ML20214W311
| ML20214W311 | |
| Person / Time | |
|---|---|
| Issue date: | 07/29/1984 |
| From: | Meyer R NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Silberberg M NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20213E209 | List:
|
| References | |
| FOIA-87-113, FOIA-87-60 NUDOCS 8706150205 | |
| Download: ML20214W311 (31) | |
Text
,
JA43-4tdb 6 2 ]$$
e' All - fff 4
. e t:s:: ss
~[,,..
- eUO.Ea : E G U Lt-TO D Y CC'.".*!SSIC'.
jQ.f,;
..as
.ct:. o : :csn
., ~ %;r [l
. -u _
i yui n neu NOTE TO:
Melvin Silberberg, Assistant Director
~
Accident Source Term Program Office, RES FROM:
Ralph 0. Meyer Accident Source Term Program Office, RES
SUBJECT:
REVIEW OF NRC AND IDCOR MELTDOWN MODELS The meltdown models in NRC's MARCH code and in IDCOR"s MAAP code have turned out to have.a strong influence on containment loads (via H, production) and on source terms (via types and quantities of materials in the melt). These meltdown models are complex, and their output appears to be controlled as much by user input as by code calculation.
Therefore, I have tried to under-stand in some detail these models as they were used for the BMI-2104 analyses and for the IDCOR standard problems.
The models are described in Enclosure 1.
Hopefully, this more detailed examination will help us reach conclusions about adequacy and focus new work in areas where it is needed most. Some. con-clusions are already apparent from reviewing these details and these conclu-s1ons are discussed in Enclosure 2.
/
e
'A,.,. ' ' '
. r Ralph 0. Meyer Research and Technical Support, ASTP0 Encicsures: As stated CC:
G. M.arinc R. Car.r.ingna:
/R. Eernerc AST:C 1.aff FSRE Staff I
i 8706150205 B70610 PDR FOIA SHOLLY87-60 PDR
~
=
-a..
I.
Descriction of NRC's MARCH Meltdown Model The meltdcwn model in MARCH in contained largely in the BOIL subroutine, but many other subrcutines (e.g., MELT, QSLUMP, ZRWATR, and HEAD) contain pieces of the model.
Figure I shows the major features of the model as it was used for the SMI-2104 analyses.
A* step-by-step commentary is given below.
Step 1 s
Switching ~cr the cladding oxidation model when a node becomes uncovered is not a profounc' occurrance.
It is simply unnecessary to run through this part of the code when a node is under water and the temperature is too low to produce significant oxidation.
Three different oxidation-rate models are available in MARCH and they are discussed in the MARCH-2 users manua1 The BMI-2104 analyses were done using the Urbanic and Heidrick model (MWORNL = 1 in Subroutine ZRWATR), but the choice of exidation-rate model makes almost no difference in the calculated results.
Ste 2 C
A value cf 2277 C was used for the fuel melting temperature (TMELT in Subroutine SCILNT) for the BMI-2104 analyses. The QUEST uncertaint study at SNL has shown that TMELT has a strong effect on overall source term results U
and that THELT may range from about 1925 C to 2475 C based on liquefaction observations. This user-input value is thus very important.
There are three meltdown models in MARCH and these models are summarred in the user's manual. Meltdown Model A (MELMOD = 1 in Subroutine MELT) was used for the BMI-21C4 analyses, and Model A by assumption prevents nadal temperatures
-*T
~
=
If node (1, j) is above water, cladding oxidation model is turned g{c c o] l, on-S te p' ' T
'i -
Q.5 J
If T(1, j) = liquefaction temp.,
node (i, j) is considered molten.
4 T(i, j) not allowed to exceed TMELT.
Oxidation of cladding con-3 j
tinues at same rate.
0 2277 C (TMELT)
Step 2 1
Y A
y If a bottom node (i,1) is molten, I
it and all molten nodes above it
(,,
fall from core to ist support grid.
Ik' Step 3 GE.ZL si f,
5(
~~
Q' Cladding oxidation model in dropped i
a n.n,
nodes is turned off.
" Drop" oxida-tion model is turned on for one time step only.
Step 4 If 1st grid temp. = M.P. of steel, molten fuel on 1st grid falls to l
2nd support grid.
1400 C (TFAILX)
Step 5 i
1,,x = 10 if If total amount of molten fuel ex-d 24
=
max ceeds specified amount, entire core falls into Bottom Head.
75% (FCOL)
Step 6 V
Fig. 1.
Diagram of meltdown model in
' If bottom head tensile stress NRC's MARCH code as used in exceeds tensile strength, all core debris and specified amounts of most BMI-2104 calculations.
steel fall out of vessel.
Step 7 e,
.,..-w,.
-,.-rw-r----c--
---vr
~ " '
-~* ~'
from exceeding TMELT except under unusual circumstances te be ciscussec belcw l
in Step 3.
Thus the user-input TMELT not only deterrines wner melting occurs,
~
~
l but also determines the maximum core temperature under most conditions.
Three user choices-governing oxidation are available for fuel in a moEen node:
(1) no metal-water reaction in a melted node, (2) no metal-water reaction above the lowest melted node in a given radial region, and (3) metal-water reaction in a melted node calculated as if fuel were still in its rod-like geometry. The latter of these options (IMWA = 3 in Subroutine BOIL) was used for the 'BMI-2104 analyses on the basis that it is just as easy to imagine enhanced oxidation on melting (exposure of fresh Zr and enhanced surface area due to wetting) as it is to imagine reduced oxidation (due to slumping and channel blockage).
IDCOR choses the equivalent of IMWA = 1 for PWRs and IMWA = 2 for BWRs thus leading to a major difference between IDCOR and NRC results.
It should be noted that these choices (IMWA = 1 and 2) imply complete blockage of a large rad,ial region of the core since a single fuel rod channel is used to represent a radial region in the MARCH and MAAP codes.
Step 3 Within meltdown Model A there are also user input values that deter-4 e w en h
molten fuel will fall from the core region onto the first gric su:::n riste (this process is often referred to as fuel slumping). One of these cre'ces (NDZDRP = 1 in Subroutine MELT), which was used for the BMI-21Cc a t'
- er, does not permit molten fuel to fall out of a node in a radial reg':r
.--il the lowest node -(j = 1) in that radial region is also molten.
>'s :-::e n re is illustrated in Figure 2.
The QUEST uncertainty study has sha.
ra re-sults are insensitive 'to the choice of NDZDRP because oneg sene #uel f alls on the first grid plate, the whole core will be in the bottom head within a few time steps (a few minutes).
Another user input in Model A specifies a minimum core fraction that must be molten before fuel can fall from the core even if the bottom node (j = 1) is molten. This fraction was set to 2% (FDROP = 0.02 in Subroutine B0IL) for the
~.%
,m
([
y
(
I i
- i. (
3 1
e c e%3,
' Q p)
Q j
3 1
g U
2' y
i
(
i (d -)
(/sh 6)
Y 1 l I
i i
A
[
r i
Fig. 2.
