, our inertia is too small leading to a more rapid coastdown than measured. -For speeds below 700 rpm, our inertia is to large leading to a slower coastdown than measured. However, the variable inertia of the LOFT pumps is atypical. Reactor coolant pumps at power plants have fixed' inertias. TABLE V.14-1 LOFT 2 RELAPSYg) (rpm) (1bf - ft ) (1b f - ft
-273 274 3200 289 274 3000 318 274 2500
( 2000 325 274 311 '274 I 1500 274 274 L 1000 216 or 34* 274 500 154 or 34 274 100 136 or 34 274 0
*The test report does not make clear whether the variable inertia flywheel was engaged or disengaged below 750 rpm. If disengas,ad, then the lower inertis value applies.,
References-(V.14-1) . "The Startup Experiment Program for the Yankee Reactor," YAEC-184, Yankee Atomic Electric Company, June 1961. (V.14-2) J. W. Loomis J. H. Phillips, A. Husain, " Reactor Coolant pump , Operation During Small Break LOCA Transients at the Yankee Nuclear hwer Station," YAEC-1437, Yankee Atomic Electric Company, July 1984. lO . l i
\
L____________ _
1 i. l
. (V.14-3)' Fernandes, R. T. , R. K. ' Sundaram, J. Chaus, A. Husain, J. W. Loomis.
~ .
Lo Schor. R. C. Harvey and R. Habert, "RELAP5YA - A Computer Program
/'N for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume Code Assessment," Yankee Atomic Electric Company Report
)
III: ~ YAEC-1300P, Vplume III, October 1982. (Proprietary) (V.14-4) L. S. Lee, C. L. Sozzi, S.'A. Allison, "BWR.Large Break Simulation g
- Test's - BWR Blowdown / Emergency Core Cooling Program," Volume 2 NUREC/CR-2229 EPRI NP-1783, CEAP-24962-2, General Electric Company, July 1982.
p . 0
.U l
1.0 I > - y - y - , ,
~ * ;. Pump Trip ,
0.8 - g :: , - og *
.s 0,3 - - .
m D
$ 0.6 .
'G6 ,,
c 1 o. 8 e O.5 @ g 0.4 - AA G n.8 A 4 0.3 - 9A - 4- I A Loop 1 Data g O Loop 4 Data A ~ 0.1 - RELAP5YA A ,
) A O.0 i i ' ' D A 2 3 4 5 6 7 8 Time (Seconds) 1.0 i s i e
- 0.9 -
Pump Trip - 0.8 - 0.7 _ i
, 0.6 - -
% RELAP5YA M" I w 0.5 !
g _
-i
, 0. 4' - , i 2*. !
y 0.3 - i N y 0.2 -
. J E
o E 0.1 - - 0.0 1 I I e a 2 3 4 5 6 7 6 Time (Seconds) ' Figure V.14-1: RELAP5YA Comparison to Yankee Rowe j O* Pump Trip Test Data i
=__________-_____________ _ _ _ _ _ _ 'i
i.
't
... ',G 10 TLTA Test 6425/R2 i , l' .
\
I' -' '"****S, 5 8 \' i
.i N Yr%
i v
.5 6 18 a s.. . - . . . . . . . . - . +- +
.a. T 1 i 4
. Jg V ,
sw 'N % j. 2 E R5YA A E 0 p. g i ! 1
!t
-2 .
0 5 to 15 TIME (seconds) FIGURE V.14-2: INTACT LOOP RECIRCULATION PUMP MASS FLOWRATE o 200 a s I l TLTA Test 6425/R2 > 160 J-t w a 120
.c O :
at . sa
- I 3: l* * * * * * * * * ** ** * * * * * * * * * * ****
a 60 E S R5YA I 40 . . . . . .. ... .... . ... ... .....p..... . i i 0 - 0 5 to 15 TIME (seconds) FIGURE V.14-3: INTACT LOOP RECIRCULATION PUMP HEAD
- - - . . - - - - - , , . --r-,- o t
- ' %);
~ _
' ' i , , , ' '
O
^
$3 -l-L " D
- Ct:b o
D 13 n n 9 E
-f - ;
a o-t. OM
" b la o
3iu
,( .g E; 8 '
- h!-
3% s r
- ou i:
4 -R -b5 8<t U No
- h
. 2 a, I$
an$ . O Y
-N 5 o b
=
m
-i .. f
_g m h o
=
- .. . I O ,cf. og.
OEI OU34dWd (ISd) d3d SS0809 380SS38d TID 3833JIO O {
ij 1 y! , 0, O
. . 0
, - - . 0 0
2 _ t _ 0
.. .63 .
0
- 0 0
2
. 3
- a 0 t 0 a u 0 D -
0
. 2 6
- - m 3 -
0 L -- 0 0 T
- 8 F -
. 1 O L
0 o t
- 0 d 0 ne t
8 1 oe sp f6 iS
- Y3 r 5L 0 ap P pm AT Cmu LF EO .M )ECo P 1
- RL S m
. ( Ar
- Y e 0.
0MP E5T
- e 0IA g O
2
. _ 1TL E.. o RL 0
- 0 5 0 -
, 0 4
. , 1 1
- V 0 e
,0 r J u
. F g
- i F
- 0
,00
_ 5
- 0
_ : 0 _ ,0
. 4 n
[: i 1
'0 0
0
. J 0
0
,m . - . ~
ng O
- g- e o aic O o =) -
8 .a a.3>QcQ. t rO5 a eW E Q
,.U a I' Ivfl! ,
1
)
I'~) . . . .
~J V- , .
l l 9 11 ~$
" . n I n o ]
~
g -h a I 1 11 e i o " i o
. il e I
". $3am ,
- N oE
. a,
~
5$
=$g*&g
~
y it a a3
<E 9 $a I c. '
da e n. D i j $
// . 3 e
,/'
.O
~
tW li o j l1 a 11 0*0b9t 0*0but 0*0bst 0*0 BOT 0*d5 0*O 0*005C 0*0bir DEI 13AdWJ (W&l) 033dS dWfkl O ~74 I i 1
O.V.15 .
.p Justify which set of pump curves will be used for SBLOCA applications:
k Bingham, Westinghouse, LOFT, Semiscale or other. A.v.is The data bases that were used to develop performance curves for the Maine Yankee. Yankee and Vermont Yankee reactor coolant pumps are Additional pump'modeling described and justified in Answer A.V.13. information is provided in answers A.I.7, A.V.12 and A.V.14. The most preferred source of information is full-scale pump The second source is the CE-EPRI manufacturer's data where available. pump data base due to its completeness (nearly 1000 steady-state tests) and similitude parameters. The Semiscale pump data base is only used to define the two-phase difference tables for curve Types 7 and 8; however, operation in these regions is very unlikely. 'Wone of the Bingham Westinghouse nor LOFT data bases are ured.
-p
.J L
/"
f
VI. FLOW REGIMES p fd- .Page 8 of Reference 9 states that "the horizontal flow regime map is ,, used rather than the vertical flow regime map when.the. angle ... is,less than 15 degrees." .. 0.VI.4
. Clarify the basis for the 15-degree switching criteria.
A.VI.4 The criterion for switching from the vertical to the horizontal flow map in RELAP5YA is when the absolute value of the angle from the horizontal is less than 15 . The main impact of switching maps is to allow for the occurrence of stratified flow. Recent work by Bornea et , al. (Reference VI.4-1) has shown that stratified flow will not occur for angles greater than about 20 from the horizontal. Hence, the choice of 15 in RELAP5YA appears reasonable. Reference (VI.4-1) Bornea. D., et al., "Cas-Liquid Flow in Inclined Tubes: Flow Pattern Transition for Upward Flow " Chemical Engineering Sciences, Voluwe 40, (1981), 0.VI .3
- l, Clarify if the stratified ficw model is also used up to angles of 15 degrees.
A.VI.5 The stratified flow model is used up to angles of 15 from the horizontal if stratified flow is calculated to exist. 5 _ _ _ _ _ _ _ _ _ _ . _ _ _ - . I
Q.V7 6 j h
\- Clarify if there are any SBLOCA analyses that will use the stratified f
)
flow model for situations where the pipe area is changing and, if so, l l l how the model dould have to be changed to handle the nonunifom area. . A.VI.6 l All the active hydrodynamic volume components in RELAPSYA (SNGLVOL, BRANCH, PIPE, etc.) are assigned a constant flow area over the length of the volume. In normal plant applications, horizontal piping runs with constant flow area are usually long enough that they are treated as separate volumes. If it becomes necessary to combine pipes of different flow areas into one volume, the usual practice 'is to preserve the combined volume and the combined length and calculate an equivalent flow area. Horizontal pipes with nonunifom flow area are not expected, but if they exist, they will also be modeled with a constant equivalent flow area. The stratified flow model utilizes f,his constant flow area. O LVI.7 l Clarify how the slip is treated for the two-phase critical flow when the upstream volume switches from nonstratified to stratified flow and what offeet this has on the break flow. LvI.7 _ in RELAP5YA, the determination of the slip between phases is primarily influenced by the interphase drag tem, F7 defined in Section 3.1 of Volume I of the RELAP5YA Code Manual. This is a junction-related parameter and its calculation depends on the junction void fraction. In horizontal coeurrent flow, when the flow is not stratified, the This junction void fraction is donored from the upstream volume. When ; applies whether the flow at the junction is choked or not choked. the flow in the upstream volume is stratified, the junction void fraction is not donored directly from the upstream volume, but is p modified to account for the stratified liquid level in the upstream t i
- - - _ . _ _ _ - . - _ _ . . . . . .I
O 'I g volume being above or below the junction location and for the mechanism of vapor pull-through or-liquid entrainment. The method for estimating b this junction void fraction is described.in-Section 2.2.1.3 of-Appendix A of Volume I of the RELAPSYA Code. Manual. The junction void fraction, along stith the phasic velocities, is also used to ompUte the junction flow. 0.VI.10 Clarify how the stopping of natural circulation and reestablishment of natural circulation has been verified analytically and experimentally. A.VI.10 please see the response to questions Q.III.1 through Q.III.4. 0.VI.11 Clarify how the effect of noncondensable gases on natural circulation
. N has been verified. ,
- b[
A.VI.11 please see the response to questions Q.III.1 through Q.III.4. O.VI.13 Submittals need to consider the degree of mixing in the vessel when the # loops each have different conditions such as break in one koop, a pressurizer voided in one loop, steam generatoe heat removal in only one loop, and flow stopped or auch less in osa loop than the other(s). If complete vessel mixing is assumed, it needs to be justified if different conditions exist in the loops. Clarify and justify how vessel mixing will be modeled for SBLOCAs when the loop conditions differ. O l
-)
4 5's A.VI.13 ' jy. l U Figures IV.2-1 and IV.2-2 represent the nodalization used in the punp trip' studies performed for the Maine Yankee and Yankee Plant at Rowe, respectively. Similar noding schemes will be used for small break calculations. The recirculation loops will-be modeled as two loops. One will represent the broken loop, the other will represent the intact-loops. Several nodes are used to model the recirculation loops accurately. The downcomer and other regions of the reactor vessel are' modeled with a number of axially stacked volumes. However, the vessel nodalization does not specifically model azimuthal variations in these areas to account for differences between the recirculation loops. This nodalization is considered adequate for SBLOCAs for the following reasons:
- 1) Only uniformly applied operator actions are used in licensing analyses. For example, reactor coolant pumps are either all running or all tripped; steam generators are either all bottled up or all available for coolinb. Even the ECCS is available in all
,,_ Therefore, the various loops are not expected recirculation loops. 'g to behave significantly different from each other.
- 2) The fluid conditions at the inlet to the core is expected to be very nearly uniform even if different loop conditions are introCuced in the upper downcomer regions. This'is because:
a) The fluid experiences v change it, direction at the injection nozzles. The exact fluid tra,4ectory in the downcomer depends on the loop flow differences and the resistance to flow in the vertical or ennular direction. ThEannalarregionisquitefree and a uniform flow is 9xpected to exit the downcomer; b) The downcomer exit flow is going to experience a change in direction f with a higher probability of mixing. The various structures present in the lower plenum area (such as a flow skirt in Maine Yankee) and the core support plates with substantial reduction in 4
)
j flow area are expected to enhance mixing; c) PWRs use an open core configuration. This geometry promotes fluid mixing in the core. No major liquid level differences are expected to exist at various 1 O azinuthal locations in the core.
- 3) A voided pressurizer does not pley a significant role for those During the phases of a LOCA where core cooling is a concern.
[
- ' pressurizer refill period, the pressurizer loop may behave differently than the other loops. But its impact on LOCA calculations.will be negligible as the core-is. expected to be fully covered during this period.
0.VI.17 J
'Page 9 of Reference 8 states: "The two-fluid nonequilibrium hydrodynamic model includes three options for simpler hydrodynamic models. These are a quasi-steady flow model, a homogeneous flow model, and a therinal equilibrium model. The two-fluid, quasi-steady, or homogeneous flow models can be used with either the nonequilibrium or Most of the equilibrium thermal models (that is, six combinations).
development effort was performed using the two-fluid nonequilibrium model". Clarify if anything but the two-fluid nonequilibrium model will be used in any part of the system for SBLOCA calculations. A Q .A.VI.17 The nonequilibrium option is generally used for volumes in our current The exceptions Yankee, Maine Yankee and Vermont Yankee plant models, to this general approach occur for several Time Dependent. Volumes Of these (TMDPVOLs) that represent boundary conditions in each model. THlBPVOLs, only the feedwater and ECCS supply volumes act as fluid The remaining TMDpVOLs l sources which deliver liquid to the systems. act as fluid sinks for which only the pressure is an important parar eter. Therefore, the cases where'the equilibrium option, rather than the nonequilibrium option, is used to represent certain boundary conditions makes no (if ference. The two-velocity ("two-fluid") option is also generally used in these plant models. The Maine Yankee input deck contains no one-velocity junctions. The. Yankee input deck contains three one-velocity junctions. These represent the connection for the flow path from the downcomer to the tube bundle region at the bottom of the steam s
\
generator in the single broken loop, the single intact loop, and the l
1 combined two intact loops. This region generally.contains single-phase The Vermont liquid;thus[theone-velocityoptionisappropriate. h . Yankee input deck contains two one-velocity junctions. 'The first represents the very small leakage path from the lower plenum to the control' rod guide tube region at the 42-inch elevat; ion in th[ vessel. ' The second represents the very small leakage path from the isolated This latter portion to the active segment of the recirculation header. junction and the nearly isolated portion of the recirculation header will be removed when the plant completes the pipe replacement program since it will no longer exist. l 0.VI.18 Clarify if there is some type of boiling droplet breakup in addition to the Weber number criteria that models mechanical breakup.
-A.VI.18_
REL.AP5YA assumes that droplets are present in the annular-mist flow regime (voids between 0.85 and 0.95) and the mist flow regime (voids v above 0.95). The average droplet size is used in estimating the In the annular-mist regime, the interphase drag in these regimes. avecage droplet size is postulated to be at its maximum and is In mist flow, the characterized by a critical Weber number criterion. a,erage droplet size is assumed to be below itt maximum value and the model for droplet size calculation 10 described in Section 3.1.3.4 of The maximum drop 1rt size, defined by a critical Reference V1.18-1. Weber criterion, is assumed to occur at a void fraction of 0.95, and corresponds to the transition point between the annual-mist and mist flow regir.cs. Droplet breakup due to boiling is not explicitly modeled, but the decrease in droplet size at void fraction above 0.95 indirectly accounts for mechanisms other than mechanical breakup. Reference Fernandez, R. T. , et al. , "RELAP5YA - A Computer program for (VI.18-1) 1," Light-Water Reactor System Thermal-Hydraulic Analysis, Volume YAEC-1300P, October 1982. ( 1
l I 0.VI.19 1?
- q. 4 s
page 33 of Reference 8 states: "The phase-wetted perimeter has been j
\ _s/
assumed proportional to void fraction," Clarify if this is done.even i for the annular flow regime and if so justify Whatjapproximations are l 2 introduced for SBLOCA calculations, j
]
A.VI.19 The wall drag calculation.in RELAp5YA is directed primarily at calculating the overall two-phase pressure drop correctly. The partitioning of the total wall drag into wall-vapor and wall-liquid components is done somewhat simplistiemily by using the void fraction to determine the phase-wetted perimeter. This approach ensures a smooth transition between single phase and two-phase conditions, but does not explicitly account for the effect of flow regimes on the partitioning of the wall drag. Thus in the annular-flow regime, even though the wall is supposed to be completely wetted by the liquid phase, all the wall drag is not apportioned to the liquid phase. ( A However, the total pressure loss due to wall drag is preserved due to the phasic sum momentum equation. The wall drag has an impact et the phasic momentum equations in RELAp5YA Which are solved for the phasic velocities. Also, the wall drag generally has a larger impact on the phase velocities at conditions of high mass flux. For pWR SBLOCA conditions, the annular flow regime generally occurs as a transition regime during the formation of two-phase levels in the system. In these situations, the annular flow regime persists only for,a short time interval and the mass fluxes are relatively low. Hence, the partitioning of the wall drag in annular flow is expected to have negligible impact on the calculation of phasic velocities for SBLOCA calculations. 0.VI.20 Equation 128 on page 35 of Reference 8 defines the HTFS two-phase friction multiplier correlation. Clarify what value of C. the correlation parameter, is used. G)'
J'
]
i: I L A.VI.20 L( j b- The correlation parameter, C, is based on the HTFS correlation
~
~
(Reference VI.20-1) for two-phase pressure drop and is' calculated as:
.. j j
(los 3p(f,) + 2.5)2 1 l 2 s C = -2 + (28.0 - 0.5 KC)exp [ C x 10 , - 2.4 i where: fp = Baroczy property index
. ( )0.2 C = mass flux
= k gfg Vg+ f Vf Reference (VI.20-1) Claxton, K. T., et al., "HTFS Correlation for Two-Phase Pressure Drop and Void Fraction in Tubes," AERE-R7162,1972.
0.V1.21 f-Page 35 of Reference 8 also states: "For countercurrent.two-phase flow, the'HTFS two-phase friction correlation is invalid. However, for countercurrent two-phase flow, the phases are approximately separated and the friction factors, h gand A , are used directly with reasonable accuracy." Clarify if any problems occur in using this approach because of the discontinuities in models that' occur in switching from countercurrent to coeurrent flow. A.VI.21 There will be some discontinuity in the vesiculation of the phasic wall drag components when the code shows a transition from countercurrent to coeurrent flow. However, this transition generally occurs at low flow O i-
-_--__._______L__._________
conditions. At these conditions, the calculation of phasic momentum is ]
~
dominated by gravity and the interphase drag component, which is smoothly behaved, and the wall dras component has minimal impact. j Hence, these discontinuities in wall drag are expected to have . .. ' negligible impact in the estimation of the overall drag on each phase. . < 0.VI.25
-Page 18 cf h ierence 13 states: . . .the assumption of zero mass transfer is que.stionable for large area changes. Large area changes give relatively large pressure changes and hence mass transfer due to P that could change the void redistribution as the fluid passes through the area change. Clarify what limitations will be placed on the use of the abrupt area change model for SBLOCAs to justify the zero mass v.ransfer assumption.
A.VI.25 No formal limitations have been placed on the use of the abrupt area change model. The partisi quotation, taken ou,t of context from the Q cited reference, and the question imply several misconceptions concerning the abrupt area change model. The following explanation is intended to clarify the role of the abrupt area change model and situations where we use it. The pressure change between two adjacent volumes connected by an unchoked junction is determined by the solution to the full set of conservation equations, not simply the area change between volumes. l Examination of Equations 231 and 232 in Appendix A of Reference VI.25-1 shows that this pressure change, (Pg
- PL b, indudes ne following phenomena:
- a. Temporal and spatial phasic acceleration.
- b. Phasic gravitational forces.
- c. Phasic wall friction forces.
i I
c--_-___ __ _ __- f
- d. Interphase drag forces.
- j. . ~,_
l
# e. Virtual mass effects. ,
2:.
- f. Interphase inass/ momentum transfer. .
Section 2.1.3.4 cf Reference VI.28-1 states that RELtp5YA and RELAp5 MODI contain models to account for local irreversible pressure losses l from abrupt area changes in the geometry of flow paths. These models provide additional terms to the two womenta equations, as discussed in' Section 2.1.3.4, even though they are not shown explicitly in the two equations cited above. These models are pd a replacement for the two momenta equhtions. The abrupt area change model provides one of two The options to account for the additional local pressure losses. RELAp5 MOD 1 authors have attempted to formulate this model in a manner that accounts for downstream void-fraction redistribution and that is consistent with the basic conservation equations (Dr. Trapp states this on page 17 of the cited reference). The authors have assumed that mass transfer is negligible in the localized region of the abrupt area
.G change. However, mass transfer is accounted for in momentum centrol Q volume as shown in Equations 231 and 232 cited above. Finally, large area changes tend to increase the loss coefficients, but do not necessarily lead to "relatively large pressure changes." The net pressure change that affects mass transfer is the sum of many effects identified above. The local pressure loss tems depend not only on the loss coefficients, but also on the magnitude of the phasic dynamic pressure. In SBLOCAs, the dynamic pressures tend toward small values as the velocities subside and the system becomes gravity dominated.
We generally do not use the abrupt area change model in plant-specific input decks, except for certain cases. Instead, we generally use the smooth area change option and enter forward and reverse loss coef ficients that are based upon standard references or plant data. The exceptions occur when we use the motor valve component available in RELAp5YA. The current SBLOCA model for the Yankee plant at Rowe uses motor valve components to represent the noneeturn valves in the primary l side cold less and the secondary side steam lines. The current SBLOCA t p I
p-aodel for the Maine Yankee plant does not use any motor valve components. :The current LOCA model for the Vermont Yankee plant uses
-[l -
motor valve components to represent the following
\s /
- a. Pump discharg,e valves in each recirculation loop.
- b. Main steam line isolation valves.
- c. Pressure regulator valve upstream of the turbine.
- d. Safety / relief valves on the main steam lines.
- e. Safety valves on the main steam lines.
0.VI.26 Clarify how the abrupt area chango model has been assessed against data for conditions expected in SBLOCA calculations. ( A.VI.26 , The abrupt area change model has only been used in ' conjunction with the motor valve component in current plant LOCA models as discussed in the answer to Question Q.VI.25. The main effects to be simuisted are the opening and/or closing time of certain valves. This can and has been assessed by verifying that the valves open, close or remain in position according to the time of their trip signals and corresponding vcive operating rates. O.VI.27 Clarify how L1 and L2 are defined for the abrupt area change model on Page 58 of Reference 8. A.VI.27 l The upstream length, L1, and downstream length, L2, used in the abrupt ig area change model are defined as follows:
]
i
l, i. s tuvansion , ' l' , 1 L1 = 0.0-N - ar , L2=10.0DIAMV(L)[ARAT(L)' Contraction
~
L1 = 1.0 DIAMV(K) ARAT(K) L2=10.0DIAMV(L)/ARAT(L) Where: K = upstream volume number. L =- downstraam volume number. DIAMV(-) = hydraulic diameter of designated volume ARAT( ) = partitioned area ratio in the designated volume associated with the abrupt area change junction.
- y' -
( Note'that we do not generally use the abrupt area change option as discussed in the answer to Question Q.VI.25. 0.VI.28 Page 103 of Reference 8 describes modeling valves using flow coef ficients rather than the abrupt area change model and converting the flow coefficients to energy loss coefficients. Clarify.how the flow coefficients are defined in terms of energy loss coefficients. A.VI.28 Information contained in Appendix A. Page 103 and Appendix B. Pages 111 to 112 of Reference VI.28-1, and Page 63 of Reference VI.28-2 indicates that tabular values of flow coefficients (C,) can be entered as a user option to account for local form losses (K) through setor and i
-__i.__.l__...___.,
l' n servo valves. The latter two references also define the method (j presumably used to convert the flow coefficients to local energy losses j (K) within the computer code. This infot1 nation is not correct. > l RELAPS HODI and RELAPSYA contain two options for mo'deling local pressure losses across these two types'of valves. The first option is to specify that the abrupt area change model is to be used for the valve. The code will then internally compute the loss coefficients f based upon the current time geometry. We generally use this option for these two types of valves. The second option is to specify that the smooth area change model is to be used for the valve. This option requires the user to enter a table of loss coefficients versus normalized valve stem position. Although the User's Hanuals indicate flow coefficient (C,) values are to be entered, the algorithm actually assumes that energy loss coefficients (K) were entered. Therefore, loss coefficients must be entered. References
./
(VI.28-1) R. T. Fernandez, R. K. Sundaram, J. Ghaus, A. Husain, J. N. Loomis, L. Schor, R. C. Harvey and R. Habert "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulie Analysis, Volume I: Code Description," Yankee Atomic Electric Company Report YAEC-1300P, Volume I (October 1982). (Proprietary) (VI.28-2) R. T. Fernandez, R. K. Sundaram, J. Chaus A. Husain, J. N. Loomis, L. Scher, R. C. Harvey and R. Habert "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume II: User's Hanual " Yankee Atomic Electric Company Report YAEC-1300P, Volume II (October 2982). (Proprietary) Q,.VI.29 Page 49 of Reference 10 states: "In addition, the Two-Dimensional Branching technique described in Section 3.2.7.2 of Appendix B is not recommended since it is known that this technique can create e nonphysical recirculation through the adjacent junctions." Clarify how
- i
? ! r 1 90 tees will be modeled and justify why nonphysical behavior will.
/
not result. { . m A.VI.29 .w
- j Ninety degree (90 ) tees are modeled using two rather'than the three junctions shown on Figure 4 in Section 3.2.7.2 of Appendix B (Reference VI.29-1). Either Junction J2 or J3 would not be used. This i technique prevents the occurrence of nonphysical recirculation between Volumes VI, V2 and V3 via Junctions J1,~J2 and J3. Appropriate loss coefficients for the branch and run junctions, based upon standard references, are calculated and entered.
References. (VI.29-1) E. T. Fernandez, R. K. Sundaram, J. Ghaus, A. Husain, J. N. Loomis, L. Schor, R. C. Harvey ar.d R. Habert, "RELAP5YA - A Computer Program for Light-Water Reactor System Themal-Hydraulic Analysis, Volume I: Code Description," Yankee Atomic Electric Company (Proprietary)
-' Report YAEC-1300P, Volume I (October 1982).
0.VI.30 Clarify how the 90 tee model has been assessed against data for single-phase and two-pha e flow conditions. A.VI.30 The 90 tee modeling technique descri ed in the answer to Question Q.VI.29 has been used extensively in both our Separate Effects and our System Effects code assessment cases. Table VI.30-1 summarizes the nodalization diagrams which show where this technique has been used. These tests include steady and transient single- and two-phase flow through tees. This modeling technique contributed toward the generally good assessment results that were obtained. O .
---------m-._ ___, .._
TABLE VI.30 ^
\' Test Facility and Wodalization Distrams With Tees -[
Figure "' Title gage pumber ,, CE Level Swell Test configuration........................ 16 2.1-7 17 2.1-8 RELAPSYA Model of CE Level Swell Test..................'.. Test Assembly Schematic for Steady-State Tests........... 66 2.3-3 67 2.3-4 Jet Pump Blowdewn Facility Isometric..................... Jet Pump Test Vessel Model for Steady-State Tests........ 68 2.3-5 76 2.3-13 Test Loop Isometric for Blowdown Tests................... 1.............. 77 2.3-14 Jet Pump Test Assembly Schematic for Test
,RELAP5YA Model for Jet Pump Blowdown - Test 1............ 78 2.3-15 82 2.3-21 Jet Pump Test Assembly Schematic for Test 2..............
RELAP5YA Model for Jet Pump Blowdown - Test 2............ 83 2.3-22 Schematic of Thermal-Hydraulic Test Facility (TNTF)...... 177' ("*g 5.1-1 U 5.1-3 RELAPSYA Model of THTF Test Section...~................... 179 Schematic of TLTA-SA Configuration....................... 234' 5.2-1 235 5.2-2 RELAPSYA Nodalization f or TLTA Test 64 25 /2. . . . . . . . . . . . . . . 267 5.2-34 Schematic of TLTA-SC Configuration....................... 268 5.2-35 RELAP5YA Nodalization for TLTA Test 6432/1............... 278 5.2-45 RELAPSYA Nodalization for TLTA Test 6441/6............... 5.3-1 LorT System Configuration.........~....................... 308 309 5.3-2 RELAP5YA LorT L3-6/L8-1 Nodalizat1on..................... i N/ t l i I L____________ __
i
~
s l l' 0.VI.31 k, Page 101 of Reference 8 discusses how gravity effects are modeled with , l
'a branch volume. Clarify how the gravity head is included. * . {
A.VI.31 The gravity head is included in the phasic sum momentum equation for all junctions. This appears in the second term on the right side of Equation (13) on Page 12 Equation (224) on Page 81 and Equation (231) on Page 83 of Appendix A in Reference VI.31-1. The gravity head term
~
is computed as variable DELPZ in Subroutine VEXPLT:
. g DELPZ = ( f)"g+(kf)"g]&2g+M f
dgSf ( + M fgf[LMZL where AZ g and4ZLare half the elevation change in the upstream and downstream volumes, respectively. This term is then added to the SUMOLD
,e variable. When horizontal stratified flow exists, then an additional term
( given by Equation (25) on Page 14 of Re(erence IV.31-1 is added to the right side of the phasic difference momentum equation. This term is then added to the DIFOLD variable in Subroutine VEXPLT. Reference (VI.31-1) Fernandez , R. T. , R. K. Sundaram, J. Chaus , A. Husain, J . N. Loomis, L. Scher, R. C. Harvey and R. Habert, "RELAP5YA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description," Yankee Atomic Electric Company Report YAEC-1300P, Volume I, October 1982. (Proprietary) 9.VI.32 Clarify if the statement is correct that "A gravity head is included in the branch volume V and V4 and results in a corresponding 3 separatic:. ef fect due to buoyancy" (Page 101 of Reference 8). 1
\
t
.,- 1 4
A.VI.32 , l (it _ I f&
, A.j .
'The cited statement is perhaps too cryptic. Gravity head terms are .
i generally accounted for in the momentum equations associated erith junctions as stet'ed in' Answer A.VI.31. Each junctTon has a momentum-cell that connects the adjoining halves of adjacent volumes. Een anL 1 adjacent volume'is vertical (for example V2 and V4 in Figure 33.of the cited reference), then its nonzero elevation change contributes a T gravity tem to the phasic sum momentum equation. ' When.an adjacent volume is horizontal.(for example V1 and V3 in Figure 33), then its zero_ elevation change will not contribute a gravity term to the phasic j sum momentum ocuation. However, when stratified flow. exists.in a j horizontal volume, then a buoyancy term is added to the difference momentum equation.. Therefore, volume V4 will always have a gravity head. term. Volume V3 will have a buoyancy term when stratified flow exists within it. 0.VI.33 page 19 of Reference 13 states: "The area change seen at a branch junction gives a crude representation of the expansion or contraction losses seen by each junction flow. But it should be noted that the partitioning used to determine the area changes is algebraic in character and not based upon'any physical principals. In fact, the partitioning fails to preserve the mass' flow of each junction stream because volumetric flow is preserved by the partitioning instead of mass flow." Clarify how much error this could introduce in SBLOCA calculations. A.VI.33 The hydrodynamics within a volume that contains two or more inlet and/or ' outlet junctions is a complicated process that generally involves local three-dimensional flow patterns. In reality, these local 3-D flow pattems detemine the expansion and contraction losses (and momentum mixing losses discussed in the anrwer to the next question). These fluid mechanical energy losses are converted to O !
In relatively small thermal energy sources by dissipation. A one-dimensional codes such as RELAPSYA, these losses are approximated - by the product of form loss coefficients (from standard references) and Although this approximation.may.not the phasic kinetic energy terms. be exact, it appears to be reasonable. based upon the generally good These code assessment results contained in Reference VI.33-1. assessment cases cover a very broad range of themal-hydraulic conditions. In particular, they include a PWR small break test (LOFT Therefore, we L3-6/8-1) and a BWR small break test (TLTA 6432/1). conclude that the error introduced by this approximation is small and acceptable. The mass We disagree with the last sentence in the cited statement. flow r&te for all junctions is preserved by inclusion and solution of thi conservation of mass equations within the limits of the code. Reference T., R. K. Sundaram, J. Chaus, A. Husain, J. N. (VI.33-1) Fernandez, R. O) u Loomis, L. Schor, R. C. Harvey and R. Habert, "RELAP5YA - A Computer Program for Light-Wa6.er Reactor System Thermal-Hydraulic Code Assessment," Yankee Atomic Electric analysis Volume III: (Proprietary) Company Report YAEC-1300P, Volume III (October 1982). 0.VI.34
**Because each junction momentum page 19 of Reference 13 also states:
equation is calculated independently o,f the other junction flows connected to the branch volume (except for the reduced area) the Clarify how neglecting these momentum mixing tems are neglected." momentum mixing terms affects SBLOCA results, and if they have been included in RELAP5YA or just in the jet pump modeling. A.VI.34 The momentum mixing terms have only been included in the jet pump model. Our studies indicate they are significant primarily at or near I
i, :. f' . i b normal jet pump operating conditions. In this region, the very high. l
~
4 velocity drive nozzle stream mixes with the such lower velocity suction nozzle stream. The momentum mixing terms have not been included for. ,. other RELAPSYA hydrodynamic components. The generally goodLeode s , i assessment results obtained over a broad range of. thermal-hydraulic l conditions indicate these terms are generally not significant except for jet pumps. O l l 4 O '
-94
J VII. ' CORE STEAM COOLING p ( 0.VII.13 ,
~
For the three ,s,teady-state fild boiling tests, Page 167 of efecence 12 states: "RELAPSYA provides a reasonable prediction of the CHF location but overpredicts the wall temperature in the post-CHF regime." Justify this statement about the CHF location because Pages 180-181 of Reference 12 only show the wall temperature data above the CHF location so the CHF location is not given. A.VII.13 J The precise location of the CHF is not presented in the Experimental Data Report (Reference VII.13-1) for the THTF steady-state tests. The exact location of the dryout point cannot be determined and is, in fact, rarely clearly defined in a rod bundle. However, the data report presents the calculated quality at the CHF location and the quality at every axial location for which post-CHF temperature are presented. [ We inferred that the dryout point was located between the thermocouple
\./ :
at which film boiling was detected and the thermocouple below it. Table VII.13-1 provides the reported quality at CHF, the inferred CHF
-location, and RELAPSYA calculated CHF location.
TABLE VII.13-1 Inferred RELAP5YA Reported Quality ,
. Expt. CHF Loc.
Test No. at CHF CHR Loc. (ft) (ft) . i 3.07.9B 0.368 4.8 - 5.2 4.7 - 6.0 3.07.9K 0.887 9.3 - 9.5 7.5 - 8.3 3.07.9X 0.84 S.2 - 8.8 6.0 - 7.5 Reference (VII.13-1) Yoder, C. L., et al., "Dispers2d Flow Film Boiling in the Rod Bundle Geometry Steady-State Heat Transfer Data and Correlation Comparisons," ORNL/8822 (Preliminary Draft).
l 0.VII.14-O( *., . For the transient film boiling test, clarify why the RELAPSYA sheath temperature data at the elevations shown on pp. 183-184 of Ref,erence 12 underpredicts the temperature during the period insnediately after CRF occurs. There is some confusion in the curves on these two pages where the same curve changes from a dashed line on the lef t part of the graph to a solid on the right and the data and RELAp5YA curves apparently cross. A.VII.14 The underprediction'by RELAP5YA of the temperature in the TNTT Test 3.08.6C during the period immediately after CHF is due to the fact that CRF is predicted to occur by about 0.5 to 1 second later than the
- experimental data. The data curves and RELAp5YA curves are shown more clearly in the attached figures.
O , 1 l l i
i Im , , , , , , , , , , , 1000
. RELAPSYA =
., 1500 .
f1400 . I DATA . 1300 .
~
5 1200 . DATA E . G 1100 . g
.o e b 1000 . .#
" I y I * - [ RELAPSYA t .
- eco .
Too . j
.I soo . .
o.o t .o 4.0 s.o s.o to.o tr.o 14.0 is.o ts.o 20 012.o 24.o to.c es.o so.o T!nt IN stC Figure 5,1-9: FRS Inner Sheath Temperature at Level E(951n.)
\* . . . . .
14 o . DATA w . g 1300 . 8 RELAP5YA 5 troo .
~[ DATA stoo .
~
/ .
C 1000 . s / v <
> \, .
$ too . ,/
I , RE' LAP 5YA . g sw . ,
# l' .
Too . soo 14.o ts .o 20.0 o.o r.o 4.o s .o s.o to.o tr.o ts.o
/
Tint IM OCC k Figure 5.1-10: FRS Inner Sheath Temperature at Level F(1191n.)
\
1 1 1 I I
,I 1.
l ;I
<m
(
\, -
l 1500' , , , , , , i 1400 . g .tsoo .
\ .
5 ' oA73 . 5 ttoo . 4 RELAPSYA , Stoo . f )
- DATA-is 1000 .
,#-J .
I o a soo . o f . s f . g soo . M 1 l I
. . . . 1'
- s. .
00 t.o 440 s.o s.o so.o it.o 14.o f!MCts.o to.o 20.o tr.o 24.o ts.o es.o 30 0 IN stC i * ( Figure 5.1-11:. FRS Inner Sheath Temperature at Level G(142 3/4 in.) u 3**0 - W ' k .
.: 3 00 .
