ML20128H934
| ML20128H934 | |
| Person / Time | |
|---|---|
| Site: | Vermont Yankee, Yankee Rowe, Maine Yankee, 05000000 |
| Issue date: | 07/01/1985 |
| From: | Papanic G YANKEE ATOMIC ELECTRIC CO. |
| To: | Zwolinski J Office of Nuclear Reactor Regulation |
| References | |
| FYR-85-12, FYR-85-72, NUDOCS 8507100314 | |
| Download: ML20128H934 (149) | |
Text
{{#Wiki_filter:- YANKEE ATO~10 ELECTRIC COMPANY Tc'=pho"" (6 ") & ' -* '00 TWX 710-380-7619 b 1671 Worcester Road, Framingham, Massachusetts 01701 2.C.2.1 ) g July 1, 1985 FYR 85-72 United States Nuclear Regulatory Commission Washington, D. C. 20555 Attention: Mr. John A. Zwolinski, Chief Operating Reactors Branch No. 5 Division of Licensing
References:
(a) License No. DPR-3 (Docket No. 50-29) (b) License No. DPR-28 (Docket No. 50-271) (c) License No. DPR-36 (Docket No. 50-309) (d) USNRC Letter to YAEC and MYAPCO, dated May 11, 1984 (e) Topical Report: "RELAP5YA - A Computer Program for Light Water Reactor System Thermal-Hydraulic Analysis," Yankee Atomic Electric Company, YAEC-1300P, October 1982 (f) MYAPCO Letter to USNRC, dated January 31, 1985 (g) YAEC Letter to USNRC, dated February 26, 1985 (h) YAEC Letter to USNRC, dated March 1, 1985 (1) YAEC Letter to USNRC, dated April 30, 1925
Subject:
Response to the NRC Questions on RELAPSYA
Dear Sir:
Reference (d) requested answers to 197 questions on YAEC's Topical Report on RELAP5YA [ Reference (e)]. Our schedule called for the submittal of responses to 79 RELAPSYA questions by July 1, 1985. Responses to 11 of these questions have already been submitted to you in Reference (1). Responses to the remaining 68 questions are provided as an attachment to this letter. The attachment contains 12 extra responses belonging to the next group of questions scheduled for August 15, 1985. These are provided to you in advance to facilitate the review process. Question Q.IK.13, instead of Q.IX 8, was erroneously identified as deleted in Reference (h). To correct this mistake, Question Q.II.8 and its response have been identified as deleted in this submittal. The response to Q.IX.13 will be provided later. We trust this information is satisfactory. However, if you have any questions please contact us. Very truly yours, 9 YANKEEATOMICELEC Py- !;i 0 ( C. Papa e 0 Senior roject Engineer - Licensing \\ g \\\\ CP/mmt u
RESPONSE TO 80 QUESTIONS ON RELAP5YA July 1, 1985 i f i f l i I
I. CONDgN8ATION HEAT TRANSFEB AND NONCONDgNSIBLs GA333 Wine potential sources of noncondensible gases are listed on Page VIII-49 of. Reference 4 for a PWR: "(1) dissolved hydrogen in the primary coolant, (2) dissolved nitrogen in the accumulator water..(3) dissolved air in the refueling water storage tank, (4) hydrogen released from the sicconium-water reaction, (5) free 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) pressuriser steam space gas." L.l.ul Clarify which of these are treated by RgLAP5YA and what rates are used to predict the amounts entering the system from each source. W Noncondensible gases are expected to have a negligible effect on the outcome of small break LOCAs. Hence, the treatment of noncondensibles in our calculations is not warranted. Appendix I.3-1 provides the justification for excluding the treatment of noncondensibles in our analysea. APPENDIX A.I-3 BOUNDING ESTIKATES OF NONCONDENSIBLE CAS SOURCES AND THEIR IMPACT ON SBLOCAS
1.0 INTRODUCTION
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 migrate 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 period of time in only very small breaks. For example, licensing analyses for breaks larger than 0.02 f t in the Maine Yankee plant show that most of the decay heat is removed through the break (Reference I.3-1). For breaks smaller 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. Plant-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 plant at Rowe. M&INE YANKEE PLANT 1. Dissolved hydronen in the primary coolantI h calculation assumed the maximum RCs concentration of 50 cc of N #IE *t "*I'# 2 (Reference 1.3-2) at STF. It is conservatively assumed that all the hydrogen is released and remains in the primary system. 2. Dissolved nitronen in the accumulator water h concentration of dissolved nitrogen is detemined from Henry's Law (Reference 1.3-3) at 104'F and 230 pois to be 2.04 x 10 1hm N /lba N 0. 2 2 This concentration is used to cateulate the nitrogen introduced during accumulator water injection. 4 3. Dissolved air in refuelina water storage tankt The pumped safety injection water from the RWsf was calculated to contain 19.9 cc of air /kg of water at STP based upon Henry's Law (Reference 1.3-3). The entire RWst inventory is assumed to be injected. 4. Hydrogen released from streonium-water reaction h calculation is based on 1% streonium reacted and 37840 fuel rods. 5. Free nitronen used to pressurise accumulators h nitrogen contained in the secumulator gas space is assumed to be released when the accumulator liquid empties. 4. Hydrogen released from radiolytic decomposition of injected watert h time when the H concentration in the primary system falls 2 below 5 cc of M,/kg of water is determined to be the time to start the cateutation of radiolytic decomposition of water. For the 0.02 ft Maine Yankee SSLOCA, a concentration of 5 cc H #kE 2 H O is reached at approximately 4700 seconds. At this time, the 2 steam generators are not relied on to remove decay heat. 7. Fission and fill gas in reactor fuelt h calculation is performed assuming all fuel rods are at end-of-life conditions. 8. Hydrogen in the make-up tank: For the Maine Yankee plant, the i make-up tank does not release water into the primary system during a 35LOCA. 9. Pressurizer vapor space _: 630 cubic feet of hydrogen at STP is assumed to be contained in the pressuriser vapor space. This represents 90% of the available gas volume. For Maine Yankee, Table 1.3-1 provides the volume at (STP) and mass of the noncondensibles from all the nine sources identified above. l It is to be recognized that the core would have to undergo tracolents I that are not predicted to occur for applicable small breaks, in order for many of the sources identified above to be important. Present Appendix X small 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 1.3-1) in area and lower, the system pressure does not reduce to the safety injection tank pressure. The only sources that may be introduced a ares (a) hydrogen dissolved in the primary coolant, (b) air dissolved in the refueling water and (c) hydrogen contained in the pressuriser vapor space. The noncondensible gas volume from these sources amounts to 2011 ft at STP, For the Maine Yankee plant, the total primary tube volume is 688 ft per steam generator. The noncondensible gas volume at che steam generator safety relief valve setpoint (1000 psia) is about 60 ft. The reactor vessel upper head volume is 1476 ft. Most of the noncondensibles are expected to migrate to the upper head without causing any impact on 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 reside there, the gas would occupy about 3% of the U-tube volume at stable conditions. YANKEE _ Pt.ANLALROWE 1. Dissolved hydrogen in the primary coolanti The calculation assumed a maximum concentration of 40 cc of H /kg of water 2 4
(Reference I.3-4) at STF. It is conservatively assumed that all the hydrogen is released and remains in the primary system. 2. Dissolved nitrogen in the accumulator water: N dissolved concentration of nitrogen is determined from Henry's Law (Reference 1.3-3) at 14.7 and 70'F to be 1.9027 x 10 1ha N /lbe ~ 2 H 0. This concentration is used to calculate the introduction of 2 nitrogen during accumulator water injection. I t 3. Dissolved air in refueling water storage tank h concentration of dissolved air is determined from Henry's Law (Reference I.3-3) to be 1.6364 x 10 lba ale /lbe H O at 14.7 pois and 130*F. 2 h entire RWST inventory is assumed to be injected. 4 4. Hydrogen released from streonium water-reactions. The calculation is based on 1% sicconium reacted and 17336 fuel rods. 5. Free nitrogen used to pressurise accumulatorst h accumulator at the Yankee plant is isolated just before the 11guld is completely depleted. This is done to avoid accumulator nitrogen injection into the reactor vessel. l 6. Hydrogen released from radiolytic decomposition of injected watert h radiolytic decomposition of water is initiated when the H 2 concentration in the primary system falls below 5 cc of H #k8 'I l 2 water. For a 1" break in Yankee Rowe, this concentration is reached at about 3000 seconds. At this time, the steam generators l are not needed to remove decay heat. l 7. Fission and fill gas in reactor fuelt h calculation is perfomed ) assuming all fuel rods are at end-of-11te conditions. I 4. Hydrogen in the make-up tankt_ For the Yankee plant, the make-up tank does not release water into the primary system during a stLOCA. l 9. _Pressuriser vapor space: 185 cubic feet of hydrogen at SYP are assumed to be contained in the pressuriser vapor space. This represents 90% of the available gas space. f l 5
For the Yankee plant at Rowe, Table 1.3-2 provides the volume (at STP) and mass of the noncondensibles from all the nine sources identified above. For breaks of 1" diameter and lower (for which the steam generator is desired to remove decay heat), the system pressure does not reduce to 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 generator safety relief valve setpoint (900 psia) is about 13 ft. The reactor vessel upper head volume is about 378 ft. Most of the noncondensibles are expected to migrate to the upper head region without 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 [FFFCT OF NONCO@MS11MLON_C@MNDATIONJEAT._HA!!H111 The impact of the noncondensibles on hout transfer and thermal hydraulics of the two p12nts is assessed in this section. It is concluded that the noncondensible gases released at the Maine Yankee and the Yankee plante during small break t,0CAs will have a negligibic effect on the hydraulic and thermal aspects of the two plants. The two effects of noncondensible gases are: 1. Reduction in the available temperature difference for given primary and secondary system pressures. 6-
2. Reduction in condensation heat transfer due to the resistance of the noncondensible saa boundary layer. The reduction in the available temperature difference is negligible as 111ustrated in Tables I.3-3 and 1.3-4 for the Maine Yankee plant, and Tables l 1.3-5 and I.3-6 for the Yankee plant at Rowe. N eo tables assume that 2011 ft (STP) at Maine Yankee and 440 ft (STP) at Yankee plant of noncondensibles are evenly distributed in either the primary system or the l steen generators. The presence of noncondensibles affacts the saturation l temperature due to the partial pressure of the noncondensible gas. h oe temperature differences of less than 3 F will not significantly affect the ses11 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, h increased l resistance is due to the fotsation of a noncondensible boundary layer. W l noncondensible gas is carried with the vapor towards the 11guld vapor interface where it accumulates. Therefore, the partial pressure of gas at the liquid vapor interface increases above that in the bulk of the mixture. This situation is shown in Figure I.3-1. W effect of noncondensibles on the l condensation heat transfer is calculated using the method outlined in Reference 1.3-6. h 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 equationt l h,(Tot - T ) = h,(Tgg 0!' -T y l +x,f,h,(r,-ro t )"an (1) g wherei 2 heat transfer coefficient in 11guld film, stu/hr wt or he 2 sensible heat transfer coefficient, Blu/hr ft 0p hs a mass transfer coefficient, ft/hr M a g 1
hgg = latent heat of vaporization, Btu /lb P, los mean partial pressure of noncondensible. = ~ AI A0 'am " I 1n P /P AI A0 Other terms are defined in Figure I.3-1. For the case where the gas mixture is stagnant, the sensible heat transferred through the diffusion layer need not be considered. Kg, the mass transfer coefficient, is defined in Reference 1.3-6 as: =0.373 f(i - 1) gH (2) 1.02D g" H MD where 2 volumetric diffusivity, ft /hr D = tube height, ft H = acceleration of gravity, ft/hr2 3 = mixture density, ib/ft3 = ha vapor density in bulk mixture, 1b/ft 3 = 3 vapor density at interface, 1b/ft i = h = mixture viscosity, 1b/ft-hr The flim heat transfer coefficient, h,, was evaluated using the Wusselt film condensation model.
- 0 fg(1-f) h,kf.25 y
g h, = 0.943 pg H(T - 7,) ag ~ where: 3 liquid density, Ib/ft 1 = fy vapor density,1b/f t 3 = .g.
liquid thermal conductivity, Btu /hr ftoy kt = hd liquid viscosity, Ib/ft-he i = All liquid properties are evaluated at: T,g = T + 0.31 (T w' ~ CI The solution can be arrived at by iterating on Tgg,until Equation (1) is satisfied. Two sets of calculations have been performed 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 Meine Yankee: a) A secondary side pressure of 1000 psia. b) 2011 ft (at STP) of noncondensible (hydrogen) are released. c)' The noncondensible was evenly dispersed throughout either the entire primary system or in the four stear. generators. I For the Yankee plant at Rowe: a) A secondary side pressure of 900 psia, 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 I.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 mixed throughout the entire primary system. The bottom curve in each graph corresponds to the case where the noncondensible was uniformly _9
distributed in the steam generators only. The graph's ordinate represents the j 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 pressure, the reduction in heat transfer is less than 10% for 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. i If the primary side begins to approach the secondary side pressure and the effect of noncondensible becomes more pronounced, the not effect will only be to stabilize at a slightly higher primary pressure (about 25 psia) than that for the case without noncondensible. The significant margin in overall heat transfer is illustrated in Figures I.3-4 and 1.3-5, Which gives the heat removal rate from the secondary side for 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 evenly throughout the primary system. The bottom curve corresponds to a case Where the noncondensible 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 smet detrimental location. j In summary, the following conclusions can be drawn. Noncondensible Gases in Steam Generators - Staanant Conditions i l l 1. The effect of the presence of noncondensible gas source on steam i generator heat transfer rate is small (10% for NY and 13% for YR), { for primary pressures that are 200 pois or higher than the steam I generator secondary side pressure. This small heat transfer degradation has no significant impact on the primary side pressure. i I f l f _ ~.., _.. -,.,.. _,,... ,____.--,._._,m., .m - 7 m m,. ,,w m.m_g
i t 2-t l' s [ 3 i, <2. For primary pressures that approach the secondary side pressure, t J the steam' generator heat tra'nsfer degradation due to the presence of noncondensi,ble gas is higher. However, this will only lead to increasing Ene primary system pressure by about 25 psia for the same' decay heat rate. i 4 sl
- 3. -The core heat transfer ir,not.affected for these cases since all w
i if > the noncondensible gas is assumed to reside in the steam generators. a '\\ \\ Noncondensible Gases Evenly Distributed - Stagnant Conditions p h '1. For these cases, the heat transfer rates in the steam generators and the core are not significantly affected. t References in (I.3-1) Liliane Schor, et al., " Justification of Reactor Coolant Pump Operation During Small Break LOCA Transients for Maine Yankee," YAEC-1423, April 1984. (I.3-2) Maine Yankee Plant, Reactor Coolant System, System Description No. 4457-010, Revision 0, Combustion Engineering, Inc. (I.3-3) Perry, Robert H and Chilton, Cecil H., " Chemical Engineers' Handbook," Fifth Edition, McGraw Hill Book Company. (I.3-4) Yankee Plant at Rowe Technical Specifications. (I.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. (I.3-6) Collier, John G., " Convective Boiling and Condensation," Second Edition, McGraw Hill Book Company. _ ~ ~
TABLE I.3-1 Sources of Noncondensibles for the Maine Yankee Plant Volume Source (STP) Mass 1. Dissolved hydrogen in the primary 515.0 ft3 2.69 lb coolant 4 2. Dissolved nitrogen in the ac:wsulator 768.0 ft3 56.0 lb water 3. Dissolved air in RWST 866.0 ft3 65.09 lb 4. Hydrogen released from zirconium 4127.0 ft3 21.57 lb water reaction 5. Free nitrogen used to pressure 96780.0 ft3 7057.0 lb accumulators 6. Hydrogen released from radiolytic N/A N/A decomposition and injected water 7. Fission and fill gas in reactor fuel 1236.9 ft3 He 10.08 lbm 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 9. Pressurizer steam space gas 630.0 ft3 3.292 lb 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.
l TABLE I.3-2 Sources of Woncondensibles for the Yankee Plant at Rowe Volume Source (STP) Hass 1. Dissolved hydrogen in the primary 117.3 ft3 0.613 lb coolant 2. Dissolved nitrogen in the accumulator 11.38 ft3 0.8298 lb water 3. Dissolved air in RWST 137.89 ft3 10.365 lb 4. Hydrogen released from zirconium 813.74 ft3 4.2526 lb water reaction 5. Free nitrogen used to pressure N/A N/A accumulators 6. Hydrogen released from radiolytic N/A N/A decomposition and injected water 7. Fission and fill gas in reactor fuel 103.37 ft3 He 1.05 lb Ar 0.05 lb N 5.85 x 10-4 lb Ke 0.044 lb Xe 0.49 lb H 9.0 x 10-5 lb 8. H2 gas in makeup tank N/A N/A 9. Pressurizer steam space gas 184.5 ft3 0.9642 lb 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...
i TABLE I.3-3 Maine Yankee Primary Temperatures With and Without Noncondensibles Assumes Woncondensibles Evenly Distributed in Primary System Primary Pressure T(OF) w/o T(OF) w/ d T(OF) (psis) 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
TABLT I.3-4 Maine Yankee Primary Temperatures With and Without Noncondensibles Assumes Noncondensibles only in Steam Generators Primary T(OF) Pressure T(OF) w/o T(OF) w/ (psia) Noncondensibles Noncondensibles 1050.0 550.53 547.26 3.27 1200.0 567.19 564.19 3.00 1400.0 587.07 584.36 2.71 1600.0 604.87 602.38 2.49 1800.0 621.02 618.72 2.30 2000.0 635.80 633.66 2.14 l I i I.
