ML20063F260

Forwards SSAR Markups Addressing Open Item F19.2.3.3.8.3-1 Re Containment Sump Design

ML20063F260
Person / Time
Site:	05200001
Issue date:	02/07/1994
From:	Fox J GENERAL ELECTRIC CO.
To:	Poslusny C Office of Nuclear Reactor Regulation
References
NUDOCS 9402140211
Download: ML20063F260 (56)

Text

{{#Wiki_filter:.- m ^' f 1 February 7.1994 Docket No. STN 52-001 Chet Posiusny, Senior Project Manager Standardization Project Directorate Associate Directorate for Advanced Reactors and License Renewal Office of Nuclear Reactor Regulation

Subject:

Submittal Supporting Accelerated ABWR Schedule - Response to Open Item F19.2.3.3.8.3 -1

Dear Chet:

Enclosed is a SSAR markup addressing the subject. open item pertaining to containment sump design. Please provide a copy of this transmittal to John Monninger. Sincerely,. Jack Fox Advanced Reactor Programs cc: Alan Beard (GE) Norman Fletcher (DOE) Joe Quirk (GE) Doug Mcdonald (GE) Jack Duncan (GE) 0 JNF94-015 n U' .gf

• l' looon 9;

9402140211 94o2o7 p JL PDR ADOCK 05200001 .N A PDR 2 r r. u. u s _ m - m _ - _ -- -_ _

_:e m m ::,.:

re:

y ,, m 1 t Contzinment Sump Shield Design . 9ED was modified to provide the information that the Staff requested and to clarify some details of the shield design. The following addresses cach concern. Opening Statement: The analysis was revised to take credit for flooding the lower drywell. 1 The probability that the lower drywell will not be flooded given debris relocation is approximately 10-4. Not crediting flooding, placed undue-requirements on the shield design and unnecessarily complicated the long-term debris solidification analysis. Additionally, the text was clarifled to state that the EDS and FDS shield roofs would contain grovisions that allow water to enter the sumps in the event that the Lower drywell is flooded. 1. Related Experimental and Analytical Work Subsection 19ED.6, "Related Experimental and Analytical Work", was added to address the references sent by the Staff. s 2. Material Properties f The analysis of channellength in Subsection 19ED.4 was modified to E account for three debris scenarios which cover the expected range of melt phenomena. Two of the scenarios were taken from the Mark Iliner failure analysis contained in NUREG/CR-5423 and represent oxidic and metallic melts. The third scenario represents the original MAAP-ABWR-case run to bound the debris ejection rate from a failed vessel. The eutectic scenario previously considered by the analysis was removed because the channels are expected to be plugged before enough core-concrete interaction can occur that an cutectic would be formed. - i Sensitivities to rnaterial properties and superheat are considered in new Subsection 19ED.5. t S. Initial Velocity of Debris The design of the FDS shield was clarified to establish that the channels would be in the shield walls which do not face the center of the lower drywell. Therefore, debris cannot enter the channel directly and the entrance velocity can be determined using the hydrostatic head calculation contained in Subsection 19ED.4.3. l 4a. Debris Tunneling Subsections 19ED.S and 19ED.5.2 were modified and Subsection 19ED.5.3 was created to establish that the EDS and FDS shields extend to the floor of the sumps to prevent debris tunneling. .i t ri m 4 M-?M-:!!7-------- c;-;t-9; ?^: 1 AM ?:t

-wts:-wae r v.wcc 15 - 't 4b. Shield Roofs The long-term analysis contained in Subsection 19ED.5.1 was modified to credit flooding of the lower drywell. When water is present, minor amounts of debris on the roof do not affect long-term solidification in the channels. 5. . Thickness of EDS Shield Subsection 19ED.5.3 was added to address the thickness of the EDS shield walls and the shield wall of the FDS without channels. I a .) l 'I l F r M: 4^?-;;=-:f87-------- 50-ci ;; 4 A :.- y;

1F.'67EE ' ~ l 23A6100 R:v. 2 ACWR sensers setery Aristysis soport i 19ED Corium Shield 19ED.1 issue During a hypothetical severe accident in the ABWR, molten core debris may be present on the lower dryvell (LD) floor. The EPRI ALWR Requirements Document specifies a 2 floor area of atleast 0.02 m /24 to promote debris coolability. This has been interpreted in the ABWR design as a requirement for an unrestricted LD floor area of 2 79 m The ABWR has two drain sumps in the periphery of the LD floor which could collect core debris during a severe accident ifingression is not prevented. Ifingression occurs, a debris bed will form in the sump which has the potential to be deeper than the bed on the LD floor. Debris coolability becomes more uncertain as the depth of a debris bed increases. The two drain sumps have different design objectives. One, the floor drain sump, is designed to collect any water which falls on the LD floor. The other, the equipment drain sump, collects water leaking from valves and piping. Both sumps have pumps and instrumentation which allow the plant operators to determine water leakage rates from various sources. Plant shutdown is required when leakage rate limits are exceeded for a certain amount of time. A more complete discussion on the water collection system can be found in Subsection 5.2.5. 19ED.2 "repend Design 'De fC F(( (EPS) A protective layer ofpf ctory bricks--a corium shield-will be built around the sumps to prevent corium hgression. The shield for the equipment drain sump /will be solid except for the inle)t and outlet piping which will go through its roof. Th floor drain sump'will be similar except that it must have channels at floor level to allow len7a cb water which falls onto the LD floor to flow into the sump. The heigh 6f the channels will be chosen so that any molten debris which reaches the inlet will freeze before it exit!d*and spills into the sump. The width and number of the channels will be chosen so that the required water flow rate during norrnal reactor operation is achievable. A l sketch of the41oordaiNnp shield is shown in Figure 19ED-1. G 8d d u'a/// j oC,rle FDS shidI, The walls of the equipmenNn4 ump shieldCLsolid shiel'd]> nly have 15Ee71$c7d"""# enough to withstand ablation, if any is expected to occur, fo{ the chosen wall material. The walls of the flooWdmp shield fehanneled chieMTmust be thick enough that molten debris flowing through the char is has sufficient residence time to ensure cwa /ff /9 claao e/f debris solidification. Both shields extend above the LD floor to an elevation greater than the expected maximurn height of the core debris bed. Thus, no significant amount of debris will collect on the shield roofs. The solid +hield w be placed diredy esop4f-the-lM-1 f*& y, . } Q gf;gQ yl 19ED 1 Conum Stueld - Amendment 32 ( 'be tw :Aa a +m e ce p c. eve,t r ciebm com ":;n -fr ---. 1- & -u a.- -7y nv.;

y 22A6100 Rev. 2 N_DWR - sunMsatoryAnalysisRepon ri; ic rT i9 E D, 2-C y ...~. s. / floor. The channeled shield will have refractory bricks embedded into the LD floor beneath the shield to prevent core-concrete interaction imohing the molten debris in The analyses presented in Subsections 19ED.4 and 19ED.5 provide a basis for sinng the codunt-3 Meld FM sei/e /d' wa /// "<t 4 proposed. design-orthc4oor-drairaump 4 e c unn- !1 r.; e 7:g,,,y-o F c .rd. e /d' wa ar > :c Aou edunefy 19ED.3 Success Criteria 4er-Papeecdoesigne//- Fe'en red /o M/ e c r/c-W'.e.a r. JorTAc the proposed $' sign-to-be-consideredourr-uM i rai:e t wafft s tmust satisfy the following e requirements: (1) Melting Point of Shield Material Above Initial Contact Temperature [Oth$ita mehing temperatur 'is The shield wall material sha c greater than the interface temperature between the debris and the shield wall. S ue cifyiny ag' ia a f c-de.rd re tc/ wart m areNe t sa cMer t (2) Channel l.ength ~ NF#*## F'V S The length of the channels in the$hield must be long enough to ensure that a plug forms in the channel before debris spills into the sump. The freezing process i3 expected to take on the order of seconds or less to complete. A channai t en y c-4 c r c.f />t e c e rt te c-;15'es 'cA;rre?u @.m (3) Shield Heigh .,, Above Lower Drywell Flonr a,ng co f~e,e ven c-debe. o nom aettec.t w, on -a a e-A e r A t e i _s The shield height above 1, e o difivus shal osen to ensurelong term debris solidificatiorfThe freezing process will be complete during the time frame when the shield walls are behaving as semi-infinite solids. In-addition;-the4hield must be 611 criou@ iv y-rat-debritirom-accumulating-on-the-roofcFthe-shield, A delp A t of A'l ntererJ saw1hs rA;.1 rer. wire m en r; (4) Shield Depthhelow Lower Dgwell Floor The-shield <lepth-below the lowcr drywell-floor-shall-be-chosen-torensure long e~, tenrrdebrissolidification. Tie wa/// c # CA E N ad d' EM i j,w ({)(5) i S kle ic s ex ten s' co the flo ors c5 'c d e. Sun}os m ,I Water-Flow-Rate P&"' *cun't el.*nf c ? de b t s /ac cie rmw'. f,; c w g,. g 2. M The total flow area of the shield channeighall be great enough to allow water l flow rates stated in the Technical Specifications without causing excessive J water pool fortnation in the lower dqvell, in =!ve i'VS f,1. : f.l (6) Chemical Resistance of Shield Walls The wall material chosen for the corium shields must have good chemical resistance to siliceous slags and reducing emironments. Resistance can be 19W2 Corium Shield-Amendment 32 7 p;. 4:3 7-e t e n_ 5.;4 3: n. 7: r

NEB 08:'94.~05:59AM G E NUCLEAR BLDG J' 5.~7/22- ~ ~ ~ ~ l Insert 19ED.2 B' th shields have provisions in their roofs to allow water to flow into the o sumps when the lower drywell is flooded. The provisions are located next l to the pedestal wall so that the debris which relocates from the vessel can not directly enter the sumps due to geometrical constraints. Additionally, the provision for the roof of the EDS shield will not affect the normal water collection capabilities of the EDS. i To prevent the debris which falls on the lower drywell floor from directly entering the FDS shield, the channels in the FDS shield are in the walls which face away from the center of the lower drywell. The FDS shield wall which faces the center of the lower drywell is solid and.does not contain any channels. ) .i I l 4 1 i .i i 1 DBM Insert 19ED.2 p.1/1 February 7,1994 l n:n : n-m -us -------- c u t-;;