6 5 . DATA y 2.50 . l E 5 t*oo ~ RELAP5YA d S . 2 t.so . I
$ 100 . t l
t c . 3 .50 , 0 00 > o.o t.o 4.0 s.o s.o 10 0 12 0fine. 14 /s In is.o ste to.o 200220240 ts.o to.o 30 0 O Figure 5.1-12: Comparison of Measured And Calculated Outlet Volumetric Flov
l Clarify shat causes the shape of .the outlet fluid temperature curve f O.VII.15
%) . after 22 s as shown in Figure 5.1-13 of Reference 12. .
A.VII.15 ,, l
)
The increase in the outlet fluid temperature, which occurs in both the test and the RELAPSYA predictions, is due to the fact that the pump was .
/
tripped at 20 seconds in the transient. The inlet flow rate fell l i At 21 seconds, the bundle power quickly after the pump was shut off. was reduced from 7.8 MW to approximately 3.4 MW over a 4-second time This interval. The inlet flow rate dropped faster than the power. resulted in the power flow mismatch that produced a fluid temperature excursion. c.VII.16_ For the quasi-steady-state boil-off test, Figure 5.1-15 of Reference 12 shows the RELAP5YA void fraction is greater than the data above the "This figure 4-foot elevation. page 169 of Reference 12 states: indicates that the interphase drag values calculated by RELAPSYA may have been somewhat high beyond the 4-foot elevation fcr this assessment case." Clarify why this is true because it would seem that high drag , values would cause more liquid to be carried upward and therefore give lower rather than higher void fractions. A.VII.16 The RELAP5YA calculation for this test, started with a full bundle with constant inlet flow and bundle power. 'A very slow boil-off transient f l of 380 seconds was run until the outlet steam flow matched the bundle inlet flow. The high drag values at the upper elevations entrained I more liquid out of the bundle during the boil-off transient. The void profile presented in Figure 5.1-15 of Reference (VII.16-1) established at about 50 seconds in the transient. We can infer that RELAP5YK calculated higher drag values than existed in the test which caused a substantial part of the initial liquid inventory to leave the bundle. Hence, more of the bundle is dry and the es1culated void fractions are j higher.
i I: 1 i Reference
/~'s
' .h (VII.16-1) Fernandez, R. T. , R. K. Sundaram, J. Chaus, A. Husain, J. N. . Loomis L. Schor. R. C. Harvey and R. Habert, "RELAPSYA,- A ~ Computer Program for Light-Water Reactor System Thermal-Hydhaulic Analysis, Volume III: Code Assessment," Yankee Atomic Electric Company Report YAEC-1300P, Volume III, October 1982. (Proprietary)
. 1 0.VII.17 Page 169 of Reference 12 also states: "A dryout CHF is calculated to occur at a much lower elevation that corresponds to the predicted two-phase mixture level in the bundle." Clarify how this statement is justified by the wall temperature and void fraction figures given on Page 186 of Reference 12.
A.VII.17 l Figure VII.17-1 from Reference VII.17-1 indicates that the inferred ! mixture level for this test is at about ,. 8.6-foot elevation. The I h 'RELAPSYA calculated mixture level at a void fraction of 0.96 is 4.6 U feet as shown in Figure 5.1-15 of Reference VII.17-2. Figure 5.1-16 shows that RELAPSYA predicted a dryout CHF that corresponds to the predicted mixture level at 4.6 feet. The test report does not specifically give the CHF location. However, we assume the experimental CHF location occurred at the reported mixture level. Reference (VII.17-1) Anklam T. M., et al., " Experimental Investigation of Uncovered Bundle Heat Transfer and Two-Pha'se Mixture Level Swell Under High Pressure, Low Heat Flux Conditions" (Final Report for TNTF Tests 3.09.101-N and 3.09.10AA-FF-DRAFT), Oak Ridge National Laboratory. l September 1981. (VII.17-2) Fernandez, R. T. , R. K. Sundaram, J. Chaus. A. Husain, J. N. Loomis. L. Schor, R. C. Harvey and R. Habert "RELAPSYA - A ; Computer Program for Light-Water Reactor System Thermal-Hydraulic f Ar.nlysis Volume III: Code Assessment," Yankee Atomic Electric {s Company Report YAEC-1300P, Volume III, October 1982. (Proprietary) l i
-100-
b ,
- h, ,
i
*4
~ ..
m CJ e4 A 44 e o
, E 20 >e o o e -! c l e .. > e O 4 w
$jl ) k4 l
0 .w
' 'N $
l
- s "
g Il * .M O o M )kh b ""
-g >
I e e l e
<1 l. ' .
t N ,\ e l'
*\ *
'.y$
-eN g 5
e i
- oO" ]
<N .
3 p
~
4
- a"
' ', i \\\
a, -[ :$ 3 ". em I
/ -
V e) O e 1
- s. .
.g4< w q> a s
,M G. M
.g- '
' ,\ \s, , .
O I
'. * -t C
. .\.% O w
\
-S- >
, g ~
e j
'. \
, *y a w E
i x' .,,;u:
- ,,i. -e
~
E 83 e m mia: m m. b ... ..*.,..,
\.O h. % *
.g
*:) E Z;E c:ol.
u>
<. t i 2.c' .
c *..*., o W 2 o-c: W (.<l
>- oio; s.,N.,%.,;'\. % -3
.. Y.,,o N E l-e OOAx o'Pi i e n n ce *v ce ro 15 on c1 n to N O I.1 0 Y lli G I O A
- -101-
. ________.-_.-____-m._m-.__. . .m
0.VII.18
/x N/ I Pages 169-170 of Reference 12 state: "The above disparity between the-calculated and measured results might be attributed to the meihod'used I in modeling this' test. The RELAPSYA calculation resulted in rs very. 4 slow boil-off transient with constant inlet flow and bundle power. The j
thermal power transferred to the fluid did not match the bundle power until 380 seconds into the transient. When they matched, the results were accepted as the values to compare to the test data. By this point in time, a substantial part of the initial liquid inventory was calculated to have left the system. In these experiments, power was applied to the bundle and then was reduced to produce peak FRS temperatures of about 1400 F (maximum temperature imposed by safety limits). The differences between the power histories applied to the bundle in the test (which are not reported), compared with the fixed l power in the RELAPSYA simulation may account for the differences between the calculated and measured results." Clarify why this test was picked for assessment purposes if the power history was unknown. (w/ I A.VII,18 , This test was selected because it was thought to represent a condition that might be encountered during SBLOCAs. The test procedure indicates that the bundle power and inventory were adjusted over a period of time until the loop stabilized. Then a 20-second data scan was taken. We simulated this test using the reported data that were obtained "after the loop was stabilised." We do not know what effect the previous unrecorded history might have had on the RELAPSYA results. However, RELAPSYA has a tendency to underpredict the two-phase mixture level for low flow conditions. This results in a conservative PCT response for slow boil-off situations expected during SBLOCA.
-102-(
s 1 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ [
c---_ j [; '! l l l l
- 1. '
f' 0.VII.19 l For the reflood tests discussed on Pages 170-172 of Reference "" 12: l Clarify how quench level is measured. A.VII.19 The quench level is determined based on the response of the FRS sheath For the TNTF tests, the quench thermocouple (Reference VII.19-1). level is defined as the point during & reflood when precursory cooling stops and a precipitous drop in surface temperature begins ! (Figure VII.19-1). Reference (VII .19.-1) Hyman, C. R. , et al. , "0RNL Small Break LOCA Heat Transfer Test Series II: High Pressure Bundle Boil-Off and Reflood Test Analysis" (Final Report for TNTF Boil-off and Reflood Tests 3.09.100 DRAFT), Oak Ridge National Laboratory, September 1981.
-103-G
7___-_-_ ,. I r .. /m IL) .
.\
j
.> - . 3 o '- TE-315RF KELVIN .
f c o,os . o *c,ot o o,os o. u o,os . ii 8 si 'g ll 'e' o o o
- ~l n R Gc 4
. M 11 f n > f a n
- f O p .c
- O 11 c
u o cn Eu 8 O n o
. " -d '8 e8, M o em P gg b ll c
t.n " = .'O 3 u n R 8 c. ; O ! is - u o o o o !O o
).
i c g ;!! u L. \ W
~
8e o
&p1 y q= >* g5o x o
- -d yEa
- ?)
a- ?c. (E o E O vi ",& x o J L oc e o 85 *i U7 ac a o en l
- a. o 58 H ll W" a- >
I n e H n o u e o -s. ,y 11 he ll o l'l o
. . .7 .
0'008 o'005 ) cobst oobot odos o'oos 0 004 o cott J S338930 JWSIC o d
-104-(
-___2_____.________-_ _ _ _ _ _ .1
( O.VII.20 Clarify how "the measured void profile at time zero was used to calculate the , initial liquid and vapor velocities" (Page 17bof-Reference 12). A.VII.20 The dynamic quality at each nodal volume is calculated from an energy balance assuming thermal equilibrium. The phasic velocities and the static quality are detemined from these dynamic qualities and the To initialize the test, the measured void fraction at these locations. following calculations were performed: I A. . Dynamic quality in each node: Q '.z(i) h sub
. ah f h .
f Where: x(i) = " dynamic equilibrium" quality in each node z(i) = elevation of the node center from the bottom of the-heated length (ft) h fg = latent heat of vaporization ( ) hsb " I"I*t '"D**"II"E I 3' Q = liner heat generation rate (Btu /ft-sec)) m = mass flow rate (Ib/sec) B. The slip ratio in each node: x(i) .1 -0((i) If ggg) ,1 - x(i) o((i) fi
-105-
Were: c((i) = nodal void fraction (from void fraction distribution at time "zero" from Reference VII) % '. L. p . 3f = saturated fluid density (1b/ft 3) S, = saturated vapor density.(1b/ft 3) C. Vapor velocity in each node: V (1) = G x(i)/f c((i) . (3) we replace x(i) in (3) by x(i) = f 4(i) W gg)3 S
, g gy ,,(gg))
Were: S = V /V g (5)
-{)-
t. to get: , V(i)'=C/[fk(i)+f(1-o((i))/S] f D. Liquid velocity: V (i) = V (i)/S (6) f F. Static quality: x (i) = (7) 2 ' + k (1 *((i)) r, au) The values of V (i). V g(i) x,(1), calculated above are used for the initial conditions.
-106-
\
i f 1 0.VII.21 [] .. Clarify why the liquid mass inventory peak agrees almost exactly with 1 I the data at 17.5 s in Figure 5.1-18 of Reference 12 when the calculated l collapsed liquid level at 17.5 s in Figure 5.1-19 of Reference 12 is J significantly below the data. k 4 A.VII.21 The test data and calculation have been re-examined to address this question. According to Reference V11.21-1 .the local void profile in the heated bundle was detemined from the nine differential pressure transducers. This experimental void profile is compared to that These calculated by RELAP5YA in Figure 5.1-17 of Reference VII.21-2. The two void profiles agree reasonably well at 20 seconds. experimental void profile was then used to calculate the collapsed This calculation yields a value of 124.9 liquid level at 20 seconds. inches which compares well with the RELAPSYA value of 125.0 inches. However, it is substantially lower than the 134.3-inch collapsed liquid 1evel at 20 seconds in the test report. We have not been able to
/Q account for this discrepancy in the reported test data.
Finally, the test report states that the liquid mass in the heated bundle was esiculated by the following fomula: liquid " f flow CLL where: f = saturated liquid density _ Afgow = bundle flow area
= collapsed liquid level ICLL :
at Their use of the saturated liquid density (47.5 lbm/f t 20 seconds) does not account for the more dense subcooled liquid (50 to 55 lbm/ft at 20 seconds) in the bundle below the two-phase region. However, their co11 speed liquid levels appear to be too large. Nevertheless, the product yields a 11guld mass of 35.4 lbm versus 34.9 lbm calculated by RELAP5YA at 20 seconds. Ox
-107-
)
/' References k
(VII.21-1) Hyman, C. R. , et al. , "0RNL Small Break LOCA Heat Transfer Test Series II: High Pressure Bundle Boil-off and Reflood Test" Analysis"(finalReportforTHTFBoil-Offand'RefloodTests 3.09.100-X-DRAFT) Oak Ridt ;e National Laboratory, September 1981. l yernandez, R. T., R. K. Sundaram, J. Chaus, A. Husain, J. N. (VII.21-2) Loomis. L. Schor R. C. Harvey and R. Habert, "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume 1: Code Description," Yankee Atomic Electric Company Report YAEC-1300P, Volume I, October 1982. (Proprietary) 0.VII.22 Clarify why the peak in the liquid mass inventory is predicted almost i exactly h RELAPSYA for Test 3.09.100 and yet is significantly underpredicted for Test 3.09.10Q. A.VII.22 The saturated versus subcooled liquid densities are about 9.5% different for reflood test 3.09.100. The collapsed liquid levels versus those detemined from the void profile test data are about 5% to f 7.5% different. Both the liquid density and collapsed liquid level enters the determination of liquid inventory in the bundle. Therefore, the combined uncertainty in the reported inventory is about 111%. The calculated liquid inventory for both reflood teste are within 111% of the test data, and show the same distinct trends as the experimental data. O.VII.28 Page 142 of Reference 11 states that for medium flow ranges the maximum l of the high and low flow heat-transfer correlations is used. This
-108-i
I l applies for rubcooled or saturated transition boiling, rubcooled or [~'\' saturated film boiling, saturated film boiling or forced convection of t
' ') superheated vapor. Clarify why this procedure is conservative if the i
maximum of the two correlations is used. .;
" . ,i A.VII.28 The P.ELAP5YA heat transfer correlations are meant to be best-estinate correlations rather than provide conservatism. In licensing analyses.
YAEC will use these correlations along with EH assumptions to ensure that heat transfer from the rods will be calculated conservatively. The logic used in the RELAP5YA selection of heat transfer correlations essentially uses a set of "high-flow" correlations with a lower bound provided by a set of " low-flow" correlations. This le because the The high-flow correlations are not applicable at very low mass fluxes. selection of the maximum of the two sets of correlations ensures that the correlations are used in their applicable ranges of flow rates. 0.VII.29
/5 k
Clarify how the reflood model has been assessed for falling film from the top and for bottom quench. A.VII.29 l The reflood model has been assessed by modeling two separate effect tests for the bottom quench, and three integral tests which included both falling film and bottom quench. I The separate effects tests are the THTF tests 3.09.100 and 3.09.10Q. The integral tests are test TLTA 6425/2, TLTA 6441/6, and LorT L8-1. 0.VII.30 Clarify how closely the assessment conditions match the expected possible SBLOCA reactor conditions. O -109-
'i
'--_-_____m.___ _ _ _ _ _ _ _ _ _ _ _
W j
- i. 2 A.VII.30
[ - The reflood tests used for the assessment of the bottom quench were part of the second series of PWR Small Break Loss of Coolant Jccidents (SBLOCA) conducted at Oak Ridge National Laboratory in the
- Thermal-Hydraulic Test Facility (THTF). The objective of the reflood tests was to study bundle quenching behavior under conditions of system pressure, linear power and flooding rate expected in a SBLOCA.'
Five reflood tests were performed from Which two were selected for assessment. Table VII.30-1 summarizes the parameters for the reflood SBLOCA II THTF tests. Average inlet flooding velocities in SBLOCA 11 ranged from a low of 2.33 in/see to a high of 4.82 in/sec. Initial system pressure ranged from 563 psia to 1092 psia. Linear power ranged from 0.304 kW/ft to 0.659 kW/ft). The tests chosen (underlined in Table VII.30-1) were at low pressure and at various flooding velocity and liner power, conditions which are expected to occur during a SBLQCA in the plants modeled. ('"N All the integral tests used in the assessment of RELAP5YA used the quench model option at ECCS initiation. Since these were transient A tests, the conditions in the test facilities varied with time. detailed account of the integral test conditions are presented in Reference VII.30-1. TABLE VII.30-1_ Summary of Initial Conditions for Reflood Tests Dundle Mass Inlet Outlet Maximum FRS Linear Heat pressure __ _ Rate (LMR)- _ Flow Subcooling - S_uperheat Temperature _ Test No. (K (OF) (K (DF)] (MPa (kW/m (kg/s (K (OF))- (psia)1 (kW/ft)1 (1bm/s)) 3.09.100 _3.88 (563) 2.03 (0.618) 0.156 (0.343) 74 (134) 198 (356) 1055 (1440) 209 (377) 1089 (1500) 3.09.10? 4.28 (621) 1.10 (0.304) 0.075 (0.164) 65 (117) 168 (303) 1027 (1390) 3.09.100 3.95 (573) 1.02 (0.311) 0.078 (0.172) 66 (118) 3.09.20R 7.34 (1065) 2.16 (0.659) 0.170 (0.373) 113 (203) 133 (239) 1033 (1400 3.09.105 7.53 (1092) 1.38 (0.421) 0.085 (0.188) 105 (189) 164 (295) 1077 (1480) O -110-
1 l Feference
- 1
%/
Fernandez, R. T., R. K. Sundaram, J. Chaus, A. Husain, J. N, (VII.30-1) Loomis. L. Schor R. C. Harvey and R. Habert, "RELAPSYA C A' Computer P'rogram for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume III: Code Assessment " Yankee Atomic Electric Company Report YAEC-1300P, Volume III. October 1982. (Proprietary) 0.VII.31 Clarify how the ability of the quench-front velocity model to track the quench front has been assessed and under what conditions it has been assessed. A.VII.31 The quench-front velocity model, which is based upon Reference VII.31-1, is part of the Rewet-Quench Model (Reflood Model). fN The assessment of this model was clarified in A.VII.29. Reference Andersen, J. C. M., et al., "NORC00L - A Model of Analysis of a (VII.31-1) BWR Under LOCA Conditions," WORMAV-D-32, Research Establishment. R150 Denmark, December 1976. O.VII.37 Appendix K Requirement I.D.3 includes the requirement that: "The effects on reflooding rate of the conspressed gas in the accumulator which is discharged following accumulator water discharge shall also be taken into account" (Page 269 of Reference 10). Clarify how the effects of the noncondensible gas are included in the reflood mo8els. ( .
-111- - - _ _ _ _ _ _ _ _ . _ _ _ _ _ . _ a
['^} A.VII.32 )
\_ / - -
Compressed gas in the accumulators will only be discharged for v.. those LOCAs which are,,large in size and do not depend upo,n the steam generators for the removal of decay heat. Therefore, the degradation , of steam generator heat transfer as a result of U-tube blanketing is of fi j no concern. 1 yor these larger SBLOCAs, the following scenarios may be postulated:
- a. Cold Let Break e
Nitrogen, due to its lighter density, would either stay in the cold leg or upper portions of the downcomer while finding'its way to the break via the downcomer annulus. If the break size is large enough, the nitrogen venting will not impact the core at all. However, fer smaller breaks, the upper downcomer annulus may temporarily be pressurized which may depress the water level in the downcomer and transfer some liquid to the core. This liquid transfer to the core will cause quenching of that portion of the core which may not have already been quenched.
- b. Not Let Break The nitrogen injected will collect into the upper downconer region. If leakage paths are available to the upper head or the upper plenum, the nitrogen will migrate to these regions on its way to the break. The core will be unaffected during this period.
If no such leakage path is available, the nitrogen may leak via the Some of the nitrogen in ruction leg with no impact on the core. This this case may also bubble through the core via the downcomer. scenario will experience a swelling of the core level and consequently quenching of those portions of the core which are still unquenched.
- c. Suetion Let Break (D end (b).
'I This break will be covered under stenaries (s)
-112-I
i-yN Thus, we conclude that the compressed gas in the accumulator which is l discharged following accumulator depletion is expected'to have no negative impact on the outcome of a SBLOCA and is not considered e in our calculations. .- 0.VII.40 Clarify what effect the following two assumptions given on page 244 of Reference 10 have on the temperatures calculated: "ruel rod deformation does not affect radii used in heat conduction and convection except in the gap region." "Zircaloy-water reaction does not change material properties of cladding." A.VII.40 This question requests clarification of the effect of two assumptions made in fuel behavior modeling. The first assumption is that " fuel rod deformation does not affect radii used in heat conduction and convection except in the gap re gior. ." For the majority of SBLOCA cases, clad rupture does not occur. Fuel rod deformation is rmall and the effect on temperatures is truly negligible. For cases in which clad temperatures approach or However, exceed rupture temperature, clad swelling can be significant. in this case, the increased fuel rod surface area available for heat removal is conservatively neglected. . The second assumption is that "zircaloy-water reaction does not change saterial properties of cladding." This assumption is not expected to have much effect for the results of cases in which zircaloy-water However, the RELAP5YA fuel behavior model is being reaction is small. modified to include oxide conductivity and heat capacity when cladding is oxidized. C\ -113-
9 k' t 0.VII.41 . Page 268 of Reference 10 states: "During small breaks, the fluid velocity in theIcore region is expected to be small. Various regions in the core are expected to consnunicate with each other and the fluid conditions at various radial locations are not expected to be significantly different at a given elevation. 'Yhe entire core, therefore, may be represented by an average core for small break calculations." Clarify for very low flows that could occur during loss of natural circulation in the loops during a SBLOCA why the fluid conditions in the high power regions could not be significantly different than in the low power regions. A.VII.41 The difference in fluid conditions between the high and low reactor . Power regions during low flow conditions will depend upon the core loading schemes. For textbook type cases utilizing homogeneous fuel 1.osding schemes, the power generation will be the lowest at the core periphery and will monotonically increase toward the evnter of the core. For such radial power profiles, the fluid conditions in the high and low power regions are going to be different. Also, for canned fuel assemblies, the fluid conditions will depend upon the assembly power due to the lack of free communication with the adjacent bundles. Fuel assemblies used in PWRs do not use assembly cans and experience free fluid communication with the adjacent bundles. Also, the fuel management schemes try to achieve uniform radial power generation rates throughout the core. Figures VII.41-1 and VII.41-2 provide typical assembly radial peaking factors for the Yank- plant at Rowe and Maine Yankee, respectively. For the Yankee plant, i.ssembly 47 is the peak l power assembly with o radial peaking factor of 1.3033. However, if one takes a region of 9 assemblies around Assembly 47 (Assemblies 36, 37, f 38, 46, 47, 48, 56, 57 and 58) the average radial peeking is only 1.002. For the Msine Yankee plant, quarter assembly 65 is the peak It power quarter assembly with a radial peaking factor of 1.32240.
.O -114- .
i
4 J 3
.k j resides in assembly comprised of quarter assemblies 51, 52, 64 and 65.
However, if one takes a region of 8 assemblies around the hot' assembly (quarter assemblies 27, 28, 29, 30, 38, 39, 40, 41, 49, 50,,51 ,52, 53,. 54, 62, 63, 64', 45, 66, 67. 77, 78, 79, 80, 81, 82.' 92, 93, 94, 95, 96
~
and 97) the average radial peaking is only 0.95. ! j We conclude that the entire core, therefore, may be represented by an ' average core for email break calculations. O . O -115-
.,____________m_ . . _ _ _ . _ . _ _ _ . _
pf I 2 3 4
.5620 .7445 .7767 .5711 1.2139 1.2475 1.3503 1.2864 f. ,
- 1.0982 1.1439 1.2357 1.1585 5 6 7 8 9 10
.7177 1.1153 1.0769 1.2763 1.!?58 .6885 1.4377 1.5000 1.2782 1.5851 1.44f1 1.3909 1.3069 1.3904 1.2020 1.4788 1.3419 1.2650
!! 12 13 14 15 16 17 18
.7139 1.1623 1.1569 1.1462 1.0473 1.0495 1.1569 .6984 1.4059 1.5043 1.4366 1.5397 1.1938 1.2653 1.5067 1.4341 1.2905 1.3974 1.3452 1.4293 1.1599 1.2047 1.3989 1.3017 19 - 20 21 22 23 24 25 26 27 28
. F75 1.1340 1.0464 1.1776 1.1799 1.!!$4 1.1760 1.1697 1.1245 .5550 1.3062 1.454' 1.2544 1.5654 1.5006 1.4334 1.5753 1.4418 1.5039 1.2265 -
1.1801 1.3490 1.1952. 1.4570 1.399S 3.3369 1.4682 1.3517 1.3950 1.1085 29 30 31 32 33 34 35 36 37 36 '
.0011 1.2715 1.0428 1.121e 1.0350 1.0346 1.1883 1.1633 1.0999 .7557
- 1. M 90 1.5754 1.1994 1.4451 1.1896 1.1849 1.5019 1.5549 1.2994 1.2649 1.253S 1.4697 1.1542 1.3474 1.1467 1.1426 1.4003 1.4457 1.2370 1.1601 39 40 41 42 43 44 45 46 47 48 7 43 1.0761 1.1410 1.1840 1.0393 1.03?! 1.1219 1.0657 1.3033 . 7963
- 1. .~"J5 1.2923 1.53S4 1.5039 1.1974 1.1899 1.4343 1.2177 1.6153 1.3922 1.1502 1.2051 1.4298 1.404& 3.1534 1.1470 1.3364 1.1828 1.5073 ,1.2732 49 50 51 52 53 54 55 56 57 58
.5810 1.!!98 1.1494 1.1768 1.1273 1.1913 1.1888. 1.0657 1.1674 .6045 1.2229 1.5016 1.4316 1.5691 1.4467 1.506311.58F3- 1.2122 1.4959 1.3319 1.1079 1.3919 1.3402 1.4581 1.3495 1.4061 31 474gl.1664 1.3880 1.2010 59 60 61 62 64 65 66 O() .7369 1.1679 1.0464 1.0614 1.1744 63 1.18141.1916 .7341 1.4445 1.5090 1.2396 1.2213 1.5647 1.4717 1.5494 1.4435 1.3145 1.4016 1.1830 1.1848 1.4576 1.3785 1.4382 1.3150 67 68 69 70 71 72
.7419 1.1541 1.2990 1.1083 1.1578 .7422 1.4208 1.4708 1.6131 1.3327 1.5443 1.4812 1.2951 1.3688 1.5047 1.2522 1.4326 1.3469
. 73 74 75 76 KSDt1Pf!!N LEI.RLE ASES.Y 1.3033 47
.6198 .R?09 .7885 .5926 ASSDULY MD.
1.3376 1.4006 1.3016 1.2699 fur.Flo. E 1.6153 47 47 1.2072 1.2E29 1.1944 1.1497 f%I. De*EL 1.5073 Figure VII.41-1 Typical Assembly Radial Power Peaking Factor for TF9 Yankee Plant I'
-116
- _ _ - _ - . s
..e ; ,1 f %v , a; k 4 , s , W ' V ' l>r i , . 1'. ..2- .3 4 .
. l?fe .**M .3't! ' .3?N o 6
h- I 5_ 6. 7 1 MrJ"
~ .2832 .502* . .i201 .Mia '
i
.$14' . 62M .7710 .7t?
I- .5474 '.612" .7526 7?t: I: -
' COE SWOIRY Or m!!P.E AC. 45te.9 ptWER pela 14 '!5 ' . Il . l' 9 le , il 12 12 E! DIP 11th ml.mL a25tst.9 .' 311 .7M? 90** 1.01.2gt6 IM431.tW' .
.W2 .4127 .Y361.05991.10!!
A V86.. O, 6? 2? 1,2940 1,;7:1 m!. 4C 8 1."5I't!!
- 3. , 45 .{6*8 +. 777 .hb
.( M's El. D4 % @ ..1.43? o IM .3?S .376! -.5773 .6tte .92R 1.04221.142". 3.16271.1t* ^'
24 - ' 21 *.- It it 20 - 21 22 23
? !
- r. .M?t .$0D ,' .7%*J 99031.11201.22M 1.2th 1.29231.2t '
'r .5106 .u01 .0"51 1.0*ff 1.3324 1.4M2 1.49?4 1.4929 1.4t< 4 4'*4 .t?'! .tt tf 1.079212?ll 1.?*421.47.1 1,420s 3.2C 1-27 . ?!
- 2*/ 34 ?$ $t ' 1 I*. 3t :.06441.17%
. 3) 32 ~
1.23041,2t171.*M15 121 1. 3.*k 1. 4' b 1.3% l. os 1.
.
- M.'
2 3!'* .74t .95N56M 3 a.2*E 1.0 l.16MM1.Y291 1.115' 1 -
. 0 .t34 .90l? 8.0M L.l?01 1.3114 1.2190 1.3444 1.3072 1.2$04 1. 44 1
2 39 40 41 42 ' 4 '44 4S ' 44 47 4327 ' ,iM 1,0't431.le?2 1.2?eg g,;7,3( l.22121.20?" 8.!9d 1.16711.171
. 6(*7 7t'e 1.22241.40N 1.42321.4tE7 3.2421.3040 8.M21.?"90 8.240
' .583 .4 M 1.ltl7 1.M03 1.3511 1.4092 1.2610 3.243 1.243 1.1994 1.2d 4 Se l - S7 - St f* to c i:-
de 50 St $2 !J S4 !I5
. *211 .41tl .P*U l.13721.164? 1.2455 1.W721 N31.31e 1.303e 1.022? .9625 1.26"
' .?Al: .414! 3.28t) 1.2149 a.27571.3F52 1.!!451.!!561.t1021.49441.0k1 1.01021.4U8 -
.3260 .5641.M431.250t 1.21441.3327 1.09571.10301.42:31.420e 1.07t? 99021.3(4 C &? 64 65 6t 61 4t 6 . 10 - 11 ' 11 13 ' 14
..3P?5 43311.11791.3D41.21331.24k 1.01c3 .9001.2@ 1.2486 .9727 - .9414 1.26 * -
.5547 7662 1.3171 1.51121.3145 1.35t* 05401.01561.43491.41?71.02t2 .98tt: 1.42**
.517[ .16% l.25161.4N5 8.25371.3222 1.04W .99571.h33 8.32941.0054 ' .97361.3M' e5 06" - 27 se t$
75 ft 77 - 7e 79 ' ' 90 F1 82 23 04 t' 74721.0ke 1.1 l12 1.2121 1.3006 '. 2663 .9745 .97901.082 bl33t 1.jt22 '
. 7tl .. 1.0223 1.2340 1.{ 1.2137
.g690 ..plli l.3157 3. 1.*3894 1.30ee 1.4*94 .432t 1.01491.02ie 1.17941.2347 8.es**
.2542 .49C .91001.ltM 1.. l.254f 1.24411.2445 1.4243 .ho* .99201.00201.823U 1.188! 8.8m
' 90 .- 91 92 9) t 94 ' - 95 96 97 90 99 100 101 102 103 ' 804
-.h64 .10H 95101.19941.Ne91.24u 1.22051.1222 1.30k 1.2570 .9514 94991.0870 8.13551.lt!
.PU .4405 !. fit? 1.4083 8.P021.2841.31D 1,4.1271.42031.343! .9594'.H06 1.49401.4165 .9t37 3.20051.24121.*2a6 .
5
.376! .6309 <l.0621.Mit 1.7M01.32381.24621. 61 .95621.14518.8931.lM* -
105 106 107 ' lof 106 110 til- 112 l'It 114 115 - lit 117 lit 186 4130 .7k? l.4tf521.21071.94ft 1.01111.]O9) 1.3028 .9948 .97511.24401.N901.lM3 5.!!% 1.80'!
.6M3 .Othe 1.299? 1.4N21.11551.0551 1.m) 149.141.03721.01561.40631.41241.21001.lt9} 1.8%!
*( .5777 .46241.19081.35211.0966 3.0t111.42491.41991.024 .9M9 8.3M3 8 3F951.16171.131. l.lMi 120 121- 122 123 IN 125 126 - 127 - 1 2 , ' 129' 130 !!! !!? 133 IM
": ? 5318 '.900? 1.1901 1.2801 1.M29 .9821 l.24561.2547. 9596 .96541.h*61.27241.12971.10071.0757 22% 1.0991 1.hti 1.49t21.1161 1.0ld 1.43221.4150 .9970 8.00031.4407
.97C .98261.37471.ht* 1.16E 1.11873.4441 1.21h 3.114e . 1.17E 1.lf 70
.69001.07961.31131.4001.10M .9401.hm 1.14'A 135 ' lE !!? $3' 139 140 ' les le 10 144 145 145 147 ist ~ 148 '
7MS 1.11231. l41.31611.267t .9737 9497 1.2414 1. 2483 1.1562 1. LM1 1.0953 1.05t9 .28:4 71.*5',46
- 41. 8.S1051.43461.0142 .99201.4027 $.444 8.2Et 8.m 1.1790 3.14M '.92'5 -
~ 1.0401 9239 1. 23 1.*E'29
- 1. 1. 3.43351.3a.M .9915 .95611.3333 3.37k 1.3m 1.3602 3.12513.07*4 90'7 3 8.260 198 id 162 161 164 lgt it'4 .
150 151 IS2 153 154 115 IS8 157 152 94791.N4 f.J61v!? 1.1M0 3.10721.07001.lE7 .M4
.l790 .MC .90k 1.229' l.kl41.20h 1.10371.2454 9774 2W .57491.10131.46441.2541.W42 8.4M41.43941.0206 97% l.4111 1.44E !.22M l.19921.16% 3.8341 .9F
..%66 .54751.04M l.P451.>451.2427 8.42061.32901.0015 .95521.38621.ht21.8483 8.135t 1.09a61.0745 14 7 - let 169 !?0 17j 172 173. 174 175 ' In 177 178 179 100 tel st2 St2
.2505 .3030 8.01@ 1.NM l.251. l.19981.0226 .97N 1.08291.0421.13351.12M l.0P531.07tKi 1.07081.08
.h45 .6330 8.27N 1,4935 8.35051.302 8.08681.02W !.17821.19971.2001.21021,17921 ~
let - if5 106 If7 10* IM 190 191 192 IM 594 195 196 197 192 199 - 20
.32tt .0011.te431.2923 l.21211.1477 .96M .94t? l.83311.'1379 8.10.4 8.10021.05651.05% l.07761.0*% I.232 4?C.75261.8631.420' 4410 .77101.20601.49?' 4 25ps 3.1964 9901 3.30631.25M !.0802 .9t'91.2MI 1.MN 1.30P7 8.37301.!M91.1 .
29' 20e ? *10 288
- 214 2:a ;gg 20! .44?4 2r0 99? 2M l.25t' 20' 30f 261 3.3 d "g 1.224!
!' 3.t44? l.11541.lN11.?!011.2tti 1.j$25 .4 8.th 1.f**t
.P I.fd.
* " I 3.24 I,4M PO217 .
's ly'Y;,
7tP 1 ?d 1.4 941.2t?! 1.?t!21.45?7 f.4'*21. 4n 8.7;4? 1.le" 1.le.. ..*;lf
.?*i
.t !! 57t!
,45t9 .nel I.in7 8.M221.2cn 1.20121.3M41.34211.17721.st491.l?tt 1.182 .9072 .91491.27f71.5M Figure VII.41-2 Typical Quarter Assembly Radial Power Peaking Factor For The Maine Yankee Plant l,
-117 6
'I VIII. METAL HEAT f .t j A
O.VIII.1 t.; .. Clarification arid justification should be given for the metal slabs associated with all of the nodes used in all SBLOCA-related analyses. A.VIII.1-Figures IV.2-1 and IV.2-2 represent the nodalization used in the pump trip studies perfonned for the Maine Yankee plant and the Yankee plant at Rowe, respectively. Similar noding schemes will be employed for the small break calculations. As is seen. Yankee uses a detailed nodalization of the plant. Also, the noding does include the representation of the reactor vessel wall, the piping wall and other 4 major internal structures present in the plant. A list of metal slabs used at various nodes will be provided with each' plant application. Clarification and justification of metal structures not represented in the analysis will also be provided at that time. 0.VIII.2 Clarify especially the modeling to be used for the pressurizer walls, upper head and upper part of the steam generator U-bends because these areas will affect the pressure during the SBLOCA refill processes. A.VIII.2 Yhe pressurizer wall and the reactor vessel wall in the upper head region will be modeled in our analyses. In addition, the major heat structures present in the upper head region will also be represented. The steam generator U-tube model is shown in Figures IV.2-1 and IV.2-2 for the Maine Yankee plant and the Yankee plant at Rowe, respectively. The U-tube region has been nodalized in detail and the U-tube bend region is explicitly represented by two heat structures.
-118-l
I II. BREAK F1,0bf
.{ *
'#- 0.11.5 Break flow depends on conditions of the' fluid reaching the break. If HPI or accumulator flow is not injected into the cold leg with the break, then the resulting break flow will be very different in the meJe1 from the actual transient because of the difference in' void fraction and tengerature of the fluid at the break. Also, HpI and accumulator water injected into the intact cold legs that bypasses the core and goes out the break needs to be modeled to accurately model break flow. Clarify how the treatment of HPI and accumulator flows affects'the break flows calculated.
A.II.5 Figures IV.2-1 and IV.2-2 represent the localization used in the pump trip studies performed for the Maine Yankee plant and the Y,3kee plant at Rowe, respectively. Similar noding schemes will be used for small break calculations. As seen in these figures, the cold legs are modeled in detail for the two plants. The ECC has been modeled to inject into the cold legs very close to their actual injection points. The break is modeled in the injection node. This nodalization maximizes the amount of ECC flow going out of the break and minimizes the depressurization rate due to the presence of two-phase fluid at the break Iccation. Therefore, we model the ECC locations accurately and select the break at a conservative location. l O.II.7 Clarify for stratified flow stich critical flow model is applied above I the interface and which is applied below it and how any switching process between the models is modeled. oN. -119 i
A.II.7 i
/M *' s When critical flow is determined to exist at a junction, then the code .
O selects the appropriate critical flow model based upon the junction void fraction, . The determination of the' junction. void fracbion for stratified flow is discussed in the answer to question Q.VI.7. For
*( < 0.05, the Subcooled Critical Flow Model is selected. For
*( g 0.05, either the Moody or the RELAP5YA Standard Two-Phase Critical Flow Model is used depending upon the option selected by the user. If the junction void fraction' lies within the transition region' j defined by 0.01 < < 0.1, then the underrelaxation scheme defined on Page 90. Appendix A of Reference 11.7-1 is used.