TABLE I.3-5 Yankee Rowe Primary Temperatures With and Without Noncondensibles Assumes Noncondensibles Evenly Distributed in Primary System Primary Pressure T(OF) w/o T(OF) w/ lh T(OF) (psia) Noncondensibles Noncondensibles 1000.0 544.58 544.05 0.53 1050.0 550.53 550.01 0.52 1200.0 567.19 566.72 0.47 1400.0 587.07 586.65 0.42 1600.0 604.87 604.49 0.38 1800.0 621.02 620.67 0.35 2000.0 635.80 635.48 0.32 I I l.___ - _, _ _ _ _.
TABLE I.3-6 Yankee Rowe Primary Temperatures With and Without Woncondensibles Assumes Woncondensibles Only in Steam Generators Primary d T(OF) Pressure T(OF) w/o T(OF) w/ (psia) Noncondensibles Noncondensibles 1000.0 544.58 541.74 2.84 1050.0 550.53 547.78 2.75 1200.0 567.19 564.66 2.53 1400.0 587.07 584.79 2.28 1600.0 604.87 602.78 2.09 1800.0 621.02 619.09 1.93 2000.0 635.80 634.01 1.79 l l l l l ! i
THE INFLUENCE OF NONCONDENSIBLES ON INTERFACIAL RESISTANCE DIFFUSION LAYER BULK MIXTURE QP s L I OPO s O s U s I D( PGI F I s L O TGO s s s s ' TGI TW s s -( PAI s O PAO TOTAL SYSTEM PRESSURE P = STEAM PARTIAL PRESSURE IN BULK, MIXTURE PGO = STEAM PARTIAL PRESSURE AT INTERFACE PGI = NONCONDENSIBLES PARTIAL PRESSURE IN BULK MIXTURE PAO = NONCONDENSIBLES PARTIAL PRESSURE AT INTERFACE PAI = WALL TEMPERATURE TW ' = BULK MIXTURE TEMPERATURE TGO = INTERFACE TEMPERATURE TGI = Figure I.3-1 The Influence of Noncondensibles on Interfacial Resistance.
l 0 0 022 S GC SR e NN e 0 k II n DD a NN 0 Y OO 6 1-- CC e 0 NN n OO 2 i NN a M // WW e 0 t a 0 R 1.- 6) ta 8R e 1 H I n S o P se 0( l b i 0 s x-6E n i 8 R de 1 n U oc S no S N 0E f o 0R t 6P c I-. 4 e f 1 f E 2 0 3 I 0 e 6 r y I 2 u 1 g i F 0 000 1 O. .g Od +d N=O O.O oHaM Hgwr 1y i i lllll
t 0 0 622 S GC SR 0 NN I I 0 DD 1 a NN a OO 0 CC 2 NN OO NN e t // 0 a WW R 0 t 3: 1 -v 6 ae 8 H C 1 ) nO R s 0I e l 0 S b i 1 nv 6P s e s( n wo ed R 1 n t o E c a n 0R t oNn .0U a f l +n 6S oP 1 4 S t e 1 c e E k ef n R f a E Y 0P 0 3 l-6 2 3 1 I eru 0 g i 0 F 6 l 0 1 0 0 0 8 ef ed +g mf O.O wHEf cw.E O , oi 1 l! (
0 0022 SCG e RS e k NN n .II 0 a D Y NDD ONN 0 e I COO b n NCC 0 i ONN a NOO 2 M NN 0 /// HHN se =o: l 0 b i 0 sn i b) e 8 R dn 1 o I cn S oN P 0( tu o 0 h i t bE i 6 R W 1 U dn S a S h t 0E i W OR x i bP u l 4 F 1 t C a E e l S I 0 0 0 4 5 0 3 i T b I A 2 e T 1 A r E ug H 0 i Y F A 0 C 0 ED 0 i' 1 ~ o.O oi o * @o+ o.ooO" U oO@ o.@o~ of n ^ MJNDE+ D v (D1_D. e % Xa3_[- HCLJ ~ 1 1_- 2 C [C E hT l l l l
o l 1 l l 1 i e o ^d M o-I" xo D H 8-mo m o O d_ i U8 o*O o-e W/0 NONCONO. M o +-+ W/ NONCOND. IN SG X o-W W/ NONCOND. IN RCS X W l> Do Y L_ d ~ __1 o v Ho C w a-IN o b ' DECAY IIEAT AT 500 SEC 800.0 10b0.0 12b0.0 14b0.0 16b0.0 18b0.0 20b0.0 2200.0 PRESSURE (PSIR) Figure I.3-5 Heat Flux With and Without Noncondensibles Yankee Plant at Rowe I
Q.T.4 Justify why any sources are omitted. A.I.4 Please see Response A.I.3. O.I.5 Clarify how mixtures of more than one noncondensible gas such as nitrogen and hydrogen are modeled, especially if fractions of each are changing in space and time. A.I.5 Please see Response A.I.3. 0.I.6 Clarify under what conditions natural circulation flow can be interrupted by the accumulation of noncondensible gas plus steam in the U-bends, and what conditions are necessary to re-establish natural circulation flow. i l A.I.6 l l Please see Response A.III.1. 0.I.7 Clarify how the pump curves are changed if noncondensibles are present. A.I.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 10CFR50.46. For example, see the answer to Question Q.I.3. b. Since large concentrations of noncondensible gases are not expected, they are excluded from our LOCA analyses. For example, 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 I.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 more head and torque degradation than the steam water data (see Figures 6-2, 6-4, 6-5 and 6-7 from Reference I.7-1 attached). However, the data scatter (particularly in the air-water data) preclude drawing quantitative conclusions with a high degree of confidence. I l. Therefore, we use pump homologous data derived from steam water tests i for all LOCA analyses. Reference I (I.7-1) Kamath, P. S., Swift, W. L., "Two-Phase Performance of Scale Models 1 of a Primary Coolant Pump," EPRI NP-2578 Electric Power Research Institute, Palo Alto, CA, September 1982. -.
13 Creare Air / Water REFERENCE I.7-1 O B&W - 1. 5 -- m Single e Phase l 2 C m 1.5 1 0 a g: m "z e m e1 Do e% 8 o c e mm a i 25-- s m* 8 a 3 5 e = m m O C o m O S e B oo mm C e m ~ o O O 8.8 f s.' 2 s.' 3 sE'pam iC5 e.6 sg7 EsD m-E tIL. m.' 9 1.'t s.1 m e O Vold fraction, aF m Et3 m -0. 5 -- I - 1. 0 - - Figure 6-2. Comparison of Creare Air / Water Head Correlation with Creare and B&W Data For V/a, Between 0.80 and 1.20. -. . _ _ _ _ _ _ _ _.,.. ~. _ _
m D Creare Air / Water e B&W REFERENCE I.7-1 1.5 -- a e m 5 a m a 5 8 e m e eg
- .s -
o 5 m go e a e e m e E E D E m a e 8 o m E go } O a e a e f e e@ m ea g s.5 -- E e m o gm m e 28 8 y eb Be e m o e e m, e j e e e o m e ? e
- e e
e oa _,e8 e e.e 8.1 5.2 s.3 8.4 sl5 sl6 e.' 7 sle sl, g,'s Vold Fraction. ar -0. 5 -- l l - 1. 9 - - Figure 6-4. Comparison of Creare Air / Water Torque Correlation with Creare Air / Water vs B&W Data For v/g Between O.80 and 1.20 . l .r
E e C-E O Cr*0re Steam / Water REFERENCE I.7-1 1.5 -- e I i.B - f e o e N e-i e e 2 e 3 e g s.5 -- i 8 e e 1 4 e a " m a a I $ E.8 ofy gfg ,f, 8I 82 5.'3 $4 s.'5 3.6 g,, Yold Fraction, or 5 I -0. 5 -- l Figure 6-5. Comparison of Creare Stean@ tater Head Correlation with Creare and C-E Data For p , 500 psia For V/g Between 0.90 and 1.10 t l t l l i 1
.+ i REFERENCE I.7-1 O Crocee Steam / Water 3.5 0 0 1.0 -- O 30O O O O 0 4 0 O c 0 0 OC0 I k O o O O h s.5 -- oO n. 3 E a2 E st et 8.3 e4 8.5 86 87 88 89 18 Vold froction. OF -0. 5 -- - 1. 0 - - Figure 6-7. Comparison of Creare Steam / Water Torque Correlation with Creare Steam / Water and C-E Data For V/g Between 0.90 and 1.10 and p < 500 psia _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
l Q.I.8 The condensation model is' described as follows on p. 35 of Ref. 10: "{g=K(1-X+X)(X,-X),where: c K = 1.0x10 (Kg/m -sec), empirical constant -5 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 steam environment. This _ causes the system pressure to be low and results in degraded heat transfer in the fuel bundle region as seen in the TLTA.and LOFT code assessment cases discussed in Section 5.0 of1 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." "The presence of a noncondensible gas is taken into account in the vaporization and condensation models described above." Clarify 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 I.3 and 1.21. However, the current RELAP5YA 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, X,-X, and therefore decrease the condensation rate. This is further explained below. s-r 1
. The static. quality of a gas. mixture.is defined by equation (77) on page 21 of' Appendix A it$ Re'ference I.8-1 as follows: ~' .X=X + (I} n v where: static quality of the gas mixture X = static quality of the noncondensible gas X = X, static quality of the vapor = 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: E U = (1-X)Uf+X + (} nn vv where: mixture specific internal energy U = liquid specific internal energy U = f noncondensible specific internal energy U = vapor specific internal energy U = y Equation (1) can be used to eliminate X in equation (2) to obtain: U = (1-X)Uf+XU + (X-X )U (3) 9 n y a
r The equilibrium quality for the gas mixture, X,, is obtained from equation (3) by. setting X!= X,, Ug = U,, U =U, g y y where U, and U, are the liquid and vapor specific internal energies g y at the saturation state corresponding.to the vapor partial pressure, Py (T,).- We then obtain: X, = (U-U, + (U,-U,)X ]/(U,-U,) (4) f y n y g The static quality difference, (X,-X), is obtained by subtracting X from equation (4) and then substituting equation (3) in to eliminate U. This yields: (X,-X) = ((1-X)(U -U,) + (X-x,) (U -U,)1/(U,-U,) (5) g f y y y g We now examine the limiting condition where the noncondensible quality, X,, approaches the fixed _ gas mixture quality, X. When X > 0.5, the
- least massive phase is the liquid. Therefore the 5 equation model sets
= U, (P ). Then equation the liquid phase to saturation conditions, U g f y (5) reduces to: (X,-X) ='(X-X,)(U -U,)/(U,-U,)- (6) y y y g which tends towards zero as X, tends toward X. When X < 0.5, the least massive phase is the gas. Therefore the 5 equation model sets the vapor phase,to saturation conditions, U = U,.(P ). Then equation (5) reduces y y y to: (X,-X) = (1-X)(U -U,)/(U,-U,) (7) g g y f 4 4 m
From pages 21 and 22 of Appendix A, Reference I.8-1: U =Uf (P, T ) g g U, = U, [P (T,)] 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,"Ug g .and the static quality difference tends to zero. Therefore, the condensation rate approaches zero as the noncondensible quality approaches the gas . quality. 'i, References . (I.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 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) .Q.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. -
0.I.23 Pages 166-168 of Reference 10 describe the energy transfer to the accumulator nitrogen as the sum of three terms: (1) the heat convected from the wall to the nitrogen, (2) the heat convected from the liquid surface to the nitrog'en, and (3) the additional energy transfer caused by condensation in the gaseous phase of the liquid that is vaporized at the interface. In the presence of mass transfer, the driving potential for heat transfer (Lewis number assumed equal to 1) is the enthalpy difference rather than the temperature difference (Reference 14). Clarify why the heat convected from the liquid water is in terms of temperature rather than enthalpy. A.I.23 The accumulator model presently incorporated into RELAP5YA corresponds to the RELAPS/ MOD 1 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: I g( w I Qcond " vap ~ D The temperature difference which appears in Equation (206) is one of the terms used to assess the water vapor concentration gradient at the liquid-gas interface. The concentration gradient is used in the l context of Fick's law with the objective of assessing the evaporation l rate M,,p, as described in Appendix I.23-1. l l l l l l [
APPENDII I.23-1 Assessment of the Evaporation Rate at the Liquid-Gas Interface Equation (206) of Appendix A to Reference I.23-1 can be derived by assuming that the vaporization at the liquid-gas interface in the accumulator can be approximated by a quasi-steady diffusion forinulation. Then, applying Fick's law, which is given by Equation (3-12) of Reference I.23-2 in the form: (1) M,p = - A, e ed* m,p y y where: h.p vaporization mass transfer rate,(Kg/s) = y 2 liquid-gas interfacial area,(m ) Aw = 2 diffusivity of water vapor in air,(m /s), equal to that in d = nitrogen These quantities are defined consistently with the nomenclature of Appendix A to Reference I.23-1. Further1more,fisthemixture(Nitrogenand vapor) density at the interface (which is assumed to be equal to the saturated is the water vapor density, , at the liquid surface). Finally, a,p y concentration of vapor at the liquid-gas interface (kg of vapor per kg of mixture), i.e.: van
- van (2)
V* 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 edo[h(P,p/P)I (3) y g y D where P is the vapor partial pressure at the interface and P D mixture or dome pressure, which is not dependent on x. The perfect gas law can be once again used to estimate: P b E1 T (4) P dx vap T D Substituting Equation (4) into Equation (3) and integrating over a diffusion length, L, leads to Equation (203) of Appendix A to D Reference 1.23-1 in the form: M,,,. - f,. A,. h.$,.(T,-T,) (s) 3 where$,.T;2 repres.nt, the equilidrium compressibility for a p.cfect gas and h. d/L in (m/s) is a mass transfer coefficient. 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.23-3, which is easily rewritten in the form of Equation (204) of Appendix A to Reference I.23-1 af ter minor algebraic transfonnations. References (I.23-1) R. T. Fernandez, R. K. Sundaram, J. Chaus, A. Husain, J. M. 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) (I.23-2) W. M. Kays, " Convective Heat and Mass Transfer," McGraw-Hill, 1966. (I.23-3) J. P. Holman, " Heat Transfer," McGraw-Hill, Fourth Edition, 1976, Chapter 11. 0.I.24 Clarify how the energy transfer from evaporation of water at the liquid surface is included. Assuming the heat-and mass-transfer analogy, the ness flow of vapor leaving the liquid surface is equal to g,(w -w )/(w - 1) (Reference 14), where g, is the mass transfer coefficient and w,is the mass fraction of water at i the interface or steWrin the main 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 the condensation rate of steam in the gas space. Therefore, the energy 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 which is derived in Appendix I.23-1) recondenses (Equation (207)) and that the heat of condensation (Equation (208)) is added to the gas space. The impact on 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 is neglected and the liquid is assumed to remain at constant temperature. In summary, the condensation energy is accounted for in the gas space. In the liquid slug, the evaporation energy is neglected because it has no effect on the injection rates. __
f 0.I.25 Justify Equation (3.4-32) of Reference 10. Please supply a copy of Reference 3.4-1 of Reference 10. .e A.I.25 / This equation corresponds to Equation (206) of' Appendix A to Reference I.23-1 in the current accumulator model. The equation is . justified in Appendix I.23-1. Requested reference is supplied as Appendix I.25-1. i g-V / ~ ms8 8 / f t I i,
4 4 _ APPENDIX I.25-1 A MECHANISTIC ACCUMULATOR MODEL FOR LIGH1 WATER REACTOR TRANSIENT ANALYSIS' K. E. Certson, D. L $8egel. V. H. Ransorn and J. A. Tropp EG&G Idaho, lac. Idaho National Eng neering Laboretwv idaho Falls. Idaho ABSTRACT V Volume A lumped parameter accumulator model has been developed and installed into the computer program z Elevation RE L AP S. The model includes heat transfer from the creek wall of the tank, from the water surface to the dome, vaporization from the water surface to the dome, and condensation of vapor in the dome. The B Coef ficient of isobaric thermal expansion list feature also includes heat addition to the a Change of nitrogen dome due to the condensing process, Dens i ty Essults from data comoarisons are presented o NOMENCLATURE Kinematic viscosity v Cross-sectional. flow area Subscripts C MJ1tiplier for heat transfer correlation c 0)ndense Cy Specific heat at constan.t volume f Liquid 0 Olameter 9 Vapor d Olffusivity K Upstream volume index (Accumulator) f Coefficient of friction L Downstream volume index (System) g Gravitational constant 1 Surge line Gr Grashof number n Ni trogen h Entha lpy S Saturated hg, h2 Heat transfer coefficients t Tank h3 Mass transfer coefficient v Vaportration k Thermal conductivity INTRODUCTION L Leng th M Mass This paper describes an accumulator model that has been developed for installation into a systems b Mass transfer rate analysis computer code (RELAP5 ), for analysis of 8 P Pressure light water reactor (LWR) behavior. This model Pr Prandit number features mechanistic relationships for heat transfer from the tank wall and water surface, condensation Q Total beat transfer in the vapor dome, and vaporization from the water R Nitrogen gal Constant T Temperaturt The RELMS computer code and its documentation. a. t Time RELAPS/M001, Code Otscription Vol. I and II by V. H. Ranson et.41., Are available fro.a the National Energy Sof tware Center. 81dg-208-Anom C-230, 9700 si Internal energy South Cass Avenue. Argonne, Illinals'. 'we4 su,,eries by ow u. s. Musiser Repaleeuf.M Ofnes of Nuclear namelesary Reneereh e COE Centract A DE,AC07 764001870. r 3 + j i i surface to the_ dome. The a:c.mulator modelFconsis ts' r of a hydrodynamic model and c heat transfer model. The.latter includes the vaoorization and - vm v. r -,,.r. ma** N condensation ef fects. The hydrodynamic model is described along with + + + ~ the heat transf er models for the accumulator. u w. Performance of the model is verified by data _ comparisons between the model results and the ,,,,,.,, h 9 LOFT L3-1 test and also a Wastinghouse Electric Corporation.tt)per Heat Injection (lHI) accumulator performance test. HYDRODYNAMIC MODEL cm. u au s s The Accumulator is modeled as a i lumped-parameter component. This model was chosen 3 over a distributed model because the spatial -gradients in the accumuletor are expected to be .. n ma. small.
