ML20211E040
| ML20211E040 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 01/31/1985 |
| From: | Ami Agrawal, Khatibrahbar, Ludewig H BROOKHAVEN NATIONAL LABORATORY |
| To: | NRC |
| Shared Package | |
| ML20209E304 | List: |
| References | |
| CON-FIN-A-3778, FOIA-86-678 NUDOCS 8610220339 | |
| Download: ML20211E040 (53) | |
Text
. _ _ _
5.
, a' m e, r m a
.../ue.,e. w g> m,x m
g, a.-
p.y;w,.an %.. -.+g y-
.:.lbg..,
.y:
- kd'
_ r N,
.-3778 1 30-85 17, !
'7
. $',. 4 *. ! O j Nj. I[.d. ' '.. [ Mfhh it ![$,{$[j_ [I'
)h $:!'} [,
' ![
.'f, q'.j"(.,)
- 3. :
- i.d4.Aj 1' N,7 ' ;
jej
,g
(. J.
r
~4.
{
'.Y?1
- s
-: h??
,...Y f.'N, $ A * *~
s'., y
. L. M1,
-~
- k
, -(
'[ f
!'l:h f
Y
$ r-Q
{ &hP.C.$$i.*
~~
~W 1.tt
- s
.I s
i I
f a
3 i.
p,...::..
g F'
d
. i....
A3'
$/
r
,j, /{
9 ND
'PRAIT.,
- 4) %
'ta V{*i '
M hl i
I c
h
?
k l
j.
WA i
k
)
~
j
b.T. / (
ed thk S. Nuclear Rogulatory Com asion
' }"' 9/
d4
' Office offlui. fear' Reic_ tor Regulati 2.' *:, '
C' 48.
d.
g tigct M
' f-f 16I L
l 8610220339 861014 C
-678 PDR
" *L
-m-w-~----
a b',E c}iNICAL REPOAT
- 'Nf.f^#NM'k.N - E *'Ik ' / i G k '
2
- ,. r-?g g 1.@i.4%['.h9.D. ti. %-@ ].,I-[AYh YIC, Y, h[N A
MM l
,.;3778 1-30-85
',-lgg$Qgiu.g'*Ej.4
- n.W. Sh y;pgfffc;p.3,.
,o
>$; hh$f.-c
?k:
hs%h ! :..
E
?~
N 7.k.y 5L t'.&
- p y
.9:. ?f' e j
W pr. \\.
v:
I
~~
' ' +
u i
f'
'k '
hf
,y I
);EV 16 I
't me.
C'.
l k
L:
'3:" -
BA
-b JDEWI h
fM=.
M..
L, ND
'PRATI'. M o
. f.fi
[Ohii f
, )h.;.'.
n l
- Q, 1,
l 1
h i
z, l
g 1
l (9 ;
l V1 u.
4 h
Bjt00K N
RY.
t 3,
~ Kf'"
, T / {"l onyw~_ng, ReyulatoryC mmthion.....
the U.S. Nuoje Prppared f
}l}
_jg$;.
g"n~%
w
~'
e 8610220339 861014 PDR C
RA1
-678 _
i l ~ I -. --
o Technical Report A-3778 1-31-85 PRELIMINARY REVIEW AND EVALUATION OF THE SEABROOK PROBABILISTIC SAFETY STUDY M. Khatib-Rahbar, A. K. Agrawal, H. Ludewig, and W. T. Pratt Department of Nuclear Energy Brookhaven National Laboratory Upton, New York 11973 January 1985 Prepared for U.S. Nuclear Regulatory Commission Washington, D. C.
20555 Contract No. DE-AC02-76CH00016 FIN No. A-3778
O o
-i i i -
ABSTRACT A preliminary Level I review of the Stabrook Probabilistic Safety Study containment failure modes and radiological source term characteristics has been completed. The review identifies the major features of the plant as they relate to risk assessment, including comparisons to the Zion, Indian Point, and Millstone-3 plant designs.
The future direction of the review is indi-cated and a list of preliminary questions is also included.
t
-v-CONTENTS Page ABSTRACT...............................
iii 1.
INTRODUCTION...........................
I 1.1 Background..........................
I 1.2 Objectives..........................
2 1.3 Organization of the Report..................
2.
2.
PLANT DESIGN AND FEATURES.....................
4 l
2.1 Plant Design.........................
4 2.2 Compa r.i son to Ot her Pl a nt s..................
9 3.
PRA REVIEW............................
11 3.1 An a l y t i c a l Me t h od s......................
11 3.2 Containment Fail ure Model..................
12 3.2.1 Leak-Before-Failure..................
12 3.2.2 Classi fication of Failure...............
13 3.2.3 Containment Leakage Model...............
18 3.3 Description of Plant Damage. States..............
20 3.4 Containment Event Tree and Accident Phenomenology......
24 3.5 Containment Matrix (C-Matrix)................
24 3.6 Ac ci de n t Sou rce Te rms.................... -
33 3.7 Source Term Uncertainty Analysis...............
40 3.8 Of f-Site Consequence Analysi s................
44 3.9 Risk of Two-Reactor Station.................
46 4.
SUMMARY
AND CONCLUSIONS......................
47 4.1 Resul ts of Level 1 Revi ew..................
47 4.2 Prel imi na ry Question s....................
48 4.3 Future Plans.........................
48 5.
REFERENCES............................
49
i o
-vi-LIST OF TABLES Table Title Page 2.1 Comparison of design characteristics...........
10 3.1 SEABROOK-1 containment liner penetrations..........
19 3.2 Frequencies of occurrence of the plant damage states....
23
~
3.3 Top events for the Seabrook containment event tree.....
25 3.4 Release categories employed in the Seabrook Station risk model.........................
27 3.5 Seab rook C ma t ri x......................
28 3.6 Simpli fied C-matrix for SPSS................
29 3.7 Seabrook point-estimate release categories.........
35 3.8 Late overpressurization failure comparison.........
38 3.9 Comparison of releases for failure to isolate containment and the by-pass sequence..................
41 3.10 Comparison of AB-c and TMLB'-c (BMI-2104) to 337 and 37...
43 3.11 Comparison of 1997 (sun) to V-sequence (Surry)........
45 i
LIST OF FIGURES Figure Title Page 3.1 Definitions of the plant damage states used in SPSS.....
22
c
' 1.
INTRODUCTION i
1.1 Background
After the accident at Three Mile Island, the Nuclear Regulatory Commis-sion (NRC) recognized the need to reexamine the capabilities of nuclear power plants to accommodate the effects of hypothetical severe accidents beyond the design basis.
This reexamination included consideration of potential design modifications to mitigate the consequences of these degraded and core melt accidents.
The Zion and Indian Point power plants were chosen to initiate this ac-tivity because of the large populations surrounding the two sites.
The con-cern was that due to the proximity of these two sites to high population den-sities, they could comprise a disproportionately high component of the total societal risk from U.S. connercial nuclear power programs.
As part of this continuing effort, programs to evaluate the risk from plant sites situated near high population centers have been set in motion, in order to introduce design modifications and mitigation features, which can re-duce the public risk.
Probabilistic Risk Assessment (PRA) studies have been undertaken by a t
number of utilities " and reviewed by Brookhaven National Laboratory (BNL) under contract to the NRC. BNL was also actively involved in preparation of a preliminary report 5 (NUREG-0850) which represented the staff's initial contri-bution to the understanding of severe accident progression and mitigation.
This report presents a preliminary evaluation of the containment failure modes and the radiological release characteristics of the Seabrook Probabilis-tic Safety Study (SPSS), which was completed by Pickard. Lowe and Garrick,
4
c
. Inc. (PLG) for the Public Service Company of New Hampshire and Yankee Atomic Electric Company in December 1983.6 1.2 Obj ectives The objectives of this review report are to provide the NRC staff with a preliminary (Level 1) review of the SPSS as part of a broader objective in-volving an in-depth review and evaluation of the technical basis for the sub-ject PRA, which will be performed over the next few months.
In particular, 4
core melt phenomenology, containment response, containment event trees, re-lease categories, and site consequence models are to be examined.
l This Level I review, which was performed over only five months' duration, highlights important features of the plant design and the SPSS as compared to the other PRAs.1
The report also provides an initial assessment of the PRA method, validity of major assumptions, and relevance and adequacy of conclusions.
Areas needing further verification and study are identified, and finally, questions for the applicant or Licensee pertaining to the SPSS are addressed.