Illustration of criterion NDZDRP = 1 for molten fuel relocation.
In (a), a central node (1, 3) 3 has become molten.
In (h), temperatures have 2.
increased and nodes (1, 2), (1, 3), (1, 4), and v
i (2, 3) are molten.
In (c), nodes (1, 1), (1, 2),
W (1,3),(1,4),(1,5),(2,2),(2,3),(2,4),
g j
and (3, 3) are molten.
Since a j = 1 node (1,1) is molten, all nodes above it fall from the core.
~.
BMI-2104 analyses.
Inasmuch as rcug-ly 55' f the core bec:ce.s rel:er. before
_;3___
NDZDRP = l is sctisfied (about 28': fer NO 0FF = 6), FOROP va'ues ir the range of 0.02 do not affect fuel movement.
The QUEST uncertainty study concluded that FCF.0F was an important parameter, contrary to the above discussion.
However, the sensitivity to FOROP did not arise directly from the fuel slumping behavior, but arcse from a temperature
- effect alluded to in Step 2.
In the OUEST study, FOROP was set to an extreme value of 0.75 (75% of the core molten), well above the roughly SSL value that would correspond to Figure 2c.
Thus, in the OUEST claculation, molten' fuel was retained in the core even though it was molten to the bcttom of the core.
This seems impossible.
Yet the important effect was not geometric; it was that temperatures in the molten nodes could now rise above TMELT.
In Mcdel A, once the heat energy in a node ' raises the temperature to TMELT and then supplies the heat of fusion (hence the importance of TFUS fcurd in QUEST), excess energy-is passed downward corresponding to slumping of the fuel.
When the bot-
. tom node (j = 1) is fully molten and the fuel is not allowed to fall, the en-i
. ergy cannot be transferred upward or sideways, so the nodal temperatures rise above TMELT.- Hence the sensitivity to FDROP found in QUEST.really shows the
-importance of the thermal arrest feature built into Model A rather than show-ing a sensitivity to the mechanics of fuel relocation.
This, I believe, is an important insight.
c
~
Step 4 During the time ster tet: a civer molten nedes is drc:cing out of the core, the molten fuel is assu e: :: :e ir s:herical : articles tnat fall into water, and
- Zr is further oxic':e: as each noce falls.
Similar exidation-rate equations i.
are used as for cladding oxidation, and the equations are described in the user's manual.
Particle diamedter (DPART = 0.5 in. in Subroutine MELT) is a user input, but the QUEST studyJYn'I"Eesults are not sensitive to this value.
a W
~,
o e*-
-e-r-- -- I.e
.v-
-s,m
-n
-e r
&w.
w-----e-
-e-
The spherical particles cay be eithe" 50megerecus in ccm::csition er they may be composed-of uc to 3 stratifiec zones (a particle core and twc shells). 'For the BMI-2104 calculaticr.s, a homogenecus particle was generally assumed, i.e.,
all of the unexidi:ec Ir (FIMCR = i in Subroutines MELT) and all of thJt Zr02 (FZ0COR = 1 in Subroutine MWDRP) were in the particle core and there were ne shells. The QUEST study included a contrasting case in which all of the unoxi-dized Zr was in an cuter shell (FZMCR = 0) and fcund'that this user option had a strong impact on overall source term results.
Still more user options a*re available in this metal-water-reacton model to force the reaction to scme specified value ranging from zero (no reaction in bottom head region) tc 100t.
For the BMI-2104 analysis, the model was allowed to calculate the degree of reaction (within the single time step) without constraints (FDCR = -1.0 in Subroutine MWDRP).
Step 5 When a group of molten nodes as shown in Figure 2c falls onto the first grid plate, their heat energy first boils away any water covering the grid and then starts raising the temperature of the grid. When the grid temperature exceeds
~
0 a user-specified temperature (TFAILX = 1400 C in Subroutine BOILNT) that c:rresponds to the melting point of stainless steel, the molten fuel on the 4 *st grid falls to the second support grid.
This heatup is rapid and failure cc:urs in one or two time steps so that the precise value of TFAILX is U*ii:C" tant.
he: 6 Ar.y one of three user-input values can signal the collapse of the entire core (molten nodes plus non-molten nodes) into the bottom head.
(1) In PWRs, the core.is supported by the core barrel; therefore, when the core barrel temperature exceeds a specified value (TFAILB = 1400 C in Subroutine BOIL),
the core collapses into the bottom head.
(2) In BWRs, the core is supported by control rod drive tubes recresented by the second support grid in MARCH;
-I therefore, when the secene gric tecperature exceeds a specified value (TFAIL2
= 1400 C in Subroutine BOIL, the core collapses into the bottom head.
(3) For.
~
~
n
either pWPs or EKFs, tnere is an intuitive assumotien that a very large molter-mass will cause core collapse.even if the above temperature criteria are not
,_~ ' ~ ~
met; tnerefore, if the total melten fraction exceeds a specified value (FCOL =
0.75 in Subroutine BOIL), the core collapses into the bottom head.
In the BMI-2104 calculations, FCOL determines core collapse in all of the PWR
/
cases and most of the BWR cases with TFAIL2 determining collapse in the balance of the BWR cases.
Since core collapse occurs in just a few time steps after core slumping (Step 5) regardless of the controlling criterion, tne particular core collapse criterion does not appear to be important.
Steo 7 Failure of the bottom head is based on strength criteria rather than temperature criteria.
Tensile stresses and tensile strengths are calculated' in the HEAD subroutine from geometry, temperature, and materials properties.
The models are described in the user's manual. When the calculated tensile stress exceeds the calculated tensile strength at any location, failure is assumed. The total core debris inventory is then dropped into the reactor cavity along with the support grid and molten portions of the bottom head steel. Several input values (such as WGRIDX and FHEAD in Subroutine HEAD) determine the amount of steel in the molten corium. This corium mix was as-sumed (also user input) to have a particle size of 0.5 in. (DP = 0.5 in, in Subreutine HOTDRP) in the BMI-2104 calculations.
The QUEST study found that particle size DP is unimportant, but variations in WGRIDX and FHEAD change the amount of dilution with steel of the core melt thus altering the core melt tem-perature and varying the subsequent core-concrete interaction.
Thus the user-input variables WGRIDX and FHEAD were found to have a strong effect on source term resul ts.
II.
Description of IDCOR's MAAP Meltdown Model We know much less about the meltdown models in MAAP than we do about those in MARCH, so this' description will be less detailed.
Noding of the core region ir. MA*P, as shown in Figure 3, is similar to the MARCH noding scheme with the following exceptions:
(a) i,x = 7 (8 for the BWR model), and j
= 10, al-g max 7-m
--w n
_ 7 I
If node (i, j) is above water, D
cladding cxidation model is turned
[CC Dh on.
