ML20059F475
| ML20059F475 | |
| Person / Time | |
|---|---|
| Site: | 05200002 |
| Issue date: | 03/29/1993 |
| From: | Espinosa R, Johnson S, Shesler A ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY |
| To: | |
| Shared Package | |
| ML20059F463 | List: |
| References | |
| NUDOCS 9311040322 | |
| Download: ML20059F475 (7) | |
Text
- - - _ - - - - ---------------- _ ----- -
APPLICATION OF ADVANCED FULL SCOPE SIMULATORS TO SEVERE ACCIDENT PRECURSOR TRAINING R. J. Espinosa, S. E. Johnson, A. T. Shester, R. L. McBeth ABB Power Plant Controls 1000 Prospect Hill Road Windsor, Ct. 06095 Presented at the 1993 Simulation MultiConference March 29 - April 1,1993 Washington D. C eSA* $0$$$ oggga[>o2 o h A PM y
APPLICATION OF ADVANCED FULL SCOPE SIMUIATORS TO SEVERE ACCIDENT PRECURSOR TRAINING R.1 Espirxwa, S. E Johnson. A. I Shcsler, R. L McBeth ABB Power Plant Controls 1000 Prmped 1 iib Fxxx!
hisor, Ct. OGy>5 ABSTRACT The Core Neutronics Model Typical nuclear full scope simulators are built such ihat they he mre neutronics model is a three dimensional neutronics detect conditions during transients or events when the range of calculation shich is tesed on mMified one-group diffusion theory.
a;plicatility of their models is exoceded. When such conditions are Phenomena modeled by the core model indude (Stesler et al.1988, encountered, an instructor station alarm will alert the instrudor and and Bollacasa and Shesler 1991)-
the remmmended action is to terminate the training sessim. %is paper looks at the potential for applying simulators with less restric-
- De fission reaction induding the produc6on, transport, and leak-tive limitations to severe acddent precursor training; specifically, the age of neutrons in the mrc.
problem of mcdcling serious acddents which enter into the severe acrident category is addressed from a training point of view, it
- local reactivity due to changes in fuel temperature, coolant den-describes first a mcdel which has been sucressfully used for this sity, and coolant temperature, and due to changes in scram and purpcr.e. It then desentes spedfic features of this mcdel and assump- control rod positions.
tions which allow the model to le used for the severe a<xadent scenario. Finally, it illustrates the application of the model using an
- tocal fuel rod temperature induding the effed of fission heat, actual serious accident scenario which was run on a full scope decay heat, heat prtduced as a result of zirc-water reaction, and simulator. heat transfer to the coolant.
IEEDUCTION .
Ac spat'a! and ccre average effects of fission product poisoning like xenon density.
Existing state-of-the-art simulato- J.necrporate models which .
have predictive capabilities for serious acddents, such as loss of .
De spatial and time dependent effects of start-up and sustairer coolant accidents (LOCA) with core uncovery. Ilowever, there is an neutron sources.
increasing demand for simulator models which have predidive capa-bility for accidents which enter into the severe accident scenario.
- De spatial effects of burnup and refocling as seen by several Severe ancident scenarios are those for which major fuel damage discrete times in the fuel cyde.
cecurs due. to high fuel and clad temperat.ucs.
- Tbc process values observed by in-core thermocouples, movable he ph sics3 of severe accidents, which are postulated to detedors, and fixed detectors, and by cx-mre source, intermediate, progress through everal distinct phenomenological stages including and power range detedors.
core meltdown, con geometry change, debris transport, and carc containment concreti. interaction, is extremely mmplex. %ere is
- Fud failure and clad oxidation due to high dad temperature, general agreement that ic-*alling a complete severe acddent capatil- induding fission product release and hydrogen production due to ity on full smpe simulators is not only unfeasible, but unnecessary the zirc water reaction.
and probably unwise.