1.3 Organization of the Report A brief review of the Seabrook design and features is presented in Chap-ter 2 along with comparisons to Zion, Indian Point, and Millstone-3 plant designs.
Chapter 3 contains the preliminary assessment of the Seabrook PRA.
Spe-i ci fically, analytical methods, containment event trees, accident phenome-nology, containment matrix, uncertainty analysis, accident source terms, and off-site consequences are reviewed.
O i '
In Chapter 4 the results of this preliminary Level i review are sum-
)
marized and areas needing further study' are also highlighted along with need 4,
i for additional information and questions to the applicant or the Licensee, i
3.
)
I i
i I
i 1
l 4
I i
i e
F l
i l
r
+
I i
,--,--,n-
...--.,,r._,__.,-,-
. 2.
PLANT DESIGN AND FEATURES In this section, those plant design features that may be important to an assessment of degraded and core melt scenarios and containment analysis are reviewed.
These important features are then compared with the Zion, Indian Point and Millstone-3 facilities in order to identify commonalities for bench-mark comparisons.
2.1 Plant Design The Seabrook Station is comprised of two nuclear units each having an
. identical nuclear steam supply system and turbine generator.
The units 'are arranged using a " sling-along" concept-which results in Unit 2 being arranged similar to Unit 1 but only moved some 500 feet west.
Each unit is a 1150 MWe (3650 MWt), 4 loop, standard Westinghouse PWR plant.
The turbine-generators are supplied by the General Electric Company and the balance of the plant is designed by United ~ Engineers and Construction.
The major structures associated with each unit are the containment struc-ture, containment enclosure, primary auxiliary building, fuel storage build-ing, control building, diesel generator building, turbine generator building, emergency feedwater pumphouse, and main steam and feedwater pipe chase.
Each containment completely encloses a reactor coolant system, and is a seismic Category I reinforced concrete structure in the form of a right verti-cal cylinder with a hemispherical dome and flat foundation mat founded on bed-rock.
The inside face is lined with a welded carbon steel plate, providing a high degree of leak tightness.
A protective 4 f t. thick concrete mat, which forms the floor of the containment, protects the liner over the foundation
. mat.
The containment structure provides biological shielding for normal and accident conditions..The approximate dimensions of the containment are:
Inside diameter 140 ft.
Inside height 219 ft.
Vertical wall thickness 4 ft. 6 in, and 4 ft. 7 1/2 in.
Dome thickness 3 ft. 6 1/8 in.
Foundation mat thickness 10 ft.
Containment penetrations are provided in the lower portion of the structure, and consist of a personnel lock and an equipment hatch / personnel lock, a fuel transfer tube and piping, electrical, instrumentation, and ventilation pene-trations.
Each containment enclosure (also known as secondary containment) sur-rounds a containment and is designed in a similar configuration as a vertical right cylindrical seismic Category I, reinforced concrete structure with dome and ring base.
The approximate dimensions of the structure are: inside diam-eter,158 ft; vertical wall thickness, varies from 1 ft, 3 in. to 3 ft; and dome thickness, I ft, 3 in.
The containment enclosure is designed to entrap, filter and then dis-charge any leakage from the containment structure.
To accomplish this, the space between the containment enclosure and the containment structure, as well as the penetration and safeguards pump areas, are maintained at a negative l
pressure following a loss-of-coolant accide1t by fans which take suction from the containment enclosure and exhaust to atmosphere through charcoal filters.
To ensure air tightness for the negative pressure, leakage through all joints and penetrations has been minimized.
- A containment spray system is utilized for post accident containment heat removal.
The containment spray system is designed to spray water containing boron and sodium. hydroxide into the containment atmosphere after a LOCA to
)
cool it.and remove iodine.
The pumps initially take suction from the re-fueling water storage tank and deliver water to the containment atmosphere through the spray headers located in the containment dome. After a prescribed amount of water is removed from the tank, the pump suction is transferred to the containment sump, and cooling is continued by recirculating sump water through the spray heat exchangers and back through the spray headers.
The spray is actuated by a containment spray actuation signal which is generated at a designated containment pressure.
The system is completely re-dundant and is designed to withstand any single failure.
The containment isolation system establishes and/or maintains isolation of the containment from the outside environment in order to prevent the re-lease of fission products.
Automatic trip isolation signals actuate the ap-propriate valves to a closed position whenever automatic safety injection oc-curs or high containment pressure is experienced.
Double barrier protection is provided for all lines that penetrate the containment boundary.
Thermal electric hydrogen recombiners reduce the concentration of hydro-gen in the post-LOCA containment atmosphere. A purge feature is also provided
.f for backup control of hydrogen.
The emergency core cooling system (ECCS) injects borated water into the reactor coolant system following a LOCA to limit core damage, metal-water reactions and fission product release, and to assure adequate shutdown margin.
The ECCS also provides continuous long-term post-accident cooling of f
-7 the core, by recirculating borated water between the containment sump and the reactor core.
The system consists of two centrifugal charging pumps, two low pressure safety injection pumps, two residual heat removal pumps and heat exchangers, and four safety injection accumulators.
The system is completely redundant, and will assure flow to the core in the event of any single failure.
The control building contains the building services necessary for con-tinuous occupancy of the control room complex by operating personnel during all operating conditions.
These building services include:
HVAC services, air purification and iodine removal, fresh air intakes, fire protection, emer-gency breathing apparatus, communications and meteorological equipment, light-ing, and housekeeping facilities.
ESF filter systems required to perform a safety-related function follow-ing a design basis accident are discussed below:
a.
The containment enclosure exhaust filter system for each unit col-lects, filters and discharges any containment leakage. The system is not normally in operation, but in the event of a LOCA, it is placed in operation and keeps the containment enclosure and the building I
volumes associated with the penetration tunnel and the ESF equipment cubicles under negative pressure to ensure all leakage from the con.
tainment structure is collected and filtered before discharge to the plant vent.
b.
One of two redundant charcoal filter exhaust trains is placed in operation in the fuel storage building whenever irradiated fuel not in a cask is being handled.
These filter units together with dampers and controls will maintain the building at a negative pressure, t
o The emergency feedwater system supplies demineralized water from the con-densate water storage tank to the four steam generators upon loss of normal feedwater flow to remove heat from the reactor coolant system.
Operation of the system will continue until the reactor coolant system pressure is reduced to a value below which the residual heat removal system can be operated.
The combination of one turbine-driven and one motor driven emergency feedwater pump provides a diversity of p'ower sources to assure delivery of condensate under emergency conditions.
The two units of the facility are interconnected to off-site power via three 345 kilovolt lines of the transmission system for the New England states.
The normal preferred source of power for each unit is its own main turbine generator.
The redundant safety feature buses of each unit are powered by two unit auxiliary transformers.
A highly reliable generator breaker is provided to isolate the generator from the unit auxiliary trans-i formers in the event of a generator trip, thereby obviating the need for a bus transfer. upon loss of turbine generator power.
In the event that the unit auxiliary transformers are not available, the redundant safety feature buses of each unit are powered by two reserve auxiliary transformers.
Upon loss of off-site power, each unit is supplied with adequate power by either of two fast-starting, diesel-engine generators.
Either diesel-engine generator and its associated safety feature bus is capable of providing adequate power for a safe shutdown under accident conditions with a concurrent loss of off-site power.
A constant supply of power to vital instruments and controls of each unit is assured through the redundant 125 volt direct current buses and their associated battery banks, battery chargers and inverters.
. 2.2 Comparison to Other Plants Table 2.1 sets forth the design characteristics of the Zion, Indian Point-2, and Millstone-3 facilities as they compare to the Seabrook station.
~
It is seen that-the containment characteristics are quite similar with the exception of containment operating pressure for Millstone-3 (subatmos-pheric design), and the use of fan coolers in Zion and Indian Point for post-accident containment cooling.
The primary system designs are nearly identical between the four units.
?
An important distinction from the post-accident containment cooling standpoint is the significantly higher amount of RWST water for Seabrook as compared to the other plants.
The Seabrook containment building basemat and the internal concrete structures are composed of basaltic-based concrete.
As concrete is heated, water vapor and other gases are released. The initial gas consists largely of carbon dioxide, the quantity of which depends on the amount of calcium car-bonate in the concrete mix.
Limestone concrete can contain up to 80% calcium carbonate by weight, which could yield up to 53 lb of carbon dioxide per cubic foot of concrete.