===ar . e er w. g p The accumulator redel and associated notations ]/: .are shown in Figure 1. The bas 1c modeling ' Figure 1. Typical accumulator. ~ assumptions are: Using these assumptions, the basic equations 1. Heat transfer from the accumulator walls, governing thermal-hydraulics of the tank and Heat and mass transfer from the 11guld are discharge line can be written as modeled using natural convection correlations assuming siellarity between 1. Conservation of mass (nitrogen) heat and mass transfer from the liquid = constant 4 o, V, (t) M . surface. n a 2. 'The nitrogen is modeled as an ideal, gas - 2. Gnservation of energy with Constant specific heat. The steam in , the dome exists at a very low partial Nitrogen pressure and'is not modeled.directiv. The energy released as a result of vapor dU g[dV +Q W M .-P condensation is transferred to the n nitrogen. Wall ~ y3. Because of the high heat capacity and f ' large mass of water below the interface, di the water is assumed to remain at its M U t v E t . initial temperature. The model for liquid flow includes 4 inertia, wall friction, form loss, and g ,2 ,[3 gravity ef fects. ,,3 (ti 22 "
. - (P - P )A, + ogg (zg
- r )A, (4)
Qt.gnj (Tg - T, ). (9) g y where f is the coef ficient of friction. The wall convective film Coefficient has been ~ defined u i g a correlation for turbulent natural 4. State Relationships convectio for vertical plates as Pygn*HRT (5) k nn hg. 0.1 (GrPr) g (10) Un. C, i. (6) n n where L is the sum of the lengths of the tank wall above the liquid surface and the radius of the tant Using Equations (1) and (6). the nitrogen top. The Grashof number is given by energy equation (Equation (2)). can be rewritten as 3 dT .gs(T-i} 9 n IIII Mn fv --PvAf y + Q. W g n where L 15 as above plus the radius of the liquid Differentiating Equation (5), and solving for the surface. pressure derivative yields In addi n to wall heat transfer, convective heat-dP P dT - P dV, transfe from W acc M a w H p H is ass e g g n g W.*K _W ~ Q W IOE to be . Equations (4),' (7), and (8) comprise the system Of*AK 2 (Tg-T) (12) h n of three differential equations used in the accumulator model. They are used numerically to where advance T,, V,, and P in time. g HEAT TRANSFER 2
- I*II The variable D is the diameter of the water surface The heat transfer to t5e nitrogen in the dome and the Grashof number is the same as above.
has been modeled as being due to two dhteet ' proc 1sses. The first mode is convective heat transfer Liquid is also assumed to evaoorate from the du3 to the temperature difference between the water surface due to the density difference between nitrogen and the accumulator walls,and,liqut(. The the vapor at the water surface and in the nitrogen l .secind 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 done is assumed to be saturated at .the gas temperature. Steam is assumed to be "Y -h3 % (og,- o ) g c nvected from the liquid surface due to the motion -sf the nitrogen. A mass and energy balance for the .h A s9 (Tg-Tf) (13) steam is then used to calculate the heat transfer to 3 ge91 the altrogen due to convection of vapor from the liquid surface and subsequent condensation in the where accumulator done.
- gg a function of W W The heat transfer from the tank wall has been temoerature.
4 - modelid with a lumped-parameter model ~ The mass transfer film coefficient, h, is 3 23., _ ~... _,,-- -.
determined assuming similarity between heat and mass CALCULATED RESULTS transfer from the water surf ace (both are conserved properties modeled by similar equations), and is The accumulator mndel described previously has given by been tested in RELAPS and gives good agreement with experimental data from LOFT Test L3-1, and a gh)l/3 (14) Westingnouse & accumulator performance h3"h2 \\ n/ test (WCAP-8475 . The comparison to the experimental data in both tests is of the form of where d is the dif fusivity of water vapor in ai pressure versus nitrogen volume. This type of dzfined as comparison makes the effects of heat transfer in the accumulator more apparent than either pressure or P 2.3 temperature versus time comparisons, and also T d - 0.239 (15) obscures most of the ef fects of uncertainties of the boundary conditions. The LOFT data used for where P, is atmospheric pressure and 0.239 is a dimensional constant that gives "d" the units of comparison are liquid level and pressure history; 2 lm /s), the Westinghouse data used for Comparison are integrated mass flow and pressure history. The mass transfer rate is then The LOFT accumulator was simulated using the f1
- 1"9 ****"d d*t**
"i *' 9'" ' l"** 15 U3 b, h hI b 'g 3 g (T7 - T ). (161 1.39 m, water volume was 1.97 m3 (0.45 m3 ~ o was 2 n 3 arge Ifne and 1.52 m, in tank), system pressure wis 4.37 HPa, and the water and frogen temperature the steam in the accumulator dome will condense as j 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 vaporization will be transferred to the nitrogen in as wall friction and form loss in the surge line. the dome. The condensation rate is calculated Time step size was I 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, providing a E - k + h (V 'g ' I II I g n The heat transfer due to this Condensation is comparison plot between the LOFT measured data and calculated results. Note that the plot includes a "g (T ) - h((T I g'g) calculated isothermal curve (uppermost curve) and a Oc* n n g calculated 15entropic curve (lowermost curve). In the early part of the blowdown (high pressure and + m, h (T ) - hp (T,) (18) low vapor volume) the expansion of the nitrogen gas g . S 5 is essentially adiabatic. As the gas expands into the accumulator dome, its tempersture decreases. The total heat transfer to the 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 pressure to decrease dU less rapidly. The calculated and esperimental gM-(191 pressure versus nitrogen volume curves are seen to 0O *Og*O g c j be in good agreement, indicatinw that the heat transfer was accurately predicted. Approximately sixty percent of the heat transfer was calculated to be due to condensation. 24 S 1 I i 1.28 i -- T Calculated Isothermal --Italculated Isothermal .... !Nbs 2o ENlrscalculation -- S calculated tsentropte _ g =---;3 aiculated Iseatropte s t s O t ' s,' s h \\ \\ y3 %N s e .64 \\ t U,'s 3 's,',., { s s .T a N -2 N .32 - N ', s s ~~ 5 ~~____,T ,~ ~ i 1 I I I I e 5 2.s 25
- 3. e
.25 .52 .77 1.03 Nitrogen Volume (m ) 9'" "** ' O N n t figure ?. LOFT L1-1 4ccumulatt
- i,.e 3.
Westinghouse sn. A. aulator. The Westinghouse lAll accumulator was the secon SLP91ARY experiment 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). DJe to indicate that heat transfer, especially from the thi large pressure drop this performance test condensing vapor in the dome, is important for resembles a large brealt type of experiment providing accurate representation of accumulator simulations. a fast accumulator transient for vert fying the The RELAPS accumulator model gives good results for accumulator model. This accumulator was simulated both slow and fast transients,15entropic or by using the measured data and performance polytropic. This technique gives the user a information contained in WCAP-847has input. The mechanistic accumulator model to use in the uncertainties for the data were not given in this nulation of light water reactors, propriatary report ~ REFERENCES The form of the cuniper[ sns between the experi-mental data and calculated results are of the form of J. p. miman, wat Transfer, 4th edition.New pressure ratios versus volume ratios in 'rd'r to o e York: McGraw-Hill Book Company, [nc.,1976, preserve the integrity of the proprietary data. pp. 280, 244-245. Figure 3 shows the results of the comparisons between th7 calculated results and the experimental data. 2. R. J. Hanks, 1959 Water Vapor Transfer in Dry The calculated isothermal (uppermost curve) and isIntropic (lowermost curve) curves are included soll, soll Science Society Procedings,1958, 22:372-374. f:r reference. The experimental data shows a siellar . behavior seen in the LOFT test, that is, the gas 3. WCAP-8479. Westinghouse Electric Corpora-expansion is initially an 1sentropic process. Then tion Nuclear Energy Systems, P. O. Box as the temperature difference between the nitrogen 355, Pittsburgh, PA 15230. and its surroundings increases, the gas pressure decreases less rapidly. The calculated results
- p. D. Rayless, J. B. surlow, R. H. Averill, are again in good agreement with the experimental results.
Experimental Data Report for LOFT leuclear 3ssall Break Experiment, MUREGICR-1145. EG8G-2007, January 1980, pp. 2R,127,140. 25 e
5. H. J. Welland, LOf T Emperiment operatine enkes any warranty, expressed or implied, or assumes Specification Small treak Test Series L3 a v legal liability or responsibility for any third levenber 1979, MC L3 ' Series (05, Rev A, pp.15 party's use, or the results of such use, of any information, apparatus, product or process disclosed g in this report, or represents that its use by such third party would not infringe privately owne'd This report was prepared as an account of work rights. The views, expressed in tnis paper are not sponsored by an agency of the lhlted States necessarily those of the U.S. Nuclear Regulatory Government. Neither the lhlted States Government Cbemission. nor av agency thereof, or any of their employees, l l l I \\ N.
0.I.26 Clarify how the analogy between heat and mass transfer was used. A.I.26 The analogy between heat and mass transfer was used as described in Appendix I.23-1. The expression for the relation between the mass 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. A.I.27 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 given by Equation (208) of Appendix A to Reference I.1 and cond again with Qcond = 0. The calculated accumulator pressures and flow rates are essentially identical as indicated in Table I.27-1 and Figures I.27-1 and I.27-2. I WITH EVAP/COND. WITHOUT EVAP/COND. TIME FLOW RATE PRESSURE FLOW RATE PRESSURE (SEC) (KG/SEC) (PA) -(KG/SEC) (PA) t 3. O. 1.28200E+06 3. 1.28200E+06 2.5003 0. 1.28200E+06 0. 1.26200E+06 5.0003 0. 1.28200E+06 0. 1.28200E+06 7.5000 106.41 1.27538E+06 106.40 1.2753eE+0e 10.000 101.51 1.26796E+06 101.50 1.26783E+06 12.500 333.27 1.25255E+06 333.21 1.25 217 E+0e 15.000 547.24 1.22266E+06 546.97 1.22157E+06 i 17.500 744.43 1.1815 4E *06 743.55 1.17898E+06 T 20.000 920.94 1.13 295 E +06 918.63 1.12778t+06 22.500 1075.5 1.08047E+06 1070.5 1.07135 E*06 25.000 1206.3 1.02711 E+04 1196.9 1.01272E+06 27.500' 1310.1 9.75263Ec% r 1294.2 9.54385E*05 30.000 1384.6 9.26660E+05 l .1360.4 5.9 8324 E+05 32.500 1422.6 8.82360E+0$ 1388.5 S.45898E*05 35.000 1383.4 8.44043E+05 1340.0 7.99057E+05 37.503 1347.3 8.11502 E+05 1293.9 7.57976E*05 40.000 1315.0 7.83389E+05 1251.4 7.21653E+05 42.500 1285.7 7.58693E+05 1212.2 6.89303E+05 45.000 1258.7 7.36662E+05 1175.6 6.60304E+05 Table I.27.1 Comparison of Calculated Maine Yankee Accumulator Pressures and Flow Rates With and Without EVAP/COND. Heat Transfer Terms
a "O O ~. S o 1-WITH EVAP/COND HEAT TRANSFER o tj - 2 L v 's N t> $ _ N (M 's e w 's re ~, s (% 's c 's o m -- WITHOUT EVAP/COND HEAT TRANSFER-oc-o s i 00 90 18 0 27 0 360 450 TIME (SEC) Figure I.27.1 Comparison of Calculated Accumulator Pressures With and Without Evap/Cond. for Maine Yankee Test
0 5 4 5 N O C ~ / P n d a A V 0 h E ,6 t ~ 3 i W T s e s U t O a R H w T ot l s I W T F e R roe E t e 0 ak F l n S ,7 2 ua N mY u ) ce A C cn R E Ai a T S dM ( e t r T ao A E lfu E M c l d H an IT Co 0 C D f/ i8 op N a 1 nv O oE C s it / ru ao P ph A mt oi V CW E 2 H 7 T 2 0 I I i9 W erug iF 0 0 o8s o83 oo8 o00 o@m o W\\CMv V)EE< bo3 u d E iD
Q.I.28 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 RELAPSYA result. 1 A.I.28 The friction factor and fom loss factor given in Table 2.4-1 apply to the marge 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.23-1. A trip logic is combined with an appropriate valve opening speed resultinr, in valve throat ratios as shown in Figure I.28-1. This valve area history is 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
o .s 4 0 I 4 tseT 0 e e I k 3 na Y / en 0 i / a M 5 I / 3 e hT / g n / 0) ir .C u I / $ E D 2 -/ S yr ( / o t // s 0E i H / 5M I a / 2I er / T A // ev 0 l / a V /- I 1 ,./ 1 8 / 2 ./ 0 I / e r I 1 ug i F o I 's 0 .o O.~ cd , 0 e.O mo o.a o - xcrcxrwe 0s-eHcx HCoxrH W>3 > 1? i )- , ;i -!ii.