Each g oup of fuel assemblics corresponding to a control rod
%c ' vndary between the senous accident category and the is modeled separately.
severe one is not a line, but a grey area; thus, there is a wide area of the severe accdent scenario where application of the emergency he three dim ~ht neutronics equations are solved using operating procedurcs is stiu an option for the operators. In order to the principle of space time separation.
provide adequate training under these conditions it is necessary to, depend on models which will represent as accurately as possible the The Recirculation ModelThermat-Hydraulics part of the physics of the severe accident which is important aid observable in the control room.
He recirculation model is a one-dimensional multiple node THE ABB S!MULATOR CORE AND repres nta m fthereadorprmuevessel,MrmlaMnes, steam imes, and steam hoe manifold. A typcal BWR/4 system is repre-THERMAIMDRAUUC MODELS sented useg the node-junction representation shown in Figure 1.
ABB Power Plant Controls has developed ard implemented The model uses a five-conservation equation thermal-hydrau-advanced core neutronics and system thermal-hydraulic models in lie formutation for the liquid 4 team system representation (two mm reat time simulators for both pressurized water reactors (PWR) and two energy, one mixture momentum). De mass and enerEy equa-bciling water readors (BWR). Ac roodels are operational in fuD tims are solved for each system ncdc. De momentum equation is scope simulators, comped simulators, and nuclear plant analyzers. solved for each of thejunctions. De two-mass, two energy equation De physical models are generic to both PWRs and BWRs. De BWR formulation permits a full repesentation of non-equi!Ibrium condi-version has been scleded as the reference simulation for subsequent tions in all parts of the system. As sudt, coexistencz of subcooled detailed dWhn_t
liquid and superheated steam can be represented in all system mm- Heat Transfer to Steam ponents.
When the core unmvers the fuel rods and steam superheat he mWel includes a complete constitutive relations package along the uncovered section of the core. He cooling mechanisms for for the calculation of the interphase mass and energy transfer.
thew conditions are heat transfer to steam ard radiation cooling.
De liquid-steam phase separation calculates the heteroge- Simulation of the effects of core uncovery is the minimum neous vpration ofliquid and steam in all parts of the system- icquirement for any serious accident simulation. He effect of core uncovery is usually visible in the control room through level, temper-he momentum equation includes elevation and pump head ature, and radiation instrumentation.
terms, form and frictional losses, momentum flax, jet pump delta pressure; it is solved locally for each junction. He ABB simulator models the complete phenomena of core uncovery. He core heat transfer mcdel has a complete heat transfer he pure and average conditions are calculated locally for package which rnodels heat transfer to steam using conclations from each node from the two mass and two energy values in each node- the literature. He model supports the calculation of heat transfer to steam over a wide range of temperatures by means of a static mWel ne integration of the mass, energy, and mcunentum equations which uses fluid and materia! paper'ies with a very large domain.
is an implicit integration of the linearized, discrete mnservation equations. %c impticit integraGon is a stable and convergent inte- Cladding Behavior at High Temperatures gration technique (Espinosa et al.1987).
'#8 I# E "* * * #* E"* #***
In addition, the model sofves conscrvaGon equations for non-condensibles (hydrogen, nitrogen, and oxygen) and radionuclides,
". * ^* h* E
"#E E*" * "E heat. "The
- U.*dat."#"#* Y ***
oxi ion rate increases with temperature. Analytical and calculates the distribution of noncondcasibles and radionuclides throughout the system (Espincsa et al.1991).
'" 'I' * *,# I" ** " '" "" * ' * * , P# **
s extremely complex due to the availability or non availatility of The Containment Model steam to sustain the oxidation, or the ability to maintain an intact configuration with expmed cladfing and extensive surface area.
De pimary containment mo$el is a multi;de node repesen-Simulation of the zirmloy water reacion is important from the tation of the drywell and wetwell or torus. A typical Mark I mntain-control room training perspective since the pesence of hydrogen in ment is represented using the nodal representation shown m Figure
- 2. he mc>$ct explicitly repescots multipic regions withm the the drywell and torus is visible 80 the operator,it saivates alarms and can cause mntainment repressurization. In response, the operators drywell, within the vents from the drywe!! to the torus, and within the are required to take action like actuation of hydrogen recombiners or torus (Choung and McBeth,1991).
venting of the containment,if required.