However, basaltic-based concrete contains very little cal-cium carbonate (3.43 w% for Seabrook) and would not release a substantial amount of carbon dioxide.5 These inherent differences in design are expected to influence the out-come of the accident events, particularly, ex-vessel core coolability and con-tainment behavior are significantly influenced by debris bed coolability, water availability and the core-concrete interaction as will be addressed in the following chapter.
. Ta bl e 2.1 Comparison of design characteristics Zion Indian Point Millstone Seabrook i
3 3
8 Design Parameters Unit 1 5 Unit 2 5 Unit 3' Unit 1.2 l
Reactor Power
[MW(t)]
3250 3030 3411 3411 i
containment Butiding:
5 3
5 2.3 x 10' 2.7 x 10
, Free Volume (ft) 2.73 x 10' 2.61 x 10 Design Pressure (psla) 62 62 59.7 67.7 Initial Pressure (osta) 15 14.7 12.7/9.1 15.2/13.2 Initial Temperature
(*F) 120 120 120/80 120/50 primary System:
3 Water Volume (ft )-
12.710 11.347 11.671 11524 Steam Volume ft )
720 720 7
751 Mass of U0 in Core Ib) 216600 216600 222739 222739 2
Mass of Steel in Core Ib) 21.000-20.407'
?
19000 Mass of Ze in Core fib) 44.5n0 44.600 45.296 45234 Mass of Bottom Head Lib) 87'.000 78.130 87.000 87000
. Bottom Head Ofameter Ift) 14.4 14.7 14.4 14.4 Bottom Head Thickness (ft) 0.45 0.44 0.45 0.45 Containment Rutiding Coolers:
Sprays yes yes yes yes Fans (with safety function) yes yes no no Acetriulator Tanks:
Total Mest of Water Ib) 200.000 173.000 348.000 213000 Initial Pressure psta) 665 665 600 615 Temperature
'F) 150 150 80
?
Refueling Water Storage Tank:
8 7
9 Total Mass of Water (Ib) 2.89 x 10' 2.89 x 10 10 1.7 x 10 Initial Pressure (psta) 14.7 14.7 12.7/9.1 15.2/13.2 Temperature
('F) 100 120 50/40 7
Reactor Cavity:
Configuration Wet Wet Dry Ory/ Wet Concrete Material Limestone Rasaltte Basaltte Basaltic 6
" 3.
PRA REVIEW In this chapter a brief review of the Seabrook Probabilistic Safety Study (SPSS) is presented.
Specifically, the analytical techniques used to analyze core meltdown phenomena and containment response are identified. Where possi-ble, parallels between this study and other existing PRA studies are set forth.
Finally, the relevance and validity of the conclusions is addressed.
4 3.1 Analytical Methods A brief description of the computer codes used to perform the transient degraded, core meltdown and containment response analyses is provided in this section.
The MARCH computer code is used to model the core and primary system -
transient behavior and to obtain mass and energy releases from the primary system until reactor vessel failure.
These mass and energy releases are then used as input to the other computer codes for analysis of containment response.
For sequences in which the reactor coolant system remains at an elevated pressure until the vessel failure (" time-phased dispersal"), the M00 MESH com-puter code is used.
This code calculates the steam and hydrogen blowdown from the reactor vessel using an isothermal ideal gas model.
The water level j
boil-off from the reactor cavity floor is modeled using a saturated critical heat flux correlation.
Additionally, the accumulator discharge following de-pressurization caused by the vessel failure is also considered.
't J
. A modified version of the CORC'ON code is used to replace the INTER sub-reutine of the MARCH code.
CORCON models the core-concrete interaction after the occurrence of dryout in the reactor cavity.
The mass and energy releases from the core-concrete interaction are transferred to the M00 MESH code for proper sequencing and integration into the overall mass and energy input to C0C0 CLASS 9 code.
C0C0 CLASS 9, a modified version of the Westinghouse C0C0 computer code utilizes the mass and energy inputs to the containment as computed by MARCH to model the containment building pressurization and hydrogen combustion phe-nomena.
This code replaces the MACE subroutine of the MARCH code.
The code also models heat transfer to the containment structures and capability for containment heat removal through containment sprays and sump recirculation.
Fission product transport and consequence calculations are performed using the CORRAL-II-and the PLG proprietary CRACIT computer
- codes, respectively.
The analytical methods used to carry out the core and containment thermal hydraulics, and fission product transport calculations are' identical to those I
used for MPSS-3.7 l.
l 3.2 Containment Failure Model 3.2.1 Leak-Before-Failure During accident sequences involving core damage, the containment struc-ture will be exposed to pressures and temperatures beyond those expected for design basis accidents (DBA).
Response of the containment building to these severe conditions is evaluated in the SPSS by employing, for the first time', a l
leak-before-failure nodel.
In this model, allowance is made for continuous l
l l
i
13-t leakage from the containment to the surroundings.
This mode of containment failure is termed local failure.
The containment leakage can occur at many locations and discontinuities such as mechanical and electrical penetrations, personnel. lock, equipment hatch, fuel transfer tube, welds, and in between the liner and concrete.
Depending upon the size of leakage area and the accident sequence, local failures may gradually relieve pressure, thereby gross con-tainment failure may be averted.
The leak-before-failure approach is more realistic compared to the gross failure approach. The extent of leakage and the resulting health consequences must, however, be carefully studied.
In order to explain this issue, it is observed that traditionally probabilistic risk assessment have made use of a " threshold"~ model.
In the threshold model, the containment is considered to remain intact' until the internal loading equals or exceeds a pressure threshold (which may also be temperature dependent), at which point it is as-sumed to suffer a failure (gross).
If the internal loading is below this threshold value, the containment is considered intact and hence the risk is j
quite low.
In the leak-before-failure model, the release of activity, which is considerably small compared with that for the gross, failure mode, must be considered in health consequences.
However, such leakages can potentially prevent the internal pressure from approaching the threshold value and thus a catastrophic or gross failure may be avoided.
3.2.2 Classification of Failure The SPSS report has classified containment failures in three categories:
1.
Containment Failure Category A.
Includes containment failures that develop a small leak that is substantially larger than the leak
-~
- e e
%,_+
. acceptable from an. intact containment but not large enough to arrest the pressure rise in the containment. Category A. failures thus cause an early increase in the rate of leakage of radionuclides over the design basis leak rate but pressurization of the containment con-tinues until either a category B or C containment failure develops.
An intact containment is defined as the one in which leakage is lim-ited to the Technical Specification value.
For Seabrook-1, this value is 0.2 w/o per day at the calculated peak accident pressure of approximately 47.0 psig.
Note that this value in SSPS has been quoted as 0.1 volume percent per day although prior to the most re-cent amendment dated August 1984, the'FSAR has cited both 0.1 volume percent.and 0.1 w/o per day.
The 10CFR50, Appendix J, mandates the allowable leakage to be quoted 'as w/o per day.
The higher value noted here is based on the August 1984 Amendment 53.*
2.
Containment Failure Category B.
Includes failure modes that develop a large enough leak area so that the pressure in the containment no longer increases.
The time during which a substantial fraction of the radionuclide source term is released is longer than approximately 1 to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />.
Category B failures include self-regulating failure l
- There appears to be substantial update / changes in the Engineered Sa fety Features flow diagram including arrangements of motor-operated valves and by-(
pass lines, which may substantially change the frequency of events.
BNL is j
not reviewing this part of the SSPS.
l 1
i
. modes where'the leak area is initially small but increases with pres-sure so that it becomes sufficient to terminate the pressure rise be-fore a Category C containment failure occurs.
The definition of
" substantial" fraction is not clear.
3.
Containment Failure Category C.
Includes those containment failure modes that develop a large leak area.
A large fraction of the total radionuclide source term is released over a period of less than 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />.
All gross failure modes are included in Category C.
Mathematically, these three failure categories can be expressed in terms of leakage area as follows:
Type A ADBA < AA<ANP Type B (1)
ANP < AB<Ap Ac>A Type C p
where ADBA = leakage area corresponding to the technical specification limit for containment leakage.
Anp = leakage area insufficient to arrest pressurization, and, Ap
= leakage area suf ficient to arrest pressurization.
3 i
A value for the leakage area corresponding to a. specified leakage rate can be obtained once the flow conditions and the orifice nature are known.
Since the expected pressures for degraded core conditions far exceeds the choked flow conditions (flow through a sharp-edged orifice is choked where the
~
internal pressure exceeds 28 psia), only the choked flow equation is noted below:
k+1 2,k-1 kfk+1[
A (6F (2)
W=
where W = discharge rate (kg/s)
A = leak area 2
P = absolute pressure (N/m )
i o = mixture density, and k = ratio of specific heats at constant pressure to that at constant 4
volume.
