ML20215N312
| ML20215N312 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 10/31/1986 |
| From: | Devincentis J PUBLIC SERVICE CO. OF NEW HAMPSHIRE |
| To: | Long S Office of Nuclear Reactor Regulation |
| References | |
| SBN-1225, NUDOCS 8611050225 | |
| Download: ML20215N312 (157) | |
Text
f i
SEABROOK STATION
~
Engineering Office October 31, 1986 8
SBN-1225 T.F.
B7.1.2 N;w Hampshire Yankee Division United States Nuclear Regulatory Commission Washington, DC 20555 Attention:
Mr. Steven M. Long, Project Manager PWR Project Directorate No. 5 Division of PWR Licensing - A
References:
(a) Facility Operating License NPF-56, Docket No. 50-443 (b) USNRC Letter, dated October 8,1986, " Request for Additional Information for Seabrook Station, Units 1 and 2, Emergency Planning Sensitivity Study", S. M. Long to R. J. Harrison (c) USNRC Letter, dated October 23, 1986, " Request for Additional Information for Seabrook Station, Units 1 and 2, Emergency Planning Sensitivity Study", S. M. Long to R. J. Ilarrison Subj ect :
Response to Request for Additional Information (RAIs)
Dear Sir:
Enclosed herewith _ are the majority of the responses of the Requests for Additional Information forwarded in References (b) and (c).
Attachment A identifies responac3 that are included in this transmittal.
Attachment B is the responses.
An additional submittal addressing the remainder of the RAIs will be forthcoming in the near future.
Very truly yours, n
. )"
John DeVincentis Director of Engineering Attachment cc: Atomic Safety and Licensing Board Service List i
Director, Office of Inspection and Enforcement United States Nuclear Regulatory Commission Washington, DC 20555 P
i 8611030225 861031 PDR ADOCK~O5000443 g
f F
PDR P.O. Box 300 Seabrook, NH 03874 Telephone (603)474-9521
f Dirn; Currzn, E;quirc P:ter J. Math ws, Mayor
~~
Harmon & Weira City Hall 2001 S. Street, N.W.
Newburyp 3 rt, MA 01950 Suite 430 Washington, D.C.
20009 Judith H. Mizner Silvergate, Gertner, Baker, Sherwin E. Turk, Esquire Fine, Good & Mizner Office of the Executive Legal Director 88 Broad St.
U. S. Nuclear Regulatory Commission Boston, MA 02110 Tenth Floor Washington, DC 20555 Calvin A. Canney City Manager Robert A. Backus, Esquire City Hall 116 Lowell Street 126 Daniel Street P. O. Box 516 Portsmouth, NH 03801 Manchester, NH 03105 Stephen E. Merrill, Esquire Philip Ahrens, Esquire Attorney General Assistant Attorney General George Dana Bisbee, Esquire Department of the Attorney General Assistant Attorney General Statehouse Station #6 25 Capitol Street Augusta, ME 04333 Concord, NH 03301-6397 Mrs. Sandra Gavutis Mr. J. P. Nadeau Chairman, Board of Selectmen Selectmen's Office RFD 1 - Box 1154 10 Central Road Kensington, NH 03827 Rye, NH 03870 Carol S. Sneider, Esquira Mr. Angie Machiros Assistant Attorney General Chairman of the Board of Selectmen Department of the Attorney General Town of Newbury One Ashburton Place, 19th Floor Newbury, MA 01950 Boston, MA 02108 Mr. William S. Lord Senator Gordon J. Humphrey Board of Selectmen U. S. Senate Town Hall - Friend Street Washington, DC 20510 Amesbury, MA 01913
( ATTN:
Tom Burack)
Richard A. Hampe, Esquire Senator Gordon J. Humphrey Hampe and McNicholas 1 Pillsbury Street 35 Pleasant Street Concord, NH 03301 Concord, NH 03301
( ATTN:
Herb Boynton)
Thomas F. Powers, III H. Joseph Flynn, Esquire Town Manager Office of General Counsel Town of Exeter Federal Emergency Management Agency 10 Front Street 500 C Street, SW Exeter, NH 03833 Washington, DC 20472 Brentwood Board of Selectmen Paul McEachern, Esquire RFD Dalton Road bbtthew T. Brock, Esquire Brentwood, NH 03833 Shaines & McEachern 25 Maplewood Avenue Gary W. Holmes, Esquire P. O.
Box 360 Holmes & Ells Po r t smou t h, NH 03801 47 Winnacunnet Road Hampton, NH 03842 Mr. Ed Thomas Robe rt Carigg FEMA Region I Town Office 442 John W. McCormack F0 & Courthouse Atlantic Avenue Boston, MA 02109 North Hampton, NH 03862
f ATTACHMENT A Responses are included in this transmittal for the following RAIs:
1 12 23 41 55 69 2
13 24 42 57 70 3
14 25 43 59 71 4
15 26 44 60 72 5
16 28 45 61 73 6
17 33 46 62 7
19 34 49 63 8
20 35 50 64 9
21 40 51 67 10 22 53 68 11 Responses to the following RAIs will be forthcoming in an additional submittal:
4 36 54 18 37 58 27 38 65 29 39 66 30 47 74 31 48 75 32 52 1
ATTACHMENT B RESPONSES TO REQUESTS FOR INFORMATION EL
fY
RAI 1
Describe how the overpressurization calculations made by SMA were checked or design _ reviewed.
RESPONSE 1 All SMA calculations on the Seabrook overpressure capacity were performed by Dr. Ralf Peek, currently on the Dept. of Civil Engineering faculty, University of' California Berkeley.
All' assumptions and methods of analysis were reviewed by Dr. D.A.
We'sley.
Numerical checks of calculations for the capacities of the controlling f ailure modea were conducted by Dr. Wesley' and other members of the SMA staf f.
The SMA report was reviewed by Dr. Torri and the SSPSA technical review board. These reviews were conducted and documented as part of a QA program that is described in Section 1 of SSPSA.
In addition a numerical check of all calculations and an independent review of assumptions is currently in progress. This independent review is scheduled for completion on November 25th, 1986.
[.
RAI 2
A meeting should be arranged with the originator of these calculations to assist the BNL reviewers in following'these calculations and under-standing the assumptions.
RESPONSE 2 A=su=ption: need to bc =ade in all engineering calculationc. The difference between the usual design calculations and this proba-bilistic evalnation of the containment integrity is that here the median ultimate strength needs to be estimated. This requires evaluating the load carrying capacity in a limit -state condition in which integrity of the liner is lost.
In most cases this involves large inelastic deformations and extensive redistribution of stresses.
This requires different assumptions compared to design code calculations, and these assumptions must be based on an understanding of the behavior of the structure in its u'timate condition.
The intent of the calculatit.is is to document in a conventional manner what was done and what asst iptinns were made, rather than to explain in detail how the structurr; is expected to behave, and why the various assumptions are justified Dr. Wesley and Dr. Peck participated in a detailed discussion of t e calculational methodology with your staff on October 16th and 17th at BNL.
b
RAI 3
Document the basis for the assumptions in the calculations.
In particular, explain the uncertainty factors assigned to various pressure capacities.
RESPONSE 3 The codes and standards used in design assure large margins of safety beyond the design and test pressures.
Ilowever ultimate capacities are not comput ed in denign. Virtually all dealgn calculations are limited to elastic conditions whereas ultimate load response involves large nonlinear effects.
Consequently, only limited data are available to quantify the uncertainties associated with many areas of the calculated ultimate pressure capacities.
Uncertainty is introduced due to inexact knowledge of the structures' material properties as well as the structure behavior at extreme loads. Where data exist, as for instance on material test properties for Seabrook,.it is a straight-forward procedure to introduce it into the overall variability for a given failure mode.
Other areas such as expected accuracy of analytical methods are estimated based on the judgement of experienced engineers, and are so noted in the calculations and the report.
In addition, as noted in the response to question 1, the bases for assumptions will be independently verified.
RAI 4
Explain the mechanism for transferring the load from the penetration sleeves to the containment wall, inparticular, the equipment hatch, when subjected to high strain conditions. Explain how the rebars around the penetrations were assessed to assure that they can resist these loads in addition to the primary pressure induced loads.
ANSWER 4 At the ultimate containment pressure of 216 psi, the force per unit length along the circumference of the equipment hatch penetration that must be transferred from the sleeve to the containment wall is 18 kips /in. The 1 3/4 in x 10 in annular plate shown in the calcu-lations (Fig 14.5, Section 14, part I, as reproduced from FSAR report),
is more than sufficient to support this load, thus preventing a blow-out of the penetration.
A typical calculation demonstrating that small diameter penetration sleeves do not punch thru the containment wall at ultimate containment capacity will be provided with the response to RAI 31.
[
RAI 5
The calculations use a rebar ultimate stain value of 4.7%, i.e., more than 21 feet of linear extension for the hoop bars.
This linear exten-sion under the high pressure load will be accommodated by formation cracks in the concrete totaling approximately 21 feet in width. Justify the assumption that the pressure loads will be carried proportionately by the linear plate and the rebars (similar to the elastic condition) in this highly cracked condition. Also address the potential for developiong a crack large enough for the local extension of the liner plate to lead to its failure at that point.
RESPONSE 5 Since the strain is distributed in the steel and concrete between the cracks, the total elongation is not all accumulated as open cracks. The reinforcing steel assures the cracks will be essentially uniformly distributed and the crack growth controlled until failure.
Cracking of the concrete and strains above yield in the liner and reinforcing steel assure that the local discontinuity elastic stresses are proportionately less important at internal pressure which result in a stress condition above yield.
Both the reinforcing steel and liner are ductile materials with flat stress-strain curves in the range of ultimate strength. This assures that brittle failures will not occur in isolated elements and the proportional strengths of the rebar and liner can be relied until failure.
The potential for significant liner strain concentration in the vicinity of concrete cracks is considered highly unlikely. The surface of the reinforcing bars is deformed in order to create a bond with the concrete whereas the liner is smooth and is in contact with the concrete on only one side. At the failure pressures of interest, the required coefficient of f riction between the liner and the concrete necessary to achieve the same concentration of strain in the liner as in the reinforcing steel is not considered credible, even if the thermal strains in the liner are neglected.
Inclusion of the thermal strains will increase the required coefficient of friction and for the liner to fail prior to the rebar significantly higher strains must accumulate in the liner since it has higher elongation.
[
RAI 6
Was compatibility of strains in the rebars and the liner plate satisfied in~the calculations? For example, the outermost hoop bars will fail before the inside bars and the liner plate reach their respective ultimate strengths. Was this fact reflected in the calculations?
In addition how is the blaxial stress-strain state of the liner plate considered.
RESPONSE 6 The controlling modes of failure in the concrete structure occur due to membrane tension.
For these modes, all reinforcing bars in a given direction have essentially equal strain at failure and the outermost hoop bars will not fail before the inside bars. The inner bars are likely to have somewhat lower strengths due to the higher temperatures toward the inside surface of the containment.
- However, the strain at which the ultimate strengths of these bars is attained is not significantly altered for temperature increases up to 7000F.
.The bi-axial stress-strain state of the liner was considered in the calculations.
A f actor of 1.73 was computed to account for the expected change in elongation properties for the liner.
RAI 7
The combined tension, shear and bending effect at base and opring line levels was not considered in the calculations (Ref. p. 35, assumption 6).
Verify that the combined effect does not change the conclusions of the analysis.
r RESPONSE 7 Bending and shear ef fects occur not only at the springline, but wherever the hoop reinforcement content changes.
However, these I
stresses are secondary in nature. This means that these stresses are not necessary to carry the internal pressure loads, they merely 3
arise from displacement incompatibilities in the membrane solution.
It was judged that such displacement incompatibilities could be accommodated if necessary by plastic rotations of the containment wall (flexural yielding due to meridional bending stresses) without j
significant loss in the load carrying capacity of the containment wall. The hoop bars especially, which are the most critical for I
carrying the membrane stresses, are unaffected by these plastic I
rotations.
Furthermore, the plastic rotations limit the shear forces which are i
developed. This is important because whereas plastic rotations are not expected to affect the integrity of the liner, the effects of extensive shear cracking could be more detrimental. The possibility 3
1 of failure due to secondary shear stresses was considered for the junction between the base slab and the containment wall.
This is the i
location where such secondary shear stresses were judged to be most l
critical. The estimated median pressure capacity for this failure mode is 408 psig as indicated in Table 3-3 of the SMA report. There-fore failure modes involving secondary stresses in the containment l
wall and dome are not considered critical.
r i
i
}
. _. _ _ _,. _,,=_ __ _ _ ___._
F
RAI 8
Since 31 cadwelds out of a total of 169 test samples failed at a stresa lower than the rebar ultimate strength and there was apparently a construc-tion problem concerning staggering of these welds, provide justification for not using a reduced ultimate strength for the rebar.
RESPONSE 8 There was no construction problem involving the staggering of cadwelds in the containment shell. Cadwelds are distributed throughout the struc-ture as specified in design.
Failure of the containment shell in the controlling hoop direction is expected to begin with vertical cracks forming in the concrete. These cracks are expected to initiate at locations of vertical, rolled section steel supports for the liner which are spaced uniformly approximately 20 inches apart around the circumference. This results in approximately 260 vertical cracks. As the pressure is increased, these cracks wil con-tinue to open until the ' ultimate capacity of the combined reinforcing bars and liner plate is reached across the crack with the lowest capacity. The location of this crack cannot be predicted since the available information predicts essentially equal probability of failure of any crack location.
However, it is important to realize-that failure does not occur due to failure of a single bar. The reinforcing bars and liner develop a ductile system with the ability to provide significant load sharing and load redistribution between the liner and a large number of reinforcing bars. An estimate of the number of reinforcing bars over which failure of a single bar is averaged can be obtained by multiplying the six equivalent bars across the shell thickness by the meridional length required for the pertur-bation damp out.
For the shell in the post yield condition, this length is conservatively estimated at greater than twice the shell characteristic length, or over a hundred bars.
The cadwelds are staggered throughout the structure and there is no way to establish the location of those cadwelds with relatively high or low capaci-ties, just as there is no way to establish the location of the crack with the lowest median rebar strength. From the cadweld test data available for Jeabrook, less than 20% of the cadwelds can be expected to have capacities below the median reinforcing bar strength, whereas by definition, 50% of the reinforcing bars will have strengths below the median. Hence, the test data shows that a conservative value of the median strength was used.
The relatively few cadwelds which may be expected to have lower capacities than the median reinforcing steel are only slightly weaker (less than 10%
reduction). This corresponds approximately to the lower bound reinforcing steel strength so that the minimum cadweld strength and the minimum rein-forcing steel strength may be expected to be approximately equal. Away from the crack, the stress in the rebar decreases due to the concrete bond, and the effect of the cad-weld strength does not influer.ce the strength of the bar once the load in the weld is :lecreased to less than the strength of the rebar across the crack.
I l
)
Even if the cadwelds were completely ineffective, the total reduction in the
~
hoop capacity would be less than 1% due to the averaging effect of the adja-cent bars. Because of the limited number of cadwelds, there is very little probability that more than one cadweld will be located in the same crack in the area where the averaging effect occurs, and an even smaller probability that two cadwelds with low capacities would be so located. Thus, the ef fect of a few randomly distributed _cadwelds with less than the median reinforcing bar strength is considered neglibible, and this effect is accounted for in the variabilities associated with the various structural failure inodes.
I
}
i i
1 i
i t
1
. ~. - -
RAI 9
The containment analysis is based on an axisymmetric geometry and loading. This is not the case due to the presence of adjoining structures such as the fuel building and main steam and -f eedwater pipe chase.
Identify.these axisymmetric conditions and assess their impact on the failure modes and analysis.
}
{
- 6 l
I RESPONSE 9 The effect of l'ocal non-axisymmetric conditions is not expected to effect the capacities computed for the axisymmetric failure modes (i.e. cylinder hoop and meridional membrane.f ailures, dome metabrane
(
failure, and base slab bending and shear failures etc.).
This is because the local effects damp out rapidly for the inelastic case.
Local failure modes such as interference between the fuel storage L
building and the containment were evaluated and found to have significantly j
higher capacities than the controlling axisymmetric f ailure mode (hoop failure).
Also refer to response to RAI 13.
l i
4 m..
..w..,4,-e.e,,-..-
-+-+,c
., - - - -..-.m~.e--.--=.,-ww~.-m,-,-r--------+, * - -
..~-w-
5 r
RAI 10
Only a sample of pipe penetrations are considered in some detail (X-23, X-26 and X-71).
The justification to consider only these should be provided.
RESPONSE 10 Virtually all the piping isometrics were reviewed by members of the SHA staff'and membera af the Yankee Atomic staff. The purpose or this review was to identif y the lines considered most likely to f ail, based on support spacing from the penetration both inside and outside the i
containment.
Based on this review " worst case" lines for a multiple pipe penetration, a thin wall (sch. 40) pipe penetration, and a thick i
wall (sch.160) pipe penetration were selected.
Evaluations of. these
{
" worst case" lines indicated that failure of the penetration (i.e.
t breach of liner integrity) was not likely to occur for the thin l
wall and multiple pipe penetrations, irrespective of relative dis-placement, although fluid leaks or flow restriction in the pipes could occur.
Thus, any further investigation of lines associated with I
this type of penetration was unnecessary. The evaluation indicated that sufficient force could be generated in a thick-walled pipe to potentially fall the penetration.
Ilowever, there are only a few schedule 160 pipes entering the containment and the expected leak path area associated with this type of failure is so small that even i
if all schedule 160 pipe penetrations should fail, the resulting leak area is insufficient to prevent a continuing increase in internal pressure with eventual failure from an independent failure mode.
Therefore detailed evaluations of individual penetrations were not 3
j warranted, once the controlloing penetrations were identified.
RAI 11
A structural evaluation of electrical penetrations should be provided.
RESPONSE 11 The electrical penetrations were reviewed and it was deLeonined their pressure capacity would not be a controlling mode of failure for the following reasons:
penetration are not subject to rigid pipe reaction which interact with containment wall dispacement it was judged that the failure of these penetrations is dominated by thermal ef f ect (i.e. leaks) a thermal analysis was performed in lieu of a structural analysis Consequently, a detailed evaluation was not conducted in order to concentrate on the more likely modes of structural failure codes.
This is consistent with the overall approach used for both the seismic and overpressure capacity evaluations where dotatted investigation is only conducted on conceivable modes of failure.
RAI 12
The basis for the leakage area assigned to the flued head at failure should be provided.
RESPONSE _12_
The leakage area due to failure of the flued head is a rather uncertain quantity as indicated by the large variability assigned to it.
As a median centered estimate, the flued head is expected to fail at a radius 0.5 in less than the inner radius of the sleeve. With a slight increase in containment pressure and the associated radially outward displacement of the containment wall, the part of the fluid head which remains attached to the pipe is pulled into the sleeve. As a result a clearance of 0.5 in is available all around, as a leak area. The logarithmic standard deviation of 0.6 indicates that the 95% confidence intervals for this clearance width are about 0.2 in and 1.3 in respectively.
Thus no claim is made that this leak area can be accurately predicted.
This uncertainty is taken into account in the probabilistic consequence analysis.
RAI 13
A more detailed evaluation of the impact of punching shear at the Fuel Transfer Building should be provided.
1 RESPONSE 13 punching shear stresses at the Fuel Storage Building begin to develop only at a (median) containment pressure of 172 psi. This pressure, at which the containment wall begine to bear against the Fuel Storage Building, is not *very approximate". The main source of uncertainty for this presaura la the uncertainty of the effect of bonding on the stress-average strain relation for tio hoop reinforcing bars, and could not be climinated by a more dettiled analysis. What is "very approximate" is the existing evaluation of the increase in containment pressure that could be supported after the containment pressure begins to bear against the Fuel Storage Bull =ing.
The large uncertainty in this increment in internal pressure is reflected by the large logarithmic standard deviation assigned to this quantity, and has been included in the risk evaluation process.
In doing so, it is found that the failure mode is not critical.
This conclusion is not expected to change even if a more detailed analysis resulted in a slightly dif f errnt value for the median pressure increment or slightly reduced uncertainties for the punchir.,t shear failure.
Furthermore, even if some of the uncertainty in modeling could be removed by a more detailed analysis, the uncertainty in strength would remain.
Finally the lack of knowledge of the behavior of construction materials under such extreme loading condition renders the applicability of the results of the most sophisticated analysis open to debate and is not warranted for this evaluation.
s- -
RAI 14
Clarify the extent to which double ended piping failures have been considered in the overall containment performance assessment. Provide isometric drawings of all piping attached to containment penetrations.
REsPONsK 14 The possibility of a pipe failure on the inside and outside of the containment due to the pressure induced displacements of the containment wall was considered.
For some of the thick walled pipes, it is likely that the penetration will fait prior to the pipe and thus relieve the load on the pipe.
For thin wall pipes, the most likely mode of failure is bending of the pipe thus reducing fluid flow but not resulting in a leak.
llowever, there is some possibility that fracture of thin walled pipes can occur, and a much reduced possibility that a fracture on both sides of the penetration can occur for a given pipe. This possiblity was considered f or penetrations X-23, X-26, and X-71 and found to have a negligible contribution to the overall risk.
For the flued head pipes (penetration X-8) however, the double-ended pipe break failure modo contributed about 55% of the total probability of this type of penetration failing before the concrete structure membrane failure mode.
Isometric drawings of all piping attached to containment penetrations were supplied to BNL in September 1986.
RAI 15
In PLG-0465, page 2-10, Figure 2-3, the conditional frequency of exceeding whole body dose vs distance apprears to be driven by the S2 source term.
If this is the case, please describe all accident sequences (internal and external events) that contribute to the frequency of the S2 source term given in Table 4-2, pg. 4-7.
In particular, define how the timing and size of containment leakage was determined for each of these classes of accident sequences. Justify the appropriateness of the bounding of each of the accidents into this particular source term.
RESPONSE 15 Our response to the overall question of how the contributors have changed f rom the SSPSA to recent updates, is comprised of the response to question 23 along with the following additional information:
ASSIGNMENT OF SEQUENCES TO RELEASE CATEGORY S2 As discussed in Section 11.6.3.(starting,,on page_,11.6-5) of the SSPSA (PLG-0300), the release categories $2, S2 and S2V were originally defined in the SSPSA to cover a class of accidents not modeled in the Reactor Saf ety Study (WASil-1400). Of these, the category S2V was found to be a significant risk contributor and was found to dominate the 200 rem dose vn distance curve la the EFZ nenaltivity study (PLG-0465). Note that fut simplicity, the release category notatalon was simplified in the sensitivity study. Hence S2 on the sensitivity study is actually the same as 52V in the SSPSA.
The S2V release category was first defined to bound the releases that could occur as a result of small penetration failure during hydrogen burn pressure spikes that would occur shortly after the time of reactor vessel melt through (see discussion on page 11.6-12 of PLG-0300). Then to avoid an excessive number of release categories (to 14), additional sequences were conservatively assigned to S2V (now S2) as well as including those in the 3FP and 7FP plant damage states (as well as some steam generator tube rupture sequences).
As noted in the response to question 23, it turned out that the 3FP and 7FP sequences fully dominated 5E.
The 3FP and 7FP sequences are dominated by station blackout, RCP seal LOCA sequences in which the release path is the 3 inch seal return line with failed open motor operated valves. There are many dif f erent sequences in but they all have the same release path. As noted above, the dominant sequences are initiated by seismic events partly because no credit was taken for operator recovery (closure of the outboard MOV manually) af ter seismic events.
As discussed in Section 11.6.6.4 the assignment of seal return path sequences to ETf is conservative becuase the actual leak rates would be much less than calculated for this category.
RAI 16
Provide justification for the liner yield stress increase from the specified yield stress of 32 kai to a mean yield stress of 45.4 ksi.
RESPONSE 16 Attached are liner plate certified materials test reports (CMTR) representing ten distinct " heats" or " charges" of liner plate material. These samples were selected at random. The average yield stress in this ten charge samples exceeds the mean yield stress of 45.4 ksi which was utilized in the analysis.
I n..ew. ~
KA&KI STUL CORPMATM MII@
CONTRACT W).: siw-snoet INSPECTION CERTIFICATE
.-u.
- T -.2.
MlZUSHIMA KAWASAKI DORI
- " " ^ * " ' * " ^ " ^ "
MANCH 25.1975 I.H. I. YOkott AM A ND.3 idORKS. _
DATE OF ISSUE *
- NAME OF CLIENT ASMF 5A316Gl4.60HT PITISntAGI-Iu M01kES STEEL CO.MATERIAL : -
DATE OF TEST
_ MARCII 25. 1975 PloACT l
sEM*oOK IDIT '. & 2
- 2 -
5501-92 Cil Aft PY IMPACT TEST 3 2E5cocot TENstLE TESTtCL= 8" wit REIXE HAltINESS OF TEST BEND Noich !
munV 8
s ac.No.
REhtARK' CHARGE NO. ROLL NO.
DIMENSIONS WEIGHT _
( KSI
(%)
t i
180*
Ave.
Min.
'(. P. I T. S.
El.
din Max.
R-At t (
)
Are.
lem 60.0 Min o.5 1-s !
3 (mm)
(ha) 32 0
-73.0 21 coop 2.59 2 46.2 64.6 31 9 60 x 3 070 m 11,205 2-st to 2 46A
( Sa-C2-9-1)
, s v u.