4 0.I.29 Clarify which curve is which and why the difference in pressures is increasing after 32 s even though the liquid levels are virtually identical. A.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 RELAP5/ 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 I nitrogen regions that would indicate the temperature difference that existed during the transient. f A.I.30 No records of temperature measurements in either region have been found.
0 ~ 54 o* h. 0 I ( b t 4 s ) e .T dne oe 0 Ck / n pa I d aY 3 v E en tiua oM 0 ht r i o I wf 3 ds ne aru h s 0) t s .C i e wr I $ E ( P 2 S dr eo ( ttaa l l 0E uu cm bM l u I ac 2I CcA T f od e nr 0 ou ss i a I d re i aM p md on N Ca O CD 0 /N 1 I a PO b 9 AC l V/ 2 EP I A e TV r UE u O 0 g AHH I i TTT '5 F AII DHH 0 0 a a of-O.v~_v ie
- om
.m og o4 n 'o gX a-Q E D i D [3.M a_ C o a a D E D O o CE nn Y - I b7
- i
0 54 0 t s i 0 i e 4 T e ) e .k d n 0 na oY C i 5 / e I 3 pn aiva E M t r 0 uo of h i 0 t s I 3 ilwev d e nL a 0) d .C hi t u i I5 E iq 2 wi S L ( ( d r eo t t 0E aa l l .0M uu i I cm 2I l u ac T C cA f od 0 e nr ou i I5 ss 1 i a re aM p N md O on C 0 C a /N i ^ PO I0 2 RC 1 V/ 9 EP 2 R I TV UE e O 0 r RHH i. u i TTT 5 g i AII F DWW 0 r. 0 " o Vd o. y. - g o t c 1NN 1_u>tJ Os O s1-a ) t &y
III. HOT-LEG OR U-BEND PHASE SEPARATION O.III.1. O.III.2. O.III.3 and 0.III.4 The procedures used 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 j removal capabilities. The imbalanced secondary condition does not preclude adequate heat l 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 III.1-3) showed j adequate heat removal and no flow stalling with up to 13% of the l primary volume filled with noncondensible gas. This is due to the fact that the noncondensible gas disperses and does not collect in the U-bends to stall the flow. Only when the noncondensible gas was injected into the steam generator at a rapid rate did the flow stalling occur. For Maine Yankee and the Yankee plant, small break LOCA conditions for which natural circulation is desired, the amount of noncondensible gas calculated to exist is about 0.5% of the primary system volume. Therefore interruption of natural circulation is not expected for small 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 Pump Operation During Small Break LOCA Transients Maine Yankee," YAEC-1423, April 1984. (III.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 Reactor Geometry Under Accident-Induced Conditions," Nuclear Science and Engineering 88, 311-320 (1984). 0.III.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. t a f
A.III.5 No special treatment is given to the heat transfer in the top of the U-bends. The standard RELAP5YA heat transfer models described in Sections 2.1.3.6 and Section 2.1.3.4, Appendix A of-Reference III.5-1 are used. The effect of noncondensible gases is not considered and has no significant impact on the heat transfer coefficient. References (III.5-1) R. T. Fernandez, et al., "RELAPSYA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description," YAEC-1300P, October 1982. (Proprietary)
IV. STEAM CENERATOR HEAT TRANSFER 0.IV.1 Page lii of Reference 15 (which modeled FLECHT 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 eliminated 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 Maine Yankee plant (Reference IV.1-1) and the Yankee plant at Rowe (Reference IV.1-2), and the simulation of the LorT L3-6 experiment (Reference IV.1-3), no such oscillations were encountered. However, if such oscillations are encountered, the time step will be reduced. References (IV.1-1) L. Schor, et al., " Justification of Reactor Coolant Pump Operation During Small Break LOCA Transients, Maine 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," l YAEC-1437, July 1984. l (IV.1-3) 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 III: Code Assessment," Yankee Atomic Electric Company Report l YAEC-1300P, Volume III (October 1982). (Proprietary) !
O.IV.2 clarify what steam generator noding will be used for SBLOCA calculations. A.IV.2 Figures IV.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. 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. 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 10 volumes and the secondary side consists of 14 volumes. All area changes, baffle plates and dome structures (i.e., separators, dryers) are accounted for in the nodalization through the loss coefficients and geometries that are entered as input. Besides the U-tubes, heat slabs representing the external walls are included in the model. YANKEE PLANT AT ROWE 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 l 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 vessel are modeled as heat structures. t
i R.puu k'r-W.: 9
- .;-I-prn.
n" - E, ! [- M i,. ..( -{ 'i- -:. t 2. O J' ? E i in v g '. trJ ,k e. CUsj ---L-q...'...- 33 a T 3 ~ [- r- ?lt: :e'- s:::.:: l. e a o ,E . %M ' ; ;- 0
- l' ' F-g E.
OJ u+ -r $ VJ h g 2 .~.. D (.,.a:p') 31. 4 g 1 a-1 2 E=. f ,1 { s ,f. b a P g i...'... g,, .g 1' r u e r ) ;-I-IfirT[T. L if s _e %.n a
- 2 e
s C e % s. o nee t l S7,l e n _ r 1 o 2 I i = = t r 7f i a m e e z F-I r t i l I e a p a d e o l r N W oo s ow t n l t no r o ee ai N kst l t i s aen e eee P a r o rrs l eee" m bpe eu ee i r e n' n esp ec e s v" e r oae kl e 0 o e 2 e hhr na c l 2 n e ppp aC Le v oao Y en oeo t dD s lll so n r a ett e o ~. ecc sA l s L( = m t t ee r et t oC i d q w_ lgab e r ee pO i bSS s oL 1 l o i o r t eee rB Y r f a l l PS r t s. w m neo a I a il e c i u sed 2 i .N m w a c o s 2 r V o n I r pug e= e e es r u g i F .e = j I .I I I I ,4 s i + is 3-s I f ni n a u +,, i 1 os + i 5 m., 2 I I 7 I e i s. e n 1 e ks s 2-4 9 r 2-i I 1 I ,,v ?
l 0.IV.3 Clarify what noding is needed to model steam generators for low flow and reflux cooling conditions in SBLOCA. A.IV.3 Figures IV.2-1 and IV.2-2 provide representative noding schemes which will be used for all the phases of a 38LOCA accident, including low flow and reflux cooling conditions. -60
V. SYSTEMS VERIFICATION AND OTHER EXPERIMENTAL VERIFICATION 0.V.12 clarify how single-phase and two-phase reverse flow through the pumps is modeled for locked or unlocked rotors. A.V.12 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 contained in Reference V.12-1. Appropriate input data would be used as discussed in the answers to Questions V.13 and V.15. For the unlocked rotor 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 valves in each cold les that prevents significant reverse flow from occurring. l Reference (V.12) 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-Hydraulle Analysis," Yankee Atomic Electric Company Report YAEC-1300P (October 1982). (proprietary) -61
0.V.13 Clarify how prototypical the data used is for full-scale pumps. e.V.13 Table V.13 summarises the single-and two-phase steady-state data 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 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). 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 Yankee pumps than the Semiscale pump. 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. EMBE foecific Speed Vermont Yankee 3802 Yankee 4687 Maine Yankee 5556 CE-EPRI 4209 semiscale 926 62-
c. This choice is consistent with the two-phase homologous data base in the principal regions of interest (Curve Types 1-6) discussed 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 Vermont Yankee pumps than the Semiscale pump parameters and design. These data are only used for Curve Types 7 and 8 in the Yankee 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-IPRI 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: a. Steam water is the two-phase fluid medium of principal interest, b. The geometry and specific speed for this pump are closer to the full scale pump parameters than those of the Semiscale pump. c. These data are more consistent with the single-phase homologous data selected for the full scale pumps. The Semiscale steam water data base is used for Curve Types 7 and 8 (positive flow negative speed) in the two-phase region. Data are entered for these two curve types for completeness. However, these regions are not important since it is unlikely that these conditions will be encountered in LWR LOCAs. In fact, this consideration is the principal reason why these regions were generally excluded from the CE-EPRI 1/5 Scale Test Matrix (Reference V.13-1). Reference (V.13-1) W. G. Kennedy et al., " Pump Two-Phase Performance Program," Volume 1 through 8, RPRI NP-1556, Electric Power Research Institute, September 1980. ~63-
TABLE V.13 Data Sources For Reactor Coolant Pump Models Curve Number 1 2 3 4 5 6 7(a) 8(a) Head Curves HAN HVN HAD llVD HAT HVT. HAR HVR Torque curves BAN BVN BAD BVD BAT BVT BAR BVR Maine Yankee Single-Phase NY MY MY MY MY(D) MY(D) MY MY Two-Phase EP EP EP EP EP(D) EPID) SS SS Yankee Rowe Single-Phase YR YR YR(c) YR(c) EP(c) gp(c) MY MY Two-Phase EP EP EPIC) EP(c) EP(c) gp(c) 33 33 Vermont Yankee Single-Phase VY VY EP EP RP EP MY MY Two-Phase EP EP EP EP EP EP SS SS Key: MY = Maine Yankee Pump Data EP = CE-EPRI Pump Data YR = Yankee Rowe Pump Data VY = Vermont Yankee Pump Data SS = Semiscale Notes: (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. (c) Yankee has noneeturn valves in each cold leg; therefore, reverse flow (this region) will not occur. %
/ 0.V.14 Clarify what range of conditions have been used to test the pump model. 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 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 roter 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 thls transient. Since each loop contains noneeturn valves, there was no significant backflow in these two loops from contiquad operation of the remaining two pumps. Plant test records contain the square root of the pressure difference between the inlet and outlet plena, normalized by the initial steady-state value. This was taken as l an indication of the " normalized flow" in each tripped loop. l These two curves are shown on Figure V.14-1 which is attached. I The Yankee SBLOCA model (Reference V.14-2) 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 l-parameters yielded similar results for both tripped loops, and are also shown on Figure V.14-1. The c'omparison 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 l
provides confidence in the homologous method for predicting 4 single-phase, transient pump performance. b. TLTA Pump Coastdown - Sinale-Phase Conditions TLTA Test 6425/R2 was a large break blowdown test with ECC 4 injection performed in the Two-Loop Test Apparatus by General 4 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 centrifugal 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 flow rate predicted by the RELAPSYA 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. l c. LOFT Pump Operation and Coastdown - Sinale-and Two-Phase Conditions LOFT Test L3-6/L8-1 was a PWR small break test that combined a long period of pump operation under single-and two-phase conditions, a subsequent trip of both pumps at 2371 seconds, and continued testing until 7469 seconds. This test has been simulated by the RELAP5YA model of the LOFT facility from 0 to 2460 seconds for code assessment purposts. The pump models used LOFT test data and a fixed inertia rather than the variable,.
inertia that exists in these components. Many calculated parameters are compared to test data in Section 5.3 of Reference V.14-3. Figures V.14-4 through V.14-6, attached, show additional comparisons of predicted pump performance parameters to available test data. 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 modeling pump two-phase performance. Figure V.14-5 compares the calculated 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 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 during the coastdown period from 2370 seconds to 2450 seconds. The two curves show similar coastdown trends, but with two i notable differences. First, the two curves are offset by about l 4.4' seconds due to the lower in'itial speed and the plotting edit frequency f or the calculation versus the measured speed / history. Second, the slope of the calculated speed history is l more negative than the test data for speeds above 1000 rpm, l about the same for speeds between 700 and 1000 rpm, and less negative 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 L 1
inertia values to the value of 274 lb -ft used for our g simulation. The slope of the pump speed equals the shaft torque i divided by the moment of inertia for a free-wheeling pump: dw/dt = 1 /I FT 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 RELAP5Yg) 2 (rpm) (lb - ft ) (1b - ft f f 3200 273 274 3000 289 274 2500 318 274 2000 325 274 1500 311 274 1000 274 274 500 216 or 34* 274 100 154 or 34 274 0 136 or 34 274
- The test report does not make clear whether the variable inertia flywheel was engaged or disengaged below 750 rpm.
If l disengaged, then the lower inertia value applies. References l l (V.14-1) "The Startup Experiment Program for the Yankee Reactor," YAEC-184, Yankee Atomu Electric Company, June 1961. (V.14-2) J. N. Loomis J. H. Phillips, A. Husain, " Reactor Coolant Pump Operation During Small Break LOCA Transients at the Yankee Nuclear Power Station," YAEC-1437, Yankee Atomic Electric Company, July 1984. , l
(V.14-3) Fernandez, R. T., R. K. Sundaram, J. Ghaus, A. Husain, J. N. Loomis, Lo 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.14-4) L. S. Lee, G. L. Sozzi, S. A. Allison, "BWR Large Break Simulation Tests - BWR Blowdown / Emergency Core Cooling Program," Volume 2, NUREG/CR-2229, EPRI NP-1783, GEAP-24962-2, General Electric Company, July 1982. l . l
1.0 0 A Q_ Q i i ^ i 0.9 ~ Pump Trip 0.8 g .c n,y 'A D rm 0.6 8 E m 0.5 8 -g 0.4 A 8 G A w4 0.3 9A a_ _s A Loop 1 Data g 9 Loop 4 Data A ~ REW5YA ' A 0.1 ) A 9 0.0 2 3 4 5 6 T 8 Time (Seconds) 1.0 l I s a 0.9 Pump Trip 0.8 0.7 0.6 RELAP5YA Y g 0.5 o 0.4 m $x 0.3 y E ey 0.2 8 OZ 0.1 0.0 I I e i i 2 3 4 5 6 7 8 Time (Seconds) Figure V.14-1: RELAP5YA Comparison to Yankee Rowe Pump Trip Test Data,
10 TLTA Test 6425/R2 ( \\ \\( "I' E 8 e E4 g ) I Yn .o 6 g 1 j DATA g y 4 + S N E 2 q 2 ( R5YA a 0 m -2 i i i 0 5 10 15 TIME (seconds) FIGURE V.14-2: INTACT LOOP RECIRCULATION PUMP MASS FLOWRATE 200 TLTA Test 6425/R2 160 -{ -- t.4 _c. 120-o< N x
- " "*""i'""
m 80 t EDA R5YA 40 l l 0 0 5 10 15 f TIME (seconds) FIGURE V.14-3: INTACT LOOP RECIRCULATION PUMP HEAD, 1 L-
-w-, -CL-DIFFERENTIAL PRESSURE ACROSS PCP (PSI) PMPHERD 120 -20. -10. D
- 10..
20 30 40, 50, 60, 70 i o O O [ s e aj r s e ~ L ~ mm EP r "? 8 E4 ? > a-R 9 E $- f' se ms an a" EE m @- o O ~O r1 O O gg er 28 E-O C, e 8 G l_ n s O ti l_ ~ u s,= s n e l-GW _ 11 ] = s h 5 l l l C
0 0002 0 0 i042 = 0 0 i0 22 = a 0 t 0 a 0 D = 02 6 3 L 0 = 0 T 08 F 1 O L = o 0 t 0 d 0 ne 8 oe A6 1 sp = Y3 iS 5L r PAT 0 ap pm 0) LF 0Cmu = EO 4EoP RL 1 C S m a A r ( Y e 0.EST = c 0MP i0IA g 31TL nE o RL
- . c 0
5 0 i0 0 4 = 1 1 = V 0 e 0 r i0 u = 0 g i F = 0 = 00 f 8 = = 0 0 = 04 L = 0 I i0 l 0 2 I i - 3 0 i 0 a. og m o o *g~ o o oo 8 l>QnQ.. ^rQQ~ ogMn Qrsa E 0r i
n a s a 1 w_ -a a a E R I I / ? I l I O / f I ~ I ' l t L I I I! 3 o I Q I It I r I I O i I HD o I I e I s e I
- aE i
i of E R UM I I 8% ma i g I S O n 03 I i l -g C
- i 1
- o.
t l 1
== i l f l 'k i I I i 0 h I 1 -h r i g; I i i It I I O i 1 I t gg I 11 i l ag j 1 R om o-Q 0%e 0*0h3g 0*aner o con 7 0*cos O'a Ogf 32Adity ( % ) 0333g c541d.