For air and water vapor mixture, k - 1.3.
If the mixture density is expressed by perfect gas law P
p=g (3) where l
l R = gas constant, and I
I T = the absolute temperature, I
then Eq. (2) becomes I
l l
l
4
. k+1 2
W=
k A
(4)
The mass of mixture can be written as M=Vp or, h
P.
(5)
M=
where V = free mixture volume in the containment.
Equations (4) and (5) can be combined to get the leakage rate as a. mass fraction, that is:
k+1 I
f=
k
. (RT)-1.5. A.
(6) 2 In other words, the leakage rate when expressed as a fraction of mass, for a f
given temperature, depends only on the leakage area.
i
[
For Seabrook, using V = 2.704 x 10 6 ft 3 and T = 296F, Eq. (6) gives:
I
(
the mass fraction leakage rate per hour = 0.00721 A (7) in l
2 where A n is the leakage area in in.
The essentially intact containment i
leakage of 0.2 w/o per day, thus, corresponds to leakage area of 0.012 in2 i
. (or, an equivalent hole of 0.12-in diameter).
Similarly, for leakage area of 4 to 10 in, the leakage rate would range from about 2.9 w/o to 7.2 w/o per hour.
These leakage values are a factor of two too low compared to those cited in the SPSS. The establishment of cut-off points for leak areas between the three types is essentially based on the release rates.
It is, therefore, important to resolve this discrepancy in leakage area computation between SPSS and Eq. (7) noted above.
3.2.3 Containment Leakage Model There are a large number of containment penetrations.
Table 3.1 lists all containment liner penetrations other than the electrical, equipment hatch and personnel lock.
These penetrations are grouped according to their func-tion.
Their sizes are also noted in this table. Note that there are 14 addi-tional spare penetrations for which size is not available.
There are a total of 64 electrical penetrations of which 14 are spare and 8 are unused.
Local containment failures (Type A and B) are believed to occur at one or more of these penetrations.
Such a failure can be due to degradation of seals around these penetrations. The overall leakage area can be representede as A (T,P,t) = A. f (T,P,t)
(8) where A
= pre-existing leakage area, f(T,P,t) = degradation factor, and t
= seal exposure duration at pressure P and temperature T.
. Table 3.1 SEABROOK-1 containment liner penetrations Penetration Penetration Nunbers Service Size x-1 to x-4 Main steam line 30" x-5 to x-8 Main feedwater 18" x-9, x-10 RHR pump suction 12" x-11 to x-13 RHR to safety injection 8"
x-14 to x-15 Containment buildup spray 8"
x-16, x-18 Containment on-line purge 8"
x-17 Hydrogenated vent header 2"
x-20 to x-23 CCW supply and return 12" x-24 to x-27 Safety injection 4"
x-28 to x-31 CVCS to pump seal injection 2"
x-32, x-34 Drain line 3",
2" x-33, x-37 CVCS 3"
x-35,x-36,x-40,{
RCS test / sample / control 1" or smaller x-52, x-71, x-72 f x-38 Combustible gas control 10" x-39 Spent fuel pool cooling 2"
x-43,x-47,x-50,{
Instrumentation lines x-57 i
x-60, x-61 From containment recirculation sump 16" x-62 Fuel transfer tube 20" x-63 to x-66 Steam generator blowdown 3"
x-67
. Service air 2"
HVAC-1,2 Containment purge supply / exhaust lines 36" x-19, x-41, x-42, x-44 to x-46, 1
x-48,x-49,x-51,(
Spare
?
x-58, x-59, x-68 to x-70 l
l l
l l
r i
I
5
. This type of formulation could be useful in obtaining a statistical distribu-tion of the leakage area.
Such a. determination can be made after careful as-sessment of location and sealant for different penetrations.
In the SSPS re-port, selective penetrations, e.g., x-25, x-26, and x-27 are used. A justifi-cation for omitting x-24 which is also a safety injection penetration (see Table 1)~ is not given. Similarly, an assessment for failure of other penetra-tions has to be made.
The 3SPS report has classified x-23, x-26, x-71 as stiff mechanical penetrations.
It appears that penetrations x-24, x-25, x-26, and x-27 are all safety injection service lines and they all should be stiff instead of just x-26.
Similar questions exist on SSPS selection of other penetrations.
The gross (Type C) containment failure modes are all due to a failure in the containment shell.
For such failure modes, potentially very large leak areas resulting in very rapid containment blowdown are possible.
An indepen-dent assessment of the SPSS failure pressur"e should be made.
All electrical penetrations are below grade.
Leakages through them are perhaps not as critical as through other containment liner penetrations.
-3.3 Description of Plant Damage States The grouping of accident sequences into plant damage states proceeds from the premise that the broad spectrum of many plant failure scenarios can be discretized into a manageable number of representative categories for which a single assessment of core and containment response will represent the response of all the individual scenarios in that category.
The plant damage states classify events in accordance to the following three parameters:
. This type of formulation could be useful in obtaining a statistical distribu-tion of the leakage area.
Such a determination can be made after careful as-sessment of location and sealant for different penetrations.
In the SSPS re-port, selective penetrations, e.g., x-25, x-26, and x-27 are used. A justifi-cation for omitting x-24 which is also a safety injection penetration (see Table I) is not given.
Similarly, an assessment for failure of other penetra-tions has to be made.
The SSPS report has classified x-23, x-26, x-71 as stiff mechanical penetrations.
It appears that penetrations x-24, x-25, x-26, and x-27 are all safety injection service lines and they all should be stiff instead of just x-26.
Similar questions exist on SSPS selection of other penetrations.
The gross (Type C) containment failure modes are all due to a failure in the containment shell.
For such failure modes, potentially very large leak areas resulting in very rapid containment blowdown are possible.
An indepen-dent assessment of the SPSS failure pressure should be made.
All electrical penetrations are below grade.
Leakages through them are perhaps not as critical as through other containment liner penetrations.
3.3 Description of Plant Damage States The grouping of accident sequences into plant damage states proceeds from the premise that the broad spectrum of many plant failure scenarios can be discretized into a manageable number of representative categories for which a single assessment of core and containment response will represent the response of all the individual scenarios in that category.
The plant damage states classify events in accordance to the following three parameters:
4
.. i -
1.
- Large Loss of Coolant Accident "S"
- Small Loss of Coolant Accident "T"
- Transient 2.
Timing of Core Melt and Conditions.at Vessel Failure
" E"
- No RWST Injection to RCS "L"
- No Emergency Feedwater "FW" - With Emergency Feedwater 3.
Availability of Containment Systems "C"
- Long-Term Containment Spray Cooling "4"
- Long-Term Spray Recirculation, No Cooling "I"
- Isolation Failure or Bypass Figure 3.1 gives th'e definition of the plant damage states and their re-spective frequencies listed in Table 3.2 as used in the SPSS risk model.
These damage states are categorized in a matrix of 8 physical conditions in the containment (numerals (1) to (8)) and 6 combinations of containment safety i
function availability (letters A to F) for a total of 48 potential plant dam-l age states. A ninth damage state type has been defined for accident sequences l
involving steam generator tube ruptures.
Figure 3.1 indicates that only 39 plant damage states can be identified as credible sequences, i
These plant damage states are the starting point for the containment event tree analysis and they define the link or interfaces with the plant analysis.
I e -
Cc4Difm3%S cQ17Assew.mf auf ACT Af f,M. of Co#. W.Lf Staaf v t 71W.of A. ACTO 4 a
58 LU.LT FMacuGee Y.S No c"""'""u"c'*"5""
ec....te r.
.. um,..
n
$";c8 g a
'",".'lE'"
',.,"c! *',*
./Jo*a
- "'*""a s
a.ar over
.. o m g m,s,
. om
. nuo o,..
o on.. o
...c A,1 o=L,v Otaw.a w cwf asFiSauawcw a.uoval onLv Caasu 19 ett los f.I T. e e8,'
seas eat MN M,N
()
"kkMM (3
@)
(25)
(83
...t.
f6sssyE
.:Mf6%
sjffdit l
A^
==
~
j (k.
(@
M vis m if.$.$.yg
/ $ff_ _gp
-Jp((fs.yj no r
Da)
,1 (ac)
(d (a)
(s)
($*)
6'l ni e
$$lds$,
62V#
('[o)
!$$<f/A 6
- 'n
( e, i)
(u)
(..)