HEAT TREATMENT :
c CHEMICAL ANALYSIS (96)
C%
C Si '
Mn P l S
Cu Ni Cr V
= 100 = nos x loo xwee x 1000 x too x seo x ses x mes x sees florentimed.. 3653 F A NB miseute.
CWUtCE NO.
Max.
13 85 Max-Max. Max 21
-10
-130 15 40 to 34 111 14 7
NN c.w.
2-stto i
[l
)
- so o Q p wa hereby c< rtify_Ihe,the masterial desseted herein has i
((
~~"~%
. _. L
A mA sua coemman ummeu &
i CONTRACT 100. *st9Fetst MlZUS!!NA KAW ASAKI DORI Kt'RASillKI JAPAN i
MYE OF (.I LENT I.11.1. YO:MIO(A N00 WitKS.
h.oJu.r : PIrr., salt-lir.:. inI! ES STII.L Co.
MATERIAL. :
Asm SA516 cit.fafT DATE OF ISSUE: sama 25.1W5 L. A. '...i 0;.11 1A3 DATE OF TEST : macel 25,1975 su m
.:3. : S r. 1- :t-u u..:.o.: 3n;.avut TENsii_E TESPG1.gn am ISUC. HARDNESS Cif Alti'Y iMI'ACT TEST Y. P.
l T. S.
E s.
OF TEST BEND Noich
- mmV 180*
Ave.
. Min.
REMARKS PEA C!!AFCE NO. ItOI.I. No.
DIMENSIONS WEIGilT
' KS!
t%)
nin 60.0 Min AIin Max.
R-At t a i<
(mm) ikg) 3:1.0
-72.0 21 05i I
2 3
A I
- i. (. -
t 2-17 a r, ta249A 9.00 x 3,070 x 11,::05 2,592 46.9 65 7 31 coop (Sn c::-9-1) i 1
i i
p;MT R0fl0NED i
A,m.
4
"~ela. scJ l
i l
CHEMICAL ANALYSIS (96)
HEAT TREAW 9iT:
C.
C Si M.i P
S Cu Ni Cr W
V
=1 CliARGE No.
13' tm
= im
= lam l' a '
. im x ino ire
= ime
= nose Plates i
j Mat 15 85 Mas M n. Mas Ilovenallood......1652*F ser N met l
._AL.. -20_.
1~1 M
L1._
l
- .171t'.
10 22 tor 12 8
3,,
i L
l2.
A10 %
j J
'"' #(
We hereby cer:6lv that alie amaterial desseted berets has 1
[
[h au ordance wieb abe raise of /s k
~L
/.
been uns&- in M
i l
l l
l
(
a
'-w n' 'y"; 64) 9
- g 8~
1
.; M:
,'t f.
f INSPECT 105sdERTIFICATE l 2.7 r.ii.s. Ta utAMA 10 3 M Mks.
m g mm.
MA 7. RIAL :
AsMS sags 6cm.WT DATE OF IS5tE:
penussaast y, 3975
.wJu: r
- a 11 r., t u.it ns tolkes stusa, cu.
w.a:...... u.a r a a, 2 a :.o.
S sn1-1 ':
DATE OF TEST : __ passasant 7.1975
.a 2.:n.
382
- >>s tassaut TESTMEs= 5" em RENC. itARRESS CHARPY BAPACT TEST E s.
OF TEST BEND Nose :
susW j
KL4 T. s.
- v. P.
i i
180' Ave.
- Min, ggt 4 ARKS M
i mi CHARCE NO. rot.L No.
DLMENSIONS WElGilt t
Mm 60.0 Min din Max.
R-At t i (mm)
( b.-)
32.0
-72.0 21 0.5 i 1
1 3
a
- - BN Enon s 9.60 m 3.op 3 11.2D5 2.592 W.6 69.1 n
G008 I
(sm-cs.9-1) i u
9.60 x 3.070 x 11.2DS 2.59 2 Om-C2 9-1)
VMg
$s.
N p,
l CHl:MICAL ANALYSIS (95)
HEN' s M Tsucs:: :
i C
Si P
S Cu Ni Cr M.
V Coq OIAltCE NO.
=,im = He
= No x sun = lyy l'a
= les
= He
- Ime x nues x go Plate :
h 15 85 M.n.
- .;.x. mx.
-135 35 O
Mermettaea..1652*F hr 30 stoutes.
12 23 to9 15 7
l : y>
~ ~ '
-..Ja
u
~-
~..
'nLE no. : medMAs(A) g KAWNK1 STEEL CORPORATION MIZUSHIMA n
..vi wnn xuusAxi oom W
CONTRACT NO.: IIN INSPECTION CERTIFICATE
==
Sumu.a I
. 1. '..:_:.:1.':A E9. 3 L
.,3. 3,., i n.,
N 27, 19
-.i :: 1. 3 L.
1, Co.
Aggg 34,,g ggg,t,,g gy DATE OF ISSUE:
g IIIMI8I 27e IV DATE OF TEST :
1
.s..
1i.Nsit r.
n.sr.ci.- O mot REIXl':. HARDNESS C11 AItl'Y IMl' ACT TEFT Y. 1*. j T.s.
- .I if TEST I4END Noich :
wwnV
!!W
% e.
Miru REMAHh kI
^"'A i.
,. it.il l.
ItlMI.NSIONS i u t.it;!!T
'I h
(48.0 Min Min Max.
R-7:
E 4
__ _ _. _ _ _ f m er..
_ _. } q' _Jy.J_ y7*'.0
- '1 05_i i
2 3
A...
' 1. I.... ' 9.o ) x 3,'#73 x 11,215 2,5.e.!
47 6 67 0 09 GND l,
(.
'-C -;-1 )
i l
i i
i t
W gjMT M'N' c
l g
g 4*ITl$-
Dr - T, e i c _
d-~~
7 i'I'!!iilCAL ANALYSIS (%)
HEAT TREATMENT b ~
Flates C
8 C Si Mn l'
- i Cu Ni Cr W
V l
1
\\
- i...
. mi
. i...
. i...i
- a,
< pc en
= tiin tun 1:i n attsw..1652*r t.r :<. =t.nou
, e. t..,..
.G.
lt if,J ~.0 s. l-
.s.
_ {
1 s I ; - l* ' 'a '
l l
,.w. m k _ _ / 11 L I_T. p% -
yrr4 11 A
113 15 8
0 l
i O
W.-
!. e cl.s ce r tily alta t slie unsteri.nl slescrM hercise lier accotilarece wills llie tales el IIte tNd q,m,,
' f
/*
l. se ris.> 1 - see
,,,,7
' e,2./ i' L --
ACO SA31W.63.
c as, u,,,cr. +
L-
'~
~=m.w
.4
@M XAV4'KI S'lHL CORPORATION MlZUS@S N NO.;2 p $ _
MtzUSil!MA KAWASAKI DORI INSPECTION CERTIFICATE sumia
- 1-l
" "^s""" '^"^"
1 SAME OF CLIENT
- 1.H.I. YOtDLAMA NO.3 WORhS.
PicJEET PITTsntfir. L-DES MOIPES STEli2. CO.
MATERIAL
- ASM _SA51(fin & trr DATE OF ISSUE: Me:cir -;.197; i
SEUt:OOK UNIT 1 & 2 DATE OF TEST - niur.it r,.
tryn i
spec.HO.s 33 3*/6001 D"
M' REDUCHARDNF.SS Cet Alt PY IMPAC r TES r l
TENSILE TESPGL-Y.P.
Ei.
OF TEST hEND Notch :
sunV
)
g u}
T. s.
i 180-Ave.
Min.
REMAM y
l CHARGE NO. ROLL NO.
DIMENSIONS WEIGHT is) 60.0 Min din
,M..x.
R-At t (
i M'n (mm)
(kg) 32 0
-72.0 21 05i i
z 3
l4.
I 1-3551 2DzptA 9.D E 3s @ 3 II'
( 5B-C2-9-1)
%,:,, n.s a rn g.xx
=-
i HEN' TREATMENT :
CIIEMICAL ANALYSIS (%)
Cg j
C Si Mn P
S Cu Ni Cr u.
V Pl*** 8
" 8" "*
" 8"
- 8" CHARGE NO.
m,
gge.a.. 3652 F for 3D solemat e.
Max.
15 85 Max.
Mas Max.
21
-30
-130 35 40
(.P'.. i @
s1 33 106 14 7
1-3551
~
g
. w,.
t
(
A c p)ay.
i x / G <.
We hereby certify that.be susterist described herein has
[
r see,i -.ae in s c.,4
.n6 6e n.
.me w
y
_,p..,g a m mesom.m.-
aan an~
2
-n e
j KAVS4K! STEEL CORPORATION a
g.,_
gw_
MIZUSHNA KAWASAKI DORI INSPECTION CERTIFICATE
" " ^ * " ' * " ^ " ^ "
sueru.
p l
I NAME OF CLIENT I.1:.I. Yo r AMA No.3 Wms.
M SA51% M DATE OF ISSUE:
EM:H 15.1975 nam;T : PITTT2rnis. des MoIMas STEEL CD.
MATERIAL :
sr.AP ocK WIT 1 & 2 wrr ro.: 5,m1.yt 3
- .
- .: 1. 3
. wut TMSILE TESTICL-8" seu REDlt RMtINESS CitAHPY IMPACT TEST l
T. S.
E l.
OF TEST BRED Notch ;
aunV Y. P.
it
(%)
AREA tuo-Ave.
Mk.
REMARK g K 3
i CHARGE NO. ROLL NO.
DIMENSIONS WEIGHT Ma 60.0 Min din Max.
R-At t (
l (mm)
(kg) 33.0 73.0 21 -
05 i 1
2 3
Ave.
i 1-3 01 spzR8A 9.60 x 3.070 x 11.805 2.592 47 9 66.8 3a ecos 6
(EE-C2-9-1)
I l
l sess=""'
Austen11e Grain 51re.....Po.8(Mir.fle.5)
MN j
Ultranos ic Test..........Jcooptaq1e
.( )
O Dimnniena Inspeetter....Jcomptahle
' _, * * <1e
^
Starfaco Inspection.......-
HEA" TR" M I CIIEMICAL ANALYSIS (96) f C
Si Mn P
S Ca Ni Cr Me V list.AL C.%
= 100 = nos x nos
= tono ioco x we w ies x ase xames x iesomennn Pieces CHARGE NO.
Name h
moessatsee.....165W ser se etnut M"
15 85 h
M" h
40 en op.sen' w
no i
SS 30 1-3601 30 83 111 18 8
( )
(
f M/. 0(.
We hereby certify stess ef:e sensorial doosr8ed lierein lies
- !w-t been made in accordun skh one rules
_ 4 /. A f 2 j_-
g
~y me.
3.; f+* &. -
a
~~ 7.
.so. : Jners as,o/6e Q
d KAWSKI STEEL C00.T&T10N MtZUSHNA @
Co m ACT NO.:
2m-:24 i
M17t*SHN A KAWASAKI DORI Kt RAS 5tlKI JAPAN I.it.I. YONDilAMA NO.3 MNtKs.
NAME OF C1. LENT DATE OF ISSUE:
W CH 23, 1973 l
fHOJECT a PITTSI1Uf0ll-DES MDIE STEEL CO.
MATERIAL ASDS SA516GR.(ort SEC1HouK UNIT 1 & 2 DATE OF TEST :
WCII 25. t'E5_.
"' A W 8 5901-382 S1'LE. !.O. : 38%$00001 TENSli.E TEWGl.- 8" *** REDUC. HARDNESS Cil AHl'Y IMI' AC T TEST V. P.
l T. S.
E l.
OF TEST BEND Notch :
menV I
CHARGE NO. ROLL NO.
DIMENSIONS WEIGitT ns!
m AREA i
180*
Ave.
Min.
REMARKS t
h 60.0 Min din Max.
R-(mm)
(lyr) 32.0
~7 2.0 21 03i -At t i i
1 2
3 A.e.
1-3706 2F985A 9.60 x 3,070 m 11,205 2.59 2
%.9 67 1 31 Goop (SB cs-9-1) a 9 60 x 3,070 x 11,205
. 2.59 2 hMO
( SB-C2-9-1) f g
t d
bY I
f L.as=="v u
l HEAT TREA1NENT :
CilEMICAl. AN ALYSIS (96) l C
Si Mn P
S Cu Ni Cr Me V
Ceq
= tm = tm
= im
= ine
= it. c = iuo
= ion x ino x scoo
= anno x 100 l
CHARGE NO.
Max.
15 85 Max.
Max.
Max.
mermetised..1652*F for 20 mimates, 21
-30
-130 35 40 s-3706 11 23 108 15 8
&WL a. \\.,,
f t w/. Q c.
-ri.-. - -.
C, We herchy certif3 that al e amaterial described herem less
[
kmp been made in accordance with the rules of the m m Nc.
-y, j g.
season.a ma#
l
.-,__.,x. e.
- a. - u.s v
w m....,
-.., _.,.. _. m.,.. _,.,_
%' g)KAWGAKI STEEL 00RPGtAT10N 9
CONTRACT No.: m p-essep MizusinMA xxwastia ooRi INSPECTION CERTIFICATE
' " " ^ " " ' * ' '^"^"
SureuEn 1--
1.94.1. YOEQtAMA 10.3 M2KS.
.MME OF Cl. LENT ASHE SA516 Cit.60MT DATE OF ISSUE:
WCI 25,1975 PIORI T i PI TTSM;I(.11-DES MDINES STEEL CD.
- gg7ggggg, SEAlbOOK Ul*1T 1 & 2 P E H 25, 575 m x t.o.
- SSrn-302 DATE OF TEST :.
dit.c. m.
i 38:E5tVXX)1 T1.NSILE TESTtGL 8"
- REDUC. HARIWiESS Cll AMPY IMPACT TEST l-Y.P.
T. S.
E l.
OF TEST BEND Notch !
munV CHARGE NO. ROLL NO.
DIMENSIONS WEIGHT t
El i
(%)
AREA g
i 11 0
- ave.
Min.
REMARKS W.
60.0 Min dik Max.
R-At t I (mm)
(bg) 32.0
-72.0 21 05i 1
2 3
4.
3-5372 2EA05A 9.60 x 3,070 x 11,205 2.59 2 46.5 65 3 32 c000
( sa.ca-9-1) g,tu lieggD:
g 5L A'If 6
~~
CHEMICAL ANALYSIS (96)
HEAT TREA'D4DIT :
C C
Si Mn P
S Cu Ni Cr W
V v%
= im x im
= im v iom w ino im x im x too x ioco x icoo gg,,
CHARCE NO.
Max.
15 55 Mu.
Max.
Max.
llormatteed..1652 F for 2 edsouten.
21
-30
-120 35 40 11 23 110 11 6
3-5372
[
b dL
-_/ N'l 0 f
[ ( -- Tkmy _
We hereby certify that time amaterial described hereisi lies n[
been made in accordance with ties rules of the -
gyg ].
ases anstees de.
'I I
uwto sua camum mzussutxs Cop N NO. 2 7k m 89 M12USHIMA KAWAS AKI DORI lNSPECTION CERTIFICATE i
Summ xua^s"'*' '^"^"
l NAME OF CLIENT I.H.I. YOKOHAMA NO.3 WCIWS.
(
MxuscT : PIrrsnutr.li Das MDINES STEEL CO. MATERIAL :
A! DER SA516GR. W ff DATE OF ISStI: Mancit 25, 1975__
i sEAltit00K UNIT 1 & 3 DATE OF TEST : MAntti 25, in75 i.
l' wonK No.: 55nt-38a snr.NO. 38aE500001 TENSILE TEST 1CL-Ba umI REIUC HARDNESS Cil Alt PY IMPACT TEST Y. P.
l T. s.
E l.
OF TEST BEND Nonch !
sunV KSI
(%)
AREA I
i 180*
Ave.
MA REMARK:
CHARCE NO. ROLL NO.
DIMENSIONS WEIGIIT h
60.0 Min kin Max.
R-At t i i
imm)
(kg) 32.0
-72.0 21 0.5 i i 1 2
3 Ar..
GOOD 3 6323 2C358A 9.60 x 3,070 x 11,205 2,592 48.5 E.4 32 (55-C2-9-1) i 01 TNT E m
- a 9.60 x 3,070 x 11,m3 2,592 Auntenit., Grain Eire.
No.8(Mi n.No.5) d.
.Acceptnbin (35 C3-9-1)
Ultrawwic Test..........
.Acceptnble Dinonnicns Innpe: tion....
.Acceptnble Surfaco Inspecti me.......
HEAT TREA1NENT :
~
CIIEMICAL ANALYSIS (96)
Im n im x im sam. iism
. Im w we x ses x teso x t:m xgh C.%
C Si Mn P
S Cu Ni Cr Me V Het A L Neeen1t h..... 1 6 2% he m ade="
Mar.
Max.
CHARCE NO.
Max-15 85 M ax.
Max.
Max.
40 34 21
-1n
-1m M
to 8
31 3 6323 11 24 110 17 8
i fM
-~~ --
We hereby certify that the amaterial described I.crein has p/ Ge e
~
cc i
been made in accordance wieli the rules of the m
(
, g4
[g.
an s assesom a so asub aar seerserna o
1 e
e.
t e
av U
.EAWAkj lg
,, ;y.
'e g*y M
^_ - 1 INSPdbI,,ypz..' M 'C k T E
- ~
~
~
=
s 3 ( er :'/
.*,e#
.MIE OF C1.!Da :
1.u.1. w e< tom les.3 w as.
- ,, f
..,.w I e n. e....=:
a..; wn.us s=.a. co.
gg: Rid :
a s s m 1Cck.6aser MTE W m *.
- ---.uar,, 39 t1 1aa m
- '-*C 1-3 '.i
@Kl.
MTE W; TEST :. _ asmag so.19,
......:.u.
3 m
' N N F m)
MRPT _ N
.mu,.
n i
Y. P.
[.4,,T.' S.
EL M
BES T Moseh
- e-
- mu#
~-
cP.AaCE m rot.t. MO.
DIMENSIONS WEIGHT
'( 1*E % i '
~WI %
('"
t ISP. ;
' Ave ' ' '
3C ' ' '
,"' R E 34 A R K Idra 1 der.e -
Min k Mast R-
- At ti
<1 (mm)
(hgi 1:*,0 W,.O c1 0.5 i
t.
3 aus.
.y;.r
.t
' o'4 17xy -
9.60 x 3,nye x 11,25 2,593 46.3 h 33 0003 r
(:.s-cw1)
'W
.1 %
, f a
%.u..avW a
9.so = 2.oro = 1s.m$
3.593 i
~-
3 (sa.c3 9 1) ha L," '
A=ma -at n egggeq g an, o
,.... sin 80mm,3Q.
l tfttsu'esen thYkst.
.....":.-/ M y.
i..,
H W N Eengpe< tieso
)
o
.........Ao>mtehh
+
Surface loupestion...
.......... M
^7
' p.
m
.n CifFMICAL ANALYSIS _(W)
HEAT TREANIT':
C Si Mn P
S Cu Ni Ce no.
V so 3..L3 Cee yajtCE V).
Ma = noe w lue
= sue to a nee v ise w I.e MFW v hMW
_Re g,g g g w
y SLs 15 "1
EL M a c. Mas
- g
- rad,
"***"I I " " * " *
- M " N ' '
m i _.-A. -l'n.. _3 3_.
?
8 n
1.)
101 13 n
e9 j -
i
}
}
.__t I
We
- h..1 3 cer tify that tlw material described herein has i
. [L. '
/
been ma!- in accordance with the rules of the svaueneir g,
/, ([),
~
f
_. r n t' n tw*-wm I citfC sh, ~ t +
6h-+
~
$;.l A '
% i.
x, J{3 t
1 w
_A..
-m N KAWASAKI D001 wW INSPECTI CATE
=S""""^*
=
- ===
d,4;
+-
. y:: s
!nE OF llENT : _d' I *
- 3 I
,I i tJJD3 4
1.i
.3 SEZL CU.
g,aA516 et.de s g ygg.
m g 3py5 M k 3fF5 Lu.A G.
t E31-M
'[g.
ggg.
i d.*J. 60.
a % 1.5G0JJ1 m m w-mmgWess CHARPY BIPid;7 M V
- v. P _.1 xT. s.
El.
W' TEST BEIS Neesh
- su#
C3 MAGE 100. rot.1. No.
DIMENSIONS W ElGIIT t mm de:
mi g i
's 180' Am-h ItEMARK:
ima
. 48.8. i Min Min Mas.
R-At i I 8
(m.)
(4 38.0 21 0.5
- I 2 I 3
A'=-
2-oo)7 HMIA 9.6e x 3,07e x 11.305 2.592 g.s l66.0 25 GOGB
("N I) s.
B 9.60 x 3,070 m 11.805 2.592
"/.
(hD-C2-9-1)
&I ~ g' n
w'e 2,
- i..,
^ ^ '
- e M Base...3A8(Mia.3do.5)
-?;-
Ultenen etejaet Leeeptab,le
... Is.
me. _n...Leeeptat le M-iF-litu.suaNl a.t Ane---E m e.
---u n.
CffF.MICAL ANALYS18 (5)
^
~
HEAT 'IltEATuse.sii; C
S4 Mn P
F.
Cu Ni Cr V M1.AL Ceq Plate e un
. un nas v lam
. som xigx3j um cysgg isq us.
. ise po
= trwm
- i ni tiennelised....1@ Sur se amin itas If 85 Mas st.,s stas
- lia, tr x,
- 1
-w
-1"i
.a_
in g_
o 2-9 e7 11 23 110 19 7
26 31 j
i We h s ct,v ocreify thst the material described licrein has
/[f),
(.{
7
,, f' f l>cen ma. i.- in acco rt'ance with the rules of the seseeentsoe ru: s'rtle: m oc,gn..
/otn'- 80 cT""
I
RAI 17
Indicate the correlation between containment failure sequences and the containment failure modes.
RESPONSE 17 terms are used to describe the f ailure behaviour of the Seabrook Three containment. These are:
Containment Failure Mode: The term containment failure mode o
is used to describe the structural member of the containment system which f ails as a result of an' overpressure condition in the containment. For example wall hoop failure or feedwater failure modes.
penetration flued head failure are containment Table 11.3-1 of the SSPSA (PLG-030f>) lists the different containment failure modes.
Containment Failure Type:.The term containment failure type o
is used to discretize the range of leak areas which are predicted for the different containment failures modes. Three containment f ailure types have been defined, namely Types A, B and C as defined in Section 11.3.1 of the SSPSA.
o Containment Failure Sequence: The term containment failure sequence is used to describe the containment f ailure associated with each For example release category S3 is associated release category.
In the SSPSA with late containment overpressure f ailure sequences.
two additional distinctions were made to characterize a release category, namely whether the containment spray system worked and whether the containment floor and reactor cavity were dry (vapori-zation release) or wet.
The release categories are described in Section 11.6.4 and in Tables 11.6-1 and 2.
Since all risk dominant accident sequences involved dry containment conditions without spray, the latter two distrinctions are dropped in the RMEPS study and in the WASH-1400 sensitivity study.
All release categories were than based on dry containment conditions without spray and the containment failure sequence was the only remaining distinction for the-release categories.
The correlation between containment f ailure modes and containment One containment failure type is given in Table 11.3-1 of the SSPSA.
failure mode. The correlation failure type is assigned to each containment between the containment failure sequences (or release categories in the failure mode is RMEPS and WASH-1400 Sensitivity Study) and the containment described below.
failure Release category S1 model an early gross containment o
sequence.
It is equivalent to PWR-1 in WASH-1400.
It is always treated as a gross (Type C) containment failure.
Release category S2 models containment failure sequences with an o
early increased leak rate.
It models a Type A containment
failure occuring at the time of vesselbreach.
From a containment failure pressure point of view, the S2 release. category is more conservative continues until a late overpressure failure mode occurs as in release category S3.
Release category S3 represents a late overpressure failure sequence.
o
-It is modeled as a linearly increasing leak rate beginning at the time when the first type A failure mode is predicted to occur and building up to a full type A failure mode (6 square inches) the time when a type B or C f ailure mode is predicted. This atis followed by a type B or C failure type at that time.
- However, because all S3 accident sequences had calculated containment f ailure times longer than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> they were always treated as Type C failures. The lowest containment f ailure time estimated in the uncertainty analysis was 22 hours2.546296e-4 days <br />0.00611 hours <br />3.637566e-5 weeks <br />8.371e-6 months <br /> and the type C f ailure assumption was retained even for this case.
Release category S4 represnets basemat melt-through f ailure o
Because of the uncertainties in the release sequences.
f ractions for basemat melt-through at a rock foundation site, all basemat melt-through failures were conservatively modeled as an S3 release category.
Release category S5 represents ' accident sequences where the containment o
remains intact.
A continuous release at a rate of 0.1 v/o per day was conservatively simulated as an equivalent instantaneous release at the time of vessel breach.
Release category S6 represents accident sequences where the o
containment is ot isolated from the beginning.
It is modeled as the largest containment penetration.which is allowed by technical specification to be temporarily open during normal operation and which communicates directly between the containment atmosphere and the environment.
This is the 8 inch diameter l
online purge penetration with a flow area of 50 square inches.
Thus in size it corresponds to a pre-existing Type B containment
~
failure.