0.V.15 Justify which set of pump curves will be used for SBLOCA applications: Bingham, Westinghouse, LOFT, Semiscale or other. A.V.15 The data bases that were used to develop performance curves for the Maine Yankee, Yankee and Vermont Yankee reactor coolant pumps are described and justified in Answer A.V.13. Additional pump modeling 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 manufacturer's data where available. The second source is the CE-EPRI 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. None of the Bingham, Westinghouse nor LOFT data bases are used. l { l i l f
VI. FLOW REGIMES 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." O.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 RELAPSYA 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., " Gas-Liquid Flow in Inclined Tubes: Flow Pattern Transition for Upward Flow " Chemical Engineering Sciences, Volume 40, (1983). O.VI.5 ~ Clarify if the stratified flow model is also used up to angles of 15 degrees. l A.VI.5 The stratified flow model is used up to angles of 15 from the L horizontal if stratified flow is calculated to exist. i l
0.VI.6 Clarify if there are any SBLOCA analyses that will use the stratified flow model for situations where the pipe area is changing and, if so, how the model would have to be changed to handle the nonuniform area. A.VI.6 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 nonuniform 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 this constant flow area. O.VI,7 Clarify how the slip is treated for the two-phase critical flow when the upstream volume switches from nonstratified to stratified flow and what effect this has on the break flow. A.VI.7 l l i In RELAP5YA, the determination of the slip between phases is primarily influenced by the interphase drag term, F defined in Section 3.1 of 7 ( 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 t junction void fraction is donored from the upstream volume. This applies whether the flow at the junction is choked or not choked. When the flow in the upstream volume is stratified, the junction void fraction is not donored directly from the upstream volume, but is modified to account for the stratified liquid level in the upstream l l
volume being above or below the junction location and for the mechanism of vapor pull-through or liquid entrainment. The method for estimating 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 with the phasic velocities, is also used to compute 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. O.VI.11 Clarify how the effect of noncondensable gases on natural circulation has been verified. 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 loop, a pressurizer voided in one loop, steam generator heat removal in only one loop, and flow stopped or much less in one 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. -
A.VI.13 Figures IV.2-1 and IV.2-2 represent the nodalization used in the pump 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 cooling. Even the ECCS is available in all recirculation loops. Therefore, the various loops are not expected 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 introduced in the upper downcomer regions. This is because:
a) The fluid experiences a change in direction at the injection t nozzles. The exact fluid trajectory in the downcomer depends on the loop flow differences and the resistance to flow in the vertical or annular direction. The annular region is quite free and a uniform flow is expected to exit the downcomer; b) The downcomer exit flow is going to experience a change in direction 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 h 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 azimuthal locations in the core. < -~
.-.._. =. l
- 3) A voided pressurizer does not play a significant role for those phases of a LOCA where core cooling is a concern. During the pressurizer refill period, the pressurizer loop may behave I
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. t 0.VI.17 "The two fluid nonequilibrium Page 9 of Reference 8 states: hydrodynamic model includes three options for simpler hydrodynamic models. These are a quasi-steady flow model, a homogeneous flow model, - and a thermal equilibrium model. The two-fluid, quasi-steady, or homogeneous flow models can be used with either the nonequilibrium or equilibrium thermal models (that is, six combinations). Most of the 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, i A.VI.17 The nonequilibrium option is generally used for volumes in our current Yankee, Maine Yankee and Vermont Yankee plant models. The exceptions to this general approach occur for several Time Dependent Volumes (TMDPVOLs) that represent boundary conditions in each model. Of these TMDPVOLs, only the feedwater and ECCS supply volumes act as fluid sources which deliver liquid to the systems. The remaining TMDPVOLs act as fluid sinks for which only the pressure is an important parameter. Therefore, the cases where the equilibrium option, rather than the nonequilibrium option, is used to represent certain boundary conditions makes no difference. 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 generator in the single broken loop, the single intact loop, and the. -r "t -e4 -w'--y-w w-w y-,,,ww-.--,,-w-w ig enw-y, y-%*-. w 3.-- emw e- - e .--u--- - - - -..e--.-.e--w www-e.-
i combined two intact loops. This region generally contains single-phase liquid; thus, the one-velocity option is appropriate. The Vermont 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 elevation in the vessel. The second represents the very small leakage path from the isolated portion to the active segment of the recirculation header. This latter 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. 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 RELAP5YA assumes that droplets are present in the annular-mist flow regime (voids between 0.85 and 0.95) and the mist flow regime (voids above 0.95). The average droplet size is used in estimating the interphase drag in these regimes. In the annular-mist regime, the average droplet size is postulated to be at its maximum and is characterized by a critical Weber number criterion. In mist flow, the average droplet size is assumed to be below its maximum value and the model for droplet size calculation is described in Section 3.1.3.4 of Reference VI.18-1. 'The maximum droplet size, defined by a critical 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 regimes. 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 (VI.18-1) Fernandez, R. T., et al., "RELAP5YA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I," YAEC-1300P, October 1982. _. _ _ _,. _ _..., _ _
\\ O.VI.19 Page 33 of Reference 8 states: "The phase-wetted perimeter has been -assumed proportional to void fraction." Clarify if this is done even for the annular flow regime and if so justify what approximations are introduced for SBLOCA calculations. 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 simplistically 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. However, the total pressure loss due to wall drag is preserved due to the phasic sum momentum equation. The wall drag has an impact on the phasic momentum equations in RELAPSYA which are solved for the phasic velocities. Also, the wall drag generally has a larger impact on the phase velocities at l conditions of high mass flux. For PWR SBLOCA conditions, the annular l 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 i Equation 128 on Page 35 of Reference 8 defines the HTFS two-phase friction multiplier correlation. Clarify what value of C, the l correlation parameter, is used. 1. . ~.,_.
A.VI.20 The correlation parameter, C, is based on the HTFS correlation (Reference VI.20-1) for two-phase pressure drop and is calculated as: 10( p) + 2.5)2 (log 2 i C = -2 + (28.0 - 0.5 XG)exp [ ) _4 G x 10 - 2.4 where: f Baroczy property index = p ( )0.2 = mass flux G = 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. O.VI.21 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, A and A, are used directly with f 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 calculation 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 -
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 drag component has minimal impact. 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 of Reference 13 states: "...the assumption of zero mass transfer is questionable for large area changes. Large area changes give relatively large pressure changes and hence mass tear 2f er 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 transfer assumption. A.VI.25 No formal limitations have been placed on the use of the abrupt area change model. Mrhe partial quotation, taken out of context from the 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. Examination of Equations 231 and 232 in Appendix A of Reference VI.25-1 nc u es the shows that this pressure change, (P ~ 'L K following phenomena: a. Temporal and spatial phasic acceleration. b. Phasic gravitational forces. i l l c. Phasic wall friction forces. _,
d. Interphase drag forces. e. Virtual mass effects. f. Interphase mass / momentum transfer. Section 2.1.3.4 of Reference VI.28-1 states that RELAP5YA and RELAP5 MODI contain models to account for local irreversible pressure losses from abrupt area changes in the geometry of flow paths. These models provide additional terms to the two momenta 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 not a replacement for the two momenta equations. The abrupt area change model provides one of two options to account for the additional local pressure losses. The RELAPS 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 change. However, mass transfer is accounted for in momentum control 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 terms depend not only on the j loss coefficients, but also on the magnitude of the phasic dynamic In SBLOCAs, the dynamic pressures tend toward small values pressure. 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 I smooth area change option and enter forward and reverse loss coefficients 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 nonreturn valves in the primary side cold legs and the secondary side steam lines. The current SBLOCA l model for the Maine Yankee plant does not use any motor valve components. The current LOCA model for the Vermont Yankee plant uses motor valve components to represent the following: a. Pump discharge 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 change 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 simulated 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 valve operating rates. 0.VI.27 Clarify how L1 and L2 are defined for the abrupt area change model on Page 58 of Reference 8. A.VI.27 The upstream length, L1, and downstream length, L2, used in the abrupt area change model are defined as follows: - - -
Expansion L1 = 0.0 L2=10.0DIAMV(L)dARAT(L) contraction L1 = 1.0 DIAMV(K) ARAT(K) L2=10.0DIAMV(L)dARAT(L) where: upstream volume number. K = downstream velutte number. L = DIAMV( ) hydraulic diameter of designated volume = partitioned area ratio in the designated volume ARAT( ) = associated with the abrupt area change junction. Note that we do not generally use the abrupt area change option as discussed in the answer to Question Q.VI.25. O.VI.28 Page 103 of Reference 8 describes modeling valves using flow coefficients 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 y user option to account for local form losses (K) through motor and - _~ servo valves. The latter two references also define the method presumably used to convert the flow coefficients to local energy losses (K) within the computer code. This information is not correct. RELAPS MOD 1 and RELAPSYA contain two options for modeling 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 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 Manuals indicate flow coefficient (C ) values are to be entered, the algorithm y 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. Chaus, A. Husain, J. M. 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) (VI.28-2) 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-Hydraulic Analysis, Volume II: User's Manual " Yankee Atomic Electric Company Report YAEC-1300P, Volume II (October 1982). (Proprietary) 0.VI.29 Page 49 of Reference 10 states: "In addition, the Two-Dimensional l Branching technique described in Section 3.2.7.2 of Appendix B is not recommended since it is known that this technique can create l j nonphysical recirculation through the adjacent junctions." Clarify how.,..
90 tees will be modeled and justify why nonphysical behavior will not result. A.VI.29 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 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) .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-Hydraulic Analysis, Volume I: Code Description," Yankee Atomic Electric Company Report YAEC-1300P, Volume I (October 1982). (Proprietary) 0.VI.30 Clarify how the 90 tee model has been assessed against data for single-Phase and two-phase flow conditions. A.VI.30 The 90 tee modeling technique described 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. TABLE VI.30-1 Test Facility and Nodalization Diagrams With Tees Figure Number Title Eggt 2.1-7 GE Level Swell Test Configuration........................ 16 2.1-8 RELAP5YA Model of GE Level Swell Test.................... 17 2.3-3 Test Assembly Schematic for Steady-State Tests........... 66 2.3-4 Jet Pump Blowdown Facility Isometric..................... 67 2.3-5 Jet Pump Test Vessel Model for Steady-State Tests........ 68 2.3-13 Test Loop Isometric for Blowdown Tests................... 76 2.3-14 Jet Pump Test Assembly Schematic for Test 1.............. 77 2.3-15 RELAP5YA Model for Jet Pump Blowdown - Test 1............ 78 2.3-21 Jet Pump Test Assembly Schematic for Test 2.............. 82 2.3-22 RELAP5YA Model for Jet Pump Blowdown - Test 2............ 83 5.1-1 Schematic of Thermal-Hydraulic Test Facility (TNTF)...... 177 5.1-3 RELAP5YA Model of TNTF Test Section...................... 179 5.2-1 Schematic of TLTA-5A Configuration....................... 234 5.2-2 RELAP5YA Nodalization for TLTA Test 6425/2............... 235 5.2-34 Schematic of TLTA-5C Configuration....................... 267 5.2-35 RELAP5YA Nodalization for TLTA Test 6432/1............... 268 5.2-45 RELAP5YA Nodalization for TLTA Test 6441/6............... 278 l l 5.3-1 LOFT System Configuration.........'....................... 308 5.3-2 RELAP5YA LOFT L3-6/L8-1 Nodalization..................... 309 I l \\ l 0.VI.31 Page 101 of Reference 8 discusses how gravity effects are modeled with 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: (d()"K+C f)K] 2K+ gg ffL L + DELPZ = C where AZ an G Z are half the elevation change in the upstream and g L downstream volumes, respectively. This term is then added to the SUMOLD variable. When horizontal stratified flow exists, then an additional term given by Equation (25) on Page 14 of Reference 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. l Loomis L. Schor, R. C. Harvey and R. Habert, "RELAP5YA - A Computer Program for Light-Water Reactor System Thermal-Hydraulic l Analysis, Volume I: Code Descript' ion," Yankee Atomic Electric ( Company Report YAEC-1300P, Volume I, October 1982. (Proprietary) l 0.VI.32 clarify if the statement is correct that "A gravity head is included in the branch volume V and V and results in a corresponding 3 4 separation effect due to buoyancy" (Page 101 of Reference 8). l l I
A.VI.32 The cited statement is perhaps too cryptic. Gravity head terms are generally accounted for in the momentum equations associated with junctions as stated in Answer A.VI.31. Each junction has a momentum cell that connects the adjoining halves of adjacent volumes. When an adjacent volume is vertical (for example V2 and V4 in Figure 33 of the cited reference), then its nonzero elevation change contributes a gravity term to the phasic sum momentum equation. When an adjacent volume is horizontal (for exemple VI and V3 in Figure 33), then its zero elevation change will not contribute a gravity term to the phasic sum momentum equation. However, when stratified flow exists in a 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. O.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 l local 3-D flow patterns determine the expansion and contraction losses (and momentum mixing losses discussed in the answer to the next question). These fluid mechanical energy losses are converted to l ' I l
/ relatively small thermal energy sources by dissipation. In one-dimensional codes such as RELAPSYA, these losses are approximated by the product of form loss coefficients (from standard references) and.- the-phasic kinetic energy terms. Although this approximation may not be' exact,itappearstobereasonablebaseduponthegenefallygood code assessment results contained in Reference VI.33-1. These assessment cases cover a very broad range of thermal-hydrauiic conditions. In particular, they include a PWR small break test (LOFT j L3-6/8-1) and a BWR small break test (TLTA 6432/1). Therefore, we, conclude that the error introduced by this approximation is small and acceptable. We disagree with the last sentence in the cited statement. The mass flow rate for all junctions is preserved by inclusion and solution of the conservation of mass equations within the limits of the code. Reference (VI.33-1) 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 Light-Water Reactor System Thermal-Hydraulic analysis, Volume III: Code Assessment," Yankee Atomic Electric Company Report YAEC-1300P, Volume III (October 1982). (Proprietary) 0.VI.34 i l Page 19 of Reference 13 also states: "Because each junction momentum I equation is calculated independently of the other junction flows connected to the branch volume (except for the reduced area) the momentum mixing terms are neglected." Clarify how neglecting these l momentum mixing terms affects SBLOCA results, and if they have been included in RELAPSYA 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 !
4 normal jet pump operating conditions. In this region, the very high velocity drive nozzle stream mixes with the much lower velocity suction nozzle stream. The momentum mixing terms have not been included for other RELAPSYA hydrodynamic components. The generally good code assessment results obtained over a broad range of thermal-hydraulic conditions indicate these terms are generally not significant except for jet pumps..-. - -,. -.