Reference (11.7-1) Fernandez, R. T., R. K. Sundaram, J. Chaus A. Husain, J. N. Loomis, L. Schor, R. C. Harvey and R. Habert, "RELAP5YA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description," Yankee Atomic Electric Company Report hd YAEC-1300P, Volume I, October 1982. (Proprietary) , i 0.II.8 Deleted. Re!9tence (II.8-1) Memo. Ed D. Throm, Reactor Systeta Branch DSI to Brain W. Sheron, Chief Reactor System Branch, DSI.,ESumary of 6/22/84 YAEC-NRC Meeting on RELAP5YA PWR SBLOCA EM Analysis," July 17, 1984. 0.II.9
- Clarify why the vesiculation showed more depressurization than the data between 10 and 60 for Marviken Test 10 and yet the calculated break flow was lower than the data for this same period (pp. 39-40 of l Reference 12).
-120-n._--.____-_ . - _ _ , _ _ _ _ _ _ _ _ . _ _ _ _ ____ _ _ _ _ _ _ _ _ _ _ _
A.II.9 The original RELAPSYA case underpredicted the pressure and escaped mass. at 60 seconds because the initial liquid level in the vessel.was. incorrectly set too low at 17.06 meters rather than 17.66 meters. This input' error has been corrected, and the two cases (RELAp5YA Standard Critical Flow Model and.RELAP5YA Moody Critical Flow Model) have been l rerun. Revision-1 for Figures 2.2-12 through 2.2-19 are attached. These revised figures also contain data uncertainty bars, where possible, from information now available in Reference II.9-1. The maximum error in the vessel pressure measurement has been estimated by the Marviken staff to be 190 kPa (Page 50 Reference II.9-1). Figure 2.2-12. Revision 1 shows that the new dome pressure overlays the test data from 5 to 65 seconds. The calculation did not predict the pressure undershoot between 0 and 5 seconds that has been attributed to delayed nucleation in the vessel, nor the brief return to suberitical flow during this period. Therefore, the blowdown period was calculated to be about 2.5 seconds longer as shown on Figure 2.2-12 near 70 MD seconds. ' Figure 2.2-13. Revision 1 compares the calculated break flow rate to that derived from the Pitot Static Method by the Marviken staff. They have estimated the error for this method falls within 18 to 115 percent in the two-phase region (Page 48, Reference II.9-1). Based upon this, we have placed 110 percent uncertainty bars on this test data. Figure 2.2-12. Revision 1 shows the calculated break flow rate lies within the data uncertainty from 10 to 65 seconds. Figure 2.2-13A compares the calculated break flow rate to that derived from the Vessel Inventory Method by the Marviken staff. They have estimated the error limit to be 16 percent for periods of clowly decreasing mass flux and 112% for periods with more rapid changes in mass flux which occurs prior to 10 seconds in certain tests with 500 mm nozzles (e.g., Test 10). Also, they state this method is not valid when the vessel level drops below 2.3 meters above the bottom. Test The comparison data shows this occurred at 57 seconds in Test 10.
-121-
i shows the esiculated break flow rates agree very.well with this test data between 10 to 57 seconds. 54 l Therefore, we conclude that RELAp5YA calculates the two-phase,. critical flowrateandpr,essurehistoryfrom10tc65secondsverywe51'f5rthis
. test based upon these comparisons to test data with uncertainty bars.
l
. Beforence 1
(11.9-1) -The Marviken rull Scale Critichl Flow Tests: Summary Report, Jaint 4 Reactor Safety Experiments in.the Marviken Power Station Sweden, NUREC/CR-2671, MIC-301, U.S. Nuclear Regulatory Commission, Ma/ 1982. 1
/
k O i i
-122-
0 _ 0 _ gv 1 ~< 1 0 v
. . 3 e w R o s l .- n F o _
Y i _ T Nl t N O a c _ I I c . o _ A Ti s i Ct d T e R Ii r _ E Dr P C EC .- N R l U Pd e - r g Ht i d
& I a l 0 1 o A .dn 1 s\
- M T Va h
A Et e wo D RS r l u F s s 1 e n _ ( 0_ 6 P r e t I1 4 e it m r o C _ D 0rd 0 5 ) 1 a
, S
( t d s n a et T S g - E n
, v 0 M e I
4 T i v P k Y rA 5 A a L E MR 2 1 0 - 3 2 2 e
- r i .
0 2 i F u g 0 1
] .
h ' 0 0 1
,= i RM Rg I 8 _
O 8H E 2s w0$,gn. L~Y _ l!. l _
;II' il' lI J
u.
't.
m. s 0 _ 0 _ 1 _
~ _
f U m u 1 0
. 3 v w e o R -
l - _ d F s _ Y o n A o Th Nl Nt Oa i t I e I c AM Ti 0 c . 9 i Ct =. T J_ Rc Ii d Ei Dr e Ct EC r N a R P Ut Pd S r l e
& 1 a t 0 d A o .dn l
~ o ~
M Tt Ai DP Ya Et RS N. N e t w ao Rl F w _ y 0 ol l a
,h 6 F c TA 4 ;, kt i
ai ~ 1 er rC
;.~ B d
7- 0 ) 0 r - 5 1 a ' r S d - ( t n Tp sa E et n TS N _ u 0 4
,i T
nA eY k5 iP _ - vA rL aE MR - d1- - 0 3
- 3 2
t a- - 2 x e r u g _ 0 i 2 F 7 x b 0 1
~
~
f/ L A n 0 a i 1
- 1 0
1 g, ,- 83 0Se 8ge 8o3 ] _m _TE[H@ 33" mmN t U.' ( l !l!ltIll l i' i
) I M %(b 'r d
o w - M h o ' t l e F YM _ T Nl 1 _ Ny O a ._ I r I c Ao Ti v Tt Ct e R n Ii M R E e D r - Cv EC s N n R n o UI
&l Pd 1 a r g i t
e N c A s .dn - i Ts A e V a Et % 0 1 d e DV RS \ P r p
\ l e -
d
- ha 0 e
M o 6 T1 ,- t w ao T Rl
" r F
_A w
" - ol
- l a -
F c . 0 i h . 5 } S kt ai ( er rC ~_ E B - M d R_ 0 r I -
~
T 1 a
, ' 0 d 4 t n sa et -
TS .
.%=
_ nA eY , kS iP
< - 0 vA 3 rL aE MR a
A 3 1 0 -
" 2 2 t
e- 2 e r
- a. u g
_ i
. t 0 1 F W
.\
[ \ I , 0
.~
0 1 8O$ 1 e g8 88 $4 - $ m 83e
, _ TM WN g;' $$
' .~Y t !
;i l lil1il
] s, 0 0 1 Nl R Oa E I c T Ti E Ct 5 M Ii - . - 0 O Dr o 2 T EC A M, I S R Pd s ,5 T6 AN r A4 TE 1 a ** D2~ AD .dnw _ M D Yao . Etl - RSF o 1 s r v
. ' -" A -
^ R e -
s n o o t i
- t c
i d
- Tl Il l e 4I I I [, P r
w o s l e d
- M o
d
~
y+- 0 5 1 no t
.w ol iF -
cl - S aa ( rc Fi s E t
]m- n di I l r oC
' o T V d
t
. 0 r 1 a d -
t n
= sa n
- et TS
- v nA
- 0 3 eY
.v
- k5 iP vA
.v rL
- .. aE MR
_~.. q 0 4 2 1 2 2 e f - r 0 u g 1 ._ - i - 1 F l o o e" n$ e. $ ..- ,
.4 .4 .4 " ?n 5 ng" eg
.;T
[ !lIl1 ' \l
jll
, J 0
0 _ 1 NJ% w o 0 l 3 .
- : d F Yo . -
_ Th Nl _ Nt Oa I e I c 1 _ AM Ti T Ct . Rc Ii . 0 v - Ei Dr 9 e Ct EC R Na R - Ut Fd s S r n h .L & 1 a o t
- Ao .dn i t
Tt Va U, , Ai DP Et RS 0 1
- i d
c e
# P r
. ' f L '
l
.i e
/ T , -
0 6 d M w o
)/ / -
3 ~
~ sl o
/
~
~
sF a Ml 0 ) a 5 d c S ei
) ( pt ei
(. . E cr M I E xC d _ T0r1 a O . e 0 4
- t n sa et d
- TS
- aA eY 0 k5 f 3 iP _
/ vA rL -
' aE .
MR
/ 2 0
5 1 2
)/ ' 2 e
r u _
/ 0 1
i F g
;// _
- 0 0
. :1 I B 3 o-S" R o Gc wy 1 ! ,!1
]ll O_ i
/
' o o
g
- hw N
+ ,
0
*. . 3 d w .
ro o rh Nl ut OF t e I aM Tl r R c C a I c ti Di 0 ct Et 8 na Ri P r ut S C a t 1 y - ao Vo
.d rt E o xi _
0 oP RM 1
% 1
.= e . -
t v T1 t i 0 ae ,-
,- 6 RR i'
I \ k ws on l o F !.
= .
t LV 0 5 ) sc si ad S M er N 7 ( kP
)
,~ r y E el M re I
a Bd _ 7( T 0M o
~ - _
. , 0 1 4
r y _ td T so eo TM J1 T , 0 3 nA eY kS iP vA
^
m rL aE
'LL. MR e,
0 2 6 1 2
- 2 l
'.IA ~
e r
- - u g
v o t i
' F
. L
- k
[ , p 1 L 0 o a 1 J 0 1 8 *T h8 8a 8S s o-pV , Tex eg $ W@ ie
,ll l
; 1
- i!
0 0 1 j U . d o h t 0 e w . - 3 M o nl Yy oF Tr i No tl I t An ca i ci Te d 3 Rv E n et 8 ri CI N P Cr Ul 1 1 e y As .d e . Ts Ae ve oo t v DV - o ae RM t RR
'l ws r on Fi o
t Ty sc si 0 ad A"" 6 Me r N kP a 2 t \ k el re
\ Bd o
0 ) 0M 1 5 y S td ( so eo
- E n TM
~V(
- 0 4
I T nA eY k5 iP
~
y vA rL aE
.- .{
MR n A _ - 0 6
- 3 1 2
- 2
^
O. e r
- 0 u g
" 2 i
- F
-.,T L-
- , ~ o t
- 'L 9
~
' i1 1~ 0 d
0 7 3 8 o8* g3 g8 83e 88e 88e
}C' /
n T o::x~ N . sCd mmf
' .i I l
l1 o o o t
. o
. s y
t w = n o i nl a i oF o , t s r tl 1 e ca c i c . n di v U et e ri R d P r n C - a 1 y
. 0 9 7 es a .d \ rn t vo uo a eo si D st ec R_M ri Pd e
Iy A~ ( o s mr eP t sl
\ \ \{1 ye Sd
-== \ o 4 0M
\ o 1
y s )S td m. N N ( so eo TM
^ A E nA o
eY s
\ oT t
r: I k5 iP vA rL N MR aE M : 7
\ 0 3
1 2 m N 2 _ x 0 i e r u g 2 F
~
o 1 t
~
] m r
! I I - f o o
3 g
.g E h I 85 .
8H - O <k egWo.
,c Il . 1
ili I! i 1 o o t n o o
'i r t e i c -
t . 5 e dl 0 0 m ee 3 2 o rd . M t P o ' . i M
, s 1 y
a6 n t4 ae vo
. d a2 tD DM a eo Dy RM o s
-r 1 l X' A -
~
v e R
- o -
l s n no oi it t c
" o s
ci ad re F r d P il _ oe a
~
Vd o 0M 1 0 ) y a 5 td S [ l so eo 1 I) ( I
/
TM n
~ & E n nA eY
/ s I k5
- o T iP o
t vA rL _ . aE MR M. e~
,V V
V 3 0 8 1 2 2 e r u A
.y 0 2
i F g A
/ o
- 1 t
/
/
o
=
g 9i e5 e e .f
. *o eo no 1 2OE "0
_o
,T l l l ll l ,
,jlilj 0 _
0 _ 1 _ O - g 0 3 . e m y _ td w _ no o ih nl at oF t e i tl 0 rM 9 e ca 1 m _ cc i c . ni di Ut et v
.g . a ri e
. s . dt P r R ,
. nS C .
- a t 1
y 0 ao .d '7 s t t vo n ai eo o DP RM st si m
. . ac Mi d -
0 d e
~ A- 6 er pP a -
cl s e r-f .
. E 0
5 ) Ed 0M 1 td o y u. S so
/ ( eo
> TM E
h" ~ M I T nA eY k5 . O 0 iP . 4 vA rL aE HR
/- - .
/
9 - 1 0 - _ 3 2 2
. e
. r u
g
, i f 0 2 F
+
.f 0 -
1 - u .
'0 ,
0 I S g s~ ' $
$C O -
g.
L 0.71.10
,m
(%J ) Clarify also Why Page 42 of Reference 12 shows the code underpredicted the escaped mass at.60 s and yet it also underpredicted the pressure at 60 s. - A.II.10 As explained in Answer A.II.9 the revised case with the correct initial liquid level in the vessel shows that the calculated pressure history agrees with the test data very well from 5 to 65 seconds. However, the calculated escaped mass still lies below the reported data at 75 seconds. Since this computer run had a small mass error (less than 0.2%) at 75 seconds, this discrepancy was investigated further. The test data shown on Figure 2.2-15 was derived by the Marviken staff from the Integral Pitot Static Method (integral of discharge mass flow rate). Since the estimated error of the Pitot Static Method is 18% to 115% for determining the two-phase flow rate, then this error is also reflected in the integral of this parameter.' Therefore, 110%
-g .m,) uncertainty bars have been placed on the reported escaped mass test data. This shows that the predicted values lie within the data uncertainty from 10 to 75 seconds.
Next, we note the initial mass in the vessel and discharge pipe is 279.644 kg. A variety of test data indicate the system was essentially filled with vspor by the time the ball valve was closed at 78 seconds. The final system pressure after 78 seconds is 2350 1 243 kPa (341 1 35 psia) from the six pressure transducers. The saturated vapor density at 2350 kPa is 11.78 kg/m . The final mass retained in the 427.33 system must have been about 5,034 kg. Therefore, the maximum m value for escaped mass at 75 seconds is 274, 610 kg. This value is within 0.065% of the 274,789 kg value predicted by the revised RELAPSYA calculation. This value is also 5.6% below the 290,000 kg value Therefore, we reported as test data at 75 seconds on the figure. conclude that RELAP5YA accurately predicted the total escaped mass from the system at 75 seconds, and that the reported test value is in error. O -133-
L 1 v f 0.II.11
~,<~w.
J(,,,/) Clarify if-the discontinuity in slip between the Moody-model and the RELAPS MODI subcooled choking model could cause flow oscillations and potentially masi" conservation problems for break flows near the 0.05 , void fraction switching point between the models. A.II.11 The original implementation of the Moody critical Flow Model in RELAPSYA included a) tabular values of the Moody critical mass flux as ! I a function of the upstream (donor cell) pressure and enthalpy, and b) i 1 Moody's theoretically derived slip ratio given by: ~! K,=(V/V)=($g/f) g This method led to computational problems and larte mass errors in the donor cell. r% Subsequently, the method of implementing the Moody critical Flow Model i
- in RELAP5YA was modified. The tabular values of Moody critical mass flux as a function of upstream pressure and entha) y were retained.
{ However, Moody's theoretically derived slip ratio was replaced by a i j solution of'the difference momentum equation using a relatively large value for the interphase dras coefficient. This approach has yielded well behaved solutions to date. It allows some slip between the phases (1.0 to about 1.3). These slip values are compatible with those calculated-by the subcooled chokins model which are near unity. The masc flow rates calculated by the Moody Two-Phase Critical Flow Model l l are still conservatively high compared to those predicted by the 1 RELAPSYA Standard Two-Phase Critical Flow Model. l l Several SBLOCA-EM calculations have been reviewed to examine the calculated break flow behavior near the 0.05 void fraction switching point. Two cases each, from the Yankee Plant Pump Trip Study (Reference 11.'11-1) and the Maine Yankee Pump Trip Study (Reference 11.11-2) are shown in the attached Figures II.11-1 and l \ ( '
-134-l
.. l 11.11-2, respectively. These results show that the transition between the Subcooled Choking Model and the Moody Two-Phase Critical Flow Model l 4
'ir relatively smooth. Note that the subcooled critical flow rates ( 4 less than 0.05) are slightly higher than the Moody two-phase' critical flow rates (0(gr'ester than 0.05). The mass error in each run was small. Therefore, we conclude that the transition between the )
Subcooled Critica1' Flow and the Moody Two-Phase Critical Flow models f yields well behaved solutions. 1 References, (11.11-1) J. N. Loomis. J. H. Phillips, A. Husain. " Reactor Coolant Pump Operation During Small Break LOCA Transients at the Yankee Nuclear Power Stations," YAEC-1437, fankee Atomic Electric Company, July 1984. j (11.11-2) L. Schor, S. Haq, A. Husain, J. Chaus, " Justification of Reactor Coolant Pump Operation During Small Break LOCA Transients for Maine Yankee," YAEC-1423, Yankee Atomic Electric Company, April 1984, t l l l l e k
-135-P
/~% . l l U '
e r _ m N 1 y $ N g l
!F l g i's #
jf t
$i f
-s=2 H; i
ull E- ; L ! Bi s d , i Eig_ q; i! - cg6 . l e <=:- I ( J u 5
) -
,,sm 7,,, - -
h "tm ex L a m 3 C (~) i I .. E __
- I y 1 :
} g_ --
# g e
-i
-2 "
gilL 25. 2 'l E8 il p e= i: H,/
,f !! i1 g s !!
gd, rim Al =
.i 8- I i
~ ~ " Mon 3M E10A (ass /n"En) agu xr27 i
-136-
l ;1 !ll 1!li;ll
~
_r - piJ -
=.
=
.* m
=.
e e.
=
r .
-{ , ., g e . s w e w n e.
=
e l f p e
- e. h . e w
s
.n . i, s.
p n.
.s
.ua ..
sam =e 3
- m. s.
- u. =. .t 3, <. n e.
- a. _ =
e r 3, 4 m c .i. p e.
,s . .
3 a.
.= p.
ag .
, n s
, s p
. =
. =
s N . e s a C M E
= -
A
. C
. . O g: e9'= I4 d i. .. I- .~ . I4 L
,5
, .E .g .i . 5- ! .B .9 .* -
B Rtrc i l S Y3 4c l
. e e
k n a Y
- e
- t
( i n a . M _
. 2
. - - 1 1
m i X
. I
= e I
r 5 . t e t.r u g
+ e s
a
.m.
. i m .e F -
" a.
. M .
c 4 e 8 .. ..
- e' e
" . .ee.s 9
..e , .e
[
*,, l.
.e. . .. e.
., G.
. . s.
~
'._ .e
*a , _ . lg f
v.: .
.=
f ' i (%
= .
e
- e. 1, 2 .. I-
.5 e E. t; i .*
-[
._I 9I. .i: *I .I .... !l se Ss s: 2I I..
C'
,*a' 3e .
8 wwe
I [ 0.II.12 . Clarify why using the Moody model only after the void fracti.on exceeds 0.05 meets the Kppendix K Requirement I.C.1.5 that'for all times after the discharging fluid has been calculated to be two-phase in composition, the discharge rate shall_be calculated by use of the Moody model. A.II.12 paragraph I.C.1.6 in Appendix K requires that the Moody Critical Flow model be used after the discharging fluid has been calculated to be two-phase in composition. Restated, one first needs to calculate that two-phase discharge conditions exist at the break location, and then use the Moody critical Flow Model thereafter. However, this requirement does not specify the void fraction nor quality of the Moody discharging fluid that is to be calculated prior to using t! Critiesi Flow model. Therefore we have selected the small void ps fraction value of 0.05 as our griterion to begin using the Moody ( model. This smali void fraction was chosen mainly to ensure continuity in the critical flow calculation between single-phase and two-phase conditions. Figures 11.11-1 and 11.11-2 from the answer to the previous question show that the rubcooled critica). flow rates are slightly higher than those calculated by the Moody Critical Flow Model Thus, we comply with Appendix K near the void fraction value of 0.05.
' requirement I.C.I.6.
O.II.14 Clarify if break-size sensitivity studies will be done for SBLOCA analyses. l A.II.14 Yes, break-size sensitivity studies will be performed and submitted on a plant-specific basis.
-138-
i
! 4 I
i 0.11.18 v. Page 89 of Reference'8 states: " Equation (245) can be derisdd from Equation (153) 'by neglecting the third C term in $ and setting C-0 (stratified) on the right side of Equation (153) and_C=e0(homogeneous) on the left side." Clarify how the opposite extremes of the virtual ! ress coefficient can be justified for use in the same equation. A .1K .18 The cited text contains errors, and should state the following: Equation (245) can be derived from Equation (153) by substituting Equations (148), (149) and (150) into Equation (153) with the following assumptions:
- a. Set C=00 (homogeneous) on the right side of Equation (149).
A (,,) b. Set C=0 (stratified) and neglect the third term on the right side of Equation (150) Justification of the model given by Equation (245) is based upon INEL's extensive experience and YAEC's assessment reruits to date. Although the two assumed values of C are significantly different, nevertheless, the resulting equation produces results that are in good agreement with test data (e.g., Marviken, LorT, TLTA). The following discussion may explain why. . Figure 15 in Appendix A of Reference 11.18-1 shows that the virtual mass coefficient in Equation (149) has a significant effect upon the two-phase equilibrium sonic speed. Thus, this coefficient will have a significant effect upon the mass mean Nach number defined by Equations (148), (149) and (154). Figure 16 (Reference 11.18-1) shows the coefficient. D, defined in Equation (150) that multiplies the relative Mach number is bounded by 10.6. The relative Mach number for i - critical flow conditions is generally an order of magnitude less than
\s/
-139-t
)
i a
?
I the mass mean Mach number due to the relatively small phasic velocity differences that develop. Thus, the second term on the left side of ~ f Equation (155) is generally much smaller than the first ters, i.e.: l M , + D M , = 11 (155) 1 V D(V* -V)g My E [ > D M,3 , Therefore, the influence of the virtual mass coefficient, C, is weaker in parameter D given by Equation (150) than it is in the sonic velocity, a, given by Equation (149) when determining critical flow conditions. This observation was checked by rerunning the Marviken Test 10
.Best-Estimate Calculation and asruming C=ocin Equation (150) consistent-with the assumed value in Equation (149). The results are shown in Figures 11.18-1 through II.18-5 as the sensitivity case indicated by g
the small dash lines.. These results show the break flow rate and discharge pipe void fraction are slightly higher. This causes the vessel dome pressure to decrease faster and moves to outside the uncertainty bars on the measured dome pressure. Both the Revision 1 prediction and the sensitivity case predict the escaped mass history within the uncertainty of the reported test data. Therefore, we believe the original model given by Equation (245) remains the best model to use for critical flow calculations at this time. Reference (11.18-1) R. T. Fernandez, R. K. Sundaram, J. Chaus, A. Husain, J.' N. Loomis. L. Schor, R. C. Harvey and R. Habert, "RELAp5YA - A Computer Program for Light-Water Reactor Syctem Thermal-Hydraulic Analysis, Vclume 1: Code Description, " Yankee Atomic Electric Company Report YAEC-1300P,' Volume I, October 1982. (Proprietary) U .
-140-
;i$!)
J _ o s) o t ~,(J%
. o s
s n o
- , i t
y c t i n .. o d i n . s e a o e r _ t t s P r i a e c C - l e c i y n d d o U e t t d P r i v L r s..g
.. w M
n i \ w a 1 t \ \n eo a . i s A ' \ rl uF t D a R v e S n e -
,\o\ s sl ea rc
- k \s o Pi
- \ - s t g ei mr Ig A 9 s D oC
- d
- - s 0 r 1 a d
0 ) t n 5 sa S et ds ( TS nA
' E eY l
r k5
% I iP vA 0
4 T rL - aE 7* % MR 1 s 8 o a 1 X I e r u g i o F s o s
]h (L o 0
o g- 4 $ eg g I sk . 8M - k Wc:)$g@ o 3)G yt b*Y
^
0 0 1
/
"(\w 0 3 s n
.- o t -
i - y c
, td i no n 0 d e
ih e 8 at o s 'r t e i a P _ rM t e c C l cc i y e ni d d Ut e t o a r i M dt P v k o l e
.i nS t
- t w a t I
i \ \ ao s \ Rl ao t t v n e
\. \ w F
ai e .S g
\ ol DP R l a
\ F c
\ i L
t 0 6 kt ai er T2 4- - . B rC
~{
d 0 r
- - c.
1 a d 0 ) t n J_ 5 S sa et r ( TS T p E 1 f nA eY k5 e-
) I iP o T vA rL t
- aE m~ MR 7 :
2
&~ 1~
8
,0 1
' 3 .
X
% I I, - e yL - r u
g
?L 0 2
i F y-5
.~ .
=W-
,o t
I
%4 g#
g 3 1 Oo
,o t
8g yy o-83e aU e
, _* u:: N 3OJA g*<C 1.. w 8
l; iI !l A , L m w
=
M . e xu-c ( a M s a d o
- n .
o yt h i t K - t e nM c i n i _ ay o e d t r t s M e _ ro i a r - _ et c C P cn i y ne d e t l e Uv dI n P r i v
% d o
n al 1 i t k N 0 e M e i J 1 t w ao as v
. s n \ .g Rl t s ae DV R e e S.
,\ \
\
\
\
w ol l a F F c
. i 0
TA 1
. m\ 6 kt ai er rC r-B 0 r 1 a d
l-b t n sa d d M )S et TS W ( E nA l r eY _ k5 xo
~
( _' % 0 4 I T iP vA rL MR aE N 3
=" 0 3
8 1 X I e m- ^ 0 2 i F r u g
,M A.
_ 0 sE 1
+N L
;0
\
m" l 0 1 c 8g 3 $" -
) .
83e 88e
$a a
, _T u:: w*& s
, ,l l
1 II, 0 0 1 n o e t s N i a
. r c C e
f'Q i t d y 5 e e t M 0 m r i 2 o P v M i t i t 4 1
, s i s a6 an . h . n _
t4 t e v n o a2 aD e e i t DM D 0 T R.. J., 8 i c
. d e
4
- X .
. r _
. P o
l l d e _ o M Tf w tI l ' o j I Is nl j l: I g , o i oF tl l .' I s ca l I I s - ac j/ -
- Ft ri i
dr C M4-i o 0 Vd 5 1 r S 0a
- (
E 1d ta n M st eS 2 = I T T A 0 nY
] /
A 4 kP iA eS vL
- rE MA v
0 H aR 3 : i 2v 4 v 8
.-v 1
g'. q 0 x t e 2 r u g i
/ F s
F f.I 0 gl 1 o o 1 e, 9 ,5 q* qo e4 4 ei-8 tk 3 c ~2 y N '
0 0 1 p 0 s 3 n _
* - o i
y t . td c no i ih n d at o e t e i e 0 r . rM t s 9 P _ e c a _ cc i C l e ni d . Ut e y d L. a r t o
. dt P i v M nS -
a t 1 t i 0 w o 1 .
~
ao . s sl
/ tl v n sF a
ai e e c ,L
.A L.
DP R S Ml d c a ei
- pt
[ f 1 0 6 E ai c, r sC d
/'. -
5 - 0r 1 a t n d -
/. /
0 5 ) S ( sa et TS nA - eY
/
- E M
kS iP I vA T rL
- 0 aE MR 4
/ -
5
~-
8 1 0 X 3 I
/ e
' r u
g
/ 0 2
i F 0
) 1
/
'0 0
1 a 3 h B E UC mME
,;T
/~N l
0.fr.19 l Justify that break-flow models bound all SBLOCA cases ecpectEd. A.II.19 RELAp5YA contains two sets of critical flow models. The standard critical flow models are used for best-estimate calculations of the break flow rate for liquid, two-phase fluid and steam conditions. The ; Moody critical flow model is used for licensing calculations of the break flow rate for two-phase fluid and steam conditions as required by Paragraph I.C.1.6 of Appendix K to 10CFR50.46. All critical flow models account for the break size and'1ocation through input data. Separate discharge coefficients can be entered as input data for liquid and for two-phase / steam discharge conditions. All critical flow models account for the upstream fluid conditions that feed the break. The special case of break flow out of a horizontal pipe containing stratified flow is discussed in the answer to Questions Q.II.7 and .O Q.VI.7. All models have been favorably assessed against a wide-range of conditions (e.g., break size, pressure, quality) reflected in the six tests identified in Table 11.19-1 that is attached. Therefore, we believe these break flow models bound all LWR LOCA cases that could reasonably be postuinted, including small break LOCAs. l
-146-
--__-.._.-__________________J
1 0
. m JABLE II.19-1 .,,
Assessment of Break Flow Models critical Flow Regima Test Facility steam . Liquid Two-Phase and Number R R, M R. M Marviken Test 10 R R CE Level Swell Test 1004-3 R R. M R. M TLTA Test 6425/R2 R R TLTA Test 6426/R1 R R R TLTA Test 6432 R R R LorT Test L3-6 Wote: R = RELAPSYA Standard Critical Flow Model M = Moody Two-Phase Critical Ficw Model o. l' O -147-i
- 1. ADDITIONAL OVESTIONS THAT ARE CONCERNED WITH SEVERAL AREAS 7
.i 0.I.2 During depressugization, the noding detail affectsithe pres [ure ' history .
af ter the pressurizer oppties. The highest temperature node containing liquid in the upper head, upper plenum, or top of the hot-leg bend (or top of a steam generator U-bend) will determine the saturation pressure at Which the fluid begins flashing. If fluid from large regions is combined in one node, the flashing will be determined by the average temperature of the large node rather than the hottest temperature that would be modeled using rmaller nodes. Additional nodal sensitivity studies are required to confirm the adequacy of the hot region nodal j detail to properly account for flashing during depressurization in a reall break LOCA" (Page 'VIII-43 of Reference 2). Clarify What sensitivity studies have been done to define the noding required.to ensure adequate detail will be modeled in SBLOCA calculations. A.I.2 A O This question refers to the noding schemes used by combustion Engineering Inc. for their LOCA analyses (Reference 1.2-1). As is seen in' Figures IV.2-1 and IV.2-2, Yankee plans to use a detailed nodalization of the plant. This detail has been utilized to represent the plant configuration accurately and to capture the important aspects of the transients. A very detailed noding of the upper head and upper-plenum region is employed. With this kind of noding, we feel that the flashing of different regions following a LOCA will be accurately modeled. As such, we cannot justify,the need to perform noding sensitivity studies to assure the adequacy of noding detail in the upper portions of the reactor vessel. Reference (I.2-1) CEN-114-p, Amendnent 1-P, " Review of Small Break Transients in CE NSSS," Pages 3.3-29 and 3.3-30," July 1979. O
-148-
~ _ _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ _
0.I.3 Appendix K. Requirement 11.2 states: **For each computer program, solution convergence shc11 be demonstrated by studies of system modeling or noding and calculational time steps (P, age 274 of Reference 10). Clarify what noding sensitivity studies have been perfonned.- A . I . 3. It has been pointed out in A.I.2 that we intend to use a detailed nodalization of the ant in our LOCA analyses. This detail has been utilized to represent .he plant configuration accurately and to capture the important aspects of the transient. With this kind of noding detail, Yankee does not see the need to perform noding sensitivity studies to demonstrate computer program solution convergence. 0.I.4 Clarify what time step sensitivity stuttes have been performed. B A.I.4 Time step sensitivity studies will be performed on a plant-specific model and documented in submittals. 1
-149-l 1
l- i l O l l l l RESPONSE TO 15 QUESTIONS ON RELAPSYA AUGUST 15, 1985
]
.I. : CONDENSATION HEAT TRANSFER AND NONCONDENSABLE CASES 1[' N 0 ; T .11' G
Clarify the assessment of the vaporization model given in Equation 2.1-20 on Page 34 of Reference 10 for cases with noncondensable gas present.-
'A.I.11 Justification-is provided in the answer to question Q.I.3 and Appendix-A.I-3 to exclude noncondensable gases in our small break LOCA analyses. -However, the: presence of a noncondensable gas will tend to reduce the vaporization rate given by the current model in RELAP5YA.
This is further explained.below.- The current vaporization model in RELAPSYA is given on Page 34 of Reference 1.18-1 as follows:
= C (G ,+ C ) [(I-I -I,)(I,-I)/fy (1)
Where:
~ ~
C = 6.4517175 x 10~ (see m Pa ), empirical constant
~
G = mixture mass flux (Kg m sec~) G = 3500.0 (Kg m
~
sec~ ), empirical constant P = total pressure (Ps) I = gas phase static quality 1, = equilibrium static quality I = noncondensable gas static quality I, = 1.0 x 10~ , empirical constant jy=vapordensity(Kgm~) Let us examine the variation ofgf with an increase of nI for the case where G , P and I are 'neld constant. From the Gibbs-Dalton a model for a gas mixture, the vapor density, , is given by: O .,. I
.c ,_
v" TOT v gas " TOT
^
n gas Since I, is a very small constant, then the quotient given by Q (K-X,-Ij/fy in Equation (1) is nearly constant. Therefore,. the variation of with an' increase-in I depends primarily upon
- the static quality difference (1, - I). The answer to Question Q.I.8 shows that this' static quality difference decreases to zero as I increased toward I. Thus, the calculated vaporization rate will also decrease toward zero as I, increases toward I.
9.11.0 1 Clarify how the. interphase. area is included and how the effect of area is modeled in changing from nonstratified to stratified flow. A.I.12 Both the vaporization and condensation models are empirical and do not include the interphare area explicitly as a variable. The vaporization model is based on data from depressurization experiments. . The-condensation model is geared mainly to provide rapid equilibration of , phase temperatures. 'Hence, there is no need to include the effect of
. changing interphase area in these models.
0.I.13 Clarify why the rate should not increase when the liquid decreases below the temperature that would make,the equilibrium quclity go to zero. A.I.13 l l The interphase mass transfer models in PELAPSYA are' written as the l-product of a conductance tem and driving force. Conceptually, the mass transfer rate should increase as the driving force increases. In RELAPSYA, the driving force is the quality difference, I, - I, but the equilibrium quality is limited to values between 0.0 and 1.0. i
1 l i Hence, if the true equilibrium quality becomes less than 0.0 (subcooled, j liquid) or greater than 1.0 (superheated vapor), it is reset to 0.0 and f p/
\.,_. 1.0,'respectively. Consequently, the driving force for interphase mass transfer is also limited. However, the conductance term in the i RELAPSYA models for mass transfer is always large. Thus, the limitation on the driving force has little impact on vaporization and condensation' rates. i 0.I.14 I
Page 32 of Reference 13 stater: "The main shortcoming of the air modeling in R5 is the lack of any modifications to the heat / mass transfer and drag packages when air is present. This is not believed to be a real problem for the dt'as packages. But the presence of noncondensables in the gaseous phase can change the heat and mass transfer rates substantially..." Clarify how the heat-transfer correlations (and in particular the ones for condensation) are modified in AELAP5YA to include the effect of noncondensable gas for flowins and n nonflowing applications. U - A.I.14: Reference 13 mentioned above is listed belou as Reference I.14-1. The heat transfer correlations in RELAP5YA are not modified to explicitly necount for the effect of noncondensable gas. However, the local state properties will be affected by noncondensables and have some impact on the heat transfer calculation. This effect of noncondensables on heat transfer does have the correct qualitative trend. It should be noted that the effect of noncondensables on heat transfer is considered to be insigni.ficant for PWR small break LOCAs. This has been addressed in the response to Question Q.I.3 (see Appendix A.I-3, Section 4.0). l' Reference (I.14-1) J. A. Trapp, " Descriptive Evaluation of RELAPS," EpRI-NP-2361, c (April 1982). ( l
1 1 1 0.I.15 ()- . Clarify if the resistance of the noncondensable gas is included as it builds up near the condensation interface for nonflowing situations. In particular, the use of a simple multiplier may not be appropriate if I the heat-transfer process becomes diffusion limited. A.I.15 4 i i t The resistance of the noncondensable gas as it builds up near the vapor-liquid interface during condensation is not explicitly modeled in RELAPSYA because it is not considered significant for PWR small break LOCAs. The justification for neglecting the added resistance to heat transfer due to a noncondensable gas layer is provided in the response to Question Q.I-3 (see Appendix A.I-3, Section 4.0'. 0.I.16 Examples and experimental bases for the heat-transfer-degradation functions that might be used should be described and justified. C.- A.I.16 The degradation of heat transfer due to presence of noncondensable gas is considered insignificant for PWR small break LOCA condition and is neglected in RELAPSYA. The justification is provided in the response to Question Q.I-3 (see Appendix A.I-3, Section 4.0). 0.I.17 Clarify and justify the temperature difference that multiplies the hea' . ansfer coefficient when the atmosphere contains a noncondensable gas. 4 l
I i
'A.I.17'
- v The temperature difference that multiplies the heat transfer-l coefficient is the difference between the wall temperature and the local saturation temperature. When noncondensable gas is present, the saturation temperature is based on the partial pressure of the vapor and will be lower than the saturation temperature corresponding to the total pressure. - Thus, the temperature difference between the wall and the bulk fluid will'be correspondingly affected (decreased for condensation, increased for vaporization) and is the appropriate temperature difference to be used in estimating heat transfer rates.