Release category S7 represents containment bypass accident seq ue nces.
o It is modled as an RHR pump seal f ailure with a combined leak area for both RHR pumps of 2.6 square inches. The frequency of the traditional V-sequence f ailure mode (RHR pipe rupture) is included in release category St.
In summary, it is noted that the analysis did not depend on the distinction between a type B or a type C overpressure containment failure, except that the inclusion of type B failures shifted the containment failure probability l
distribution to lower pressures compared to the case where the type B f ail 1res are not considered.
I
RAI 18
Provide the basis for concluding that the sight glasses in the hatches will not fail under high containment temperature and pressure conditions.
ANSWER 18 The sight glass-in the personnel hatch was. tested by its. supplier, Owen Corning Co., under the following conditions:
150 psig Pressure
=
5500F Temperature
=
In addition the pressure was cycled.from 0 psig to 150 psig ten times i
at a constant' temperature of 5500F.
The Owens' Corning data sheet is attached.
We are currently persuing discussion with Corning Glass to determine if any testing has been done above these valves.
i i
O e
4 f
,.. -.,.,...,,;.___...___,..__,_,__-,....,... _ _. -...,, _. _ _.. _ _.. -. _... _ _ - _ _ _. _ - _ _. ~. _ _, _.., _ _,, _ _ _.
RAI 19
l Document the effect that the recent update in seismic f ragilities will have on the conclusions of the PSA results.
RESPONSE 19 Seismic sequences dominate release categories S2 and S6 in the Risk Management and Emergency Planning Study (RMEPS). The response to question #23 discusses the principal contribution to early health risk and explains how these release categories contribute to early health risk.
The following explains how the seismic f ragility update is expected to change the frequency of S2 and S6; it is expected that the effect on the frequency of all other release categories will be insignificant.
A complete requantification will be included in the probabilistic ' safety assesment (PSA) update now in progress and' planned for completion in 1987.
In the complete seismic risk analysis, a point estimate analysis is first performed using the plant event trees that are quantified for several discrete values of ground acceleration.
From the point estimate results, dominant sequences initiated by seismic events are identified; then, these sequences are reanalyzed using a computer code called SEIS4. This code is described in the SSPSA Section 4 and 9.
In SEIS4, the seismici ty curves and fragility curves are appropriately combined and uncertainties in these curves are propagated to obtain uncertainty distributions on the final. result, which is either a core melt or plant damage state frequ'ency contribution.
In the following approximate analysis, the point estimate step is bypassed, so some assumptions are made about dominant sequences.
Hence, these results are only rough approximations and should only be used for order-of-magnitude estimates.
Again, a complete reanalysis of seismic events is currently in progress and is l
planned for completion in 1987.
1.1 RELEASE CATEGORY S2 This release category is dominated by earthquake and transient initiating events. These sequences can_be simply represented as OG(DT + DG + SSPS)
(1) where OG
= Offsite Power Fragility DT
= Diesel Generator Day Tank Fragility DG
= Diesel Generator Fragility SSPS = Solid State Protection System (SSPS) Fragility (actually 120V AC power panel required for SSPS success) and only seismic unavailabilities are included.
Also, earthquake and large loss of coolant accident (LOCA) initiating events provide a small contribution and can be represented as LL*0G*(DT + DG + SSPS)
(2) where LL = Large LOCA Fragility
.. ~. - -.
- - ~
Equation (1) was quantified with the SEIS4 computer code and resulted
.in the following annual cqre melt frequency:
Mean
= 2.84 x.10-5~
Variance
= 2.24 x 10-9 Based on the fragility update, SSPS and DT can be dropped f rom the model, based on significantly higher capacities.
However, a relay chatter fragility at a relatively lower capacity has been identified in the 4,160V switchgear. This chatter could have a negative ef fect; e.g.,
trip out the diesels.
Until the consequences of this chatter are evaluated, it is assumed that the chatter fails both diesels.
Therefore, Equation (1) can.be changed as follows:
OG*(chatter + DG)
(3) where Chatter = Relay Chatter Fragility (4,160V switchgear)
Quantifying equations (3) for annual core melt frequency with SEIS4 results in Mean
= 1.8 x 10-5 Variance
= 9.58 x 10-10 Comparing the quantification of Equations (1) and (3) shows.a slight reduction (less than a f actor of 2) in f requency. However, this assumes the chatter fails the diesels without recovery.
An ongoing relay chatter review will determine whether this particular-chatter is a real concern.
In addition, this review will determine whether there are any other relay chatters that should be considered in the model.
1.2 RELEASE CATEGORY S6 This. release category is dominated by earthquake and transient initiating events.
These sequences can be simply represented as NOG *SSPS (4) where NOG = Of f site Power Available (negation of OG - fragility)
As described above under release category S2, the solid state protection i
system can be dropped from the model.
Therefore, the simple model in Equation (4) would go to zero. To actually determine the new S6 f requency, the whole plant model needs to be'requantified and unraveled to obtain' new dominant sequences and frequene.les.
- However, the trend is a reduced frequency unless the ongoing relay chatter review identifies new sequences.
i
RAI 20
Assess the impact on risk of assuming that the containment capability corresponds to the pressure which produces 1% strain in the containment wall.
RESPONSE 20 A sensitivity analysis on the' risk results in the SSPSA, the RMEPS and the WASH-1400 Sensitivity Study was performed to determine whether any of the basic conclusions with respect to risk or emergency planning requirements would change substantially if the containment was postulated to fail at a deformation strain of 1 percent. It is emphasized that there is no basis in experimental data or analysis to support a 1 percent strain failure con-dition as being a reasonable definition of failure or that containment failure at this condition is any more likely than what is implied by Figure 11.3-14 in the SSPSA or Figure 4-7 in the RMEPS.
Since the' failure definition is arbitrary, there is no meaning in assessing
~
uncertainty in this failure condition and it has been postulated to occur with certainty when the containment pressure reaches that pressure value where the calculated deformation strain reaches 1 percent according to Figure 4-2 in Appendix H.1 of the SSPSA. This corresponds to a maximum radial displacement of the containment wall of 8.4 inches and it occurs at a pressure of 175 psig or 190 psia. This corresponds to the low temperature wet containment condition in Figure 11.3-14 of the SSPSA or Figure 4-7 in the RMEPS. Since in the analysis no distinction is made between type B and type C containment failures (see answer to question 17), the composite wet containment probability distribution (solid curve) is stepped from the calculated value of 0.2 at 190 psia to 1.0 (guaranteed failure). Note this implies that the containment failure analysis in the SSPSA predicted a 20 percent probability for_ containment failure at the pressure corresponding to the 1 percent deformation strain. For dry containment conditions (dotted curve), the elevated temperature conditions reduce the strength of the innermost steel layers and the 1 percent deformation strains are~ reached at a pressure of 158 psig or 173 psia. An containment failures which in the original containment failure analysis were predicted at pressures below these 1 percent strain pressures'were retained without modifications. The modified containment failure distributions of the 1 percent strain failure sensitivity analysis are shown in the attached Figure 20-1.
Next, the. impact of this change on the calculated risk and on the conclu-sions with respect to emergency planning were. examined. An impact would result in one of two ways.
Containment event tree split fractions would change where these were determined from the containment failure pressure probability distribution.
Secondly, the timing and release magnitude of late overpressure failure release categories (S2 and S3) would change due to earlier containment failure times. However, changes would only result in cases which depended on containment failure pressures above 173 psia for dry cases.
I
The impact on release categories was examined first. New probability distributions for the time of containment failure were generated for dry containment sequences. These are shown in Figure 20-2.
These curves do not exhibit a steep change because the uncertainty in the containment pressure versus time relationship calculated by the MARCH and COC0 CLASS 9 codes is still a valid consideration. The net effect of these changes is to reduce the time of release for release category S3B in the RMEPS from 89 hours0.00103 days <br />0.0247 hours <br />1.471561e-4 weeks <br />3.38645e-5 months <br /> to 70 hours8.101852e-4 days <br />0.0194 hours <br />1.157407e-4 weeks <br />2.6635e-5 months <br /> and to reduce the release duration of release category S2B-3 from 56 hours6.481481e-4 days <br />0.0156 hours <br />9.259259e-5 weeks <br />2.1308e-5 months <br /> to 37 hours4.282407e-4 days <br />0.0103 hours <br />6.117725e-5 weeks <br />1.40785e-5 months <br />. No changes result in either the conservative source terms in RMEPS or in the source terms for the WASH-1400 Sensitivity Study, because for all these source terms, the release timing was assessed as occurring before the time when the pressure in the containment reaches the 1 percent strain level. The release fraction factor for particulate radionuclides is shown in Figure 11.6-3 of the SSPSA. It is shown that no significant change occurs between 70 and 90 hours0.00104 days <br />0.025 hours <br />1.488095e-4 weeks <br />3.4245e-5 months <br />. Therefore, there would be no change in the release fractions as a result of the above noted change in release timing. Furthermore, in both cases (S3B and S2B-3), the warning
~
times are still much longer than required for complete evacuation.
It is thus concluded that no change whatsoever can be identified in any of the consequences calculated either in the SSPSA or in RMEPS or in the WASH-1400 sensitivity study as a result of postulating containment failure at a 1 per-cent deformation strain.
Lastly, the impact on containment split fractions-was examined in Section
~
11.7 of the SSPSA. In no case is a split fraction dependent on a contain-
~
ment failure pressure in excess of the 1 percent strain value for dry con-dition (173 psia). The split fractions for two top events in the containment event tree are affected. The split fraction for top event 10B on Table 11.7-8 in the SSPSA shifts slightly to increase the probability of late overpressure failure and correspondingly decrease the probability of basemat melt through. However, as explained in the answer to question 17, all basemat melt through cases are conservatively treated as late overpressure failures and this change therefore has no impact on the results. Secondly, the split fractions for top event 12A in Table 11.7-8 in the SSPSA would shift significantly to increase the probability of type C containment failure and decrease the probability 'of type B containment failure. However, since in the analysis all late over pressure failures were treated as type C failures, this change also has no effect on any of the results.
Overall, it is concluded that the assumption of containment failure at a pressure corresponding to 1 percent deformation strain has no discernible ef fect on any of the results and conclusions documented in either the SSPSA or the RMEPS or in the WASH-1400 Sensitivity Study. This conclusion can be traced to three distinct reasons:
(1) no early pressure transients reach a magnitude of 190 psia, (2) the reduction in release timing for late overpressure is insignificant with respect to warning times and release fractions, and (3) conservative analysis assumptions in the containment event tree quantification aosorb any effect which would otherwise be visible in the release category frequencies.
l 13,
l-Tenem D
= assumer
- DEV e
,e A
G 9
w' =-
e
~
l
.,,g2 E
2 8
W
=!
~
F 0
8 ' '
l f
t I
10
-140 10 0 tem 3W 220 240 280 Pptmust (Pesal FIGURE 20-1.
COMPOSITE CONTAINMENT FAILURE PROBABILITY DISTRIBUTIONS FOR THE 1% STRAIN CONTAINMENT FATLURE SENSITIVITY CASE
w, Sk 1
se -
AE 75 sa 5.9 0,5 =
l 5
em
.N8 t
y 1
i Il I
I I
I I
m
[
s ao e
as a
1m no no Tuus OF fit 8.EAM M FIGURE.20-2.
CUMULATIVE PROBABILITY DISTRIBUTIONS FOR LATE OVERPRESSURE FAILURE TIME IN DRY SEQUENCES - 1% STRAIN CONTAINMENT FAILURE SENSITIVITY CASE
RAI 21
What is the impact on risk f rom accidents during shutdown and refueling when the containment function may not be available?
RESPONSE 21 The purpose of this technical note is to address the risk from accident seq ue nces that could potentially initiate during plant shutdown at Seabrook Station.
Specifically, this note is intended to answer a question posed by the NRC staff during their review of the Risk Management and Emergency Planning Study (Ref erence 1), the companion sensitivity syudy (Reference 2), and Question Number 21 (Reference 3).
All work performed to date to identify and to assess the risk of potential accidents at Seabrook station has concerned itself primarily with scenarios that could initiate at or near full power operation.
In the original full scope PSA (Reference 4), the coverage of accident sequences in terms of initiating events, the possibilities for system success and failure states, and the treatment of dependent events met or exceeded those of other published PSAs.
This coverage was certainly greater than was possible during the seventies when the Reactor Safety Study was performed.
A judgment normally made in a PSA, and made in the SSPSA, is that the level of risk associated with accidents that could initiate during f ull power operation, however small, rt
is substantially greater than that associated with accidents that occur during shutdown. There are many reasons to support this judgment including the fact that at full power there is a greater level of RCS stored energy, after-heat level-and inventory of radionuclides than the case with plant shutdown. There is also generall.y more time available to recover from adverse situations during shutdown.
Several years after the SSPSA was completed a research project was performed for the Electric Power Research Institute in which the risk of accidents at the Zion nuclear plant during plant shutdown and RHR system operation was assessed (Reference 5). The only risk parameter quantified in this study was core melt frequency. The results in comparison with the results of the Zion plant PSA (Reference 6) for power operation events that had been completed previously by.the same PSA team are as follows.
Core Damage Frequency-Description Mean Median Cold Shutdown (Reference 5) 1.8 x 10-5 2.6 x 10-6 Power Operations (Reference 6)-
6.7 x 10-5 5.0 x 10-6 Hence, the core melt frequency from cold shutdown events at Zion is less likely but more uncertain than that from power operations. The Zion cold shutdown study did not address consequences of these events; it only addressed the frequency of core damage events.
2 1447P102986
The Zion cold shutdown study examined plant shutdown and startup procedures in detail to identify a wide spectrum of potential accident sequences that could originate and develop during plant shutdown.
It also made use'of an in-depth review of in-plant records and information that covered 10 refueling outages, 24 maintenance outages, and some 27,888 hours0.0103 days <br />0.247 hours <br />0.00147 weeks <br />3.37884e-4 months <br /> of RHR system operations. Several person-years of effort went into the Zion investigation.
It is of interest in this note to address the risk frnm plant shutdown events at Seabrook Station,:which like the Zion plant, is a four-loop Westinghouse PWR with a large dry containment.
In the brief time available, it is not possible to complete the kind of in-depth examination that was described in Reference 5.
On the other hand, for the purpose of addressing the implications on emergency planning, it is not sufficient to measure risk simply ~1n terms of core damage frequency.
With this perspective in mind, the objectives of this response are to:
o Provide an order of magnitude estimate of the frequency of core damage events that could initiate at Seabrook Station during plant
- shutdown, o
Estimate the frequency of-the above events that result in containment bypass, containment high leakage, or containment intact end states.
e Account for important specific and unique features of the Seabrook plant hardware and procedures.
3 1447P102986
Provide a suitable allowance' for uncertainties associated with a o
preliminary level of analysis through the appropriate use of conservative assumptions.
Provide for a reasonable level of accountability of operating e
experience with events that have occurred in similar plants during plant shutdown.
APPROACH The approach taken to address shutdown loss of cooling events at Seabrook Station was first to review the Zion study (Reference 5), to compare the design and operational features of Zion and Seabrook, and to identify key differences important to the determination of shutdown cooling risk.
Based on this review and the key differences that were identified, a determination was made of the extent to which all or part of the NSAC-84 results for Zion could be applied to Seabrook.
In cases where Seabrook specific features indicate a reduced level of risk, appropriate corrections were made to the Zion results. Finally, a quantification was made of sequences that could occur at Seabrook Station at a higher frequency than that assessed for Zion.
In summary, the risk of shutdown cooling events at Seabrook Station was evaluated as follows:
Seabrook Risk = Zion Risk per NSAC-84
- Portion of Zion Risk Not Applicable to Seabrook
+ Portion of Seabrook Risk Not Applicable to Zion 4
1447P102986 fL
In other words, there are some design and operational features common to Zion and Seabrook and some' unique to each plant. The enhanced features of Seabrook were accounted for by reducing the risk contribution of selected dominant sequences in the Zion results. This resulted'in a reduction of the core damage frequency evaluated in NSAC-84.
Then,'the enhanced features of Zion were accounted for by adding to those results a separate Seabrook specific analysis of accident sequences that were not important in the Zion results because of its unique enhanced features.
The above process resulted in a balanced and unbiased albeit conservative assessment for Seabrook Station that was especially designed to make maximum and appropriate use of the Zion results for core damage frequency. Then, all the resultant core damage sequences were evaluated to determine the frequeny of three types of core damage release states: core damage with intact containment, sma;; bypass, and large bypass.
Finally bounding estimates were made of the contributions of shutdown. loss of cooling events to the-200-rem dose versus distance curves in References 1 and 2.
COMPARSION OF ZION AND SEABROOK DESIGN FEATURES The ability to respond ti this question quickly f acilitated-by the f act that key ' plant and systems analysts of the Zion PSA team played major roles on the Seabrook PSA team.
The design features of the respective plants were compared from two perspectives.
First, the major differences between the two plants were noted based in our general understanding of the plants, systems, components, and PSA results.
Second, the 34 dominant accident sequences for Zion shutdown cooling
between these systems and the frontline systems, such as the RilR system.
For one thing, Ll4ere are ways to utilize equipment on Unit 2 for Unit I and vice versa at Zion that are not possible at Seabrook. These differences stem from the fact that modern design criteria, to which Seabrook was designed and Zion was not, call for a strict physical separation between redundant trains of safety-related systems and greatly reduce the opportunities for lining up cross-train pump and heat exchanger combination.
In othere words, there are more success paths in the older plants such as Zion.
Ironically, the introduction of these more restrictive design criteria in Seabrook produces a relative advantage for Zion in this regard. Therefore, we would expect to see a higher contribution from sequences involvong. cross-train combinatior.s of electric power, service water, component cooling water, and RllR systems at Seabrook, relative to Zion.
- 3. Other Plant Differences.
The remaining plant differences that were identified could be significant in the determination of the risk of power operation events, but are not found to be significant with respect to shutdown cooling risk. These differences include those in the containment heat removal systems (different configurations of containment spray and f an cooler systems), use of solid state versus relay technology in the safeguards actuation system at Seabrook and the ability to utilize Unit 2 equipment for Unit 1 and vice versa at Zion. There is a high degree of similarity between Zion and Seabrook in the procedures that govern shutdown operation. Of the differences in this area, there are distinct advantages to Seabrook (e.g. some of the local manual valve operations at Zion are performed remote manually from the control room at Seabrook.
UTILIZATION OF NSAC-84 RESULTS FOR SEABROOK Following the design and procedures review and comparison, the dominant sequences from Table 6-1 in NSAC-84 were reviewed for applicability to Seabrook Station. The following conclusions were reached.
Because of the similarity between the plants and the procedures, the e
dominant sequences from Table 6-1 are generally applicable to
- Seabrook, e'
The NSAC-84 sequences would be expected to occur at the same frequency at Seabrook, except for those sequences involving inadvertent closure of RHR suction path MOVs and those involving combinations of support system faults and RHR train failures.
The sequences involving suction path.MOV closures 'would occur at a lower frequency at Seabrook because Seabrook has a separate suction path for each pump. The frequency of valve closures was calculated as part of Top Event RM in NSAC-84. This top event asks whether RHR cooling is maintained during maintenance and refueling outages. The cause table for this event is shown in Table 1 (adapted from Table 5-5 of NSAC-84). Also 8
1447P102986 7
shown in the table is a correction factor that shows the effect of two drop lines at Seabrook in lowering the frequency of " hardware failures" and " human errors." The derivation of the correction factors is explained below.
-For spurious valve closure to cause a loss of RHR cooling at Seabrook Station, it is necessary to postulate either a common cause event
~1nvolving one valve in each suction path, or a coincidence of a single valve closure and maintenance being performed on the other RHR train (these could also be maintenance in a support system of the'other RHR train, but these sequences are aparately accou..ced for below). The correction factor for this cause of RM is given by BMOV + (1-BMOV)'( 5) QRHRM =.072 where
-SMOV = MOV Common Cause Parameter =.043 from 35PSA Section 6 QRHRM = Maintenance Unavailability of a Single' RHR Pump Train During Shutdown
= 6.1 x 10-2 Based on Zion Data in NSAC-84 The factor of.5 is the chance that the maintenance is being done in one of two specific trains.
The correction factor for errors in inverter switching is given by
.5 Q
=.031 RHRM 9
1447P102986
The result of the above corrections to this top event is a reduction in the failure. frequency to a factor of.145.- This factor was applied to applicable sequences in Table 6-1 and the following results were obtained:
Core Melt Frequency Results Mean Median NSAC-84 Results:for Cold 1.8 x'10-5 2.6 x 10-6 Shutdown Results Corrected for two RHR 7.6 x 10-6 1.1 x 10-6*
Suction Paths
- Estimated as source factor reduction as calculated for mean results.
Hence, because of the dominance of the valve closure events in the NSAC-84 results, the effect of having two suction paths is a reduction of core damage frequency' of the NSAC-84 sequences. at Seabrook by a factor of about 2.
ANALYSIS OF SEABROOK SUPPORT STATE SEQUENCES Because of differences in the support system interfaces with the RHR system and because these particular differences are unfavorable for Seabro'ok, separate event tree analyses were performed to cover these events for Seabrook. The following initially events were selected for this analysis.
10 1447P102986
Designator initiatingEvent LOSP Loss of Offsite Power LIRH Loss of One RHR Train L1PC Loss of One PCC ' Train L2PC Loss of Both PCC Trains LISW Loss of One Service Water Train L2SW Loss of Both Service Water Trains As shown in Figure 1, these initiating events were first analyzed in support system event trees whose sequences result in one of five different plant states, The plant states denote the number of RHR trains and safety grade AC power trains rendered unavailable by the combination of the initiating event and support system failures. These states together with the sequences borrowed from NSAC-84 as corrected for Seabrook were then_ fed into a frontline system event tree, which considers additional events needed to resolve the end states of the event sequences in terms of release categories. This main line event tree is
~
based in the event sequence diagram in Figure 2.
In this analysis, the NSAC-84 sequences were assigned to support. state R2E0 (loss of both trains of RHR with both trains of AC power available).
11 lt47P102986
The' event tree quantifications f ar L1RH,' L1PC, LISW, and LOSP are shown in Figures 3, 4, 5, and 6, respectively. The quantifications were based
~
on the SSPSA and RMEPS results for the support systems and initiators, except for maintenance unavailability. Train B of all systems was assumed to be unavailable for maintenance with a conservative value of unavailability of 0.1..This more than accounts for the higher chance of maintenance during plant shutdown. The initiating events L2PC and L2SW are assigned directly to support state R2E0 because of a very small char a of electric power failure with no loss of offsite power. The results of the analysis up to the point of support state are presented in Table 2, which is orgaiiized into three types of events:
Type 1 is events with one RHR train unavailable (RIE0, R1E1); Type 2, with two RHR trains unavailable (R2E0, R2El, and R2E2), and Type 3 is the set of NSAC-84 sequences. When the sequences are combined according to support state, the following results are obtained.
Mean Requency Support State (events per reactor-year)
R1EO 1.7 x 10-1 R1El 3.3 x 10-5 R2E0 2.0 x 10-3 R2E1 2.0 x 10-7 2E2 3.8 x 10-7 R
RXEY = Sequence with X RHR trains and Y electric power train unavailable.
The event sequence diagram in Figure 2 defines the possible progression for Type 1, 2, and 3 support state sequences. For Type 1 and 12 1447P102986 ru
2 sequences, consideration is given to operator recovery to prevent core melt.
On the other hand, such consideration is not made for Type 3 because such actions are already considered in NSAC-84. Next, the RO event question whether the RCS pressure boundary is open initially; i.e., vessel head or steam generator manway cover is removed.
For RO closed sequences, the ESD tracks the possible
-developement of interfacing LOCA conditions either through check valve closures or.RilR system repressurization via Top Events CV, RV, and 11C.
For pressure bonndary open sequences or all other nonbypass sequences, consideration is given to whether large and small penetrations are initially open and, when open, whether or not operator actions to secure these penetrations are successful.
Since'the containment sprays are not tracked in the ESD for simplicity, successfully isolated sequences could result in either a containment _ intact (SS) or delayed overpressurization (S3). 'Those with large or small bypass sequences are assigned to S6 and S2, respectively.
EVENT TREE QUANTIFICATION 1
-With reference to Figure 1,.the event trees were quantified in two
. stages. First the support ' system event trees were quantified for l
each. initiating event resulting in the quantification of the uncon-ditional frequencies of 5 different RHR~ support states.
RIEO, RlEl, R2EO, RZEI_and R2EZ (where RXEY is the state in whcih X RHR trains and and Y trains of safety grade AC electric power are rendered unavailable.
4 Then, the main line event tree was quantified 6 times, one for each RHR state and a separate quantification for the sequences borrowen from the NSAC-84 results. The derivatior. of the event tree split fractions i
for each event tree quantification.
Support System Event Trees (Figure 3, 4, 5 and 6)
The support system event trees wer'e quantified for the following initiating events.
i
.LIRH '- loss of 1 RHR train LIPC - loss of'l PCC train L2PC - loss of both PCC trains LISW - loss of I service water train I
L2WS - loss of b'oth service water trains LOSP - loss of off site power LlRH 4
The frequency of Icss of 1 RRR train during shutdown was estimated using the f ollowing model.
JLIRH = N I
RHR R11R where JTLlRH frequency of the initiating event
=
(events per reactor year) 4
)RHR failure rate of 1 RRR train
=
(dominated by the RHR pump) tRHR the number of hours per year in shutdown
=
Note the mission of MOV closure events in the above is by design, these events are included in the " type 3 events" borrowed from NSAC-84 and corrected for Seabrook having 2 RHR suction paths from the RCS.