De model solves conservation equations of mass and enerl7 for each of the nodes for liquid, steam, and noncondensibles. %c he core model calculates the local zircaloy-water reaction using the local ekdding temperature and mcdels from the literature.
noncondensibles include hydrogen, oxygen, and nitrogen. De con-Ac core model assumes that an intact mre configuration is main-centration ofcach noncondensible constituent is explicitly calculated.
Discharge of liquid or steam through leaks from the recuculation tained. %e zircaloy-water reaction is locally interrupted when the system into the mrresponding nodes is explicitly roodeled, complete cladding region has been oxidized. %c bydrogen gener-ated on the mre surface is released to the vessel. Den,it is transferred from the vessel to the containment or to the mnnecting systems Local temperatures are calculated for each of the nodes. De,r- through the open paths or leakt raedynamic non-equilibriurn betweca the liquid and gas phases is medeled. %crmal stratification in the drywell is also mcxleled. Heat ne recirculation and contamment models calculate the effect scurces and sinks from the wa!!s and across the air cmlers arc of the noncondensibles in the pessure calculation. He containment included.
model also calculates the local distn'bution of nonmndensibics within A modensation model of the steam discharged into the liqmd the drywell and torus. Rus, local alarms detecting the presence of hydrogen can be activated.
of the torus is implemented.
%e total pressure is calculated from the steam and non-cundensibics partial pressures.
As the core Seats up, the dadding loca!!y ha!!oons, ruptures, he radionuclides are partitioned such that the gaseous iso- Partiany a completely blocking the path for the coc8 3nt, and opens topes evolve completely to the vapor space, and prticulates remain a path for the release of tbc gap fission poducts to fu molant. De entrained in the liquid phase. Italogens are partitioned to both the radionuclides whidi are released from the fuel roc. are then trans-vapor and liquid phases. ported to all the system components whida are connected and have an open path from the vessel.
PHY9CS OFTHE SEVERE ACCDENTSCENARO Re physics of fuel failure is usuaDy represented in the form his sedian describes important physical phenomena which of correlations, function oflocal mnditions such as system pressure, coeur daring severe accidents (Fauske & Assodates 1992). It dis. rod temperature, and several additional parameters such as beating cusses the significance of the individual specific phenomena from the rate and circumferential temperature gradients in the rod. Experimen-control room perspective, and finally discusses the implementation tal data usua!!y exhibits a large degree of scatter.
of the model in the ABB simu!ator.
The principal thermal effect of dad ballooning and rupture is that it changes loca!!y the gapconductivity, heat transfer area, channe!
resistence to flow, and thus, coding flow and heat transfer coeffi- Quenching and Sys1em Repressurization cients. A large degree of uncertainty exists in the determination of the magnitude of these parameters. Implementation of actions to cool an overheated system and r t
to remver from an acrident may result in significant repessurization Simulation of the relcase and transport of radionuclides are the of the system due to rapid stcam formation. Such actions can bc most important phencunena from a training perspective since the restart of reactor coolant pumps in PWRs or initiation of PfCS operators fully depend on alanns on the control room panels to injecion in PWRs and BWRs. ne process of quenching is again a determine excessive concentration of radionuclides anywhere m the very complex phenornenon and certainly is a function of whether or 9 tern. not the cure has toen damaged. In addition, the direction of core qucnching, w hethcr predominantly from the top (e.g., due to core ne ABB simulator ignores the efTect of dad ballooning or spray or upper head injection) or from tbc tuttom, presents an rupture on the heat transfer calculations. His only has an effed on additional dimendonal to an already complex phenomenon.