Step 1 V
If T(1, j) = liquefaction temp.,
4 oxidation is oxidation is also stopped in stopped above and node (1, j).
below node (i, j).
3 Channel blocked, q
L
Y If T(1, j) = M.P. of UO2, node (1. j) is considered molten and C-i
~
mass is placed in adiabatic " limbo" node.
sa g
[
N IQ,.
2850 C Step 3 M
If accumulated molten mass exceeds
~
specified value (50% in PWR, 20% in BWR), fuel from " limbo" node falls c.
gj l to support grid.
r--
Step 4 O
(1 PWR BWR After specified Grid fails, clock time, and rate of grid fails and flow down CR i
= 7 ( W R) drops fuel, drive tube is max calculated.
i
= 8 (BWR)
Fuel arrives in bottom head.
max d
= 10 max Step 5 l
V l
If temp. of instrument tube or CR drive tube = M.P. of steel, vessel Fig. 3.
Diagram of meltdown model 0
in IDCOR's MAAP code as 1527 C Step 6 i
l used for standard problems.
V
~
Rate of egress of molten corium is calculated based on size of pene-tration.
Step 7
~ x; g-
though apparently only 3. radial nodes were usec fer :ne stancarc cr ble.,(j',
~,-
the PWR core model in MAAP has only one grid support :! ate, arc (c) the EKR core model in MAAP has a rather detailed icwer plenum region in place of the icwer grid suppcrt plate in Figure 1. Figure 3 also shcws schematically the major features of the MAAP meltdcwn models, which are ciscussed below.
These models are located principally in the HEATUP subroutine of MAAP.
Step 1 Zircaloy cladding reacts with steam in the uncovered core nodes.
The rate equation of Cathcart, et al. is used below 1580 C and that of Baker and Just is used for higher temperatures. These equations are similar to the Urbanic and Heidrick equations used in MARCH.
Considering the insensitivity to small changes in oxidation kinetics, Step 1 in the MAAP analysis is equivalent to Step 1 in the MARCH analysis.
Step 2 In the IDCOR document it is said that, when the temperature in a node reaches the melting point of Zircaloy, oxidation in the node is stopped.
Stopping the oxidation reaction is based on the assumption that molten cladding will block off the coolant channels in the node such that steam car t: Icnger enter the fuel in that node.
The melting temperature of Zi,rcalcy "! ;'ver in the IDCOR U
document as 2123 K (1900 C). The code developers at FAI rave sir.ce raised that temperature and are now using a cutoff temperatu s :' :::- :. althougr they usually check results for cutoff temperatures u: :: :: -0 This revised temperature would correspond more closely to the licus'a:-':
- e. era:ure of U
Zircaloy-clad fuel (1925 C to 2475 C).
In the HEATUP'/PWR model, when the oxidation cutoff temperature is reached, steam and hydrogen generated below the blocked node are diverted to other unblocked chanels.
This is quite similar to the unused option IMWA = 1 in 5tep2ofthedescriptionofMARCH.
In the HEATUP/BWR model, when the oxidation cutoff temperature is reached, the BWR fuel channel box is considered to be blocked.
Thus oxidation is stopped,
Lin the affected ncce and.in all ncdes directly abcvs it: ec further steam cr hydrogen flow is permitted in that radial region ef 'the r: cal' scheme. 'This
' "~2 algorithm is similar to the unused option IMWA = 2 ir Ste: 2 of the cescrip-tion of MARCH.
In summary, then, MAAp produces hydrogen as the cladding heats up to a C
liquefaction temperature of 2027 C and then stops producing hydrogen.
MARCH produces hydrogen as the cladding heats up to a somewhat higher licuefaction U
temperature of 2277 C and then continues producing 'hydrecen at that temperature until the mefted fuel falls out of the core.
This is. a major dif-ference between the codes.
Step 3 When the nodal temperature reaches the user-specified UO melting temperature 2
of 2850 C, the node is assumed to be molten and the fuel mass is transferred to an artificial " limbo" node where the temperature (stored energy) rises adiabatically.
Inasmuch as steam flow, hydrogen flow, and hydrogen generation were terminated at the lower liquefaction temperature, moving this fuel into a limbo node has little further effect on the analysis.
f!'
s Step 4 When,the amount of mo'lten fuel in : e -bc noce reaches a user-scecified fraction (50% for PWRs and 20'. fcr bli. the molter fuel is crc::ed cnto the grid support plate.
Thus, while t.e 5: ar,ics of dropping fusi :c the grid seem very artificial in MAAp arc :.':e cifferert frer te.:se ir. MARCH, the result is about the same; melter. #ue: f alls to the grid plate wher. A to i of the core has melted.
Step 5
.In the PWR model, a clock is started when molten fuel is drepped on the grid plate. After one or two minutes (user input), the accumulated reiter fuel is dropped to the vessel bottom head.
_ ! nn,
,,.., m - ~.
m n nm
- +
L
[f Q
yk " ^"
t e
This part of the SWR moce' is ra:-
cre rechanistic arc is contained in subroutines FREEZE, ICRUST, are c_CL.
As socn as nelter fuel is droppe'd on
~
"I the grid, this weaker BWR gric is assured to fail anc a hele sire is specifiec by the user. Mciten fuel ficws trrcugh the grid plate hele and runs down, partially freezing as it goes are forming a crust on the drive tubes.
While the BWR model conjurs up images of molasses running in a forest, the result is essentially the same as for the PWR nadel; molten fuel reaches the bottom head in one to two minutes.
Step 6 Subroutine VFAIL calculates vessel failure when the tenperature of a CR drive tube (BWR) or an instrument tube (BWR and PWR) reach the melting point of carbon steel (about 1527 C).
For a Cerbustion Engineering reactor, which has no lower head penetrat. ions, the user specifies the time of failure.
Step 7 Subroutine VFAIL also calculates the rata of corium flowing through the vessel penetration, the rate of increase in penetration size due to ablation, and the rate of addition of molten steel to the corium from the enlarging penetration.
': cther steel (e.g., from the grid support structures) is added to tns t'
Initial penetration diameter is taken as the diameter of the failed CF. c 've tube or instrument tube except for Combustion Engineering reactors #c" </':
'.! is specifisc b) One user.
e.
.n, Conclusiens 1.-
The number of user options in MARCH appears staggering and leads 5 the feeling that the code user can get any result desired by simply altering in-put. This appearance is largely illusory, however, since most of the switches have been threwn as surely as if they were hard wired.
For example, the " user option" to centinue oxidation after the onset of liquefaction (IMWA =3) is now part and parcel of the NRC model and is not an option within the framework of NRC source term analysis.
Many other options (a) have been either permanently chosen on the basis of technical evidence (e.g., MELMOD = 1), (b) provide materials properties (e.g.,
TMELT, TFUS), or (c) don't really matter much (e.g., MWORNL, NDZDRP, FDROP, TFAILX,TFAIL2,TFAILB,DPART,DP).