2 The time on RHR, t RHR, is estimated using zion experience, whcih is viewed t
as a conservative assumption for Seabrook. The reason for this view is that the zion experience is worse than average for PWRs and does reflect
N-h'
/
the generally higher availability factors of the Yankee system of plants (Maine Yankee, Vermont Yankee, Yankee Nuclear Power Station).
In the~
first 16 reacter-years of experience of zion 1 and 2 there were 12 refueling outages of average duration 1,992 hrs and 3.05 maintenance outages per unit-year of average duration 488 hrs.
12 refuelings t o,m
+ 3.05 maintence events / reactor-year
=
16 reactor years x 488 hrs / outage = 2982 hrs / year Hence, in the first 16 reactor-years at zion the plant was shut down about 34% of the time.
We conservatively assume the same value for the plant lifetime at Seabrook.
The support system event tree quantification for LIRH is shown in Figure 3.
In.this event tree. it is assumed that the plant is initially being cooled with RHR train A and train B is in the standby.
Becuase of the strict train-wise dependencies at.Seabrrok Station, critical operation of RHR train a precludes unavailability of service water and PCC trains A, since both are.needed to operate RHR train A.
In normal power operation, the ~
unavailability of single service water and PCC trains is very low-no greater than 10-3 t o 10 -2 per train.
However, during plant shutdown, the unavailability due to maintenance is generally higher. For example, at sion the plant specific data shows single train maintenance unavailabilities of RHR, SW and PCC the range of.03 to.06.
It is conservatively assumed in the analysis that all safety grade non-operating subsystem have shutdown maintenance unava11 abilities of 10-1 This is greater than any shutdown maintence unavailabilities observed in the data.
LIPC The loss of one PCC train initiating event is analyzed in Figure 4.
The initiating-event frequency is estimated with the following model.
21 ks +)pe )
=
= failure rate of 1 PCC pump = 3.4(.5) per hour f rom PLG-0300 Section 6.
f tRHR = (.34) (8760) based on LIRH analysis
)s
= failure rate to start of standby PCC pump
= 2.4 x 10-3 from PLG-0300 Section 6.
[=meantimeto repair the initcally running pump
= 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> per PLG-0300 Section 6.
6.3 x 10-4/ reactor-year Hence
=
LIPC
The LIPC' event-tree is quantified using information from the RMEPS and SSPSA for service water train A and the'same train B maintenance assumptions as-with the LIRH event.
i, 4
LISW The co'nfiguration and failure, rate of.the SW cystem are comparable to the
.SW system, i.e., each train has an operating and a standby pump.
The event tree for LISW is quantified in Figure S.
The initiating event frequency is the same as that for'LIPC.
Note that this analysis includes the tunnel SW system only..The SW cooling tower system is considered in the subsequent recovery analysis.
LOSP The loss of offsite power event tree is quantified in Figure 6.-
The initiating event f requency is estimated using the following model:
-)#LOSP =
[
where LOSP = f requency of LOSP =.135 events / year (SSPSA Section 6)
.tRHR = time on RHR =.35 (see LlRH above)
EPR,= frequency of non-recovery of LOSP before core damage =.01 (assumed) i The above' assessment for EPR can be compared with EPR-1 in the SSPSA, a i.
value of.03 for full power operation.
The.01 value is viewed as
' conservative in comparison with EPR-1 since the time constants for core
{
recovery are much longer during plant shutdown.
i The event tree split f raction for LOSP in Figure 6 are based on the results of the SSPSA and RMEPS for. train A and 10% maintenance unavailability for train B used for all shutdown loss of cooling event's.
i L2PC and L2SW The initiating event f requency for loss of both trains of PCC and -loss of both trains of service water are estimated with the following model.
>E2PC =
f L2PC kL2SW=)I L2SW tRHR
' Po
' po where
~
JI = frequency of same event during power operation f rom SSPSA t po = hours of power operation assumed in SSPSA (8760)
I
,. - +
e r
., - - r
..--,-,-w
---,n.
,,.--e~-w..
~,.,.m
,,,---,,-n--
._,n.----..--.--e.
,.--.~,a-,-
Top Event RH For the LIRH and other RIA0 support states sequences, it is assumed that train A is initially used to provide RHR and train B is in standby.
- Hence, it is the train A subsystem that is involved in the initiating event.
Fo r these conditions, the failure to provide continued RHR cooling is estimated f rom the following model.
m+hs+
Ril =
PR p where
[m=RiiRpumstrainmaintenanceunavailabilityduringshutdown
= 6 x 10-from Zion data in NSAC-84
%ps=standbypumpfailuretostart vote
= 3.3(-3) f rom SSPSA Section 6 hpr=runningpumpfailuretorunvote
= 3.4 x 10 5/ hour from SSPSA Section 6 (0p=meantimetorepairtheinitiallyfailedpump=21hoursfrom SSPSA Section 6
[RH=6(-2)+3.3(-3)+3.4(-5)(21)=6.4(-2)
For RIAL, the same model is used except the standby pump must run longer to cover the repair time of a diesel generator - assumed to be one week.
.IIence, for R2EO, R2El, and R2E2,
./RH=.1 ORil - 6(-2) + 3.3(-3) + 3.4(-5)(168) = 6.9(-2)
Top Event OH RIE0 represents the most ideal conditions and minimum stress levels of these considered f or OM for these conditions, OM is estimated f rom:
OM = del = SC = BF where DEI = operators f all to recognize that RCS heat removal should be restored after running RCS pump stops.
SC = operators f all to align and restart a core cooling system BF = operators fail to provide long term makeup to chargining system Using appropriate values from Table 5-6 of NSAC-80, the following quantification is made:
OM = 1.0(-5) + 5.0(-4) + 1.0(-5) = 5.2(-4)
As shown in Yable 3, higher values are used for the remaining support states to reflect different and progressively greater st ress and comparison levels going through the sequence RIED, R2EO, RIEL, R2El, R2E2.
For type 3 events, OM = 1 to avoid double counting recovery already considered ir NSAC-84.
Top Event RO The fraction of time the pressure boundary is open is keyed to the time assumed for RHR shutdown cooling.
For consistency, since Zion data was used to quantify the latter, it must be used to quantify the former. - From Table 3-4 of NSAC-84, the total time the RCS is opened during maintenance outages is 5,014 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br />.
From table 3-1, the RCS open time during refueling outages is above 3,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />. Hence, overall the 31,687 hours0.00795 days <br />0.191 hours <br />0.00114 weeks <br />2.614035e-4 months <br /> of the Zion outage experience, the~RCS was opened (8014/31,687) =.25 of the time.
Top Event CV The frequency of CV is quantified per our response to question 47.
= 5.5 x 10 4 Top Event RV This event is quantified in RMEPS; it is estimated using:
/RV=21RV=_4'.8(-5) wherel RV is the failure rate of each RHR relief value = 2.4(-5) from SSPSA Section 6 Top Event MC For both trains of AC power available, the top event is estimated using the following model.
~ MC = 2(1MOV2+B MOV)+4()CV2+BhCV).
MOV where
) MOV failure rate (fail to close on demand)
= for MOVs - 4.3(-3)'per SSPSA
}CV=failurerate(failtocloseondemand) for check valves = 5.5(-4) per question 75 response.
MOV;bCV = Beta factors for each type of valve =.1 hence
[MC = 2[(4.3 x 10 3)2 + (4.3 x 10 3)(.1)s + 4[(5.5 x 10 4)2
+ (5.5 x 10 4)(.1)e = 1.1(-3)
For one AC power bus available (RIEL and R2E1) there is only one MOV on each suction path potentially available. For these states
[MC = 21 MOV + 4 ( ~}CVBCV)
= 8.8(-3) fMC=1 For two AC power basas unavailable A
r-- - - - - -
,. - ~,
.,.-e-
-._-....e._-
.---,,,,---=w,,w
,my,-~n
-+-~,
na,wae,
N Top Event LI This event question whether any large penetrations are open initially.
Large is defined as a total equivalent single opening of greater than 3" in diamter.
Examples of such penetrations are equipment hatch, personnel hatch, and contai nment purge penetrations. These penetrations maybe opened during shutdown unless fuel is being moved.
y The chance that large penetrations are opened is highly dependent on ht reason f or the shutdown.
If the reason is refueling, steam generator manintenance or othere maintenance on reactor coolant system somponents, it is-likely that large penetrations such as the equipment hatch will open.
If on the other hand, the outage occurs due to need to repair or maintain equipment outside the con t ai nmen t, (e.g. turbine generator related maintenance) there would not be a compelling reason to open up large penetrations in the containment.
To reflect the above considerations, LI in assessed as a function of the status;of event RD.
If RO is true (reactor coolant system is opened), it is assumed that LI is true (large penetrations are open) 90% of the time.
If RO is not true (reactor coolant system is closed), it 2s assumed that LI is true on 10% of the time.
Note that at Zion, of the 8,014 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> during shutdown that the RCS was opened, the fuel was being shuffled for 1600 hours0.0185 days <br />0.444 hours <br />0.00265 weeks <br />6.088e-4 months <br /> (roughly 160 hrs per refueling outage).
Hence 80 percent of the time that the RCS was opened, it would have been permitted by tech specs to have the equipment hatch open.
Top Event OL Given a large penetration is opened initially, the event questions whether the operators successfully close the penetrations before a potential release situation could develop. The probability of successful recovery is assessed as dependent on the RHR support state, i.e. the combination of the initiating event and the response of the plant support systems. At different support states then would be different levels of stress and confusion to inhibit operator recovery actions.
To provide an indication of the amount of time availabic to close the equipment hatch or other large penetrations, the time to core. damage, taken as the time to uncover the core was estimated for the following cases:
Cases Time to core uncovery (hr) 1.
Reactor vessel head open with water level at hot leg nozzle midplane A.
Loss of cooling at 2
days af ter shutdown 0.8 B.
Loss of cooling at 30 days af ter shutdown 2.6 2.
RCS filled at pressure < 425 psig with A.
Loss of cooling at 1 day af ter shutdown 5.4 B.
Loss of cooling at 10 days after shutdown 14 C.
Loss of cooling at 30 days af ter shutdown 22 3.
Water at refueling level with A.
Loss of cooling at 5 days after shutdown 72 l
B.
Loss of cooling at 30 days af ter shutdown 162 L
It is not known how quickly the equipment hatch can be secured. Our current information is it would take several hours _to attach, and up to 8-12 hours to secure all the bolts -and establish a tight seal. Just how quickly this process can'be accelerated is uncertain. To address this uncertainty, a base c' se - and a bounding case. are performed.
In the base case, it is assumed that a
the mean time to close the hatch is 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.
In the bounding analysis, it is assumed that the hatch remains open with a probability of 1.
From the Zion data in NSAC-84, there were 10 reflueing outages and 24 forced maintenance outages resulting in an average outage duration of about 39 days.
~
Of the entire 31,687 hours0.00795 days <br />0.191 hours <br />0.00114 weeks <br />2.614035e-4 months <br /> of outages, roughly 1% of the time the RCS was drained, 5% of the time refueling was taking place, 5% of the time the plant was not on RHR, and in most of the remaining 89% of the time, the reactor system was filled on RHR.
Based on the above recovery time,'and assuming a i
4-hour hatch recovery time, it is seen that with the RCS drained to the hot leg nozzle midplanes, the chances of successful hatch recovery are not very high.
While under all other conditions, the chances are high. Therefore, the' base case and low stress levels, a value of.01 is used for failure to isolate large penetrations. For degraded RRR states, this valve is increased to correspond with higher stress levels, as indicated in Table 3.
In the bounding case, a' failure frequency of 1 is assured for all states.
Top Events Si and OS It is conservatively assumed that small penetrations are open 90% of the time and the chances of recovery are assessed at levels compoarable to those for OL, even though all small penetrations can be isolated quickly.
2 Results The results of this preliminary analysis of shutdown loss of cooling' events are shown in Table 4 for the base case assumptions on event OL.
To bound the consequences of these events, accident sequences were assigned to the existing-PSA release categories, even though the release f ractions for shutdown events would be expected to be considerably lower than those calculated for power operation events.
Based on what is believed to be a very conservative set of assumptions in this base case, the impact of shutdown events is assessed to result in no greater than a 14% increase in core melt frequency, and an 18%
increase in category S6 frequency.
For the bounding case of no credit for event OL, the frequency of S6 would increase to about 5 x 10-6/ year.
l The impact of these boumding estimates of shutdown events on the dose vs distance curves for 50 rem and 200 rem whole body gamma doses are shown in Figures 13 and 14 f or the base case OL and bounding case OL assumptions, respectively. Our best current statement of risk levels is represented by Figure 13.
As seen f rom this figure, the addition of shutdown events impacts the right tails of these curves, but the combined results at 1 mile are still less than the NUREG-0396 valves at 10 miles.
Even with no credit for equipment hatch recovery as assumed Figure 14, the combined shutdown and power operation results fall below the NUREG-0396 10 mile levels at Inss than 2 miles.
Hence, even a very conservative analysis of these events does not impact the conclusions of the sensitivity study.
It is expected that a more detailed investigation of l
these events would result in much lower levels of risk than either set of l
results presented here.
-~
e
-r1
- 4..,
-37 REFERENCES I
-1.
Pickard, Lowe and Garrick, Inc., Westinghouse Electric Corporation, and Fauske and Associates, Inc., "Seabrook Station Risk Management and Emergency Planning Study," prepared for Public Service Co'mpany of New Hampshire, New Hampshire Yankee Division, PLG-0432, December 1985.
2.
Pickard, Lowe and Garrick, Inc., "Seabrook Station Emergency Planning Sensitivity Study," prepared for Public Service Company of New L
I Hampshire, New Hampshire Yankee Division, PLG-0465, April 1986.
4
)
1 3.
Letter from S. M. Long, U.S. Nuclear Regulatory Commission Staff, to R. J. Harrison, Public Service Company of New Hampshire, " Request for 1
Additional Information...," Docket Nos. 50-443 and 50-444.
4.
Pickard, Lowe and Garrick, Inc., "Seabrook Station Probabilistic Safety Assessment," prepared for Public Service Company of New Hampshire and Yankee Atomic Electric Company, PLG-0300, December 1983.
5.
Bley, D. C., and J. W. Stetkar, " Zion Nuclear Plant Residual Heat Removal PRA," EPRI/NSAC Report NSAC-84, July 1985.
6.
Pickard, Lowe and Garrick, Inc., Westinghouse Electric Corporation, and Fauske & Associates, Inc., " Zion Probabilistic Safety Study,"
prepared for Commonwealth Edison Company, September 1981.
. ~
-_. ~....- _-
j i
i Table 1.
CORRECTION OF NSAC-84 RESULTS FOR RHR LOSS l
TO ACCOUNT FOR 2' SUCTION PATHS I
l 4
l 1
Revi sed Seabrook l
4 Failure Cause Mean Value Dominant Contributor Correction Failure
{
Factor Frequency i
. Hardware Failures 6.08-2 Spurious Closure of RH8701.or RH8702
+.072 4 38-3 i.
j Maintenance 7.37-3 Running.RHR Pump Fails with Standby Pump x1 7.37-3
}
Out for Maintenance 1
i Human Errors 6.00-2 Errors During TSS 15.6.36 or Inverter x.031 1.86-3 j
' Switching (RH8701 or RH8702 close)
Support System 2.94-6 Component Cooling Water Heat Exchange r x1 2.94-6 l
Failures Failures Dependent Component 6.03-3 RHR Pumps, Fail During Operation x.1 6.03-3 3
Failures b
l l
l i
Total 1.34-1 1.96-2 l
l l
1 i
i i
l
TABLE 2.
CLASSIFICATION OF SUPPORT MODEL ACCIDENT SE MAIN LINE MODEL QUANTIFICATION CASES Impact Vector P
r Classif cation A
8 A
B Type 1 Events - One RHR Trafn Made Unavaliable L1RH
(
1.7-1 X
R1E0 L1PC*PR 5.0-5 X
RIED LOSP*G8M*01R 2.0-5 X
X R1El LOSP* gal *DIR 1.3-5 X
X RIE1 LISW*SR 5.1-6 X
R1E0 LOSP*PSM*PR 3.4-6 X
RIE0 LOSP*W8M*SR 3.8-7 X
X R1El LOSP*PA2*PR 8.9-C X
RIE0 LOSP*WA3*SR 5.6-8 X
X R1El L1PC*WA3*SR 1.8-9 X
RIE0 Type 2 Events - Two RHR Trafns Made Unavailable s
LIRH*P8M*PR 1.8-3 X
X R2E0 L1RH*W8M*SR 2.0-4 X
X R2E0 LIPC*P8M*PR 5.5-6 X
X R2E0 LISW*WBM*SR 6.3-7 X
X R2E0 LISW*P8M*SR 5.7-7 X
X R2E0 L OSP
- gal *G8M*D2R 3.5-7 X
X X
X R2E2 LOSP* gal *P8M*DIR*PR 1.4-7 X
X X
R2E1 LOSP*PA2*GBM*01R 6.0-8 X
X X
R2E1 LOSP* gal *W8M*DIR*SR 1.5-8 X
X X
X R2E2 i
L OSP*PA2*P8M*PR 9.8-9 X
X R2E0 t
LOSP*WA3*W8M*SR 6.9-9 X
X X
X R2E2 LO!P*WA3*P8M*SR 6.2-9 X
X X
R2E1 1
LOSP*WA3*G8M*01R*SP 3.9-9 X
X X
X R2E2 LOSP *PA2*WSM*SR 1.1-9 X
X X
R2E1 L1PC*WA3*WSM*SR 2.2 X X R2E0 L1PC*WA3*P8M*SR 2.0-10 X X R2E0 Type 3 Events - From 7.6-6 X
X NSAC-84 R2E0 t
NOTE:
Exponential notation is f.e.,1.7-1 = 1.7 x 10-I. indicated in abbreviated form; n
1448P102886
TABLE 3.
SUMMARY
OF SPLIT FRACTIONS FOR MAIN LINE EVENT TREE Event Tree Quantification Cases Main Line Event Tree Type 1 Type 2 Type 3-Top Event (NSAC-84).
R1E0 R1El R2E0 R2E1 R2E2 RR 6.4-2 6.9-2 1
1 1
1 VFf 5.2-4 1.0-2 2.0-3 1.0-1 1
1 TRT
.75
.75
.75
.75
.75
.75 Elr 5.5-4 5.5-4 5.5-4 5.5-4 5.5-4 5.5-4 10 7 4.8-5 4.8-5 4.8-5 4.8-5 4.8-5 4.8-5 FRT 1.1-3 8.8-3 1.1-3 8.8-3 1
1.1-3 ETlR0
.90
.90
.90
.90
.90
.90 Il?lTUI
.10
.10
.10
.10
.10
.10 IR 1.0-2 3.0-2 1.0-2 3.0-2
.10 1.0-2 El-
.90
.90
.30
.90
.90
.90 IBi 1.0-2
.10 1.0-2
.10 1
1.0-2 X = Event Success Y = Event Failure NOTE: Exponential notation is indicated in abbreviated form; i.e., 6.4-2 = 6.4.x 10-2, 1448P102886
TABLE 4.
KEY RESULTS OF SHUTDOWN SEQUENCES FOR SEABROOK STATION Release Category Event Tree S5 or'S3 S2 S6 R1E0 5.6-6 5.0-8 1.9-8' RIE1
.1.9-8 1.7-9 2.0-10 I
R2E0
'4.0-6 3.5-8 1.4-8 i
R2E1 1.8-8 1.8 1.9-10 R2E2 3.2-8 3.3-7 1.2-8 Type 3 7.5-6 6.8-8 2.6-8 Total for Shutdown 1.7-5 4.9-7 7.1-8 Events l
Total for Power 1.1-4 2.0-5 3.2-7 Operation Events i
Percent Increase 13.3 2.4 18.2 with Shutdown Events NOTE: Exponential notation is indicated in j
abbreviated form; i.e.,
5.6-6 = 5.6 x 10-6, i
1448P102886 l
RHR PLANT STATES SUPPORT SYSTEMS (STATUS OF RHR AND MAIN LINE EVENTSEQUENCE fN!TIATING EVENTS EVENTTREES AC POWER SYSTEMS)
EVENTTREE END/ RELEASE STATES LOSS OF 1 RHR TRAIN L1RH R1EO y
NOCORE
+
i LOSS OF 1 PCC 4
TRAIN L1PC CORE R1E1 DAMAGE WITH NO BYPASS SUPPORTTREE FOR LOSSOF 2 PCC TRAINS L2PC AVAILABLE i
(FIGURES 3,4, AND 5)
LOSS OF 1 SW EVENTTREE CORE TRAIN LISW R2EO y
OUANTIFIED FOR DAMAGE WITH m
l d L EACH RHR PLANT SMALL BYPASS STATE S2 (FICURES 7 THROUGH 12)
LOSS OF 2 SW 4
TRAINS L2SW CORE DAMAGE WITH I
R2E1 LARGE BYPASS LOSS OF SUPPORTTREE FOR S6 OFFSITE PC#ER
> LOSSOFOFFSITE LOSP POWER (FIGURE 6)
APPLICABLE CORE R2E2 DAMAGE EVENTS FROM NSAC-84 FIGURE 1.
ACCIDENT SEQUENCE EVENT TREE MODEL FOR SEABROOK STATION SHUTDOWN COOLING EVENTS
9 5
vl!
i Cs
$5#
M5Y EMI i!!!*
9 c_l:
,. 8
'n BS 33' N_ !"
a 8
E b I
ik.
{a lli u
o l!
=
t 1lll 101'"h
~
1
{
i it.el5l i, i +
r)
I g
I irl is A
1l8
!!b "
,a v
il i,
(
W iiil 11 lllb!!
e e
..lil1 I
=!!lli!
ho,j1,;[inlil u@v ia
,gj s !,
l i
m
- lill li l
-e
-e
-i u'8ll
~
INITIATING SERVICE SER'/ ICE PRIMARY PRIMARY SEGENCE EVENT WATER WATER CC wONENT COWONENT FFE1ENCY NUMBER &
L1RH TRAIN A TRAIN B COOLINGA COOLNG B (EVENTS PER RHR TRAING WA WB PA PE REACTOR MADE YEAR)
UNAVAILABLE
.20/ YEAR n 1
m
.90 a1 m
.90 1
.185 1
w<
s-se x,
.10 2
1.8 ( 2) 2 0
m 3
0 1
v 4
0 2
.10 1
~5 2.0 (-2) 2
_ _, g_
0 8
0 2
O
- w- -a 7
0 1
8 0
2
--w-
--w-9 0
2 GF = GUARANTEED FAILURES I
FIGURE 3.
SUPPORT SYSTEM EVENT TREE QUANTIFICATION FOR LOSS OF ONE RHR TRAIN (LIRH)
~
INITIATING SERVICE SERVICE PRIMARY PRIMARY h
FREQUENCY NUf4ER M SEQUENCE EVENT WATER WATER COWONENT COWCNENT W L1PC TRAIN A TRAIN B COOUNGA COOUNG B (EVENTS PER RHR TRAINS WA WB REACTOR MADE PA PB YEAR)
UNAVAILABLE 6.3 (-4) n - 1 n.90 no n.90 1
0 0
s-s-
se se
.10 2
0 1
1 n.90 3
5.0 ( 4) 1 v
.10 4
5.5 ( 5) 2
.10 0
Q
,, gp,,
5 0
1 1
,g_
8 6.2 ( 5) 2 3.5 ( 4) n.90
.90 7
1.8 ( 7) 1
_ _ g_
.10 8
2.0 (-8) 2 10
,,,,,,g.,_,,,,_g,,
9 2.2 ( 8) 2 GF = GUARANTEED FAILURES l
i FIGURE 4.
SUPPORT SYSTEM EVENT TREE QUA.1TIFICATION FOR LOSS OF ONE PCC TRAIN (LIPC) l
)
L L
'l fL
0 SEQUSCE INITIATING SERVICE SERVICE PRIMARY PRIMARY h
FPEOUENCY NUfeEROF EVDg WATER WATER COWONENT N
6!
(EVENTS PER RHR TRAINS LISW TRAIN A TRAIN 8 COOUNGA COOUNGB {
REACTOR MADE WA WB PA PB YEAR)
UNAVAILABLE 6.3 (-4) g0 g
n n
1 0
0 2
0 1
n 3
0 1
v 4
0 2
5 0
t o
_ _ w_
8 0
2
_g_
1 n.90
.90
_7 5.1 (-4) 1
_ g,
.10 8
5.7 (-5) 2
.10 9
8.3 ( 5) 2
_, g_ _ _ g_
GF = GUARANTEED FAILURES FIGURE 5.