a n e n-

FEB 03 '94 06:59At1 G E fVLEAR BLDG J P.8/22 23A6100 Riv. 2 ABWR. stedordsserrA=trsisR:p:st determined to a first degree by comparing the Gibb's free energy of the oxides which make up the shield wall and the oxides present in core debris. $P'e cP ' / 7 vss af tje SA,. /d watt pat er gt gag ifrc ata. e (gj da/sn6l N(fy((7) Seismic Adequacy d'3 Iffaire eien. N hc3 The seismic adequacy of the corium shields will be determined in the detailed w[k-c g/ev design phase. Adequacy should be easily met because the shields are at the $p c,4 r' lowest point in the containment. Missile generation is not an issue because the (C,, r l F-.>. shields are not near any vital equipment. .w Subsection 19ED.6 contains calculations which demonstrate satisfactionY requirements (1) through (4) for a chosen channel height of1 cm. F 19ED.4 Analysia d OhieldheezingAbiliQC Aao fr o / t.ef,gy{ Heat transfer and phase change anagses gpresented in this subsection to determine 3 the ffasibilityef-a-ehann cicd 3R 3 t, f : 8nTlIf88ehis ingression into the 41ooe A FN E' p' n[ grain sump. Two4mgrag[eghysidered4issy freeze front analysis wa performed for early *imes ( econds or4esskto determine the time required to form a s gryhI '{ plug. The4gdrm-ability-of4 plug 4o remainseLi Ocwouined wiu3 medr5 tate 'Hi 'p' d -analysis. T-he Freese p roin g asiafi.re'.1 if e valmreef O r three tb/ sic h d. d e/c p c tad egected & dehris S c.enartof 19ED.4.1 Assumptions Cehr/> guramcerX The major assumptions invoked in the analyhs and their bases foller (1) Molten debris enters the channel with negligible superheat. h Molten debris interacts with structural material (steel, concrete, etc.) and the _~c r 8 ' p;- \\ lower dr>well environment as it passes from the vessel, contacts the LD floor % accco4* g M 'hr[,M# $ g \\, superheat,and4arwesuh4n-euteede-form 4 d c2 N and spreads to the shield. This interaction depletes the molten debris of any g,3 c f L ngldugmuymmf-L M 3 # i" g c,d Wore-debns-which4ra.nsndergene44ttle4rrterauion4npproximately 500 R i i g f en /N',, digmficantimerardon ii ie muucic iloor reduces die-debrirmeld"6 - p;// M 7,1 ja<f mperatur, to mpproximately 1700 E C' J ' i, e 6' c ^j ) 1 \\ 4 6f{/>g,y During the freezing process, the temperature profile of the solidified debris Q

p

\\ rapidly obtains its steady state value. This assumption introduces little inaccuracy because: rferinal pnd'act/Wey rP -ede-(a) ihe heat enduccon.coeflieient4rt-the-solidified debris is signifundy c larger than that of the shield materialp% meszr d'e4c/s scense,'cr' see M rec ren 19 57h V. % (b) the depth of the solidifi4 debris is considerably % 6an the4eight nf > the-shield + oir (y / c/nj rdar, & e vSerma/ Sm conr w r or vie <rebets 1,, cle e.f a a e t is sma t c oiwared r 19ED b f.j;*'J* s^==we n.ap nm.Ma C' ?I/14 V'%1ct. 7sh o'y Corium Shield-Amendment 32 e.Q.e,oW'cet eryJum kek by comgar e pe e ege. rw ? f ws va g,h; 9 *i :' _ _.h" 4'#M Y 'f f "Y 'h % f'*#'Y'** I '#7### 73:

FEB 03'**.4 07r00AM G E NUCLEAR BLDG J P.9/22 23A6100 Rev.1 ABWR sunderssarmAutrsis noen /' (3) Heat transfer within the char:nel and shielo is one dimensiona]. The height of each channel is much less than its length. The heat transfer in the shield material is low enough that any heat transferred from debris contacting the shield wall outside of the channel does not affect the temperature along the channel untillong after a plug has formed. Any heat + transfer to the shield material between adjacent channels enhances the debris freezing process. (4) The shield wall acts as a semi-infinite slab with an initial temperature of 350 K - 1 during the initial freezing process, s Aave been retected ArcA = lac it is The properties of shield aan.e4 tao 4,e a poor conductor of heat. The penetration depth during the short duration of the freezing process is on the order of ten millimeters. The smallincreases in LD temperature prior to the 'l presence of core debris does not significantly alter the shield temperature from its value during normal plant operation. i (5) Core debris is not expected to enter the LD until about two hours after ~ accident initiation at which time the decay heat level is approximately 1 % of rated power. Core debris will not enter the lower drywell before about two hours for any credible severe accident (Subsection 19E.2.2). (6) The decay heat generation in the debris is negligible compued to the rate of latent heat generation during the freezing process. This assumption was verified during the analysis. (7) The thermal conductivity and thermal diffushity of debris in solid and liquid phases are the same. (8) The contact resistance between the bricks was assumed to be negligible. i Contact resistance will be controlled in the detailed design by varying the-thickness of the bricks or by using a high temperature binder between the bricks. Thicker bricks tend to minimize overall contact resistance by reducing the number of contact points. Some contact resistance may be acceptable in the final design if the composite thermal conducthity is high enough that the shields provide short and long-term debris solidification. i 19EO4 Corium Shield-Amendment 31 ry:n 4:9 gt-;e r. e c ;- p ; ; e 77;3

FEB 08 '94 07:00AM G E ffJCLEAR BLDG J-P.10/22 22A6100 Rn. 3 o ADWR Srsadard Safety Analysis Report (9) The corium shields were assumed to be structurally stable. Structumi stability is only an issue during the initial onslaught of debris into the lower drywell. After debris comes into contact with the shields, a crust will form and it will tend to grow in time. Crust formation eliminates buoyancy forces and will hold the individual bricks into place. 19ED.4.2 Initial Freezing of Molten Debris in Channel If the floor drain sump shield fulfills its design objective, a debris plug will fonn in the channel before molten corium has a chance to traverse the channel and reach theg 6mp. Molten debns enters the channel at a significantly elevated temperature (,1/00 K d-?)#tY2500 K) compared to the shield wall (~ 350 K).The walls absorb heat from the debris /'8 g/ because of the large temperature difference. Since the debris contains negligible y'

1. g superheat, any heat loss by the debris results in freezing. Freeze fronts start at the

(# channel walls and move towarti the center of the channel. The4eading-edgMthed-. JreezeJront-willatayat the-melting 4cmperature-of-thedebri/The freezing process is symmetric about the centerline of the channel because the same amount of heat is transferred through each wall while they are behaving as semi-infmite slabs. The channel walls behave as semi-infinite slabs during the freezing process because the heat - I conduction rate through the wall material is low compared to the release rate oflatent heat. A sketch of the freezing process is shown in Figure 19ED 2. (1) Freezing Time The temperature profile in the p[tusMReference 19ED i9D6 reaches its steady state shape, is: gLf / 2i T, - Tr. m x T, + T, m t T,(x) = 1-x + -+ (19ED-1) 2k 2 ' L, 2 g where: e temperature within the crust ded"If T(x) ver-Naf fre ~ x -crust coordinate measured from the crust centerline l heat density of the crun de4ru j q Le half thickness of the crust = kr thermal conductivity of debris T, interface temperature between the wall and debris = Corium Shield-Amendment 33 19 W 8 i i T F 0ff ?-?:f-lE!'-------+

. - C E - 9 4

D:

A?- Fl*

• "'TEB 08 '94. 07:01r41 G E tilCLEAR-BLDG J.

P.11/22"

~ 22A6100Mov.1 -ACWR standardsafetyAnaksisReport Tr,m melting temperature ofdebris = The energy balance at the freeze front is: dT-q[h = -k (19ED-2) _ g dx,, _t;. where: the latent heat flux = The latent heat flux is: dx q"h " dt Pcm lh ND h where: crust thickness x = e t time density of debris Pcm = h debris latent heat of fusion lh = Combining these two equations, evaluating the temperature gradient and rearranging yields: dx* "k x-d (T -T)-d2 (19ED-4) - = dt p h 'm cm lh.*c This is a non-linear, non-homogeneous, firstorder differential. equation. Before effort is expended to solve it, the relative magnitudes of the tenns containing the crust thickness will be determined to see if either one dominates. The initial interface temperature between the. wall of the channel and the debris can be approximated by assuming both the debris and the shield wall behave as semi-infinite solids. The resulting temperature will be somewhat less than the actual interface temperature because the freezing process will force the crust to stay closer to its initial temperature than it would ifit were a 19CD4 Corium Sbleld-Amendment 31 F L :.: '4Did:t-15Ei-------. 00-0 ?-94 10 : 1 A:: T1.

m-- m 2sAs100 nov. t .ABWR sondedsareryAnso sesnevert r i f i 7 semi-infinite solid body only experiencing conduction. The contact temperature between the debris and the channel wall (Reference 19ED-2), assuming semi-infinite bodies,is: $ 3 T,mj@pc)co, + T jppc), r i 1 T= (19ED-5) )(kpc),+ j(kpc), where: specifc heat c = -- L debris material properties h r[e ' aMre rla / r y'e R - l cm (x', de wat .w cl me deM:, ) wall material properties w = t[& coo m/oed,*nd 192D-2, reye:rhely J Talg /c r /fM ~/ Using-the<lebrispropertiesfotmdirrTabic 19E.2 l'i, Imy nrPammm,,fv, -< __. Su.mplo.on Analyst, aud upresenta6: wa!! r ug.k. ruuud h. --C 9-Table-19ED:t, the interface temperature is estimated to be4390-L. f be k WEf ASO i 80 C. The debris energy generation density can be found by assumm]g a ecay heat - ao level and a total amount of corium. The density is: kh cm E q== (19ED 6) S :: c I where: Qe - decayheatlevel mem = total mass of corium,235 Mg, Assuming the decay heat level is approximately 1% of rated power yields: 6 8 i = 1.5 x 10 MW / m E t i The two terms inside the brackets in Equation 19ED-4 can now be evaluated. I For a channel height of I cm (x.ma - 0.5 cm) and a debris melting c temperature of1700 K, these values are: i k (T -T,) = 1.86 x 10' W / m*

r. m c

Corium Shield-Amendment 31 ' i 19ED-7 1 7 7 :r' s c e - G M - : E * * -------- C -M-94 l M* .II2' e

~ FEB 08 '94J07:01AM G E tOCLE R BLDG J P.13/22 ~. nAssoon, s ? j Argyg? A w my9,y, 7. s. x* 3 t q2 = 3.8 x 10 W / m Therefore, the term containing the temperature difference across the crust is 3 ouch larger than the one ccntaining the heat generation rate. The? i temperature profile in the channel system ignoring energy generation in the debris is shown in Figure 19ED-2. Equation 19ED-4 can be simplified to: dr* k (T -T,) (19ED-7) - = dt pc,h Xlh e Sohing this equation with the initial condition that x (t = 0) = 0, reveals: c i (2k (T,-T,) t f f, 9 ,',N Pcm g h i This equation can be rearranged to determine the time required to freeze debris in a channel of height H. The freezing time is: o 2 oEcm lh t (19ED-9) freeze = k (T, m-T ) 8 f f s (2) Interface Temperature. T, The interface temperature between the debris and the channel wall can be j t determined by equating the heat flux from the crust to that which the crust ; can absorb. The heat flux from the crust is: I l dT* l q~rus = -kg (19ED-10) l x=x/2 i which evaluates to: i l 1 qx, k-i f 9"rus: = - + - (Tf, m - T,). (19ED-11) j C f As shown predously, the temperature-difference term dominates the energy-- generation term in this equation for small channel heights. Herefore, the.. crust heat flux can be simplifled to: 19EO 8. Corium Chield-Amendment 31 S y y [1] 43i-@*{=}(?"~*-~--.- ))~73-94

EiI ~ AIU '

I35'

FEB 09 '94 07:01A!1 G E TOCLEAR BLDG J P.14/22 22A6100 Rev.1 ACWR standantsotaryAnstystsnapart k q" rust (T, - T,) (19ED-12) = g, C 1r serting the expression fcr x in Equation 19ED 8 and rearranging yields: e kpc,hth ( f, m "~ :) q,Cru81 (19ED-13) t The heat flux (Reference 19ED-3) absorbed by the channel wall can be approximated by that which a semi-infinite solid body can absorb. This flux is: k (T -T ) g q ", = (19ED-14) jna,t where: 1 thermal diffusivity of the wall material ot = w Combining Equation 19ED-13 and Equation 19ED-14 produces a relationship governing the interface temperature. It is: T* - TI 'nk p**h "ih

• t (19ED-15)