If we did not characterize these code variables as options, the appearance o' arbitrariness might go away.
f 2.
All of the MARCH variables found to be important in the QUEST study (TMELT, TFUS, FDROP, FZMCR, WGRIDX and FHEAD) are in the meltdown part of MARCH, further underscoring the importance of this fuel behavior model.
3.
The sensitivity to FDROP found in the QUEST study is misleading inasmuch as the sensitivity is not due to the direct effect of material relocatien, but rather to the indirect effect of changing the limit on core temperatures (the thermal-arrest feature).
4 Up to the peint of vessel failure, the important parameters that remain uncertain appear to be (a) the melting point (or range) of fuel, (b) the associated heat of fusion, (c) maximum core temperatures (therma.1 arrest),
and (d) the rate of oxidation of melting and molten fuel debris.
By contrast, the mechanics of fuel relocation appear to be rather unimportant.
In this light we might question whether is is reasonable to expect important questions tc be' answered by SCDAP meltdown modeling, on which MELPROG and MELCOR will ultimately depend. The LIOSOL, L10 CON, and LIOSHR subroutines in SCDAP 12 -
.m cw s :
..,: s, u.~ -
> e v u.
c..e L
r 9
.. =,
calculate the amount and position of relocated fuel and cladding alcrg with
~
-the associated change in flow area and thus appear to address the less -
- a::---
important geometrical part of the meltdown phenomenon.
The important parameters listed above probably require experimental measurements and may not be ameanable to analytical deduction.
4 t
4 e
e-
-O e
I i
I et o
e 13 -
t 4
+
t
,o-ones ROUTING AND TRANSMITTAL SUP 4/25/84 m mome, esce symned. mem numw, inniens o t.
humming, Agency / Pest) 3 g,
Baranowsky Burdick-g Ernst a.
4, g,
cc: Cunningnam Action.
File Note and itetum f u _;:
For Cleersnee Per Conversation As E:;_ _ _^ f For Correction Propero Itepsy circulate For Your information See Me t- - - : ^
signature
"- ^ ^
Justify
^ ^^
AEMAR8LS As requested by Bernero, attached is an issue paper on Plant Categorization to be discussed at the May 15-17 NRC/IDCOR Technical Exchange Meeting.
If you agree with the content of the paper, please initial the signoff sheet and send it up the ladder.
Otherwise, return with consnents for modification.
A text editor will make the. editorial. corrections.
J Due date to Bernero is 4/27/84.
An issue paper on Quantification of Sequence Likelihood is forthcoming.
l l
00 NOT use this form as a Rf.Coltb of approwels, concurrences, disposals, esserences, and somner actions PROtt: (Neme, org. sym6el, Agency / Post)
Reem No.--Sids.
TONY C. ENG
%g cost-ter OPTIONAL FOfted 41 (Rev. 7-76)
'N"
- 11 #
- sto sees o - ses.srs (232) me se
o s
ISSUE 2.1.1 PLANT CATEGORIZATION 1.
Description of the Issue The NRC Severe Accident Policy Statement states that the existing plants pose no undue risk to public health, safety and property.
In support of this state-ment, ongoing nuclear safety prcgrams are being conducted to provide additional assurance that this is so. One method for providing additional assurance is to use the insights from existing PRAs and infer them to all LWRs.
This extrapolation of PRA insights to all LWRs necessitates that all LWRs be categorized into generic classes in order to enhance the manageability of the supporting research.
The NRC Accident Sequence Evaluation Program (ASEP) is extending the reference plant concept that is presently used by the NRC by incorporating more plant characteristics and using the probabilistic approach to plant categorization.
2.
Implications of the Issue to Regulatory Questions-
- The ability to develop and justify the grouping of plants has a great effect on j
the NRC regulatory decision process.
Since it is desired to confirm the policy statement on severe accidents on as generic a basis as possible for all operating i
and near term operating plants, the issue of developing plant classes is highly important. There is a direct relationship between the issue and the regulatory l
questions. The position of the issue can affect the answers to:
i How safe are the existing plants with respect to severe accidents?
How can the level of protection for severe accidents be increased?
What additional research or information is needed?
2
- r-Is additional protection for severe accidents needed or desirable?
3.
Subissues Is it possible to group plants having similar dominant accident sequence risk characteris'ics?
t Do the existing PRAs (12) provide enough information to make inferences on all existing plants?
Can insights drawn from plant categorization be used in severe accident rulemaking?
If so, how should uncertainties be considered?
Given the associated uncertainties, can plant categorization be used to identify weaknesses in plant design?
What level of information is appropriate for plant categorization--function, systems, or component level?
What characteristics of plants should be used in the categorization?
Can event trees for plant classes be formulated?
If so, how useful are they?
Plants are categorized based on dominant accident sequences and plant systems needed to mitigate the accidents, consequences resulted from these incidents are not considered. Does this have any impact on regulatory decisions?
What are the other potential uses of plant categories?
4.
Status of Understanding The NRC and IDCOR currently are using a similar deterministic approach to plant categorization.
A reference plant is chosen, usually a plant with a PRA, to
,-_.__.~__.--,m.c-
__r,-
.m..,,
~
3
. represent a group of plants that have similar containment design.
The reference plants selected by the NRC are:
Peach Bottom (BWR Mark I); Limerick (BWR Mark II); Grand Gulf (BWR Mark III); Zion (PWR Large Dry);' Sequoyah (PWR Ice.Conden-ser); and Surry (PWR Subatmospheric). These reference plants cover the contain-ment types of all the operating and near term operating plants.
The IDCOR reference plants are a partial set of the NRC reference plants and they are j
Zion, Sequoyah, Peach Bottom and Grand Gulf.
Plants with PWR Subatmospheric and BWR Mark II containment are not covered by IDCOR.
Since the reference plant concept depends solely on the engineering assumption that plants can be grouped by containmen,t type, the ASEP work extends the reference plant concept, uses a more systematic approach to plant grouping, and accounts for more plant characteristics besides containment type.
This produces 4
a hierarchy of plant groups that are formulated based on similarity of plant function / system response to the accident initiators, risk characteristics of the dominant accident sequences, and systems design differences. At each level,
' the plant groups can be further categorized by their containment types. The attached figure represents a possible plant class hierarchy.
ASEP formulates the plant groups from the " top down" and " bottom up" t
perspectives. The " top down" approach groups plants using the event tree methodology and the " bottom up" approach groups plants based on similar system design and similar system reliability characteristies.
The first step in plant z.
categorization is to divide the LWRs into PWR and BWR groups.