This situation is illustrated (for condensation) in Figure I.3-1 of Appendix A.I-3 in the1 response to Question Q.I-3. However, as also discussed in the response to Question Q.I-3, we propose to neglect the presence of noncondensable gas in PWR small break analyses. 0.I.18 Clarify how the condensation is modeled at the liquid interface in the pressurizer with and without a noncondensable gas present. A.I.18-The pressurizer is modeled by RELAPSYA using the normal hydrodynamic components available in the code (SNGLVOL, BRANCH or PIPE components). In these components, RELAP5YA uses the condensation model described in Rquation 2.1 '21 of Reference I.18-1. The effect of noncondensables on 1.he condensation rate as calculated by'this model is discussed in the response to Question Q.I.8. Reference (I.18-1) R. T. Fernandez, et al., "RELAPSYA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume I," YAEC Report YAEC-.1300P, October 1982.
LJ.19
. ,y-( j_ Clarify how the condensation process is modeled for pressurizer sprays ;v with and without a noncondensable gas present.
A.I.19 The pressurizer spray is modeled in RELAPSYA as a liquid-filled
-junction injecting into volumes comprising the upper. regions of the a pressurizer. The effect of noncondensables on the condensation rate will be as given by the model described in Equation 2.1-21 of Reference I.19-1 and as further discussed in the response to Question Q.I-8.
During most of the small break analyses the reactor coolant pumps are expected to be tripped. The pressurizer sprays will not be operable when the reactor coolant pumps are off and the condensation process associated with the sprays will not be effective. Reference (I.19-1) R. T.>Fernandez, et al., "RELAPSYA - A Computer Program fu LWR
' System Thermal-Hydraulic Analysis, Volume I," YAEC Report YAEC-1300P, October 1982.
O.I.20 Clarify how the condensation process is modeled between the levels of a stratified flow with and without a noncondensable gas present. l A.I.20 RELAPSYA does not explicitly account for the geometry of the flow ! pattern in calculating condensation rates. The condensation model is , I described in Equation 2.1-21 of Reference I.20-1. The effect of j noncondensables on the condensation rate is disetssed in the response f to Question Q.I.B. i I _ _ . _ _ _ ________._.__________________._._________J
i!
- Weference -
(I .' 20-1) R. T. Fernandez...et al.. "RELAP5YA - A' Computer Program for LWR-System Thermal-Hydraulic Analysis, Volume I," YAEC-1300P, October j
~1982. i
,i I
o, O - O ->- t
4 I V. SYSTEMS VERIFICATION AND OTHER EXPERIMENTAL VERIFICATION 0.V.8 Figure 5.3-5 of Reference 12 shows that the break flow was overpredicted from 200 to 800 seconds, and yet Figure 5.3-9 of Reference 12 shows the change in calculated total mass to be less than the data from the same period. Also Figure 5.3-3 of Reference 12 shows the calculated pressure agreed almost exactly with the data during this time period so the HPIS flow would also be expected to agree very closely with the data. Clarify why then if the break flow was overpredicted during this period, the change of the total mass was less than the data. A.V.8 There appears to be an inconsistency in the reported data. Table V.8-1 presents system mass values at.200 and 800 seconds for both the LOFT L3-6 data and the RELAP5YA calculation. In addition, the average flow rates in a,nd out of the system are also listed. Using these values, a mass balance calculation was made to check the system mass numbers presented in Figure 5.3-9. Comparison of the two system mass numbers shows that the RELAP5YA calculation agrees with the system mass obtained from a mass balance while the L3-6 data does not. The
' difference between the two system masses for the L3-6 data is about )
i 1,300 lbm. Taking the worst combination of the measurement uncertainties.only accounts for about 350 lbm (Uncertainties: j Inventory i 110 lba, break flow i 15%, HPSI flow i 0.044 lbm/sec, RCP injection i 0.035 lbm/sec). Therefore, there is an unexplained inconsistency in the reported test data. I 1 i t-i 3
l i j! l lI ' s se e e O ae b b MS l l m0 5 2 3 e0 6 t8 4 0 s yt 4 4 Sa e e sc' c a a ae n b b MS a l l ml m0oa 2 1 e0rB 7 2 t8f s s 7, 0, yt s 5 4 Se a M tnao'n cb a cb a li el el ot s s oc /0 /0 Ce a3 m3 t1 b1 rd oI r I l l l t 6 a 6 a cp 1t 1t am eu 2. T o
- 2. T o
RP 1 m m 8 b b w el cl V o e e O l s2 s2 ._ F /5 /5 R L m5 m5 - B S b b I ll ll A P a a T H 2t 2t o o
- 9. T 9. T m m b b w l l o
l e0 e4 F e8 e8 k /s 6, / s2 a m4 m5 e b b r ll ll B a a
- 8. ot 8. ot 7T 8T s
se ae m m MS b l b l m0 e0 0 3 t2 7 2 s 7, 6, yt Sa 9 8 A Y 5 P O 6 a 3 a LD t A. I E R 9 D L I s 1I
il VI. FLOW REGIMES
'O.VI.8 The rates of phase separation in flowing fluid will determine if steam and noncondensable gas will separate from the liquid at the top of the candy canes for B&W plants or the top of the U-bends in CE and }{
plants. Clarify how the phase separation ).s modeled as the two-phase fluid flows around the high-point bends. A.VI.8 No special treatment is given to the phase separation at the top of the steam generator U-bends. h standard RELAPSYA interphase drag models described in Section 3.i of Reference VI.8-1 are used. Ecicience (VI.8-1) R. T. Fernan6ez, et al., "RELAP5YA - A Computer Program for Light i Water Reactor Thermal-Hydraulic Analysis, Volume I: Code Description," YAEC-1300P, October 1982. (Proprietary) j I 0.VI.40 Clarify and justify the verification provided by systems and experimental comparisons for interruption and restart of natural circulation. A.VI.40 Please see the response to Questions Q.III.1 through Q.III.4 as provided in our July 1,1985 submittal. s _ _ _ _ _ _ _ - - _ - _ _ _ _ _ _ _ _ _ _ . _ . 1
1 i i VII.= CORE STEAM COOLING ~ 0.VII.9' ' l l Clarify the range in conditions for the core level assessment tests. compared with the expected range in SBLOCA calculations. ]
'A.VII.9 Reference VII.9-1 presents an extensive assessment of BELAP5YA compared to many separate effects and integral test results dich address relevant thermal-hydraulic phenomena including core level. The most {
relevant tests for core level assessment are the boil-off tests. The boil-off tests used for assessment were:
- 1. The quasi steady-state boil-off Test 3.09.10I conducted-at Oak Ridge National Laboratory in the Thermal-Hydraulic Test Facility (TNTF). The objective of the boil-off test series was to study the heat transfer and mixture level swell under SBLOCA conditions in s
Pressurized Water Reactors (PWRs); , i l The conditions for Test 3.09.10I are shown below: System pressure 650 psi Mass fit.x 2.19.10 lbm Linear power / rod 0.68 kW/lb l
- 2. The systes; boil-off experiment, Test 6441/6, conducted at General Electric, San Jose in the Two-Loop Test Assembly (TLTA) facility.
1 L The boil-off tests attempted to simulate system conditions which I might occur during a small break LOCA in a BWR, if none of the Emert,ency Core Cooling Systems, including the ADS were available. In these tests, the recirculation loops were blocked off and the liquid inventory was slowly boiled off at a constant pressure and constant bundle power. The power level was representative of decay i l
1 heat in a BWR. The main objective of these tests was to evaluate heat transfer in a partially covered bundle at decay power levels and low flows. l The phenomena and~the range of cono3tions encountered in this test will also be encountered during SBLOCAs in PWRs before the accumulator actuation if the flow from the high power safety
. injection pumps bypasses the core.
The initial conditions for Test 6441/6 are: System pressure 395 1 10 psia Bundle power 250 i 2 kW Initial-two-phase level Bundle top One of the integral tests used in the assessment of RELAP5YA simulates a small break LOCAs with core uncovery Test L3-6/L8-1 was a small break test performed in the Loss of Fluid Test (LOFT) p facility at the' Idaho National Laboratory. Test L3-6, which was a SBLOCA with the pumps running, was extended into a more severe transient (LB-1) which produced core uncovery and heatup. For the performance of Experiment L3-6/L8-1, the LOFT facility was configured to simulate a small break equivalent to a 4-inch diameter rupture in the cold leg of a large (approximately 1,000 MWe) commercial Pressurized Water Reactor. In Experiment L3-6, the primary coolant pumps were not tripped until the hot les depressurized to 311.83 psia. The High Pressure Injection System l (HPSI) flow was then terminated and the break left open. Experiment LB-1 started when the primary coolant pumps were i tripped. When tho maximum fuel cladding temperature reached 600.53 F, core reflood was initiated. l
\
Experiment'L3-6/L8-1 was initiated from primary coolant system. conditions of: I( Hot les temperature 579.1 1 3.24 F' Cold leg. temperature 544.5 1 3.24 F Hot les pressure 2,156.7 1 20.3 psia Intact loop flow rate 1,065.48 1 5.72 lb/see Power level 5011150 Maximum linear heat generation rate 16.06 i'1.12 kW/ft since this is a transient test. the conditions in the test facility varied with time. A detailed account of the integral test conditions are presented in Reference VII.9-1. Reference (VII.9-1) Fernandez, R. T., R. K. Sundaram, J. Chaus, A. Husain, J. N. Loomis,- L. Schor, R.'C. Harvey and R. Habert, "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume III: Code Assessment," Yankee Atoalc Electric Company Report YAEC-1300P, Volume III (October 1982). (Proprietary) l O-E- -- --- - _ - I
K N X.: ADDITIONAL OUESTIONS THAT ARE CONCERNED WITH SEVERAL AREAS JX-
- r. Q.K.7 Since the. release of the version.of RELAPS used as the bases for your small break evaluation model, numerous corrections and updates to:
RELAPS have been made to correct errors and improve the numerics of the-RELAPS computer program. Provide a-description of the quality
' assurance program used by you to incorporate required changes in your.
'EM model-(e.g., corrections of coding errors). Include in this description the~ program to validate your EM model when changes are made.
A:.kl' Yankee Atomic Electric Company utilizes a Quality Assurance Manual.to provide guidelines. to tho' engineers to assure that design changes and
. additions are consistent with safety-and licensing requirements. In particular, instructions WE-103 and WE-108 are the two' guidelines relevant to.the issue of the quality assurance of computer codes and O engineering analyses.
l WE-103 provides guidelines for the performance, review, approval.and control of engineering calculations and analyses. WE=108 provides guidelines for development, acqisisition, modification, verification, testing and approval of computer codes that will be used for WE-103 calculations. Copies of WE-103 and WE-108 are attached. 1 i O
b:3 ( (-* ' d. I: YANKEE NUCLEAR. SERVICES DIVISION FRAMINGHAM, MASSACHUSETTS Engineering Instruction ENGINEERING CALCULATIONS AND ANALYSES TITLE: . 103 INSTRUCTION NO: 0 - REVIEWS (Initial) NE DEPT. E.E. DEPT. QA DEPT. DATE REV. P.E. DEPT.. APPROVAL
' i *r' ? //W Th7 7 BW 3cc ')h k 'L -bGt lw
=
w M M 'Lr Adw 'de s azu 66% Chl 'l,7.r t As li,ap , TbH Bel Jhit /dL.,r. Mg kh c s
-r u VSL 9% -% n. kip ,,%
L 4 ew WJ M Sh %# ,24 , 7 mL .f U jf)Q gnff fy,j . ,;jj, , u - 1 O . i
CALCULATIONS AND ANALYSES e WE-103 1.0 PURPOSE 1.1 To provide guidance for the performance, review, approval, and control of engineering calculations and analyses. 2.0 DISCUSSION 2.1 Definitions 2.1.1 Design Analysis Review - To confim, substantiate, and assure that a calculation or analysis has been perfomed in confomance with the specified requirements. Acceptable methods include critical second - party reviews, comparison to alternate calculations or analyses, or qualification tests. O - 2.1.2 Check - To determine the reasonableness or accuracy of a calculation or analysis by investigation, comparison, or examination. A check is usually perfomed in cases where a calculation or analysis is not related to a specific design or design change. or when a design analysis review has already been perforned. 2.2 Use of this Instruction assures that calculations and analyses are performed in a logical manner, so that a competent individual can review the work without recourse to the originator. t.3 This Instruction is used for the performance of calculations and analyses during engineering design activities, and for the review . of calculations and analyses which were performed by engineering or by outside design organizations. It may also be used when it is requested that calculations and analyses be performed (1) to verify that operating plants conform to their technical specifications or WE-103-1 1
licensing requirements, or (2) in responsa to o reguictory es:ncy, when the action or response involves a change in the technical
/9 specifications or an unreviewed safety question. .U 2.4 Specific calculations and/or analyses occasionally have to be performed in accordance with requirements of codes, standards, design criteria, specifications, procedures, instructions, etc.,
wh!.ch differ from those found in this Instruction. In such cases, originators are directed to identify and reference the non-standard requirements and their sources. 3.0 INSTRUCTICWS 3.1 Calculations and Analyses shall be assigned and identified as follows:
- 3.1.1 Originators shall be designated by a person whose title or function is Senior Engineer or above.
3.1.2 The originator shall establish and maintain a work file with which the calculation or analysis can be identified. tO All pertinent documentation accumulated in the work file as part of the work activity shall be included in the calculation documentation. There is no requirement to maintain a work file once a calculation is approved. l 3.1.3 To assure raady reference and retrieval of work, the calculation or analysis chs11 be assigned a title which will allow the work to be identified with the plant, For l' project, or request for which the work was initiated. example, the title may include an EDCR nusber, Work Order
- rumber, a specification numb' e r, or similar identification, such as " Pipe Stress Calculation for YY EDCR 83-2". .
. 3.1.4 An identifying number shall be obtained from the Document Control Center (DCC).
O WE-103-2 l
\
3.2 The fomat and content of a calculation or analysis shall conform ; to the following requirements: 3.2.1 The title page (Figure 1) shall contain the work activity title; the same of the' applicable project or. Plant;
.j calculation number; eignatures and dates-indicating-performance'(or revision), review, and approval of the work.
'3.2.2 Successive pages including attachments shall contain the .
i calculation number, page number, and date; and shall l encompass the following, as applicable: (see Figure 2)
- 1) A problem description Which includes the objective of -
the analysis or calculation -
- 2) The intended method of solution.
l
.t
- 3) Any references (such as drawings, specifications, or instructions) required to accomplish the activity.
- 4) Sources of equations and the units of parameters shall ,
be identified, except where the equations and units are common engineering knowledge and are thus not taken from a specific reference. l
- 5) Any design inputs used in the calculation or analysis, including specific references and sources.
l
- 6) Assumptions used to perfom the calculations and their sources must be clearly stated or referenced in enough detail to allow competent, independent review.
- 7) The actual calculation or analysis. If a computer ,
code is used, the originator shall provide a statement
^
as to the status of code verification (code Level II j or method of verification) sad that the code is i D 4 WE-103-3 i
i appropriate for the application. See. Figure 3 for guidance. Data extracted from computer codes shall be identified and reference shall.be made to the computer
-(
program used. Whenever it is impractical to include the computer printout in the activity work flie, the originator shall it.clude or reference a listing of the input data with a synopsis of the output in the activity work file. 7
- 6) A summs. y of the results or conclusions. The originator shall be responsible for stating that the objective of the analysis / calculation was achieved, if appropriate, f-
- 9) Review form, if required.
3.2.3 Figure 3 provides guidance for the use of computer codes in calculations and analyses. Depending on the status level of the code, the guidance assures that the code has been 7 properly verified and is appropriate for the application. , If this assurance'is not obtained, the originator is directed to WE-108 for guidance in obtaining, writing, or modifying a code to Level II status. Unverified codes (or portions of codes) and engineering aids (such as calculators, programmable calculators, , handwritten programs, etc.) may be used in calculations and analyses if the design analysis review required by Section 3.3 of Wg-103 will provide the proper verification. 3.2.4 As the guidance given in this Instruction must be general
- enough to apply to many different situations, the suggested format for the calculation or analysis should be followed ,
I WE-103-4 I i
[^ l .*
)
to th2 cxtent that it can be cpplied to th) spscific ' activity. The content may be changed as necessary to give sufficient information to accomplish the desired objective. 3.3 Calculations and/or analyses shall be reviewed and approved according to the following Instructions: l'
- 3.3.1 Reviewers shall be designated by a person whose title or function is genior Engineer or above.
i 3.3.2 If a calculation or analysis is related to an initial dseign or design change, a design analysis review (see Definition) shall be perfomed by an individual other than ' the originator or his supervisor. The depth of review shall be determined by the reviewer, considering the safety significance, complexity, degree of standardization, state-of-the-art, and similarity to previous work. When altemate calculations are performed, the appropriateness of assumptions, input data, and node or other calculational method used in the original calculation shall be addressed. O 3.3.3 The review of mechanical piping and structural analyses shall be performed using the. attached checklist, Table I. The items on this list shall be addressed in the review. 3.3.4 All checks and reviews shall be documented as follows: ! l
- 1) If a criti, cal second-party review was performed in ,
accordance with section 3.3.2 above, and the 7 reviewer's comments are readily resolved with the originator, the reviewer shall document his review by '
- signing and dating the cover sheet.
a O WE-103-5 i j
- d I
)
2)- If extensive or major. connents are involved, or if a method other than critical second-party review is involved, the results of the review and the method , used shall be documented on Fom WE-103-1 or by a 7 Standard Nemorandum, Ws-005, addressed to the ' L originator and included in the work activity file. After resolution of comments, the reviewer'shall sign " and date the cover sheet. l i 3.3.5 If, af ter consultation with the originator, a reviewer deter 1aines that any portion of a calculation or an analysis is unacceptable, he shall issue a Standard Memorandum to the individual who assigned the task, explaining his objections. The issued Standard Memorandum shall be included in the work activity file. 3.3.6 The originator shall incorporate or resolve all coments by reviewers. In the event resolution is not possible, the ! appropriate Manager (s) shall provide resolution.
" Resolutions shall bb documented on the review form, or by l7 memo, and signed by the involved individuals.
3.3.7 If a calculation or analysis is not,related t to an initial design or design change (i.e., analysis of plant perfomance), then the depth of review shall be determined by the reviewer to the degree necessary to address any safety significance, as discussed in section 3.3.2. 3.3.8 If a calculation or analysis has already had a design analysis review puformed (such as work done by an outside design organization), check or review is not required, but may be performed. If the reviewer does perform a check or review, review of his work is not required. , i O , WE-1C3-6 l t { i _ _ _ _ _. I
3.3.9 Upon completicn of the r; view pracoss, the esiculction er analysis shall be approved by the originator's Manager. .' - 3.4 Calculations and analyses shall be revised according to the following Instructions: 3.4.1- Revisions to approved calculations and analyses shall be reviewed and approved by the same discipline (s) or group (s) that performed the original review and approval, unless otherwise designated by the Project or Department Manager. The designated organization and all reviewing organizations shall have. access to partin.ent background information, have demonstrated competence in the specific area of interest, and have adequate understanding of the requirements and intent of the original calculation or analysis. 3.4.2 Calculations or analyses which have been transferred to DCC as design records may be revised or supplemented, provided the revisions are reviewed and approved in accordance with this EI. The revision shall be transferred to DCC in
- accordance with DCC transfer requirements of WE-002.
3.5 The distribution of approved calculations and analyses shall be as follows: i 3.5.1 Forward to DCC in accordance with the internal interface and DCC transfer requirements of WE-002.
- 3.5.2 Copies of completed work, or where more appropriate, a standard Memorandum describing the work and results, should be sent to those who:
- 1) contributed to the work content. .
O e O. WE-103-7 i
- 2) are affected'by its results.
O 3) are responsible for any implementation necessitated by the work. If a atmo is written, it shall be included in the work file. 3.5.3 Revisions shall receive the same distribution as the original calculation or analysis. 3.6 Deficiencies with this EI shall be reported according to Instructions found in WE-001. 4.0 RECORDS 4.1 Calculations or analyses which are engineering design records, shall be controlled by DCC.
- 4.2 calculations or analyses which provide technical backup for answers
- to NRC questions or verify plant operation within licensing requirements are not design records, but shall be forwarded to DCC for retention.
L ' l O eindicates guidelines which are not QA requirements. WE-103-8 ;
a i PAGES I PAGE 1 GF l
-I
- e IMS No.
( RECORD TYPE W.0./P.O. 30. I i YAIKEE ATOMIC RLECTRIC COMPANY ANALYSIS / CALCULATION FOR TITLE CYCLE PLANT CALCULATION NUMBER I l PREPARED BY/DATE , REVIEWED BY/DATE lAPPROVEDBY/DATE ( i !
' ORICINAL ,
I l
, REVISION 1 3
jREVIsIoN2 '
$ REVISION 3 .
l i KEYWORDS I 9 8 9 FIGURE 1 WE-103-9 t
l FORMAT FOR CALCULATIONS AND ANALYSES
/~~g us)
- 1. Calculation No.
- 2. Structure / System / Component to which calculation applitet 3, Froblem Description Including objective and intended method, results of literature searches and other background data
- 4. Details of Analysis:
Design Inputs, including sources Assumptions, including sources Appropriate figures, sketches, etc. Calculation / Analysis Identification of computer codes used
- 5. Results/ Conclusion
- 6. References r
( Including original work request, correspondence, texts, pertinent memos, drawings
- 7. Attachments 7
Including (as applicable) review forms, computer output, technical information, associated correspondence, and memos 9 FIGURE 2 r I
\s. ,
WE-103-10
\
( .
- d e
d o e c t t e
. e r ,$
d r o f0 e at f t n i s se t e ec td t e t p
?
r 5 d r e eg WQ_ e d pe s e et . t n g d t e i l ti
.l e ne e i
.se
- i. i
_ i t f i e n d el ed w .y n i. d d o i ene et ene t _t m
. p e
t id e e t s ee h c d e t e
- e p s e ec ve eosp t s l
o d t ar be e oep h 3 e a v s ae t t
'l e e s 3 d , ft it S ot e e 3 t g ed e t e c
_ n e D C e n ie e b mu t n nt es ti sh il ee st r
?
I dey l s e rb d n e e n t t F e ss e lt e e O f c e el l e e fed e3 o o do we f ch e e v db e et ne rer?i3 , e0 n1 S.
. S ,
t t s s , d el t l e t e Y t ofe tot fsd s? s e hfe t g t ye sE iW _ 9 3 1 T E L ed c e o r f l dl ei o tt ei
. d e
o 1 ediee derl s oietu cfi et w v e ws nset otts Cl uedse E D dt eh t l pt e e sh c r e oi 5 i t t pe*t dupn sg e cs e le V e T r o eeel A T ats s ** A h t m rehv wt er h e S o y t
. t t
e t t b e t e d e , e d e d i d a C e . " e C C C t t i T J] _ I I l t e r e f t e I l e v v v e e ta et
- l. I s l -
t e d v ys e s s t e T au m e e d ' eb o d ot c Co - e n - h t s I l
. d' e
e f _ e c 7 .
%b nt ee dn l o oi 9
1 t 1 r u
?
es d r? ef et on gt e g e s oet s p v e e f o e e v ers yet Y ct a eci T o s t he l. ti ei e l s l ne t ,l p _ s d ev p _. U e ,o e iw9 e ee n i 7 r ,e e p .h h a t t j] s
?
et ? ee se t i st m siu t t sfd e y tel sehs ete s e e v s e emtt f. Y re are e eel - d th y o otdt c o ee e Cf sf - ouas h t - e s te f e O a e es
? .
3 8 1 dn. 3 1 eeE 1 etW t - wtee s s 3 e U 01 eet e C hb ar t - t c f E ol e e st at Oyi e dat o cn. ca ie r s r e a. ltIf
- _ _ _ - . . _ - _ . _, __ _ _ _ _ _ _ = _ -___ ______ _-
TABLE I MECHANICAL ANALYSIS VERIFICATION CMECKLIST 'O' V I. MATERIAL PROPERTIES
- a. Are all material properties used in the analysis consistent for the material required and the design temperature
- b. Are material property cources adequately referenced?
II. MODEL GBOMETRY
- a. Are lengths and coordinates consistent with reference documentation
- b. Are the elements employed consistant with the structure modeled?
- c. Are eccentricities properly accounted fort
- d. Are boundary conditions or support functions adequately represented?
- e. Are units consistent?
- f. Is mess / weight distribution consistent with the type of analysis performed?
III. ELEMENT PR0pERTIES *
- O a. Is orientation of elements consistent with structure geometry?
- b. Are cross sectional properties consistent with dimensions or pipe schedule?
- c. Are all necessary properties either calculated or referenced?
- d. Is the derivation of " equivalent" substructures clear and done properlp Have " cookbook" formulas been used correctly?
- e. If stress intensification factors are used, have they been properly calculated or referenced and appliedt
- f. Have all model calculations been checked for calculational accuracy?
. g. Has transformation of neutral axes been properly donet
- h. Are all assumptions of rigid sttuctures proper? ,
9 O WE-103-12 j
i IV. LOADING
- a. _ Have all required loadings on the system been considered (pressure, accelerations, thermal, dead weight, etc.)? Are loading combinations done properly?
- b. Are ankhor' motions properly accounted fort
- c. Daes the loading used consider adequately:
- 1. Locations (building, elevation, etc.)
- 2. Design loats required (FgAR, envelope, etc.) '
- 3. Dynamic amplifications for static analysis (resonance, '
multimode effect)
- 4. Desping consistent with FgAR or regulatory guide standards 5.- Uncertainties (i.e., peak broadening)
- 6. Required frequency content
- d. Is the source of all loadings properly referenced?
Y. MSICW CALCULATIONS
- a. Are procedures followed suitable to reqe.irements of analysis?
- b. Have all hand calculations been verified? ,
- c. Are computer codes uniquely written for calculation properly verified?-
O- d. Are all computer codes employed familiar to the verifier and employed in a proper manner?
- e. Do all calculation results seem consistent with the problem geometry and the input loading?
VI. DESIGN CRITERIA
- a. Is the acceptance criteria consistent with required design codes *
- b. Are required design codes consistent with plant FSAR requirements?
- c. If current standards / criteria are to be followed, are up-to-date ;
documents being used? i
- d. Have any deviations which could affect acceptance criteria been aseguately accounted for (i.e., anchor bolt spacing)?
- e. Are standards followed well documented?
1 O WE-103-13 l t
.. . J 4
VII. REPORT FORMAT
.e o. Does the calculation comply with the curren'. revision of WE-103?
i
\ b. Is the calculation well Scumented and traceablet
- c. Are all changes to exis, & design or new designs requi.ed documented in one summary locationi
- d. Are the major assumptions stated? Does the verifier concur that all assumptions are reasonable for the analysist 1
O . O 4 9 e n WE-103-14
\
.* e l' + l l: FORM WE-103-1 Revision 1 REVIEW FORM I- .'
CALCULATION No. l TITLE: i 1 b Mechanical Analysis Verification Checklist was used. C000 TENTS RESOLUTION 4 r J Reviewer / kate O t
,7_
!(
..f L/ :
YANKEE NUCLEAR SERVICES DIVISION FRAMINGHAM, MASSACHUSETTS Engineering Department Instruction 2 COMP 1TTER CODES TITLE: WE-108
' INSTRUCTION NO:
o - REVIEWS (Initial) APPRoyAL DATE REV. P.E. DEPT. N.E. DEPT. E.E. DEFI. QA DEPT.
~*
($b /1 WP vfsH i e b .4 *'
, 9 2
3 4 5 6 I l h O
COMP"TER CODES I\. WE-108 1.0 PURPOSE 1.1 To provide guidance for development, acquisition, modification, verification, testing, and approval of computer codes that will be used for WE-103 calculations. 2.0 DISCUSSION 2.1 Engineering Instruction WE-103 requires that computer codes used in calculatiota and analyses have been verified and are appropriate for their applications. This Engineering Instruction (EI) provides the requirements to assure proper verification and documentation for codes develope 1 within YNSD and for those obtained externally, and to assure proper controls fcr the use of computer codes. O This Engineering Instruction provides instructions and guidance in meeting the procedu.*al requirements. Additional guidance for complex tasks is provided in YAEC-1336, the systems Development Stancneds Handbook. This handbook contains very detailed instructions for all aspects of computer code development, testing, and documentation. This Engineering Instruction (EI) provides a summary of these instructions to provide guidance in meeting the procedural requirements. The handbook along with ANS 10,4, Guidelines for the Verification and Validation of Scientific and Enaineerint Computer Prograss (and other referenced guidance
. documents) should be used if more detailed guidance is required.
4 e w O WE-108-1 i L_ __ __ _ _ _ _
E-I6 2.2. Computer code status categories are defined as follown 1
/ j
' A.4
- 2. 2.1 - Level 1 - Computer codes which are in process of being developed, modified, or tested. This includes codes obtained externally which have not been verified internally for a given application.
2.2.2 Level II - Computer codes which have been verified and have documentation to demonstrate verification and to identify the scopa or limitations of code usage. l f Verliication is the act of confirming the methodology (logic) used to calculate or process data is correct per the specifications and requirements stated. 2.2.3 Level III - Computer codes which were previously Level II and then superseded. Documentation for these codes identifies known restrictions and options that are no longer usable. Level III codes may be used in conjunction with their previous usage provided their use is approved , by the Project Engineering Manager or Group Manager. 2.2.4 Level IV - Computer codes which are out-of-date, inappropriate, or-no longer applicable.and are kept for historical purposes. Such codes should be stored in l source coce format. 2.2.5 Computer codes used prior to January I 1984 may be
" grandfathered" to Level II or III on the basis of f previous satisfactory usage, provided the documentation
- contains a statement of the basis for usage and verification signed by the Project Engineering Manager er Group Manager, and a description of the scope or
. applicability of the code. Codes not " grandfathered" shall be assigned a Level I or Level IV status.
'O, W-108-2 i
L l' 2.3 The use of this Engineering Instruction is not required for code verification if an existing code or program can be verified through the normal design analysis review required by the use of WE-103 for calculations and analyses. 3.0 INSTRUCTIONS 3.1 General 3.1.1 The intent of this Engineering Instruction is to assure that computer codes are obtained, prepared, and used in a systenstic manner, properly verified, and provided with i documentation to demonstrate verification'n, and allow the scope of applicability to be detetuined for future usage. Note: Applicability applies not only to technical scope; even if technically correct, consideration must be sf%3n to other restrictions such as plant-specific ! or non-generic limitations. 3.1.2 The overall process for the development and use of computer codes is shown in Figure 1. The basic requirements are delineated in the following step-by-step instructions. Guidance for accomplishing these requirements is provided in Table 1. The cognizant engineer shall determine the extent to which this guidance is applied for each application dep'nding e on safety significance, conqplexity, degree of standardization. state-of-the-art, and similarity to previous work.
. 3.2 Initiation (see Table 1)
~
3.2.1 When it has been determined that there is a need for computer code capabilities presently unavailable on the system as Level II, a Code Cognizant Engineer (CCg) shall be assigned by the project Engineering Manager or Group Manager. O WE-108-3 I i 1 - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ J
I I 3.2.2 The CCE shall' perform an evaluation in which he will l ,A) . document the problem in a Statement of Problem and describe the requirements that the application aust satisfy in a Requirements Specification. From this evaluation the CCE will determine if an existing code should be modified or a new code should be developed. l 3.2.3 The CCE shall obtain authorization to acquire a code or to have a code developed or modified in accordance'with the 1 steps outlined in Technical Administrative Guide No. 9, l
" Policy for Control of Computer Hardware and Software i Expansion".
3.2.4 If purchase of an existing code is authorized, the CCE shall initiate the purchase of the code in accordance with instructions found in WE-205. 3.2.5 If the CCE is authorized to prepare or modify a code himself, or if a datailed specification is required for acquisition of a code, he shall proceed to Section 3.3. 3.3 Development - Level I (See Table I) 3.3.1 When authorization has been obtained to prepare or modify a computer code, the CCE shall prepare a design specification. He may request assistance from the Corgputer Services Department (CSD) to the extent he deems necessary for the application. 3.3.2 The specification shall be given a peer review or
' walkthrough by entineer(s) in the same discipline as the CCE. Reviewers are appointed by the CCE's manager or Lead Engineer.
The CCE shall revise the specification to L' arporate (or 3.3.3 resolve) the comments of the reviewer (s). The final O version of the specification shall be signed by the CCE, reviewer (s), and the CCE's manager or Lead Engineer. WE-108-4
I f 3.3.4 If.the cp cificcticn is to be us:d fer purchas3 of a i l' computer code, the CCE shall initiate the purchase of the code in accordance with the instructions found in WE-205. L ' 3.3.5 The CCE shall proceed with the design, documentation (see I Figure 20, and testing of the code. The CCE may request the services of CSD to the extent he deems necessary for the application, as follows: 1 1 1 (1) Prepare a standard useorandum (or service Request) { from the CCE's Department Manager (or Pro. ject j
- Manager) to the CSD Manager, providing the scope of services requested and attaching the specification.
(2) Notify the Software Control Library (SCL) of the planned code addition or modification by filling out the appropriate forms according to CSD procedures. (3) During the development and acceptance of the code, the CCE shall provide technical assistance for CSD as requhsted, and shall participate in the testing of the code to assure that the requirements of the specification are met. 3.3.6 Final design review and femal approval to Level II shall be as follows: (1) Final design and testing documentation (see Figure 2 for documentation contents) shall be independently reviewed by an engineer in the same discipline as the CCE using the review guidelines set forth in WE-103 for the review and documentation of review for calculations and analyses. If the code was obtained from an external source, the 4 CCE shall determine the extent of testing and review required depending on the amount of verification and O WE-108-5 1 k
r 1 1 i \ \ I l validation supplied by the vendor, whether the vendor
]
l D has an approved quality assurance program, the j I complexity of the design, and the safety-related _ status of the code. The CCE shall resolt'e any verification and validation deficiencies in the vendor's code and documentation. The resolution shall be documented. l' (2) The reviewer shall document completion of 'the review by signing and dating the cover form (see Figure 3). (3) Upon completion of the review process, the documentation package shall be furwarded to the CCE's manager for approval and signature. Completion of the cover ferm constitutes final engineering cpproval and designates a code as certified for Level II usage. (4) Completion and return to CSD of the appropriate SCL forms designates a code as ready for inclusion in the (
- Software Control Library.
(5) The CCE shall forward the approved documentation package to DCC in accordance with the internal interface and DCC transfer requirements of WE-002. 3.4 Usare - Level II 3.4.1 Codes that have been certified as Level II may be used in accordance with WE-103 for design calculations and analyses or other work requiring cuality assurance. 3.4.2 If a code certified for Level II usage replaces or revises , an existing code and the CCE determines that continued use
- of 'the earlier code is either inadvisable or unwarranted, I
he shall contact the Software control Library to request archival status for the application. O WE-108-6 i
l The CCE shall coordinate the archival procedure to ensure that there will be no impact on any other group or
- .g Q department, l
The CCE chall be responsible for following the usage of ! 3.4.3 those codes which he verified for Level II usage, and he shall be notified according to CSD procedures of any discrepancies in such codes. (1) Note discrepancies.
- 4 (2) Document notification of other users.
(3) Acquire inforination regarding the impact of these discrepancies. (4) Assess impact. (5) Accumulate all records in a program file. 3.4.4 If it is determined that modification is necessary to resolve discrepancies and that futurir use and availability I of the code is desired, the CCE shall implement the change process by returning to the initial instructions of this procedure and following through where applicable. 3 . 5, Levels III and IV 3.5.1 Codes assigned archival status are defined as either Level III - Controlled Use or Level IV - Archive. (1) Codes designated as Controlled Use status may be . accessed with Group Ilanager or Project Engineering Manager approval. (2) Codes designated as Level IV have Archive status. O WE-108-7 l 1
4.0: arcoRos' y j 4.1 .When computer codes are used in _ calculations and analyses.: the records requirements of WE-103 shall apply.
~
i. 1
'l O
lft , e 9 O WE-108-8
\
se ,eler.est et user l 1._ e i.iu.i s,e if n.ii . p e Stepe. Schedvie. luess s j e New Or modify 4 -
\v/ eseige Code f.v.f aa.t hat.=*r ,
1 I sese rces and ' Authertsetten To beveler 5 Avallette _
.. . . . . . . . . . . . . . . . . . . . . . ~ . - - - . . ~ ~ ...*.~~"*"***"***"***"*"
LEYLL I <r g(VCtDPNPT g g bevelop Specifttetter. l 1 l leftist aseten bewiew l - l Soternal Sevree . 8*.meuse Develsoment i
/ \ ' I l Street see l l heJtfr Code Baeede tevelop/noitt, Dw .sst.ts T.
. .; . see,. n ..: .
see, Teest., . m e. etic.I r . ; i e versfu tier. o heete Sese;. . .. M Ms41ty Spec. l e Apollechtitty ; e n.. ne. 6. . e Scopa + - .-d Docuser.t &cview ' fspel Destga Eevsv.
. ,e. ...
Fermal Aeprewal ' o scope /8tet. Cheate* !
*To towel 11 I
. . . . . . . . . . . . . . . . . . . . . . . - . . . . . . - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~ . - ~ ~ . -
L! s'E 6 II ,,
/ t*t Wt s, .
hap &ece or novase saist6ng Coset - e e I ' l Feise, besse e B
- 4. - e
. kNe Discrepancies l Minor Assuavlste L 4 Potentist beterde Msjer Fatent8si 1'
mettfr Users Aseees
. Get T.11 l*rece sepect mater ...
y
.. .............. _ ... .. - -- - . ,........ ~- . . . . . . . . . . . . . . . .