= 2 T - T, 2k f, m t w Sohing this equation for T, using the quadratic formula yields: -(c,- 2T ) i (c -2T )2 - 4 (T - c,T,) 2 i o g f, T, = (19ED-16) where: the square of the right hand side of Equation 19ED-15 c o Negative solutions of this equation are physically impossible. For a Tr,m of 1700 K and a T of 330 K, the interface temperature is 1560 K. Similarly, the i interface temperature is 2180 K for Trm = 2500 K and Tg = 330 K. Since this temperature is higher than the value for two semi-infinite solid bodies coming into contact, the dominance of the temperature difference term in Equations 19ED-4 and 19ED-11 should be reverified. The Corwm Shioid-Amendment 31 19ED 9 TEc"

5-;;?-ic? ---~~---

0:-ct-24 A:' r:

FEB 08('94T 07:02AM G E NUCLEAR BLDG J1 - P.15'22 : 2sAet00 nov.1 '. ' ^ 'ACWW . Senadenf SatoryAnalysis Report '. a m.. t heat-generation and temperature-difference terms for a interface temperature of 1560 K,md channel half height of 0.5 centimeters are: 8 2 (T 4 ) = 8.4 x 10 W / m r, m 3 e x i 8 2 (g * = 3.8 x 10 W / m 2 Even though the dominance is not as great as before, the temperature-difference term is still significantly greater than the hiat generation tertn and the assumptions made previously are still valid. i 19ED.4.3 Required ChannelI.ength toinsure Freezing ,,r m A -O'edre r/07 Thegropagation rate of the freeze front was determined inJSubsection 19ED.4.2.This allowed

• determination of the time to cm.p?ctly &c :: de dadfin a channel of-j.[specified height. A siruple approximation of the channellength, required to provide this residence time, is the product of the initial molten debris velocity and the freezing time. This ap' oximation would predict shield dimensions considerably larger than actually regt ed. A more realistic channel length can be obtained by considering the

~ i reduction in channel flow area as debris freezes. In the remainder of this subsection, the followmg parameters will be determined: debris velocity at channel entrance, a a chann'el area decrease resulting from debris freezing, s. average channel debris velocity, and a h' the required channellength to insure plug formation at the channel entrance l a before corium ingression into the sump. i 4 (1) Debris Velocityat Channel Entrance The possibility exists that molten debris will not even enter the channel a6er. it has come into contact with the shield wall. Debris which is spreading across _- Li the lower drywell floor will have at least a thin crust formed on its leading edge. j If the flow energy of the advancing debris front is not great enough to break j this crust and overcome surface tension on the length scale of the channel I height, debris will not enter the channel. Unfortunately, the physics of crust .] formation is not currently understood well enough to support this argument without a great deal of uncertainty. ~' i .i. I 19ED 10 Corium Shield-Amendment 31 mu e n : m- "---- du-a :r;in m m-

FEB 08 '94 07:02AM G E HUCLEAR BLDG J-P.16/22 23A6100Cev.1 'A&WR^ SaunderdSofetyAnalysis Report SI's c. t ybe: c/wtitte I/ Ct r c in r Ale /cl cua/ff cuficD DrC ACT fa c<'n f q k e c/d fccaree- ),y 4 ceit t er o5 c ,The entrance velocity will be governed by the height of codum outside of the ' channel. Assuming that the debris spreads uniformly across the lower dqweU floor, the height of debris can be obtained by integrating the volumetdc expulsion rate of corium from the vessel dhided by the floor area of the lower i g drvwell. Tonservative over prediction of debris depth can be obtained by N T/d cygv/, ion multiplying the maximum expulsion rate by time and dividing by area. The rgef Ae upper bound of the expulsion rate was shown in Subsection 19EE.6.2.2 to b t;*b rC8 L Scenar',Of trancd f The velocity i"nhe channelbutw ce redsctiewdee :: dchh n ccdag can w /// 6e c c M /et~d be conservatively over predicted by ignonng fdctional effecu. This velocity is: I f,, g,,j, fec cien O 1l/ v*(t) =./2gAz(t) { (19ED-17) / g0 gf re 6 ~> c.c er Fric'Cic/M I W d' u 6 Cena Ay,,, where: g /,o a (c/ cc c W " If. Y l 'C 18 Y velocity at the entrance of the channel) b e a .f rcd ~ t ulc-c aI e co(4 v, 7 = b c ca efe-T:bc. Ot n1e W P g ft 6 'bx$- gravitational acceleration constant - Vi f Cc fe,FY A g = j O MI ge b r',' i de debeu y if Az height of debris in the lower dowell g,rff ;g g g = d Fd#6 C4 //# Expanding debds height yields: / Q S i t l-rce z d. ; 1 !2gm*** t {" v (t) = (19ED-18) Ecm ^1d where d tn ye, = 4naximunrejection rate of corium from a failed vessel j aid floor area of the lower drywell = (2) Channel Area Decrease Resulting From Debds Freezing y Sino-the-entrancea ainchv h =<e"m-L'e remaiuonstan he mass flow rate of corium in the channel decreases in time due to the area reduction resulting from debris freezing. A conceptual picture of this area reduction process is - _j shown in Figure 19ED-3. Conservationefeassteqt9rae '6t them!!c'emte j i 40(odum*ntedng-the<hann elper-unit-le n gthn< ens tan tahroug hout-th e,_JL >2- --<hannet The mass flow rate at the entrance of the channel and at the location downstream where the debris front hasjust anived is: j Cotium Shield - Arnersdment 31 19ED 11 FE?!( (C?-9;!-!{&'-------- QM-94 1 P ; !. A!l 7I#

gg wp gyagggi g g tqtcLg0c;f puGG y-P.17/22 1 23AG100 Mov,1 ' ABWR - s standantsareryAnnirstsa ut n ,( rn (t) = p,,v,(t)H (t) = pcm'o(I)N + 1r,- . (19ED-19) 3 i o where: .r thj time vaging mass flow rate per unit width at the entrance of the channel H3 time varying entrance flow height of the channel = time varying velocity at the downstream location in the v o channel where molten debris hasjust arrived H unobstructed height of the channel = o h ena.rs Cree s/ny rat e e s debr.'s per t y,. = u d t~ This equation requires that: vJh4 /d v4e c/ta o o e / v,(t) /5;c y, (t) = H* H (t) - (19ED-20) - 3 t"s *ll o r-The entrance flow heightis: (.. k. y H (t) = H,-2x,(t) (19ED 21) g Inserting the relationship for x, found in Equation 19ED 8 into this expression yields: '8k (T - T') t H (t) = H - J (19ED-22) 3 h N Pem th The product of this equation and the width of the shield channel describes the reduction of channel inlet flow area with time. (3) Average Channel Debris Velocity The velocity of the leading edge o molten debris in the channel can be I obtained by combining Equations 19ED-20 and 19ED-22. It is: i 8k (Tr, m - T,) t" ) v (t) = v (t) 13 r ( (19ED-23) M H p,hlh -d M CT YG ~' 2ao be /b \\ y [_ Q p sm & a 19ED 12 Corium Shield-Amendment 31 \\ p y, :.. 4:.p;n -ie5 -------- c:-C?-;4

.

' At:

F:~

van 9rn 23A6100 Rev.1 VABWR aamrantsehnyretysisneport i / /- The average velocity of debris between the entrance of the channel and the { leading edge of molten corium is: t t t J-1, [*,(t) d t k v(t) = S \\ 8 g o (19ED-24) { \\ Evaluating this integral yields: f \\ L f3 v(t) da,d a b H, t - (19ED-25) gy, I where: i -t 4 E ves a = - Pcm^ld [ 5 2k (Tf,m - T,) g b = - E h cm th This is thhverhelocity of the molten debris into the shleid channel. En se rY / FED Y<7 p Required Channel Length to Insure Freezing - 1 The channel length, required to ensure a plug forms at the channel entrance before debris spills into the surnps,Is: j i wAere MT5ere D ,1 Lfreeze = 7(tfreeze) tfreeze g g a A Y, RED j n t ^ b

• o o,

O m e. j g t ,,o freeze _ H I** I n In seri7s ftEA % % tPED. g 5 gng 77ep, y t I; } 19ED.5 Long-Term Ability.of-Debr!s to 8 main 4elid c g,fr,hp,# n/e.///e/d' //,/o/4-Initial debris solidification was considered in Subsection 19ED.4. The requirements for ' keeping the debris in the channel frozen for an extended period of time (at least 24 hours) will be determined in this subsection. The height of the uppershield w above Corium Shield-Amendment 31 19ED 13 l . r np 4cs-9:?-itei------- 024?-94 10'!! AM ?!*

FEB 09.'94 -07:03AM G E-NUCLEAR BLDG J P.19/22 ~ q. g Insert 19ED.4.3 (4) Mass of Debris Frozen in Channel The time varying mass of debris freezing in the channel per unit width can be approximated by /bfr

• x hfe s'

wh et-e : A': c rort recv'onal et rea-o# Nt e zen debru. The cross sectional area can be related to the crust thickness by modifying Equation 19ED 8 to account for the variable residency time of the debris at various vertical locations in the channel. This process yields: A' 2 b.}lI/z - 7/n ' & Wk 'rd

• L = fy c/ from che cAa/mel en Wancd ro r4e

/ec4:6'nc e e o;b el6 de 4 <-u fe.Ac veeti caI eocrelimce /neaturec/ V by = Prom rfe en west ce of de cAanned Evaluating this integral yields: A'=fb,Mc)C#. Combining this result with Equations 19ED-25 and 19ED-26 yields: 1 i F (2;) = h a W ~ #

• C 2 be ? 60 g o

No hqo W = c .Wo i i + '),* pts 7 i DBM Insert 19ED.4.3

p. VI February 7,1994 u m m n-:e r--------

c:- n->;

e n;

c tauos y W. p 22 Insert 19ED.4.4 19ED.4.4 Channel Lengths for Different Melt Scenarios The analysis to determine the channellength required to ensure that a plug forms in the channel prior to debris entedng the sump is contained in Subsection 19ED.4.3. The analysis will be executed in this subsection for three different melt scenarios which cover the range of expected core melt conditions. The scenarios differ in debris composition, debris material properties, initial temperature of the debris and the ejection rate of debris from the failed vessel. The impact of debris superheat will be considered in a sensitivity study contained in Subsection 19ED.4.5.2. Scenarios I and II were taken from NUREG/CR-5423 (Reference 19ED-7).- These scenarios represent predominately oxicid and metalk melts, respectively. The Peach Bottom Atomic Power Station was used for specifica in the NUREG/CR-5423 calculations. However, the similarities between Peach Bottom and the ABWR in core composition and vessel geometry allow the NUREG/CR-5423 core melt parameters to be applied in this analysis. Scenario Iis based on MAAP calculations which predict that there is a significant amount of molten debris available for relocation at the time of vessel failure. The debris release in this scenario is consistent with the core composition, corresponding to approximately 30% by weight of zirconium. Further,20% of the available zirconium was assumed to be oxidized. i Scenario II is based on BWRSAR calculations which predict that debris ) which initially relocates to the lower plenum of the vessel is quenched. Subsequent water depletion results in the remelting of the debris and local- ) failure of the vessel. Since the metallic constituents of the debris remelt-i before the oxicid constituents, the initial debris pour from the vessel is primarily metallic. For both scenarios, only the initial molten material relocation is important in determining the required channel length because the channels will plug in a relatively sbort period of time (less than 10 seconds). Only the maximum initial debris relocation rates were utilized so that the calculated channel lengths would be conservatively over-estimated. Subsequent mohen material releases from the vessel will go into filling the lower drywell with debds and have no bearing on plug formation. Thus, the long term debris relocation parameters discussed in NUREG/CR-5423 are of no consequence to this analysis. The third scenario considered, Scenario M, represents the MAAP-ABVG analysis that was used to predict a bounding debris ejection rate from a failed vessel, see Subsection 19EB.6.2.2. MAAP-ABWR predicts that the mass flow rate from the vessel jumps to approximately 1000 kg/see at DBM Insert 19ED.4.4 p.1/4 February 7,1994 n m e m - : m----- mu-u 1:: o e n:

FEB 09 '94 07:04AM G E' NUCLEAR BLDG J P.21/22 vessel failure and then increases to 6000 kg/see in eight seconds, see Figure 19EB-9. The debris release is essentially complete (flow rate 0 kg/sec)in ten reconds. In order to avcid determining the debris entrance velocity into the channel for this complicated flow rate profile, the maximum debris relocation rate was assumed to prevail throughout the plug formation process. This assumption will lead to a conservative over-estims' ion of the required channel.ength, s Another melt scenario commonly considered in calculations involving ex-vessel core debris is the formation of eutectics as a r'esult of core-concrete interactions. However, consideration of eutetics is not required in the analysis of plug formation. In order for the debris properties to be changed significantly, core-concrete interactions must increase the debris mass by at least 10% The time required for this to happen is longer than the time required for plug formation for a quicidy spreading debris front. Alternatively, if the debris front is spreading slowly eriough to allow a significant amount of core-concrete interaction, the leading edge of the debris will have a thick crust and have a height greater than the channel height. Thus, debris will not even enter the channel. The parameters describing each of the three scenarios considered is contained in Table 19ED-2. The debris material properties were determined using the constituent material properties contained in Table 19ED-3. Inserting the scenario parameters contained in Table 19ED-2 into the analysis contained in Subsection 19ED.4.3 results in the channel lengths contained in Table 19ED-4. All of the required channel lengths are less than 0.5 meters. i l 1 DBM Insert 19ED.4.4

p. 2/4 February 7,1994 1, g : c e _ m _ : u _ _ _._.

e_a_n m :3 y r;.

FEB 08 '94 - 07:05ft1 G E TO'LEfR BLDG J P.22/22 Table 19ED 2 Scenario Parameters Scenario I* Scenario II* Scenario M** a Flow Rate (m / min) 4 0.7 42 Debris Temperature (K) 2850 1800 2500 Composition (w/o): UO2 70 4-61 ZrO2 10 0 3 Zr 20 47 24 Fe 0 35 ~ Cr 0 8 ~ Ni 0 6 carbon steel *** 12 i ~ ~ Material Properties: density (kg/m3) 8900 7300 8500 specific heat (J/kgK) 960 705 800 thermal conductivity (W/mK) 6 26 12 heat of fusion (MJ/kg) 0.31 0.26 0.28 Scenarios I and II correspond to the Scenarios I and II defined in NUREG/CR 5423 (Reference 19ED-7).' Scenario M corresponds to the MAAP-ABWR case run to maximize debris ejection rate, see Subsection 19EB.6.2.2.

• • • MAAP uses carbon steel instead ofits constituents Fe, Cr and Ni, DBM Insert 19ED.4.4

p. 3/4 February 7,1994 r u n u s-nt-: m --------

e:-a-a

. g r;;

3 cs ca'st 07: S t1 G E T E LcAR R.DG J ~' ~- P.2/34 j Table 19ED 3 Constituent Material Properties

• thermal density specific heat conductivity heat of fusion (kg/m3)

(J/kgK) (W/mK) (MJ/kg) i UO2 '10100 1000 3.3 0.27 ZrO2-5600 991 3 0.71 j Zr 6500 780 18 0.25 Fe 7800 570 35 0.27 Cr 7200 781 35 0.26 l Ni 8900 609 35 0.30 I carbon steel ** 8000 795 35 0.25 Material properties from NUREG/CR-5423 (Reference 19ED-7),-MAAP 3 User's Manual (Reference 19ED-8) and the CRC Handbook (Reference-l 19ED-9) i

• • MAAP uses carbon steelinstead ofits constituents Fe, Cr and Ni.

1 I i Table 19ED-4 i' Results of Channel Length Calculation Scenario I Scenario II Scenario M Interface Temperature (K) 1880 1580 1900 Freeze Time (sec) 5.7 4.2 4.2 ~ Channel Velocity (m/sec) 0.03 .0.01 0.07 Required Channel Length (m) 0.15 0.04 0.30-DBM Insert 19ED.4.4

p. 4/4 February 7,1994 nmm_s:t_:n________.

x_

TfB 08 '94 07:16#ti G E NUCLEAR BLDG J P.3/34 Insert 19ED.4.5 19ED.4.5 Sensitivity to Melt Parameters The three scenarios considered in Subsection 19ED.4.4 were chosen to represent a wide range of possible debris characteristics. The calculation of required channel length depends on the debris flow rate, the debris s temperature, the channel height and several material properties of the debris and wall material. This subsection will evaluate the sensitivity of - the calculation to these parameters. The final part of this subsection will evaluate the impact of debris superheat. 19ED.4.5.1 Material Properties The channel length required to ensure freezing is dependent on both debris and channel wall material properties. The relevant debris material properties are density, latent heat of fusion and thermal conductivity. For the channel wall, the relevant properties are thermal conductivity and thermal diffusivity. The sensitivity of the channel length calculation to these material propertiem will be determined in this subsection. Additionally, the sensitivity to debris temperature and debris flow rate from the vessel will also be determined. The sensitivity of the channellength calculation to these parameters will be estimated by varying each parameter, except debris temperature, by i.0%. Debris temperature will be varied by i100K. The wide variations iu material properties of the three base scenarios take into account deviations outside of this range and combinations of deviations. The results of varying these parameters are contained in Table 19ED-5. The following discussion will describe the effect on channel length due to increasing each parameter. The reverse of the effect described holds true for decreasing the parameter. Increasing debris density, latent heat of fusion or flow rate increases the amount of energy that must be transferred to the channel walls before a plug forms, and, as a result, increases the required channel leng+h. 1 i Increasing the debris thermal conductivity increases the rate at which the debris can transfer its energy, and decreases the required channel length. Increasing the debris temperature increases the debris to channel wall temperature difference and, as a result, the rate of heat loss by the debris. Since this analysis assumes that the debris enters the channel with negligible superheat, increasing the debris temperature reduces the required channel length. l The impact on channel length due to variations in wall material properties - is a direct result on the change in the interface temperature between the-debris and the channel wall. Increasing the wall thermal conductivity decreases the interface temperature and results in a shorter channel DBM Insert 19ED.4.5 p.1/8 February 7,1994 n w m - m - m -------- a-m-n a : a s #.u rr

MCL4 LG TJ W7:16AM G E TOCLEAR BLDG J ' P.4/34 1 length. Conversely, increasing the wall thermal diffusivity increases the y interface temperature and results in a longer channel length. Decreasing the interface temperature increases the temperature difference driving heat flow from the debris, and results in more rapid freezing. Alternatively, increasing the interface temperature decreases the interface temperature, and results in a longer freezing length. The interface temperature is decreased by increasing wall conductivity and s increased by increasing wall thermal diffusivity. Thus, increasing the wall thermal conductivity decreaseo the required channel length, while the opposite is true for increasing wall thermal diffusivity. The three parameters which have the largest impact on channel length are debris density, debris latent heat of fusion and wall thermal conductivity. The impact of the debris properties are not surprising because they directly impact the amount of energy that must be removed from the debris in order for plug formation to occur. The impact of decreasing thermal conductivity does not have any significant impact in reality because a lower bound of alumina thermal conductivity was used in the base analysis. Thus, a ten-percent decrease in the wall thermal conductivity is physically unrealistic. As can be seen in Table 19ED-5, none of the parameter variations resulted in required channel lengths greater than 0.5 meters. Therefore, a channel length of 0.5 meters supplies enough margin to account for uncertainties in material properties. 19ED.4.5.2 Impact of Superheat The channellength analysis contained in Subsection 19ED.4.3 assumes that the debris enters the channel with negligible superheat. The analysis contained in this subsection considers the effects of superheat on the corium freezing process. First, the credible amount of superheat for each melt scenario is discussed. Then, the change in energy due to including superheat will be calculated. Finally, the impact on freezing time will be determined. Freezing time can be used to correlate the length of channel required-the greater the time required for freezing, the longer the channel. Since freezing time dictates channel length for a given corium velocity and superheat is not expected to affect debris velocity, the analysis will conclude at freezing time and not determine channel length. According to NUREG/CR-5423 (Reference 19ED-7), the initial superheat of - Scenario I is expected to be negligible "because of several concurrent reasons: (a) convective heat transfer to boundaries--typically less than 100 *C can be sustained at decay heat levels, (b) continuous melting and incorporation of boundary material, and (c) heat losses to water and control rod guides during the relocation through the lower plenum." The upper bound of reasonable superheat that the debris could obtain at the time of vessel failure was specified to be 125 C. The most probable superheat was limited to less than 50 C. Due to similarities in Scenario I DBM Insert 19ED.4.5

p. 2/8 February 7,1994 m m - m - w ----.---

FEB 08 '94 07:17AM.G E NUCLEAR BLDG J P.5/34 and Scenario M (quick penetration failure after delayed core plate failure and mostly oxidic rnelt), the superheat of the two cases can be assumed to be the same. Scenario IIis expected in NUREG/CR-5423 to have more superheat than Scenario I because the molten mass does not have as much opportunity to lose hast to the lower head. The upper bound of reasonable superheat that the Scenario Il debris could obtain at the time of vessel failure was specified to be 150 'C. The most probable superheat was limited to less than 100 *C. A5ter the debris exits the vessel in any of the scenarios, the debris will lose some ofits heat before it reaches the corium shields. The heat loss will be by radiation and convection to the lower drywell environment.and structures. Additionally the debris pool on the lower drywell floor will conduct heat to the lower drywell floor. These heat losses will remove some, if not all, of the debris superheat. This analysis conservatively assumes that the debris enters the shield wall channels with the same superheat it had when exiting the vessel. 19ED.4.5.2.1 Change in Energy The fractional change in debris energy due to adding superheat can be obtained by comparing the energy content of the debris with superheat to the energy content without superheat. This yields (insert equation A for 19ED.4.5) Table 19ED-6 shows the percent change in energy which results from assuming different amounts of superheat for the three scenarios. The energy comparison indicates that a significant amount of superheat (greater than 25 C) could impact the channel length. However, for superheats on the order of a few degrees, there is negligible impact. 19ED.4.5.2.2 Channel Length with Superheat A simplified analysis of the channellength required to ensure plug formation before debris ingression into the sumps will be presented in this subsection. The superheat temperature will be assumed to decrease the fusion point of the melt, not increasing the temperature of the melt exiting the vessel. This will conservatively result in longer required channel lengths, as indicated by Subsection 19ED.4.5.1. Balancing the energy required to be removed from the debris to freeze and the energy to be absorbed by the shield wall, and assuming that the debris in the channel behaves as a lumped mass, yields e'f t freeze ~ ~4w DBM Insert 19ED.4.5

p. 3/8 February 7,1994 m.. w. w w. -- -

u. n wa w

V1o# BW "94 Nfi1@AM G E NUCLEAR BLDG J P. 6/M where ef = energy required to be removed from the debris in order for freezing to occur EP (h n +c 6T) i p = 2 t Q.= time averaged heat flux into either the upper or lower channel wall (semi-infinite body), see Equation 19ED-14 for the instantaneous heat flux 2k (T -T;) w ~ Vnct,t Combining all the portions of thi6 squation yields -2 H pcm(hth +c aT)Vnot, o p ~ 4kw(T -T ) 3 i The plug formation time for the three scenarios is presented in Table 19ED-7 for various amounts of superheat. Note that this equation produces the same result as Equation 19ED-9 if there is no superheat. The average velocity in the channel can be obtained by following the same methodology used in Subsection 19ED.4.3 for the case without superheat. Performing this analysis yields (insert equation B for 19ED.4.5) Note that this equation and Equation 19ED-29 are identical except for the definition of the parameter bo which describes the ratio of heat removal capability to the energy that must be removed for freezing to occur. The channellength required to ensure freezing is sirnply the product of the average velocity and the freezing time. The required channellengths for each of the three scenarios is shown in Table 19ED 8 for various amounts of superheat. All of the Scenario I and II required channel lengths are less than 0.5 meters. The required channel lengths for Scenario M exceed 0.5 meters for superheats greater than approximately 70 *C. However, for a channel length of 0.5 meters, the amount of debris that can enter the sump under these conditions is small. 4 The average velocity of the debris in the channel for Scenario M is less than approximately 0.1 m/s. Assuming that the total channel width is 2 ' meters, the amount of debris that can enter the sump is less than 0.006 m3 (0.2 cm average depth) for superheats up to 125 *C. This amount of debris ] entering tha sump does not pose a threat to containment integrity. 4 l DBM Insert 19ED.4.5 p.4/8 February 7,1994 u c u m - a n - u n -----