VII. CORE STEAM COOLING 0.VII.13 For the three steady-state film boiling tests, Page 167 of Reference 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 The precise location of the CHF is not presented in the Experimental Data Report (Reference VII.13-1) for the TNTF 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. T_AK. Q. 13-1 Inferred RELAPSYA Reported Quality Expt. CHF Loc. l Test No. at CHF CHR Loc. (ft) (ft) 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 8.2 - 8.8 6.0 - 7.5 Reference
-(VII.13-1) Yoder, G. L., et al., " Dispersed Flow Film Boiling in the Rod Bundle Geometry Steady-State Heat Transfer Data and Correlation Comparisons," ORNL/8822 (Preliminary Draft). I L e,----
0.VII.14 For the transient film boiling test, clarify why the RELAP5YA sheath temperature data at the elevations shown on pp. 183-184 of Reference 12 underpredicts the temperature during the period immediately after CHF 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 RELAPSYA curves apparently cross. A.VII.14 The underprediction by RELAPSYA of the temperature in the THTF Test 3.08.6C during the period immediately after CHF is due to the fact that CHF is predicted to occur by about 0.5 to 1 second later than the experimental data. The data curves and RELAPSYA curves are shown.nore clearly in the attached figures. l l l l l
1700 i t800 RELAP5YA ' s 1400 g = ,' 1300 DATA g IN ~ DATAI 5 I C 1100 i/ d 1 C 1000 (l' b m RELAP5YA g M 800 700 800 00 20 40 s.O 8.o 10 012 0 } O j8 18.0 20 0 22 0 24 0 2s.0 u.0 m 7 g Figure 5.1-9: FRS Inner Sheath Temperature at Level E(951n.) 1500 e ~ ~ ~ ~ DATA ~ 1300 g 8 RELAPSYA 5 1200 DATA i 1 00 f ~ / E 1000 / d l' m b 3 RELAP5YA h 800 o' 2 7M 800 8.0 80 lo.0 12 0 14 0 13 0 18 0 20 0 7gg gp g Figure 5.1-10: FRS Inner Sheath Temperature at Level F(1191n.). O
1500 RELAP5YA N t400 _ N-1300 g M 5 1200 DATA 7 5 1100 /4 RELAPSYA e, j 3 1000 DATA d / c 900 / o I / h m j Y I 4 700 J8 800 e i e e e i e e i i 00 20 4;0 80 00 10012014018018.020022.024028028.0300 TIM IN SEC Figure 5.1-11: FRS Inner Sheath Temperature at Level G(142 3/4 in.) y 3 *" u Wk A .3.00 6 E y 2 50 DATA 3 RELAP5YA e5 1 50 I 8> t.00 f y I d B .50 0.00 e i e e e 0.0 20 4.0 80 80 10.012014018018020022024028028.0300 TIM IN SEC Figure 5.1-12: Comparison of Measured And Calculated Outlet 3 Volumetric Flow
i 0.VII.15 Clarify what causes the shape of the outlet fluid temperature curve after 22 s as shown in Figure 5.1-13 of Reference 12. A.VII.15 The increase in the outlet fluid temperature, which occurs in both the test and the RELAP5YA predictions, is due to the fact that the pump was tripped at 20 seconds in the transient. The inlet flow rate fell quickly after the pump was shut off. At 21 seconds, the bundle power was reduced from 7.8 IW to approximately 3.4 MW over a 4-second time interval. The inlet flow rate dropped faster than the power. This resulted in the power flow mismatch that produced a fluid temperature excursion. i O.VII.16 4 J For the quasi-steady-state boil-off test, Figure 5.1-15 of Reference 12 shows the RELAPSYA void fraction is greater than the data above the 4-foot elevation. Page 169 of Reference 12 states: "This figure indicates that the interphase drag values calculated by RELAP5YA may have been somewhat high beyond the 4-foot elevation for this assessment i 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 of 380 seconds was run until the outlet steam flow matched the bundle inlet flow. The high drag values at the upper elevations entrained 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 RELAPSYA l 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 calculated void fractions are higher. !
Reference (VII.16-1) 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 III: Code Assessment," Yankee Atomic Electric Company Report YAEC-1300P, Volume III, October 1982. (Proprietary) 0.VII.17 Page 169 of Referer.ce 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 Figure VII.17-1 from Reference VII.17-1 indicates that the inferred mixture level for this test is at shout ,,8.6-foot elevation. The RELAP5YA calculated mixture level at a void fraction of 0.96 is 4.6 feet as shown in Figure 5.1-15 of Reference VII.17-2, Figure 5.1-16 shows that RELAP5YA 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-Phase Mixture Level Swell Under High Pressure, Low Heat Flux Conditions" (Final Report for THTF Tests 3.09.10I-N and 3.09.10AA-FF-DRAFT), Oak Ridge National Laboratory, 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 Analysis, Volume III: Code Assessment " Yankee Atomic Electric Company Report YAEC-1300P, Volume III, October 1982. (Proprietary) -100-l-
RI I D utssuReo b
- -Ch g.
O DRIFT FLU,X O v v 4 WILSON -"'"' ~~~'" ""9 X YElf~ ~~~~ 3- .O...cARDNER 1 ~. . ~ ~ " ' " " + O~ "O ~~~~~ _V. 955DNEN] . p,.. / e 0 O 'O / , '. -- $,g. g-v,- x w g, 7' o l o,, - D llf.x k MIXTURE LEVEL $ 8- ,' /. b l'!. t o* > d- ,/, - 8 l/. Y./, 3-3- 3
- I 8 lb8 14'Om jdes l$0m Figure VII.17-1 Experimental and Predicted Void Fraction Profiles; Test 3.09.101.
1 0.VII.18 Pages 169-170 of Reference 12 state: "The above disparity between the calculated and measured results might be attributed to the method used in modeling this test. The RELAPSYA calculation resulted in a very slow boil-off transient with constant inlet flow and bundle power. The 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 I temperatures of about 1400 F (maximum temperature imposed by safety Itmits). The differences between the power histories applied to the 4 bundle in the test (which are not reported), compared with the fixed power in the RELAP5YA 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. 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 scar /*yas taken. We simulated this test using the reported data that were obtained "after" the loop was stabilized." We do not know what effe:t the previous unrecorded history might have had on the RELAPSYA results. However, RELAP5YA 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-
~ ~ 0.VII.19 For the reflood tests discussed on Pages 170-172 of Reference 12: Clarify how quench level is measured. A.VII.19 The quench level is determined based on the response of the FRS sheath thet1aocouple (Reference VII.19-1). For the THTF tests, the quench o level is defined as the point during a 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 THTF Boil-Off and Reflood Tests 3.09.100-1-DRAFT), Oak Ridge National Laboratory, September 1981. l -103-l t
THTF SBLOCR II REFLOOD TEST 3.09.100 9N:======== ~0 1 9 E-PRECURSORY-a COOLING e 9 H i 0-y QUENCH _] m Ui s9 a g-c3 e e g- -g x A p mb d Z M 9 U $~ e-8 o -l 9 g-i -=============- 9 -10.0 -S.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 I TIME (S) i s Figure VII.19-1 Example of Precursory Cooling and Quenching Indicated I by FRS Sheath Thermocouple Response
0.VII.20 Clarify how "the measured void profile at time zero was used to calculate the initial liquid and vapor velocities" (Page 170 of 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 determined from these dynamic qualities and the measured void fraction at these locations. To initialize the test, the following calculations were performed: A. Dynamic quality in each node: Q *.z(i) h x(i) = (1) fg fg where: x(i) " dynamic equilibrium" quality in each node = elevation of the node center from the bottom of the z(i) = heated length (ft) latent heat of vaporization (8 ) h = g inlet subcooling ( ) h = aub liner heat generation rate (Btu /ft-sec)) l Q = i mass flow rate (1b/sec) i a = l B. The slip ratio in each node: i x(i) .1-c((i)Ef ( (2) l " 1 - x(i) a((i) g -tos-
l where: c((i) nodal void fraction (from void fraction distribution at = time "zero" from Reference VII) 3 f saturated fluid density (lb/ft ) = 3 saturated vapor density.(lb/ft ) g = C. Vapor velocity in each node: V(i)=Gx(i)/fe((i) (3) we replace x(i) in (3) by (i) S
- (
" y e((i)S + {g(1 -c((i)) where: S = Y /V (5) g to get: V(i)=G/[fO((i)+[g(1-c((i))/S] D. Liquid velocity: V (i) = V (1)/S (6) g F. Static quality: .0 x (i) = II -*(III} (7} 1.0+fg d(i) The values of V (i) V (i) x,(i), calculated above are used g for the initial conditions. -106-
i l 0.VII.21 i 4 Clarify Why the liquid mass inventory peak agrees almost exactly with the data at 17.5 s in Figure 5.1-18 of Reference 12 when the calculated collapsed liquid level at 17.5 s in Figure 5.1-19 of Reference 12 is significantly below the data. A.VII.21 i The test data and calculation have been re-examined to address this question. According to Reference VII.21-1, the local void profile in the heated bundle was determined from the nine differential pressure transducers. This experimental void profile is compared to that calculated by RELAP5YA in Figure 5.1-17 of Reference VII.21-2. These two void profiles agree reasonably well at 20 seconds. The experimental void profile was then used to calculate the collapsed liquid level at 20 seconds. This calculation yields a value of 124.9 inches which compares well with the RELAP5YA value of 125.0 inches. l However, it is substantially lower than the 134.3-inch collapsed liquid level at 20 seconds in the test report. We have not been able to j account for this discrepancy in the reported test data. Finally, the test report states that the liquid mass in the heated bundle was calculated by the following formula: liquid " f gg,,ICLL A where: saturated liquid density f = i bundle flow area
- Agio,
= i collapsed liquid level ZCLL = Their use of the saturated liquid density (47.5 lbm/f t at 1 20 seconds) does not account for the more dense subcooled liquid (50 to 55 lba/ft at 20 seconds) in the bundle below the two-phase region. However, their collapsed liquid levels appear to be too large. I Nevertheless, the product yields a liquid mass of 35.4 lbm versus 34.9 lba calculated by RELAP5YA at 20 seconds. -107-
References (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" (Final Report for THTF Boil-Off and Reflood Tests 3.09.100-I-DRAFT), Oak Ridge National Laboratory September 1981. (VII.21-2) 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 Analysis, Volume I: 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 exactly by RELAP5YA 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 determined from the void profile test data are about 5% to 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 tests are within 111% of the test data, and show the same distinct trends as the experimental data. 0.VII.28 Page 142 of Reference 11 states that for medium flow ranges the maximum of the high and low flow heat-transfer correlations is used. This -108-
applies for subcooled or saturated transition boiling, subcooled or saturated film boiling, saturated film boiling or forced convection of superheated vapor. Clarify why this procedure is conservative if the maximum of the two correlations is used. A.VII.28 The RELAP5YA heat transfer correlations are meant to be best-estimate correlations rather than provide conservatism. In licensing analyses, YAEC will use these correlations along with EM assumptions to ensure that heat transfer from the rods will be calculated conservatively. The logie used in the RELAPSYA 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 is because the high-flow correlations are not applicable at very low mass fluxos. The selection of the maximum of the two sats of correlations ensures that the correlations are used in their applicable ranges of flow rates. 0.VII.29 Clarify how the reflood model has been assessed for falling film from the top and for bottom quench. A.VII.29 The reflood model has been assessed by modeling two separate effect l tests for the bottom quench, and three integral tests which included l both falling film and bottom quench. The separate effects tests are the THTF tests 3.09.100 and 3.09.10Q. 1 i The integral tests are test TLTA 6425/2, TLTA 6441/6, and LOFT L8-1. l 0.VII.30 l l Clarify how closely the assessment conditions match the expected possible SBLOCA reactor conditions. -109-
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 Accidents (SBLOCA) conducted at Oak Ridge National Laboratory in the Thermal-Hydraulic Test Facility (TNTF). 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 TNTF tests. Average inlet flooding velocities in SBLOCA II ranged from a low of 2.33 in/see to a high of 4.82 in/sec. Initial system pressure ranged from 563 pois to 1092 psia. Linear power ranged from 0.304 W/ft to 0.659 W/f t). 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 SBLOCA in the plants modeled. All the integral tests used in the assessment of RELAP5YA used the quench model option at ECCS initiation. Since these were transient tests, the conditions in the test facilities varied with time. A 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 Linear Heat Bundle Mass Inlet Outlet Maximum FRS Test No. Pressure Rate (LHR) Flow Subcoolina Superheat Temperature (MPs (W/m (kg/s (K (OF)] (K (OF) [K (OF)) (psia)1 (W/f t)1 (ibm /s)) 3.09.100 3.88 (563) 2.03 (0.618) 0.156 (0.343) 74 (134) 198 (356) 1055 (1440) ~3.09.10P 4.28 (621) 1.10 (0.304) 0.075 (0.164) 65 (117) 209 (377) 1089 (1500) 3.09.10Q 3.95 (573) 1.02 (0.311) 0.078 (0.172) 66 (118) 168 (303) 1027 (1390) 3.09.10R 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) -110-
Reference (VII.30-1) 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 III: Cc.!e Assessment," Yankee Atomic Electric Company Report YAEC-1300P, Volume III, October 1982. (Proprietary) i 0.VII.31 e 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). The assessment of this model was clarified in A.VII.29. Reference (VII.31-1) Andersen, J. C. M., et al., "NORC00L - A Model of Analysis of a BWR Under LOCA Conditions," NORHAV-D-32. Research Establishment, R150, Denmark, December 1976. 1 0.VII.32 Appendix K Requirement I.D.3 includes the requirement that: "The effects on reflooding rate of the compressed 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 models. -111- }
a A.VII.32 Compressed gas in the accumulators will only be discharged for those LOCAs which are large in size and do not depend upon the steam generators for the removal of decay heat. Therefore, the degradation 4 of steam generator heat transfer as a result of U-tube blanketing is of d no concern. For these larger SBLOCAs, the following scenarios may be postulated: s. Cold Len Break l Bitrogen, due to its lighter density, would either stay in the cold les or upper portions of the downconer while finding its way to the break via the downconer annulus. If the break size is large enough, the nitrogen venting will not impact the core at all. However, for smaller breaks, the upper downcomer annulus may temporarily be pressurized which may depress the water level in the downconer and transfer some liquid to the core. This liquid transfer to the core will cause quenching of that portion of the j core which may not have already been quenched. i i b. Not Let Break l The nitrogen injected will collect into the upper downconer region. If leakage paths are available to the upper head or the I upper plenum, the nitrogen will migrate to these regions on its way l to the break. The core will bc unaffected during this period. i If no such leakage path is available, the nitrogen may leak via the r j suction leg with no impact on the core. Some of the nitrogen in this case may also bubble through the core via the downcomer. This scenario will experience a swellies of the core level and consequently quenching of those portions of the core which are i I still unquenched. l c. Suetion Let Break tr l This break will be covered under scenarios (a) and (b). l l l -112-l