/ ")
v.: m
(.4)
'?l,'.7'N.".'Z7,'#."*...
I)
4c)
Do'.'
b:v:M] "R"n!"';;ig,
[S#)% l "f,' *,,"ct,",can',o',7,,'"ar'a l
l ;;5;;>'.',',',',%;A""' *'"'a Plant State Represents 2A/6A AEC 4A/8A TEC.SEC 2C/6C AE4 4C/8C TE4 10 AE 20/60 AL 30/70 SE TE/TEFW t
40/80 SL.TL 2E/6E AECI 4E/8E TECI 1F V
2F/6F AEI 3F/7F SEI 4F/8F SLI Figure 3.1 Definitions of the plant damage states used in SPSS.
. Table 3.2 Frequencies of occurrence of the plant damage states Frequency Frequency Plant Damage (events'per Plant Damage (events per State reactoryear)
State reactor year) 10 3.03-7) 6A 3.41 -7) 1F 1.89 -6) 6C 3.57 -10) 1FA 6.10 -11) 6D 2.49 -7) 1FP 8.52(-7) 6E 5.30(-14 2A 1.85(-6) 6F 2.08(-16 2C 1.91(-9) 6FA 1.11(-11 20 2.53 -7) 6FP 1.34(-12 2E 1.40-13) 70 7.06(-5) 2F 1.06 -13) 7F 3.55(-8) 2FA 3.10 -1 )
7FP 1.09 -5) 2FP 1.58 -1 )
8A 4.50 -
30 1.94 -5 8C 4.29 -
3F 5.00 -
80 5.51 -
3FP 6.21 -
8E 5.02-11) 4A 1.28 -
8F 1.02 -10) 4C 1.65 -
8FP 1.95 -7) 40 2.79 -
9A 7.51-10) 4E 2.24-11) 9C 3.62 -13) 4F 2.25 -13) 90 9.09(-9) 4FP 1.18(-7)
TOTAL 2.30(-4)
NOTE: Exponential notation is indicated in abbreviated form; i.e., 3.03(-7) = 3.03 x 10-7,
. 3.4 Containment Event Tree and Accident Phenomenology An important step towards the development of the containment matrix in-volves the quantification of branch point probabilities in the containment event tree.
These probabilities depend heavily on the analyses of degraded and core melt phenomenology and the containment building response described in Appendix H of the SPSS.6 The SPSS toitainment event tree uses the twelve top events identified in Table 3.3 t <.djor phenomenological phases which could occur with respect to the formation and location of core debris.
Specifically, these processes are grouped into four phases, namely:
(1) phenomena occurring while the core is still in place: (2) phenomena occurring while the core is located below the lower grid plate but is still in the reactor vessel; (3) phenomena occurring with the core debris located in the reactor cavity and on the containment floor; and (4) the phenomena involving long-term cooling of the containment and/or basemat penetration.
3.5 Containment Matrix (C-Matrix)
The twelve top events in the Seabrook containment event tree are sum-marized in Table 3.3.
A negative response at any of the five nodes (4, 8,10, 11, and 12) in the containment event tree results in the failure of the con-tainment building by a variety of failure modes.
Each of these failure modes results in a particular radiological release category.
For those paths that do not have a negative response at any of the five nodes, the path will even-tually result in no failure of the containment.
The containment event tree then links the plant damage states to a range of possible containment failure
1
. Table 3.3 Top events for the Seabrook containment event tree
. Number Top Event Accident Phase 1
Plant State Initiator 2
Debris Cooled in Place Debris in Vessel 3
No H Burn 2
4 Containment Intact 5
Debris Dispersed from Debris in Reactor Cavity Cavity 6
Debris Cooled 7
No H Burn 2
8 Containment Intact 9
No Late Burn Long-Term Behavior 10 Containment Shell Intact 11 Basemat Intact 12 Benign Containment Failure Failure Mode (Small Leak)
. modes via the various paths through the tree.
For a given tree, each path
' ends in a conditional probability (CP) of occurrence, and these cps should swa to unity. The quantification of an event tree is the process by which all the paths are combined to give the conditional probabilities of the various re-
~
lease categories.
In SPSS, fourteen release categories are used for the quan-tification as summarized in lable 3.4.
Note that two of these release cate-gories (namely S5 and 55) ccrrespond to intact / isolated containment. Fission product release for this category would, therefore, be via normal leakage paths in the containment (and enclosure) building.
Table 3.5 is a reproduction of the Canatrix for the SPSS.
It lists the conditional prooabilities of the release categories given the plant damage state.
A simplification of the C-matrix is obtained in Table 3.6 by disregarding all the very low probability values.
BNL staff do not have the opportunity to perform independent confirmatory calculations of accident progression and containment response.
Instead, the l
knowledge gained from in-depth review of similar risk studies " is used to guide this assessment.
The mode and timing of containment failure cannot be calculated with a great degree of accuracy.
Judgements must be made about the nature of the dominant phenomena and about the magnitude of several important parameters.
Furthermore, the codes and methods used for these calculations are approximate and do not model all of the detailed phenomena.
Fortunately, risk estimates are not sensitive to minor variations in failure mode and timing.
It is im-portant, however, to properly characterize the major attributes of failure
. l Table 3.4 Release categories employed in the Seabrook Station risk model Release Category Release
- Group Category Definition SS Containment intact / isolated with enclosure Containment air handling filtration working.
Intact / Isolated
^
S5 Same as SS but with enclosure air handling filtration not working.
S2 Early containment leakage with late over-pressurization failure and containment building sprays working.
If Same as S2, but with containment building spray not working.
T2V Same as Tf, but with an additional vaporiza-tion component of the source tern.
S3 Late overpressurization failure of the con-Long-Term tainment with no early leakage and contain.
Containment ment building sprays working.
Failure U
Same as S3, but with containment building sprays not working.
53V Same as U, but with an additional vaporiza-tion component of the source term.
S4 Easemat penetration failure, sprays operating T4V Containment basemat melt-through with con-tainment building sprays not working and ad-ditional vaporization component of the source term.
S6 Containment bypass or isolation failure with containment building sprays-working.
T67 Same as S6, but with containment building sprays not working and an additional vapori-Early zation component of the source term.
Containment Failure / Bypass S1 Early containment failure due to steam explo-sion or hydrogen burn with containment building sprays working.
H Same as S1, but with containment building sprays not working.
- S denotes applicability to Seabrook Station: number corresponds with contain-ment failure mode: bar denotes nonfunctioning of containment building sprays; and V denotes achievement of sustained elevated core debris temperatures and associated vaporization release.
1 e.
Table 3.5 Seabrook C matrix p
Release Category state 51 52 53 55 56 If U
Il 55 Ut UV Rv MW Ev-d ID 4.9-5 8.3-5 0.60 0.40 1r IFP 1.0 1.0 IFA 1.0 8I I
2A/64 3.4-5 1.4-4 1.0-2 g 0.990 5.0-6 t
2C 3.9-5 1.1-4
.9999 2D/60 3.9-5 1.1-4 0.9998 1.0-4 CI I
2U6E 1.0 2F/t/
1.0(dl 2F P/6FP 1.0 2FA/6FA 1.0(al 3D/70 2.0-6 7.5-5 0.89 0.11 3F/7F 1.0 3FP/7FP 1.0 4A/8A 3.1-6 1.3-4 5.2-3(b) o,995 5.0-6 s
4C/8C 1.2-6 6.8-5 1.0 N
40/80 1.1-6 3.1-5 0.9999 5.2-5ICI
?
4E/8E 1.0 4F/8F 1.0(d) 4FP/8FP 1.0 9A 1.0 9C 1.0 90 1.0 a.
Consequences from 51 are greater than from II.
b.
Conservative assignment instead of 54.
c.
Conservative assignment instead of 54 d.
Conservative assignment instead of M.
NOTES:
1.
A blank indicates 0.0.
2.
Exponential notation is indicated in abbreviated fons; i.e., 4.9-5 = 4.9 x 10-5, 9
~.
Table 3.6 Simplified C-matrix for SPSS Release Category Plant State 51 52 53 55 56 32 Yi SiV IIV T4V
'EV Ed ID 0.60 0.40 IF 1.0 IFP 1.0 IFA 1.0 2A/6A 1.0-2 0.99 2C 1.0 2D/60 1.0 2E/6E 1.0 2F/6F 1.0 2FP/6FP 1.0 to 2FA/6FA 1.0 y
3D/70 0.89 0.11 3F/7F 1.0 3FP/7FP 1.0 4A/8A 5.0-3 0.995 4C/8C 1.0 4D/80 1.0 4E/8E 1.0 4F/8F 1.0 4FP/8FP 1.0 -
9A 1.0 9C 1.0 90 1.0
. mechanisms; (1) whether the failure is early or late, (2) whether it is by overpressurization, bypass, or basemat melt-through and, (3) whether or not radionuclide removal systems are effective.