- c e-n m :9 a m

ws 3 j i 19ED.4.5.2.3 Conclusion of Superheat Impact A channel length of 0.5 meters will prevent significant debris ingression into the sumps for credible superheats in all three debris scenarios. No debris is expected to enter the sump for either the Scenario I or II melts, while a small amount may enter for a Scenario M melt which is superheated in excess of 70 *C. Based on the methodology of NUREG/CR-t 5423, the superheat in a Scenario M type melt is probably limite.d to less than 50 C. For Scenario M melts with superheats outside the probable range, only a small amount of debris will enter the sump and the average depth will be limited to approximately 0.1 cm. Therefore, the corium shield ~ for the floor drain sump will perform its function even if the core debris exiting the vessel is superheated. s l I i i i DBM Insert 19ED.4.5

p. 5/8 February 7,1994 n.m 4:s-;:e_:ee __ ____.

.;_,,_y

r: yLay L(e ss pt:18AM G E td.JCLEAR BLDG J fP.@/34 ~ .t ?= . 9' ffff [ tymp est 5 mr //. 1 s g + /l'I [ p /) (

1. l g -nc<e c

/}1 iff C,P b ? LE where in - /k'ff e r~ d'en,;7 t cc / Ud , hp @cn c., , ea capac,. cy oi'- / / ~ C E kJ M f b{ ~ g ? /)7 c o n. 4 ^ GN f u,ccrkC~tT /~ ~ Clc bo, ~ b.m - C 2 C 0,- ,y g dc .e

b,

.r--7 / + 7 2/ C 2G (*r - r:) i w/,' ere bo = ,Ac <n {& & C.yA 'fr,' dj ' l n :n 4 : 5-9: e -: e s,-----

-: ? - :. ;

r:: 3.u r:-

ums vwm Table 19ED 5 Effect of Parameter Variations Channel Length (m) Average SenI Sen II Sen M Effect _ _ _ Base Case 0.15 0.04 0.30 Debris: density + 10% 0.18 0.05 0.35 +19% density - 10% 0.12 0.03 0.24 -26% latent heat of fusion + 10% 0.18 0.05 0.37 +20% latent heat of fusion - 10% 0.12 0.03 0.23 -28% flow rate + 10% 0.15 0.04 0.31 +5% flow rate - 10% 0.14 0.04 0.28 -5% thermal conductivity + 10% 0.14 0.04 0.28 -6% thermal conductivity - 10% 0.16 0.04 0.32 +6% temperature + 100K 0.14 0.03 0.27 -13% temperature - 100K 0.16 0.05 0.33 +12% Wall: thermal conductivity + 10% 0.13 0.03 0.25 -18% thermal conductivity - 10% 0.17 0.05 0.36 +17% thermal diffusivity + 10% 0.16 0.04 0.32 +8% thermal diffusivity - 10% 0.14 0.03 0.27 -10% i I DBM Insert 19ED.4.5

p. 6/8 February 7,1994 u n u s-;;t-nt -___ ___

_a _gy

,mnu, f Table 19ED 6 Change in Energy due to Superheat Change in Energy (%) Superheat' (*C) Scenario I Scenario II Scenario M O O O O 5 2 1 1 10 3 3 3 25 8 7 7 50 15 14 14 100 31 27 29 125-39 34 36 150 41 a Table 19ED 7 Plug Formation Times with Superheat Plug Formation Time (sec) Superheat ( C) Scenario I Scenario II Scenario M J 0 5.7 4.2 4.2 5 5.9 4.3 4.3 l i 10 6.1 4.4 44 i J 25 6.7 4.8 4.8 i 50 7.7 5.4 ~5.4 q 100 9.9 6.7 6.9 i -l 125 11.1 7.5 7.7 150 8.3 I DBM Insert 19ED.4.5

p. 7/8 February 7,1994

. i 710?' 4 0 ? - ? ! ? - 'I f ! '* - - - - - - ~ ~ c'-7?-94 t*: p 41; y;

emcin ? p: -1 ~ ;,. .r Table 19ED-8 r-Required Channel Lengths with Superheat -i Required Channel Length (m) -i Superheat ( C)' - Scenario I Scenario II ~ Scenario M .j 0 0.15' O.04 0.30 s 5 0.15 0.04 0.31 \\ 10 0.16 0.04 0.32 25 0.18 0.05 0.36 50 0.23 0.06 0.44 -100 0.33 0.08 0.63 125 0.39 0.09 0.74 -150 0.11- ~ f.; j, DBM Insert 19ED.4.5

p. 8/8 February 7,1994 n m u t -,::- n t y __- _ __ _ _.

e: ;,_,, g,,,

v1DflG9 "54 ' WT:19AM G LC NUCLEAR MLDG J P.12/34 { t 0 s., insert 19ED.4.6 19ED.4.6 Conclusion of Channel Length Analysis A channellength of 0.5 meters provides adequate assurance that molten debris that enters the floor drain sump corium shield will form a plug prior to debris spilling into the sump. Three debris melt scenarios were s considered which bound the credible melt compositions and the credible debris ejection rates from the vessel. Two of the scenarios represented the oxcidic and metallic melts used in the Mark Iliner failure analysis, NUREG/CR-5423. The third scenario was developed with MAAP-ABWR to provide a gross over-estimation of the maximum debris ejection rate from a failed vessel. A sensitivity analysis demonstrated that a channel length of 0.5 meters provides enough margin to account in uncertainties in material properties. Additionally, the impact of superheat was shown to be minimal. t ( I i ( DBM Insert 19ED.4.6 p.1/1 February 7,1994 n m us ;:2-:n -- -----

- e e-n mae n:

p. 2PJ73%

22A6100 M:v. 2 ACWR smowentsdoryAurrubnorm eW eZeMQI CE L:cKsef. ,;rh G[#sn - /dh J e fsuey* hon!6y, caall y 'ct,our aae W

. selgrec

Insert 19ED.5.1 A To help assure the integrity of the roof of the corium shield, c.he upper shield wall should be tall enough to prevent the debris colletting on the lower drywell floor from flowing on top of the shield. Debris fa' ling directly 1 on the roof during relocation from the vessel does not pose a threat to roof integrity because the amount of debris that can fall on the roofis small. This is a result of the CRD and lower drywell configurations. The length and density of the CRDs in the lower drywell prevents the debris from exiting the CRD grid with any significant horizontal velocity component. The sumps are on the periphery of the lower drywell floor. Thus, debris which relocates from the vessel will not fall directly on the sump roofs. Insert 19ED.5.1 B After debris relocation into the lower drywell, the lower drywell will almost always be flooded with water either by active systems (e.g., the firewater addition system) or by the passive flooder. The probability that the lower drywell will not be flooded after debris relocation is approximately 10-4. This probability is low enough that the case of a non-flooded lower drywell can be excluded from consideration. 4 The water in the lower drywell will provide long-term cooling to the debris on the floor, to any debris that is on the roof of the corium shield, and to the corium shield walls. Additionally, since the roof of the corium shield allows water flow into the sump when the lower drywell is flooded, the inner walls of the shield will be cooled. The heat transfer from the shield walls to the water is effective in preventing the debris frozen in the channels from remelting. Therefore, if the lower drywell is flooded, long-term solidification of the debris in the channels is assured and debris ingression into the sump is prevented. To meet the requirements set forth in this subsection, the upper shield wall is specified to be 0A meters. j 4 - DBM Insert 19ED.5.1 p.1/1 February 7,1994 1 m w-:.: m _ ___ .n m

m i 22A6E008 y.2 ABWR samutantsetory Annorsis n: vert ~ 4' ($'. n fN height of the upper wall N Huw N temperature of the upper wall in contact with debtis Ti .s. tempenture of the upper wallin contact with the lower drywell To = environment .y Natural convection governs the temperature of the wall in contact with the lower . drywell emironment. The heat flux from the top of the wall can be written as: q", = E (T,- T ) (19ED 30) g where: natural convection heat transfer coefHcient E = temperature of the lower drywell emironment Tu The natumi convection heat tansfer coefficient depends on the Rayleigh number.The Rayleigh numberis: g (T,-T )L id a (19ED-31) Rat = va where: Rayleigh number

Rai,

= [ thermal expansion coefficient of steam = 1/T assuming ideal id = gas behavior kinematic viscosity of steam v = thermal diffusMty ofsteam a, = characteristic length of the shield top L 3 The characteristic length of a horizontal heated plate is one halfits width. (Reference 19ED-4).The Door drain sump is approximately one meter wide; therefore, the characteristic length of the shield roofis 0.5 meters. Evaluating the Rayleigh number for satur'ted steam at ultimate containment pressure (1.34 MPa (180 psig), ( 463.1 K (190'C)] yields: \\ 1 % D-15 Corium Shield - Amendertent 32 t M DWWWD$

m w j i 22A8700R;v.f ( ,.ACWR standertsareryAnno ateneport l i v 4 i 4 -.. s. c {w$ ,9 ? Rag = 5.5x10 K-1(T,-T ) . 09W2) 8 id or 10 s Rat, s 1013, the Nusselt number (Reference 19ED-4) for an upward-facing f 7 ii 9 heated plate undergoing natt.ral comection is: i 09 33) f Nut = 0.15Ra t 'j The average natural convection heat transfer coefficient is: l u h=k Nu (19ED 34) -' L L L, where: j 4 t thermal conducthity of steam in the lower dr)well l i k = Combining Equations 19ED-30 and 19ED-32 through 19ED-34 yields: q"uw a 8.79 (T,- T ) W/m Y OW435) id which can be rearranged to: l i j

1. 4 e

u 9** T, = Tg+ K (19ED 36)- (8.79 W / m Inserting this into Equation 19ED-29 yields: @4