44 VEL !!! . C4BP790LLED f$t 3 ts LEV 18 3% Onde Artf*Priete . per Applicattant a 9 tan aern',}J, 8 Ses t' 1r Use Wnder Centtelo
- Aethnve Set Te Des 3
a FIGURE 1 ! p WE-108-9 , 8 : 1 l 1 1 k
DOCUMENTATION
- O Purpose / Objective - This is a description of the reasons behind the initial A>
s calculation or revision. Discussion - This is a description of the initial calculation or revision. This should include consideration of scope, applicability, and Itaitations, including plant-specific acceptability. Input and output - This section usually contains the code manual normally provided with the code. As a minimum, will contain an input,and output description and requirements. Programmers Information - This should include a compiled listing of the code on fiche using appropriate update options for cross referencing. Operational Information - This section should contain hardware requirements, job control language (JCL), and update information as applicable. Validation - Items a through d below should be repeated for each test cJse. a) Description of test case b) Test case input with JCL c) Test case output d) Test case evaluation Data Retrieval Information - This should include f. ape number, file numbers, and file names, as applicable, software Control Information - This will include the Software Control Library Forms, if applicable. User Feedback - The intent of this section is to provide a location for user comments if they are accumulated. These comments may be incorporated in following revisions. It is not necessary that this section be reviewed and j approved. S FIGURE 2 WE-108-10
- Q)
4 4 PACE 1 0F PACES IMS NO. g-
'O- RECORD TYPE j s
v.0. /P.O . NO. _ YANKEE ATOMIC ELECTRIC COMPANY. COMPUTER CODE FOR. TITLE _ . CYCLE PLANT NUMBER PREPARED BY/DATE ' REVIELT.D BY/DATE APPROVED BY/DATE 0' is\ aRicINA1 s I REVISION 1 l REVISION 2 l 1 REVISION 3 : I KEWORDS . m 4 9 ! U nouRE 3 1 . WE-108- 11 1
y -. , TABLE 1 Outdance f or the Dewalesment and Das of e-ter f.eder Il l . betailed Guidsete
. (g - Ramuiremente of WE-100 a m rr h idence j
( }l . 4
'3.1<GBWetAL Cognisent engineer determines estent to 3.1.3 taglooer uses owidence okich guldence applies depending eat safety etsnitteenee
,1esit, l
tagree of standardization . 4 Biallerity to previous wort B, j
.3.3 .151TIAT!0N #
3.3.1 appolet este cognLsent m gteser (Cct) , dNS 10.4 prepare statement of problem and 3.3.3 .Sct performe evaluetten seguiremente spee1flestten 80s Handbook enelyse current ersten phase 1 penettiitty Stue
, pages a to 42 dnalyse requirements and propero receasendetlen prepare pre 1Lainary desLen for proposed erstes
- q. Betablish costo, tenefits and schedules propero possibility report Deliverables: .
J o Punettonal specificetten !
.(
- e Generet erstem design
' N,.s TAC Wo. 9 Recommend punhaeo/medificatLst 3.3.3 CCs obtelne authorisetten Begueste Lavelvlag easte of less then $1.000 require only approval by Cg0 geneger sequest for authorization requires the following steps:
Verify staller offerte not esaderway Determine coalistic ersten life prete eeste provide motifisetten to CSD before a new progres er erstem to esquired or developed prepare request to justify coteameadation ettein Doper ment geneger approval . Butalt request to genegensat Review Board in the fe11ewies formata punctional seguiremente Breten description
- App 11estion(s) postures and benefite
- procurement seats support teste f,:
mejor implementatten alleetenes potential future surension WE-108- 12 f
\
e TAs1.t 1 (continued) euidance f or the bevel;;mt and tise of computer Codes .f\ \ ] Dete11ed Cuideete 'V Sumaser Guidante Reeu1reneets ef WE-100 Wr-203 propere supply requisitten 3.2.4 If purchase. Cet initiates process , for purchese ! WE-108 proceed to Section 3.3 3.2.5 If modificatten er Ireseratten. CCE starte development instructions (Level 1) 3.3 DevataputeT . lavsk 1 sDs Mandbook Design detailed Logical erstem: phase 11 - Detail syster 3.3.1 cct tend /or CsD) prepares specificotten WLat functlena will be performed Design pages 43 to ?$ What circumstances the functiets will be performed under What data vill be acted open What processing will be required What output will be produced Design detailed physical erstem plan the Laplementation Determine test and operational environment Revise costs, benefits, and schedules 4 C propere system Specification Report Deliverables: e Design specification o Acceptance test plan o Systes test plan 3.3.3 specification is reviewed Reviewers aseurs t. hat the design specification: Completely defines the proposed code or erstem Will produce a functlenal code or erstem at the estimated cost and othedule . l por guldence in performing specification vs-108 l review one attached tables la Ws.108: { Table 2 eutdance for Deguirements verification Table 3 fasidence for verification of Test plan Table a cuidance for Design verification Table 5 eutdance for verification of the progran Documentation 3.3.3 Spec 4fiestion is revised and approved Reviewers prev 1de comments to CCE for incorporation er resolution l
\
pinal version signed by properor, reviewers. and manager er tend Sngineer A i ) Wr-205 (_/ 3.3.4 If purchase, cct ini' iates t process prepare eupply requisition for purehaee WE-108-13
\
t
. TABLE 1 (continued) i cuidance for the bevelsoment'and use of r.amouter Cedes
- l. l -.
summarr Guidance totalled euleance I ' aneutrements of Ut-los 808 Mendbook
= ' 3.3.5 CCE (and/or CSD) proceeds w!'th Des 15n tede and test ersten Phase 111 -
onceding. developing /sedifying. . .. Programmir4 and documenting. and test!Es. Propere procedures and senduct training Procedures of code Pages 76 to 94 Settsfy operettone requirements Conduct systen test CCE participates la certifisetten of the erstem test results - per sutdance la perferetas preten test see ettethod tables in 5-100: Table 6 h1dence for Seures code verifisetten Table 7 h1dence for Ver1f tsetten of Progres Integration Table 8' eutdance for Program va11detten Table 9 outdance for verification of Test teoutts beliverstdgs e Test d systen/ report
' o User 's manual ,
o Tratsing manual ,j por formes, see rigure 2 WE-108 (1) CCE roguests CSD services prepare standard asserandus WB-005 6 (or Service Request) and attac5 specification Propero appropriate forms: CSD procedures (2) CCE notifies Softwere Control Library of planned additten Wr SCL Forms modification of code. o SCL purport Bequest e Product Installation Request l (if initial Eastallation of code) l e modification Sequest (if modification of previously installed code) e Test assults soport setet signatures of CCE and reviewer are met entered until final testing to semplete e Product lasta1Latien telease (3) CCE essiste in development and Complete procedures and senduct formal 808 Mendbook acceptance process training treintes phase TV-System Acceptant
- Pages 95-99 Conduct acceptance test of operational erstem ,
ces participates la sortif tsatten of
, erstem test results
- Deliverables e Operstlenal erst.en I
/
WE-108-14 i 1 h
TABLE 1 (continued) Cuidance for the Develcoment and Use of Computer Codes
/%
$ ) Beautrements of WE-100 ggumsrv Cuidance Detailed Cuideats l
3.3.6 pinal design review and formal approvst to tavel 11 (1) Independent review by er.gtneer Use reviewing guidelines delineated WE.103 in sene discipline os ccR in Ws.103
- j i
por documentatten contents, see pigure 3 WE.108 If code obtained outernally. CCE Depth of rewtow**dependent~*og;, j determines estent .pf Aesting .. - .- w- 1 required Amounts of verifitotten end validation i eupplied by vender Whether vender has an approved quality assurance program j Ceaplomity of design ' I safety-related status of code . J CCE resolves deficiencies in vendor ; documentation and documents resolutlen ! (2) Beviewer docuannts coegiletion g!gns and dates form. Pisure 3 WE.108 of review (3) Certification of code for CCE's maneser approves documentation 14 vel !! usage package. Signs and dates fors, pigure 3 (4) Designation of code es ready Appropriate forms completed CED procedures for inclusion in gCL SCL poems (3) Documentation package Use transfer steps delineated in vs-002 WS-002 V
/
transf.rr.d to oCC .. 3.4 VgAct . LEYCL !! 3.a.1 Code may be used for design analyses Use rwrieving guidelines de11mented in Ws.103 WE.10J er work requiring QA spyroval o verify that code is Level 11 o Verify that sede is appropriate 3.a.3 tulsting code is replaced CCE initiates archival procese CSD prweedus es or revised CCE decides continued weed of code Centact Software Centrol Library is inadvisable Cecedinate orchival precedure 3.a.3 .CCE felleve usage Wpdate procedures and senduct training 3Ds dandbec'x ) at fe11oe-on 1ecettens phase V - . Implemenietten and Suptc Provide post-installetten support pedes 100 104 prepare for post.Lastellation oudtt Install operational systen at felleeHen 1ecettens per guidance in performing installation. Ws.10e see attached table in WE.100 [ Table 10 Guidance for Verificotten of Installation packase WE-108-15 i
l TABLE 1 (continued)
- I cul0 ente for the Develeoment and Use of Computer Codes l
.[ Rl Detailed Culdswee l
\M Reeutremente of Wr-108 Summarr cuidance !
l sete discrepancies CSD proceduree- I 3.4.3 (continued) - Usere shell notify i CCE of discrepancies Document notification of other weets Cet fv11 laspott j Asease layett Aceianulate recoste la ps , gram file . 3.4.4 CCR Egentifies need to modify blodifisetten to toquis,d: W -100 , Boters to beginatas of Procedure 15-104 and follow where applitable 3.5 AaCMavAL ETAtus . LevtLs Ill AND tv (1) Controlled Use Used under sentrol Aveliable for see with Group tLenager er Project Ragineering kanager approval (2) Archive Archive status (
\
s. 4.0 RECORD $ 4.1 Centrol Seguiremente For codes used in seleulations and We-103 analyses, use records requirements delineated in W5-103 G
\
WE-108- 16
I i I
. ' TABLE 2 f n fggigence f or seavirewts eer!flestien =
I ion .
- 1. Does the Requironient specification can' form to required documentat 1 standardst ,
i
- o. Are all required sections presentf i Does each section contain s!! el the required information? f
- b. .j Is the format as requiredt 1 c.
Does the requirement Specification reflect en understanding of the 'j
- 2. i
- problean to be soleedt
- s. Are the requirements consisteet erith the stateneet of probles for the progreat b.
Are the models shot are specified sprapriate for the probles to De soleedt
- c. Are the numerical techniques that ere specified appropriate for the problee to be solvedt d,
Are the sigoritNns that are specified appropriate for the problem to be solvedt
- e. Move program functions been partitioned in a menner consistent with the problem to be solvedt i.
Will the program os specified solve the problest
- 3. Are the requirements completet
- s. Do the requirements include all functions es11ed f or or implied by i the Statement of problest
\, /
b. Do the requirements identif y all program interf aces and fully specif y required program eehavior with respect to eacht d c, Are ett program inputs identified and described to the essent neede to design the progreat d. Are all required program outputs identified and described to the extent needed to design the progres* into s. Does the specification describe the operational envirotunent ethich the program ewst fitt f. Does the specification include all desired quality requirements, e.g.. requirements for performance, reliability eccuracy. Portability, maintainability, veer friendliness g. Does the specification referener all desired development stenderes' h. Does the specification include acceptence criteria for the progreat
- 1, Does the specification include timing and sising constraints if oppilcable'
- j. Does the specifiestion include required behowler in the f ace of
- improper inputs and other enamelous conditlens?
- s. Are the (requirements correct?
of the prettesi for
- a. Are oli requirements consistent with the statement the progreet b.
Are all requirements consletent with documentes descriptions and into which the known properties of the operational envirofunent program ewst fatt k'E-108- 3 7 i
l l' l l TABLE 2 (continued) (,,I 2 M/ Ouldence for seeutrueents verification
- c. Do interf see regulroments agree with documented descriptions and known properties of the interf acint elementst
- d. Do input requirements correctly describe all inputs whose fo
- e. Do output requirements correctly describe all outputs adhe f.
Do requirements eencerning models, algorithms. and eunerleal
- tenniques stree with stenderd references, where applicablet S. Are the requirements coasistenti
- a. Es the Deguirement specification fece of internal contredictionst b.
Are the models, algorithms, and musaarital techniques that are specified mathematically compatibles c. Are input and output formats consistent to the estent possiblet d. Are the requirements for eimitar or reisted functions cor sistenti e. Are the accurecies required of inputs ~ computations, evtput. etc., compatibles f. Are the style of presentation and the level of detail consistent
- throuthout the documentt 6.
Are the requirements clear and uneAnb!guoust A
- s. Can all requirements be interpreted?
- b. Con all requirements be interpreted in only one wayt
- c. Are the requires. -te organised and presented in a way that promotes clarity (for esampse, use of tables and lists in piece of text. where applicable)?
- d. Are the requirements sufficiently detailed to prevent misteterpretationi e.
Does the specification dif f erentlate between program requirements and other information provided in the specificatient
- t. Are the requirements feasiblet
- a. Are the specified models. algorithms, and numerical techniques within the gtete of the arti
- b. Can they be implemented within the constraints imposed on the system and on the developonent affortt
- c. Are the gustity attributes specified for the program schievable both
- individually and as a grwpt .
d. Are the required functions attelnable within the eva11sble resourcest .
- 8. Does the requirement Specifiestion contain adequate provision for prograr.
talidst10,n and acceptances
- a. Is each requirement testablet
- b. Are acceptance criterie specified?
D WD-108 .18
c---- . _. ._ 1 l ,. l l~ TABLE 2 (continued) l
,c m .
Cuidance for tenuirements verification l l
.(
c'.. Are the acceptance criteria consistent with:
- tesults obtained free staller computer progress
. C1sesical solutions
. : Accepted experimental results
- #melytical results published in technical literature f I
i
. tenctanark problems ,
- 9. Does the requirement Specification eyeid placing undue' constraints en pr* gram , design and implementatient
- e. Does the specification avoid telling how the program is to te designed er implemented?
- b. If it does piece constraints en design er implementation, is there justification f or including such constraints in the requirements?
l N. i 4 0 G O 9 VE-108- 19 i
p TABLE 3
/
Culdence for verificettom of the Test plan
- 1. Is the Test plan description completet
- a. Is the oef twere product to be tested fully identified?
- b. Are the scope and objectives of the Test pasn identifiedi l
q
- c. Are the scope and objectives consistent with the Seguirement Specificattent
- d. Are f.he Beguilement Specification end other documents required for
, Test plan devoteenant and esecution referenced?
- 1. Es the teetles approach to be fe11ewed describedt
- a. Are requirements to be tested identifiedt
- b. Acceptance criteria definedt c.
Art methods veed for indicating complisace identifiedt j
- 3. Are the test probless' definitions adequatet i
- s. Is the bests for selection specifiedt b.
1s the description of Lost programs coupletet
- c. por complex applications. is there et least one problee that erill i provide unequivocs! enslysis of results with minions haan judgementf d, Does esth probles have known and eccepted eveultst g Are the numerical securacy limits of each problee consistent with the ,
e. k software product? f. is the application range of the sof twere product. as defined in the Requirement specifiestion, adequately covered by the set of test problems?
- s. Are s!! types of problems that are identified for solution by the software incl.eded in the set of test problems or escluded for a specified. ve116 reasont
- a. Is each testable requirement adequately covered?
- a. 1s at least one test case proved for each requirements l J
- b. If the requirement covers a range of venues or esymbilities. are test !
cases identified to cover the rente adegustely?
- c. Is the rationale for selecting test cases clear and validt
- d. Does the testing methodelegy establish an unambiguous mean e.
Is the metrix of test cases versus requirements completet
- f. Are redundant test cases oveldedt S. Are the to,st cose descriptions completedt l le each test case derived free e documented testing approacht s.
b. Are the test esses tens! stent with the documented testing approacht
- c. Does the test case entrie clear 3r establish the relationship between iet t.st.d, tesi c.ses .nd ,e,uirements s
WE-106- 20 1
i-T ABLE 3 (continued)
~~s pyj(ente,[er verification of the test Plan
,[ j.
L/ 6. is the specification for each test case completet
- Unique id6 ratification i
- Obiettive(s)
- Input
- Espected results
- Svaluation criterie
- Seletten to other tests
- Sationale for test estup
- e. Are Were test cases specified that are espresentative of senditions under which the program will be stuisoif
- 6. Is the specification for each test asse adequatet
- a. Is input detail sufficient for imput preparetient
- b. Are espected results explicit and specified erith sufficient accuracy?
- c. Do the evolustion criteria provide unambiguous poss/f ail status for each test cesst
- d. Is the method of comparing test output with expected results feasiblet e, Are dependencies between test cases descrlini and adequately specifiedt
- f. Are date libraries identifiedt g Are the operating environment requirements fossiblet
- 7. Is the plon f or building, updating. and maintaining the test date base s,
edequatet i *
- a. Is every test case represented?
- b. Does the input agree with test esse specifications
- c. Are output options correctly specifiedt
- d. Is all output required for evaluation of test case results specified?
- e. Are full references and all job control languese consistent with specifications
- f. Is each job stream esecutablet 3
Is each input case correctly identifiedt
- h. Are input properation instructions in occord with the program documentatient f
- 4. Are the test procedures complete'
- e. Are the procedures specified in suf ficient detail to permit third-party esecutient
- b. Do the procedures cover building, updating, and use of the list date beset
- t. Will all required test results be produced'
- d. Are instructions provided for disposition of test resultst
- e. Do the procedures cover preparation and woe of all necessary date .
flies and external support programst f'~ f. Do the procedures provide for configuration control of the test date base, dete files, and enternal programs, consle tent with the v&v plant k I l bu:-108- 21 I _ _ _ _ _ _ _ _ _ _ . _ . _ _ _ __ _ _ _ . . . - _ _ J
5
/*g TAst.g 3 (continued)
[ . culdence f or verification ef,}he , test Plan
- s. Are the precedures automated Wherever possible to pintelse human errect h, is the evaluetten methodelegy for each test case describedt
, i.
Is the use of discrepancy report forms specified in the procedurest J.' tu o procedure for terunning test cases includedt
- Can the procedures be performad within the planned resourcest
- 10. Is the plan for reporties test results adequatet
- a. Se there e receemovided foemet for discrepancy reportst
- b. Is t.he test case les focust specifiedt
- c. Is the informatten seguired en the test esse les sufficient to deemment date end time of esecutions, observettens that is11ures (discrepenties) have occurred, and reopens!.bility for test ouecutient
- d. le an adequate mothed of summerlains test results providedt
- e. Will the test results provide the 1eput needed to demonstrate that the progree meets its acceptance criteriet
- f. Are there precedures to tentrol totesting and document then$e?
- g. Is the method for essembling individual test results into the Test Report describedt g
( . 1 4 WE-106- 22
\
\
-__ - - _ .J
1
?
f
- l. ,
I 1 . L TAgLE 4 gg culdence for Desite verificetten j l
- 1. Does the Design Specification conform to required documet.tation standardst '1 s, Are all required sections presentt
- b. Does sech section contain all ef the required informatient
- c. Ic the format as required?
- 2. Is the Design specifical.lon traceable to the Seguirement specificatient J
, )
- a. Are all requirements 1sglemented in the designi
- b. De all mopects of the design beve their basis in the requiremantet I i
- c. Aew the emanarical tectwsiques that are specified appropriate for the problem to be esteedt d, Are the algorithms that are specified appropriate for the problem to be solvedt
- e. Mas the program design been partitlened in a menner consistent with the problem to be solvedt
- f. Will the program as designed umet the requirements 3, is the design coq 1 stet
- a. Does,the design implement required program beherfor with respect to each program interfacet
- b. Are all program inputs, outputs, and data base elements identif tet and described to the extent needed to code the progrant O. .
V' c. Does the specification describe the operational environment into wh!ch the program swet fatt
- d. Are all necessary processing steps included?
- e. Arc all possible outcomes of each decision point designed?
f. Does the des 1gn tate into account all expected situations and conditionst
- g. Does the design specify appropriate behavior in the face of enerpected or leproper inputs and other anomalous conditions?
- h. Does the specification reference all desired programing standards?
- 4. to the design correct?
- s. %s the design logic sound, that 1s. Will the program do what is intended?
- b. Is the design consistent with documented descriptions and known properties of the operational environment into which the program sust fitt *
- c. Do interface designs agree with documented descriptlens and known properties of LM interfacing elementst 4.
Does the document correctly describe a!! inputs, outputs, and date base elements Whose format. centent, dets rate, etc., are mot at the discretion of the designert
- e. De the models, sigorithms, and numerical techn14ues used in the design agree with standard references. Where applicablet
[ b . WE-108-23
\
l d TABLt 4 (continued)
^{ cuidance for bestan verification 1 I
- 5. Is the design internally consistent?
- a. Is the eecument free of internal contradletionst 7 i
b. Are the osdels. algorstles, and muserical techniques that are precified mathematically cougatiblet t. Are input and output formats sensistent to the artent possiblet
- d. Are the eosigns for sla11er er related functlens consistentt l
- e. Are the accurecies and units of taputs. data base elements and eutpute that are wood together La aceputations or logical decisions eespotiblet f.
Are the style of presenteilen and the level of detait consistent throughout the secumentt 6, is the design clear and unambiguous a. Can all design information be interpreted?
- b. Can all design information be interpreted in only one wayt
- c. Is the design ergenised and presented in a way that promotes clarity (for esasyle, use of tables and lists in place of test, where applicable)? u d, 1s the design sufficiently detsited to prevent oisinte g retatient
- e. Does the specification differentiate between program design and other
'O information provi6ed in the specifiestion?
(w a
- f. Is the design feasiblet
- a. Are the specified models, algorithms and numerical techniques within the state of the art?
- b. Can they be septemented within the constraints imposed on the system and on the development efforti
- c. Are the functions as designed isrplementable within the evallable resourcast I
8 (t l WE-108- 24
\
i i TABLE 5 I /m\
/ Ouldance for verification of the Proaram Documentation
- 1. - Does the Frogram Documentation confers to AsSI 5413 1974 and/or other documentation standards imposed on the documentt
- a. Are all required sections presentt
]
- b. Does each settien contain all requ2 red infermatient
. . . . _, s. Is the format as requiredt
.. y
. 2. le the infeemstion in the Program Documentation senststent with that in the Boquirement Specificatient
- 3. Is the information in the Program Deementation consistent with that in the Design specificatient.
ed. Is the information in the Program Documentation an accurate reflection of the seded progreat
- 5. Is the description of user input adequate for test planninst
- a. Are all inputs specifiedt
- 6. Are formats fully specifiedt
- c. Are relid ranges of input values specified?
- 6. Are valid data rates specified. if applicablet l
- e. Are valid input sequences specified as applicablet i p 6. Is the information in the Program Documentation internally consistent?
I ()- =
- e. Is the document free of internal contradictions
- b. Al*e the style of presentation and the level of detoll consistent throughout the documentt
- t. Is the infomation in the Program Documentation clear and unambiguous?
- a. Can all of the information be inter 7retedt
- b. Can all of the information be interpreted in only one way?
- c. Is the information organised and ptesentwo in a way that promotec clarityt
- d. Is the information sufficiently detailed to prevent saisintesyretation?
*Not appilcable in the desi5n phase k'E-108- 25
(. ./ TABLE 6
./
Culdance for source code Verification Does the coding conform to specified standards and procedurest 1.
- a. Does the coding conform to Ass! Stoddard ABS 10.2 en pf*gressing practicast
- b. Does the coding conform to Ass! standards on the s'J ing language, if app 11 tablet
- c. Does the soding eenform to other specific development standardst
- 2. Are suffittent samment statements provided to give en adequate description of each routinet .
- s. Are Esquut and output variables serrectly describedt
- b. Are can6 tents essed in the subroutine describedt
- c. Are the various calculation and tasks emplainedt
- d. Are the swading and writing of 1/0 (11es clearly explainedt
- e. Are special coding features such as mised mode, word pecking, and non-Ass! seeing clearly identified and explainedt
- 3. Is the coding clearly understandable
- a. Are cosqplex teding structures avoidedt
- b. Are consistent indenting and relsted techniques used to enhance clarityt
- c. Is self-endifying tede avoidedt
- a. Is the source code logically consistent with the Design Specificatient ,
- s. Are all foetures of the deslEn fully and correctly implemented in the codet
- b. Do all features of the coded program have their basis in the Design Specificatient S. Are all variables properly specified and useet
- a. Is the program free of unused variables'
- b. Are all variables initialisedt
- c. are array subscripts consistentf
- d. Are loop variables within boundst
- e. Are constants torrectly specified*
- f. Are proper units use6 with each verlable?
- 6. Is there satisfactory error checkingt
- a. Are input items checked for limitst
- b. Are esternal data files checked to assure that the terrect data file is being road shd the data is in proper fennett
- c. Are results of cattulations checked to be reasonable valuest
- d. Dces arrer detertlen result in consistent error messages and recovery?
- f. Do all avbrouttna calls trenefer data variables correctlyt
- a. Are the number of variables and the (Ype of each vertable the same in both the calling and called routinest Are labeled common vertebte names, type. 1ecation, and sise of arrays b.
consistent throughout the problem? ) i 1 ( f WE-108- 26 l 1 1
TASLt 6 (continued) Ouidance for Source Code verification
- 8. Is the data reed from euh (!!e consistent with the dets written to that f11ef
- a. Are the smanber and type of verlebles consistentt
- b. Are unit swabers consistentt
- 9. Do unit test reewits show that
- e. Bach htine test.ed for major logic patha within the routinet
- b. Both routine was checked for appropriate sintam, earlmen, and overese sets of variablest
- c. Seit statements were wood to print intermediate resultet d.. The module reproduces identical resulte with identical inputt D .
e e e O . WE-108-27
\
TABLE f Guidance f or Werif teetion of protraP intetretto9
/],
N.) 3, Does the integrated program confors to the system resource requirements
- a. Does the program emot storage requirements for sea 11 core, extended sore disk and tapet b.
Does the progren meet time end sising requitoesntst 2. Does the integrated pregeen interf ace.ptwperly with outernal (!!sst
, s.
Does the pr*gran properly road and writo ester at filost
- b. Does the program properly road weer input delat 3.
Arw all pieces of the Amteststed program properly identifiedt
- a. Wee the oeurce tede been verifiedt
- b. Was the tempiler been identified?
c, ansee special user libraries been eerifiedt
- d. Maes system libraries been identifiedt
- e. Mas the looder been identifiedt
- e. Does the progres !!nk together in a consistent menner?
- s. Are there any missing subroutines b.
Does the module linkage specification create o properly linked prograat !
- c. Are all routines leaded into proper segmentst d.
Are global and local labeled common blocks properly specified for each segeent? f
\
5 Are the interlaces between functiohal units correcit
- e. Are labeled seemons consistentt
- b. Are argument lists passed consistently
- c. Are I/O data file names consistentt
- 4. Are 1#0 ettwetures consistent and correct?
6, 1s the control language for building the integrated prograz correctt
- s. Are proper compiler options used?
b Are proper libraries speelfledt e, Are loading options consistent for initialization of vertebles med obtaining lead maps?
- 7. 1s the control language used for execution proport
- e. Are s11 files properly stetifiedt
- b. Are esecution Line limits correct?
- c. Are externst date files properly attsched and seeed? .
- 8. Are the special date librarles that are used for esecution correct?
- a. Do the Jibreries confers to the Design Specification in strweture and
- fortett b.
Are there suf ficient data in the libreries f or proper esecutient c. Con additional dets be added to the !!braries est!!yt O) ( w/ WE-108- 2 8 1
4
- TABLg 9
',8 'culdence for verification of test mosults
~
,/ - .
- 1. Does the Test Report comply with the format specified in the Test plant
- a. Does it provide templete identificotten of the program t'estedt
,. - s ,1 . - c,,, . . , t T.e , _ rt, j
- c. Does it reference the Test plan and any other relevant docume4tet i
- d. Does it prev!de a complete and occurate description of it.e test !
, envireaments i
= Location
- aquipment
- Support softssere used ,
i
- e. Does it describe and justify each deviation free the Test plant .
- f. Does it provide e owmmary of test results? I
- g. Does it provide en ovatustion of the progreat
- h. Does it prowlde recommendstlens for rotesting and/or program acceptaneet
- 1. Does it provide a detailed description of the results of each test casef J. Does it include a copy of the test case' legt
- k. Does it include all discrepancy reporte prepared during the testintt
- 2. Is the information in the Test Report.an accurate reflection of the
,O testing performedt
\ . a. Does the summary of test results occurately esflect the test outputst
- b. Is the evaluation of the protram a realistic and eccurate reflection of the test resultst .
)
- c. Are the recommendations regarding retesting and/or ecceptance sound l and based on the test results?
- d. Do the descriptions of test results accurately reflect actual test outputst
- e. Is the test case log complete and consistent with actus! test outputs?
- f. Are the discrepancy reports coglete and consistent with actual test outputst
- 3. Neve all test esses been esecuted correctlyt
- a. Does the test case leg indicats performance of each test case in the specified environment. using specified test procedurest
- b. Is there an explanatten for any deviation free the specified test enviremment er precedurset
- c. Es thers a discrepancy report for each deviation free expected resultar
. d. Were correct inpute need for each test caset
- e. Are the output values for sech test case occurately reported?
VE-108-30 i l I _ _ _ _ _ _ _ _ _ _ _ _ _ _ i
TABLE 10 ?i
,A 'I auidance for verifteetion of Installation pnekaae
- 1. Are su* . Are sufficient heteriale erellable en the program installation tape to
,,ppj g ,1,.,
permit rebuilding of the installed progreat
- r. Ae- Are the necessary pieces from the following list avs11ebiet e.
- Sourse sede Seer-supp11ed library routines
" .. ' spetano library routines a
p le 11mkage specifications
. I M etrweters esfinittore end date base meterials
- Ntrol Leagunge for inste11stien
- Date libraries to be used by the progren
- Test cases
- Centrol lenguage for esecution
- Output results from the test seses
- b. Are the format and content of the tape properly identified la the installation procedures for esey reading of the (11ost
- c. Are the installation procedures clearly understandable to allow
$natellation and checkoutt t
- 2. Can the progres be rebuilt from the installation packaget
[ k Con the program source be recogiled in the some menner es before?
\ e.
- b. Con the program be reloaded into en esecutable program in the same menner es before?
- c. If there are changes in rebuilding. de these changes effect the functional operstlen of the progreat
- 3. Do the test cases produce results identical to output supplied with the installation pockeget
- e. Can all test cases be performedt
- b. Are all results identical to prevle,.h resultst
- c. Are differences in results clearly understood and justified (e,.ch as eiew dote and time en printed output)?
WE-108- 31 i t I
;I
!I t
! i l l'
. TA4LE i (continued) . /% > -l Ouidance for Verification of Proarse interretton I
- 9. Neve sufficient edits been produced to verify the processing of date and j trenesission of date between and among modulest j 1
- a. have princirst 1/0 flies been checkedt I b. Neve labeled esemon blocks been checkede
-c. have variables passed to routines been checkedt
- 4. Neve Principal results been checkett q
' .i 1
a 1 4
)
TABLE 6 Ouldence for Protrem Validetion I
- 1. Has each requirement been tested adequatelyt
- e. Does the set of test results corresponding to each requirement cover the range adequately?
- b. Mes each test result for this requirement estisfied its eyelvetion criterief
- c. Does the sembination of test case results for this requirement meet the acceptance criteriet
- 2. Is the total set of requirements mett
- e. Is every acceptance criterion met settsf actorilyt
- b. Are there any test results that indicate wnrepeatable, unreliable or unewpected pro 6 rem behevlor?
t
- c. Are the test results consistent with the initial Statement of Problem for the prograaf B
e
/
r
\
WE-108-29 t 1
- s. ,e '.
RESPONSE TO 43 QUESTIONS ON RELAP5YA NOVEMBER 1, 1985 l I w i O
i l a I. CONDENSATION HEAT TRANSFER AND NONCONDENSABLE CASES g, - %,/ ' 0.I.1 Clarify the concern expressed in the following from Page 22 of-Reference 13 where R5 stands for RELAPS: "In R5 the mass transfer is modeled with a simple relaxation formula making proportional to (X-X,). Two different proportionality functions are used; one for condensation and one for flashing. Both of these relaxation rates have been adjusted to fit experimental data obtained from pressure change-dominated experiments (vs. wall heat-dominated tests), mostly depressurization tests. More experimental j comparisons are needed to test the condensation rates. Because the i present_model has been adjusted to fit the pressure change-related mass transfer rates, it is not suitable for wall heat flux-induced
' mass transfer (i.e. , mass transfer caused by subcooled boiling or any other wall heat transfer mechanism has not been adequately addressed in RS). Because the mass transfer rates due to pressure
. change'are larger than the corresponding rates due to wall heat flux, R5 at present calculates most heat flux-generated mass transfer probams as nearly equilibrium situations. In developing the modeling needed to address mass transfer due to wall heat flux mechanisms, it appears that some of the detailed partitioning of energy that was avoided in the five equation formulation will have to be examined and included in the Q model."
A . I .1 ' The interphase mass transfer models in RELAP5YA are similar to RELAPS/MODl. Hence, the concern expressed in the question is valid, but mainly for realistic (best-estimate) LOCA analyses. In licensing analyses of reactor systems, the code assessment work at Yankee Atomic Electric Company (YAEC) has shown that the impact of the models is in the conservative direction. This has been discussed in Section 2.1.3.1 of Reference I.1-1. 1 Reference , (I.1-1) Fernandez, R. T. , et al. , "RELAPSYA - A Computer Program O for Light-Water Reactor (LWR) System Thermal-Hydraulic V Analysis, Volume I," YAEC-1300P, January 1983. i
4 .' o 0.I.9 ,The condensation model is described as follows on Page 35 of Y"'N ' Reference 10: Y)' [T = -K (1-X+Kc(e~ K = 1.0 x 10 (Kg/m -sec), empirical constant I = 1.0 x 10~ , empirical constant.
"From the code assessment, we have inferred that this model~tends to overpredict the condensation rate when subcooled ECC is injected into a steam environment. This causes the system pressure to be low, and results in destaded heat transfer in the fuel bundle region' as seen in the Two-Loop Test Apparatus (TLTA) and Loss of Fluid Test (LOFT) code assessment cases discussed in Section 5.0 of' Volume III. An improved model in this area is desirable for best-estimate analyses. For licensing analyses, the degraded heat transfer effect and the conservative assumptions imposed by Appendix K yield
(~ conservative results." , V Clarify why this equation is conservative for licensing analyses, including its effects on ECC injection, pressurizer refill and condensing any vapor in high points of the system. A.I.9 The conditions of pressurizer refill and condensation of vapor in high points of the system are not important from the point of view of small break LOCA analysis since these occur beyond the point of core recovery and are, therefore, beyond the scope of the analysis. For the case of ECC injection, the above equation may not be conservative. However, the input assumptions used along with this condensation model will be such as to ensure conservative ECC injection flow rates. These ECC modeling guidelines are described in A.II.5 to A.II.8. 0.I.31 Clarify for the LOFT L3-1 accumulator comparison in Section 2.4.2 of Reference 12, Which curves are the data and which are the RELAPSYA calculated results. t
l l A.I.31 As described in A.I.23 (July 198s submittal), the accumulator model 7
.{ i described in Chapter 3 of Reference I.31-1 has been replaced by the RELAPS/ MOD 1 Cycle 18 model. Therefore, the LOFT L3-1 accumulator response has been recalculated using the revised medel as part of the system test prediction. Comparisons of the RELAPSYA results to the LOFT L3-1 data are presented in Appendix V.1-1.
1 Figures V.1-6 and V.1-7 present comparisons of the calculated and measured accumulator level and pressure. As a result of a later calculated accumulator injection time and the assumption of a higher ECCS water temperature, it is hard to assess the accuracy of the accumulator model from these plots. For this reason, the curves were reconstructed to show accumulator level versus accumulator pressure. Figure I.31-1 compares the RELAPSYA calculation to the LOFT L3-1 data. The comparison shows that RELAP5YA accurately predicts the accumulator expansion. Reference [\ Q) (I.31-1) Fernandez, R. T. , R. K. Sundaram, J. Ghaus, A. Husain, J. N. Loomis L. Schor, R. C. Harvey and R. Habert, "RELAP5YA
- A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description,"
Yankee Atomic Electric Company Report YAEC-1300P, Volume I, October 1982 (Proprietary).
/
V} t' 1 i
\
s . 0 _.. b t C 2 b. Ii id
- b 3
E-R- ' RU- ' TS AS- ' DE ' R-1P- ' 3R-LU-T b. TR ii id FL- O) OU '4 I R LM- ' U ' S OC- P TC- ' ( A
. N -
' E
~~ BS- R s
SV I -
' U
' S RL ' S AE- E PV- R ME i6
- b. P OL ri C - 0 R '6 RO- '
YT- ' 6R PL- ' HU-LM 1 ' EU- R- ' RC Y3 C- 5L ' R- P AT ' LF ii EO RL b.
' 6
- 0
'6 M
= '
= '
b.
. - . - . . . - :...:..:- ::: :- C' 0
0 0 0 0 0 0
- 0. 7 6 6 4 3 2 1 0
_ + sw , J_w:. w a_ - isi
. 0.I.32 "As part' of model verification studies performed using test data
'~
'from the LOFT facility, the sensitivity.of accumulator model response to'the size and location of-the node attached to the accumulator.should be assessed" (Page VIII-59'of Reference 4).