SUPPORT SYSTEM EVENT TREE QUANTIFICATION FOR LOSS OF ONE SERVICE WATER ' TRAIN (LISW)
o e
PRMARY PRMARY SEGJENCE INITIATING N
POAER POWER WATER WATER M
COOUNG (EVENTS PER RHR TRAINS TRAIN A TRAIN 8 TRAIN A TRAIN 8 TRAIN A TRAIN 8 FEACTOR MADE GA G8 WA WB PA PB
. YEAR)
UNAVAILABLE 4 6 (-4) n.93 a.90 m.98 n.90 n-1 m.90 1
3.1 (-4) 0 g
ms ms
\\1
%s rr rr
.10 2
3 4 (-5) 1 2.9 (-3) n.90 3
89 (7)
~ t' PA2
.10 4
9 8 (-8) 2
.10 m-1 5
'3 8 ( 5) 1
_, g, 2.9 (-3) 6 1.1 ( 7) 2
,g, PA2 18 (2) g 90 90 7
5 6 (-6) 1
,g,
.10 8
62(7) 2
.10
'9 69 (7) 2 7,
.10 98
-t to 41 (-5) 1
_ g,
_,y_
2.9 (-3) 11 12 (7) 2 g.
PA2 18 2)
,,,,,, gp,, _,, g,, _,, 7,,
12 7.7 ( 7) 2 73 (2) 90
.90 Q
,, op,,
.90 13 25 (5) 1
,y,
.10 14 2.7 ( 6) 2
.10
,,,g,,,,,__g,,_
15 30 (6) 2
.10
_, gg,
,, gg,
,, cp,,
16 34 (6) 2 LO6P = LOSS OF OFFSITE POWER WITH NO RECOVERY BEFORE CORE DAMAGE GF = GUARANTEED FAILURE FIGURE 6.
SUPPORT SYSTEM EVENT TREE QUANTIFICATION FOR LOSS OF 0FFSITE POWER (LOSP) m
f a
1 leefftATuse RHR OPGRATOst RCS Rwn RNR Run Lames OPERATOR BMALL OPSR ^ Tost 9'
geouessCg SVGesT1 COOLNs0 ACTIONS TO PRENuRG CMGCat ROLisp SUCTIOss PeNETA.;Teosse ACTeosseTO PesseTRiveOsse ACTsosse TO tesO PRecuessCV
$UPPCsBT hoheerTAessGO Pettvess?
00ussOARY valves Valves asove eseeftALLv CLoseLanes sNsTtCLLv CLoeg ensatt.
37;T3 teveerfs Pess i
I TV,ATO gT asSLT OPGeo CLOSS OPGRATE CLOSG SGCURGO PENGTRATIONS SECURfD PENGTRATIONe i REACTOst 83 Ret One RO CV Rv hsC
~ Le OL Si 08 I
YSAmt
'8 6** c*
ou c*
2 ou 2
9 c8
__-------C
C
A0 m,, =9 4
M t.? f.75 1A4-n a
33 9.7 (4 m-t m.t.
e a
t., e-n wr wr m
m,,-t w
t..
9 A e-n et us 1 A e-n e
a t., e m -t g-t
_ _,4 m
_ _ _ g.t.
te n
. 9 e.n
.00 m=9 ft 98
'43449 sv 9.0 f-n It at 44049
.te mal m, a.to 13 se 0.744
~ ~ *
- mr A0 m,=1 to SS 3.9 f-ft s
18 f 31 ft St 8.1646 9.0 f-29 te se 3.? f-se 42041 m, = 0 m,,.00 C.te et es 3.34499 3
.00 m, = 9 10 se 3.t f.tel 3
t.0 f 25 et 33 33 6.stl
.90 at a.te M
M f.? 1939 m
- * * * ~
se rr A0 et to ge 3 8 (.99) m we 1.0 0-n St st in t.tn tDi n tt 88 t.f t-tn t t f-t' to te t o f-tM 8 ' I'#
Rf8VLTS E 80
- f A (4)
Z St
- SJl.el ZSS*24le FIGURE 12. EVENT TREE QUANTIFICATION FOR TYPE 3 EVENTS. EVENTS TAKEN FROM NSAC-84 THAT ARE APPLICABLE TO $EABR00K STATION
e gestTIATwee mesa op0AAton mCs num num num LAmos openATom sesALL crenATom HouGNCE 0v08871 CoeuseG ACTeoses To poemune CNecir never suCTioso PGNGTm TioNG ACT*0NSTo PSNE Tm..TeDast ACTH>880 70 GNO PRecuggeCY SupponT asaessTAsasco entveer?
goussoAny valves valves asovo ameTtALLY CLosELAmos aNeve/ LLv CLoes sasALL gT_Tg gevenTg ega STAft ang(T opeN CLOSE optmATE CLost SECusto ptNETmAftoms secuato ptNETearposee meACTom niet RN ons no Cv av mC u
og se os R
TaAmi
" ' - " ^ 88 0"
'88-"
2 0"
" ' - =
O -'
n.=
. C
C
- 2 w
.00 a =9 4
98 1J 6-79 3e 9.9 t-M 9
99 12 f49 at n.to 0
08 930-79 00 m
se me 9
SS 9.914p te n, e =9 9.0 4-M e
St 9.10-el 9 O f-M 9
80 tJ 841 70 n =1 n =1
.00
.to te M
9 A 8-79 00 m.t il 98 3 A 040 se th f.M 11 St sat 49
.te m =9 n.te 13 M
4ttal mr mv
.00 a =t to N
33879' wr 1.0 f M tt 92 t A f-el l e f-M 14 80 4 t t-el 4 8 f41 a =9 m.90
.te 17 98 14 f.tH se me r
.N A =k t$
E
$ $ $-IO 3r t O f-M 19 St 94ftM 10 n =t n.te 10 08 f.0 6 9M sr we
.00 a =9 ft e6 98190 we t O f-M tt si SAftM 908-21 it to 106921 90 10 f.t H
'8 N
to 10W at9ULTS Zesela(49 Z80 etAl4)
Ist a s.elas FIGURE 7.
EVENT TREE QUANTIFICATION FOR R1E0 - EVENTS THAT FAIL ONE RHR TRAIN AND NO TRAINS OF SAFETY GRADE AC POWER
isstTIATiss0 RNA OPfRATOst DCS mwn mum num Lamos opem ' TOR SMALL OPeRATO84 NE
- foueNCe evewes Cocume Actione ro passovne Cwecu nouse suction pomera.rio=e Actiome to pomera.rio=e ACrione ro )
susoom7 MAeNTAfNeo PeeveNT 00VseOAmy watves Valves Move INe fs A LL Y CLose LAmos e: ers LLY cLOSESwALL 4 e=o paeove=cy g7ATS MELT ope 4 Close OFeN~Te CLose SeCuneO Pe = e TR -.TtONS SeCUReo Pe =e TR A fiONS 37.78 feveart rem 8
OL J
08 k
veARI meACTOR 1
O as e a
- s. w a ~'
s2 w
* ' - = ~n
_ c
' =
DO m,.00 e
M 44 e.9m m
.t.
S 82 p.9 e.ttl
.se m.et m,.se e
se es e.ie w
T, 20 m, 20 t
H eb s.et s
,te 0
et 4J e.100 s..e n e
a u,.
.=
g.'
O
.0 n
u u -si m,
s is n
... w m,,.ie n
u s...
.00 m,.00 to M
t a e.et s
,e n
, s e...
3a e.n ie
=
.. e..
. s e-ei m,.
m,,.e
_,,,,-C...
u a
u i.i.,
s
..a m,, e.
,e a
ee.n
,e n
e e...
.ie m,.e, m,.i.
so a
e e.....
s s
n si n
e A e.9 9 sr
.to 22 et 7 3 tes o e-t) 23 m
3 g e.,s) 8 8 '-8' 8*
7 8 88 M
88 840*
- ReSULTS Z 8.*lA44)
Z 82
- 1.t f41
% N
- 2A 6100 FIGURE 8.
EVENT TREE QUANTIFICATION FOR it.E'.
900sTtATueG AMA OpenAfon 8000 mum num num Lames OpenAfon sesALL OpenAT0m I
300uessCS OVessTt 000L8800 AC70810 70 poeMune CNeCu motsee sucylog PameTmArposee ACT*0sse 70 PoweTo Te0Ne ACT*0NS 70 -
8880 PascusesCV pupp0m?
hsA885TAs0000 MISVGBf7 00UI00ARY VALVt3 V LVfe t#Ove INtTt ' LLY CLOGELames eN87 TALLY CLOOG 9hAALL ST-.TE ($VENTg pgg ST TE teELT ope 4 CL0se OPinATE (LOSE SECumEO PENETR.TIONS SECURE 0 PENGTm.TB0888 $
meACTost m300 feet Oss se0 CV Av asc LI OL tub 08 a
VEAmt
'A
- '* Oa 1
0=
e O-'
_.- 3 0=
sAr*
- '* Om g'*
_c
=
t***
90 m =1 4
OS 9Af4 me u e-a e
u u s-tm m
m,, -t m,,.t.
=
u e.
m a
at e-n m,,-,
teem e
a u
i A e-m
.Ae4
.n g -t g -t
_,,, g A.
_ _ _, g.to ie a
usn A0 ot 99 88 sat 4 m
se t A f.S tt Of SA S.sp 90 m =9 m,te 13 M
3Af4 As a.9 to 08 f.7 6-79 se 4 0 f-fl to et 3.7 f.ep tot-M te se 2004 4 4 f4 n, =9 m 90
.te ff te SJ f.f tl.
s we r
88 m =t it SS tJ f-top f
wr 9 A t-B 10 92 120-99 I
I
.te m,, -t m,.te a
es u s-tm
]
00 m -t it se
' 2 0-999 re J
t.0 t-il it 83 1J f-in 9 O f-n 12 se t A t-in 6
M M _
M _
$ 8 l-48 fe te it.? f-40 Re9VLTS EOS*48le I st.,s A se ZeeetA14 FIGURE 9.
EVENT TREE QUANTIFICATION FOR R2E0. EVENTS THAT FA!L TWO RHR TRAIN $ AND NO TRAINS OF $AFETY GRADE AC POWER
e i
10 eft 1ATIIs0 Rout EpWBATOA AW RNA RMS RMM (AR$$
OPORATgm OgsALL
, OPgAATOR gggggggg eTn. THpes ACTHps0TO PgesermATIOese ACTH>esTO SNO PageweseCV SYS887f OOOLHet ACT*0NSTO pm W G CNecst meLISP SUCTiO,N Pgm,NeveALLv CLoseLAmes smit ALLv Close sesALL 97.79 (evesste Pem amov SU880ft?
48AtesTAspsEO PRevels?
SOUNOAny Valves Valves STATE RsGLT opese CLost OPemATE CLoes stCURED PtNETnATeONS SECumto PtNefmafice86 RSACTO84 m303 met One no Cy my asc La OL
- s Os 5
veAnt 0't e, '
18W
=
c*
=
- ^t-'*
c=
x"
_x" 30 6
4 M
e n, a 9
9 93 8e44
.88 n.80 n.te e
M T.? 4 00
~ ~ ~ * * * -
me me 7
88 8
n,, 0
.30 t
e er es.
.ie e
es ese4
'O M
34 4 g.1 g.1
, _g m
_ _ _ z.10
.n m
n.
n M
e
%v 9
12 Bf 13 4.?)
13 98 2 A f-08 90 a De n, e.10 mr
.00 a0 to 88 9
sr 9
16 St
$J t-4D te te tsf4 90 tr M
e
,. _ 4.<e m
s e <* no n
me me
.00 AD 18 88 9
we t
it O2 0
,te n.go n.10 20 SS e
si sr 90 n0 ft SS 0
wi 1
22 82 0
.10 23 et 0
24 M
14 f-Hf M
98 18 f.909 6 8 f-49 MtsVLTS Ise a.t t.
I43 a 3J17)
Ist
- tJ f-el I
FIGURE 11. EVENT TREE QUANTIFICATION FOR R2E2. EVENTS THAT FA!L TWO TRAINS OF RHR AND TWO TRAINS OF $AFETY GRADE AC POWER
n e
sufftAftNS RMR OPORATOR RCs RNR RNA pun Lames OPgAATOR eMALL OPERATOR ggOug80CE 3
/
COOLN00 ACT10NO TO PR W S CMSCIE K ELISP SUCTION PSNGTit 710800 ACTIONS TO PONGTai.TiONG ACTIOneg TO 8800 pegougNCY T
SAA4NTAINGO pit 9VG007 OOuse0ARY WALves VALvet esOve sNITI ALLY CLOSELAp08 sNsTI ALLv' CLOSSSMALL gTAfg tevgNTg Pest staff asELT OPEN CLOct OPERATE CLOSS secumfD PENETRAftONS SECusto PENETR ATIONS RO *CTOft R281 RM One RO CV RV asC LI OL ss OS 5
ftARD o'c 0'
88 671 2
o't ta t-71 8
3
- h"'
a2e c"
_c*
.90 m.90 4
M 4.1 f-100
%r 10 S
S1 434-99)
.90 m.97 n.10 8
88 4A l-tGB we wa A0 n AD
?
M SA f-Si wa
.te 8
82 4 A f-tel 3 0 f-21 8
SS S A 4-109
.76 n.t n.t
.90 90 90 M
1A(4
.00 m.00 ft M
t.9 f St nr 90 12 52 1.2 8-80 10 n A7 m.90 93 SS 9 A (-100 sr ir 30 a 90 to SS t.3 (4 se
.to te si e A f.ies
$ 0 f-M 18 SS 434-199 4 4 f-8) n.t a to
.to 17 98 G A f-tet A0 mm te M
. A e-TM
.te 19 91 GA f 949 10 n.97 n.to 30 te 7 A l-t St mr mr 90 n.00 ft 9f 8.7 f-tal se
.90 22 82 eJ f-tet 3 0 f-M 23 30 tJ f-tSt
- b"
- J I'
-_3..
'4
- 2 b'"
.LL'
ResuLTS 70s
- ta f-4)
%52
- I A f-Of Ise
- 1 A 9:0e l
l l
FIGURE 10. EyrNT TREE QUANTIFICATION FOR R2E1 - EVENTS THAT FAIL TWO RHR TRAINS AND ONE TRAIN OF SAFETY GRADE AC POWER t
I
I DINTIATWee RMet OPGRATOR RC$
RMR RMM RMM LARGE OPERATOR
' suaLL OPERATOR es0Useste EvGefff C00Lisse ACT10seB TO PR m e CMEC4 peLISP DUCTsose PeseETRATIosse Acta 0meTO PoweTRATs08se ACTrosse TO elec PneousesCV SUPPORT AAAINTAtesso PReveer?
00Ups0ARY WALVG8 WALVE8 Move BNITt? LLY CLOOGLAmos pNefgJ LLY CL0gg ensALL STATO tevessyg pget STATO esELT OPtse CLOSS OPER'TE CLOSE SGCURec PtNETR ' TNDNS SECURtc PENGTRATI0ses RgACTOR k
TYPE 3 Re4 Oes 18 0 Cy 5.v asc
' Lt OL aW 08 vtARS I
'8'*' O' t
0" o
I c'
=
=
^E g
g '"
8 88
'I I"'I A0 a=9 4
98 1.? l-79 mr 128-3 9
Of 1.744
.00 m=t m.to e
SS t.7 f-ft we wr
-.80 n,=1
?
es 9A44 120-5 8
82 1ASM I A l-2 SS 1.? f-el g-t g-t
_ _ g.
_ _ _ g.10 30 m
..t e-n
.=
ii
=
44 4
.=
m,-1 148 2 12 82 44 6-el
.t0 m,,-t m,,.90 13 m
.7:4 30 n =1 14 SS 8.9 t-79 se 1 A t-2 16 St 8.114 1 A f-21 te SS 8.7 l-el 4A44) n =1 n.30 16 17 N
SA f-tti me se e
As n=t 19 M
2.26100 9.0 l-M 19 92 2.24-92 n, 10
.te ot 20 N
2.? f-f M n
we w
.30 m *f 29 SG 2A l-Ifl
\\r 1.0 t-M 22 52 12ttn 1 A 4-21 23 as 2.7 t-tal t.t (-m M
se 3.0 f-tM 6 S l-4) 26 98 2.1 f4 ResuLTS E SS *72(4)
Z St * $4 t 81 Z M = 24 (-el FIGURE 12. EVENT TREE QUANTIFICATION FOR TYPE 3 EVENTS - EVENTS TAKEN FROM NSAC-84 THAT ARE APPLICABLE TO SEABROOK STATION
1 i
iiiieiig e
i i e i isig i
i i i i s i_
~
NUREG-0396 en H
Z SENSITMTYSTUDY FOR EVENTS AT POWER ~
- g
......._. - eve,,,
Me sd2 gW 01 :
az
$h b-~%
bui
\\
\\
\\
50 REM We
\\
\\
88
\\
\\
OW
\\
k gh 0.01 7 g
h
\\
\\,
gm g
~
O l
}
w l
- ==4,=.( I[,So Re%.
A 86D-lMg i
=
I bI
' ieil 0.001 I
e t I eI ei 1
10 100 1,000 DISTANCE (MILES)
F,eurc/3 COMPARISON OF 200 REM AND 50 REM DOSE VERSUS DISTANCE C
)
CONTRIBUTIONS FROM SHUTDOWN EVENTS l
1
,, m, g
!i; SENSITMTYSTUDY g
FOR EVENTS AT POWER
>. g
....-.-. SNuroOwN EVENTS fH4 WITH NO CREDIT
~
4Gj FOR EQUIPMENT HATCH RECOVERY d'
18 o'
W 95 y$
T-s w ur
\\
!a
-%. (%/
~
88 s\\
en oo' i
i \\'
be
\\ \\\\ s
\\
1" l
\\\\
\\
Sw l
s\\
\\
98 st
's Ob
/
20o }.,}
t.
oo l REM b l/
g 200 REM 0.001
I 1
10 100 1,000 DISTANCE (MILES)
' gg /d COMPARISON OF 200 REM AND 50 REM DOSE VERSUS DISTANCE CONSERVATIVE ASSUMMION OF NO CREDIT FOR OPERATOR REC 0VE EQUIPMENT HATCH
i l
RAI 22
It is the staf f's understanding that preexisting violations of containment i
I integrity were " included" in the PSA by assuming the average ef fect was to raise the containment lesk rate to the design basis value of 0.1%/ day.
Compare this assumption with the containment. integrity violation a.
data presented in NUREG/CR 4220.
b.
What contributions would these containment integrity violation data make to the probabilities for each of the release categories (assume the S5W category is redistributed over all the appropriate categories by the conditional probabilities of preexisting leakage paths of the size appropriate for each category).
RESPONSE 22 As discussed in Section 11.6.3 of the SSPSA (PLG-0300), release category SS represents accident sequences where the containment remains intact (does not fall). All containment failures and bypasses, including failure to isolate the containment are included in the other release categories (S1, S2, S3, S4, and S6). Preexisting con.ainment leakage is quantified in S5 by assuming an effective average leak rate equivalent to the containment design leak rate of 0.1% per day.
As shown in Section 11.6.4.1, this is a reasonably good approximation when compared to Weinstein's Containment Integrity availability work (reference 11.6-18 of PLG-0300 - also referenced in NUREG/CR-4220). This leakage is also imposed on category S3 but is increased to reflect the predicted higher containment pressures. Weinstein's work found PWR containment integrity to be available 97.3% of the time and an average leak rate of 31 times design basis the rest of the time.
a.
NUREG/CR-4220 suggests that PWR containment integrity availability could be as low as 71% with leakage in the range of 1 to 10 times allowable 29% of the time.
The probability of larger leaks (28 square inches) is estimated to be in the range of 0.001 to 0.01 with a point estimate of.005.
Section 9 of NUREG/CR-4220 states that these are upper bound estimates of containment unavailability.
Section 6.1 indicates that the 0.29 unavailability is based on 215 events in 740 reactor years where type B and C leak testing results exceeded 60% of allowable leakage; a leakage duration of one year is simply assumed for each event.
Section 6.2 indicates that the.01 to.001 chance of a 28 square inch leak is based on 4 events in 740 reactor years ranging from "small drilled holes" to an open six inch valve; again a one year duration is simply assumed with little, if any, basis. There is insufficient information presented in this document to assess the applicability to Seabrook of each event.
We believe it is highly conservative to assume that 0.29 events I,
per year will occur at Seabrook in which 60% of allowable leakage is exceeded to a level one to ten times allowable leakage and that such conditions will exist 29% of the time.
It is also highly
conservative to assume that 0.005 events per year ranging "f rom small drilled holes to an open six inch valve" represents a 0.01 to 0.001 chance of a 28 square inch leak.
The reasons why we believe the application of these assumptions to Seabrook to be conservative are as follows.
- First, it seems extreme to assume that evidence of exceeding Technical Specification limits with LERS will mean that the extent of the leakage is ten times the limit.
Also, events involving mispositioning of manual valves are' subjected to monthly surveillance testing at Seabrook, therefore, one year is an inappropriate fault duration for Seabrook.
Many of the possibilities for mispositioned valves will be covered by automatic containment isolation actuation, whose failures are included in the PSA.
With regard to the four large leakage events noted in sections 4.1.7 and 6.2 of the NUREG, the information provided is insuf ficient to compare.directly with Seabrook; however, it may be that none of these events are applicable to Seabrook (especially for durations of one year) for the following reasons.
1.
The containment purge valves at Seabrook are leak tested every six months or less; their position is checked monthly; they are actuated valves which receive containment isolation signals (failure of containment isolation signals / valves is included in the risk models).
2.
Seabrook's containment is three to four feet thick; it is difficult to imagine that a hole could inadvertantly be drilled through it, never mind go undetected.
3.
At Seabrook, valves, flanges, penetrations, airlocks, etc., are leak tested and position checked after any maintenance activities on them.
4.
Manual isolation valves outside containment are position checked every month. All isolation valve positions are checked before return to power operation and at least once per 18 months.
The upper bound frequency of pre-existing leakage derived in NUREC/CR-4220 is greater than that included in the Seabrook PSA; however, they are overly conservative estimates which are inappropriate for use in the Seabrook risk models.
b.
Regardless of the above, the ef fects of applying the NUREG/CR-4220 preexisting leakage estimates to Seabrook was evaluated; it was found that even these upper bound estimates would not result in any of the emergency planning risk criteria to be exceeded.
The effect of an assumed small preexisting leakeage, one to ten times allowable, was evaluated as follows. Assume a preexisting leak of ten times design leak rate 100% of the time.
This would increase the source terms for SS and the first part of S3 by a factor of ten.
Since the source terms for the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of S3 are greater than SS, S3 will be evaluated.
(For the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, S3 and S5 are the same leak size but the driving containment pressure is higher for S3.)
Using Tables 4-14 of PLG-0432 and 4-4 of PLG-0465, it can be seen
that the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of the release for S2C clearly envelopes ten times the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of -S3W in terms of source terms; release timing is also conservatively enveloped. Table D-1 of PLG-0432 showsi that S2C '
is an insignificant. contributor to early fatality risk and,-in fact, S2C has zero. consequences (early f atalities). _ Table 1, which follows, is a partial reproduction of the no evacuation case risk summary table for the EPZ Sensitivity Study (PLG-0465) which was provided in the response to question 23; however, the S3 and S5 releases have been replaced by S2C.
(The S2C source term represents 50 to'100 times the Seabrook maximum design Icakage rate and assumes it exists 100% of the time.
In.other words, Table 1 represents a 100% chance of five to ten times the preexisting " Technical Specification Violations" leakage predicted to occur in the NUREG 30% of the. time.) As Table 1 shows,
'this extremely conservative case still shows zero early f atalities and a small contribution to early injuries.
NUREG/CR-4220 estimates an upper bound of SE-3 for "large leakage" which is conservatively assumed to be a six inch valve (or 28 square inch hole).
W'e can conservatively bound the risk contribution of this by' assigning this frequency to release category S6W.
In Figure 1, we plot the upperbound effects of both small and large leaks in the 200- and 50-rem dose versus distance cerves based on the assignment of small leaks to S2-CM and the large leaks to S6W according to the NUREG-4220 probabilities.
While we don't agree that these results are reasonable for Seabrook, they show that the NUREG-0396 results at ten miles do not occur at Seabrook at one
. mile or less.
Hence, a conservative interpretation of the NUREG/CR-4220 experience with pre-existing holes has no impact on the conclusions of the sensitivity study.
i d
f i
O s
TABLE 1
}
P l
j FILERI#tt P9 NIP!t5 i
St40400r IP2 CeltteuGEt TAIUtsftM W-1400 soutCE 'IRR IEW17!VITY Ifulf IOURCE 79W1 atC-fl ButtflTUtil Fet SIW AND SSIAT 10 $1MULATE LA89tt LGE Apt ilEUtt IWC PHInl!LITT Intl.T IARLV fWiE0!0 OTNtt itfAL
--- Agengit Allt -.-. - PflCtWT40t R!N -
rut 49 Il CATESORY liff fatBJtuCT P.C. :MC FAfpliff !Nigtf CanCEGI CenttWI CANCEtt RUE litL.CAf FAYAL!ff lumf Canett FATALITY Ntfif DECIt 4 1WC34f104 Calil B 1440 SOLACE ttOS
--- COtfleutuctI.-.-
--- MW t!N ----- -- PtKINT F RIN --
447 81W-9 81W 0 6.0E-09 L
1 765.6 1000.7 6476.8 714.6 940.1 647 Sie 3.H-46
- 4. IE-N 3.It-46 0.1
.0
.6 453 Itu-0 IEW 9 f.It45 1
1 1M.1 354.7 til5.1 643.3 699.1 653 888 f.et43 1.88-04 !.5t-et M.t 74.I 5.t 45E
$6W-4 86W 0 6.3E47 1
1 196.3 1700.1 1681.E 1151.4 1805.3 450 Bee ~ 0.It90 1.M-03 3.H-01 0.9 0.9 38.0 189 ft-n-8 83U 0 1.6E46 1
1 0.0 f.4 576.9 1035.7 1004.1 tM 88W l
4.N46 1.1t49 f.6t46 13.7 7.3 0.3 i
434 Tat-0 37W 0 6.N-08 1
1 4.0 0.6 10.t 13.7 14.t 654 97W 0.4E40 f.6!40 0.ft-91 0.0
.0
.0 IM ttt-N-0 95Mt 0 t.10-44 L
1 0.0 9.6 576.9 1053.9 1904.7 139 95 HAT 0.At90 1.1t41 1.8841 0.6 7.0 61.6 l
. ~.. -..