the magnitude of the clad temperature, which in any way is a param-eter which is tot directly avaibble to the operators in the control he act of quenching a superheated mre produces pressure f room. De efTect of cladding rupture on the fission product release and temperature responses which are visible in the contrd soorn. It is fully modeled. %c model calculates kral fuel failure as a function is a key physica, phenomenon which must be modeled by full scope of system pressure and local rod temperatures, and calculates the simulators. F release ofradonuclides tothe coolant. Ac trampartofradionuclides throughout the rest of the system is then calculated using flow he ABB simulator's mphisticated thermal-hydraulics and netwcxk calmbtions which include nold radionuclides concentra- heat transfer modcts calculate a first principles' response of the 140 0. system during quenching of the core. Again, the assumption that the core geometry remains intact aficcts the rate of total quenching, but Fuel Behavior at High Temperatures the physics remains the stune. But since there is a large degree of uncertainty in the speed of quenching due to the complexity of the When the state of core degradation reaches temperatures suf- physics, a first principles' resporne like the one which is calculated ficiently high for mciting of the reactor fuel, the core material is is extremely useful for training in full smpe simulators.
expected to slump and compact the core geometry, and possibly anitiate a process of tdocation of the matenal within the core or within c m A i the reactor vesscL he process of fuel melting and relocation is a very FUL1 SCOPE SIMULATOR complex set of processes. different materials in the core melt at dificrent temperatures and materials can retxate either heteroge- A severe type accident scenario was run on a full smpe neously or homogenous!y with other molten substances. De snotten simulator. This case illustrates the capabilities of the model and its debris can then solidify, obstruct the coolant channels, thereby blod- rbility to successfuUy execute transients whidi go beyond the enve-ing the flow of steam and influencing the rate and extent of hydrogen 5 ge normally associated with major accidents type. It also generation. Le molten debris relocated to the bottorn of the core wcmonstrates the valuable academic and practical training which can I would continue to heat up and would eventually lead to the faI!ure of be achieved by running this type of transient on full scope smutators.
the mre support structures. As the molten core debris fa!!s into a lower plenum containing water,it can produce significant amounts He transient is a large break IDCA in one of the rectrculation i of steam and an inarcase in the rate of hydrogen production which lines in a BWR. After the system has recovered at the end of the l can result in pressure spikcs and, possibly, steam explosions. Accu- blowdown due to actuation of the emergency core cooling system j mutation of molten core detris in the lower plenum can then result in (ECCS), it is assumed that ECCS is lost. This produces mre uncov-thermal attack of the vessel wa!! by the high temperature debris. cry, major core superheating and core damage. ECCS is subsequently Finally, if the molten debris penetrates the vessel wall completely it restored and the phenomenon of quenching with the consequent will initiate a process of concrete crosion. system repressurization is demonstrated.
Scenarios leading to the above processes are what have been he specifics of the acddent scenario are as foUcus-traditionaDy called the severe ao:ident scenarios. he physics of the obove processes is extremely complex and can only be formulated in = 0 Large break LOCA initiated j e hypothetical manner. Since it is not possible tolearn details of fuel configuration from available plant instrumentation, there is little t= 5 minutes ECCS terminated j incentive to model the physics of fuel melting and core relocation on training simulators. t = B0 minutes ECCS restored De ABB core and recirculation models assume that the core, t = 180 minutes Transient terminated reactor vessel, and vessel internal structures remain intact during the 2
severe accident scenano. His implies first that the core can always De dynamics of the model is illustrated by key parameters be flooded (if it rew ets). Second, simulating a severe acddent pre. like dome pressure (Figure 3), total system mass (Figure 4), maximum cursor event with the intact geometry assumption affects mostly the clad temperature (Figure 5), drywell pressure (Figure 6), and hydro-magnitude of key processes which are simulated (heat up rate, oxida. gen concentration in the drywell and torus (Figures 7 and 8).
tion rate, etc.)and not the direction of the response. Furthermore,any mcdel of any accuracy and sophistication will also han a very large Initia!!y the system depressurizes to containment pressure i degree of uncertainty in the accuracy and timing of the calculation of (Figure 3), and the mass inventory in the system (Figure 4), and level processes related to fuel melting and relocation effects, and thus in reccwcr due to ECCS aduation. %c drywell pressure peaks carfy the cakulation of prameters which are observable in the mntrol during the bicwdown (Figure 6).
ruom.