Clarify how this has.been done'in the submittal because the Maine Yankee and LOFT L3-1 analyses modeled only the accumulators wit.h pressure boundary conditions. A.I.32 As mentioned in A.I.31, the LOFT L3-1 accumulator response was reanalyzed as part of the integral test prediction. In addition, a node size cenv',civity was performed for this test, the results of which are presented in A.II.S. No assessment of the sensitivity to the injection location has been made since only the actual location has been used in model assessment and will be used.for plant applications (see A.II.3). L l O
\
II. NONEOUILIBRIUM EFFECT.9 Q) 0.11.1 If a volume contains subcooled liquid with the quality less than l 0.5, clarify what happens to the liquid energy if enough liquid is removed from the volume to make the quality greater than 0.5 so that the liquid temperature is then jumped up to the saturation temperature. A.II.1 RELAp5YA will generally predict significant nonequilibrium effects only under conditions of rapid pressure transients. Under all other conditions, RELAp5YA will predict very little nonequilibrium between the phases. Hence, if a volume initially contains subcooled liquid, at a quality of less than 0.5 (but greater than 0.0), the amount of subcooling calculated by RELAp5YA will be very small. If this system is perturbed in such a way to cause the quality to exceed 0.5 and the liquid to be saturated, the effect of the perturbation on the liquid energy is also expected to be small. Because of the minimum-phase-at-saturation assumption in the code, (m) 0.11.2 any subcooled water injected into a node containing some steam would condence instantaneously all the steam needed to bring it to saturation as long as the liquid remained the minimum phase. This can cause a rapid drop in pressure that can cause the ECC injection rates to be overestimated because the delivery rate increases as the system pressure decreases. Also, the amount of steam available and the resulting depressurizatian is dependent on the node size. Clarify how overestimation of ECC injection rates is avoided. A.II.2 Based on the RELAp5YA calculation of LOFT tests L3-1 and L3-6, it was found that raising the ECC water temperature to 200 F mitigated the artificial rapid pressure decrease in the injection mode. Thus, in pWR applications, tmising the ECC temperature to near saturation (corresponding to the containment pressure) provides a method for avoiding overestimation of ECC injection rates. A further discussion of the guidelines used for modeling ECC injection ( is provided in the answers to questions II.5 through 11.8. t
_ _ _ _ _ _ _ _ - - . _ _ _ - _ - _ , _ _ _ = _ - _ _ _ - _ ___ L I
~ 0.II.t' If alternate injection locations are used, clarify that the heat 'I transfer in the core is not overpredicted because of higher pressure than would hr.ve occurred with cold les ECC injection.
A.II.3 Alternate. injection locations will not be used. A realistic representation of the injection location will be used in plant applications. 0.II.5 Clarify how the' condensation rate is affected by the size of.the volume and also the sizes of the adjacent volumes. A.II.5 To assess the sensitivity of the condensation rate and the ECC injection flow rate to node size, a node size sensitivity was performed for LOFT Test L3-1. The base case calculation, described in Appendix V.1-1 and A.II.6, used an ECC water temperature of 200 F. The sensitivity case consisted of reducing the node size of the cold les injection node and the adjacent volume by one half as shown in Figure II.5-1. This sensitivity calculation also used an ECC water temperature of 200 F. . The primary system pressure, the accumulator pressure and the accumulator level are presented in Figures-II.5-2 through II.5-4. For each plot, the node size sensitivity case is compared to both the base RELAPSYA case, as well as, the test data. The results show no significant difference in the two RELAPSYA calculations. This 1 isplies that the node size has no significant !.mpact on the RELAPSYA calculated condensation rate for these conditions. 0.11.6 Justify that the injection model will result in conservative values for the ECC flop rates. A.II.6 In modeling the accumulator injection for LOFT Test L3-1, RELAPSYA calculated a sharp drop in primary pressure, unlike the data, slons with an overprediction of the decrease in the accumulator level during some periods of injection. To prevent the overprediction of the accumulator flow rate, the temperature of the ECC water was artificially raised to limit the condensation rate (see Appendix
I.' l l< t d i I . f Os i 143 151 e 152
-+
605 ECC Downcomer Bace Case Cold Leg Nodalization 0 - 143 : 151 - + 154 - 155 - + 152 *
--+
605 ECC Downcomer Cold Leg Nodalization Used for Sensitivity Calculation Figure II.5-1: RELAPSYA Cold Leg Nodalizations Used in LOFT L3-1 Prediction
\
8 _ _ _ _ _ _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ ._ .i
O * . 0
. ( ~ - . . ~ - _i 0 0 -
8 1 M 0 S 0 S 0 7 1EE 1 o S0 _ 0 t 3R0 0 LBN _ - 0 8 1 1 - TAR 3 FYY 0 Ls
. O55 0 n LRR 0 5 To m '. 1 Fi Ot 0 L a 0 l c: _ - 0 4
nu 1 ic l 0 ea 0 rC 0 u sy 3 R 1 st T 0 ei A rv D . .r 0 0 2 Pi t 1 ' 1 mi
- 0 es 3
L . - 0 0 1 ) t n se yS . T F 1 C S e 0 E yd O 0 S ro L . 0 ( aN 0 m O E id ON 1 T . 0 M I rn
. 0 T P a O 0 9 f e S os I 0 a R nB A . 0 0
o P 8 sA M iY r5 O 0 aP C . 0 pA A 0 7 mL Y oE CR S 0 P 0 R . 0 : L 8 2 E - R 0 5
. 0 I i0 I 5
0 e r
. 0 i0 u 4 g i
0 F
. 0 0
3 0 0 0 2 _ 0
. 0
- 0 1
l 0
" 0 og Rh o~
o g..
- 0. q3 9g-k' Rg O :
5 A ga@cgi 8J9O AoS 9E s. z-
,e.
llj jl,i il
- y. .,
l l l l O ; o
. . . . . . i g
i t 9
> h o
-l- o o, "
-h-IE i = 33a
$3 o S8a a cm g *O C -
e >. E- $U e
-R -
- 3 ea 3-l * $$
-8 og H = 3m L O -me 3- -l 0$3 o
H o
"s*oe ,fT g - -l $M$ %)m o-e o H*
- O to 5- -l @E e
- as 8- WE
- o. <
C it
~l BA M
L I l O ON g E
- h. a :
n a E A n -g ;
' -me -
Ima o* mco *
' 3 aa z -g y Hcc 3 o
eMM a== -g n <t o l
-l u <t l L o I
\ \ ~8-i r A h 9 o
0*0S9 0*dOB 0*dSG 0*005 0*d9 C*d@ 0*dSC C*dQC 0*0SC (WISd) 380SS38d 801970WD309 O
-to-I
.L r:_
f . - . _ _ . : ... _ : .:_.~ :::.:3
'_ - - - - - 0
_ d
, g
)
8 i d
,g
)
6 i R , T d
)
A ,g D 4 i 1 , ' 3 L - . d 3 T ,g F , 2 i O L , '
)
O
, C T ,
' E dS
)(
N ig O O , IE S M I . ' I R . ' T R P . M t
'O O i C C . b R .
Y . ' 6 P . R iI S 'O C L S E . '6 R 1EE SD '
. 3RO '
LBN TRa 'O iI FYY ' C' O55 '4 LRR
. e *-
. e *
'0 i1 '
C'
'2
- - - ~ .
- - _ . _ _ _ ...... . :: _ ..: _.!:- ;. : . - c 0 0 0 0 0 0 0 6 6 4 3 2 1 0
*g :w>w,_ : aosa>ax"8a:
i_..'
V.1-1). An ECC water temperature of 200 F was selected since it
~'} is closer to the saturation temperature corresponding to containment pressure yet low enough to avoid the injection of superheated liquid.
Figure V.1-6 compares the accumulator level calculated by RELAp5YA. using a temperature of 200 F for the ECC water, to the LOFT Test L3-1 data. Since there is no measurement of the accumulator flow rate, the flow rate will have to be inferred by the rate of change in the level. The calculated rate of level decrease is lower than that in the data implying a lower volumetric flow rate. In addition, raising the temperature reduces the density of the ECC water and reduces the mass flow rate (for the same volumetric flow). For plant applications, the ECC water temperature will also be raised to 200 F to allow for conservative ECC flow rates. O.II.7 Clarify what guidelines will be used for injection node size in SBLOCA analyses.
,~\
)
~ A.II.7 The only guideline used in RELAp5YA nodalization of Reactor Systems is to provide sufficient detail to capture physical phenomena. This general guideline is also used in determining node sizes around the ECC injection locations. The node size sensitivity study discussed in A.II.5 has shown that this guideline provided an acceptable nodalization for LOFT Test L3-1. Similar nodalization schemes will be used in plant applications.
O.II.8 Clarify how the effects of injection node size and location have been assessed against data and how the range of data compares with expected SBLOCA conditions. A.II.8 Our test facility assessment models and LWR System models contain relatively detailed nodalizations that use actual ECC injection locations. These nodalizations allow the pressure and flow to communicate readily between surrounding volumes. Thus we expect the depressurization rate to be sensitive to the ECC condensation rate, [~'}
but not sensitive to the ECC injection node size. This has been 1
_______ i
confirmed by the injection node size sensitivity case' described ..n
= /* Answer A.II.5 above. No' sensitivity studies have been performed I
with alternate ECC injection locations since'we have used actual injection locations in our assessment and LWR System models. Two integral tests, LOFT L3-1 and L3-6/L8-1, have been used.to assess RELAPSYA for PWR SBLOCA applications. These tests were q scaled to represent a four-inch equivalent cold leg break in an LWR. This break size is'close to the limiting' break size expected for our PWR SBLOCAs. Table II.8-1 compares the nominal' accumulator conditions for LOFT Test L3-1 to those for the Yankee and Maine Yankee PWRs. This comparison shows that the L3-1 accumulator flow
~
contains.more subcooling than either plant due to the higher initial pressure and lower fluid temperature. Therefore, with respect to condensation rates, the LOFT accumulator bounds the plant-application. Three TLTA tests, consisting of two large break-and one small 7 G experiments, have been used to assess RELAP5YA for the BWR full. ..
'V) t break spectrum. These are described in Reference II.9-1. We.
believe these tests span the range of conditions for BWR LOCA applications. TABLE II.8-1 Nominal PWR Accumulator conditions Pressure Temperature Subcooling Plant (psia) (OF)- (OF) Maine Yankee 220 110.0 280.0 Yankee Rowe 488 100.0 364.5 LOFT L3-1 634 88.5 403.7 O : i
\
l _ _ ______jf
i j I l 1 0.11.9 Page VIII-29 of T.eference 4 says that submittals need "...to confirm j
/~'N that HPI and accumulator flows during small breaks will not result j i f
/
in the formation of water slugs, or if they do, to show that the structural design bases of the primary system include loads due to: ) (a) water slug inertial motion (b) water slug impact and (c) pressure oscillations due to steam condensation." Clarify how RELAPSYA models water slugs if they occur and how sensitive this modeling is to noding. A.II.9 We assume that the formation of water slugs is defined as the phenomenon of periodically filling a fluid volume completely with water. In the RELAPSYA analyses carried out so far at YAEC, there have only been a few cases where water slugs have been calculated to occur. However, in these cases there was also a buildup of mass error, and when corrective action was taken to minimize mass error ! (see response to Question X.1), water slug formation was not observed. All the RELAPSYA calculations described in Reference II.9-1 and the additional calculations described in response to
/N Questions V.1, V.2., V.3 and VI.9 have acceptable mass errQrs and 1 J
~ have not shown occurrences of water slugs. The same principles used in developing the nodalization for these calculations will be applied in plant calculations.
Reference (11.9-1) Fernandez, R. T., et al., "RELAPSYA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume III," YAEC-1300 P. January 1983 0.II.10 Pcge 13 of Reference 8 states: "The thermal energy source term associated with the virtual mess acceleration term is not necessarily positive or zero. However, consideration of the second law of thermodynamics dictates that such rust be the case. This contradiction is not resolved at this time, and since the energy dissipation associated with virtual nass acceleration force is rN small, this energy source term is presently omitted." Clarify ( )
\/ approximately what the energy error is, based on this assumption.
j
.j, A.II.10 This broad hypothetical question is difficult to answer precisely.
hU for all conditions that might be encountered in PWR SBLOCAs and the BWR full break spectrum. However, the RELAP5YA code has been assessed against numerous separate effects and integral system. tests that are discussed in Reference II.10-1. These cases span a very. wide range of conditions that might be encountered in PWR SBLOCAs or the BWR full-break spectrum. In all assessment cases, the-last term. on the right-hand side of Equation-15.in the cited reference was neglected. -The favorable RELAPSYA code assessment results' indicate-this appears to be a reasonable assumption. Therefore, we believe this term has a negligible effect on the thermal energy equation. Reference (II.10-1)- Fernandez, R. T., R. K. Sundaram, J. Ghaus, A. Husain, J. N. Loomis, L. Schor, R. C. Harvey and R. Habert, "RELAPSYA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume III: Code rh Assessment," YAEC-1300P,, Volume III (October 1982), (Proprietary). V
-is-5
_ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _l
l l, l l IV. STEAM GENERATOR HEAT TRANSFER 7s
) {
^
0.IV.4 Page 301 of Reference 12 states: "The RELAPSYA calculation I generally provided a reasonable analysis of the LOFT system response for Test L3-6. The only major discrepancy between the calculation and the data seems to have been in the underprediction of pressure in the Secondary Coolant System. This is believed to be due to overprediction of condensation rates. The calculated mass error in the primary system was, at most, 20 lbm, which is well within the data uncertainty. The Secondary System mass error was considerably-higher. This may have resulted from the overprediction of steam condensation." Clarify why this overprediction of steam generator secondary condensation rates and large secondary mass errors will not produce nonconservative results for SBLOCAs. A.IV.4 To address the impact of underpredicting the secondary coolant system pressure for small break LOCAs, the RELAPSYA calculation of LOFT Test L3-6 was reanalyzed. In this reanalysis, the steam generator secondary, side pressure was set to match the data and the impact on the primary system conditions were investigated. Figure IV.4-1 compares the original and revised RELAPSYA secondary system pressures to the LOFT L3-6 data. At about 1800 seconds into the transient, the original and revised RELAP5YA secondary pressures differ by as much as 300 psi. The primary system pressures for the two RELAPSYA calculations are compared to the LOFT data in Figure IV.4-2. Despite the large differences in the calculated secondary pressures, the difference in the primary pressures is no more than 30 psi. This small impact on the primary system is also illustrated in the comparison of the primary system inventories presented in Figure IV.4-3. The revised RELAPSYA system inventory is only 200 lbm lower than the original. The overall effect of these differences in primary system conditions on small break LOCAs is best determined by comparing the cladding temperature responses for LOFT Test L8-1.
/~}
V The starting point of L8-1 is the final system conditions of L3-6. Therefore, the
~16-t
differences in primary system conditions will be carried through'to Test ~L8-1. Figure IV.4-4 presents the peak clad temperature
/s(N\') responses for the original and reviced RELAPSYA calculations compared to the L8-1 test data. The two RELAP5YA cladding temperature responses are almost identical with the only difference occurring during the recovery period. The reason for this difference is a result of accumulator modeling differences described in A.V.9.
'Since there is virtually no difference in the two calculated peak cladding temperatures, underpredicting the secondary system pressure is not expected to produce nonconservative results for small break-LOCAs.
0.IV.8 Page 15 of Reference 16 says: "The results of a FLECHT SEASET steam generator separate effects test analysis using RELAPS/MODl. . . suggests that the nonequilibrium behavior on the secondary side of a steam generator in reverse heat transfer cannot (N be correctly calculated with MOD 1. The presence of such a subcooled Q layer on the tube sheet is closely associated with the propagation of a ' quench front' up the insides of the U-tubes. This quench front cannot be calculated without a reflood model containing a nonequilibrium heat transfer correlation package together with a moving fine-aesh temperature grid..." Clarify how secondary-to-primary heat transfer'is treated in RELAP5YA. A.IV.8 No special treatment is given to the secondary-to-primary heat transfer. The standard RELAPSYA heat transfer models described in Sections 2.1.3.6 and 2.1.3.4, Appendix A of Reference IV.8-1, are ) used. The nonequilibrium behavior described above was seen in LOFT Test L3-6 and not predicted in our RELAPSYA calculation. Reference (IV.8-1) Fernandez, R. T. , et al. , "RELAP5YA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume I: Code (~ '
\ Description," YAEC-1300P, October 1982 (Proprietary) 1
lil I lil L 0
~ - . _ ~ - - ~ _ - ~ _ - 0 0
0 2
= 0 _
0
= b 4
2 _ _ 6 _
= 0 3
0 L b 2 2 T F s . O L n
= -
0 o b 0 inti 0 a -
= 2 e lu r c u l 0 s a R 0 s C T . b e P. 9 rAY D 1 P 6 m5 P teAL
- 0 3 0 s L . b 8 y E 1 S R T
P y de O 0 r s L a i 0 dn v - O . $ T 1 o Re 1
) c N C e O S dn S 0.
0 S f E a I . R 5( 2 ol A P 1 E no ian M M s g O I i C 0. T r 0 a r i R b 0 pO - Y 1 m o S Cot P R O. L : E do 1 R L S - DA 4 EN V SI I 6IG O.
. VI o e 3ER d O r LRO u g
TRR i PYY F O55 O. - LRR do
.= i
_ = 0
. 0 d
3 0 p 0
,g o. ~ ~ - - - ~ -
o N = g o g 9 0 o si s 9 r o- l e$ ara ~ g 9lW
~
2~E uohg: wE8 aSgmC: 9 Ewb - 1i ~ EI lt!
. _ _ _ - - _ _ _ _ _ _ _ - _ _ _ _ _ _ _ - - _ _ _ _ _ _ _ _ . _ ~ . _
\
.1 l
1 i o g I g5b w
~$
o* l i co - c3 4 e>- l be$ cc S T {$$ smm "h 3 a i
<e n [.
9 Smc i
- at U .
j 05 c 9 e-o 23 86 7 o. s-b- -$ ~ 0$
~ a
'5 a ww C o 8 5 J
m t, a- -g
- u s..a .
O =- , 93 m-a: -u a %. e O' a 8
.g 9 ::
C
~ $3 ou g UO c: o*
d. e .H g 3 o 3 "I 8 i o tu d o 2 i . . i . . . . . . 0 0*00tZ 0*000C 0*0081 0*0091 0*0061 0*0061 0*0001 D*000 C*000 0*006 0*00C 0*C (WISd) 380SS38d 83*110H cl0011091NI i
,s.,x. .
k
? O n .. n n t n ...in f o . nnlei n'u n e e t.... ..n.f, nn n n t n en n ,t n ., n e it i n.n...t.u..n ,.t..u ,,,iit n..., n d o n n u, m I i i I i 4 i 4 e 4 8 4
.N g
.c
. m O
tk "N Ik'
. ,,,3 ..
IP'
-O
-- ww . - n a Q w C.D co 8>w qi w t<
. Id M , A to MO lb, O c Ao I-*C C qq C4. N > -O c#
g -- O Lnin w lH D ^ @ \ H , -2MM (D y& C g- a t T3 11 O id . u o .o.e , uc g>" M-og y ' CU e -O 4 (t?* -
.,JU.- m - D c >.,
N m
.T g a
- p>. . E<
ea kM. O c
. # 133 to D:
g" -O A I c e- gG =? p L- .g tLJ A .to
" try M e4
%s 9 >
p r- E., j u) *UJ "
.o td W
Nj i
"E Z. -
- -- DE O E
uc C w~
~p o
M Qg CfJ . ce o CO - . o .c - U C De e4 n" "g e4 & e -- -
-o et
>J.
gc C
.CD o, o E
g H" OO g O. . UW J Ld
.O ..
I M ~- - D ( l . .tD l . . f. l . .
"Q c
.o, y e
g t _a
. .N
. 1. I f f I f f f f f f _
su u n q m u n..a u m uninin u npuun np uu.iu p uunupuu nupu n u ngu n mug un ningun n eugu n su n w O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O L") N
- O CD CD N CD 10 + CT) N *
/^
- e e -
\.
(W97) SSUW 18101 i
1
!l e
O 0
. . - . - - - - . 0 0
. 2 L
DA - 1 EN 8 SI L . 1IG 0 _ . VI I O T _ 8ER 8 F _ LRO 1 O L TAA PYY n O55 i LRR 0 e- e
. 0 r s 8
1 u n t a a o i r t e pa m lu 0 e A 0 T lc T . 4 a A 1 gC n D i g1 d Y A 1
- d l 5 8 # , P L 0 aA CL
. 0 )
T 2 1 C6 E P O E2 R S0 d L ( e 8 s O EG OT . N O S e0 0 0 1 M5 I Te d R o d i v e I Nna R f A ol P M n a O . 0 o in C a0 8 s g i A r ir Y a pO 5 P mo A Cot L 0 - E u0 R . 8 : 4 4
- _ V I
0 s0 e 4 r u g i F L : 0 2 0 0
- - - - ~ ~ -
o g" ag og 9@ og O 9g 6 Mga - EH Cz9 a8 8
*0 s$
O.8 8m 8e t oh o iS3 _
i 0.IV.9 Clarify how the condensation heat transfer that_could occur in the. primary side of the steam generator U-tubes during SBLOCAs has been. { '
%/ assessed.
A.IV.9 The RELAPSYA c.alculation of condensation heat transfer in vertical flow is based on two' commonly used correlations: For laminar flow the Nussel theory of film condensation.as
' documented by Collier in Reference IV.9-1 is used.
For turbulent flow the correlation of-Carpenter and Colburn documented by Collier (Reference IV.9-1) is used in the primary side of the steam generator U-tubes. The condensation heat transfer was assessed in benchmarking the code against integral tests. Two SBLOCA tests performed in the LOFT facility, Test L3-1 and L3-6' were assessed using RELAP5YA. The system behavior in these tests which depended in part on condensation heat transfer in the steam em generator U-tubes, was adequ,ately predicted by RELAPSYA. Reference (IV.9-1) J. G. Collier, " Convective Boiling and Condensation," McGraw-Hill Book Company, London (1972). m i
I V. SYSTEMS VERIFICATION AND OTHER EXPERIMENTAL VERIFICATION
/ T; i
' II.K.3.30 Page 3-17/ of Reference 1 states that: "The revision should account for comparisons with experimental data, including data from the LOFT Test and Semiscale Test facilities." "In addition to the modeling concerns identified, the task force also concluded that, in. light of the THI-2 accident, additional systems verification of the small-break LOCA model as required by II.4 of Appendix K to 10CFR50 was needed. This included providing predictions of Semiscale Test S-07-10B, LOFT Test L3-1, and providing experimental verification of the various modes of single-phase and two-phase natural circulation predicted to occur in each vendor's reactor during small-break LOCAs."
i Page 3-178 of Reference 1 states that: "Other separate effects tests l (e.g., ORNL core uncovery tests) and future tests, as appropriate, should also be factored into this assessment." Three types of integral system tests are discussed in Reference 12. (~ These include tests a) the Thermal-Hydraulic Test Facility (THTF) at ORNL, t' tests at the TLTA at General Electric, and a single LOFT test. The THTF tests were not modeled as integral tests, but rather only the test section was modeled using boundary conditions from the experiments, so questions on the THTF tests are presented in Section VII, along with other questions related to l core heat transfer. The TLTA is for BWRs, so it was not reviewed for PWR SBLOCA applications. Therefore, LOFT Test L3-6/L8-1 was the only integral PWR calculation included in the submittal. 0.V.1 Clarify why additional LOFT tests and also semiscale tests were not simulated as recommended in II.K.3.30 (Reference 1). It would appear the semiscale MOD-2A Tests S-NC-5 and S-NC-6 that involve injection of a noncondensable gas for both tests and natural i circulation for Test S-NC-5 and a reflux cooling for Test S-NC-6 would be especially useful. A.V.1 Additional assessments of RELAP5YA have been made. A prediction of LOFT small b:' *k Test L3-1 has been performed and is contained in [' 5 Appendix V.1-1. To assess the various modes of natural circulation, I .
i semiscale test S-NC-2 has been analyzed. This test contains a range. [] of power levels and primary system inventories. The results are
contained in Appendix VI.9. This test was chosen instead of S-NC-5 or S-NC-6 since noncondensable gas has been determined to be insignificant for our SBLOCAs (see A.I.3 and A.17I.1 to A.III.5, July 1985).
I l' , i
\,.
l N
\
l l \ ,
^?
l 1
- l. "
- Appendix V.I-1 y
f s
# ' Introduction 1 i To provide additional assessment'of integral PWR.small break phenomena, a RELAP5YA calculation of LOFT Test L3-1 was performed. The. main difference between this test and L3-6, which was previously analyzed in Reference V.I-1,-
are as follows: l a)' The reactor coolant pumps were tripped at the start of the accident,- b) The break was located in the inactive broken loop, and c) -Accumulator injection was allowed during system depressurization. LOFT FacilitX A description of the LOFT facility is contained in Section 5.3.1 of Reference V.I--1. RELAP5YA Model The RELAPSYA input used to model LOFT Test L3-1 was based on the input previously used to analyze Test L3-6 (see Section 5.3.2 of Reference V.I-1). Therefore, only the changes made to the model will be discussed and are listed below: o The cold les break location was changed from the intact loop for L3-6 to the broken loop for L3-1. o The break junction was modeled with a discharge coefficient of 1.0 for both subcooled and two-phase critical flow. o The ECCS injection was directed into the intact loop cold leg as per the test configuration. i ____________________________________.m_.____m_ _ .____________ _ _ _ . _ __._ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ . _ _ _ _ _ . _ _ _ _ _ _ . _ _ _ _ _ . . . . . _
pg , wa s o;~ The accumulator and' associated' piping were modeled to inject into the cold les.- o .The temperature'of'the ECCS water and the' auxiliary feedwater was
. artificially raised to 200 F.to limit the steam condensation rate-t in the cold leg injection volume.
- o. The reactor coolant pump flywheel' system with its variable inertia was modeled through a code modification.
o 4 Initial conditions were entered to' match the measured initial condit8cus of the L3-1 test. small Break Test L3-1 LOFT Test L3-1 was designed to simulate a 2.5% small break (4-inch. equivalent).in the cold leg of a large PWR. The actual break diameter was
-0.6374 inch and was located in the broken loop' cold leg. Safety injection was provided by means'of:the HPIS and the accumulator directed into the intact- ,
loop cold leg. The reactor coolant pumps were tripped at break initiation.
~
i The experimental procedure for LOFT L3-1 was as follows. -The reactor was operated for 92 hours prior to the initiation of the experiment. The reactor was then scramned 2.15 seconds before the break was opened. Electrical power to the reactor coolant pumps was terminated at blowdown. initiation and the pumps began to coastdown under the influence of a f3ywheel fsystem. Below a speed of 750 rpm, the flywheel system was disengaged. Coastdown of the reactor coolant pumps was completed at about 19 seconds. During the blowdown, ECC injection was directed into the intact loop
- cold leg. The HPIS flow started automatically at 4.6 seconds While the accumulator injection started at 634 seconds When the system pressure dropped below 633.8 psia. At 1,741 seconds, the accumulator and surge line emptied and the nitrogen cover gas entered the system. The auxiliary feedwater pump started at about 75 seconds and operated for 30 minutes. The fuel rods remained well cooled throughout this text. A more complete description of Test L3-1 is provided in the Experiment Data Report (Reference V.I-2).
i
4 J) The results from the RELAPSYA prediction of' LOFT Test L3-1 are presented in Figures V.1-1 through V.1-7. The calculation was ended slightly beyond the measured accumulator depletion time since this generally corresponds to core recovery in PWR SBLOCAs. Figure V.I-1 compares the "
- predicted and measured primary system pressure. The comparison shows very
. good agreement for the first 350 seconds. Between 350 and 500 seconde, the calculated decrease in pressure is less than the data. It is suspected that this slower depressurization is a result of a lower break flow quality and the subsequent lower energy removal rate. Beyond 500 seconds, the calculated and measured depressurization rates are similar, although the calculated pressure remains about 40 to 100 psia above the LOFT data.
The secondary system press"res are compared in Figure V.I-2. For the
~
first 725 seconds, the predicted secondary pressure agrees with the measured value. From 725 seconds on,.the code underpredicts the data. This underprediction is suspected to result from an overestimate of the mixing of.
' the auxiliary feedwater and a resulting overprediction of the steam condensation rate. This ov'ercondensation occurs in spite of the fact that the auxiliary fee'dwater temperature was raised to 200 F. However, it has been-determined by a sensitivity study for LOFT L3-6 that the steam generator secondary side conditions have little impact on the primary system response (see A.IV.4).
Figure V.I.-3 shows the RELAPSYA calculated break flow rate compared to the LOFT L3-1 results. The comparison shows that RELAPSYA predicts the overall trends of the data. However, between about 350 to 1,000 seconds, the calculated break flow rate is slightly higher than the data.
'The calculated intact loop and broken loop cold leg densities are compared to the LOFT L3-1 data in Figures V.I-4 and V.I-5. For the first 630 seconds, the LOFT cold leg data show a gradual decrease in density followed by an increase due to accumulator injection. The data also suggests that steam ,
. voids were uniformly distributed between the intact and broken cold legs throughout the test as shown by comparing the center beam data from each {
loop. The RELAP5YA calculation shows a more abrupt density decrease at about 150 seconds in the broken loop and 250 seconds in the intact loop. This suggests somewhat more vertical stratification was predicted for the primary is
system early in the test than indicated'in the data. Beyond 630 seconds,
/N f RELAPSYA calculates a lower density in both cold legs than indicated by the
' data. We believe these differences'are due to the following:
l
- a. The predicted break flow rate from 350 to 1,000 seconds in somewhat higher than the-data.
- b. The predicted accumulator flow rate is lower than.the data and !
initiates about 100 seconds later. This is discussed further below. 1
- c. These two effects imply that we calculate a lower system mass inventory than probably occurred in the test, and would explain the lower. cold les densities that were calculated. However, the test report does not contain the mass inventory history to confirm this observation. ;
l 1 Figure V.I-6 presents the calculated.and measured accumulator level. ;
-(
,q Since there was no measurement of,the accumulator flow rate, the accumulator
'N_) ' level provides an indirect means to determine volumetric flow. First, the comparison shows that the calculated time of injection is about 100 seconds later than the data as a result of a higher calculated primary system pressure. Second, the RELAPSYA calculated level decreases at a slower rate than the data, indicating a lower volumetric flow rate. The main cause for ;
this lower calculated flow rate is the assumption of a higher ECC water temperature, that is 200 F used in the calculation versus 88.5 F used in the test. This higher temperature reduces the condensation rate and thereby reduces the pressure difference between the accumulator and the primary ) system. The reason the ECC temperature was raised was to: 1) prevent an unphysical rapid drop in primary system pressure, and 2) ensure conservative ECC injection flow rates. Figure V.I-7 compares the calculated and measured accumulator pressure. The results are similar to the level comparison presented in Figure V.I-6. O J
F. , i a Conclusion.
.js 4
("^)L
'The RELAPSYA calculation generally provided a reasonable analysis of:
<e
' the. LOFT System response for Test L3-1. For those areas where the RELAP5YA
I i. calculation' deviated from the data (i.e. , densities and accumulator level), the code results tended to be in the conservative direction. In particular,;
'we believe we underpredicted the primary system-inventory.
References (V.I-1) :R. T. Fernander, R. K. Sundaram, J. Ghaus A. Husain, J. N. Loomis, L. Schor, R. C. Harvey and R. Habert, "RELAPSYA - A., Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume III: Code Assessment," Yankee Atomic Electric Company' Report YAEC-1300P, Volume III (October 1982) (Proprietary). (V.I-2) P. D. Bayless, J. B. Marlow and R. H. Averill, " Experiment Data Report for LOFT Nuclear Small Break Experiment L3-1," NUREG/CR-1145, January 1980. (V.I-3) LOFT L3-1 Data Tapes.
.{s
k
.4
/
f R I _ l I I I E f I : o
- . . -H -
e, ft .- es m
-d -
/ e MS -p
- n **
- a u ~
5 e O
- Q (0
I (D a- i,
- u-c
- e $
o- -g g-
- - e8 a-i o
.h e o- M..
e ~
" m. 9 O
e g mm tw f"% 0
# '$ a e s .h
- .s a v -
E- .E s t g a.* D. E e I O E tu O
* -h ~
e e o "
- m. .- T e .
3 k O g
.E =
O +
.- m l O
. .g
, o.
.g e
.g ese D
5 5 4 8 4 5 s 6 0*00EE 5 0*000t 4 0*0081 0*0091 0*0061 0 00tt 0*0001 0*000 0*000 0*006 0*00C (UISd) 3Bf!SS3Bd 83110H d0011091NI i
,s-~5 t
v) , , , , , , , , _ O g o C e 4 :
-h
$3 o_
'b i ,
_g n - MS 9 n -l u 9 .3
-$ e u
9 E
-% e e -
e 9 e5 U ~ e
~k-30 ap
' 9 #3n
- 3. 8
= ku 9
' " $6 3-a H
-Q-s @*3 8
og to ha (n z. -Ku" M *?
\ S E 9b E-a:: -
-g 8-' .$A
!!=
o o 0d o, O ~ eU e "N 88
$ o a=- -g li d
x 9
~
'h o u
- b
-g C o
o o 0~00CI D*0bt1 0*0bOI C*dO6 O' 0* dol 0*dOB 0* DOS 0*b 0*00C ( (U~Sd) 380SS38d W31 SIS 1890N003S t
l 1 L 1 u .I L I L ' .{x / O. i i i i i 7 g o l w. 1
~ e o=
Cl o I u e un - g s a- to _.2 c- o_. - o es o a u_
- u.a o E3 I
-O a e S
);. 5 1
to a a .! 3 9 o .: H o I _ p)" 0] u_ ' - o ma o - J 3
- o. m mg o a omos
- o L_3 ew g -
oz - SS j; o 9g eT
$- [] -g I 03
=~
cr H c.; A [ 1 9 s r_ a . o _8 a o m C o.
>_ _a LD Q a_
[ _.) o. a_ _a e - g o. i 2 i i i i 0'09 0'0S O'0F O'0E 0'0E 0'01 0*O O coas/wel) alua Mola >isasa
\ !
.l la O 0 Nh,
,L
~ - ~ -
0 ey M [d a M. a L 0 n e i 5 t y 0 i 0 s 0 n e h
. 4 1
D 0 g en
. 'O
}
L o i A A .
'0 dt T l a A 6 0 ol D . 3 Cu c 1 0 pl - oa 3 6 0 oC L . 6B0 1 L 111 1 A tY T A000 0 cS F Y000 aP O S- - I. X) tA L . PCCC i lC nL RPPP E I ER O L- -
OTN O S I ECEE RDDD ax M(E
- 0. S f
' oT N o1 oo n
- 0. Ii3s -
t R e.x d rL A . e a P pT M 0 mF O 0 oO C . d 7 CL A Y S P d O. o 4 A . e 1 R . L V E Y 0 R 5 o e P d R 3 r L u g E O. i R o F
._ d i 0
{ d 0 3 1 0
) d 0
3 i 0 0
. d1 0
[ l,l 0 g CW od _ O
.7 aY o
> Ug5o ,
.3e ty -
i
), +
s o e e e e e n , e e j ~
- 4j .
% -l 1-m-
m -
=
1 s . 4 -i-i H ; 9 e j
;[ -g i a - - x
<n-
- . o h
O - 1
? Y%
E s -g 5 si o* j
~
Y 3 h 6 o o, l G. I. '.- _g a { q U 3-o a_ a $E - 5 ,-
-g E .!!
og 6'P -- g> -j:em: e O ;g. -
.H_m -
g: - T e5 ,87
-8" 03
= 1 at 3' -g oS
$ $ t' s
o
,U. l
. e- -g d = 7 e _
q > 2- % -g e
] E
_= _=g 3w 9 M -R r
\
f -g n ox wh ok ok ok ok ok ok ok os ao tr. w W97) MISN30 637 0703 dOO7 N3X088 i
\
_ _ - _ _ _ . - - I
t ' I,7 1 i I i l I y~x i 1 I
-( O' N_/ saeeeeeaeinaeeeeaaafneeeeaeaefaeeeeeaeefeaaeeeaeafeeeee eee 4 4 3 8 3 m
O ! i
-u I
- . .j
- . \
O
. - D e
"O
-- D (D
.e
. . g
. . ~' C 0
E . ed -e
> "O ea C - D > fa O 9 0 e4
-w 45 M U I
= . ua O -
yy n - . e
.J *
~O es <
p . O c> L N U O
- f, $4
.J - .
m u4
< h1 g . . O g g
- 14.8 w "O CO OO g -- ~
_g , -
;% . ~* e ts.1 8"n t ) .$...-
g M m La 4
. Q). . . W g
L - . O sgU E: Oo o ~~ ~
~O UA O . ,10 I
g . - g
>= . . s in e gg, . . .
g -- g O > W - w _O 8 E " Ct n MS : S b (L.
- s-. .
h O o g
. E .J ,,9 O
-- . .g
. .N O
....rs....I.........t g i ........I......as.f.........f i 3 g ..... .. O O O O O O e e s e a e O.
to to
- M N
- O
\ ua 13Aa1 volvinunaae
\
l l1 Il\!J O 0 _
- - _ - - - l - 0 _
0 9 . 1 _ 0 _ 0
. 5 7 _
1 0 _ 0
. h 8
1 0 _ 0 _
. 5 5
1 7 0 0 0 4 i n e 1 rn 0 . uo si 0 st 0 3 ea A 1 rl T 0 P u A 0 c D 0 rl 2 1 oa tC 1 0 a
- lA 3 0 0
uY L . 1 mS 1 uP T 0 cA F' cL O 0 AE L . 0
- 0) R 1C f O oot T 0. SC n ON 0( o1
, 0 9 s -
O .E i 3 S M rL I 0. I a 0T pT R A . 0 8 mF P oO M 0 CL O C .