i i I Thus, we conclude that the compressed gas in the accumulator which is discharged following accumulator depletion is expected to have no negative impact on the outcome of a BBLOCA and is not considered in our esteulations. i 0.VII.40 clarify what effect the following two assumptions given on Page 244 of Reference 10 have on the temperatures calculated: " Fuel rod deforination 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." 4 A.VII.40 This question requests clarification of the effect of two assumptions made in fuel behavior modeling. k 4 The first assumption is that " fuel rod deformation does not affect i radii used in heat conduction and convection except in the gap region." For the majority of SBLOCA cases, clad rupture does not i occur. Fuel rod deformation is small and the effect on temperatures is I truly negligible. For cases in which clad temperatures approach or exceed rupture temperature, clad swelling can be significant. However, f in this case, the increased fuel rod surface area available for heat I removal is conservatively neglected.. The second assumption is that "zircaloy-water reaction does not change material properties of cladding." This assumption is not expected to have much effect for the results of cases in which zirealoy-water reaction is small. However, the RgLAp5YA fuel behavior model is being f modified to include oxide conductivity and heat capacity when cladding is oxidised. t i -113-i i ....,-..___,.m,,
0.VII.41 Page 268 of Reference 10 states: "During small breaks, the fluid velocity in the core region is expected to be small. Various regions in the core are expected to communicate with each other and the fluid conditions at various radial locations are not expected to be significantly different at a given elevation. The entire core, therefore, may be represented by an average core for spall 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 loading schemes, the power generation will be the lowest at the core periphery and will monotonically increase toward the center 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 Yankee plant at Rowe and Maine Yankee, respectively. For the Yankee plant, Assembly 47 is the peak power assembly with a radial peaking factor of 1.3033. However, if one takes a region of 9 assemblies around Assembly 47 (Assemblies 36, 37, 38, 46, 47, 48, 56, 57 and 58) the average radial peaking is only 1.002. For the Maine Yankee plant, quarter assembly 65 is the peak power quarter assembly with a radial peaking factor of 1.32240. It -114-
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 assesblies 27, 28, 29, 30, 38, 39, 40, 41, 49, 50, 51, 52, 53, 54, 62, 63, 64, 65, 66, 67, 77, 78, 79, 80, 81, 82; 92, 93, 94, 95, 96 and 97) the average radial peaking is only 0.95. We conclude that the entire core, therefore, may be represented by an average core for small break calculations. -115-
I 2 3 4 .5620.7445.7767.5711 1.2139 1.2475 1.3503 1.2864 1.0982 1.1439 1.2357 1.1565 5 6 7 8 9 10 .7177 1.!!53 1.0769 1.2763 1.1258.6885 1.4377 1.5000 1.2782 1.5851 1.4461 1.3909 1.3069 1.3904 1.2020 1.4788 1.3419 1.2650 Il 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 1.2805 1.3974 1.3452 1.4293 1.1599 1.2047 1.3989 1.3017 19 20 21 22 23 24 25 26 27 28 .5v75 1.1340 1.0464 1.1776 1.1799 1.!!54 1.1760 1.1697 1.1245.5550 1.3062 1.4548 1.2544 1.5654 1.5006 1.4334 1.5753 1.4418 1.5039 1.2268 1.8801 1.3496 1.1952 1.4570 1.3998 1.3369 1.4642 1.3517 1.3950 1.1085 29 30 31 32 33 34 35 36 37 38 .80!! 1.2715 1.0428 1.1216 1.0350 1.0346 1.1983 1.1633 1.0999.7557 1.M90 1.5754 1.1994 1.4451 1.1896 1.1849 1.5019 1.5549 1.2994 1.2649 1.2536 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 7eJ3 1.0761 1.1410 1.1840 1.0393 1.03E'l 3.1219 1.0657 f.3033 7963 1.2535 1.2923 1.5394 1.5039 1.1974 1.1899 f.4343 1.2177 f 6153 1.3922 1.1502 1.2051 1.4293 1.4048 1.1534 1.1470 1.3384 1.1828 1.5073,1.2732 49 50 51 52 53 54 55 i 56 57 56 .5810 1 1198 1 1494 1.1768 1.1273 1.1913 ' t.1888.1.0657 1.1674 1.2229 1.5016 I.4316 1.5691 1.4467 1.5063II.58F3 1.2122 1.4959 1.4045 3319 1.1079 1.3919 1.3402 1.4581 1.3495 1.4061 )l.4743 1.166e 1.280 1.2010 4 59 40 61 62 63 64 65 66 .7369 1.1679 1.0484 1.0614 1.1744 1.1814 1.1916.7341 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 1.2980 1.1083 1.1578 7422 1.4200 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 IESCRIPf!!3i Mr.VALLE ASSDSLY .6198.8209.7885.5P26 3033 47 1 47 1 3018 1.2699 MI. <. (153 1 33762072 1.2829 1 1.4008 1944 1.1497 MI. 73 47 l Figure VII.41-1 Typical Assembly Radial Power Peaking Factor for The Yankee Plant ? -116-
I 2 3 4 . 90.'504.32??.35M ba4 47t$ 4m I'.N76.44 fit 45 0 5 6 7 t .39??.502' 4201.M9 .5744.6?M 7710.7t? (GE $tMIRV F MAIPts FLO ANMy,y PMI DATA .5474.6120.M26 774 : IEtulP71% mt.vAJ.T A N MEtt ? 10 11 12 13 14 "15 16 t' A MMEtv AW. 6! MI. hEl k22 b5224) .{696.St42 4127."f561.041.1011 1.f.72? 1.23601.27:1le. . 777. b. l!!! 65 .e?4) nr.I. Dea 1.47250 139 .25?9.376!.5773.6198.9227 1.0432 1.1435 1.1627 1.17 7 lt 19 20 21 .%n.7h3.90031.IIN 1.22h) I.2333 t.29}S3 1.25I' 24 e .3476 .510.6601.8251 1.0'!? 8.33241.4t421.49241.49?' t.4t?4 . 4**4 .43'?.t!!t 1.07921.2'!? 1.29431.422? 1.420* 1.2$2? 27 28 29 3' 31 32 33 24 35 3t 27 9,).}A991.0'0S1.!!5. 35' 7458.950C . 0644117% 1.2M4126171.;35 515 .'l21 :.{t t lje .3 .3H2,.h4 .29351.36541,3291 1.MJ 1. 04 . 240.5M.90471.064' .1901 1.3108 1.;?00 1.3444 1.20?! .2504 41p7 39 40 4: 42 43 44 45 46 47 41 .4M l.0343 .lm t.42321.4?*71.30421 '3040
- .3%2
. 2'90I.2?'e 1.27961.22121. 075 l.!9?? l.1476 :.{t?; 77i .6097.7850 8.2324 40N .53 3.7639 1.1817 .340? 1.3511 1.4092 1.2680 1.2425 . 24 3 . 1994 .2012 49 M 51 52 $3 54 55 to 57 58 !9 to it .22tl 414. e+.,7 1.137 l.1643 1.245% .614),1.2183 1.216I' 1.2757 1.3e*4 t.04721.N25 f.1601.M3e 1.0227.9685 1.2W 1.1145 1.1156 l.l102 1.4944 1.064 1.0103 1. 4579 . 2*A l .3240.5254 1.0945 1. 3 08 1.2148 1.3327 1.09571.10M !.43331.42061.0787.99021.3Est 42 43 64 65 64 47 (J 69 70 71 72 73 74 .3e*5.6331 '.l379 :.32261.2133 .245L 1.0103.98201.26901.24M.9727 94141.26M .5597 7?42 .3171 '.5112 1.3145 .35 0 :.054) 1.01541.4149 f.41971.0262.9M: 1.437' .5378.7650 .2516 l.43051.337 .3222 1.0400.9957 1.3633 1.3294 1.0054.9736 1.264; 75 74 77 79 79 80 t 82 83 84 e5 86 t7 8e t' .1490. jet .747 .27t(.114 1.0223'0360 1 27)5 ,.2137 2?40 1. 6* 1 .itt 1.2121Joi 1.3086 t.4*94 432* 1.0149 1.0216 1.1794 . 2347.{M'" 1.3088 .2663.9745 9780 1.0835
- .1334
. 3157 .254.4902 .9100 .it331.2142 . 2548 .244! 1.2445 1.4243 .3609.99201.00201.12M .1838.l m 90 91 9; 93 94 95 96 97 98 99 100 101 102 103 104 .2664. 5064 95191.1904 l.2449
- 2466
- .7.'05 l.. 2 1.30 % 1.2570.9514 94?9 1.0970 1 2412 1.2249 13851.13tl 2932.460' t.11631.4093
.2002. 3" 4 . 3130 . 127 t.49401.4165.9837 A .9806 1.2005 1. .3768.6359 1.0655 1.1416 '.3340 '.3238 :.2468 '. 468 1.4203 1.3438.9596.9562 1.1451 1 1925 1.1 9 5 105 106 107 loe 109 110 lit 112 ll2 114 115 116 tif lit 115 .4130.7367 l.04521.2M71.04!!
- 01111.3091 1.3028.9941.9751 1.24401.24901.lN31.11M l.100:
.4343.8856 I.2993 1.4243 1.1155 .0551 1.5003. 4934 1. 0372 1. 0156 1. 4063 1. 4124 1. 2100 1. l e93 1.1 % I .5m.86241.19081.3521 1.0966 .0411 1.4249 :.4199 1.0276.9939 1.3343 1.3??5 1.1617 1.1312 1.1391 120 121 122 123 124 IM 126 127 128 129 IM 13I 132 133 IN .5116.9007 1.190 l.2901 1.0429.982 . 2656. 2547 9596.9654 l.24.96 1.2724
- .1297
.1007 .075 .72591.0991 ).X41 .494 * .4900 1.0796 1.3113 :.409) 1.ll61 1.0160 : 4322 4150 9970 1.0003. 4417 1.4461 . 106 .1735 .1470 9782.9826 :.37471.36t9 :.I'635 l .1187 :.1140 1.1034.9%0 :.3604 .34?9 135 Ik 137 1 35 139 140 tot 141 143 144 145 146 147 148 to 7345 .'307 .2214
- .3161
- 2678.9737 9497 8.9414 1.0401
.h.t4 .M46 .5105. 4344 1.0142.9'20 1.I327.2683;.g.62..{254 8.1790NS 1.09"3 .0569.gJ 4476 .1444 9229 l .2s03 .2643 .4335 .3630.9915.9581 1.3313 :.3736 .1775 .86021.1251..CM4.93M 150 151 152 153 154 155 154 157 158 119 140 141 162 143 164 165 19 790.3933.90 % l.*29' !.2614 i.2076 1.3037 ;.2494.9n4 94791.24761.27171.120 l.1072 l.07801.0*A7.994 + .5749 1.1013 1.3644.3tte .304* 1.4944 4184 1.0209.97 % 1.4111 1.4456 1.2293.it92 .16 % 1.1341.9W .} '.5475 1.0434 1.3945 .3446. 242) 1.4206 .32901.0015.95321.33821.3t921.1683 l.1358 : 09861.0745.91!7 167 168 149 170 17 172 l?1 174 1 75 174 L77 L78 179 180 ft! It? 182 . 505.50301.0160 :.29341.251j 119P6
- .0226 97241.08291.0%21.im l.in3 1.09531.07001.0788
.0012
- .19
- 9
.h45.63301.2724 4935 8.3505 1.3002 .0068 1.0260 1 1718 1.1997 1.2092 1 2102 : 1792 1.16 % 1.1578..1643 '.3464 .3476.6120 1.14 % 42291.MH 1.2425 .07871.00521.12241.14431.16031.1632 .3561.0*901.0994,.8046 .2%4 1 94 195 186 187 19e 199 190 198 tt2 193 194 195 196 197 199 1 99 . 3298.6201 1.0443
- .2923 :.2121 1.1677 9684 9492
.3231 .1379
- 1. :124 1.
002 '.0565 .0556 .0774 .0"6 .2252 47t2.77101.2860 497' .30631.Me91.0102.98?9 .2N! .2406 Off I. 730 1399 1313 .1504 .l?60 4042 4400.M261.1625 420' .25v41.1994 9901.9734 .1983 :.1919 . 341. It3 .0790 .0775 .1012 :.3216 . ? s: 201 202.M.*F457.{k'4..{ht .2101 '.2t{2 f.{$29 . ),h' 0 Y
- 0 26f 77 20?
909 218 .0{. !.y l 213 {14 2t?
- tt
't' .$?,64?4 t l 1. 3 .0N 91tf.gN't . j 'M .234! 94: .It! .7f 37 .t9U .M.Itt4 .tf .t** 4h 4 *2 . 24? . P*! . lH e l .M 1.3923..!07)..}0i} 4!77 .34261.l?N .100.l?tt.ll? 9072.914' .27t? .?S?? 97t! w 'i .4V !.P e 451. 774l 3t'44 Figure VII.41-2 Typical Quarter Assembly Radial Power Peaking Factor For The Maine Yankee Plant -117-
VIII. METAL HgAT 0.VIII.1 Clarification and justification should be given for the metal slabs associated with all of the nodes used in all SBLOCA-related analyses. A.VIII.1 s Figures IV.2-1 and IV.2-2 represent the nodalization used in the pump trip studies performed 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 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. O.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 The pressurizar 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. -110-
II. BRgAK FLOW 0.II.5 Break flow depends on conditions of the f.uid reaching the break. If HPI or accumulator flow is not injected i.ito the cold leg with the break, then the resulting break flow will be very different in the model from the actual transient because of the difference in void fraction and temperature of the fluid at the break. Also, HPI and accutaulator 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 nodalization used in the pump trip studies performed for the Maine Yankee plant and the Yankee 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 ICC 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 location. Therefore, we model the ECC locations accurately and select the break at a conservative location. O.II.7 Clarify for stratified flow which critical flow model is applied above the interface and which is applied below it and how any switching process between the models is modeled. -I19-
. _ _ _ _ _ _ = A.II.7 When critical flow is determined to exist at a junction, then the code selects the appropriate critical flow model based upon the junction void fraction, The determination of the junction void fraction for stratified flow is discussed in the answer to question Q.VI.7. For e( < 0.05, the Subcooled Critical Flow Model is selected. For e( 1 0.05, either the Moody or the RELAP5YA Standard Two-Phase critical Flow Model is used depending upon the option selected by the If the junction void fraction lies within the transition region user. defined by 0.01 <t( < 0.1, then the underrelaxation scheme defined on Page 90, Appendix A of Reference II.7-1 is used. ]teference (11.7-1) 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 Light-Water Reactor System Thermal-Hydraulic Analysis, Volume I: Code Description," Yankee Atomic Electric Company Report YAEC-1300P, Volume I, October 1982. (Proprietary) 1 0.II.8 Deleted. Reference (II.8-1) Memo. Ed D. Throm, Reactor System Branch DSI to Brain W. Sheron, Chief, Reactor System Branch, DSI, Sumary of 6/22/84 YAEC-NRC Meeting on RELAP5YA PWR SBLOCA EM Analysis," July 17, 1984. 0.II.9 l Clarify why the calculation 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 Reference 12). -120-
c. A.II.9 4 I The original RELAP5YA case unde' predicted the pressure and escaped mass r 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 correc'ted, and the two cases (RELAPSYA Standard Critical Flow Model and RELAPSYA Moody Critical Flow Model) have been 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 IX.9-1. The maximum error in the vessel pressure measurement has been estimated by the Marviken staff to be 190 kPa (Page 50, Reference IX.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 seconds. Figure 2.2-13, Stvision 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 IX.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 slowly 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 data shows this occurred at 57 seconds in Test 10. The comparison -121-
shows the calculated break flow rates agree very well with this test data between 10 to 57 seconds. Therefore, we conclude that RELAP5YA calculates the two-phase critical flow rate and pressure history from 10 to 65 seconds very well for this test based upon these comparisons to test data with uncertainty bars. Reference (II.9-1) The Marviken Full Scale Critical Flow Tests: Summary Report, Joint Reactor Safety Experiments in the Marviken Power Station Sweden, NUREG/CR-2671, MXC-301, U.S. Nuclear Regulatory Commission, May 1982. 6 -122-
M i 1 M v e w R o l s F n Y o T Nl i N Oa t I I c c A Ti W i T Ct d R Ii e E Dr Pr C EC N R U Pd l r e I ig 1 a i t d .d \\ o A n h i i T Va 1 \\ A Et w e D RS o r l u F ss k e la 0 r 6 c P 4 i I ' t e i m r o C D d 0 r 0 1 a ) 5 d S t n s ( a e t T S E n A n e Y I k S 0 T i P 4 vrAL aE M R T 2 1 0 3 2 2 e rug i 0 F 2 0 1 D ] . g [ 0 0 1 g4 i gR og 8 l* 2s wo$&n. tuY
001 1 0 3 v w e o R l d F s Yo n Th Nl o Nt O a i I e I c t AM Ti 0 c T Ct 8 i Rc Ii d Ei D r e Ct EC r Na R P Ut Pd S r N l 1 a e t .d 0 d n Ao 7 o Tt V a ~ \\. N M Ai Et e DP RS t w ao Rl F wol t 0 l a T 3 i h 6 F c 4, i kt ai '.I er rC B d r 0 0 r i 5 1 a ) S d _ r t n ( .p sa et E 3 TS - _w 1r I nA T eY - r 0 kS = 4 i P vA r-rL aE 9 g 1 MR 3 0 1 3 2 T.p 2 e rug 0 i 2 F T. 1 0 a 1 A. /% L U 0 A 01 83c 88e g8e g 88 E* o-TM- & 3b' $n l;u8 l i
83c T DATA & UNCERTAINTY l-Vessel Inventory Method 88" REV. 1 PREDICTION Standard Critical Flow T e 8 eg = = uJ h. I L g a8 I T "- 8 W m un T n N cem g m,=LMR-2 _ Wah r r\\ 8 \\ n \\ % n o -10 0 10 20 30 40 50 60 '70 80 SO 100 TIrlE (S) Figure 2.2-13A Marviken Test 10 Break Flow Rate RELAPSYA Standard Critical Flow Model Predictions-Rev.1
00 1 Nl R O a E I c T Ti E Ct 5 M Ii 0 O Dr 2 T EC 0 M, I R 3 A S Pd T6 AN r A4 TE 1 a D2 AD .d M D nw Vao Etl RSF 0 1 8 4 v o" e R -s no 0 i 1 tc i Tl d il l e r jI h lI P U 0 l 6 e ~ do /# M w no ol iF 0 t ) 5 cl
- M S
aa ( rc Fi E t n di ir I oC T V 1 0 d 4 0 r 1 a d t n ~ sa et TS y ~ A 0 nA 3 eY
- y k5
~ A v iP vA _v rL ^_. aE q MR 0 4 2 1 2 2 er u 0 l 1 g )1 i F 0 0 1 ee 9* 3* n.* e5 vd y n?- s E tN$ S ,9'
001 wo 0 l 3 d F Yo Th Nl Nt Oa I e I c AM Ti 1 T Ct Rc I i 0 Ei Dr 8 v Ct EC e N a R R Ut Pd S r s 'ls. 1 a n t .d o Ao n i Tt V a t Ai Et 0 c DP RS I* i / de r P l le d 0 o 6 { 1 M / w /. ~ o sl sF a Ml 0 a ) 5 d c ,/ ) S ei ( pt ai E cr xC M E I d T 0 r 0 1 a 4 / d t n l sa et TS nA eY y 0 k5 3 iP / vA rL aE [. O MR 3 / 5 1 0 2 2 2 er / ug 0 i 1 F
- //
.:0 01 E 3 }a S o-5 mm C 'Me l
001 0 d w 9 Y o o Th Nl Nt OF I e I AM Tl T Ca Rc I c Ei Di Ct Et 08 N a Ri Ut P r S C 1 t y Ao .d Tt V o Ai E o DP RM 01 I e. K 1 t v 0 ae 1 i n 6 RR 4 ws on e \\ l o \\ Fi t sc si .r 0 ad ) 5 7 S M er ( kP p a e T E el 1 re r I Bd T o 0 0M r 4 1 y td so p eo TM nA T eY 0 k5 k a 3 iP s vA ^ rL L. MR aE %^P 6 ~ 0 1 i 2 2 ^ 2 ^^ -Fa. eru r g T 0 i 1 F L i, , r - ( l y o 0 l l A 01 83c g8e gS ss Oo 3u _ Tg N& $' =< br
' do hte w M o nl Yy oF Tr i N o tl I t c a An i c Te d i Rv et E n ri CI P r N C Ul 1 e y 1 As .d Ts v o e Ae e o t v DV nM ae RR 4 ws on l o Fi t Ty sc si M ad Me x r kP a k el \\ \\ re \\ Bd o 0M %) 1 \\ S y td ( so ,~ E eo TM M I nA T eY 0 k5 4 i P vA rL aE = - s MR A 0 6 3 1 A 2 ^ 2 e r u 0 g 2 t i F ,~ T I n o t 4 A i M h 0 01 83e 88e 88e 8S o$ ,o 8O o- . Th ~ w Nc 3O mgQ i c e C? !l L
} 001 0 3 y t w n o i nl a oF t i 0 r tl 8 e ca 1 c i c n di U et v ri e d P r R n C a 1 y 0 t vo \\': 7 es a .d rn a eo uo D R. M si st ec ri Pd I7 A~ ( 0 e 6 mr eP t \\ki sl -== \\ ye \\ Sd \\ o \\^ 0M 1 0 y ) 5 td N N S so ( eo A TM ^ E nA \\ 1 eY f k5 I A 0 T iP 4 vA s rL aE MR M \\ 7 1 0 N 2 3 2 k erug 0 i 2 F i 1 01 Ef-0 01 8M y-4 g. E l l 8" - <k-hWn. h?