~.
j TOTAL t.7E06 M'AL 3.M-03 1.H4l I.M4l 100.9 100.0 100.0
-.m
_ _ r --
i IN THE ISOYE IPRDO MET FOR M lH4M IERI!!!VITV Sfp0Y M CeuMOUENCl3 0F $0lflCE filWI SIC lf-0 FROM TE lifel CA'.CUuf!0ef i
tall gultf!TU?tt POR till CSGt0UNII O' BOURCE TERill 888 40 8384f. fHil 81EWLAfff A LEAK RAff WHICM II 3170 !M fil45 LMIER TMil 14
]
CeutatsgM M3!su gests Law pit. tytp w!fu 71sts (Amet istt4K flegit 18 W CMult JW 1NE EARLT FATAL.lTV RIK. f4Ett II A4007 A 15 I IllCREAft j
in M tem,y lit ytt n!st, ellvtVit TN!I II A 3RALL IWCAD$t COMPAAtB V5 THE Itapet !N LEM 44M. M EAftLY WURV Ritt II Sf!LL CLiegt.V 00RINATO If RELEAN CAft90t? Ilv. TB!I ii00L4 NA81 TNAT 7NE IIIPACT IN THE 50 Eli Cutti AIS M 100 RDI CIIRft liOULD N $MLL.
t 4
j
I i
i e i i i il i
e i i e isig i
i i i i,,_
NUREG-0396 m$z SENSITIVITY 8$
STUDY
~
$h
~
e ma s ma s -e m e BOUND OF THE
_g EFFECTS OF A gb CONSERVATIVE
,wW I
INTERPRETATION J2 OF NUREG/CR-Ow 4220 DATA ho 0.1 7
g u
e<
I z
az wy I
d5 T~i%
[
mus I \\
o m$
g g
\\
50 REM O
oB i
\\
!S
\\i
\\
Ow
\\-l
)
Wb 0.01 t-co M
\\1 1
Q
-... p-
\\
gw g
50
- g***,J
~
~
/
k 50 REM
-#aLli
', g/
200 REM g
g I
I 0.001 1
10 100 1,000 DISTANCE (MILES)
[cu f /
B0UNDING THE CONSERVATIVE INTERPRETATION OF NUREG/CR-4220 DATA r
PRE-EX! STING CONTAINMENT LEAKAGE
RAI 23
a.
Provide a narrative description that quantitatively delineates the dominant contributors to the dose prob-ability vs distance curves and the early fatality prob-ability curves. The dominant release categories should be specified and the dominant accident sequences contrib-uting to each of these release categories should be specified. The probability of occurence of each release category should be stated.
These data should be provided for the current study and for the original PSA results.
Changes between the two studies should be attributed to specific differences in the analysis.
b.
Provide a set of early fatality conditional probability curves for each release category, assuming evacuation distances of 1 mile and 2 miles.
c.
Provide the conditional mean risk of early f atality for each of the curves provided in b.
RESPONSE 23 The following describes the principal contributors to early health risk at Seabrook Station, as determined in the original probabilistic safety assessment (PSA) in 1983 (PLG-0300), in the PSA updates of 1985 (PLG-0432),
and in the sensitivity study (PLG-0465). The risk measures of interest here are the early fatality risk curves and the frequency of exceedance of dose and distance curves for the whole body dose of 200 rem.
The three (3) parts of this question are addressed collectively.
i ll 1.
SSPSA'RESULTS (1983) 1.1.
EARLY FATALITY RISK CURVES The only results available for 'early health risk in the Seabrook j
Station Probabilistic Safety Assessment (SSPSA) (PLG-0300) assume
.a 10-mile evacuation zone.
Significant: and costly analysis would be required to produce these results assuming evacuation distances i
of 1 and 2 miles.
The contributors to risk can be expressed in a number of different ways. Alternative ways to group-accident sequences are.to group the individual sequences by initiating '
event, by plant damage state, and by release category. A graphic display of how sequences grouped by release category contribute to j
the mean risk of early fatalities in the original PSA is shown in Figure 1, which.is taken f rom Figure 13.2-la in PLG-0300. - As seen in this figure, release category S6 (large isolation failure) makes a small contribution, and all other categories make negligible contributions that are at frequ'ency levels below 10-9 per reactor 4
year. Note that the~ mean risk curve, whose-contributions are being discussed here, is the mean of a family of curves that characterize f
uncertainty in the~ risk estimate. This f amily is shown in Figure 2, which is reproduced f rom Figure 13.1-Sa in PLG-0300. The fact that the mean curve falls well outside the median (.50) risk curve indicates large uncertainties.
These uncertainties are due to uncertainties in estimating the accident f requencies, source terms,
and site model parameters.
The conditional risk curves for each release.
category can be found in Figure 3.2-2A of the SSPSA. (See page 13.2-78 of PLG-0300.)
f A tats.; sr representation of the information in Figure 1 is provided in Table 1, which is adopted from Table' 13.2-7a in PLG-0300. -This table 1
shows that more than 99% of the mean risk curve comes from release I
category S6.
Most of the remaining contribution comes from S2.
Only t
in the extreme right hand tail, at frequencies below-10-9 per reactor year, does another category appear, S1 (early containment -rupture due I
to steam explosion, early overpressure, or external missile).
The next step in breaking down the SSPSA-risk contributors is to examine the contribution of sequences grouped by initiating event.
Because nearly all of the early health risk comes from S6 and a
- small contribution from, S2, it is more efficient to confine our
. search to these release categories.
The initiating events that
[
make significant contributions-to S6 and S2 are provided in the table below, which was adapted from Table 13.2-5a in PLG-0300. As can be seen, release category S6, which indicates early f atality risk, is, in' turn, dominated by the interf acing system loss of coolant accident -
i (LOCA) (V-sequence).
i I
I l-
Initiating Event Percent Contribution S6 l
S2 V, Interfacing LOCA 76.
95 EL, Seismic LOCA 2.
5 Others
< 1.
<1 l
l l
l Total l
100.
I 100.
I In a similar fashion, accident sequences can' be grouped with respect to plant damage states (sometimes referred to as a bin). The following is -
the plant damage state breakdown of the S6 and S2 release categories.
Plant Percent Contribution l
Damage State
- i S6 i
S2 IF 78 0
IFP 0
5 3F 21 0
3FP 0
35 7F 1
0 7FP 0
60 Others
'<1
<1 l
Total 1
100 100 In comparing the previous two tables, note that 100% of the interf acing i
systems LOCAs were modeled as IF states. Hence, of the 24% of the f
seismic contribution to S6, 21% terminated in plant damage state 3F, 2%
in IF, and 1% in 7F.
Hence, most of -the overall risk contribution comes from the interfacing system LOCA initiator and plant damage state 1F.
Essentially, all the remainder are seismically initiated sequences ending in plant damage state 3F.
Note that all the FP states, and have more
~
than 99% of the release category S2 and are dominated by similar sequences with the same release path and characteristics, namely Station Blackout with a RCP seal LOCA and failed open seal return line isolation valves.
The result is high early leakage and delayed overpressurization of the containment.
- See Table 1-2 in the Seabrook Station Risk Management and Emergency Planning Study (RMEPS) (PLG-0432) for definitions. Numbers denote containment and reactor coolant system conditions at time of reactor vessel melt-through; letters denote status of containment systems and leak paths.
F states are isolation failures or bypasses more than 3 inches in diameter; FP states are isolation failures less than 3 inches in diameter.
The final step in breaking down the early health risk is to examine specific accident sequences.
In the SSPSA, an accident sequence is a single path that can be traced through the plant event trees from the point of entry (the initiating event) to the point of termination (the plant damage state). As with other PRAs, the interf acing system LOCA was analyzed as a single sequence. That is, the event was analyzed as an initiating event and assigned directly to the most severe plant damage state considered in the study and denoted as IF.
This reflects the conservative assumption that multiple f ailures of the interf acing valves automatically result in a core melt and early large containment bypass.
All other initiating events were modeled through the plant event trees, which include more than 4.5 billion sequences counting all the initiating events, the plant damage states, and paths connecting them through the plant event trees.
Therefore, the above contributions from the
'V-sequence are from a single sequence, whereas the seismic contributions come from many sequences.
The nature of the specific sequences initiated by seismic events is next described. Of the seismically induced transients that make up 22% of the f requency of release category S6, the single sequence having the greatest frequency makes up only about one-fourth of this contribution and was analyzed as follows.
l l
Event l
Frequency Earthquake occurs (.3g) 1.1 x 10 4/ year Offsite Power Does Not Fail
.35 Solid State Protection Fails
.041 Charging Pumps Fall
.88 Containment Is Initially in Purge Mode
.10 Emergency Core Cooling System, Containment Isolation, and Containment Sprays Fail [ dependent failures resulting from loss of solid state protection system (SSPS)W l.0 Other Equipment Does Not Fail
.86 Total l
1.2 x 10 7/ year l
The remaining three-fourths of the seismic sequences in S6 are made up f rom a large number of sequences, some involving loss of of f site power and others involving f ailures of other equipment.
.~
~
In a similar fashion, the seismic contributors to release category S2 are also spread over many sequences. The single most frequent sequence in this category is a seismically induced loss of.offsite power and f ailure of both diesels due to either seismic causes or independent causes.
This sequence appears several times in the-scenario identification tables (in Section 13.2 of PLG-0300) once
' for each discrete range of ground acceleration.. The total frequency of this' sequence. summed over all values of ground acceleration is 6.9 x 10-6 per reactor year, or about 40% of the total ~ release category frequency.
In summary, the early h~ealth risk curves in the SSPSA, which were only performed for a 10-mile evacuation zone, were dominated by the interf acing LOCA sequence (about 76% contribution to the mean exceedance f requency in the risk curves). Most of.the remaining contributions come from seismically induced sequences with release
-paths _either through the purge lines (the S6 sequences) or through the.
reactor coolant pump seal return line (the S2 sequences).
1.2 DOSE VERSUS DISTANCE CURVES
.The 200 rem and 50 rem dose versus distance curves that correspond with the SSPSA results are compared with the RMEPS and sensitivity study in
~ Figure 3.
The SSPSA curves are dominated by release categories S2 and S6.
I 1
l l
l i
I
..... ~.. -. -.__. _,. _ _ _..
2.
PSA UPDATE RESULTS (RISK MANAGEMENT AND EMERGENCY PLANNING STUDY, 1985) 2.1 EARLY FATALITY RISK CURVES In the RMEPS update of the Seabrook PSA, the following changes were made that had an impact on the risk levels and the ordering of the risk contributors.
Plant Model Changes Item 1.
The single sequence int 2rf acing LOCA model was replaced by a two-event tree model, one for suction side and one for injection side residual heat removal / reactor coolant system
( RHR/RCS) interf acing valve ruptures. This led to a reduction in the f requency of plant damage state IF and the addition of three new p1'nt damage states (IFV, IFPV, and 7 FPV).
a Two plant damage states (IFPV and 7FPV) were added to model new scenarios with a submerged RHR pump seal bypass.
This, in turn, led to the introduction of a new release category, S7, which takes credit for decontamination and scrubbing in the source term de te rmination. Plant damage state IFV contains interfacing LOCAs resulting from unsubmerged piping failures.
Item 2.
A conservatism in the treatment of certain seismically initiated sequences in release category S6 (plant damage states IF, 3F, and 7F) was eliminated.
In the updated results, credit was taken for loss of instrument air to the air-operated valves
( A0Vs) in the purge lines on loss of of f site power; hence, a high probability of purge isolation valve closures in these instances. This resulted in a shift in some of the frequency of release category S6 to S2 because, when the large purge valves are assumed to close, there remain small open. lines with motor-operated valves that fail in these same sequences. There still remain some seismic sequences in S6 with the purge isolation failure. Those that remain either involve a no loss of offsite power condition or mechanical f ailure of the purge valves.
Item 3. In support of the ef fort to optimize plant technical' specifications (PLG-0431), the PSA systems modeled were revised to incorporate revision to the technical specifications and a more complete treatment of common cause failures.
This led to many minor changes to individual sequence frequencies, with the most significant change being an increase to the unavailability of the primary component cooling system. This led to a slight increase in core melt frequency from 2.3 x 10 4 to 2.7 x 10 4, most of which occurred in plant damage state 8D.
Item 4.
The updated results take credit for recovery of certain containment systems (principally the containment building spray) during core melt scenarios initiated by loss of of f site power and
involving a station blackout.
Both the original and updated results took credit for recovery of electric power prior to and in prevention of core melt.
This new recovery action results in a small shift in frequency from release category S3 (gradual containment overpressure) to SS (containment intact).
This change does not appreciably affect the results of the RMEPS or sensitivity studies since neither S3 nor SS contributes to early health risk with at least 1 mile of evacuation.
Containment Model Changes Item 5.
Uncertainties in source terms were reasssessed for all release categories with the net ef fect of a reduction in the mean source terms for all categories.
Item 6.
A new release category and three new plant states for interfacing system LOCA scenarios were added.
Item 7.
Interfacing system LOCAs resulting in unsubmerged RHR piping failures (plant state IFV) were reassigned from release category S6 to S1 (small conservative ef fect).
Site Model Changes Item 8.
Site model uncertainties were reassessed (minor effect).
Item 9.
The evacuation distance and sheltering assumptions were varied.
Item 10.
The Unit 2 construction workers were eliminated f rom the population distribution.
Although all the above changes contribute in some way to dif ferences in the updated results, the ones that had the most significant impact on early healch risk are items 1, 5, 6, and 9.
A more quantitative picture of the significance of each change is provided below.
The results in the RMEPS update for early fatality risk are presented in Tables 2, 3, and 4 for evacuation cases of no evacuation, 1-mile evacuation, and 2-mile evacuation, respectively. These are the comparison tables for Table 1 and the original SSPSA results. The conditional risk curves for each release category and evacuation distance can be found in Appendix C of PLG-0432. There are two kinds of differences exhibited in the new tables.
One is that the risk levels (exceedance-frequency values) are lower although less evacuation is assumed, and, as expected, the levels decrease as the evacuation zone is increased from 0 to 2 miles. The other difference is that several new release categories, in addition to S6 and S2, appear as making significant contributions:
S3, S7, and SI.
Release category S3 contributes only under the assumption of no evacuation. This result is viewed as purely academic because the time of release for S3 issome 89 hours0.00103 days <br />0.0247 hours <br />1.471561e-4 weeks <br />3.38645e-5 months <br /> after the initiating event, during which even ad hoc protective actions would be affective.
For 1 or 2-mile evacuation, release categories S2, S6, S7, and S1 are significant. The contribution of S2 only appears in the low consequence,
I relatively high frequency portions of the risk curve.
In comparison of these results with Table 1, the shift in the ranking of contributors is due to the following.
1.
The frequency of S6 in RMEPS is lower because of the deletion of the interf acing LOCA and some seismically initiated sequences with station blackout.
2.
The frequency of S2 increased slightly from the same seismic sequence noted in 1.
3.
Some of the old V-sequence frequency formerly categorized in S6 is now in S7.
While source terms in S7 are lower than S6, they are still great enough for potentially fatal doses.
4.
The frequency of S1 increased due to the addition of the pipe break type V-sequences formerly categorized in S6 and to a smaller extent by a reassessment of some turbine missile scenarios that was done since the RMEPS.
The contributions of plant damage states and initiating events to all updated release categories are shown in Tables 5 and 6, respectively.
Tables 7 and 8 define the codes used for initiating events.
About 78% of the scenarios in category S1 are pipe break type interfacing LOCAs that are assigned to plant damage state IFV.
The remaining scenarios in S1 include aircraft and turbine missile scenarios that f ail the containment in plant damage states IFA, 2FA, and 6FA and a wide spectrum of transient and LOCA scenarios with containment f ailure due to reactor vessel steam ex pl os io ns.
The contributors to category S2 are the same as those in the SSPSA; namely, seismically induced station blackout with a failed open small penetration.
Release category S6 is now dominated by seismically induced accident sequences with no loss of offsite power and failure of the SSPS system with an assumed containment purge in progress. No credit for operator recovery of any system or component is taken for any seismic sequence, including those that now dominate S2 and S6.
Release category S7 is composed wholly of new interf acing LOCA scenarios in which the RHR piping remains intact and the bypass occurs via a degraded and submerged RHR pump seal.
In assessing the uncertainty on the source term for S7, a 10% probability was assigned to the possibilitly that the leak path would not be submerged. From the information provided in RMEPS, i t is clear that there would be no contribution to early health risk f rom S7 if only best estimate (submerged) source terms had been used.
Similarly, had the conservative source terms not been used for the remaining release categories, the risk levels calculated in RMEPS would have been much lower than they were.
In fact, on the basis of using the best estimate source terms only, release category S1 is the only category that produced any potential for 200-rem doses and, hence, any potential for early fatalities.
2.2 DOSE VERSUS DISTANCE CURVE In the RMEPS results, there was found to be very little potential for 200-rem doses, even close to the site.
As seen in Figure 2-9 of RMEPS
(PLG-0432), the frequency of exceedance scale had to be' extended from.001 to.0001 to pick up the mean risk of exceeding the 200-rem dose shown on the curve. The median curve for the 200-rem dose was off-scale. The contributions to the mean risk at various distances are indicated in the table below.
Percent Contribution to l
Release 200-Rem Exceedance Frequency Category (Figure 2-9 in PLG-0432) l 1 mile 1.5 miles 2 miles!
l l
SI 1
3 6
S2 13 34 0
S3 67 3
0 SS 0
0 0
S6 17 50 73 S7 2
9 20 l
l By comparing these results with those in Table 2, it is seen that the 200-rem risk has the same set of release category contributors as the early fatality risk curve for no evacuation.
Category S3 dominates at I mile.
Again, this result is largely academic.
It is difficult to envision, even if no emergency plans existed, that any individual would be in a position to receive a large dose more than 3 days after the initiating event. At 2 miles, the 200-rem i
curve is dominated by S6, with smaller contributions by S7 and SI.
l Hence, the overall picture of the risk contributors is the. same for the early fatality risk curves and the 200-rem dose versus distance curves.
2.3 CONDITIONAL MEAN RISK Appendix D of PLG-0432 provides mean risk summary tables for early fatality risk as well as cancer risk.
l l
i
4 3.
SENSITIVITY STUDY UPDATE -
L In the'Seabrook Station Emergency Planning Sensitivity Study (PLG-0465',
1986), there were no changes made to the plant model.
Source terms were revised to reflect the Reactor Safety Study (WASH-1400) source term methodology; i.e., were calculated using the CORRAL computer program for the Seabrook plant configuration. These CORRAL source terms 'had been developed during the original SSPSA.
CRACIT computer program runs were made using best estimate (median) modeling assumptions, and median accident frequencies were used for consistency with NUREG-0396 and WASIl-1400. However, unlike NUREG-0396 and WASH-1400, the full treatment of dependent and external events in the Seabrook results was lef t unchanged.
3.1 EARLY FATILITY RISK CURVES The early fatality risk curves for 0, I and 2-mile evacuations are plotted in Figure 2-1 of PLG-0465. The conditional early fatality risk curves for each. release category are found in Appendix B of this
-report.
The contributors by release category are shown in the. tables below for no evacuation, 1-mile evacuation, and 2-mile evacuation.
i-RESULTS FOR NO EVACUATION Release
- Percent Contribution' to Category Early Fatality Risk Curve 1 Fatality 1 100 Fatalities 1,000 Fatalities S1
<1
<1
<1 S2 100 99 99 i
S6
.<1 1
1 Others
<1
<1
<1 1
I Total 1
100 100 100 l
RESULTS FOR l-MILE EVACUATION l
l Release Percent Contribution to Category Early Fatality Risk Curve 1 Fatality I 100 Fatalities 1,000 Fatalities S1
<1
<1
<4 S2' 99 95 0
S6 1
5 96 Others
<1
<1
<1
\\
l l
Total 100 l
100 100 4
g
- r-e----
e-
+
-c--
e---
--e
=-+----'*--v
---r+
==--r-e-ew-ev-+-w-em~vt-*---=~wm-- " -wes+
- =
~a e-w v
--'-e
FESULTS FOR 2-MILE EVACUATION Percent Contribution to Release Early Fatality Risk Curve Category l
1 Fatality 100 Fatalities 1,000 Fatalities S1 2
2
.S2 0
0 S6 98 96 8
Others
<1
< l'
\\
l Total 100 10 0 100
\\
l l
- Results below 10-9'per reactor year not shown.
By comparing these results with the RMEPS results, it can be seen that one chief difference is that S2 now has a more dominating impact than it did in RMEPS, especially for the 0 and 1-mile evacuation cases.
Category S6 dominates the low frequency tail of the 1-mile curve and completely dominates the 2-mile results. The other chief difference is that categories S3 and S7 no longer make a significant contribution to early health risk and the percent of contribution of S1 is reduced somewhat. These differences stem from the fact that application of the. WASH-1400 source term methodology did not have~ uniform impact on all the source terms. The application of this methodology appears to have increased the S2 source term more than the others.
In addition, the' RMEPS results for early health risk are heavily influenced by the conservative source terms used in that study. For category S7, the conservative RMEPS source term assumed no credit for a flooded RHR vault, while such credit was taken in the sensitivity study to make the analysis consistent with WASH-1400.
In WASH-1400, credit was taken for suppression pool scrubbing in some boiling water scenarios.
In Figure 3, the 200-rem and 50-rem dose versus distance curves are compared between NUREG-0396, the Sensitivity Study, RMEPS, and the original SSPSA.
The 200-rem curves for the latter two studies are of f scale. As can be seen f rom this. figure, the Sensitivity Study results fully bound the R!!EPS and SSPSA results.
3.2. DOSE VERSUS DISTANCE CURVES The 200-rem dose versus distance curve is fully dominated by release category S2, with very small contributions from S6 and SI.
The contri-butions of S6 and S1 occur below the level of conditional core melt frequency at which the curves are cut off in NUREG-0396 (.001).
i 1.
3.3 CONDITIONAL MEAN RISK Table 9 provides a mean risk' summary table for the sensitivity study
[
results.
Column number 5 provides the information requested in Part C of this question..The corresponding information for RMEPS can be found
'n Appendix D of RMEPS. There are no results in the SSPSA that assume either a 1-mile or 2-mile evacuation.- The results presented in Table 9 confirm that. release categories S2W and S6W completely dominate the risk at Seabrook Station if the WASH-1400 source term methodology is used to define source terms. The only additional small contribution to risk is made by release category S1W for the early fatality risk with a 2-mile evacuation distance.
Figures 4 and 5 show the decrease in the early fatality risk as a f unction of evacuation distance, comparing the mean risk results from the RMEPS study (PLG-0432).and the EPZ sensitivity study (PLG-0465),
which used WASH-1400 based source terms.. Figure 4 compares the risk reduction for the two cases on an absolute basis, and, in Figure 5, the risk reduction is normalized to the no-evacuation case for each study. The results-indicate that, for the WASH-1400 source terms
~
(PLG-0465), the acute fatality risk decreases even more rapidly in the first 2 miles than for the RMEPS baseline case.
Furthermore, the risk for all WASH-1400 source te'rm cases remains below the safety goal risk.
r d
i
. _ - - - ~ -. -.
i TABLE 1.
OF EARLY FATALITIES AS CALCULATED IN Number of Early fatalities (percent contribution of release category) 1 10 100 1,000 10,000 S6 (98.98)
S6 (98.8)
S6 (99.4)
S6(99.4)
S6 (99.5)
S2 (0.92)
S2 (1.10)
S2 (0.52)
S2 (0.49)
S1 (0.5)
Others (<.1)
Others (<.2)
Others (<.1)
Others (<.1)
Others (0) f[c*,{"*",cy of 4.60-7 3.87-7 3.14-7 1.78-7 6.26-10 NOTE:
Exponential notation is ind i.e., 4.60-7 = 4.60 x 10-7.icated in abbreviated form; t
n
I i
TAllLE 2.
CONTRIBUTIONS OF RELEASE CATEGORIES TO RISK OF EARLY FATALITIES BASED ON RMEPS UPDATE - NO EVACUATION CASE (EO)
Number of Early Fatalities (percent contribution of release category) 1 10 100 1,000 10,000 S2 (40.1)
S3 (62.7)
S3 (50.0)
S6 (67.6)
S7 (98.5)
S3 (35.4)
S6 (27.6)
S6 (32.7)
S7 (24.9)
S1 (1.5)
S6 (18.3)
S7 (6.7)
S7 (13.3)
S1 (7.5)
Others (< 7)
Others (< 3)
Others (< 4)
Others (= 0.0)
Others (= 0.0) 11p la t nl Frequency of 1.40-/
7.91-8 2.98-8 4.41-9 2.53-11 Exceedance Resu ts 4.60-7 3.87-7 3.14-7 1.78-7 6.26-10 NOTE:
Exponential notation is indicated in abbreviated form; 1.e.,1.40-7 = 1.40 x 10-7
TABLE 3.