- 0 0 :
7 7 A - Y 0 1 5 - 0 . P . 0 V A 0 L r e E 0 r R 0 u g 1 0 5 i r A- F Y3 0 SL .0
. P 0
- AT 4 LP EO 0 r RL 0
- e 0 3
- 0
- e 0 0
2 r 0 0
. 0 1
0
; 0 eg 9E =@ og o( R@ og o O, n_ ot5t((Y n awenw on4 gHb8Ba O 88sE _
,w*I
{1r jIll t
0.V.2 clarify other tests or ac!_'url plant transients that might be L/~' available to verify other meeas or parts of areas that. require b' i verification. A.V.2 Please see the response to Q.V.1. 0.V.3 Provide justification for not performing analyses for the other tests and plant transients that could provide additional verification. A.V.3 Please see the response to Q.V.1. 0.V.9 Figure 5.3-14 of Reference 12 shows that RELAPSIA overpredicts the primary pressure decrease caused by condensation following accumulator injection. Clarify how this condensation will be modeled for SBLOCA calculations to prevent too much condensation occurring with the resulting overprediction of ECC injection rates. . r ( A.V.9 The ECC modeling guidelines described in A.II.5 to II.8 recommends artificially raising the ECC water temperature used in the analysis close'to the saturation temperature corresponding to the containment pressure. The-recommended temperature is about 200 F to avoid the possible injection of superheated water. Figure 5.3-14 described in the above question refers to the RELAPSYA prediction of LOFT Test LB-1. To check the proposed ECC modeling guidelines, the L8-1 test prediction was reat.alyzed with the following changes: a) The accumulator water temperature was raised from an original value of 88.5 F to 200 F. b) The accumulator model was replaced by the RELAPS/ MOD 1 Cycle 18 model (see A.I.23, July 1985).
>O
c) The initial conditions for Test L8-1 were taken from a L3-6 run in Which the secondary system pressure was set- to match the test data. This was'done to. consolidate computer runs. Figure V.9-1 presents the original and-revised primary system pressure compared to the LB-1 test data. The revised RELAP5YA primary pressure is: slightly above the original calculation as a result of a higher secondary system pressure. However, both calculations are somewhat below the measured data. As the accumulator injection occurs, the original RELAP5YA calculation tends to underpredict the system pressure while the revised calculation follows the data reasonably well. Figure'V.9-2 compares the peak node cladding temperatures for the two RELAP5YA. calculations and the test data. The revised calculation shows a slightly later quench time than either the original calculation or the test data. In LOFT Test L8-1. the accumulator injection was delayed. This created a higher initial pressure difference between the accumulator and the primary system that resulted in higher accumulator flow rates. This atypical test procedure causes more rapid condensation than expected for PWRs. For this reason, raising the ECC water temperature to 200 F should limit the condensation and injection rates for PWR SBLOCAs. l l
~\
[O ! l
,t
j! ; W M
-y 0
0
.' 0 5
2 0 0 5 4 2 0 1 0 - 5 4 8 2 L T s 0 F n 0 O o Li
- d2 na t
= il
= 0 ec u
= 0 rl A =
5 4 ua 2 sC T ^ = s A = 0 eA D . rY 0 P5 1 4 P
- 2 mA 8 : : eL t E L .
0 sR u-y T b) Sde F O - 4 2 C E ys L 1 S ri 0 av T O . 0 ( me
. 54 E iRr N 2 MI Pd O T f an S 0 o
I . 0 R l A 5 4 na P 2 on si M i g O 0 ri C . 0 ar i 4 pO A 2 m Y oo 5 0 Ct P A . 0 0 L 4 E 2 1 R L 9 DA 0 CN 0 V SI $ 3 e 1IG 8ER LRO VI 0 2 g i r u
~ 0 F TAA $
3 PYY 2 O55
.- LRR 0
= c. =
;; N2 0
.c 0
. n, 0 5
0 l " o0 oN W~ odg o
- - - - ~
5 3 2 0
~
9g og 0
" _g o B" w58wE <aw" "2 a- aH o$ od -
A
,wwi l l1 l I l i
I!J_ _ O _ 0
- . - - - - 0 3
2 DA EN L " 1 SI - 1IG 0 B
. VI D L 8ER I 8
T LRO 1 F' TRA O FYY L O55 LRR s-_ 0 i n
. .e- -
D e
- I 9
- 1 r e us t n ao ri et pa A 0 ml T
51 eu A Tc l D ga 1 lq> nC i
- dA 8 # % dY L 0 aS lP T . N) CA F I C 6E L O E 2R L S 0
( O 8 e d T 0 ME SiGs _ ONO S 0 I d1 T ee dR v I R o _ Nd _ A n P f a M o O . 0 l C 5 na 8 on _ A si Y i g ri 5 ar P pO A m L 0 oo E - Ct R N 2
-_ 9 0 V '
M e r u g i F
- r o d
0
~ . = - 0
=g o h'
~
, ~ ~
a.$ g~ ag ag o ak oY ad O b MMCb 5H Cz Q D O gP8m d8 ~ Lo' .
VI FLOW REGIMES t ) 0.VI.1 Areas where horizontal countercurrent and/or stratified flow of steam and water may occur are in the hot legs, cold legs and pressurizer surge line. Countercurrent flow could occur in the hot leg with steam going from the vessel to the U-tube steam genc ator, condensing and then flowing back along the hot leg to the vessel (reflux cooling). Countercurrent flow might also occur with potential slugs of liquid introduced by cold ECC water injected into a largely vapor-filled cold leg. It might also occur in the pressurizer surge line. Clarify when countercurrent and/or stratified flow will occur and how the transition from nonstratified to stratified flow 's modeled. A.VI.1 Countercurrent vapor-liquid flow is calculated in RELAP5YA as a consequence of the solution of the basic governing equations. Stratified flow is assumed only in horizontal components and only when the stratification criterion is met. This is described in the [] horizontal flow regime map shown on Page 28 of Aphendix A of Reference VI.1-1. A transition region between stratified and nonstratified flow is also shown in the flow regime map. Within this region, the interphase drag is calculated by interpolating on mass flux between the stratified and nonstratified regime boundaries. This interpolation provides continuity for the transition to either regime. Reference (VI.1-1) Fernandez, R. T., et al., "RELAPSYA - A Computer Program , for LWR System Thermal-Hydraulic Analysis, Volume I," YAEC-1300P, January 1983 0.VI.2 Page 14 of Reference 8 states: "The additional force term which arises for stratified flow geometry in horizontal pipe is added to the basic equation when the flow is established to be stratified [D 'Y'^ from flow regime considerations." Clarify how momentum surges are avoided when the stratified flow model is turned on and off. 3
. i
A.VI.2 In a typical RELAPSYA plant model, there are only a few horizontal components which generally represent horizontal runs of reactor
' ^)
\/ coolant pipes. It is not obvious that if the flow goes from nonstratified to stratified in one of these components, e momentum surge will ensue that will have significant impact on overall system behavior. The inclusion of the extra pressure gradient term in stratified flow is an attempt to improve the raodeling of stratified flow phenomena (refer to Page 14, Appendix A. Reference VI.2-1).
This term only appenes in the difference momentum equation Which primarily determines the relative velocity between the phases. This equation contains other terms (for example, the interphase drag term) which are smoothly behaved between nonstratified and stratified flows. Hence, we do not believe that a transition from nonstratified flow to stratified flow will result in a nonphysical momentum surge. It should be pointed out that momentum surges can occur due to a variety of reasons (for example, propagation of density waves). In f3J general, no attempt is made to avoid them if they are not K' accompanied by nonphysical numerical behavior and buildup of mass error. 0.VI.3 Page 32 of Reference 8 states that: for the virtual mass coefficient, C "It may be appropriate to assume that C=0 should be used for separated or stratified flows. At present, the value of C defined by Equations (109) or (110) is used without regard to the flow regime." Clarify how much different the separated or stratified flows would be if C=0 was used. l 1 i A.VI.3 A value of C=0 might be appropriate for completely separated or j stratified flows only when the two phases do not significantly ) interact. However, this will only occur when (a) the phasic l velocities are small, (b) the temporal derivative of the relative ) velocity is small and (c) the spatial gradients of the phasic l 1 velocities are small. For these conditions, the dynamic drag terms ! l /'~s in the momentum difference and the thermal energy equations that contain C will be small (see Equation 107 of the cited reference). ; 1 t . I l
l'1 g ..-
.j i
i i Therefore, we expect small phasic flow rates for completely { separated or stratified flows regardless of the value for the-
, jf]
(> virtual mass coefficient C. j
.l 1
0.VI.9 Clarify how'the vr.rious modes of single-phase and two-phase natural circulation have been verified analytically and experimentally. A.VI.9 To verify RELAP5YA's' ability to predict the various modes of natural circulation, an assessment of Semiscale Natural Circulation Test, (S-NC-2) has.been performed and is described in Appendix A.VI.9. O l O
1 V t' Appendix A.VI.9 r W
- Introduction-p An assessment of Semiscale Natural Circulation Test 2'(S-NC-2) has been performed to' demonstrate that' RELAPSYA predicts the single-phase, two-phase.
and reflux modes of natural circulation. Semiscale~ Facility The Semiscale MOD-2A test facility at the Idaho National Engineering Laboratory was used to investigate:the thermal-hydraulic phenomena accompanying various hypothesized loss-of-coolant accidents in a PWR System. The Standsed Semiscale MOD-2A System (see Figure VI.9-1) is approximately a . 1/1700 scale model of a four-loop PWR. It consists of two primary coolant loops connected to a pressure vessel which has an external downcomer. The intact loop is scaled to represent three loops while the broken loop. represents a single loop. The pr6ssure vessel contains electrical heaters to-simulate the reactor core. . Several separate effect tests were performed to examine the important system parameters during natural circulation. The Semiscale Natural circulation Test 2 (S-NC-2) was performed using a subsystem of the MOD-2A facility (see Figure VI.9-2). This subsystem had the broken loop, the upper head and the intact loop pump removed. Removing the broken loop makes the facility a single-loop system. Removing the upper head reduced the primary system volume, but did not affect the natural circulation modes since the bottom of the upper head was located above the hot legs. The intact loop pump was replaced by a spool piece to eliminate leakage from the pump. Hydraulic resistances were simulated by an orifice in the line. S-NC-2 Test Description The S-NC-2 experiments were performed in the MOD-2A facility to examine various modes of natural circulation (Reference VI.9-1). Three cases at different power levels (30, 60 and 100 kW) were examined. At each power level, the system started at 100% primary mass inventory with the pressurizer 6
\
i I establishing subcooled fluid conditions. The pressurizer was then valved out' l ,N. of the system and mass was drained from the primary system in discrete amounts. The primary pressure was allowed to float while core power and steam generator secondary conditions were held constant. . Test data were recorded when steady-state conditions were established at each mass inventory. At 100% inventory, single-phase natural circulation existed with subcooled liquid circulating through the systera. As liquid was drained from h the primary system, the liquid level in the vessel dropped. Once the level dropped below the hot leg elevation, bubbles appeared-in the hot les and two-phase natural circulation was established. As the mass inventory was further reduced, the flow rate increased until bubbles were observed in the steam generator. outlet. With continued reduction of mass inventory, reflux flows were observed in both rides of the steam generator. All modes of natural circulation were capable of removing core heat. In addition, with the exception of the reflux mode, no dryout or heater rod temperature excursion was observed. RELAPSYA Model The RELAP5YA input used to model the Semiscale Natural Circulation Test 2 (S-NC-2) is based on the RELAP5/ MOD 1 base model described in Reference VI.9-2. The RELAPS nodalization for S-NC-2 is shown in Figure VI.9-3. The following changes were made to the base model to obtain the RELAP5YA model: (a) The pressurizer component was replaced by a time-dependent volume for the 100% mass inventory cases. This change is not significant since the time dependent volume, like the pressurizer, establishes the desired primary pressure when the system is at 100% mass inventory. For primary mass inventories less than 100%, the pressurizer in the experiment was valved out and the corresponding time-dependent volume in the RELAPSYA model was eliminated. O i
7 1-
;n i
p ib')- All the passive heat structures were removed from SNLA's RELAPS. model-except the heat structures that model the' steam generator L tbes . - This' change was made to simplify the model and should be ' p insignificant since S-NC-2 consisted of steady-state tests, In addition, the tests were conducted'with strip heaters designed to offset the environmental-heat losses. Therefore, the assumption-of an adiabatic boundary in RELAPSYA adequately' represents these heat structures. Cases Analyzed The S-NC-2 experiments were performed at three power levels (30, 60 and-100 kW). We analyzed the 30 kW and 100 kW cases at various primary system inventories. The results-for these two cases are presented below.-
,Results Figures VI.9-4'through VI.9-6 conqpare the measured and calculat'ed hot-leg temperature', primary system pressure and mass flow rate as a function of ,
primary mass inventory for the 30 kW case. .The calculated hot les temperature and primary pressure compare well with the measurements. As in the data, the primary pressure decreases bec.ause the hot leg is at saturation conditions. The maximum flow in the loop was predicted wt.en the inlet flow to the steam generator U-tubes was two-phast and the outlet flow was single-phase fluid.
- The peak flow rate was overpredicted by about 25% and also occurred at a slightly higher inventory than observed in the experiment. As the system mass inventory was further reduced, the calculated mass flow rate also decreased,
- closely following the experimental trend, The single-phase, two-phase and reflux mods. of natural circulation were predicted reasonably well by RELAP5YA. Reflux was calculated and experimentally seen at the lowest inventory (approximately 60%).
At some system inventories, oscillations in mass flow rate were calculated. The range of oscillations at these inventories is indicated by vertical bars in Figure VI.9-6. The experiment shows flow oscillations at some system inventories as well. The exact magnitude and period of these oscillations is difficult to determine due to poor resolution of presented experimental data [ Reference VI.9-1).
Figures VI.9-7 through VI.9-9.show the measured and calculated hot leg
/~] temperature, primary system pressure and mass flow rate as a function of '# system inventory for the 100 kW case. As in the 30 kW case, the calculated temperatures and pressures compare well with the measurements. The maximum flow in the loop was predicted when the inlet flow to the steam generator U-tubes was two-phase and the outlet flow was single-phase fluid. The maximum flow in the loop was overpredicted by about 25% and occurred at a slightly lower inventory than the e::perimental data. As the system inventory was further reduced, the calculated mass flow rates also decreased. The experiment identiiled reflux cooling to occur for the lowest inventory (56.9%)
shown in Figure VI.9-9. This mode was not calculated to occur at this inventory. However, at a lower inventory (55%), periodic countercurrent flow was calculated to occur in the hot legs accompanied by periodic core heatup. This is indicative of the beginning of the reflux cooling mode. As in the 30 kW case, mass flow oscillations were calculated at some inventories. The cause of these oscillations are also believed to te due to an unsteady calculation of void distribution. (N) Conclusions , t xs The results of our natural circulation calculations show that RELAP5YA qualitatively describes all modes of natural circulation correctly. Tho calculated mass flow rates were generally higher than the data. The impact of these higher flow rates should not be significant from the point of view of LOCA analysis, since adequate core cooling is maintained for the single-phase and two-phase modes of natural circulation. l
}
References ] (VI.9-1) O'Connel, Thomas M. , NUREG/CR-2454, EGG-2141, " Experimental Data Report for Semiscale MOD-2A Natural Circulation Tests S-NC-2B, S-NC-3 and S-NC-4B," December 1981. (VI.9-2) McGlaun, J. M., and L. M. Kmetyk, NUREG/CR-3258, SAND 83-0833, "RELAp5 Assessment: Semiscale Natural Circulation Tests
./~'N S-NC-2 and S-NC-7," May 1983.
% j'
\
L O.VII.3 f- s Pages 18-20 of Reference 12 show the RELAP5YA results compared with (_.,) the test data for a GE level swell test. Again, the predictions are under the data for low void fractions. Clarify why at 10 s, 100,s, 150 s the RELAPSYA void fractions decrease significantly in the upward direction-in the volume just before the steep increase'in void fraction occurs. A._VII.3 RELAPSYA does not contain models to predict.and track two-phase levels within a volume. In the calculation of the interphase drag term, F ,, at junction locations, a volume-weighted void fraction 7 is used. This would result in some smearing of liquid above the-location of a two-phase level and, hence, some inaccuracy in the prediction of interphase drag near regions of steep void gradients. Also, RELAP5YA uses a donoring scheme to determine void fractions and fluid properties at junction locations. This would also have some impact near regions of sharp void gradients because of the occurrence of countercurrent finws. It is believed that the above characte .stics combine to yit ~ che void profiles referred to in c-the question. Howeveb, the di) in the void prediction below the
- (,,/ -
level does not propagate to other regions and did not result in a buildup of mass error. In the calculation, the location of the level is assumed at the location where the void fraction reaches a value near unity. In general, the calculeted level is always below that observed in the data. This is in the conservative direction and believed to be acceptable for licensing analyses. O.VII.4 Clarify why the void fraction at the 10 and 12 ft elevations takes so much longer to reach 1.0 for the calculations than the data. A.VII.4 There is an error in Figure 2.1-14. The data are the solid lines and the RELAP5YA calculations are the dashed lines (the reverse is indicated by the figure). The calculated void fraction at a given location reaches unity faster than indicated by the data. The corrected figure is attached. m 5
'**' . . . . . . . . y r w .
s'
/
; .a . ,s .
\J s s
.e . s -
t
.7 . .
f E
.s .
DATA ,/ . 1 i RELAPSYA l .s . j . l .4 ,.
/
/
j .
.s . f .
1
.t .
#,a~~ . )
.a#
.1 . .
/
o.o ' ' ' ' ' ' ' ' o.o 1.o t.o s .'o4.o s AItAl. .'o s.'o , 7.o s .'o e.'o 10 0 11 0 12.o ts.o 14.c CLt.'YAT1 set. IrT Figure 2.1-13: Axial Void Distribution at 150 Seconds for GE Level Swell Test
- 1. .
I g
*# ,# DATA -
/ --- narm
, a
.so .
l a 12 ft.
/ 10 ft. .
.To l .
8 I E .so . ,' y' . C l t i i W a *50 - ta I . b I3 i O II d j .4o . . l
\ t
.so . l I s .
If I
\4
.to .,s o ' , a'% *g ,7 A
2 ft,. ^ o '
.to 87 , s ~ *- *- a- * *- #*- *- * *- 'i-
- L
- b " *- *- *- *- *
- r o.o0- ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
o to to So 4o 5o So To to to too 110 120 13o 140 15o 180 170 100 T!ftE IStri
> Figure 2.1-14: Nodal Void Fraction Versus Time i
V 0.VII.6 Page 78 of Reference 10 states: "RELAPSYA uses a large finite value (10 N-S/m ) for F in the void fraction range 0.0<=( <0.001 ('^ 7 b- t'o ensure homogeneous flow behavior. To provide continuity with < this scheme, a narrow-range of void fraction is defined as 0.001<of<0.01. In this void fraction range, F is caledaud by 7 exponen ial interpolation between 10 N-S/m and the value calculated by the undistorted bubble model at e( = 0.01." Clarify if the F is used directly or if a relaxation method is applied. 7 A.VII.6 Please see A.VII.8. O.VII.7 Clarify if this large change in F with 7 small changes in void fraction can lead to instabilities. A.VII.7 Please see A.VII.8. 6.VII.8 Page 101 of Reference 10 states: "After the interphase drag term F is calculated for a particular junction in subroutine FIDRAG, 7 m it is averaged over space and time." Clarify how this averaging is k done and how long it takes for a very large F 7to change to the much lower values for void fractions greater than 0.01. A.VII.8 The interphase dras coefficient 7F is first calculated according to the methods described in Section 3.1 of Reference VII.8-1 for each hydrodynamic junction using junction properties. This value, designated as Fg , is further adjusted as follows: Civen F 7 ,* for all junctions:
- a. Volume-based F are calculated using inlet and outlet 7
junctions: F 7
= 1/2 [F 7y +
IJ y IN O
- b. Junction-based F are recalculated using volumetric weighting 7
~~
of volume-based F values of adjacent volumes: 7 s
** V F
IJ
= IVK
- K+F7 V , L L
K+ L i
- c. At each junction, the F 7is time-averaged as an arithmetic average between old time value and the latest value to give the new time value:
= F +
F"IJ IJ F"IJ j 2 Thus, in response to Q.VII.6, the F7 calculated by the models in Section 3.1 of Reference VII.8-1 (designated above as 7F ,9 are j not used directly, but some under relaxation is provided as ! described above. In response to Q.VII.7, the above scheme will bias the final value of Fg towards the larger value between F g ** and F" . In general, larger values of F g will provide , p/ C stronger coupling between the phases and, hence, less oscillatory behavior. Thus the above averaging scheme is believed to help in mitigating instabilities. In response to Q.VII.8, the impact of averaging F7 as described above will depend on the node and time step sizes, as well as the type of problem being analyzed. The motivation behind this averaging procedure has been discussed in Section 3.1.3 of Reference VII.8-1. Its largest impact will be when a junction undergoes a change between near-single-phase conditions (voids below 20 4 N-S/m ) to two-phase 0.001 or above 0.999. F7 = 10 conditions (voids between 0.01 and 0.99, F7= 10 to 10 N-S/m ). The time-averaging scheme will allow this transition over about 50 time steps. For small break LOCA analyses, the time step sizes are typically 0.01 to 0.05 seconds. Thus, the impact on the transient will be a delay of 0.5 to 2.5 seconds in the transition between near-single-phase and two-phase conditions. This g (/ appears reasonable for small break LOCA transients which generally occur over a time scale of 10 to 10 seconds. i
Reference
. ,.- g
( ) (VII.8-1) Fernandez, R. T., et al., "RELAPSYA -'A Computer Program for LWR Thermal-Hydraulic Analyses, Volume I," YAEC-1300P, January 1983. O.VII.23 Page 2-74 of Reference 19 describes the Yankee Atomic earlier experience in modeling a Lehigh University low flow film boiling test with RELAPS: "In this case, the code underpredicted the wall temperature profile because the heat' transfer coefficient (based en T ) was overpredicted by approximately a factor of two. Also, the code does .not predict the thermal nonequilibrium - (approximately 400 F of vapor superheat) measured in the test. This illustrates the need for improvements in modeling low flow film boiling heat transfer and nonequilibrium vappe generation rates." Clarify if RELAP5YA>has been used to model this same test because the test provided such a severe test for RELAP5. A.VII.23 The reference cited in the question deals ,with YAEC's initial Essessment of RELAr*>/ MOD 1 for predicting post-CHF heat transfer phenomena. Subsequent review of the Lehigh test indicated that there were many uncertainties in modeling the Lehigh test facility, especially with regard to the " hot patch" at the test section inlet designed to generate post-CHF conditions in the test section. , Hence, alternate tests have been used for RELAP5YA assessment, described in Reference VII.23-1. l Reference (VII.23-1) Fernandez, R. T., et al., "RELAP5YA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume III," YAEC-1300P, January 1983. i
)
0.VII.25 Clarify how detailed the noding needs to be to get proper superheating rat.her than forcing the steam to stay at saturation ; because it is the minimum phase in a large volume partially filled with liquid. A.VII.25 In general, RELAPSYA will not predict vapor superheating to occur in
'~~
3 the presence of the liquid phase. Hence, to calculate superheating ,
)
~- of steam, the void fraction has to reach 1.0. This can happen if ) 1 (a) the noding is detailed enough to show the existence of a two-phase level, if it can occur; and (b) the heat supplied to the i volumes above the level is high enough to completely evaporate entrained droplets. .This can be seen in the RELAPSYA assessment i against the THTF boiloff test, 3.09.10I, described in Section 5.1.5 of Reference VII.25-1, and further illustrated in Figures 5.1-15 and 5.1-16 of Reference VII.25-1.
Reference (VII.25-1) Fernandez, R. T., et al., "RELAPSYA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume III," YAEC-1300P, January 1983. 0.VII.26 Clarify if forcing entrained liquid drops to be at saturation reduces the superheating of the vapor. r) L.) A.VII.26 Entrained liquid drops, whether at saturation or not, will reduce superheating of the vapor. Further, in PWR LOCA analyses, it is difficult to envision situations of significance where entrained liquid droplets will not be at, or very close to, the saturation temperature. The vapor temperature is, however, more influenced by the interphase mass transfer coefficient than by the liqua.d temperature. RELAPSYA generally uses a very high mass transfer j coefficient and, hence, will predict very little vapor superheat in the presence of entrained liquid droplets. This results in higher ! vapor generation rates. However, this leads to conservative results for LWR LOCA analyses as discussed in Section 2.1.3.1 of Reference VII.8-1. 0.VII.27 Clarify and justify that the range of CHF data and post-CHF data covered by the experiments includes the range of conditiond expected j'~3 for SSLOCA calculations.
! /
v 5
i 1 A.VII.27 Reference VII.27-1 presents the assessment of the CHF and post-CHF ]: fg mcdels implemented in RELAP5YA. k.) - CHF Assessment i The criteria considered in selecting the steady-state CHF tests include the range of thermal-hydraulic conditions expected for SBLOCA calculations. Twelve tests were selected to assess the CHF options. The tests were performed.in three different facilities as follows:
- 1. Five tests in the medium pressure heat transfer flow J. cop at the Chemical En61neering Research Laboratory of Columbia University (References VII.27-2 and 3).
- 2. Four tests in the Nine Rod Test Section at General Electric (Reference VII.27-4).
< , 3. Three tests in the Thermal-Hydraulic Test Facility at Oak Ridge L(. National Laboratory (Reference VII.27-5).
The range of conditions for all the steady-state tests is summarized in Table VII.27-1. Further assessment of the CHF models Implemented in RELAPSYA has been performed for the system tests described in Section 5.0 of Reference VII.27-1. Post-CHF Assessment The most relevant tests for the post-CHF assessment are:
- 1. The steady-state film boiling tests, 3.07.9B, 9K and 9X, conducted at Oak Ridge National Laboratory in the Thermal-Hydraulic Test Facility. These tests are described in Table VII.27-2. The tests provide steady-state film boiling heat transfer data in rod bundle geometry. These data were used in assessing film boiling heat transfer correlations. The 7-~s
( ,) heated pin diameter and rod pitch are typical of later
-Se-
\
x g i
, generation PWRs with 17x17 fuel bundles. The conditions in'the tests K and I are representative of conditions expected during: \ -~ SBLOCA transients.
- 2. The quasi-steady-state bolloff test, 3.09.10I, conducted at Oak l Ridge. National Leboratory in the TNTF (Reference VII.27-6). .
The objective of the bolloff test series was to study the heat ]
. transfer and mixture < level swell under SBLOCA conditions in Pressurized Water Reactors (PWRs).
The conditions for Test 3.09.10I are shown below: System pressure 650 psi-Mass flux 2.19 x 10 lbm/hr-ft Linear power / rod 0.68 kW/ft
- 3. The system bolloff experiment, Test 6441/6, conducted at General Electric, San Jose in the TLTA facility (Reference VII.27-7).
The bolloff tests attempted to simulate system conditions which might occur during a small break LOCA in'a BWR if none of the Emergency Core Cooling Systcms, as well as the ADS, were available. In these tests, the recirculation loops were blocked and the liquid inventory was slowly boiled off at a constant pressure and constant bundle power. The power level was representative of decay heat in a BWR. The main objective of these tests was to evaluate heat transfer in a partially covered bundle at decay power levels and low flows. The phenomena and the range of conditions encountered in this test might be encountered during SBLOCAs in PWRs before the accumulator actuation if the flow from the high pressure safety
.O injection pumps bypasses the core.
I i l
~;
The initial conditions for Test 6441/6 are: s-n 1 ! ). M- . System pressure 395 110 psia-Bundle power 250 12 kW Initial two-phase level Bundle top LOFT Tests L3-6/L8-1, used in the assessment of RELAPSYA, simulate a small break LOCA with core uncovery (Reference VII.27-8). Test L3-6, which was a SBLOCA with the pumps running, was extended into a more severe transient, Test L8-1, which produced core uncovery and heatup. These experiments were configured to simulate a small break equivalent to a 4-inch diameter rupture in the cold les of a large (approximately 1,000 MWe) commercial pressurized water reactor. In Test L3-6, the primary coolant pumps were operating until the hot leg depressurized.to 311.83 psia. The coolant pumps and High Pressure Injection System (HPSI) flow were then terminated and the break lef t open. Experiment L8-1 p started when the primary coolant pumps were tripped. When the , U maximum' fuel cladding temperature reached 600.53 F, core reflood was initiated. Experiment L3-6/L8-1 was initiated from primary coolant system conditions of: Hot les temperature 579.1 1 3.24 F cold leg temperature 544.5 1 3.24 F Hot leg pressure 2,156.7120.3 psia Intact loop flow rate 1,065.48 5.72 lb/sec Power level 50 i 1 MW Maximum linear heat generatien rate 16.06 i 1.12 kW/ft A detailed account of the integral test conditions is presented in Reference VII.27-1.
.A
(
1 l i
\
l l REFERENCES A.1.9-1 R. T. Fernandez and H. C. da Silva, " Vermont Yankee Bh'R
.'\,. '
Loss-of-Coolant Accident Licensing Analysis Method," YAEC-1547, June.'1986. l l i l 3 i i O i i
V t- .i
.J
)
TABLE A.I.9-1 . Summary of the Vermont Yankee Small Break Accident' Assumptions CASE EW
.p-O 1. Small recirculation discharge' break-(0.05 ft 2) at 4.0E-6 seconds.
- 2. Loss of auxiliary power occurs at 4-.0E-6 seconds. .
- 3. Reactor scrams after 0.5-second delay from first.RPS signal. Scram curve 678-EOC is used.
'A. Feedwater coasts down to 0.0 lbm/sec at 5.0 seconds.
- 5. MSIVs close.in 10.0 seconds after isolation signal plus 0.5-second delay.-
- 6. Recirculation pumps in A and B loops coast down with decreasing power from loss of MG sets.
- 7. ADS may actuate if appropriate' signals exist. Thereafter, ADS cycles open/close at 12 psid between steam line and drywell any time ADS criteria are currently met.
- 8. HPCI steam turbine admission valve fails to open on demand. Thus HPCI fails to inject. (This is the single failure.)
- 9. No credit for RCIC operation.
- 10. Two LPCS Systems inject on demand.
- 11. LPCI-A injection valve opens upon demand.
- 12. LPCI-B injection valve opens upon demand.
- 13. Drywell pressure and quality are assumed constant at 16.4 psia and 1.0 for fluid sink conditions. High drywell pressure is conservatively estimated to occur at 18.4 seconds for this case by a containment calculation.
- 14. Wetwell pressure and temperature are assumed constant at 14.7 psia and 1650F for fluid source and sink conditions.
- 15. EM point reactor kinetics initially at 1,664 MWth.
- 16. EM core heat transfer.
- 17. Passive heat structures are included.
- 18. Moody two-phase crit 3 cal flow model used at the break location. )
- 19. 1971 ANS Decay Heat Standard plus 20%.
r TABLE A.1.9-2
~~ Sequence of Events for Small Break Case EW F
Event Time (sec).
- 1. . Break opens. 4.0E-6
- 2. Loss of auxiliary power. -4.0E-6
- 3. Control rod insertion initiated 0.5 second beyond estimated, 3.56 RPS underfrequency reactor trip signal. '
- 4. MSIVs begin to close. 3.56 5; Feedwater flow coasts down to zero. 5.0
- 6. MSIVs completely closed. 13.56
- 7. Low-low level signal. 15.6
- 8. Recirculation pump motors trip on low frequency at their 17.0 MG sets.
- 9. High drywell pressure. 18.45
- 10. Turbine stop valve begins to close. 18.9 O 11. Turbine bypass valve begins to open. 19.05
- 12. Turbine bypass valve completely open. 19.65
- 13. ADS valves open. 138.4
- 14. Earliest nodal CHF. 222.0
- 15. LPCS injection begins. 338.4
- 16. Recirculation loop discharge valves begin to close. 331.8
- 17. Minimum primary system inventory (161,927 lb) occurs, 337.6
- 18. Peak clad temperature occurs (inner 893.70F; 351.4 outer 891.80F).
- 19. LPCI begins to inject. 335.6
- 20. Average core and high power regions are well cooled. 374.0 0 -
- _ - _ - _ . _ - _ - . _ _ . __ - - _ _ _ _ -_ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ __}
VDMONT YANkCC etSSS L11W5fND #0001 CRSC CWs SNHLL ORCAK LOCR fePCN01W E ItCSLTS ItCCIRC LOOP DISOPMRGIC P1PC BRCN (0.05 f'72)
. ,/ bo V) YR l 4
9 l E E o h O x 3-o d 15.0 20.0 35.0 0.0 5.0 10.0 TIMC (SCO)
'D Figure A.1.9-1: Reactor Power History (SBLOCA-EW) 'V '
o d - n o -- d_ n o l US
-R g.
1 l e - h o f~ l o 10.0 15.0 20.0 25.0 0.0 5.0 ( ) IIMC (SCC) v Figure A.'l.9 2: Net Reactf.vity (SBLOCA-EW) i I . 1 L :
l:n .. P VERftDNT YMEE N$s$ LICDeSING MODEL
. @JC CWs S'tRLL 8RCAK LOCA RPPpoig g Egir.ts .
, RCCIRC LOOP DISOeREC PIPC ORCM 10.05 0721 l; J f L j. - -
g No - y <
./ . ,.
o s' N-N
\ e- *TCCDWRTER e--eNRIN STEMLINE g, - \,
N ',
\,
"o \-
ag . E 'N
- N Ld \_-'-
15.o 20.0 25.0 c.o E'.0 10.0 TINC (SCC)
~
Figure A.1.9-3: Feed and Main Steam Flows (SBLOCA-EW) o N l_ ^ - h-
=-
Yo ( s g- 3 l
* \ ~~
EE l o x - _ _ 3 o S as.c 5.0 10.0 ts.o 30.0 c.o TINC (SCC) Figure A.1.9-4: Vessel Water Level (SBLOCA-EW)
~
vemwi ve*TC 96SS L1CCNSles pC001 cast rue se.t. tWtrnK LOCA fVPCNDIN K RCSLTS
, ftCCIRC LOOP 01NMRGC PIPC 6 (0Xa ml / '; g j -
E Mo I N o g N' j ._._ a w N' \ s . ro.O 3m.0 450.0 o.o 90.0 180.0 TINC (SCC) r Vessel Pressure History (SBLOCA-EW) Figure A.1.9-5: (]h = 5o v> R - b~ e-e SRV1 FLOW
'3 l e--e fl0S TLOW i
=o 'I
_ q f h ! He \ p - I s t, w % 4,(
! t i-- I- ::::::::: : ::
G ' 3 ;; o.o so.o too.o TINC (SCC) no.o 380.0 450.0 rigure A.I.9-6: S/RV and ADS Flowrates (SBLOCA-EW) _ _ _ _ - . _ _ _ _ _ _ _ _ _ _ _ _ l
, 1;-
WERNDNT YfMT.C 1635 LICDit!W 19011, ORSC Du 9NRL18RERK LOCA frPDef t K acsts.ts 8CEIRC LOOP Of5DfWtK PIPC 5tERK 10.D5 fT2) a i N P l l N d e 100.0 Wo.0 380.0 e.0 C.D 80.0 71MC (SCC) Figure A.I.9-7: Break Flowrate (SBLOCA-EW) ( n I d f E p . 2 Ej. J
- l. 4 ,-
j-i o,o p0.0 880 0 8D iso.c O so.o TIME (SCC) Figure A.I.9-8: Break Junctions Void Fraction (SBLOCA-EW)
i VCRNONT YfWKCC NS$$ LITN$1NG N00CL CMSC CWs SNRll BRCRK LOCR MPf'CNDix K ACstA.T$
)
, RCCIRC LOOP DISCmRGC FIPC 9RCPW (0.05 FT29 f
I n ' Y, 4
" ~
l i
~
M (
$a$
ld h g
)
6-c::: c.o so.o too.o
- d no.o sec.o 450.0 TIMC (SCC)
,A rigure A.1.9-9: LPCS Flowrate (SBLOCA-EW) a 8 i o
'!I i is g g
.N o hg.
, i O m--eLPCI A ll l
~*
o--cLPCI B l
.' l l
* .i
,' I n
-. t p
.. 'qi t g
14 T I ! U ,
$ I
, /:
l i l
- l I
i
; .a.
i , l l d _ _ . _ __ __ _____ _ TINC (SCC) Figure A.I.9-10: LPCI A and B Flowrates (SBLOCA-EW)
. i ERNMI YR*lE NS$$ LICCNSING N00Q, CRSC CWs SNRLL BRCRK LACR RPPCNO]X K RCSLLTS y1RCt.00POfSCHRRGEPIPCORCN(0.050T21
; r -4.. ga E
o m h e-sa Y N Co e No. o 270.0 360.0 450.0 0.0 90.0 180.0 TIMC (SEC) Figure A.'1.9-11: Net Flowrate into NSSS (SELOCA-EW) we x8 o N N w
/
lg- _- - m
\ .N ._
/ !
0 - N/ a I b.
.? =
o sec.o ese,o 80 0 to'0UD wo.o TINC (SEC) Figure A.I.9-12: NSSS Fluid Mass Inventory (SBLOCA-EW)
i t i ytmwr wuct wss3 t.1ccustNG 9000l, f msC tv smu. entw t.oca nPPENDIX k RCSu,TS ftCCIRC (DOP. DISCHRRGC PIPC RCRg (o.05 FT21
/^^\ : Dq
** l I.j 0- a--c BYPRSS REGION 0- oDPPER PLENUM me g.
., 1
}
u) [ I
$5 _9 _\
in. )
\
\ ,
h we g ,-
- o d I
- - s 1 A 3 gi t p e
= == w too.o ' a/o.o aso.o esa.o o.o e6.o TINC (SEC)
[ Figure A.1.9-13: Bypass and Upper Plenum Fluid Mass (SBLOCA-EW) b
'ii' 5!
a .
- b m; e - j J
3- . N j d= o l E
$.-.o.....
i .... ) .. ....,q0 r".