The assessment of the containment response and failure mechanisms is based on the general understanding of the accident phenomenology and the con-i tainment design characteristics discussed earlier.
The phenomena of interest may be summarized as follows:
Early Failure (S1,3T) which can result from a steam explosion or an early hy-drogen burn are believed to be unlikely.
Although explosions in the reactor vessel lower plenum are highly probable, the resulting mechanical energy would be limited by the fraction of the core which could participate in a single ex-plosion and by the efficiency of the process.
In recent PRA reviews,4,7 we have assigned a conditional probability of 10-4 to steam explosion induced containment failure; however, steam explosions would have negligible effect on risk, and consequently the applicants 5 x 10-4 value is not included in the simplified C-matrix.
The conditional probability for an early containment failure due to ex-ternal events (i.e., aircraft crashes) is assigned 1 in the SPSS as shown in Table 3.6.
This simply indicates that an aircraft crash into the con linment is assumed to fail the containment structure with certainty, which is believed to be conservative.
Early Containment Leakage (S2, T2, T2V) without the gross failure of the con-tainment building is only expected to be likely for the non-isolated steam
i-i.
31-l
}
4 generator tube rupture event with containment sprays available (S2), and for j
large break LOCA sequences'with RWST injection in the absence of sprays (S2),
~
and for dry cavity sequences with a vaporization release (SE).
i j
late Overpressurization Failure (S3, Tf, ~53V) can occur due to steam produc.
i tion in a wet cavity or noncondensible gas production as a result of core-con-j crete interaction for a dry cavity situation.
For sequences in which early i
and intermediate failure is not expected to occur, and for which containment j
sprays are inoperable, failure is expected to be a certainty.
The conditional probability for a late overpressurization failure with a vaporization release (dry cavity) is shown to be 0.60, this results from the j
relative competition between the late overpressure failure and the basemat penetration (S4Y) for accident sequences without the containment sprays.
l!
~
The failure time for the late overpressurization failure mode is much i
longer than previously calculated for other large dry containments.[1 8]
This is as a result of the very high failure pressure for the Seabrook con-tainment.
As a consequence of this high containment failure pressure, (median pres-sure of 211 for wet and 187 psia for dry
- sequences) it is difficult to challenge the containment integrity by any conceivable event.
Hydrogen def-4 j
lagration either, early in the accident sequence or later after vessel failure when steam condensation occurring as a result of reactivation of sprays (due i
I l
- For dry sequences, only primary system water inventory is available in the containment.
In this case, containment atmosphere becomes superheated and, at failure, the temperature can exceed 700'F.
I i
- m--+-
~.-,.-w,..
_--,wy...,m,--,,m,
.eww._mm.,w-.,-,,-em,-.,,-v,,,w,,,,,m,,
- m. mw--ww
. to regaining of AC power), or other natural heat sink mechanisms,7 which can produce a deinerted atmosphere can not challenge the containment integrity.
Basemat Penetration Failure (S4,lDR) can only result in the absence of con-tainment heat removal systems (sprays) for a dry cavity.
The conditional probability of the basemat melt through is always less than the late overpres-surization failure. This is believed to be a conservative assessment.
No Failure (SS 55) would result for all sequences with full spray operation.
The radiological releases are thus limited to the design basis leakage with essentially negligible off-site consequences.
Containment Isolation Failure (S6,13Y) is represented by an 8-inch diameter purge line.
The accident sequences where the containment is either not isola-ted or bypassed (Event V) are assigned a conditional probability of unity to this release category.
An interfacing systems LOCA (V sequence) results from valve disc rupture or disc failing open for series valves that normally separate the high pres-sure system.
This event results in a LOCA in which the reactor coolant by-passes the containment and results in a loss-of-coolant outside the contain-ment.
Furthermore, the concurrent assumed loss of RHR and coolant make-up capability leads to severe core damage.
In the SPSS, three possible inter-facing systems LOCA sequences have been found and discussed.
These are 1.
01sc rupture of the check valves in the cold-leg injection lines of the RHR.
1
. l 2.
Disc rupture of the two series motor-operated valves in the normal RHR hot-leg suction.
3.
Disc rupture of the motor-operated valve equipped with a stem mounted limit switch and " disc failing open while indicated closed" in the other motor-operated valve in the nonnal RHR hot-leg suction.
For the V-sequence, the core melts early with a low RCS pressure and a dry reactor cavity at vessel melt-through.
The low pessure injection system injects the RWST into the RCS, which discharges into the RHR vaults outside the containment.
The sump remains dry and recirculation is not possible.
The core and containment phenomenology used to arrive at the split frac-tions for the containment event tree and thus the C-matrix are in general agreement with the other previously reviewed studies.[ '3'"3 Further.
more, the unusually high strength of the Seabrook containment, reduce the im-pact of sensitivities caused by uncertainties in the accident progression as demonstrated in the SPSS.6 However, should the claimed strength of the con-tainment be reduced to levels comparable to other large dry containment struc-tures, the impact of uncertainties on accident phenomenology may become signi.
ficantly more pronounced, as discussed in other recent PRA reviews.7 3.6 Accident Source Terms In this section the approach utilized in the Seabrook Probabilistic Safety Study (SPSS) to determine the fraction of fission products originally in the core which can leak to the outside environment will be outlined.
The fission product source to the environment as calculated by t'ils approach for each release category will also be discussed.
- - - _ _ - ~.
I 4
i i
As in the Reactor Safety Study (RSS)9 the CORRAL-II code was used in the l
SPSS for determining fission product leakage to the environment.
This code 1
takes input from the thermal. hydraulic' analysis carried out for the contain-
}
ment atmosphere.
In addition, it needs the time-dependent emission of fission products. 'The fission products were assumed to be released in distinct phases as suggested in the RSS, namely, the Gap, Melt, and Vaporization phases.
The i
time dependence of these phases is determined by the timing of core heatup, l
primary system failure, and core / concrete interaction.
The methods used in i
j the SPSS differ from the RSS methods in the following ways:
1)
The treatment of iodine was changed and iodine was treated as cesium j
iodide.
This was accomplished by merely using the same fraction of 1
core inventory released for both the cesium group and the iodine 3
- group, 1
2)
Leakage releases are represented by a multi. puff model, 1
3)
An uncertainty analysis was carried out in which it was attempted to i
j account for shortcomings in the RSS methods.
In general, the net re-r I
sult of the SPSS analysis was to reduce the fractional release of j
particulate fission products.
This will be discussed in more detail i
later.
In all, fourteen releases were determined ranging from con-j tainment bypass sequence to the no-fail sequence.
i These release categories were evaluated by considering the containment l
l failure mode, the availability of the spray system, and whether or not the i
cavity was wet or dry.
Table 3.4 shows a description of the release catego-3 ries and Table 3.7 shows the point-estimate releases as determined by the i
j methods outlined above.
Containment failure mode Si corresponds to a gross 1
1
Table 3.7 Seabrook point-estimate release categorie5 Ma Ant Release Time Duration Warning Time Energy Release Cft S#4# #
I"*"'5I Ih*d'5I Ih*"f5I*
10' N/
Release Fractions by Group r
-t e s) xe 1-2" Cs Te Ba Ao La M
MC 1.9 0.5 0.35
< 10.0 10.
.94
.023
.023
.24
.0033
.41 9.8-5 52 AlC 2.6 1.0 1.9
< 10.0 10.
.89 215 2.1 5 4.4 6 2.9-6 8.8-7 8.8-8 51 IC 4 66.1 0.5 62.5 210.0 10.
.90 1.0-7 1.0-7 1.9 8 1.3-8 3.8-9 3.8-10 15 ric 1.9 24 0.35
< 10.0 10.
.0091 3.5-8 3.5-8 6.1 9 4.0-9 1.2-9 1.2-10 ItCl 4.5 4.0 4.0
< 10.0 10.
.90
.0036.0036
.00067.00044
.0001) 1.3 5 if At 1.4 0.5 0.3 210.0 10.
.94
.75
.75
.39
.093
.46
.0028 5!-I 7.3 U-2 9.1 6.2
< 10.0 10.