oo t no i r t e c t i 5 e dl 0 m ee o 2 o rd s M t P o i M s 1 a6 n y t4 ae .d a2 tD vo DM a eo Dy RM os 1 X 4 I .v e R o l sn no oi it t c ci ad o re s F r P d iloe Vd o 0M AT! 1 o y s) td 1 Il l 3 2 so ( eo ~ TM EM nA / eY I k5 T 0 iP A-aE 4 vA rL M-MR 8 y 1 A 0 3 v 2 4 v 2 -v e r A_. _~ q ug i o F t / / f l 0 1 I = o = o 1 ,4
- ~
n* .4 c4 t?- aS[3 8 i
001 09 y td w no o ih nl at oF t e i rM tl 08 e ca 1 cc i c ni di Ut et v a ri e dt P r R nS C a 1 t y 0 ao .d 1 s t t vo n ai eo o DP RM st si ac Mi d d e A 0 er 6 pPacl E se Ed o 0M 7 1 0 y ) 5 td l S so ( eo e TM k" ~ E f~ l nA r eY I kS T iP 04 vA l) rL aE MR 9 1 0 3 2 2 1 eru g r 0 i 2 F 01 . 0 01 H S a S" S o- ~E-C ,{s t i{ !illl
0.11.10 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% uncertainty bars have been placed on the reported escaped mass tast 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 vapor 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. Tne 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 reported as test data at 75 seconds on the figure. Therefore, we conclude that RELAP5YA accurately predicted the total escaped mass from the system at 75 seconds, and that the reported test value is in error. ~133-l t. r *-
0.II.11 Clarify if the discontinuity in slip between the Moody model and the RELAPS MODI subcooled choking model could cause flow oscillations and potentially mass 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 RELAP5YA included a) tabular values of the Moody critical mass flux as a function of the upstream (donor cell) pressure and enthalpy, and b) Moody's theoretically derived slip ratio given by: x,.(v,/v,)-(f,er,)2' This method led to computational problems and large mass errors in the donor cell. Subsequently, the method of implementing the Moody Critical Flow Model in RELAPSYA was modified. The tabular values of Moody critical mass flux as a function of upstream pressure and enthalpy were retained. However, Moody's theoretically derived slip ratio was replaced by a 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 calculatSd by the subcooled choking model which are near unity. The mass flow rates calculated by the Moody Two-Phase Critical Flow Model are still conservatively high compared to those predicted by the RELAPSYA Standard Two-Phase Critical Flow Model. 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 IX.11-1) and the Maine Yankee Pump Trip Study (Reference IX.11-2) are shown in the attached Figures 11.11-1 and -134-
11.11-2, respectively. These results show that the transition between the Subcooled choking Model and the Moody Two-Phase Critical Flow Model is 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 (Mgreater than 0.05). The mass error in each run was small. Therefore, we conclude that the transition between the Subcooled Critical Flow and the Moody Two-Phase critical Flow models yields well behaved solutions. 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, Yankee Atomic Electric Company, July 1984. (I1.11-2) L. Schor, S. Haq, A. Husain, J. Ghaus, " Justification of Reactor Coolant Pump Operation During Small Break LOCA Transients for Maine Yankee," YAEC-1423, Yankee Atomic Electric Company, April 1984. l I I -135-l l l l l
on a m e n nu EM I q T i S / (EM E V_ I l M A e I A ME ET n K H R m em R A 1 3 BK )2 E_L 1 BtE E_L_I V C2 )2 B C S E M_D_M
- e_i, S
E1 C_D W S1 A AD S ( C T (4 4 M_I E D A E r t EJF Ms u N Mu I u R_ INK s A e s R_ N. O 1 e sTs 1 u I C - _1 I F Y_5 f, I T E s A R A A 1 R 1R w F / D s IOY es s a a C se M E -ACO s L a B I I g g 3 S 3 3 3 3 S ENR l z e 9B tna l P ee kn s a n Y e m r f 1 b ,f 1 1 r X es E I se M IT er S K5 l YV u j I AM d i g P 0r K i 0EI i ER T E A F 2B E )3 2g 2 L R 1 ) - E_I._ V C1 IL C1 SD E SDB E _ (S 1 A W eS1 A C_D_M T C_ dN n(4 4 D A f m Ee IO_K; N MUF eEs W_ N_ T IMn Ma u u O E I T r. LI Te I s t iF I A F IOC I WR WA TB TR F D IO e V e m s e S 3 I 3 g 3 i I g g 3 k2R , lm x 2 eB nc'
e. = e ..s e e. s ee w o. = eeas e. s. = c e e s n s a a c s e. r e = g e e s al I s n. e u a r e s e =M ,.e er sf A a
- e. t.
e-e e 3 e. i, ta aK s 'pC 3, at c4 a u sc a f sM t e o. e. e e i e m i s r e = e i-5 e. e. s e a e s e. e e s = se aC e. M s E ee -A e. C e O .N .u
- g
. l' .t .e .g I; .3 .C e k-E4 .bo' J' L B s: I I S ee kna Y en ia e. M eno g e. 2 se s 1 e 1 a e.e m X a e. s I m s e.. y ee. e re e. se s r s e e s i u r a a m se e-n j g ri e. e i a w i ee e a r F e
- t s
a r sc N e. e iw a e s a. re e r o s e n s
- e. s e.
e so r aa e s s m c i oi e a ses
- esa 3
a ss c nr e aa e n. as es sa e. me se s. +s r .as o, a a s .t ~ e e e e v. se s o. ssu s a ae e e v s 0 g e. ,s e F r a s. J e m o sa i e r e a e e e.es e e
- ese e.se e
s l.. e e. e l. e". Ig C; i*
- ! 3.!
3 .gI I 8 .4 g-e?* l, i e s g 33 .d 13s e, 3I
0.II.12 Clarify why using the Moody model only after the void fraction exceeds 0.05 meets the Appendix K Requirement I.C.1.B 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.l.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 calcultte 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 discharging fluid that is to be calculated prior to using the Moody critical Flow model. Therefore we have selected the small void fraction value of 0.05 as our criterion to begin using the Moody model. This small void fraction was chosen nr. inly 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 subcooled critical flow rates are slightly higher than those calculated by the Moody Critical Flow Model near the void fraction value of 0.05. Thus, we comply with Appendix K requirement I.C.1.6. 0.II.14 Clarify if break-size sensitivity studies will be done for SBLOCA analyses. A.II.14 Yes, break-size sensitivity studies will be performed and submitted on a plant-specific basis. -138-
0.II.18 Page 89 of Reference 8 states: " Equation (245) can be derived from Equation (153) by neglecting the third C term in D'and setting C=0 (stratified) on the right side of Equation (153) and C=oc(homogeneous) on the left side." Clarify how the opposite extremes of the virtual mass coefficient can be justified for use in the same equation. A.TI.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=33 (homogeneous) on the right side of Equation (149). 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 results 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, LOFT, TLTA). The following discussion may explain why. Figure 15 in Appendix A of Reference II.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 critical flow conditions is generally an order of magnitude less than -139-
the mass mean Nach number due to the relatively small phasic velocity j differences that develop. Thus, the second term on the left side of Equation (155) is generally much smaller than the first term, i.e.: M +DMr " il (155) y V D(V -V) E [ > D M, %_ My 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 assuming 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 the small dash lines. These results shoe the break flow rate ans 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 came 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 (II.18-1) 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) -140-
001 0 3 sno i y t t c n i 0 d i n 8 e a o e t t s r r i a P e c C c i l n d y e U e t d r i o d P v } M i n i ,\\ W \\ i w a 1 t \\ eo i h I rl a s \\ uF t v n ,\\e1 s a e e sl D R S \\ ea rc ( 0, Pi 5 6 t 4 O N ei mr f \\ oC D d 0 r g% 1 a d 0 t n ) 5 hs S sa et s TS ( A nA E eY s n k5 i P I N 0 T vA 4 rL aE N MR s 1 8 0 1 3 9 X I e rug i 02 F ot 3f-0 ot 8H g~ 3 gS gM g v"- n<k w$$Mn. iE7
00 1 03 sno y t td i no c ih n i at o e 0 d t e i s 8 e rM t a r e c C P cc i ni d y l Ut e t e a r i d N d t P v o nS i M O a 1 t e I t w t i N \\ ao ao s t t v n \\. \\ Rl ai e .e \\ F DP R S w g ol \\ l a f N F c 0 i 6 kt T 2 1 ai er g rC g 0 r g B d 1 a d 0 t n ) 5 sa S - r et ( p TS E nA r-u M eY I k5 i P T 0 vA 4 rL aE -= MR d-L. 2 8 0 1 3 %a X I er ug i 02 F A E ot i d o n 0 a ot 83e ,84 88v gg 8g g4 og o-n T!$. ww g mOJL y8 i a 1un i l
001 O d S o s h n yt o t e i nM t i n c ay o e i t r t s 0 d ro i a 8 e et c C r cn i P ne d y U v e t l n r i e dI P v d al 1 t; \\ n i o M g. 0 e e i 1 t w as s t s v n N ao ae e e DV R S, N \\ Rl F \\ w s ol l a \\ F c i 06 kt T1 1 ai er rC E B d 0 r 1 a 'b d 0 t n ) d 5 sa E S et TS ( .6 E nA M eY k5 I iP T vA 0 rL t. M aE ~ MR = ~ 3 =m 8 0 1 3 ~ M-X I ^ erug i 0 F 2 M a r -C 0 u 1 .i .%~ f i I 0 2 0 1 83e 88e oac 8g 8a ,N o- _ To:~ J I s3" $e c ue i_0' l ,1I llt
oo g no e t s i a r c C e i t d y m 5 e e t 0 m r i 2 o P v M t i i 1 t s i s s n a6 an t4 t e v n o a2 aD e e i 0 t DM D J S., 8 c T i d X 4 e r P ~ le 0 d 1 o M Tf w t II f ' g nl o j ) l' oF I.' w i 0 tl I 6 ca I l Ig ac f)+- ri Ft i dr i C o 0 V,d ) 5 r y2 S 0a ( 1d n E ta st M eS I T T A 0 b-kP 4 nY eS k i A v vL rE o aR A H u - v A 03 v A 4 v -v 8 1 a q x e I 0 e 2 ru g / i F / K o t I = 0 = 0 1
- o 9o
.,o 9-Ta t?- t 2 ~t asa i l 1 i
001 03 sno i y t td c no i ih n d at o e t e i e 0 r rM t s 8 P e c a cc i C l ni d e t Ut e y d x' a r t o d t P v M nS i a 1 t w t i O o I ao s sl / tl v n sF ai e e a (- DP R S Ml a ~ s)/ d c ei ~ pt 0 ai 4 ~ 6 c, r )// sC E d E 0 r f. 1 a d t n sa j-[, 0) et 5 TS S ( nK f* eY E k5 l iP r vA I rL T 0 aE 4 MR / .-(l ~ 5 8 1 y 0 X 3 I er ,I u g i F 02 ot .:0 o t R g" G S a o-E u)"@ s i Z v.s
0.II.19 Justify that break-flow models bound all SBLOCA cases expected. A.II.19 RELAPSYA 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 location 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 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 identifled in Table II.19-1 that is attached. Therefore, we f believe these break flow models bound all LWR LOCA cases that could reasonably be postulated, including small break LOCAs. -146-
TABLE II.19-1 Assessment of Break Flow Models Test Facility Critical Flow Regime and Number Liquid Two-Phase Steam Marviken Test 10 R R, M RM GE Level Swell R R Test 1004-3 TLTA Test 6425/R2 R R, M R, M TLTA Test 6426/R1 R R R TLTA Test 6432 R R R LOFT Test L3-6 R R Note: R = RELAP5YA Standard Critical Flow Model M = Moody Two-Phase Critical Flow Model -147-
I. ADDITIONAL QUESTIONS THAT ARE CONCERNED WITH SEVERAL AREAS 0.I.2 During depressurization, the noding detail affects.the pressure history af ter the pressurizer empties. 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 smaller nodes. Additional nodal sensitivity studies are required to confirm the adequacy of the hot region nodal detail to properly account for flashing during depressurization in a small 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 This question refers to the noding schemes used by combustion Engineering Inc. for their LOCA analyses (Reference I.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, Amendment 1-P, " Review of Small Break Transients in CE NSSS " Pages 3.3-29 and 3.3-30," July 1979. -148-
0.I.3 Appendix K. Requirement II.2 states: "For each computer program, solution convergence shall be demonstrated by studies of system modeling or noding and calculational time steps (Page 274 of Reference 10). Clarify what noding sensitivity studies have been performed. A.I.3 It has been pointed out in A.I.2 that we intend to use a detailed nodalization of the plant in our LOCA analyses. This detail has been I utilized to represent the 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. O.I.4 -u' Clarify what time step sensitivity studies have been performed. A.I.4 4 l Time step sensitivity studies will be parformed on a plant-specific model and documented in submittalo. i l -149- .}}