CONTRIBUTIONS OF RELEASE CATEGORIES ~TO RISK OF EARLY FATALITIES BASED ON RMEPS UPDATE'- 1-MILE EVACUATION i
Number of Early Fatalities (percent contribution of release category) i 1
10 100 1,000 10,000 l
S2 (71.7)
S6 (65.5)
S6 (50.7)
S7 (62.4)
S7 (98.5) i S6 (18.6)
S7(26.4)
S7 (41.2) 56 (28.3) 51 (1.5)
S7 (7.4)
S1 (8.1)
S1 (8.1)
S1 (9.3)
Others (< 3)
Others (=.0.0)
Others (= 0.0)
Others (= 0)
Others'(= 0) i U pila t il Frequency of 7.44-8 1.64-8 8.25-9 1.59-9 2.53-11 Exceedance 1
i Resu ts 4.60-7 3.87-7 3.14-7 1.78'-7 6.26-10 NOTE:
Exponential notation is indicated in abbreviated form; i.e..
7.44-8 = 7 44 x 10-8 I
I i
t l
TABLE 4.
i CONTRIBUTIONS OF RELEASE CATEGORIES TO RISK OF BASE 0 ON RMEPS UPDATE2-MILE EVACUATION ZDNE (E2)
Number of Early Fatalities i
(percent contribution of release category) j 1
1 10 100 1,000 10,000 S2 (46.0)
S7 (55.2)
S7-(65.3)
S6 (72.9)
S7 (43.3)
S6 (33.3)
S6 (23.9)
S7 (19.1)
S1 (10.7)
S1 (11.5)
S1 (10.8)
S1 (8.0) 1 Others (= 0)
Others (= 0)
Others (= 0)
Others (= 0)
Upla L. il Frequency of 4.96-9 3.07-9 1.17-9 2.28-10
{
Exceedance
\\
0.0 Resu ts 4.60-7 3.87-7 3.14-7 1.78-7 i
1 6.26-10 f
NOTE:
Exponential notation is indicated in abbreviated form; 1 e j
.. 4.96-9 = 4.96 x 10-9 I
e
j
.i I
i l
e i
TABLE 5.
MAKING MAJOR CONTRIBUTIONS TO RISK OF l
RMEPS UPDATE RESULTS l
i
{
Major Risk and Core Melt frequency Contributing Release Categories (percent contribution of plant damage states to release categories)
S1 S2 S3 SS S6 S7 IFV(77.5)l 7FP (49.5) 80 (72.0) 8A (82.1) 3F (92.5)
IFPV (69.5)
IFA (7.3) 3FP (43.8)
FD (14.6) 4A (15.0) 7F (7.3) 7FPV (31.5) 8A (5.0)
IFP (6.6) 3D (11.3) 2A (1.6) 2FA (3.7) bfA (1.3)
Others (5.2)
Others (< 1)
Others (< 3).
Others (< 2)
Others (< 1)
Others (0.0) t 1
l Release Category 6.00-9 2.02-5 1.43-4 1.17-4 3.00-7.
- Frequency 3.93-8 NOTE:
Exponential notation is indicated in abbreviated form; i.e.
, 6.00-9 = 6.00 x 10-9 l
i
~
I. I.
- - =
=.
- ~ e
~e e ~
S
- O O *" 4J- - =,"
J4 g
y
- 2.,. =* 3 t 5 W h
a w
3 o
m m.
s a m
- 4 3
e g
3004 : 43 H
J
" g.4 *.4 "a a
T 3
9 k
3
.e M== M y
e
- w e
e QC W
2 2 2.E.l.l:
oe
>= d g
k g
".. w w w a
=
S D.=X :3 w w wW
- s. A _
=.
o w
- ! 4 : O 2
ROO 4 00
$32N.5 ' E.
)
d.
2 w-H w
w o
as (Mbg 2355.5.55355*
e-.
.m E 3
-, ~. - g
- v. 2 W3 : =. =.~.~. =,.
m
= wj
=.
gg
-a w <
w w w w W e W
E2 k-2
- a. o ass
~a U
O >.
3 m
O "hO" E2g My 5
23
- y y
JJ JJ2.
h 5
C o
4 M
Wg yw
- 2. I.
.. 2 i
e ZK J2 4 4 '.
l w u.
g m
g3
=
g=
l CH 7 *
.J S
7 w _J g
I Mg
. w w w w a
~
w 2
KW 5
5 b 5 5,5, w as
- e. '
QO
.I O
1." 2. :. 2
~ " "
O T
8 v" a 3
2 2
og 3
a w w w a
4
- C O E
t l
- u. g a a om 1
=
m 3
OO m a::
=
a
^:
2 a
=*
om g
-o h
- I 4
an si
=
n e
zr
. u
> = > =
~
5 %
g-
~
e.
.=
3 m 3 a
o m
w
=
3 3.t
=
- t-.
t "O
T.
4
@ et*
5"' k 2
2 e
4 wm 3
y 3
cc w 4c 8
=
42 m
I
~
=
~
1,, :-
E D
- C 3
w a
5
- I a.
?2
=
- 7
.%.f.4 s p m%
.": k 47 a
4 2o:241 23 J J
TABLE 7.
BINNING OF INITIATING EVENTS THAT HAVE New' Initiating Events 8 inned SSPSA Initiating Events Title Frequency (events / year)
Title Frequency (events / year)
EXTAC 2.70-6 FSRAC 5.19-7 FCRAC 2.10-6 FL2SG 8.50-8 EXTLP 1.20-3 FTBLP 6.00-4 FLLP 3.20-4 TCTL 2.76-4 EXTCR 5.43-7 TMCR 3.98-7 MCR 5.80-9 ACR 1.39-7 TLPCC 1.82-5 LPCC 1.39-6 FSRCC 3.60-6 FCRCC 9.00-6 FPCC 4.20-6 l
TMPCC 1.27-8 MPCC 5.46-9 TLSW 6.22-6 LOSW 2.52-6 FCRSW 2.10-6 FLSW 1.60-6 TLCV 4.18-1 LCV 4.18-1 TMLCV 8.30-5 l
NOTE:
Exponential notation is igdicated in abbreviated form; i.e., 2.70-6 = 2.70 x 10.
O
TABLE 8.
INITIATING' EVENT CATEGORIES SELECTED FOR QUANTIFICATION OF THE SEABROOK STATION RISK MODEL FO
~
Sheet 1 of 2 Group Initiating Event Categories Selected
-Code for Separate Quantification Designator e Loss of Coolant 1.
Excessive LOCA ELOCA Inventory 2
Large LOCA LLOCA 3.
Medium LOCA 4.
Small LOCA MLOCA(a).
SLOCA 5.
Interfacing Systems LOCA V
6.
Steam Generator Tube Rupture SGTR(a) e General 7
~. Turbine ~ Trip TT(b )
9.
Total Loss of Main Feedwater TLMFW((c)
'10 Partial Loss of Main Feedwater PLMFW c)
- 11. Excessive Feedwater Flow EXFW(b)
- 12. Loss of Condenser Vacuum LCV(b) 13 Closure of One Main Steam IMSIV(b) i Isolation Valve (MSIV) 14 Closure of All MSIVs AMSIV
- 15. Core Power Excursion CPEXC
- 16. Loss of Primary Flow LOPF(b) i
- 17. Steam Line Break Inside Containment SLBI
- 18. Steam Line Break Outside Containment SLB0
- 19. Main Steam Relief Valve Opening '
MSRV 20.
Inadvertent Safety Injection SI e Common Cause Initiating Events
- Support
- 21. Loss of Offsite Power LOSP(d)
System Faults
- 22. Loss of One DC Bus L10C
- 23. Total Loss of Service Wa:er LOSW
- 24. Total loss of Component Cooling LPCC Water
- Seismic
- 25. 0.7g Seismic LOCA E.7L Events
- 26. 1.0g Seismic LOCA El.0L
- 27. 0.29 Seismic Loss of Offsite Power E.2T(e)
- 28. 0.39 Seismic Loss of Offsite Power E.3T(e)
- 29. 0.49 Seismic Loss of Offsite Power E.4T(e) l Transient without scram scenarios are represented by a separate code, ASLOC.
a.
Transient without scram scenarios are represented by a separate code, ATT.
b.
I Transient without scram scenarios are represented by a separate code, ALOMF.
c.
Transient without scram scenarios are represented by a separate code, ALOSP.
d.
Transient without scram scenarios are represented by a separate code, ExA, e.
x =.2,.3,
.4,.5,.7, 1.0.
t
[
(.
m
TA8LE 8 (continued)
Sheet 2 of 2 Group Initiating Event Categories Selected for Separate Quantification Code Designator
- 30. 0.5g Seismic Loss of Offsite Power E.5T(8)
- 31. 0.7g Seismic Loss of Offsite Power
- 32. 1.0g Seismic Loss of Offsite Power E.7T(*)
- Fires 33.
Cable Spreading Room - PCC Loss El.0T(*)
FSRCC
- 34. Cable Spreading Room - AC Power loss FSRAC 35.
Control Room - PCC Loss Control Room - Service Water Loss FCRCC 36 FCRSW 37 Control Room - AC Power Loss
- 38. Electrical Tunnel 1 FCRAC
- 39. Electrical Tunnel 3 FET1 40'7 PCC Area FET3 FPCC 41 Turbine Building - Loss of Offsite Power FTBLP
- Turbine 42 Steam Line Break Missile 43 Large LOCA
'TMSL8 TMLL 44 Loss of Condenser Vacuum TMLCV 45 Control Room Impact TMCR
- 46. Condensate Storage Tank Impact Loss of PCC TMCST 47.
TMPCC
- Tornado
- 48. Loss of Offsite Power and One Missile Diesel Generator MELF
- 49. Loss of PCC MPCC
- 50. Control Room Impact MCR
- Aircraft
$1. Containment Impact Crash
- 52. Control Room Impact APC ACR
- 53. Primary Auxiliary Build:ng Impact APA8
- Flooding 54 Loss of Offsite Power l
FLLP 55 Loss of Offsite Power aad One-Switchgear Room FLISG
$6. Loss of Offsite Power r 1 l
Two Switchgear Rooms FL2SG
- 57. Loss of Offsite Power 2 Service Water Pumps FLSW -
- Others
- 58. Truck Crash into Transmission Lines TCTL
Transient without scram scenarios are represented c.
,ATT.
d.
Transient without scram scenarios are represented by a separate code, e.
x =.2,.3,
.4,.5,.7, 1.0.
i
{
\\
~
9 eiEa1
! O. P. 4. M. O. a. ' O. 9 I l l O. o. 4 4 0. *.
O. e I O. n. P. 4. O. =. gI O. 8 ll 8 8 8 E 8 0 4. O. a e Ws e
e s
a e
We eX eY mon O OeY MOM O OeM NO4 04Oe IY IO44e e
e in P
e.m.
Os
= e.En.
Oa
= a.in.
8 O. e e in iOP4e e=
1 e
e.* e e
eguaEi eg ag e
a egu4 e
a e
I e
i e
t B
a e
a y
eW e la i e IL e IL E
I I ed 0
go eO eO sO eO e
I
! ES e
og aC e
a e
a e
I
'4 e
a U-e>
eH i
e e-e>
e g
a eZ>sZ
=. fu. O. n. O. O.' O. e Z
- n. n. o.. O. O.
O. e 2 CD. O. O. N.O. O.,' O. e W Z>eO.
e.u
>.W'"'"
aWPeW eWi eW e
.i W H t
. e su-au 040n00' OuuIODOMOO Oeu
' 4 0 0 t9 0 0 li O a au=l EW eEJeE Ot e
O..B a Ov eu4eW I.O. a E I
(9 4
0eE P
80a IEJ t
s eW' i
~
eW s-e FIW4 t
CD o e L e= a 0.
i eA1.
e G. '
e I
e i.n. i dL t-w e e 4e a
t a
3 3
u a
I C
i e
po aa4st ea i
e 1 I
.e Q8 I ta, a
a e
a
<a va
- 5' e
s e
e a
w
.iea 4,undon nu e,
4,u n t o n, m.
4mn#nn um.
, i e
, m m iu.
M >-
e 8. v.s 0 0 0 0 0. o, O.
O. O. C. O O O.,
O..e,
W e
l 000000 ll O. e 38,
- o. 0 0
<a 8W s
e e
y "H'
55:'
e WWM n - 8 8. W. ' s. e: x 8 8. M. 8.W:'
n - P 8. W.
8 8.N 8 n o.
- 8.. M.l 8..: W:5<, 8. n. 8. e:
mmm
.s usu
..x 3,W.
u u, zw
.. n.
-en
- e.n.
4.
.e m
aW..
.e ae a
ZH<
a>
a i
e-a i
e
..W O~U e3 eZ i
eZ e2 a
3 e a
a eJ>s4 4f90400
' fi e 4 i 440400
. 4s4 4 0 0 Fl O O l na V J>IM#na wwz aOp a
e n -.e W i, 0 0 0 0 0 0 o.at 000000 c
O..eE 000000 O. es En 00 I B
in-.000 Ha i
)
E-~
l t.d :.
' 888888
- 8. :
888.W.88 N..:
' 888.".88.is.:
e J. i d ! 8. W. W. :
t a
e
. e EZ4 eICa ti fu O 4 O O e a 46 e 19 e M tu O t1 O O ne e
. MOO 4OO
- h. W D i 4e 4
l ICIMn4E e
e ei e >
l 3 B6 e e
MO e
s i
I e
a e
O i
e e
O O >C Z
=:
e e
i e
a e >
i s
e a I e I a 1 1
e a
w z u.3 e J tn a s
=. n. n. fu. a.
a
, =.. n. e. m. a e I
- 4. n. n. N. =. s e
I uuue pw aqga I,i.
e e i
ag g
a 3
444e ogy a e-w e 1
B O&O440 s
000440e a
e gg g g 9 O W B En U. O P O n.4 0 4&
Q
- 5 (n 4 e1 4
a En 44 4
a W Wtaab m 4,,,,,,
e t-Z N Lai h P4 f.u.
eW 0* 4 fu.
eW P4 I
8 e
4eU l
eU e
E.3E eU 3
e O =. e a ueZ e2o aZ e
e ZEEa H CL e a
e e W.i aW s
e e
Z N
O B
I E D 1
e3 I
I O
8 8 N8 O >-
e - e aO sO I
e
' 4 4 0 0.* 0 0 voc e>>aw
- 4. =. O. t9 0 0 eW 4 0 0 19 0 0 3 eW i
e gg e > a in a cn oyg I
weZ DNO4CO eZ MPO-OO,
a in e
eZ nOOMOct a
e Js0 em m so a
4 so 4
4 e a g
gm gwsgu.n-o gu o
n go o
W -.
ei
>- x a
- u..
ei B.
zw a a
eae a, e a ei e
o e
a e
z e
e a
e e
e a
.n#n 4
ts entn 4
de enen.4
.#a i
p a ua ooooOO Oe
!oooOoO c, e ooOOoO oe a Ze es e e e e ee i a e e e e e as
- e e e eee t
e i
e e
W tal W W W Lal Wa W W talW W tal WB 4
e a
O. *. e. n. M. a.tal W tai td W laf I' n e laf E O. =. a. n. M. =.
- n. e O. =. 4. o. n. a.
- n. u MH e
e i
s e
3 e
MM e
W5 4 fu.* 4 4 g
.g w
e Ee
! fu e 4(U 44a tu e 4 fu 4 4 4 Si e a
e a
e mM th B e in u
e in 4
e o
a a
sW s
eW e
e o
's e
e in a tn 3
e e
a e4 s4 i e og e U 0- e eu aui e
O CO s q in. a En OOOOOO i
e
- i. **=-
i a
- i. tutu m m m o e
s a:C e'>.
sZ sZ s
i 1.aJ W eWQeWt e.o.
s.O.
a i
a JM e
e an e u
i e
e to e
e4e e>
aH e
1 a:g t,aJ B
eu4 e4 eC e
p BW>u a
a3 as e
i e in E. a Z e i
eu e
au e
e c 0 e.0,.e 3 3 3 3 3 4 ed 23333c a4 23333 e
8 woe g.
tum4nI e>
tu m 4 o I e>
- gu (9 4 ts e
i a Jtage > g ingnangnEnen eW (nEngnEnin n i aW tn gn in gn in e
eW>aqs an 8
(41 i
e in i a
aE4e3 t-BW aW 4
e e ueu e
a e>
e.J e.J e
e e41 e==
a Jer JeI ae e
eW 48 4e la e
Ze i is gi si tu 4 keW f 40o4
>eO n 0* n ID #
> e i
e 3eO i 4 en o n n e O e Z e 4 n ei n n u Oe3 t 4 r3 n n yl O8 e Ea2 44444 eeaOI44444 1 eeee44444 0>e 2
104 I
l 1
g 4
10 e
d>
e 10-5 c
m 6
s h,10-6 W<
" "h_ _ _
ALL RELEASE CATEGORIES u
h m
O
%s
@ to-7
'^
N S6
\\
E
\\
S
\\
f
\\
i
\\\\
10-8 _
\\
\\
I
\\
S2
\\
\\
_.\\
l 10-8 I
1 100 101 2
10 103 4
10 NUMBER OF EARLY FATALITIES FIGURE 1.
CONTRIBUTION OF RELEASE CATEGORIES TO RIS
'EARLY FATALITIES (MEAN VALUES) AS CALC rt
.= x 10*3 -
g g
g.
g 4
10 w
il:a Ej _
10-5 ai w
o>
=8 W<
@g 10'O Um 0.95
.5E gg
- e..........,,**
0.90 g$
3 2 w w io-7 0.70 w
E MEAN RISK CURVE 10A-0.30
~
0.10 9
I l%
I I
0 I
2 3
4 0
10 10 10 j0
)0 10 EARLY FATALITIES FIGURE 2.
RISK OF EARLY FATALITIES WITH UNCERTAINTIE FROM SSPSA (PLG-0300, 1983)
1 i
i e
iiiiig i
i i i i i sig i
i i i i i..
~
m NUREG 0396 Ez
SENSITIVITY STUDY go
=======8-SSPSA SO REM (200 REM ~
k g
OFFSCALE)
.g RueeS ReSutTS 50 neu "j 2 (200 REM OFFSCALE) om 58
=
o c<
z5z
~
N Mai
.\\
- u. o 50 REM 200 Og
\\
~
$Q
\\
k OW
\\
n
";o I
)
W 9:
0.01 E
r
\\
\\
l 1
9m l
tg 9 12 l
\\
oo i
I oo 1
I L
=.k,,(
200 REM
'I \\'
'I
' ' ' ' I 0.001 1
10 100 1,000 DISTANCE (MILES) l i
fi6cce 3 COMPARSION OF SSPSA, SENSITIVITY STUDY, AND RMEPS WITH NUREG-0396-200 REM AND 50-REM WHOLE BODY DOSE PLOTS FOR N0 IMMEDIATE PROTECTIVE ACE 0NS 1
i
4 10 g
g g
NRC SAFETY GOAL FOR INDIVIDUAL RISK MULTIPLIED By POPULATION WITHIN 1 MILE OF SITE BOUNDARY 10-3 EPZ SENSITIVITY STUDY (PLG - 0465)
.W-1400 SOURCE TERMS kw 10~4 er P
3 h
5*
8
< 10-5 k
i RMEPS (PLG - 0432)
BEST ESTIMATE MEAN RISK 10-6 _
REDUCTION IN 2-MILE EVACUATION RISK WITH i
SHEETERING TO 10 MILES lx 10~7 I
I O
2 4
6 8
10 EVACUATION DISTANCE (MILES)
FIGURE 4 ACUTE FATALITY RISK AS A FUNCTION OF PROTECTIVE ACTION (From RMEPS; PLG-0432)
~
100 I
80 da Q 60 RM EPS (PLG - 0432) g BEST ESTIMATE MEAN RISK s-m O
40 5
O EPZ SENSITIVITY STUDY a-(PLG - 0465)
W 1400 SOURCE TERMS 20 -
9 I
0 0
1 2
4 6
8 10 EVACUATION DISTANCE (MILES)
^;
FIGURE 5 RISK REDUCTION AS A FUNCTION OF EVACUATION DISTANCE AT SEABROOK STATION O
y y
RAI 24
Provide a quantitative description of the ef fects of the following dif ferences between the original PSA and the current study:
- a. ' reduction in probability of core-melt V sequences
- b. factor of 1000 scrubbing of releases through RHR seals
The. effects should be described in terms of dif ferences in risk curves for early fatalities and for 200 rem vs distance.
RESPONSE 24
Reference:
Response 23 Re: 200 REM-and early fatalities
. In response to part a) of question 24, the following highlights the key factors that result in a major reduction in risk levels for the core melt V-sequence in the updated Seabrook probabilistic safety assessment (PSA) results [per the Risk Management and Emergency Planning St'udy (RMEPS), PLG-0432, 19855 in comparison with the Seabrock Station Probabilistic Safety Assessment (SSPSA) (PLG-0300,1983). Qualitatively, the key differences fall into three main areas of the analysis:
initiating
+
event frequency.
The response to 24 b. will be provided in the response to 30.
plant response to various types of interf acing loss of coolant accident (LOCA) scenarios, and operator actions to prevent core melt and isolate the bypass.
Description of each of these areas follows.
~
1.
INITIATING EVENT FREQUENCY The initiating event frequency model in the SSPSA conMdered four residual heat removal (RHR) cold leg injection paths, each having two series check valves, and two RHR hot leg suction paths, each having two series motor-operated valves (MOV).
The check valve model considered successive, independent rupturas of, first, the inboard and, second, the outboard check valve.
The second failure was assumed to occur at the same rate as the first at any random time between the first failure and the next test (refueling). The MOV model included a similar sequence of ruptures (inboard, then outboard), as well' as the possibility that the outboard valve is already open when the inboard valve fails. The updated model included all the above failure modes plus several more. For the c
check valve failures, it was conservatively assumed that the outboard + inboard sequence could also occur and that it would occur at the same rate as the inboard + outboard sequence.
In addition, the possibility of instantaneous failure of the second valve in a sequence at the time the first valve failed was also considered. Hence, the model used in the update is more complete. The net effect of those model enhancements is worth about a factor of 2 increase in the frequency of a given leak size.
A second difference between the two studies was the definition of the initiating event.
In the SSPSA, the V-sequence initiator was defined as a, major rupture leading to RHR overpressurization.
In the RMEPS, any
.ructure with a leak flow exceeding 150 gpm (capacity of one charging pump) was considered an initiator. Such flows are not capable of overpressurizing the RHR system when the RHR relief valves operate properly.
A third difference in the initiating event frequency was in the treatment l
of check valve data. The SSPSA used check valve rupture data--actually l
zero failures in a large sample of component hours per population--taken from the Indian Point 2 and 3 PSAs.
In RMEPS, a different approach was.
based on a frequency-magnitude correlation of nuclear grade RHR and reactor coolant system (RCS) check valve experience in U.S. pressurized water reactors (PWR).
These data are documented in RMEPS and in a separate submittal.
To put the corresponding analyses on a common footing, the frequencies of RHR overpressurization events can be compared as follows:
SSPSA:
1.8 x 10-6/ reactor year.
RMEPS Update:
7.1 x 10 / reactor year (leak > 1,800 gpm).
Thus, the net effect of the model differences (which have an increasing effect in the update) and the data treatment (which has a decreasing effect in the update) is a reduction in the frequency of valve ruptures leading to RHR pressurization by a factor of 2 to 3.
Hence, if no other changes would have been made to this analysis, the V-sequence risk contribution (and its early release frequency contribution) would have decreased by this same factor.
FL
PLANT RESPONSE The plant response to RHR interfacing valve ruptures in the SSP 5A and in most previously published probabilistic risk assessments (PRA) on PWRs has been treated rather simply, according to the following assumptions, without consideration of their incremental probability.
Valve ruptures produce a shock wave with peak dynamic pressures e
significantly greater than the RCS pressure that travels down the low pressure RHR piping.
RHR piping ruptures outside the containment.
e RCS and refueling water storage tank (RWST) inventories leak outside o
the containment via a piping break. '
Core melt occurs with unsubmerged bypass.
s No credit is taken for any operator actions.
e Therefore, the plant response to the V-sequence in the SSPSA was treated as a single sequence.
It was assigned to plant damage state 1F, which, in turn, was assigned to release category S6.
In the RMEPS update, a large number of alternative scenarios were identified to provide a more complete picture of plant response. The most important variables introduced in the update to consider alternative plant responses are the size of the valve ruptures that initiate the event, the response of the RHR relief valves inside the containment, the pressure capacity of RHR low pressure piping, the response of RHR pump seals to overpressure, and the configuration of the RHR pump vaults with regard to source term implications.
The major differences in plant response to RHR interfacing valve ruptures, as modeled in the SSPSA and the RMEPS update, are illustrated in Figure 1.
This figure is a highly simplified version of the event sequence model that was developed and quantified in the RMEPS update.