%e- - 1;^9 ,e + w w + o'e ei r
v' )., o,a a a--eLOKER PLENUM i- WI-~ ,e--oCEG7 REGION b l , eso.o c.o so.o too.o 270.0 seo.o f-i TINC (SCC) Figure A.1.9-14: CAGT and Lower Plenum Fluid Mass (SFLOCA-EW)
,_ . -References I h
, \_s .
VII.27-1 Fernandez, R. T. , R. K. Sundaram, J. Ghaus A. Husain, J. N. Loomis, 1 L. Schor, R. C. Harvey and R. Habert "RELAPSYA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume III: Code Assessment," YAEC-1300P, Volume III (October 1982) (Proprietary). VII.27-2 " Critical Heat Flux Correlation for CE Fuel Assemblies with Standard Spacer Grids, Parts 1, 2. Non-Uniform Axial Power Distribution," Combustion Engineering Topical Report CENPD-207, June 1976 VII.27-3 Electric Power Research Institute Report, EPRI-RP-813-1 (to be published). VII.27-4 'Janssen, E., "Two-Phase Flow and Heat Transfer in Multirod Geometries, Final Report," General Electric Company Report GEAP-13347, March 1971 _(,A) VII.27-5 Yoder, G. L., et al., " Dispersed Flow Film Boiling in Rod Bundle Geometry - Steady-State Heat Transfer Data and Correlation Comparisons," ORNL/5822 (to be published). VII.27-6 'Anklam, T. M., et al., " Experimental Investigations of Uncovered Bundle Heat Transfer and Two-Phase Mixture Level Swell Under High Pressure, Low Heat Flux Conditions" (Final Report for TRTF Tests 3.09.10I-N and 3.09.10AA-FF-DRAFT), Oak Ridge National Laboratory. September 1981 VII.27-7 Seady. D. S., and R. Muralidharan, "PWR Low-Flow and Bundle Uncovr.ry Test and Analysis." EPRI Report No. NP-1781, June 1962 , I l VII.27-8 Bayless, P. D., and J. M. Carpenter, " Experiment Data Report for - LOFT Nuclear 3 mall Breah Experiment L3-6 and Severe Core Transient Experiment L8-1," N'JREG/CR-1868, January 1931.
~A U
5
l l Table VII.27-1 I Steady-State CHF Test Conditions , Average Average Heated Rod- Heated Pressure Mass Flux Heat Flux Diameter Length Experiment .(psi) (Mlb/hr-ft 2) (MBtu/hr-ft2) (in.). :(in.) Columbia 1,500 - 2,005 1.968 - 2.008 0.282 - 0.436 0.382 .150.0 CE Nine Rod 997 - 1,005 0.249 - 1.248 0.289 - 0.522 0.570 72.0 ORNL THTF 635 - 1,849
- 166 - 0.525 0.14 - 0.29 0.374 144.0
-Table VII.27-2 THTF Steady-State Film Boiling Data Ranges O Pressure Mass Flux Heat Flux Test No. (psia) 2 (lbm/hr-ft 2. (Btu /hr-ft 2),
3.07.9B 1,849 5.25 x 10 2.9 x 10 3.07.9K 635 1.66 x 10 1.4 x 10 3.07.9K 872 2.50 x 10 1.9 x 10 0 I
0.VII.38 Page 215 of Reference 10 describes the calculation of internal gas
- pressure for the fuel behavior model, Clarify what values of the user-specified parameter,A Typ, are recommended for the range of SBLOCA transients expected to be modeled.
A.VII.38 The RELAPSYA fuel behavior internal pressure model has been improved to achieve greater accuracy through more detailed modeling. The former fuel behavior model assumed that the fuel rod internal pressure was proportional to the absolute coolant temperature adjacent to the fuel rod plenum plus a user-specified offset. The accuracy of the former model was limited by the specified temperature offset and the neglect of geometric changes. The new internal pressure model addresses both of these limitations. The new model is based on the fuel behavior models in RELAP4/ MOD 3 (Reference 1) and TOODEE-2-EM (Reference 2). The only difference is that RELAP5YA neglects fuel pellet cracks. The internal pressure calculation in RELAP5YA is now baseo on the volume and temperature in the fuel-clad gap and plenum. The plenum gas temperature is assumed to b,e the temperature of the adjacent coolant plus an '
" offset. Yankee uses a 2 F offset in our current licensing calculations for large break LOCAs. The same offset value of 2 F will be used for small break LOCAs as well. !
The fuel rod internal gas pressure is calculated via a two-region model using the ideal gas law. The model considers a plenum region and a f.nel-cladding gap region. The internal pressure, Pid, is calculated by the ideal gas law as: R . M T int " P i=1 gi (v 4
.___7-_____
where: (_- ' total gram-moles of gas in the fuel rod M = T R = universal gas constant V-p
= plenum region volume (m )
Tp = plenum region gas. temperature (des-K) (assumed to be adjacent coolant temperature plus a user-specific offset) n = total number of axial fuel segments V = fuel clad gap volume of axial segment i T = fuel clad gap temperature of axial segment l' The plenum volume is calculated using the hot dimensions of the fuel and cladding such that: Vp= _V - TT (rg +4 rg ) ALg +[I(rd +O # c) A L c where: V
- 3 l
V = initial free volume in the plenum region (m ) po e initial cladding inner radius (m) d-Ar = change in cladding inner radius taken as that of the top axial node (m) e g= initial fuel outer radius (m) Ar g = change in fuel outer radius taken as that of the top axial node 6L = total length change of the cladding due to thermal expansion (m) AL g= total length change of the fuel column stack due to thermal effects (sum of axial thermal expansion at the pellet dish node (m) N]./ l- \
~
;I p; : References LO'
'u ' (VII.38-1) "WREM: ' Water Reactor Evaluation Method (Revision'1),"- NUREG-75/056 Division of Technical ~ Review, U.S. Nuclear Regulatory Cotanission (USNRC), ' Washington, D.C. , May 1975. (VII.38-2) 'Lauben, G. N., "T00DEE2-EM A Two-Dimensional Time-Dependent Fuel Element Thermal Analysic Program," Division of Technical Review, USNRC,'May 1975. O
l i IX. BREAK FLOW
- l p
i
$' 'h O.II.1 Investigations and experiments "...have demonstrated a wide ;
1 variation in mass flux as a function of break geometry. Mass flux j i was shown to be influenced by the degree of curvature at the break j inlet, flow passage diameter, flow passage length and the ratio of j the break diameter to the vessal diameter. Correlations incorporating all thsse factors are 'not available at the present time. Moreover, small break geometries postulated for reactor systems could range from splits in pipes to double-ended breaks restrained by pipe supports, and could include full ruptures in small diameter pipes" (Page VIII-40 of Reference 2). One approach the submittals might use to account for uncertainties in break flow would be to model a spectrum of break sizes with enough size range to account for the uncertainties in break flow models. However, note that the uncertainty for liquid break flow is different than for two-phase break flow. Clarify how the uncertainties described above will be treated in SBLOCA analyses. (' Nm A.IK.1 YAEC will provide a break spectrum study in our plant-specific LOCA analysis submittals as required by Paragraph I.C.1 of Appendix K to 10CFR50.46. Each break spectrum study is aimed at achieving the following goals:
- a. Identify the maximum cladding temperature for each accident category to assure that 10CFR50.46 criteria are met.
- b. Account for break flow modeling uncertainties.
i
- c. Minimize the number of esses to be analyzed in order to achieve the above.
Each break spectrum study will use the RELAPSYA subcooled choking model for liquid discharge and the Moody critical flow codel for the jN two-phase and steam discharge in our licensing submittals.
\- Substantial evidence shows that the uncertainties for each model,
derived by comparing predicted to measured break flow rates, are f] er'entially the same. Therefore, a separate variation of the discharge coefficients for each break size is not necessary since the variation of break size within the spectrum will account for the uncertainty of the two break flow models. This is further explained below. Subcooled Discharge Coefficient Variation Durinz Blowdown Comparison of predictions by the RELAPSYA subcooled choking model to subcooled discharge data from LOFT and TLTA yields variable discharge coefficients during subcooled blowdown thnt generally lie in the range of: 0.6gCD-SUB k* where D-SUB " EST JiYA-SUB* , V This is shown in Figures 5.2-5, 5.2-26 and 5.3-5 (attached) from Reference II.1-1. These tests used relatively small nozzles as shown in Table IX.1-1. This same trend is shown in Abdollahian's comparison of predictions from the Burnell subcooled choking model to Marviken subcooled discharge data. This is shown in Figures 5.1 through 5.4 (attached) from Reference IX.1-2. Please note that the RELAPSYA subcooled choking model is very similar to the Burnell model (References II.1-3 and II.1-4). These Marviken tests were conducted with relatively large nozzles as shown in Table II.1-1.
)Londy Discharge Coefficient Variation During Blowdown Comparison of predictions from the Moody Model to two-phase s discharge data from Marviken also yields variable discharge coefficients that generally lie in the range of 4
i lt
~j 0.6 p D-MOODY
- l (q
. wj where I
{ C D-20
- EST b ODY' This is shown by Abdo11ahian in Figures S.1 through 5.4 where he has used the upstream thermodynamic state at each point in time.
BREAK FLOW MODELING UNCERTAINTIES The preponderance of these data show that the discharge coefficients for each model applied over a range of conditions are essentially j comparable and lie in the range from about 0.6 to 1.0. These discharge coefficients are directly related to the uncertainty of each break flow model as follows: Uncertainty, % = ODEL ~ EST} t . EST
= ~ *
( CD Since the discharge coefficients for both models' span a similar range, then we infer that these models contain similar uncertainties that can be treated as equal. This implies that parametric studies could be performed using the same discharge coefficient for each model to bound the uncertainties. However, this same result is obtained when we vary the break size within the break spectrum i study. Therefore, we conclude it is not necessary to separately vary the discharge coefficients for each break size. We further observe that if a break size is dominated by either subcooled or two-phase discharge, then variation of the break size alone is sufficient to bound the uncertainty of the dominant model j
. since that variation is equivalent to a variation of the discharge
.() coefficient. Finally, we note that errors in break flow modeling
+. . - - _ _ _ _ _ _ .
'l i
can be somewhat self-correcting. This stems from the fact that both
! (' the subcooled and two-phase discharge flow rates decrease with-( _ pressure. Thus, if a model predicts too large a flow, the' predicted pressure will drop more rapidly and reduce the flow. The converse also occurs. This may explain why computer code predictions with fixed discharge coefficients often yield results that compare well to test data, provided that the fixed discharge coefficient is reasonably close to the mean variation of actual values.
Based upon this background, YAEC proposes to use the following method to account for break flow modeling uncertainties within the
' break spectrum study:
- 1. Define the break size spectrum for.each accident category based upon plant hardware features, system capabilities and single failure assumption.
- 2. Analyze selected break sizes that span this range. Use single and two-phase discharge , coefficients of unity, and the Moody
{ 'two-phase critical flow mcdel. Determine the break size within this spectrum thLc yields the maximum cladding temperature. References (IK.1-1) Fernandez, R. T. , R. K. Sundaram, J. Ghaus. A. Husain,
'J. N. Loomis, L. Schor. R. C. Harvey and R. Habert, "RELAPSYA - A Computer Program for LWR System Thermal-Hydraulic Analysis, Volume III: Code Assessment,"
YAEC-1300P, Volume III, October 1982 (Proprietary). (II.1-2) Abdollahian, D., et al., " Critical Flow Data Review and Analysis," EPRI NP-2192, Electric Power Research Institute, Palo Alto, CA, January 1982 (II.1-3) Tong, L. S., " Boiling Heat Transfer and Two-Phase Flow,"
/ John Wiley & Sons, Inc., New York, 1965 t
h" <, . , c. ) 9 V l-(11.1-4) Ranson,.V. H., R. J. Wagner. J. A. Trapp, K.-E..Carlson,- 73
.* D. M. Kiser, H. H. Kuo, H. Chow, R.!A.. Nelson and S. W. James, "RELAPS/ MOD 1 Code Manual, Volume 1: System Models and Numerical'. Methods," NUREG/CR-1826 EG&G Idaho, Inc., March 1982 Pages.54 and 55
.(
i
g @ Li Table IX.1-1 h Critical Flow Test' Nozzle Geometry ( .,. Nozzle Diameter. , Aspect Ratio Test (un) (L/D)
-TLTA 6425~and 6426 18.9- 9.42 LOFT L3-6 16.2 3.34 Marviken Test.4. 509 .3.1 g : Marviken. Test 24 500- 0.33 Marviken Test 18 300 3.7 '~
Marviken Test 6 300 1.0
- (~
A o O t
<( 0 1 3 5,
)
C E 9. _ S 0
/ (
) )
M 3 2. A I B T L (2 (3 - L T
=
O L
- i r -
F S _M /
' 0 3
E s T 2 N T R A l D A F 2 0 I* G Y 1 5 k T P C g A U L E B N S U 4 g R R , d i 1 p S C
> i i
5 2 l t 6 h u~
,N b 0 5
1
)
S ( c e t S f o v / # E T ! m I 3 E T H 2 3
= A
* [
( 9 M _ T i C g,
- O ,
e 5 8
- s. 0 # 2 s
y 2 l 0
- 7 5
, 8 g T d 2 S G 2 -
9 7 D i F 0 0 C
+
& j h
i
,I $ ,
0 1 0 0 0 2
~ /
Q d @! ; i I
,b 0
1 3 I 0 C , E = S
/
M Y
- B
- L E0 O
- 5(
( L = F 0 3
? 4 2 N y O s
'l I g .
T G C U 5 N U S s c e R s . I - o C ) c
, e
' ; S 6 (
2 4 c c 0 5 E t 3 1 M m 6
/
c . I T T m e ) S k s t E / u A T 1 3 h q ( Y 5 P A L i E 9 R 2 s o - e s
=
e, o
= 7- -
a y s - s s 0 7 o _ 5 g
- s
_ r - G a o _ 9 w c o G etLQv e e-
~ -
i' I
;J r
_ . (I _ - 0 1 0 0 0 ~2 _ 2 -
- g. 5v g ,S .
s ic f'
1l lll11
-(~
o o
.- i .
s
, t e
s g , e
, t
- s 0
g 0 n
=
. 2 no ii
= . t
,. ea
, tl
* . 0 s au 3 i 7 Rc l
1 wa
- . l oC
' g i FA S s .
. 0 0 kS Y
l' - o 5 aP o . 1
) eA rL
=.
. . c BE
. ) e R S f
_- c d
< . e(
s oot y g e
. c n t E n
- i e / b . i ~
v
=- c s =
M I o6 s - e
/ /
/
b l A) . T i3
- b
. P
.=5 (
i s o rL a pT i 6 ( - 0 e g 1 5 t mF v v 3 o si 2 1 oO CL v 2 -n A Y y
= 5 i
* , P 0 5
= A 5 -
g u e L' E
. 7 3 W s n o s
R
. 5 o s .
e
- s. a 3 e r e
g a c c e s u g 6 c . i o A F L
. F a N"F T O
v L e f
~ e_ M
.f{ .
.g V
U V i
\ )
$: 1 _ .: s x a n 0 n S e es
..amsK" m
eRx ._ ge 5Me qb - m y' lil,lllI. , l:!
,<'~% \
s TEST 4, D = 509 nrn, L/D = 3.1 - 3.0
\ a Data m HEM -
-- - Moody Slip
_d O \( - Burnell
~
d 10 - d a '\ ' h, - 2.0
) .
e i
's _ s a y ~~ --- ~~ 1 m . ~~~ m s , '% k 9
_3 _ _$
~
$ % 8 t
g
.g% _ _ ,_ _ * % H k, a
t [ g
\ E 0.5 -
o
- o. w 4 %
)-
2 21 i 1.0 66A66?6 A Aaggs
'h~ l ~
E 4 V Svecootch -> u 4--- Snregerth r b 0 I I I I o 0 10 20 30 40 53 Time (S) Figure 5.1 Mass Flow History for Test 4 l
\
_x_ _ _ _ _ _ _ _ _ _ .- _ _ _ _ _ . - - _ _ _ _ . - . _ _ _ _ _
/3 k) 1.5 TEST 24, D = 500 m, L/D = 0.33
- A Data 3.0 .
- -- Moody Slip ,
\
- - - Burnell -
t 2 2.5 1.0 kdy AL ~~ 2.0
~
g .- b 6A* '\ h x x At G m - E q- Ei 6 g Q 5 g m7-'-s N -
.1.5 5
x p? N N
?
!e E j 0.5 --
'g.{ m 9 __,, ,_ i 10 ff o. $,
n +
~
- < A sg 4A 41% (u l
,l ~ 0.5 7 s l
\
T 3B *W -> h + -s.4r m reb : 0 I I I l j l o f 0 10 26 30 A0 50 50 Time (S) Figure 5.2 Mass flow History for Test 24 i
Y' i A 0.5 TEST 18. D = 300 m. L/D - 3.7
\,,q\ 6 Data N _
pp 'N HEN
.- Moody Slip
\
) \ _ . .Burnell 0.4 -
N \ e E\ y
~
A >
\ -
o3
\
t-
' s _ ~ ._ - ,_
Il *
\ *
, 4 a, 1
'S n
(~~. .
%}
- b ~"
05
~
a~-<-~~-,4 g O g.D N -
^
Q., $ C 5 k _ h* E i 0.1 - k .,e SATgliA Wb 4 '
~( 4 y SyBCooldD ->
Q f 1. o '
] 1 I I O O 20 30 40 50 60 79 80 90 Time (S) rigure 5,3 Mass flow History for W N O
\
1 1 l-0.5 - TEST 6, D = 300 m. L/D = 1.0 A s a Data i 1.0 . HEM s - - Hoody Slip 6 k I -Burnell O.4 - f 3,\\ db - A 6'% .3 o b - x 0.3 - \ u y \ g\ 3 y g \ A -_ _ _ - g a b \ *M . b g o] f" x a
,__a 13 nn 0.5 g 0.2 3O k 4 1
O= e a f g
.s
!l.
~
0.1 . 4 l i
< St/BCool Eb _ ,,
f &SATJAA@ 4 D a . v i D , j l 1 I I I I T 0 o O 10 20 30 40 50 60 70 80 i Time (S) ! Figure 5.4 Mass Flow History for Test 6 O 4
l. 0.IX.2. 3-and'4 L. r3 Page 2-7 of Reference 3 states: "The staff' finds that the predicted. flow. , through the Power Operated Relief Valve =(PORV) has a large uncertainty when the flow is two-phase in composition. Section 2.1.2 of NUREG-0578 requires
.that PORV and safety valves be qualified to perform under conditions of both solid water and two-phase flow."
-0.IK.2 Clarify the uncertainty in the RELAP5YA models;for single-phase and two-phase flow through PORV and safety valves.
0.IK.3 ' clarify-the effect of liquid entrainment and the need'for detailed modeling below the PORY or safety valves. 0.IK.4 Clarify how the flow through valves has been assessed. A.IX.2, 3 and 4
'.The answers to Q.IX.2,.Q.IX.3 and Q.IK.4 are provided below for each plant
. ' L(I~h's,) model. Maine Yankee (MY) MY has two PORVs and three Safety Valves (SVs)~ located en the top of the pressurizer. The two PORVs are set to'opan at 2,385 psia. The three SVs are set to open et 2,485, 2.510 and 2,535 psia, respectively. Fone of these velvss actuate during SDLOCAs dua to their high pressure setpoints and the generally decreasing primary system pressure from 2,250 psia. Therefore, they are not modeled for GBLOCAs. A stuck-open PORY ira snalyzed by adding c. TRIP VALVE (latched open) te model this cocponent. The valve passes steam because the pressurizer spray is terminated by the decreasing primary system pressure. The valve throat area and a discharge coefficient of one are used to maximize the steam flow rate through this path. A more
- realistic value would lie in the range of 0.6 to 0.9 based upon our assessment of the critical flow model for steam.
1
MY has six Secondary Safety Valves (SSVs) located on each of the
~
three main steam lines that exit from the steam generators. The g }. group on each steam line is set to open as follows: 1 at 1,000 psia, 1 at 1,020 psia, 2 at 1,035 psia and 2 at 1,050 psia. These valves pass steam when actuated during SBLOCAs. Each group of SSVs is modeled using a TMDPJUN component where the steam mass flow rate is specified as a function of the upstream pressure. This incorporates a discharge coefficient to match the Technical Specification ratings for these valves. Yankee itowe (YR) YR has one PORV (2,400 psig setpoint) and two SVs (2,485 and 2,560 psig setpoints) located on top of the pressurizer. None of these valves actuate during SBLOCAs due to their high pressure setpoints and the generally decreasing primary system pressure from 2,000 psig. Therefore, they are not modeled for SBLOCAs. A stuck open PORV is analyzed in a similar manner to that described []' for MY above. YR has three SSVs located on each of the four main steam lines from the sLuam generators. Each group is modeled using a TMDPJUN component in a similar manner as that described for MY above. ! Vermont Yankee (VY) VY has one SV cach on two of the four main : steam lines. The SYs are [ nominally set at 1,225 i 15 psig. These valves do not operate during sny LOCAs (small to large) due to their high pressure setpoints and the generally decreasing primary system pressure from f 1,020 pria. VY has one Safety / Relief Valve (S/RV) on each of the four main ster.m l i lines. The S/RV setpoints are as fo11cws: I at 1,080 psis, 2 at
,q 1,090 psig and 1 at 1,100 psig. For small break LOCAs, usually one
- valve is sufficient to provide pressure relief prior to actuation of I
I
x the Automatic Depressurization System (ADS). During this period, C' wet to dry steam (static quality > 0.8) passes'through the valve. After ADS is initiated, the static quality'in the upstream steam line may drop as low as 0.5 and then increase toward unity. The S/RVs are modeled as motor valves with a short opening and-closing time based on test data. The valve throat areas and a. discharge coefficient of unity are used for the junctions-in order to maximize the loss of fluid. Comparison of predicted-to rated flow rates shows that a more realistic value for the discharge coefficient is 0.84. 1 0.IX.6 page 3-17 of Reference 17 states: "The flow from assumed small breaks is strongly affected by stratification. The mass discharge rate is dependent upon whether or not the break is covered by the liquid-vapor interface. Vapor pull-through and liquid entrainment occur and the resulting flow is highly complex. Semi-empirical
.models are required for these effects." Clarify how the break flow
[]. effects of stratification have been assessed and what the resulting-uncertainties are for break flows in cold legs where stratified flow exists. A.IX.0 As pointed cut in the question, stratification ef fects can lead to highly complex two-phase flows at the break. Reliable models for calculating t.he phenomena of vapor pull-through and liquid enten5nment are currently under development. RELAp5YA contalcr. an approximate method to account for the effect of stratification on break flow. This it discussed in the response to Question VI.7. An assessment of this method is provided in the response to Question V.1 which discussna the RELAp5YA calculation of the LOFT L3-1 experiment. Further assessment and 17proud modeling of the phenomena would indeed be desirable for realistic analysis of SBLOCAs. However, for licensing annlyses, a spectrum of br(sk sizes will be investigated. It is believed that the break spectrum j i analysis would cover the uncertainties in modeling stes.tification effects on break flow. i _. . _ = . - ___ ____ __-_ _ - _ ___ - _ .
n, - 1 0.IK.13 Page 261'of Reference'10 states: "AnalysesLwith discharge coefficients ranging.from 0.6 to 1.0 will be performed for'large. 0()'$I-
- A
' - ' breaks in boiling. water' reactors. . However.:small break. analyses.for PWRs, as well as for BWRs, will'be performed with a discharge coefficient of 1.0." Justify why only one'value of discharge
' coefficient needs to be. considered for SBLOCAs.
A.II.13' See the answer to Question IX.1. i i
X. ADDITIONAL OUESTIONS THAT ARE CONCERNED WITH SEVERAL AREAS l m (' ) ' Page 19 of Reference 21 states that: "Under some conditions, RELAPS l O.X.1 does a poor job of conserving mass and/or energy. These errors can I dominate the trans' nt history. The problem seems to lie in the I basic formulation of the finite-difference equations and donor cell j 1 determination." Large mass errors were also observed in a calculation reported in Reference 22. Clarify how mass errors are either avoided or affect the results presented in Reference 12 and what will be done to eliminate or avoid them in SBLOCA calculations. j l A.X.1 If the mass error becomes large, the numerical solution becomes questionable. The results shown in Reference X.1-1 and the additional calculations furnished in response to Questions V.1, V.2, V 3 and VI.9, have low mass error (less than about 10%). Techniques that have been successful in minimizing mass error are (a) reducing time step size and (b) appropriate renodalization without sacrificing phys,1 cal modeling. These techniques will be
'^'
used to keep mass errors to a minimum in plant calculations. It should be noted that mass errors (or energy errors) are a basic feature of most finite-difference solution schemes used for two-phase flow snalysis and, hence, they cannot be. eliminated completely without introducing other numerical compromises. Reference (I.1-1) Fernandez, R. T. , et si. "RELAPSYA - A Computer Program for LWR System Therral-Hydraulic Analysis, Volume III," YAEC-1300P, January 1983 0.X.5 A better appreciation for the verificatico of SBLOCA modtis provided by the previous comparisons is needcd. Clarify the areas verified by the comparisons; for example, by an itemized list of area and aspects of the area verified for each comparison. p v . 1 i 1 I l
[:. N .
.); ;;
h . ,. , A.I.5 Table'I.5-1 lists the area'and. aspects of area verified by the tests presented.in Reference 1.5-1. .In' addition,~two new' tests were added E.[ l
~ as part of the RELAPSYA assessment, the semiscale test S-NC-2 and
, the LOFT Test L3-1. r: - i y O
/ l H f
_ u Ct s . y a - an X X _ r t ea _ d sHr _ y o T H P _ l _ a m F _ r H X I' X X e C _ h T e r - nn o soo _ C lii ett vcu _ eab L ri Fr e t K I X X X X rd s oli C oD V k aw eo X rl BF s 1 rc oli t al 5 mamu arra
. I eeer s t nhd seTy G H K :
T n o i lt aa rl uu t c ar Ni C n ol ian ti o cci eat jf a n r s_ I eq t t Asdn CI n C o Z&C r ne of is t n aa l sr l nT _ ae - _ Wdt na oe CH c i r X t c - td et e y sn - t s p t3 l d mea iae o s - s Es al T stTI o l e4 n at t ei K aS t L eT0 e0 is l s tFg u - f0 s v 0 k1 be ae S ng Qyf1 e C el 1 i mT rT ei . d o T G Ll vt u e Ftl7 Fal 9 I O 2 w9 [ e . rs B C t l F GD I nF eH Tai 0 HRG1 rTei0 HRG1
h t c K' en K K.
~
we eu RO - s c r i F e l H f m., /. . (. d u a r Cts t ea sHr an K X X y o T H P l a - m F r H K X X e C h T e C r o soo lii ett vcu nn eab Lri Fr K e t K X rds oil cod V k aw eo X K X rl BF
)
d
't rc s
n o oli c t al ( mamu arra X K 1 eeer
- t nhd .,V 5K seTy G H e
E n L o B i A lt T aa rl X uu K K
.t c ar Ni c
n ol ian tio cci eat jf a nrs I en K K X t e X X And CI n C o E&C r ne of is tn aa l sr l nT K K K ae Wdt na oe CH 1 V d0 0Q . f 2 3 L A1
- 1. 01 f 6 o
o ot/ t1 eC t C/ O2 l9 l n1 s L3 s - l N t s fR3 e 0.0 9 l e4 om4 Bi6 e8 TL a-cS e T B4 S6 e 3 r / s - T Ft Aet T6 it T At Tsd Tps F - ms ee F Ts Re t ibG R lx G i 0, 4 3 s W ] T _
0.X.6 The staff's experience with advanced thermal-hydraulic computer
') programs has shown an important sensitivity to modeling of the steam
'~
generators when analyzing small break loss-of-coolant accidents. In specific, the modeling of liquid entrainment, condensation and hydraulic resistances (e.g., flow regime maps) could significantly depress the mixture level in the core. This phenomenon has been observed in Semiscale Experiment S-UT-8. Recognizing Semiscale's atypicality, the staff nevertheless believes this phenomenon to be real and, therefore, plausible in a full-scale reactor. It is for this reason that we request integral experimental validation of your computer program to predict this phenomenon, should it occur in a full-scale reactor. Validation with the Semiscale experiment would be acceptable. Use of other integral experiments for validations requires that these experiments simulate this hydraulic behavior. A.X.6 Introduction The staff requested that YAEC validate the RELAP5YA computer program , ') against the S-UT-08 exptriment. The, request was made because this
~
test uncovered certain phenomena that were not previously observed to this degree in other UT non-UHI tests. Specifically, it appeared that total core uncovery occurred prior to clearing of the loop seals. This uncovery was attributed to complex thermal-hydraulic phenomena in the steam generators which depressed the core coolant by developing significant resistance to steam venting. RELAPSYA has been used for a SBLOCA analysis of the Maine Yankee PWR. This calculation showed that RELAPSYA can predict the phenomena observed in S-UT-08. Thus, we believe that a simulation of Test S-UT-08 is not necessary to further demonstrate this code capability. The S-UT-08 test results and the RELAP5YA results for the Maine Yankee SBLOCA case are described below. Description of Phenomena Observed in Test S-UT-08 The phenomena observed in S-UT-08 are described in detail in [' ) \ / Reference'I.6.1. The important aspects of the test behavior are
'~' briefly described below.
pi Semiscale Test S-UT-08' simulated a SBLOCA in a Westinghouse PWR u
, w' ; resulting from a 5% communicative break in one of the cold legs.
N The break was located between the broken-loop primary coolant pump and the inlet to the downcomer. The experiment included early pump trip, high and low pressure ECC injection and accumulator injection
~
into the cold legs of the broken and intact loops. After the break was initiated, the primary pressure rapidly decreased and approached the secondary side pressure due to. SCRAM. and pump trip. The loop flows coasted down and natural circulation was quickly attained. As the pressure continued to decrease slowly, the' loop void fractions increased and an asymmetry developed in the liquid level between the upside and downside of both the U-tube steam generators. Figure X.6-1 shows the collapsed levels in the upside and downside of the broken-loop steam generator. At about 85' seconds, the upside level became larger than the downside level due to CCFL at the steam generator entrance. This caused steam flow to be blocked and resulted in a depression of the core level seen in ()' Figure I.6-2. The core level continued to decrease until the loop seals cleared at about 210 seconds. This caused the liquid held up in the steam generators to be swept up and over the top of the U-tubes. The ECC water was then able to recover the core level. Subsequently the core underwent a slow boiloff followed by recovery due to accumulator injection starting at about 500 seconds. There were two core heatup periods. The first heatup, at around 200 seconds, was caused by the core level depression associated with the liquid holdup in the steam generators. The second heatup was caused by the. core bolloff. The core thermal behavior can be seen in Figures X.6-3, X.6-4 and X.6-5 which show cladding temperature responses of the hot rod at three axial elevations. Observation of S-UT-08 Type Behavior in Maine Yankee Calculation RELAp5YA was used to assess the Maine Yankee plant behavior for various postulated break sizes as part of the Maine Yankee Pump Trip Study (Reference 1). The worst break size, 0.05 ft , was used to (s perform a pump trip time sensitivity. Two times were chosen to trip t
g , , , . .- f lE the pungs, two and ten minutes _ following the HPSI . injection. -Both-Y) these- transients exhibit similar behavior as the Test S-UT-08. ' We Q' :
~
will concentrate our discussion on the transient with the RC~ pumps tripped two. minutes following HPSI injection. The major assumptions utilized in this the analysis are presented below:
- 1. Off-site Powtr Available - The assumptions of continued operation of RCPs during a'small break LOCA event requires the availability of off-site power.
- 2. 102% Steady-State power Operation.
- 3. . Only one HPSI and one LPSI Pump Available - Of the two high-pressure pumps which are energized. automatically on safety injection actuation signal, it was assumed that only one was available. Due to the assumption o'l off-site power availability, the delay in starting the pump is 0.9 *econd.
- 4. One auxiliary feedwater pump available.
- 5. ANS decay heat + 20% uncertainty.
- 6. No Steam Dump and Bypass is Available - The steam relief is through the safety relief valves only.
- 7. Metal Water Reaction - The Baker-Just model was used for metal water reaction.
- 8. Appendix K (Reference 7) Lockout Options - These flags force a degraded heat transfer calculation during the blowdown phase even in circumstances where calculated local conditions indicate that rewetting occurs.
- 9. Break Location - The break was located in the cold leg pump discharge pipe.
1
[ -{ .a i l The Maine Yankee nodalization is given in Figure X.6-6. The trends-
~( qo predicted by RELAp5YA for some of the Maine Yankee system parameters f r
L/ are compared with the corresponding trends observed in S-UT-08. J Similar to S-UT-08, after break initiation, the primary pressure rapidly decreased to near the secondary pressure and then continued
- to decrease- at a slower rate. The loop flows coasted down, the loop void fractions. increased and two-phase natural circulation was quickly established. The steam generator liquid holdup phenomena-are remarkably similar to those observed in S-UT-08. Figure X.6-7 shows the collapsed levels in the upside and downside of the broken loop steam generator. It can be seen that at about 400 seconds, the upside level begins to be higher than the downside level, indicating liquid holdup and a consequent blockage of the steam venting path.
Figure X.6-8 shows the same behavior for the intact loop steam generator. Consequently, there was a rapid core level depression beginning at about 400 seconds shown in Figure X.6-9. The level continued to drop below the bottom of the core. At about 700 r
.( seconds, the loop seals cleared and the, core level was recovered.
A Figure X.6-10 shows the void fractions in the loop seals. The clearing of the loop seals is indicated by the void fractions increasing rapidly from near zero to near unity at about 700 seconds. Beyond this time, the core underwent a steady bolloff and the core level decreased steadily. At about 2,000 seconds, accumulator injection began and the core level began to recover (Figure X.6-9). However, as seen in Figure X.6-10, the loop soal in the broken loop refilled at this time thus eliminating one path for steam venting. Hence core recovery was slower than that observed in S-UT-08. Similar to S-UT-08, there were two core heatup periods. Figure X.6-11 shows the cladding temperature response of the hot rod. The ; first heatup is associated with the core level depression caused by liquid holdup in the steam generators and is seen to occur between 400 and 700 seconds. The second heatup is associated with the slow core boiloff and is seen beyond about 1,000 seconds. The refilling of the loop seal in the broken loop and the ensuing blockage of i
! s 's ; 'j
.:3 ;
. steam venting resulted in an intermittent quenching of the core seen
'O in Figure X.6-11 beyond about 2,100 seconds.
N) Conclusion Test S-UT-08 exhibited certain phenomena which were not encountered
.to this degree in other tests. Phenomena very similar_ to that'in-Test S-UT-08 were also observed in a RELAP5YA calculation for a Maine Yankee.SBLOCA. This calculation demonstrates that RELAPSYA has the capability-of predicting the phenomena observed Tn S-UT-08.-
Therefore, we propose that further validation of RELAP5YA against S-UT-08 test is not'necessary. Reference (I.6.1) Robert Fujita, " TRAC-PFl/ MOD 1 POST-TEST Analysis of Semiscale Small Break Test S-UT-08," Proceedings of Third International Topical Meeting on Reactor Thermal-Hydraulics, Newport, Rhode Island, October 1985. O ii
_ - _ - _ - _ = - - - - -_ . _ _ . o 1200 . 4 3 1000 - 0' ~ 800 - d Upside O -
- a. co0 -
O
@. Downside ~
400 - w ~ 0 200 - sa 8 o _i
-2000 200 400 600 800 Figure X.6.1 Intact Loop Steam-Generator collapsed Liquid Level (S-UT-08) %/
400 '- 4 g - g - O 8 300 .- _ d g .. a 0 200- - B 3 . . i00 - s y - O -
-60 ' I - f ' I O 200 400 600 800 TIME (s)
O ,1.... x.e.2 c. . co11.... L1,.1e L.. 1 < ...-ee, t 1
i
.f~
s/., _ 1000 g , ,. L 800 - -
- g. .
Is3 600 - - n L_ a: 400-~ ' I ' ' ' ' O 000 -400 600 800 TME(s)- Figure X.6-3: Heater Rod Clad Temperature at the ( 137-cm Elevation (S-UT-08) 1000 i . 2 800 - ~ R g . - o, 600 -
. 6 l
l 400 ' ' I O 200 400 600 800-TIME (s) Figure X.6.4: Heater Rod Clad Temperature at the 208-cm Elevation (S-UT-08) t
.. s., .- . {s
.i j
l lL ' LO 8 1000- , ,
- 900 -
M
. W 800 -
M
< 700 -
W - - . H -600 -
-O o -
M 600 - e - ' 400 > ' 600 800 O 200 400 Figure X.6-5: Heater-rod Temperature at the 253-cm Elevation (S-UT-08)- i
~
7 N s MP heL o t e s I
,o 8 "'#
t ;.sr C u C 9 3;. A
- .mv 8
<9 .
e
~
. e I
m 1 8
' 'x u
4 ;: .t 3, ,
~*
'9 n
. 4 ,
1 M3- o I c, ,5 i M : 1
. ,o - e e
F 1 9 v e l 4 f
- u..
o c a k Y n a
- ,, n-8
# e 0
l
, n 3 i 0 - a 2 l o M 1 o, r
- t :
! : f f
- n :a- . - 5 0 o n
d 1
- o i s I - l ti r +l l l I l I
. as
'g*ot* zy
,b L 6 i l E E g s* * . 'n 5 l a 3 5 4 r an s _
E V s l l l I l i
. ~ ..
M2 l
> r*3 dA o
NA C x 't d *. 8
. 8
. ~ 6T AO e
l 1 YL t 5 B s - PS n } 6 A u s 1 o -
'a L r E o 0 :- d s
s 52 2 l-n :i - y s Rf a o 8 o P e - s : 7 e 6 o Q1~. - dig 1 f } s 6 I 6 e e o -
@t a
J "}}