From the " top i
down" perspective, the PWR and BWR groups are subdivided according to findings O
e-
-,-v.--..,
1,-,--
..e,-g,,,e-,,,--.,
---..--,,--.----.m,
---m_..,,w.m--,ve,ywe
..--...-,e+,----c y-.m.~,..w,,-,---e-
4 s
ASEP PUWT CLASS.. HtE& AKCHY +
' f,m WS f /4M'T wA7ER.AEAC7083 (L.N$
l 7~WO TYfts of lwr $
Swg PwR i
PLA </T classes BA:EC CA/
gwR=
$/M/t AK k
k j usc77cA4+L REC / MdC gg, g L.l, 3.
- /~o T NI T*tA 7~tM$ C Vth'T3 I
I PLAntr cussses BAsco en own-sun sin. 9K sysreu ResPCA4E L1,5 a.1 4 To sftctfte /Nrr~ tart % qvtrir.:
i.
PLAur cLAsse: BASEp cw 0"L 6UL
$/M/LAA RI$X CHANAC7T/ITT/C ?
U'b d'4 CF Dw'r/ art 1~ Acciocur secusvec 4
SwA sun 8'd# " ""# #### #^#
Or vets TY 05 Dars t 4 Arg cp A'S ggy OcFEN5G -/At -OrpTIf l
$ ysTEM s A/ECESSARY TC HIT!4RTC 00M/A*A ss7" ACC/0CAf" s E G GEA/CE.5
- A+ cucl level t f Br="Jjs cap k.hdhp_ ss+eweisxf <% emamed s%aa h
5 t,-:..
generated in the development of functional event trees.
Functional event t'rees are developed for the PWR and BWR groups in response to the general initiating events of LOCA and transient.
Plant groups are formed at this level based on 4
the difference of the preventive or mitigative functional responses or the sequence end-states (success or core damage). ASEP does not expect this level of plant categorization to, result in many plant groups since over the past 15 years or so the philosophy of both PWR and BWR accident prevention and mitigation has not fundamentally changed.
That is, the needed functions are generally recognized as being the same.
To develop these functional event trees and 4
descriptions of the functions and sequences, ASEP analyzes and synthesizes the functionaleventtreespresentedinexisiingPRAs. At least four functional i
event trees are developed, two for each of the PWR and BWR groups.
One of each i
group is for the general initiating event of LOCA, and the other, for the general initiating event of transient.
Plant grouping at this level is not definite until the grouping at the lower level (e.g., systemic event tree approach) is completed since the total accident sequence delineation process is iterative.
l The next level in the " top down" approach is the grouping of plants based on their system response to the initiating events. As mentioned earlier, the philosophy of accident prevention and mitigation is generally reorganized as being the same for both the PWRs and BWRs, but the strategy to achieve this j
philosophy'is different among plants.
By strategy, we mean the systems to perform the functions.
Classes of plants are developed based on the system
)
combinations required to either prevent or mitigate a core melt ensuing a given specific initiating event (or set of similar initiating events).
l l
___a,__,_
6 Further refinement in plant grouping is considered since system dependencies vary among plants.
For example, containment spray recirculation at PWRs can be performed.by the spray injection pumps at some plants, low pressure injection pumps at others, and is a completely separate system at some others.
Another source of refinement is the possible different end states for a given accident sequence; that is, the sucgess at one plant may be core damage at another.
Plant categorization using the " top down" approach is not completed until the merging of plant classes from the " bottom up" approach since the subsystem characteristics of the lower plant class levels may affect the higher level plant groupings (e.g., the structure of the systemic trees can differ depending on whether or not main feedwater pumps are motor or turbine driven; in the latter, closure of the MSIVs or loss of the power conversion system causes failure of main feedwater).
l The " bottom up" approach uses twelve existing PRAs as the starting point to l
develop plant classes.
Dominant accident sequences are identified as well as l
l the plant systems and support systems needed to mitigate these dominant acci-l_
dent sequences. A plant survey is conducted to obtain accident mitigation fluid and electrical system for all the systems and support systems for as many plants as possible. After the system drawings were collected, they are simpli-fied based on past PRA insights in terms of major flow / energy paths, major active components, important passive components, etc. After the system had been simplified, generic system configurations are formulated by comparing all systems (e.g., compare all AFW systems) considering significant differences in the redundancy, diversity, or support system dependencies for each system of
7 interest.
For example, all AFW systems are grouped into 20 generic AFW confi-
-gurations.
After-the generic system configurations were formulated, plant groups are for-mulated in accordance with the diversity in designs of the key defense-in-depth systems necessary to mitigate the dominant accident sequences. The PWRs having certain design features in common were organized into 29 plant groups and BWR into 15 plant groups.
Some insights can be drawn from this level of plant categorization.
First, for a particular systen, the number of configurations for the surveyed plants can be large.
For example, 20 different AFWS configurations are identified for the surveyed PWRs and 7 RHR systems for BWRs.
Service water systems are found to be essentially all plant-specific. Without accounting for service water system variations, the differences in systems still resulted in 29 PWR pant groups (from 72 PWRs) and 15 BWR plant groups (from 31 BWRs). This is a dramatic evidence of the lack of standardized designs in the U.S.
i At the next level, the plant groups based on system design differences are coalesced based on their risk characteristics computed thiough accident sequence quantification, sensitivity, uncertainty, and recovery analyses.
These characteristics include the best estimate, their upper and lower bounds, and the dominant factors that drive the sequence likelihood.
e e
t i
-.-----,.,--,,,,,-w
---...,.,.-,.,.,,,,,--.,em,,_,,-,-w. - -,,.,,,,,
,.,.,,,.,,.--~.,-.,,-.,,..,,.--,,..,,,e._
8
- ~ _.
The plant groups derived from the " top down" approach, which are developed deductively by safety phil'osophy and strategy, are then merged with plant groups derived from the " bottom up" approach, which are developed inductively by safety tactics (subsystem traits) and dominant risk characteristics (sequence.
likelihood range and dominant factors).
Ideally, there should be a smooth meshing of the top down and bottom up approaches.
Practically, this may not initially occur since either approach may overestimate or underestimate the cc.
important of the multitude of possible similarities and differences.
It may be beneficial to have a plant in one plant group for one type of accident and in another for another type of accident. A coherent plant categorization hierarchy is to be formulated after interactive refinement among all levels of plant gro"ps.
This plant categorization hierarchy can be expanded by SARRP to incorporate P
containment types for each plant group / at various levels.
5.
NRC Position Since the research on plant categorization is still in progress, the NRC cannot provide a position on this issue at this time. Therefore, the NRC is relying on the reference plant concept (deterministic approach) in its policy making l
until the research results are completed in late 1984.
The reference plant concept is consistent with the IDCOR approach to plant categorization.
6.
Continuing Confirmatory Work Research on plant categorization will continue to support the NRC position on plant categorization.
l
-~=m-+-------- u
n Issue Number 2.1.3 Title Quantification of Sequence Likelihood Prepared by:
T. Enq Signature of Author:
Draft i l
Date 5/7/84 Contractor / Consultants F. Harper, SNL Review Process (List here contractor who will supply a statement on issue relative toIDCORwork)
Initial Date Branch Chief _
Asst. Div. Dir..