• j r

a y* K (19ED-37). -l q",= T, - Tg ' uw (8.79 W / m,s This equation can be solved iteratively to determine the heat flux which'can be i transferred through the upper wall of a given wall height. The wall heigh. quirement for transfenir:g a given heat flux is: 44 ~ ~ r a y H,,s," T;-Tg K (19ED-SB) ' 9uw (8.79 W / m s L .,\\ te decay heat levelin the ABWR,24 hours after accident initiation,is approximately 0.6hTh etric heat generation rate of debris at this time can be determine 10EO-16. MH - Amendment 31 l l 770tt t.:?-p;t-;sti-----. - g - t, g g 4 3 7. ;; g - y,

venwrwer~WrTzaw s Lc wufcurgilaags y F. insa 22As100 nov.3 ABWR snodsntsatoryAn:tysis neport ~ 1 using Equation 19ED-6. It is 0.9 MW/m. The debris /wallinterface temperature which s will guarantee that the debris remains frozen is 1700 K. The temperature of saturated l steam at ultimate containment pressure is 453.1 K (180'C). The height of the upper ) [ shield wall. which will transfer all of the heat generated in the channel for these g condidons. is: k* H,, 5 (19ED-39) 8.52 W / m,K If the upper shield wall satisfies this inequality, it will be capable' of transferring all of the heat generated by debris in the channel; and, as a result, guarantee long tenn debns solidification even if the lower dqwea has not been flooded. To be acceptable, the height of the shield wall m*ust satisfy the inequalities in Equations 19ED 39 and 19ED-28. 19ED.5.2 Lower Shield Wall (Below Lower Drywell Floor) m/r-4 CdaM6N ne side of the lower shield wall is in contact with debris and the other is in direct [ contact with the basemat. The basemat is constructed of concrete. A conservative ,[ estimate of the lower shield wall depth can be made by assuming that concrete acts like a perfect insulator. Thus, no heat is allowed to pass from the shield wall to the basemat. 3 The boundary condition between the debris and the wall is conservatively assumed to ~ be constant heat flux. The initial burst of' energy into the shield wall, caused by debris I freezing, has ample time to distribute itself throughout the wall. With these boundary conditions, the temperature distribution in the lower wall can be determined analytically. The analytical solution will provide a means for determining the time required for each of the interfaces to reach their allowable temperature limits for a given heat flux.The wall /basemat interface temperature should not exceeded the melting point ofconcrete~ (1450 K). Continued debris solidification is guaranteed if the wall / debris interface temperature does not exceed 1700 K. The wall will be sized so that the limits are not exceeded during the fint 24 houn after initial debris solidification. The upper shield wall will be sized so that it can transfer the full decay heat load after 24 hours has elapsed, as discussed in Subsection 19ED.5.1. The temperature distribution in a slab (Reference 19EIF5), subjected to constant hea flux at one surface (x = H ) and insulated at the other (x = 0), is: S 7 pO ~ p te 6 J 19ED-17 Cchum Shield-Amendment 33 - m m m... = .m

v:wa 3 6 ..- a ?pa.f. ~;

9... :.

r Insert 19ED.5.2 As stated in Subsection 19ED.5.1, long-term solidification of the debris in the channels is assured due to heat transfer from the shield walls to the water which has filled the lower drywell due to initiation of an active system or the passive flooder. Therefore, the only requirements for the lower shield wall are to absorb the initial energy released by the debris during the freezing' process and to prevent tunneling of the debris beneath the shield when significant core concrete interaction has.not already occurred. 'q!...v. During the freezing process, the channel' walls behav'e as semi-infinite bodies with a penetration depth less than a centimeter. Thus,.the lower shield wall needs to have a depth of at least one centimeter to meet the a' requirement for initial debris freezing.- ~ 4. If core-concrete interaction is occurring, the potential exists that the lower drywell floor will be eroded to a depth below the lower shield wall. If this occurs, the debris could tunnel into the sump. Thie ' concern is completely climinated by specifying that the shield wall extends down to the f oor of the sump. ,y { l o l DBM Insert 19ED.5.2 p.1/1 February 7,1994 n: m - : w = _____. m.y y, 7;.

4-qq. 23A6100 R:v.1 BWR standardsatoryAassysis Report ~ ~ p -- q" t q" Hg* TIw (x, t) - T1. i* = p c H k + w p, w tw w (19ED 40) 2 Eit/H'i, nnx' ' 3 x' - H 2 { (--1)" -% ~ iw e cos 2 g y* \\ 6 H,,, l x n=2 n where: temperature distribution in the lower shield wall Tw a i adjusted initial temperature of the shidld wall T jw = i heat flux through the lower shield wall q"1w = specific heat of the shield wall e,w = p depth of the lower shield wall below the lower drywell floor Hjw The maximum temperature at each interface is achieved as t --+ w. The maximum h.. temperatures at the wall / debris interface, Tw/d, and the wall /basemat interface, Tlw/b> are: l 9]wN]w 9fw Tw/d = T;,3, +.cp,,Hh Sk, and: qf,t q",H, i w/b = T,3, + pw p, w T ~ t C H 6k tw w The heat flux through the lower wall is bounded by one-half of the heat flux generit;;d in the channelwhen the sizes of the upper and lower wall are comparable. The actual heat flux will be lei because the upper wall is free to co-ect to the lower drywell emironment and will accept more heat flux than the lower wall. The maximum heat generation in the channel corresponds to a decay heat level of 17o. Since decay heat decreases with time, using the maximum value bounds the temperature response of the lower shield wall. Using Equation 19ED-6, one-half of the heat flux generated in the channelis: j .J ( Corium Shield-Amendment $1 19ED 18 Fi c 4 c c -;;9 : gr____

..,VLQTLSW "T4 Ffs 2M1 V E iAJCLEAR BLDG J P.20>34 2SA6100 Mov. 2 ABWR manMsanyAnatnisnem .. ~ Q geg-( (9EO. %$ .l [ 91/2 chan

• ( 9Ew) limiting C!n where:

m = total corium mass. cm The initial temperature of the shield wall should be adjusted to account for the energy it absorbs dudng the debds freezing process. If both shield walls have the same thickness, the adjusted temperature is: Pc,h H, th (19ED-44) 1,1w = T + 2p,c,H, T 3 p, 3 Equations 19ED-41 through 19ED-44 can be used to detennine if a chosen lower shiefd wall depth will satisfy the requirement of keeping the debris in the channel frozen for at least 24 hours. After 24 hours has elapsed, the upper shield wall will be able to remove the entire amount of heat generated in the channel (Subsection 19ED.5.1), J The process for determining an acceptable wall depth proceeds as follows. First a wall' depth is chosen which is comparable to the upper shield wall height.Then, the adjusted initial temperature and heat loads are calculated using Equations 19ED-44 and I 19ED-43, respectively. The interface temperatures at 24 hours are determined by Equations 19ED-41 and 19ED-42. If Tw/d < 1700 K and Tw/b < 1450 K, the chosen depth is acceptable. If not, a new depth is chosen and the process repeated until an acceptable depth is determined. An example of this procedure is given in part (4) of Subsection 19ED.6. l 19EDfr h6factica ef CeMgn-Requiremerits: ;o: //m m T Fr,'er M ef Esizing requirements for the floor drain corium shield were set forth in Subsections 19ED.4 and 19ED.5 based on a chosen channel height. The satisfaction of these requirements by the nominal design is demonstrated in this subsection. The selected channel height, H,is one centimeter. Representative shield wall material o properties are shown in Table 19ED-1. (1) Melting Point of Shield Material Above Initial Contact Temperature The initial contact temperature between the debris and the channel wall given -- A in Equation 19ED-16 is: f p, ec t t uD. C Corium Shield-Amendment 32 19ED-19 yp:u 4: 9-gt-:5y -------- r:-:: ;;

ne y;

ya;:f.;g.g g-g.g p, a,34,.. , vts Lea '94 07:23An a c nuctmW a.DG J w y., y - (,. ',, 'J.;. 5,V'k*; 7 , a.@.? e .y g ' gt:;;.,i C f i.> % ~ c . n u. .c. 5. .r 't Insert 19ED.5.3 7 y ; & ;.e Q :r T.Q... w y IDED.5.3 Shield Wnlls without Channels J J O i .o .. y. 7.,.. .r The discussion in most of this attachment has focus 6d on the shield walls with channels. This subsection will address ~ the requirements for the shield walls without channels. The thicknes' ; height add depth will be s specified. "?@f5QifDy The corium shield walls without chanEei's only lfieed t6 be thick enough to aQu W:;.W: provide a long-term, stable interface between the debrid oh the lower ~ drywell floor and the interior of the suidps.?ls' dis' cussed in' Subsection 19ED.5.1, only the long-term scenario with Eflooded lowe'r drywell needs ' $ D M( 2 f J M; ?WMf. s ~ to be considered. MMd The wall thickness needed to transfer a 'given heat flux'under steady-state ! q.g; p.L.N.l:cptp'Mi'y conditions is k,(Tw,g-,T,,1) $0fl &T 7;?.p t y gg.o ? gx = ~ ",h."hkfftl$YV Oh!R. WfpW: surface temperature "of Wall irid.ontact with debris where Tw,o = surface temperature df wall on~ m%jKF 't M M p M 3p sidelof shield, and g s' Tw'i = heat flux from tha de@bn$,Q h.L.M a ' ffMt

s. v.

fe-qs = Assuming that the water in the sump is'fMMnl% 'J ~ dt thrie atm6 spheres, the inner shield wall will achieve a temperature'bf~al prodmiitely 4-10 K to allow i ~ nucleate boiling, To avoid ablation, the%11 surfactuiniperature in contact with debris must be less than th6:ineltirig tiinperature of the shield wall material (approximately 2280 Kfor~aliimina). MkM%h $hw. The heat flux from the debris bed'in thdloser drywell can'be ~ approximated by assuming that the decay lie ~at level is one-percent and that all the surfaces of the bed have the%Aine' best flux; This approximation yields: /jys . j 7, ;f y.g. qg = 2501si/m?..[% M. $.N x The actual heat flux to the wall may be significantly' lower due to enhanced heat transfer from the debris bed to the. overlying pool of water or separation of the bed from the wall. Evaluating the wall thickness for these co'n tions yields f Ax = 3 cm. DBM Insert 19ED.5.3 p.1/2 February 7,1994 n: m-m-av---- nam n

FEB 08 '94 ~07:24AM G E ltJCLEAR BLDG J' [ ~, ' ~,~ ^ ' ~ P.22/34 4 ...j v : . :t p, ~+ ;....,- i This wall thickness provides a stable interface between the debris bed and ,3' a water Slled sump. If the wall is thicker, it will ablate to this thickness and then establish a stable interface. To provide mnrgin for any erosion due to initial debds contact with the wall, the thickness of the shield walls without channels is specified to be 10 cm. The reasoning contained in Subsections 19ED.5.1 and 19ED.5.2 regarding the height and depth of the shield walls with channels also applies to the 5 walls withat channels. The height of the shield walls should be 0.4 meters which is greater than the maximum height of the debris bed in the lower drywell. The shield walls should extend to the floor of the sump to prevent debris tunneling. s () 4 DBM Insert 19ED.5.3

p. 2/2 February 7,1994 n m a-s:t-r --------

. :- o -.4 u :s e 7::