When ECES is interrupted, the mass inventory ditys (Figure 4). Core uncovery occurs followed by the onset of core superbeating
i (Figure 5). When the cxte reaches temperatures of the order of 1503 SUWARY deg F (Figure 5) hydrogen begins to le generated in significant qumtitics due to the zirc-water reaction, and the hydrogen concentra- his paper has demonstrated first that the ABB full smpe tem in the drywc!I and torus increases (Figures 7 and 8). By the time simulators can be uscd for applications which go far beyond the the core reaches approximately 2200 deg F sufficient hydrogen is traditional tounds assumed for advanced full scope simulators. De being generated to start producing a pressurization efTect in the loundsduring accident analysis arc not those of accidents with simple system (Figure 3). core uncuvery but those which enter into the severe accident scenario category. Although much of specific phenomena usually asscriated During this time interval, the high radiation and hydrogen high with severe accident scenarios is not modeled, such as fuci deforma-concentration alarms would have alcried the operators of the state of tion and mciting, vessci failure, corium formation, the important i
the cue. Actions spccified in the emergency operating procedures phenomena from the training point of view are repre.sented, such as j (EOP)can be executed dunug training sessions. As the wre contin- fuel failure, dad oxidation, stility to run at very high fuct and clad ucs to heat up, ahhough clad temperatures are not availabic to the tcmperatures,and core rewetincluding system reptessurization. %c operators through the panels, and the core exit thcrmocouples instru- physics of the sevcir accident scenario is extremely complex, and the i ments may have been damaged, these parameters are still available uncertainty of any calculation is very large. He assumptions whid: '
to the instructor through the instmctor station, so they, and any other have been implemented are such that they permit the training of paramctcr associated with the system can be made available to the operators during scenarios which can be considered to be of the severe operators or trainees, so that they can corrrlate what they see on the type by mcdehng only the phenomena which is important from the pancis, uith what is actually happening in the corc. training point of view.
Fotlowing ECCS restoration as water re-enters the core, large REFERENCES
, amcents of steam are generated which produce a system repressuriza-tion (Hgure 3) as well as a major change in the conantration of steam Choung K. C., R. l McBeth,1991 *A Primary Containment ard noncondensibics in 1e containment (Figure 7). Finally, after the AfodelforaBHRSimulator*,1991 Society for Computer Simulatico a
cxnc cluenches, water exits the break into the drywell, stcam con. Multioonference, New Orleans louisiana.
denses on the cold ECCS water, and the pressure drops.
Bollacasa D., A. T. Shesler,1991, *A 3.D Advanced Core his transient was run essentially without any pmtulated Neutronics AfodctforSimulatorf,1991 Society for Computer Sim- )
- operator actions, except for those taken to restart the ECCS system. ulation Multiconference, New Orleans, Imuisiana.
A closer analysis of the reference event, however, shows the need for operator participation in control and recovery of the event. In fact, a Espinosa R. J., S. leichtberg, P. K. Doherty,1987.
- Fast and blind actuation of the ECCS system without direct involvement by Supcr.FastNSSSSimulatim Using implicit Fiw Equation Afodels*,
the operators can cause, as did occur during the reference event, ne International Topical Meeting on Advances in Reactor Physics, unacceptably high pressures in the containment during vrwet of the Mathematics and Computation, Paris, France.
system which could jeopardize the integrity of the containment. For evolutions of the type simulated during the trfertnce event, operators Espinosa R. J., D. Trumpier, C. DeVoe, S. K. Sim, S. E.
must keep the systcm under control either by contro!!ing if Wble Johnson,1991. *1mplementation of an Advaneca NSSS Simulation the rate of rewet of the system, or by venting the containment. Afodelin the CE7R4N Environment *,1991 Society for Computer i Simulation Multiconference, New Orleans, Louisiana.