Division Dir.
Tech. Series Div. Director Date Sent to ACRS Date Sent to IDCOR M d - Orc av) d 4 ~ L cL Y_
zmt W p ~ 7 Ad A e
i ISSUE 2.1.3 OUANTIFICATION OF SE00ENCE LIKELIHOOD
-1.
' Description of Issue This issue paper deals with the quantification of accident sequence likelihood, which is en integral-part of PRA.
Its function is to provide infomation about i
plant design, operation, and safety that may not have been identified by tradi-tional deteministic analysis. These qualitative and quantitative insights enhance the understanding of plant design strengths and weaknesses, the impor-tance of assumptions about accident phenomena, operational strengths and weak-
~
nesses, source of uncertainty and their implications, and the relative importance of contributors to plant risk. This issu'e paper has a significant interface-j with other n':t; xx-mt issue papers including:
Plant Categorization
/
i (2.1.1);IdentificationofAccidentSequences(2.1.2);EquipmentPerformance and Success Criteria (2.1.4); Influence of Operator Action on Accident Sequence
-(2.1.5); Qualitative Assessment of Severe Accident Likelihood (2.2.2); and 1
Effects of Uncertainties and Sensitivities on Estimates of Severe Accident Likelihoods (2.2.7).
To support the NRC Severe Accident Policy Statement, the l
Accident Sequence Evaluation Program (ASEP) performs quantification of accident sequence likelihood to provide numerical measures for the selection of dominant I
accident sequences, categorization of plants, and determination of dominant i
factors that drive plant risk. Bottom line results are not as important as the l
insights gained from the ASEP work.
i 1
a
2 2.
Implications of the Issue to Regulatorv Questions The position of this issue, collectively with the issues mentioned earlier, can affect the answers to the following regulatory questions:
I How safe are the existing plants with respect to severe accidents?
How can the level of protection for severe accidents be increased?
What additional research or information is needed?
Is additional protection for severe accidents needed or desirable?
3.
Subissues What statistical meaning can be associated with the generic estimate of sequence / core melt frequencies and associated uncertainties?
Can decision be made based on the generic estimate of sequence / core melt frequencies and associated uncertainties?
How can operational experience be used to assess the likelihood of severe accidents?
How will sequence likelihood be used in plant categorization?
How should the results of existing PRAs and other studies be used to estimate accident sequence likelihood in existing plants?
What levels of detail is sufficient for generic quantification?
What is the data base for generic quantification?
a Can external event initiators be treated?
How are TMI fixes incorporated?
e
3 m.
4.
Status of Understanding ASEP is the primary research to support this issue.
The quantification of accident sequence likelihood takes two phases.
The first phase is qualitative and plant specific in nature and is performed to provide an interim reassess-ment of the core melt frequencies of selected PRAs as imediate input to the NRC source term work and th9 Severe Accident Risk Reduction Program.
The second phase is quantitative and generic in nature and uses PRA system modeling (fault trees) for quantification of sequence likelihood for the entire LWR population. This work 6 r ' C;;.' 7:rt :' ^EE" mrP O supportsthe
/
several issue papers indicated earlier, especially Plant Categorization.
Where appropriate, the results of the second phase will replace the qualitative results.
The qualitative approach uses the insights gained from existing PRAs and infor-mation sources including Station Blackout Study, Accident Precursor Study, human factors work and safety analysis reports as well as known TMI fixes.
These insights are applied to all the dominant accident sequences identified by ASEP from existing PRAs. These insights are applied to the initiating event frequency and system failure probabilities of the accident sequence.
A detailed es.n l,e to.s d description of the qualitative approach and results is fc.nd in an " Interim se Report on Accident Sequence Likelihood Reassessment" dated August 1983 by Sandia National Laboratories and Science Applications, Inc.
The qualitative approach shows, in general, the sequence frequencies of the examined plants are slightly lower than the original PRA estimates.
For the
~
4 PWR plants, the small LOCA with loss of coolant injection sequences have slightly higher frequencies due to the examination of operational experi_ence that showed the small break frequency is about 10-2 instead of 10-3 For both the PWR and BWR plants, the total loss of AC power types of sequences hav.e slightly higher frequencies based on the insights gai'ned from the long term blackoutanalysisperformed,undertheNRC$tation$1ackoutwork.
External v
event initiators are not included in the qualitative approach te - - fiv_.h er e=
- -- - '"- ::d since they were not included in the original PRAs.
To build up from the qualitative approach, ASEP is presently applying the fault tree modeling technique to the generic quantification of sequence likelihood.
The purpose of the generic quantification is not to generate bottom line numbers but to develop insights on the factors that drive sequence likelihood. These factors are used to group LWR 51nto plant classes as described in the Plant
/
Categorization issue paper.
Before the quantification precess, ASEP identifies a set of PWR and BWR dominant accident sequences,from existing PRAs and from the previous qualitative assess-
/
ment of the sequence likelihood as potential " candidates" for the LWR population.
j
,f Plant system and support system drawings are collected from FSARs for all the plants and they are verified by the utilities for accuracy.
For each system of interest, the drawings are simplified and collapsed considering significant differences in the redundancy, diversity, or support system dependencies.
For example, all AFW systems are grouped into 20 generic AFW configurations (see the Plant Categorization issue paper for more details).
Fault tree models are
s.,.
I.
5 developed for all the generic system and support system configurations with mecept 4*4 most at the train level of detail.
The basic events icAded in the fault tree v
models are those found significant past PRAs.
A data base is developed from data sources including PRAs, NUREG/CR Data Summaries, LERs, NPRDS. These data are analyzed for their generic applicability by ASEP and the uncertainties of these data are discussed in Issue Paper 2.2.7, Effects of Uncertainties and Sensitivities on Estimates of Severe Accident Likelihood, ch plant class v.
that is grouped by systems similarity Lc5accidentsequenceisquantified)'
The initiating events that are being considered are those associated with the dominant accident sequences " candidates" from past PRAs. Therefore, external initiators are not considered by ASEP.
The transient initiating event fre-quencies are from the " Transient Event Data Base Study" by EG&G and the LOCA initiating event frequencies are from the ASEP " Interim Report on Accident Sequence Likelihood Reassessment."
The SETS /SEP computer codes are used to produce best estimate and J d uncer-tainty range of the sequence likelihood.
Sensitivity analysis is perfonned on the sequence cut sets to determine event importance and uncertainty.
The results of the quantification and sensitivity snalyses will be used to derive v
insights for each plant class.
These insights provide generic information on the risk characteristics of the plant classes including their dominant accident sequences, sequence frequency range, and dominant factors that drive the sequence likelihood.
Using these insights, the plant classes can be further grouped based on similar risk characteristics.
l
^
l l
6
- 5.
NRC Position
~~
A firm NRC position cannot be made since supporting research is still in pro-The'NRCfeelsthatthegenericquantificationofsequencelikelihood gress.
cannot replace plant specific quantification since the analysis.is based on accident sequences that were found to be dominant from existing-PRAs.