'rca ce 794 'c7:24AM G E TdCLEAR BLDG J ^ ~ ~' P.23)34 ~ ~ .( Insert 19ED.6 19ED.6 Related Experimental and Analytical Work The freezing of molten fuel in narrow channels and tubos has been studied previously in regards to core disruptive accidents in liquid metal fast breeder reactors (References 19ED-10,19ED-11,19ED-12 and 19ED 13) r and in regards to ceramic core retention devices (Reference 19ED-13). This subsection will review these works for application to the channel freezing analysis contained in Subsection 19ED.4. Cheung and Baker (Reference 19ED-10) analytically and experimentally studied the transient flow and freezing of molten core debris in coolant channels of a liquid metal fast breeder reactor. Their data analysis determined the impact of several parameters on the penetration depth of the fuel into the channel. The derived variations are in general agreement with the trends shown in the sensitivity study contained in Subsection 19ED.4.5. This work culminated in the determination of penetration depths for several coolant channel diameters. The material properties of Scenario M compare somewhat favorably to the material properties used by Cheung and Baker. However, they used a channel flow velocity of 100 cm/sec. Modifying their results to account for velocity differences yields penetration distances from 20 to 66 cm for channel diameters between 0.64 and 1.27 cm and a debris temperature of 2770 *C. This result compares b<. well to the results determined for Scenario M-a penetration depth of 30 cm for a 1 cm channel. Fieg, et. al., (Reference 19ED 12) performed channel plugging experiments at the Karlsruhe THEFIS facility in Germany using alundna and alumina-iron melts as fuel simulants. The results indicated that the conduction-limited crust growth is an adequate hypothesia for modeling the penetration and freezing of molten fuel. The basis of the this crust-growth model is that a stable crust forms at the channel boundary and then grows continually until it clogs the channel. This is also the basis of the model developed in Subsection 19ED.4. The experimental results presented by Fieg cannot readily be compared to the corium shield because the experimental velocities are so much higher (2.2 m/s to 4.4 m/s compared to 0.01 m/s to 0.1 m/s). However, Fieg's findings that increasing wall temperatures and/or driving pressures increases penetration depth are consistent with the model developed in Subsection 19ED.4. Soussan, et. al., (Reference 19ED-11) compared the results of experiments performed at AAE Winfrith and CEN Grenoble with the BUCOGEL code developed at Cadarache. The comparison revealed that penetration depths are over predicted using the conduction freezing model and under predicted using the bulk freezing model. The freezing of UO2 was shown to consistent with the conduction freezing mode. Alternatively, freezing of ( molybdenum was determined to undergo bulk freezing. This would tend to DBM Insert 19ED.6 p.1/2 February 7,1994 b h h*h ' (

• h ** h h (I

d

$@ @B #@s @7:25AM G E NUCLEAR BLDG J P.24/34 indicate that the analysis in Scbsection 19ED.4 over predicts the free, zing (.) of metallic melts such as Scenario II. However, the overall impact to the analysis is negligible because Scenario II does is not limiting. PLUCM (Reference 19ED-13) was developed to analyze freezing in a variety of geometries including the gaps between the ceramic bricks of a core retention device. The model in 19ED.4 is similar to the PLUGM model for " Thin Slit Geometry - No Crust with a Noumelting, Infinitely-Thick s Wall". The primary difference is that 19ED.4 is somewhat more simplified to allow a closed form analytical solution, whereas PLUGM must be solved numerically. Unfortunately, the example contained in Reference 19ED-13 for a thin slit geometry does not lend itself to comparison with the corium shield analysis because the example models a vertical channel with a high entrance velocity. The past investigations into the freezing of molten fuel in narrow channels tend to support the modeling and results of the channellength analysis contained in Subsection 19ED.4. hr e -( DBM Insert 19ED.6

p. 2/2 February 7,1994 m 4:g;;e:u ______ _

a_:ge; . :, a 3:;

6 RutLLrte r,ityg y (9.25/34 23A6100 R:v. 2 ACWR sundedsureryAnotysia nopon ~ - (c, - 2T;) c - 2T ) * - 4 (T* - c,T,,) o 3 g T*== (19ED-16) 2 / c where: h "w' 'nk p m th fc O g 2k, s The parameters required to evaluate this equati9n are: f initial temperature of the shield wall, SSO K Ti = debds freezing temperature ranges from 1700 K to T,m = f 2500 K debris thermal conductivity,30 W/mK kr = 3 density of corium,9000 kg/m pcm = 5 debris latent heat of fusion,2.7 x 10 J/kg h = lh ~ thermal diffushity of the shield wall material, a a w 4 2 representative value of 1.48 x 10 m /sec will be used ) thermal conducthity of shieH wall material, a - kw = representative value of 4 W/mK will be used Evaluating Equation 19ED-16 yields interface temperatures of 1560 K and 2180 K for debds melting temperatures of 1700 K and 2500 K, respectively. The melting temperature of the representative shield matedal is over 2200 K; therefore,it passes this test. (2) Channel Length The equations needed to determine the channellength required to ensure that a plu5 s formed at the entrance of the charmel before debds spills into i the sump are Equations 19ED-9 and 19ED.26. b I (reeze

• o freere freeze

~ w Corium Shield - Amendment 32 tsCD 20 FE" 4 0 E - :- ! *. - j f 5 7 - - - - - - +

-??-E4

:_5 AM r!?

] iridft@ 79#J t#1P4sfRR @ E (KU%LL%1 LGN6 X ,s

@.$y34 2 SAC 100 nov. t *

.~ ' smedentsennyAnalysin serat IABWR '.(, .Pc,hth gg g g). E =_ rreeze 8 (TCm-T) a . s I_ 4 2gRd a = - 5 p,Ald - J $y l } 2k &g,,-T,) 5 f --,and b = - 3$ pc,hth 3.x. the assumed channel height (0.01m). - H = o The maximum length results wheriTr,,- 1705 K. The contact temperature was shown previously to be 1560 Kfor this freezing temperature. The other - j parameters required to evaluate these equations are: l. P, .- y e., ejection rate ofdebris from a failed yessel,6000 kg/sec 4, = m t , I'.;,. (conservan,ve maximum), ~l 4 { -4.- s-,q .l I minimum floor area of the lower drywell, speci5ed as A, min ld 0.02 m /MW in thiEPRI ALWRRequirements 2 th 2 Document; itis c' qual to 79 m, Using these parameters, the plug formition time is 7.2 seconds and the i required channel length is 1.06 meters. This length was determined using a highly conservative corium discharge rate. The analysis assumed a constant discharge rate equal to maximum discharge rate predicted using a highly conservative model. The actual dischaq;e rate will be lower. If the length requirement is highly restrictive, the discharge rate could be refined with additional efrort. ,..o..., t (3) Shield Height, H w, Above Lower Dr>well Floor u The height requirements for the upper shield wall are given in Equations 19ED-28 and 19ED-39. These equations are. .) H,,2 0.33 m (19ED.28). l ? and: 19ED 21 Corium Shield-Amendment 31 IECM 4C$-92$ '{$" --+~. -- g7;gg g4 ,g,.g .1

' ~ TEB 08n*94 H 07:26Ali G E ftJCLEAR BLDG"J ~ P!27/34 R ~ -~2 san 100M N.1 ~_ ACWR sanederdseveryArksisnaport k, ~ H,5 (19ED 59) 8.52 W / m,K -1 For a wall conductivity of 4 W/mK, these inequalities require: } i ( 0.33 m s H,s0.47 m (19ED46) y a A height of 0.4 meters is chosen. Shield Depth, H y, Below Lower Dr)well Floor Lj (4) i The lower sideld wall should be sized according to Equations 19ED41 - .j through 19ED44. An initial height of 0.4 meters is chosen to begin the - detennination of acceptability. The adjusted initial temperature of the lower g shield wall, accounting for energy absorption during debris freezing,is: Pc,h H, th { T " = T, + I1 2 p,c,,H, (19ED-44) p 3 = 341K 3 where: density of wall material, pw d specific heat ofwall material. -{ c,w = p 5 The limiting heat flux through the lower wall is: O .dh cm l% E o ( q",) limiting 2 m,, (19EM) = t 2 = 7520(W/m ) 1 where: 1% of decay heat,39.26 MW j Q1%,dh ^ mass of corium,235,000 kg m = em The wall / debris,Tw/d, and the wa!!/basemat, Twfd, interface temperatures - .k., are given by: 1 Corium Shield-Amendment 31 19ED 22 'F E O M 4 0 5 - 9 : 5 - 15 9 - ------- - g,.;p-p; 3 9 7., g y

FEITIf8~'94 07:27At1 G E NUCLEAR BLDG J P.28/34 23A6100 Rev.1 ABWR stnadardsatoirAnahnutsseron .t**'

9fwH, 9iwI j

+ (19ED41) Tw/d = T1,1w + p c H Sk w p, w lw w P8 1 and: q(t q(H, 3 (19ED42) Tw/b = Ti. lw + pw p, w e H 6k Iw w where: time assumed to be 24 hours. t Evaluating these expressions yields Tw/d = 1190 K and Tw/b = 820 K. Since these temperatures meet the requirements for long-term debris solidification (Tw/d < 1700 K and Tw/b < 1450 E), the chosen wall depth is acceptable. (5) Summary ofShield Requirements A proposed floor drain sump corium shield with a specified channel height of (S. one centimeter and wall material properties shown in Table 19ED-1 will ~~ prevent codum ingression into the sump ifit meets the following i requirements: Minimum melting point ofshield material: 2180 E, s Channel Length: 1.06 m, a Height above lower drywell floor: 0.4 m, a Depth below lower drywell floor: 0.4 m. e 19ED.7 Detaileci Design issues N During detailed design of the ABWR, the exact shield matedal and shield dimensions will be chosen. Cursory examination of material properties indicate alumina may be an 1' acceptable wall material. The requirements for the shield are stated in Subsection 19ED.S. Example calculations of the requirements are shown in Subsection 19ED.6. Interference with under-vessel servicing equipment will be considered in determining if the proposed dimensions are acceptable;ifnot, the sizing / process will be redone for a new channel height and/or a new shield wall material. The / number and width of channels in the shield will be chosen to meet the design requirements for water flow into the sump during normal ABWR plant operation Corium Shleld-Amendment 31 19CD.23 77:

n-;_

-;ts:------ -

- ?-;;

e r-

M ' 21As100nov.1 ^ ACWR steadentseveryAntrais neverr ,i 19ED.1) References 19ED 1 Frank P. Incropera and David P. DeWitt, fundamentals ofHeat and Mass Transfer, 2nd Ed., John Wiley and Sons,1985, pp. 85-86. s 19ED-2 Glen E. Myers, AnalyticalMethods in Condudion Heat' Transfer, Genium Publishing Corp., Schenectady, NY,1987, p. 202. 19ED 3 Frank P. Incropera and David P. DeWitt, fundamentals ofHeat and Mass Transfer, 2nd Ed., John Wiley and Sons,1985, p. 203, 19ED-4 Frank P. Incropera and David P. DeWitt Fundamentals ofHrar and Mass Tmnsfer, 2nd Ed., John Wiley and Sons,1985, pp. 433435. 19ED-5 H.S. Carslaw andJ.C.Jeager, Conduaion ofHeat in Solids,2nd Ed., Oxford University Press,1959, pp.112-113. 19ED-6 Mad's StandaniHandbookforMechanicalEngineers,8th Ed., Theodore Baumeister, Editor-in-Chief, McGraw-Hill Book Company,1978, pp. 6171 to 4177. N ,qgy-;r z & 7%e04eo1; W./2 Anara o*r;Y"i Nkn ad i

u. Raws "rhe Prohaht/<

' W U n e'~ f #" Maes-r cedca/g estt," Atcesh& /c/-Fr.25.. in a. Augesv 199/. teed-8 (up e rep ffE.Vl) ) i95D 'l " MC (4n boot o.P datir cry and' i=%y1. cd & Prerr, Doce hwn,fM.rdt, tqet gq59 tc cI1e unf, R 5., and Ba / der, 2-., " T^rentoh r feceaisy,. of t-<'y.o icfr in 7~ule Floc ~' .Alo c(n:t i fe<*est c e ad ~ Eq <*^ cer a y, 4 o,,m'in l'~ $ /7 7 4 Soussos, E, Sc.Awaett, H.,2iMayon,2?, g.,1s' Der ede5 B I9ED it ei f'r-o,C% p cfan a d frce c) S s'-fo fc e M MTChiQ/I 7 s cerf recccC*fc.n o 5 Ep r/in esCaf sk'd/hr l'. re.oceedty s e5 rde /79'e> ni erm tion a-( t ikt c ac'ea cec <.Grfey Heec/y fnowdied, ( uvah,ingarva-t4 nre I Corium Shield - Amendment 31 1p50 24 7 7 0'4 4 0 E -?; 5 : t r-- n_;g_p; 3;. 9 C ~ r M-

FEB 0394 07:23AM G E NUCLEAR BLIXT, J r !., vq. P.30/34 c... # @. -: . s l ". ' i.e. ' o o nlf.;';lhN$l$. ~b.b ^ '. w:~m y y.;,;g. q p ;; .7 C. 19ED-12 G. Fieg, M. Muschke, L Schub and H. Werle,'" Penetration and Freezing Phenomena of Ceramic Melts Into Pin-Bundles", Proceeding of the 1990 International Fast Reactor Safety Meeting, Snowbird, Utah, August 12-16, 1990. 19ED-13 M. Pilch and P.K. Mast, "PLUGM: A Coupled Thermal-hydraulic Computer Model for Freezing Melt Flow in a Channel". i NUREG/CR-3190, SAND 82-1580, Sandia National Laboratories, Albuquerque, NM 1982. 4 4 ** ee \\. ? }, { ' ' 4[*

• {-;*I*---+.--.