.15
.092
.092
.017
.011
.0034
.00034 20.3 17.
19.2
< 10.0 10.
.24
.093
.093
.017
.012
.0034
.00034 U-3 29.3 1.2 29.2
< 10.0 10.
.51
.12
.12
.023
.015
.0046
.00016 5? Total A1.
7.3 27.3 6.2 10.
.90
.31
.31
.057
.038
.011
.0011 e
51 At 27.2 0.5 26.4 210.0 10.
.90
.12
.12
.022
.015
.0044
.00044 g
55 IEC 4.3 24 0.6
< 10.0 10.
.014 5.2-7 5.2-7 9.5-8 6.3-8 1.98 1.9 9
)
5.TI-I 2.2 1.5 1.9
< 10.0 10.
.05
.037
.037
.0069
.0045
.0014
.00014 SIl-2 6.2 7.2 5.9
< 10.0 10
.10
.012
.072
.0080
.0079
.0062
.0010 124-3 35.2 78.0 34.9
< 10.0 10.
.85
.20
.20
.30
.022
.018
.0030 52r Total AE 2.2 68.7 1.9 10.
1.0
.31
.31
.32
.034
.025
.0012 pri TE 81.5 0.5 76.2 210.0 10.
1.0
.024
.024
.030
.0026
.0023
.00039 Mt aC 50.0 0.5 49.6 210.0 10 1.0
.058
.058
.072
.0062
.0054
.00031 SDr-1 SC4-2 2.2 1.0 1.7
.35 10.
.15
.11
.11
.020
.014
.0041
.00041 4.2 3.0 3.7
.31 10.
.31
.14
.14
.026
.017
.0052
.00051 ist-3 11.2 10.0 10.7
.21 10.
.51
.18
.18
.36
.017
.024
.0044 Cd Total 5El 2.2 14.0 1.7
.26 10.
.97
.43
.43
.40
.018
.033
.0053
- 31E
Esponential notation is shown in abbreviated form; 1.e., 2.1 5 = 2.1 a 10-5,
- ilcental ioJine - not used, all todine is treated as Cst *aascJ on time of gap release eicept for 56 and 35V where it is ba t
e
' failure of the containment, resulting from a steam explosion, early pressure spikes, or early hydrogen burns.
Failure mode S2 represents a loss of con-tainment function early in' the accident sequence.
This loss of function takes the form of an increase in the leak rate to 40% per day where it stays until the containment fails due to overpressurization.
Failure mode S3 represents a late overpressurization failure of the containment driven by decay heat or late hydrogen burns.
Failure mode S4 represents a basemat melt-through, S5 represents no containment failure and the leak rate is limited to the contain-ment design basis leak rate.
Finally, failure mode 56 represents sequences where the containment is failed or bypassed as part of the initiating event.
The second parameter considered in defining the source term is the avail-ability of sprays. This is determined by the plant damage states.
Those re-lease categories with operating spray systems are designated Si to S6, while thosewithspraysystemsnotoperatingaredesignated5ItoS3I.
The third and final parameter considered in differentiating between source terms distinguishes between wet and dry cavities.
In the case of dry cavities a vaporization release due to core / concrete interactions will occur, while for wet cavities the core debris is assumed to be quenched or the water in the cavity will scrub the vaporization release thus effectively reducing the release to zero.
The release categories which include a vaporization re-lease include a "V" in their designation as shown in Table 3.4.
Fran the point of view of risk it was founds that S2V, 53, 53V, and 56V were dominant either for acute or latent health effects.
In view of this re-sult these four categories will be considered in more detail.
Release categories S3 and S3V have late overpressurization failure modes, with no spray systems operating and differ only in the omission or inclusion
. of a vaporization release, respectively.
The containment at Seabt ook is cal-culated to fail at a median pressure of 211 psia for ' wet sequences and 187 psia for dry sequences.
At this pressure a gross failure is expected re-sulting in a puff release of approximately.5 hr release duration.
From Table 3.7 it is seen that the T3 and S3V sequences fail at 27.5 hrs and 81.5 hrs.,
respectively. These failure times are several hours later than was calculated for Indian Point, Zion, and Millstone-3.
The primary reason for the later failure in this case is due to the superior strength of the containment struc-ture.
Table 3.8 compares the U, 3Ti release parameters with similar para-meters for the other three reactors mentioned above.
Note that a fair com-parison should set (0l+10) equal to (Cs-Rb), since iodine was treated as Cs!.
It is seen that I, Cs, and Ba groups for U are approximately half the other releases, while the Te, Ru, and 1.a groups are low by approximately an order of magnitude.
This difference is due to the latter failure time, allowing more time for settling and the absence of a vaporization release, which dominates the release of Te, Ru, and t.a.
A similar comparison for the 53 release indi-cates a uniform reduction of approximately an order of magnitude for all species.
The reduction is entirely due to the late failure time for this sequence.
Release category 527 is associated with early containment failure in which the containment function is compromised by increasing the leakage area in such a way that the leak rate increases from.1% per day to 40% per day.
This release rate is not enough to prevent an ultimate overpressuriza-tion failure.
This release is modelled as a multi. puff release.
The first puff corresponds to the release up to the time when vaporization starts (melt + gap).
The second puf f includes the period of vaporization release and the third puff is equivalent to an overpressurization failure at the time of
+ e Table 3.8 Late overpressurization failure comparison M111 stone-37 Zion / Indian 5 Indian 3 Seabrook8 Point Study Point TI T37 M-7 TMLB' 2RW Xe 9.0(-1) 1.0 9 (-1) 9.6(-1) 1.0 10+1 1.2(-1) 2.4(-2) 1.5(-1) 1.05(-1) 9.3(-2)
Cs-Rb 1.2(-1) 2.4(-2) 3.0(-1) 3.4(-1) 2.6(-1)
Te-Sb 2.2(-2) 3.0(-2) 3.0(-1) 3.8(-1) 4.4(-1)
Ba-Sr 1.5(-2) 2.6(-3) 3.0(-2) 3.7(-2) 2.5(-2)
Ru 4.4(-3) 2.3(-3) 2.0(-2) 2.9(-2) 2.9(-2)
La
'4.4(-4) 3.9(-4) 4.0(-3) 4.9(-3) 1.0(-2)
T(release).
27.2 81.5 20 (hrs)
T(duration) 0.50 0.50 0.50 0.50 (hrs)
Energy 300E7 300E7 540E6 150E6 (Btu /hr) 9 s
. catastrophic containment failure.
In this model the duration of the melt re-lease is seen to be 3.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, vaporization release 7.2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> and the re-maining release 78.0 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
It is nots clear that the melt release in this case is 5.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />, however, it does not seem to be unreasonable.
A 7.2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> duration for the vaporization release is not consistent with the RSS,9 which only allows 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> for this phase. Finally, it is not clear how the 78 hours9.027778e-4 days <br />0.0217 hours <br />1.289683e-4 weeks <br />2.9679e-5 months <br /> for the last phase was determined.
The release duration for a single puff, e
which is the sum of the above three phases leads to a release time of 88.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> which seems extraordinarily long. Our recommendation would be to reduce these times to be more consistent with RSS methods.
The total release of fission products from the sequences can be compared to the M-4 release determined for the Millstone-3 study.
This comparison is made in Table 3.9.
It is seen that, once adjustments are made for the dif-ferent ways in which iodine are treated, the S2V release is approximately half the M-4 release.
Without the benefit of a calculation, it is difficult to judge whether_the differences are reasonable.
However, a possible reason for this reduction is the credit taken for the enclosure building surrounding the actual containment building.
This feature is unique to the Seabrook contain-ment structure.
Release category 55V has binned into it an isolation failure corres-ponding to an 8" diameter breach in containment and the interfacing LOCA (V-sequence).
This sequence is also represented by a multi-puff release.
In l
this case as in the previous case, the total release time is long compared to acceptable limits of the RSS9 consequence model.
Our recommendation would be to reduce these times to more reasonable values.
[
. The release fraction can be compared (Table 3.9) to the M-4 release from Millstone-3, PWR-2 for the RSS and the V-sequence from the RSSMAP study for Surry.10 Except for the iodine group, it is seen ~ that the release fractions
~
are comparable.
If the iodine group were set equal to the cesium group value, it is seen that the value for S6V would be the lowest release fraction.