Poten-ca.Jd be-tial outlying dominant accident sequences pr missed.
Therefore, the generic approach cannot be used to messure plant performance against a specific criterion
- or safety goal.
However, the insights gained from the generic quantification of sequence likelihood for all LWRs are valuable infomation to the NRC.
The application of the generic insights to a specific plant does not necessarily mean that the plant requires modification or has a significant safety problem.
The generic insights only flag an issue for further study on a plant specific basis. Also, current estimates of sequence likelihood, either generic and plant specific, provide a good basis for selecting dominant sequences for deteministic analysis, but they are not adequate for use with quantitative criteria because of the uncertainties involved.
2 6.
Continuing Confirmatory Work Research on the quantification of sequence likelihood will continue to support the NRC position, either through further work in ASEP or through more plant-specific PRAs.
- )
f
_.. _ _ _ _ _ _ _ _, _. _ _. ~
UNITED STATES
, /y neo.,';o,,
NUCLEAR REGULATORY COMMISSION g
g WASHINGTON. D. C. 20555
- a R.ry/}
From the desk of g/<f/7/
TONY ENG 8
7xi--4 t'~
u~Sq &
L.a.e A
2.R.B 5 2g/d T & l q -< S E h E N & Y 1@ ^Y g ascp,.c? ',' =1sf CA h2 C A cAA'~ d aff.rA 4.13,1; 7
f 7
4 4_L; a gA' oc = n ad.J' sf~ CocS) a-t'a/
~
a
,14 f t'd ut c Q d L' a. ocS LJ Jj c>w
<<1]
a m st/. rb~yl J Q 4. L d o c.s, 2 h e & d w& y 8 y-2 & a. 7 i fA eA J
%..- w.
- FSA2 m -/L 1 yhdd.3AA u ~ p
.t ly t
- g s,s e a m. c, 10 cre p~t so. 7i dd n db ~ a~ -<~4+ f Y WM' gQ m m
4 4 d '. - = Q ~ A K3.< A b J, f
~LAR kJ4, e t zo., MM. r.]y% &,'& oLh.
44&
J= L r zl L
~ d - - s a ~ y &. r.4 L p
- ~ e s.. ~ a s Lwn~ kg - A ~~th, M^
v.tf 4 y 6 @ a U d y y M.Ad.
m.
O'
- l. '
[
~TE:
TECHNOLOGY for ENERGY CORPORATION '
September 9, 1983
'Dr.-Denwood Ross, Deputy Director Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Comission 7915 Eastern Avenue Silver Springs, MD 20910
Dear Denny:
Agenda for NRC/IDCOR Technical Exchange Meetings Enclosed are proposed' schedules and agendas for five NRC/IDCOR 2
Technical Exchange meetings.
- gives the proposed schedule, with target dates for transmittal of information.
These dates have been tentatively set based on discussions with Dr. John Larkins of your staff. gives the proposed agendas for each meeting, with a list of IDCOR technical task reports relevant to the subject of each meeting.
The numbering of the items'is consistent with Table 3 "Organiza,
4' tion of the Technical Issues" of the August 5,1983, draft report of the SARP Senior Review Group, " Severe Accident Decisions for Existing Nuclear Power Plants."
Notice that Meeting 3 is on codes and models, which is not identified as such in Table 3 of the SARP report.
We are sending separately the relevant technical tasks reports f
for meeting No. I except for 12.3 and 15.1, which we will send next week.
We would appreciate it if you could send to us appropriate documents describing NRC work on these subjects.
Meanwhile, we will be working with John Larkins on arrangements for this first meeting.
Sincerely, 9
1, Anthony R. Buhl, Vice President Engineering AR8:gf W'
cc:
R. Bernero I /
G. Edgar.
M I/ ' l M. Fontana J. Larkins M. Leverett R. Mattson C. Reed J. Siegel Nss 100 983-041 L
cwrnattrem stirrumsrm rsr nnnu~mm nrruwrmia
r Tentative Scheduie for NRC/IDCOR Technical Exchange Meetings Date for Date of Information Meeting Subject Meeting Transmittal 1.
Sequence identification, progression of core melt, and loading of containment 11/7-10/83 9/12/83 2.
Fission product release and transport 12/5-8/83 10/15/83 3.
Codes and methods 1/9-12/84 11/15/83 4.
Response of containment and other essential equipment, and integrated assessment of safety 1/30-2/4/83 11/30/83 5.
Assessment of possible changes in design, operations and emergency preparedness
- 2/20-23/84 12/20/83
- 0ffsite emergency planning is not wi. thin the scope of IDCOR 6
1 e
e O
~
Enclcsure 2-Meeting 1 AGENDd
. e-
.i
.1.
Identification of the important accident sequences
" ]
1.1 Plant categorization
- 1.2 Identification of accident sequences for deterministic and probabilistic analysis 1.3 Quantification of sequence likelihood for probabilistic analysis 2.
Progression' of core melt in the reactor coolant system 2.1 Calculation of reactor coolant system thermal and hydraulic behavior 2.2 Rate and magnitude of hydrogen production in-vessel 2.3 Interactions between' fuel debris and structures surrounding or support-ing the fuel 2.4 Interactions between fuel debris and the reactor vessel or vessel penetrations 2.5 Lieklihood and magnitude of in-vessel steam explosions or spikes 2.6 Debris coolability in the vessel 3.
Loading of the containment 3.1 Calculation of containment temperature and pressure response
)
3.2 Hydrogen generation, combustion, and control Rate and magnitude of combustible gas production ex-vessel Distribution of combustible gases Conditions leading to and resulting from deflagration Conditions leading to and resulting from diffusion flames Conditions leading to and resulting from detonation 3.3 Molten fuel - coolant interactions Likelihood and magnitude of ex-vessel steam explosions or spikes Debris coolability in ex-vessel locations Debris dispersal following vessel failure 3.4 Molten fuel - structural interactions Between fuel debris and containment shell Between fuel debris and containment floor Between fuel debris and internal containment structures Generation of non-condensible gases Generation of Combustible gases i
Extent of penetration of floor 4
4 I
e
+~--e,,,
,,,,,,,,, -. ~ ~ ~ ~, - - -
w-e-
~nn.,-_,-,-n,-
.vn
,.g-p---,-,-------w--,,.
r-
-, w nsw w
b ;.,'.'
REPORTS:
TASK NO.
TITLE AUTHOR-
~2.1-9;"
Ground Rules TEC
(\\b Q ).
3.1 " 'i A
Define Initial Likely Dominant Sequences TEC._(tocp) 3.2 - }
Assess Dominant Sequences EI hgg) i
' 3.3 Assess Dominant Sequences Update EI (nq 12.1 -G...