,. { gc g,4

*.. ap go

• 7-+

'N' l 2348100 Mov. 1' ' ABWR - aandedsareerAar ubnepari r . ** w I' Alum *na. Table 19ED-1 Material Properties of *RepresentativeRefr;;'. / &!d-3 and Concrete ReprenantatMriek. A/gm,'on. g Proporty (Reference 19ED-6) Concrete s Melting Temperature (K) 2200 J.2 go 1450 3 Density (kg/m ) 2700 2300 Thermal ConductMty (W/mK) 4-9 1,3 Specific Heat (J/kgK) 1000 800 2 Thermal DiffusMty (m /s) 1.48 x 10-8 7.5 x 10-7 ).

s..

Corium Shield-Amendment 31 19ED-2S rien ::E-9:5-:59 --------

e - e t - N 1c

09 An

7.- 'FCB 08 '91107:28At1 G E TLICLEAR BLDG J P.32/34 l

^

1. f 2.wsren 1

'ABWR smadantsavnyAnstruis nerat x t Od s Lower Dryw NAf' 1 Corium Shield I FOL s O .j i h Qlag/lelff Chann I r. .k' M W.Eh'[ 4'jk i 0 .::t ? h T

1 % ra

'y A ,. +

;h l ^y.,;l1Got?$,f:-lA&,'$ l;; t. g:..

g. np T . ~;.ygi; y x w <C Sump f r.:,... ap. .. m. I b

O

^ lQ ',;;b2.M:$Q 1

+

- g ( (Q ,g .#. m s .r e Concrete Fill i 5 k h . d A b dif@.$ji%k{% ifQi 7 m. _;f .e; g ce nver h f D b .Y, l ' (o uJer Containment Bounday ...d e wed a) side' view I f hannels _ 3.;=^^ 'I. _ f:"% \\ m m m m 'c Nu' M ;, X WN]

D.

., i 5, > R }::. O.; o ti - v . Tf?$ p - xy;ts.;;g.g g ij l ,g. y >y. ; "f.a Rina hy spy, g ConcreteFill W/;n.i. #

.e u'

.xx m ts;m a Lgg

.' ey ex fx n,2 v.a. x :t s b) front view (7 Figure 19ED-1 Conceptual Design of Lower Drywell Floor Drain Sump Shield 1 SED.26 carium Shield-Amendment 31 neu m-m-m - c: os-u a e, ,.n i

rte ce w e m u,c,et. au.c+ m e #.,,,3.;,,..j.;.3 . _ p4.g g wwa_ .. v :m. s.v 2W4%q y o zustoonov.W,9 44 y pg3: fs

q,,q.,, m.. e: a...,,.udedsdotyAue,'ysis nore.t s

...ABWR* m e m.- Jg ;, ~ * - 4s. 7*$..., s s 4d e .e ..I k y g. v .'4 4 e .a. a y,u. :..x. ~, p).m:n.. .p. v, . w,+ I.* ig[U,'*' }f,{,* :3. l,.

• v, %;;p.2; 4m..w. %., "

e e ...1 Yc....'.W ~, ru..n.. 7l%.: 9;. Sh.i.e,ld. Wall ,,,n,s $4. % y(~ seml4nhite .:fd,ydgs M, #9%u,st.,."4@~4% w,o,d ),L. Ci -% Solid B Molten Debris

e....

a 4,% w ,+ g m --. mzw.run caaaccauxacaucua tutunautuutututuu;. 1 yr$f-/ gyMg

tttunutuuntt'ututt"?:ututtatuunulutua! :

rcg3 r - c ' c., y.~ %.1' w a

tutuutunnunuuuu

e.g.. r -C AA,.,l% .stttutuututunnunta: m 4 ~

tututututttttuuuuMS; dek.33,,M'fmM"R

+ r gwnnngj y..rlfh

_ g

. >pp

tunutunautununut

m.*h.c.b A_S wsE.WL_m,

Muttuntunnuutunu!

Tr.m

UtMMMthMthuttunut-

%g-J ~ ~ . 'f.. -kpKfRV^i.5'*-*JG

Muuntuntutunutut$j i

. K 4.y

3..

• %gw

.4 1

tutugtututununuttu

auustuuuuuuuuuu.

.y d, my! @Q' " .M

uutuuuuuuuutunui..

e, _, _ er

uuuuuuuuuutunu

(~. - JX6 ) 7-A utu utu ututtuuS ..; w.5 -. ']., _ _ T[uunuutunnuun!, 3 h.ngnf.=m;-=M M,5

d uuuuuuuuuuu

,4; ; . c ~m m,, W wnunatuuuuuuu:! q;> m ! %w % g 4.' 1 ; p M u +pq.<-

uhm D-

"'uununutun J 'i g.y k.3M 4*

aututuusa"""""?? 7 1,

s-j w, h4 KV4

• '. 4..; d[W,b.2 ['

ututuuuuuuuuGhW i

,.z y r). ' - a

MutuutututunnutuS p

nuttuutuututuutut{!

A*. g. -7 J. 4 jym.c s unutununutununu!, e -

tuntutuututtttttttuts gt,

a s v (..uuuntunuuntunua: t.. - :. 4 l g.- y ,1- ,, %,w. y 'i$gm

utttggtuttunutunuu8 7

a ~'.y.

tunutuuttuutunuuts.

p... 1... e, g .nm mmnm>mm>m r m; . p.,s.my.a *;.~, ;gg,g c ?,4,

.g.

Channel Freeze Front, i, 7.;g;7S.R g Centerline ,'.#.#.'.I.'/. ,.,a i r x, =, ,"s] 8 [.jp N, % ~ i / ' \\, Figure 19ED 2 Temperature Profile in Channel Region Codum Shield-Amendment 31 19ED-27 1 T t, e, i

• 4 eog;$_t g5'"'..,.__..

-.**_4 g ,Q ' r-*

A UN N $ f.[ ' } ' *.,. ",(, h d'* { 4p "[.,k e g,

u94N&(qy[y,&7,

. 'fr;n ' > T4 # L *. 'w ., ~ 23487M b ft.g ' (4., de ifa'sneesNser,d,, ys..'jA.hl,sh'.,i_ ,.-a. i d

,,, a. ' {"p'... p ',

'7,,,,.,g..Q,. A:,TW; -ct.1'tlLa w an v. 4 4r y AMR ' ~ 9.(*.T.. ', \\ s ..t.c 1 4. Lf,'o '$Y., ;l.f..J...A* ' s ~ Q ' ' v- -,,. e .a% 'y%,:pg*E.,5~/:.:'s. I

,, i, k

. *.. ;y/.* 9 ?> M.,4 r. r-g.,

• r[. 6//.*.!* i r. t

.c,',, ? *...'a,'; '. ;.1 ',,i_Q :. ),, L.;'; * ' 'W ;,. j '..,t

• /

\\ 6 =,r.,* ,. '.f., y . 'c. . s ; ' r.: ,,t., 1 ..<c g Te s. ( g ' ' ; [.. r < < .,4 ~ s.' j a -r ' g'. ';s'o J.,.,.,. 4 'I' '.1,. :,,..: .,s .+ pse,;- ed a L ' ;..

• 3

, g m.yy.

• .y.s,. ; %','J. '.' ? ]-

. '... f ' '/,b.A' h

• 8 ' f.'.

gr f s > 3 ~ ~ S,i:3f' 't t. -e s,.. .~ ll. f,E,.$<,* .. '.,. <... :. t... 'gg

• l

'. 'j t .g.. . g..

7..

..,.,,4, .3 3 , 3 va,., ..,,t' 2 4. '. P y3 .,.. ~ a ~. ! 4 !. s h. r 4." at;- . %..e e s.t of ss >.s. ~.., s ..e ^* e.'N:,4 Qf'p 'h k*~4 (; * * '. ' . ~

• i y,.,

.. r,7 f fs;, [qg, ', .~< r.

1... % v, g *,,, "..

-4.,"C '

• • f" r,.,

.?4.... .. t ' v,..,;,gnl1$,A.i".f ?y '. '/ %); J, a'i',

.v ;,

? . f t.S,w, y.-y.y;,.... n,.. >.,.y. c..- .. n.. . d. f G C l $. ? r.y..'.......Chan%. W& L. ' ~ e' e a:W h.w,e nelWall e . Incr.e.a.s. i.n~g Tu..ne'..Y..y fg.,:,. y?p.. ,,p.,f.w.,.e.r.v... .... n._.:...,.. ;. ~ A Molten ..r..a fs... m u.e y-... Debn.s 7 0Q g,1' 4,..jg p;,3.,'.,,@, d: e$.. eq A..h. of y 0H(t) ..$ frNi .m.~ o. e Eron:..a@,e.,Y W N; 4A ... u Freez.a Velocity t- . 'M m s,...;'.. ;. '.1L - u g '.e gW,f_,, ?4. J ', y:,. t'9 L f,,,,',._*,, Y c 9 A , witMgp,s..,wec-v, v..- .g f ~ 's n. . n,1.:.y.l.1 if.,s,. s .......;p.. . 3 s .r-

m.,.

. a....m.,.

a m... a ,.y '{ .., -.. c;. y. ..g- .q e 4 95 1 Figure 19ED-3 Channel Flow Height Reduction During Freeze Process 19ED 28 Corium Shield-Amendment 31 rnt: ;M -?;!-;ts ---------

;-et-;;

2 7. : r a: r;

FEB 08' '94 '~ 06:56Ft1 G E NJCLEAR BLDG J P.1/22' 9 \\\\ ga GENuclewEnergy ABWR Date 2 h/9A To C kh 70 v L, a n y Fax No. This page plus _53 page(s) From Jmek l~oy Mali Code 7 BL 175 Curtner Avenue San Jose, CA 95125 Phone (408) 925 4 9 h 4 FAX (408) 925-1193 or (408) 925-1687 Subject Oge-T4e - F \\ 9 '2 3. 3. e. 3 ( 3 ST Pu h) Message l FK!' 4 : ? - ; ; 5 - ; i e 7, -. - - - -- 77.,79 94 jy. 7_ .s. ,}}