Simulation of manetreers of this type can be a valuable mm- !
ponent of operatar training. First, they broaden the experience of Fauske & Associates, Inc.,1992. 'Sewre Accident Afanage-operators in gaining direct experience with the panels during events ment Guidance TechnicolBadr Report, Volumes 1 and 2*, Electric ,
which are totally foreign to their common experience. Semnd,it is Power Research Institute, Palo Alto, Califomia.
i e remaiable tool for serious or severe accident instruction and j training, since it gives the operator an rpportunity to gain direct Shesler A. T., S. G. Wagner, G.12nhart, M. M. Tolmazin, 1 hands-on experience with the physics of the scenario. Such simula. 1988. 'A Three-DimensiomlBMW Core Neutronics Afodc1for Real j tor.hased t aining can reinforce knowledge gained through class. IIme Simulatim*, XIV Spanish Nuclear Society Meeting, Marbella, rcom. based severe accident training. Sp:un.
I l
1
. .._ .- - ,.-.n -. .i s .- , _. _-
J
}
5 1
i i
l s4 Csup amma j
.s-as sh SCACTORtgg I (3) l m Menamu 6 b ese itsPW sacesCau seggg i .. _ . . . _ .
g i i.
m -- n 1 1
.. M ..
- .. m i.4 ___ _ ;
.* mactess !
4en . .... +7; 70-1 1-sea ww7s e n.rt wacts
- 1. . .
se a,s. . es ,. .
,, ,,3 9g,yg ,,,g,
.m m.~ !.
- ** # war s*** maa ,
j.
_ ._, _ ... .. Lp ;
am mmm e-l (83 ;- ij (5)
- f. ..
W-~
ei . . . . . . - -
0 ::
i l
kl *
. ... .. . .. .. .: ;is 7, [.-_- ] , . . <
i a
L. A 1, Lr& .
.....~...: ;- L..........
3* t \
I I :T L I I ir> ! "'
!! t 's I
,. t .. ; it --
.: :: I o
" l- . I <
, y= = n '
<c.-- . - .q - -
e T. 1
,,, ....t..:t...l..
u 9.. :..:.... (*i i
i
- . - i 10f t I (12/13 1 i
Figure 1: NSSSNodalization Figure 2: Primary Containment Nodalization i r
t
\
umm . nonao ao 1000.00 ; -
l i
i smaanso 2 "
l
. l
- m - : !
- iit : 1
{ E .mmemo i % C '
- E
'coJo i 7 5t
- $ smomm - i
~ 2 l
G eacm i ;
- mm 2 ;
I socio i sooooomo j t i om ..
- L -
..,s....
em. .....
. ..... ....... ...... . - om ...acom a4 o som incm uom ...amm Time (Min)
Time (Min)
Figure 3: Domelhssure Figure 4: Total Mass in the Systesa
.J
I
.. - l l
l I
,, sec co I i
i 1
i m; ix oo l u
cn e W y M S
E e a uom - 3 som :
16 o B : 8 a -
E Q-e -
H -
mcm : ** b cm .. ..... ...... , , , , . . ... Em . .- .. .. . ' . .....v. ..... .
c oo so oo ioam iso m n ao em som ioom tw oo ... room Time (Min) Time (Min) l'igurr 5: Maximum Clad Temgmtuar Figun 6: Dry
- ell 15tssure o so em .
c O
. T
- O .
3 om - 3 o.is -
E -
8
- u. u. "
e (9 E 8 -
2 - 2 -
8 on : 8 nie : "
O
- O c . c O O g H . E 1
- .5 D 1o Om C C o
. O O - 0 l 1
cm . .. .. .. ...... .. .. ....
Om . ... .. ............ ... ...
om.. sem imm isooo amm om som imm inom ... zoom ,
Time (Min) Time (Min) l i figurr7: DrywellIlydmgen Concentration figurr 8: Torus Ilydrogen C<mcentration I 1
i l
l l
1