3.7 Source Term Uncertainty Analysis In this section we will briefly describe the uncertainty analysis carried out for the four dominant accident sequences and, where possible compare the fission product leakage to the environment to more mechanistic determina-tions. There are two contributors to the uncertainty in release characteriza-tion. First, uncertainty in time parameters which are influenced by:
- 1) Prediction of key event times, and
- 2) The mix of accident sequences binned into a release category.
Second, uncertainties in release fractions, which are influenced by:
- 1) Analysis methods and data, and
- 2) Uncertainties in timing of key events.
The above principles were used to determine source term multipliers which would give a range of fission product leakage to the environment. A probabil-t ity is associated with each source term, and for late overpressurization fail-ure modes (S3, S3V, and S2V) the following discrete probability distribution is used, i.e.,
Subcategory Probability U-a
.02 U-b
.08 SI-c
.30 U-d
.60 l
l l
Table 3.9 Comparison of releases for failure to isolate containment and the by-pass sequence w
Seabrook6 Millstone-37 RSS
- RSSMAP 10 9
377 357 M-4 PWR-2 V-Sequence Xe 1.0 9.7(-1) 9.0(-1) 1.0 1.0 OI+I 3.1(-1) 4.3(-1) 2.0(-1) 7.0(-1) 4.8(-1)
Cs-Rb 3.1(-1) 4.3(-1) 6.0(-1) 5.0(-1) 7.9(-1)
Te-Sb 3.2(-1) 4.0(-1) 5.0(-1) 3.0(-1) 4.4(-1)
Ba-Sr 3.4(-2) 4.8(-2) 7.0(-2) 6.0(-2) 9.0(-2)
Ru 2.5(-2) 3.3(-2) 5.0(-2) 2.0(-2) 4.0(-2)
La 4.2(-3) 5.3(-3) 7.0(-3) 4.0(-3) 6.0(-3)
(release) 2.2 2.2 0.20 2.5 2.5 T
t rs) h T(duration) 88.7 14 2.0 1.0 1.0 (hrs)
Energy (Btu /hr) 140E6 4E6 70E6 20E6 0.5E6 1
- The same as M1A release category in Millstone-3.7 4
4
- This indicates that.there is an 8% confidence level that 53-b correctly de-fines the source term for the 53 release category.
The results of this analysis for the overpressurization failure modes is:
Particulate Release Factor (multiplier)
Pr'obability M
UiT S2V
.02
.22-
.63
.17
.08
.071
.22
.07
.30
.024
.065
.02
.60
.0071
.021
.007 Fran this-table it is seen that for the most likely release, i.e.,
"d",
the reduction factors of the source term are substantial.
The first two releases can be compared to releases published in BMI-2104 Volume V (Surry) for the TMLB'-c and AB-c sequences.
These two sequences cor-respond to late containment failures and are both binned into S3 or S3V se-quences.- A comparison of these sequences is shown on Table 3.10.
From this table it is evident that for the volatile species, Xe, Cs, and I, the release categories S3 and 53Y bracket or exceeds the mechanistic estimates carried out i
in.BMI-2104 for both TMLB' and AB sequences.
However, for the less volatile species Te, Ba, Ru, and La the releases of the AB sequence is the only one bracketed or superseded by the 53 and S3V releases.
The release fraction de-.
termined for the TMLB' sequence are higher than all the 53' and S3V releases.
This discrepancy is primarily due to the comparatively early failure time.
It is felt that agglomeration and settling would reduce the source for the TMLB'
+
4 s
Table 3.10 Comparison of AB-c and TMLB'-c (BMI-2104) to STf and TT Release Fractions Release.
Probability Release Category Time (hrs)
Xe Cs I
Te Ba Ru La S3V-a
.02 28 1.0 1.5(-2)
- 1. 5(.-2) 1.9(-2) 1.6(-3) 1.5(-3) 2.5(-4)
.08 36 9.0(-1) 5.3(-3) 5.3(-3).
6.6(-3) 5.7(-4) 5.1(-4) 8.6(-5)
.30 54 8.0(-1).
1.6(-3) 1.6(-3) 2.0(-3) 1.7(-4) 1.5(-4) 2.5(-5)
.60 89 7.0(-1) 5.0(-4) 5.0(-4) 6.3(-4) 5.5(-5) 4.8(-5) 8.2(-6)
D.
Ti-a
.02 22 1.0 2.6(-2) 2.6(-2) 4.9(-3)'
3.3(-3) 9.7(-4) 9.7(-5)
SI-b
.08 28 9.0(-1) 8.5(-3) 8.5(-3) 1.6(-3) 1.1(-3) 3.1(-4) 3.1(-5)
Si-c
.30 34 8.0(-1) 2.9(-3) 2.9(-3) 5.3(-4) 3.6(-4) 1.1(-4) 1.1(-5)
Ti-d
.60 53 7.0(-1) 8.5(-4) 8.5(-4) 1.6(-4) 1.1(-4) 3.1(-5) 3.1(-6)
TMLB'-c 12 1.'0 2.8(-3) 6.0(-4) 8.5(-2) 1.7(-2) 2.4(-5) 4.3(-4)
AB-c 24 1.0 4.8(-5) 4.7(-5) 4.0(-5) 4.9(-5) 2.4(-7) 3.6(-5) l
, m \\
e
,+
w sequence to values close to those reported for S3 and S3V. No comparative se-quence for S2V was analyzed in BMI-2104.
In the case of the S6V release category a different probability distribu-tion was used.
This change reflects the break location, which initiates the V-sequence. This break could be either in the hot-leg (b release subcategory) or the cold-leg (c release subcategory).
This sequence is modeled as a multi-puff release and each puff is treated separately.. In this comparison only the sum of the release will be considered, since no adequate method of analyzing a multi-puff release is readily available.
Table 3.11 shows a com-parison between the totals of the various S6V releases and two V-sequence ~ re-leases computed for Surry and published in BMI-2104.
One of the V-sequences is " dry", implying no water in the path of the release and the other is " wet",
implying that the releases passes through 3 feet of water befor'e entering the atmosphere.
From this comparison it can be seen that all the releases, except Cs for the " dry" V-sequence are bracketed by the S6V releases.
3.8 Off-Site Consequence Analysis The site consequence analysis aims at estimating the number of health and economic effects (consequences) in the population surrounding the Seabrook 1
Station due to radioactive atmospheric releases as a result of a core melt accident.
In SPSS a computer model called CRACIT (Calculation of Reactor Accident Consequences Including Trajectory) is used for this purpose.
The CRACIT code requires in'put for site specific data (population distribution, economic parameters, topography), meteorological data (wind case, wind speed, rain),
o
[
Table 3.11 Comparison of'S6V (sum) to V-sequence (Surry) 1 Release Fractions i
Release Probability Category Xe Cs I
Te Ba Ru La T6V-a
.02
.97 4.3(-1) 4.3(-1) 4.06(-1) 4.2(-2) 3.32(-2) 5.3(-3)
S6V.b
.45
.97 2.95(-1) 2.95(-1) 1.36(-1) 3.53(-2) 1.52(-2) 2.0(-3)
L I
- --S6V-c
.45
.97 1.295(-1) 1.295(-1) 3.2(-2) 1.593(-2) 5.2(-3) 5.3(-4)
.08
.97 5.2(-2) 5.2(-2) 1.3(-2) 6.6(-3) 2.0(-3) 2.2(-4)
V (dry) 1.0 5.52(-1) 1.99('1) 1.2(-1)
V 1.0 1.04(-1) 3.84(-2) 2.5(-2) i
^
- Individually not reported.
i
o o
4s-
\\
plume characteristics (isotopic content, physical description), and population
~
response (evacuation parameters).
A detailed review of the SPSS site-consequence approach will be made by l
the NRC staff, and is therefore not addressed in this report.
3.9 Risk of Two-Reactor Station The SPSS uses the results of the single unit assessment to arrive at the integral statement of risk for the two-unit station as described in Section 13.3 of the report.
The multi-unit event sequence model takes into account the various depen-dencies and common cause factors in arriving at th'e sequence frequencies.
The consequences of the double-reactor unit accidents were estimated by scaling up the mean S matrix damage scale by a factor of 2.
With respect to the timing of the releases, the double reactor events were modeled as simul-taneous releases. This is considered to be a' conservative assumption.
The resulting risk estimates for the early and latent health effects for the two-unit station at Seabrook shows that:
(a) neglecting the double-unit events, the early fatalities would not have been underestimated by an appre-ciable amount of accident frequencies greater than 10-9 per plant year; and I
(b) for' latent health effects, the single-reactor events also make the greatest contribution over much of the range, and dominate the results above 10-5 events per station year.
.