(1) Hydrogen Generation During Severe Core Damage Sequences FAI (p ey (2) Hydrogen Generation During Severe Core Damage Sequences EPRI d -
(3) An Assessment of Existing Data On Zirconium Oxidation Under Hypothetical Accident Conditions in LWR's ANL 12.2 Hydrogen Distribution in Reactor 7
i Containment Buildings EPRI 12.3 Hydrogen Combustion in Reactor-Containment Buildings EPRI It d 14.
(1) Key Phenomological Models for Assessing Non-Explosive Steam Generation Rates FAI
' " #1 (2) Key Phenomological Models for Assessing Non-Explosive Steam Generation Rates FAI WvI i
15.1 (1) In Vessel Core Melt Progression Phenomena FAI g, y }
(2) Phenomological'a'ndModeling Background l
for the BWR arid PWR Heatup Codes
( D EPRI 15.2 (1) Effects of a Hypothetical Core Melt Accident on a PWR Vessel with Top Entry Instruments EPRI (0 y \\
(2) Debris Coolability Vessel Penetration and Debris Dispersal diFAI 15.3 Core-Concrete Interactions FAI Wy) l
\\
--,y y,,----,,__...y,-.,y-_,,-_w,.,--
--._,-,y-- -, -.,,
-~.m
,,--,,w
.y--
..i
' Meeting 2 AGENDA 5.
Fission product release and transport 5.1 Rate and magnitude of release of fission products from fuel in-vessel 5.2 Deposition of fission products during in-vessel transport 5.3 Rate and magnitude of release of radionuclides from fuel ex-vessel 5.4 Deposition of fission products in containment due to natural processes 5.5 Effect.of engineered safety features on fission product retention 5.6 Ex-containment consequence analysis REPORTS:
TASK NO TITLE AUTHOR
- 11.1, 11.4, 11.5 Estimation of Fission Product and Core-Material Source Characteristics EPRI 11.2 Identifying Pathways on Fission Product Transport IMPELL 11.3 Fission Product Transport in Degraded Core Accidents IMPELL 9
4 e
e t
Meeting 3 AGENDA
-Codes and Methods REPORTS:
TASK NO.
TITLE AUTHOR
~
11.3 Fission Product Transport in Degraded Core Accidents IMPELL
- 15.1 (1) In Vessel Core Melt Progression Phenomena FAI (2)' Analysis of In-Vessel Core
~
Melt Progression, Volumes I-IV EPRI 16.1 Assess Available Codes, Define Use, and Follow and Support Ongoing Activities JAYCOR 16.1.1 Riview of MAAP/BWR Code JAYCOR Review of MAAP/PWR Code JAYCOR Review of RETAIN Code JAYCOR 16.2 MAAP Users Manual FAI e
0 m
e
-.y-v
,-.---,.m.
-.-me,-y---
y
-.-_---y-.--
-,.--,.-,-_,,..,v.,.--
Meeting 4 AGENDA 4.
Response of the containment and'other essential equipment 4.1 Reliability of early containment isolation 4.2 Character and likelihood of early containment failures 4.3 Character and likelihood of later containment failures 4.4 Equipment and instrumentation survivability 6.
Integrated assessment of safety 6.1 Calculated response of typical (reference) plants to selected severe 4
accidents 6.2 Qualitative assessment of likelihood and capability for severe accidents 6.3 Integrated probabilistic risk assessments for reference plants 6.4 Generic conclusions independent of design 6.5 Conclusions applicable only to reference plants i
6.6 Residual uncertainty for other plants 6.7 Analysis required to qualify other plants to the generic conclusions i
REPORTS:
TASK NO.
TITLE AUTHOR 4.1 Containment Event Trees for PWRs and BWRs SAI 4.2 Update System / Containment Event Trees TEC 7.1 Baseline Risk Profile for Current Generation Plants EI 10.1 Containment Structural Capability TEC i
17.
Equipment Survivability in a Degraded Core Environment NUS i
18.1 Evaluate Atmospheric and Liquid l
Pathway Dose NUS
{
21.1 Risk Reduction Potential EI 23.1 Inte (1) grated Containment Analysis Peach Bottom PEco (2) Sequoyah TVA (3) Grand Gulf MSS (4) Zion W
i 23.2 Technical Integration of Containment Analyses TEC i
a
- i..**
i 1
ting 5 g.
AGENDA 7.
Assessment of possible changes in design, operation and emergency preparedness
- 7.1 Operational reliability 7.2 Design i:hanges.to better prevent acciden,ts 7.3 Improvements in severe accident management capability 7.4 Design and emergency preparedness changes for severe accident consequence mitigation 7.5 Selection of cost-benefit methods and criteria REPORTS:
. TASK'NO TITLE AUTHOR 1.1, 1.2 Safety Goals TEC e
5.1 Human Error:
Effects of Dominant
-Accident Sequences TEC 6.1 Risk Significant Profile for ESF and Other Equipment EI 7.1 Baseline Risk Profi,le for Current Generation Plants EI 8.1 Effect of Post-TMI Changes on the Overall Risk Profile EI 9.1 Preventive Methods to Arrest Sequences of Events Prior to Core Damage NUS 13.
Evaluation of Means to Prevent, Suppress, or Control Hydrogen Burning in Reactor Containment S. Levy 19.1 Alternate Containment Concepts Bechtel 20.1 Core Retention Devices Offshore Power 21.1 Risk Reduction Potential EI 22.1 Identification of Safe Stable States FAI a
23.1 (1) grated Containment Analysis Inte Peach Bottom PEco (2) Sequoyah TVA (3) Grand Gulf MSS (4) Zion W
}
- Offsite emergency planning is not within the scope of IDCOR.
8
4 23.2 Technical Integration of Containment Analyses.
TEC 24.1, 24.3 Operator Recovery Actions During Severe Accident Secuences NUS 24.2 Human Reliability Assessment HRA 4
6 0
4 o
6 -
b 6
y.
y-1
- -.c,..,mv-c-
y v-
,- +--,-, -
v i
+r--r
g,..
?,-
[
REPORT NO.
MEETING NINBER 1
2 3
4 5
1.1. 1.2 X.
. 2_
2.1 X
3.1 X
3.2 X
3.3 X
4.1 X
4.2 X
5.1 X
6.1 X
7.1 X
X 8.1 X
9.1 X
10.1 X
11.1,11.4.11.5 X
11.2 X
11.3 X
X
=r._
12.1 X
12.2 X
12.3 X
13 X
14 X
15.1 X
X 15.2 X
15.3 X
., y..
.3
' REPORT NO.
MEETING NUMBER
-1 2
3 4
5 16.1 X
16.1.1 X
17 X
18.1 X
19.1 X
20.1 X
21.1 X
X 22.1 X
23.1 X
X 23.2 X
X 24.1, 24.3 X
24.2 X
6 e
e m
G v-e um v y w-w' we
' - ~ -
w e
v+e w
-v-w*-
v--44------e-re--
m----w-m-
--