The chief differences in the update in this regard are a lower frequency of unsubmerged pipe rupture-type bypasses because of the high capacity of the RHR piping. A more likely outcome is a submerged bypass via the RHR pump seals.
l l
l
OPERATOR RESPONSE Because of a different treatment of hardware and plant response, the potential for operator actions to mitigate the effects of the interfacing valve ruptures was appropriately considered in the update. The two key actions, which are illustrated in Figure 1, are those to prevent melt and to isolate the bypass.
If the RHR piping remains intact, there is a high chance, as assesed in RMEPS, that the operators would prevent core melt whether or not the bypass was isolated. The key is to diagnose the bypass at the RHR pump seal and to provide long-term makeup of coolant to the RWST.
If this action is not successful, there is some chance that the operator can isolate the bypass, but only for the discharge check valve rupture case (VI).
The net effects of the major update factors are summarized in Table 1.
1 f
TABLE 1.
IMPACT OF KEY FACTORS IN UPDATED V-SEQUENCE ANALYSIS Factor SSPSA RMEPS 1.8 x 10-6 7.1 x 10-7 Frequency of RHR System Overpressurization g,,c(,rYear Reac
- r. Year Percent of Overpressurization -
Events that End with:
e No Core Melt 0
93 e Melt with No Bypass 0
~1 Melt with RHR Seal Submerged 0
~5 e
Bypass Melt with Unsubmerged RHR Pipe 100
~1 e
Rupture Bypass I
i
e.
FACING t
VALVE YES L E AK WITHIN VALVE I
LEAKAGE r'
CHARGING W IMME ATE D
- tw Gru CAPACITY OFETY L
$sGNIF 8 CSSPSA:
DlONOT ASSESS T l
RME PS: 7.7 X lt#/r vt f f RHR$YSTEM )
VALVE LE AKAGE YES PRESSURE RISE LOCA
$ 1,000 GPM,
VIA R T.
r LlulTEO RF SOTH
(
RV LIFT
}
p,g NO
'[
SSPSA: I S X 10-6/yr T
( RMEPS: 7.1a 10-7/vr j SSPSA DID NOT ASSESS T 3p RMEPS: 7 0 X 10-8/yr j RHR SYSTE M PRESSURl2ATION (PE AKS AT 2,250 psea)
SSPSA: DlO NOT ASSESS T RMEPS: 4 8 X 10-7/vr j if RHR P PING RE MAINS YES
- m T S YES INTACT TO PREVEN T 7'
SUCCESSFUL 7
MELT TE RMINATION NO NO SSPSA: 18 X 10-6/yr )
RMEPS. 4 6 X 1&9 r j
/y u
CORE ME LT WITH OPE RATOR UNSUBME RGE D ACTIONS TO YES CORE ISOLATE y
MELT WITH
$5 OR $3 syr*%
SYPASS NO SYPASS NO S1 (56 IN SSPSA)
($$PSA DID NOT ASSESS \\
RMEPS 3 7 x to 8/vs j
[ SSPSA: Ol') NOT ASSESS 3
(
RMEPS: 6 X 10-8/yr j
P CDHL Mt i I Wille butlMt H(st D HHHbEAt BYPASS
$1 FIGURE 1.
SIMPLIFIED EVENT SEQUENCE DIAGRAM FOR V-SEQUENCE P
RAI 25
Provide a list of all paths for loss of RCS inventory outside containment.
Show how these have been considered with respect to LOCA and with respect to containment bypass for radioactive materials following core damage.
RESPONSE 25 Loss of RCS inventory outside containment could occur via a flow path which directly links the RCS to systems outside containment or in conjunction with the secondary side of the plant assuming steam generator tube rupture has occurred.
In either case, a specific containment penetration can be identified as being associated with that flow path.
The Containment Penetration Analysis (PLG-300, Table D.13-3) in the SSPSA accounts for all sne_h penetrations and either provides or references a description of the equipment failure or maloperation necessary to bypass containment. Penetrations associated with flow paths which would be operating during an accident are also listed. Penetrations which are associated with flow paths that could line the RCS directly.
with systems outside containment are:
X-9, 10, 11, 12, 13, 24 through 31, 33, 35, and 37.
Penetrations which are associated with flow paths that could link the RCS with systems outside containment assuming a steam generator tube rupture are:
X-1 through 8, 63 through 66.
An evaluation of LOCA outside containment was performed for the SSPSA, examining each line which communicated with the RCS and penetrated the con t ai nme nt.
Based on that evaluation, six lines were considered in detail - four RHR cold leg injection lines and two hot leg suction lines.
(See SSPSA 56.6.3.2.1).
These are the classic "V-sequence" lines first discussed in WASH-1400.
This evaluation of LOCA outside containment was not documented explicitly in.the SSPSA but can be reconstructed f rom the SSPSA 5D.13.
Table D.13-3 contains a list of all the containment penetrations with the related isolation valves and affected system.
In order to be of interest in the evaluation of LOCA outside containment, the line must not only penetrate the containment to the atmosphere but also communicate with the RCS.
Each penetration '
discussed below.
Reference is made to Table D.13-3 in SSPSA, PLC-300.
LOCA OUTSIDE CONTAINMENT Penetrations Containment Quantified Penetration System in SSPSA Comments X-1, X-2, Main Steam Yes Quantified in 55.3.11, X-3, X-4 Steam Generator Tube Rupture with secondary side leak to atmosphere.
X-5, X-6 Main Feedwater No See Table D.13-3 X-7, X-8
" Comments" X-9, X-10 RCS/ RHR Yes Quantified in 56.6.3.2.1, RHR hot leg suction.
X-II, X-12 RRR Yes Quantified in 56.6.3.2.1, RHR cold leg injection.
X-13 RHR No Discussed in 56.6.3.2.1, RHR hot leg injection.
X-14, X-15 Containment No No direct communication Building Spray with the RCS.
X-16, X-18 Containment No No direct communication Online Purge with the RCS.
X-17 Equipment Vent No No direct communication System with the RCS.
X-20, X-21, PCCW No No direct communication X-22, X-2 3 with the RCS.
X-24, X-25, Safety Injection No Each line has at least X-26, X-2 7 two check valves and one normally closed MOV in series.
X-28, X-29, CVCS No See Table D.13-3 X-30, X-31,
" Comments.
X-33 X-32, X-34 Floor and No No direct communication Equipment Drain with the RCS.
X-35 SI No 3/4" test line, has two normally closed A0Vs in series.
RCS No 1/2" sampling lines X-36 Deminieralized No No direct communication Water with the RCS.
m Penetrations Contai nment
. Quantified Penetration System in SSPSA,
Comments 9-Nitrogen Gas No One check valve and four normally closed valves.
Reactor Makeup No Two check valves and two Water normally closed valves.
i X-37 CVCS No See Table D.13-3
" Comments".
X-38 Combustible Gas No No direct communication Control Fire with the RCS.
Protection X-39 Spent Fuel Pool No See Table D.13-3 Cooling and Cleanup
" Comments".
i X-40 Nitrogen Gas No No direct communication with the RCS.
j RCS Sampling No No direct communication with the RCS.
)-
X-4 3, X-47 RCS, S1 No See Table D.13-3 4
-X-50
" Comments".
l X-52 Post Accident No See Table D.13-3 l
Monitoring
" Comments".
i i
X-57; SS, S1 No See Table D.13-3
" Comments".
X-60 CBS No See Table D.13-3 i
" Comments".
j X-63, X-64, S/G Blowdown Yes Quantified in $5.3.11.4 X-65, X-66 Steam Generator Tube
~
Rupture with secondary i
side leak.
See top event IV (p.5.3-95).
j X-67 Service Air No No direct communication i
with the RCS.
X-71, X-72 Combustible Gas No No direct communication Control with the RCS.
ilVAC-1, Containment Air No No direct communication HVAC-2 Purge with the RCS.
i 4
.e Nominal Diameter Penetration #
Isolation Valve Inches Comments
.X-72 CGC-V-10 1"
Manual locked closed valve that isolated Train A H2 analyzer input line.
X-72 CGC-V-3 1"
Manual locked closed valve that isolated Train A H2 analyzer return line.
X-39 SF-V-86, 87 2"
Manual locked closed valves.
Would only be utilized in conj unction with the refueling canal skimmer pump.
X-67 SA-V-229, 1042 2"
Manual locked closed valve outside containment, normally closed valve inside containment.
Note: Test connections were not considered in this analysis. Test connections are 3/4" with normally closed manual valves and a pipe cap.
Per the Seabrook Station Technical Specifications there are verified closed every 31 days.
RAI 26
Indicate the extent to which the effect of local deflagration / detonation of hydrogen gas concentration is localized areas both inside and outside the containment has been considered in the assessment of risk.
Include a discussion of how weak areas of containment have been considered in your assessment, for example, the containment is considerably weaker in its resistance to pressure loading from outside the containment.
RESPONSE 26 A separate probabilistic analysis of the effects of hydrogen combustion has been performed for each plant damage state. The analysis accounted for uncertainties'in the hydrogen generation, release from the primary system, ignition, and containment atmospheric conditions. Vessel breach discharge burns and global burns at different times in the accident progression were treated separtely. All hydrogen burns were treated as instantaneous adiaba-tic combustion events. The analysis is documented in Section 11.5.2 and in Tables 1.7-1 and 11.7-7 of the SSPSA (PLG-0300). Local hydrogen deflagra-tions or detonations were considered and dismissed as requiring conditions of nearly stagnant or quiescent atmospheres which are not considered cre-dible under accident conditions. In a large dry containment thermal and mass transfer induced mixing of the containment atmosphere under accident conditions is considered assured particularly on those accident phases where rapid releases of hydrogen into the containment are possible such as at vessel breach.
Weak areas in the containment with respect to the capability to contain hydrogen burns could not be identified. A maximum adiabatic post burn pre-sure of 128 psia was determined for a limiting vessel breach discharge burn.
The lowest containment failure pressure was identified'at 181 psia for the type A failure (0.5 square inch leak area).
Hydrogen burns outside the containment in the annulus between the primary and secondary containment could be postulated.
Such hydrogen burns would have no. impact.on the calculated risk or conclusions with respect to emergency planning requirements for two reasons:
1.
In order for hydrogen to accumulate to a flammable mixture in the annu-lus region, the hydrogen must be released from the containment and this requires that the containment is already failed. Furthermore, the con-centration in the annulus would be lower than in the containment due to the additional mixing with the annulus air.
2.
A hydrogen burn in t he annulus would impose an equal load on the enclo-sure building and on the containment building. Even though the contain-ment may be weaker for external loads than for internal loads, the external load capacity is definitely much greater than the internal load capacity of the contatnment enclosure building which was evaluated in the SSPSA Appendix H.1, Section 6.
The weak elements in the enclosure pressure boundary are identified as the HEPA filters and the sheet metal duct work, each of which is expected to fail at a pressure between 1 and 2 psid.
RAI 28
Identify any penetrations connected directly into the containment atmospheric which rely on any remote manual or manual valves for isolation.
Response 28 Nominal Diameter Per.etration #
Isolation Valve Inches Comments X-14 CBS-V-1 1 8"
To CBS spray rings. Normal X-15 CBS-V-17 8"
closed motor operated valves open on Hi-3 containment pressure or manual. Valve closure is manual.
X-60 CBS-V-14 16" Containment sump return line MOV.
X-61 CBS-V-8 16" Normally closed. Valves open automatically on ECCS/CBS recircu-lation signal indicated by 2/4 Lo-Lo level in the RWST. Valve closure in manual.
X-38 CGC-V-43 2"
Compressed air line to containment CGC-V-44 2"
for pressure testing. Manual locked CGC-V-45 10" closed valves.
Line is used only for containment pressure testing.
X-38 FP-V-592 4"
Manual locked closed valve outside containment, normally closed valve inside containment.
X-71 CGC-V-36 2"
Manual locked closed valve.
This valve is in series with CGC-V-28 located inside containment which will isolate on a containment isolation signal.
X-71 CGC-V-32 1"
Manual locked closed valve that isolates the Train B H2 analyzer input line.
X-71 CGC-V-24 1"
Manual locked closed valve that isolates the Train B H2 analyzer return line.
X-72 CGC-V-15 2"
Manual locked closed valve.
This valve is in series with CGC-V-14 located inside containment which will isolate on a containment isolation signal.
Nominal Diameter Penetration #
Isolation Valve Inches Comments X-72 CGC-V-10 1"
Manual locked closed valve that isolated. Train A H2 analyzer input line.
X-72 CGC-V-3 1"
Manual locked closed valve that isolated Train A H2 analyzer return-line.
X-39
.SF-V-86, 87 2"
Manual locked closed valves.
Would only be utilized in conj unction with the refueling canal skimmer, pump.
X-67 SA-V-229, 1042 2"
Manual locked closed valve outside containment, normally closed valve inside containment.
. Note: Test connections were not considered in this analysis. Test connections are 3/4" with normally closed manual valves and a pipe cap.
Per the Seabrook Station Technical Specifications there are verified closed every 31 days.
RAI 33
Confirm that a complete and independent check will be performed for the containment strength calculations that served as the basis for the EPZ sensitivity study.
RESPONSE 33 A complete and independent check will be performed for the containment strength calculation. This effort will be completed on approximately November 25, 1986.
1 l
.RAI 34 Fully address the effect of uncertainty in the ' ultimate strength of Cadweld splices on.the pressure capacity of the containment.
'As discussed in the meeting your response should. address potential, non-ductile failure of the Cadweld splices.
RESPONSE 34 See response to RAI 8.
v RAI 35-Assess the response of the containment sump encapsulation vessel on the containment. integrity.
RESPONSE 35 The sump encapsulation vessel is'not a part of the primary containment pressure boundary.
See attached drawings for penetration X-60 (typical of X-61 also) 9763-SL-X60-01 9763-D-801212 Inside containment.CBS line 1212-1-151 is welded by partial penetration welds to a flat plate penetration closure.
In turn, the flat plate is welded to the penetration sleeve.
Containment pressure.is present in the piping up to Motor Operated Valve CBS V-14.
The encapsulation vessel surrounds this valve and is present only as a secondary boundary to contain valve stem leakage, if any.
This penetration is-also not subject to the large shell deflections estimated in the SitA ultim' te. capacity analysis since the line penetrates containment a
below the basemat elevation, (-31'-6").
P
. q* p O R M, A ] ) bl D y....
u E.
f f. 7g I
$I
=
g$2
,3 yf 3. 6-a '
4w i
j 6-
- c I
o_
'. G, %sMir. ~ WR;;
/
u) b*
/
W
- [~f
[
6
~
,( // 7 '//0, j '
' Sif fiTA/i &
t.d N
-E c
e O*w X
f
/
g 00 o J
[
t.O h 3
d A
E.19% - 9005,
{
f %-i):2,
~
~
Y3W
/
// 75/0 lt 0
?
a=~ sm L
zo'd.
[
g lg _ y y b
.r.
, f 'd.'
'r ' 2 ('.",,
t 4
s WB"Q E ?
/r %.
- --i'h.)
Y
_ i A-
..:. y -,g,N e
=
~ lq ! l Q t E s-e]f COZ I-CES-12 f 2 + \\5l-Ib'- i
' 'fE I Q,.f,
g
$ f,5 l
a
~
C BS-it t2-02 to jif r
H{..
7
\\g U /.
{
M o
,e e su s
8 o o O
- na t
e fo so ? y $'o sh S@NEZD 'O' fg
{ {BS-Ill2+!5141) l i 5 Q g
WEAt77E**s2
't, **;
/
w
- 797.g..ggi (E 256 - 276 j
- I Nstt. pipe somer sar.
6 e e e #
1 4 o J J o e=
m g e ur s soa gt.s w AT.
('C85-s
- Dil?R )
5 $ 5 $ 5 $
x O
kcuno7 rA}
A
,/.
9prx:/
V' w fDop E.-
f -50 i
lf=34<g l
< urwo
> + ----~ o
-j q
s 3
MSci.
cza.s o c e t:
r
$$${ !
[/
La r h2078 I g
-Q*
e N
//,
- !L'"IR HWi 9.* 3 i2 2 f o.,h C'
w gic I
$2$$
~
/Q,7j" h j)ghl*I 1
- i tt's-)
f' b
s1.D
\\
s 2e e ss
,3 uta w
s y
oJ( DETalL 2
,0 G;
N N'
V' PGnIETRATION$ s TilNNEL.
}[
f } } y n
m k
'^
LOC A T)QtJ: CONTAINMENT P80E
,4Q p cgbe 3
l
, (9j!
an h;
5 % 3DU 5 a._
a%
e,. a.
h, !
- E 5 cc>t car 4 Nhh wd
+ r k 2
+
m
~
=
. I y
r ct es novoo et awutmm S
,,., si E I
-- f -- - _
_ _~7 rms: coetto 35 ~
h EN Y $
d yt.8 lm ig=l -l M.;-]
en t
nn:%D ' um n
.?
a
_a
m.
.nM..a i. '. - tr,,,
- w...... *..,,.,
?m N.,..
"6.w
- i &,y.. d_
~
i
,q..-
-l.h,k
- f.. $
, hk j 'a.,h.,( )g g gg g j
u
... M. O,, j ]4 "[E qhjQh.
h j
- df m etiM pq y t.h.g;.d, wit. -
~..- 4.
L
..y -.
p!
m mm;y
.a w j j
.i.
,ji(
vq:.
m iunx n li m i-1))p,.
3. gs
. t.p
.. 4, % m!l g!
s.
a
.~g l...
..- n Lg i, y,Mjt
.%.g.,;.i.
balilf re:3:3!!!ib;"qi,[ :. -
- f s.
a t
m'E aui
- a
'a -
., i '
i 6a T,y.y..lhg;i li.g' e i h 3
- 8
~
f.y(l 'd*{Y/.
':.t ! !
i IIIIIIIIIIli!!!!.1 4.!
-(
3" d 9.l 'j l 1
u 4
C,.,t yp*
h [
I
~r
- c.
L l n s. v. : p 3,
5,.
h \\'.c/
\\.
\\rsk.
- sw p.s y
- . ~ I..] j i
'fl5.[:,,l8'l1$ Z!I.d
/
i 7
/ tr.
- n
(
e lK
\\,
,N.
.. _.. '. - ;..}
r._. g
?
'U:..
,5
' ' ~
Y g '(* (/..,,.:h.g\\h. \\.'$ '.
\\
e
' \\,
.6
.e--M.
.+
,g g \\
7
.v p \\h h.
g
'W E..b
'.4 4-fi
\\
j s.i:
/
- g MV K, i; i, \\ / e'q,.
Ng_ $ ~. Pg-t Q:r i.,).@p'A
,g
.: 4,
% Q(
v
"*; A 1 Q'tk t * ( s
, j !
4
.,)..,
'Ie.
y
.M q
/ f" L'"
\\-
O
-,vve m'9
>c
/
,su,e
- .m::
= r..7;.,.
\\
s-
..a /
- y. ;
- ,L -...
4 q.;;.
s,i y'\\
,.{
't "Y
\\
t
- n. I 1
ft
-x+,,. ;
.V.,.
s;,%
- !?
?
\\
3 I
g I ' 1/
Mgg I
' $* b.'
l'a - 5
, [',
sdgy~4
/
kl g,
n J
(
v b,.
is s"T e
- x. -
,, (. l i
>g.
- :.' s.
E' E
I y
t,[&,pp.
. c,,,
c. :.
-q.g' -
,N
/
'4 i,)
\\
f L g'
/
~ 'e - \\
s 4
s 4 r,.
l'
'T n
e\\
a e
l i
g i
susji
..a r r.
/
, j j.,
g
/
r.. \\ r i '.,i 2.t,/
.4
. 4,
\\
,, j/
4
/.
s >'.
\\,
$g ','
.f.I'
\\
d
'/
s.i.a
\\
j 3s (gy'.,..e.
\\
.,. N
..y /.
s
@c.
- .${3 x.\\i.,
_gg i
..,y
.A
. /p,s m;
x
- '~
\\
).' x..
4h
' \\,l
//
\\
, o e.
4
.\\
/
)
,...i-g.
54 % /%
v j\\
s
?
Og Aj
,j \\W
- x/
\\.
t.g= X.
mn m.,
v --
s, y
!q ve1 S,
~~
RAI 40
What indication are available if RHR is lost during shutdown (e.g. spurious closure of suction valve)?
RESPONSE 40 This is dependent primarily on the type of f ailure.that causes the loss of the Residual lleat Removal System. The list below indicates some various alarms and indications for two scenarios; I
RilR Pump trip on overcurrent (faulty overcurrent relay)
RHR PUMP TRIP BREAKER LOCK 0UT alarm on the Video alarm system (VAS) 2 Spurious f ailure CLOSED of an RHR suction valve.
RHR LOW-LOW FLOW alarm on the VAS (335 GPM)
RRR LOW FLOW on overhead alarm annunciator (555 GPM)
RilR HEAT EXCHANGER LOW DISCHARGE FLOW alarm on the VAS (1500 GPM)
Abnormal current indication on the Main Control Board Valve Position on the Main Control Board Other indications and alarms of a malfunction in the RHR available to the operator at the Main Control Board are; INDICATIONS (for both 'A' and 'B' train) mlR pump amperage (MCB and MPCS)
RilR pump discharge flow (MCB and MPCS)
MIR pump discharge pressure (MCB and MPCS)
RilR heat exchanger inlet temperature (MPCS point)
WIR outlet' exchanger outlet temperature (MPCS point)
RHR pump motor winding temperature (MPCS point)
RHR pump radial bearing temperature (MPCS point)
RllR pump thrust bearing temperature (MPCS point)
RilR pump PRIMARY COMPONENT COOLING WATER flow (MPCS point)
RilR heat exchanger PRIMARY COMPONENT COOLING WATER flow (MPCS point)
ALARMS (for both ' A' and 'B' train)
RilR pump motor winding high temperature (VAS)
MIR pump radial bearing high temperature (VAS)
RllR pump thrust bearing high temperature (VAS)
RHR pump discharge pressure high (VAS)
RilR pump PRIMARY COMPONENT COOLING WATER low flow alarm (VAS)
RHR heat exchanger PRIMARY COMPONENT WATER low flow alarm (VAS)
RAI 41
What indication is available for vessel level during shutdown and refueling modes?
RESPONSE 41 As the plant is brought out of the standby condition and' into hot shutdown and cold shutdown, the Reactor Vessel Level Instrumentation System (RVLIS) indicates at the main control board the level in the vessel.
This system is available both with the RC puups cperating (dynamic-range) and without the RC pumps running ( f ull-range). The RVLIS has two electrically-independent trains of instrumentation and indication. The RVLIS is not available when the head is off the vessel. During shutdown when the vessel head is off the reactor vessel, or the RCS is otherwise open to the atmosphere, two methods of vessel level determination are available.
First, a clear plastic tube is connected to one of the loop drain lines and routed vertically to an elevation above the reactor vessel head to serve as a level guage.
The RCS is drained either to just below the vessel flange (refueling) or to the midpoint of the loop piping (steam generator maintenance). The loop drain line connection is made at the bottom point of the RC piping. Therefore, this method of temporary level measurement is available throughout the range of draining required for all maintenance which may take place while fuel is in the vessel. The second method of vessel level determination is by level tranmitter RC-LT-9405 which indicates on the MCB (BF) by RC-L1-9405. This method is also available only when the RCS is open to atmospheric pressure, and is valved-out during aormal operation.
A final metod of vessel level determination is by the Refueling Cavity Level Transmit ter SF-LT-2629. SF-LI-2629 indicates on the MCB (BF). This method is available and applicable only when the refueling cavity is in communication with the RCS; that is, when the cavity level is above the reactor vessel flange with the head of f and the fuel transfer tube gate open.
Y s
i
RAI 42
Does loss of power to the pressure transmitter that provides input to the autoclosure. interlock for,RHR suction valve cuase the valves to close?
t RESPONSE 42 2
Loss of power to the pressure transmitter that provides input to the autoclosure interlock for RHR suction valve will not cause the valve to close.
1 1
i l
i i
i s
}
i i
+
RAI 43
To what. level (s) is the RCS drained for maintence activities while shut-down with fuel in the vessel? What level is nece'ssary to maintain connection with the ultimate heat sink?
RESPONSE 43 For maintenance and refueling the RCS may be drained to the level. required for the activity. The two typical stopping points are 6-inches below the reactor vessel head and to the midplane of the reactor coolant loop piping for steam generator, reactor coolant pump and valve maintenance.
The midplane'of the reactor coolant loop piping is the minimum level to ensure proper operation of the Residual Heat Removal System.
Estimates of the time that might be expected at the different water levels based on Zion data is provided in response to RAI 21.
RAI'44 Describe the availability of the SI pumps while shut down. How difficult would it be to restore the SI function to respond to transients during shutdown and refueling conditions? Consider maintenance of the SI system in your response.
RESPONSE 44 The Safety Injection (SI) pumps (intermediate head safety injection) are to be SAFETY TAGGED (with a CAUTION TAG) with the motor circuit breakers open and rack-out within four hours af ter entering MODE 4 (average RCS temperature <325 degrees F but >200 degrees F) Ojt any one of four RCS cold legs temperatures <325 degrees F, whichever comes first.
To restore the SI function of these pumps, the SAFETY TAGS must be removed from the breaker (s) and the breaker (s) racked into the OPERATE position.
They may then be started manually or automatically by a Safety Injection Actuation Signal.
A more complete discussion of this subject is in response RAI 21.
h 5
4 I
.m.
-._,__,_--_._.._,..m._,___,_,__..,--
a
\\
r 1
j
RAI 45
}
Provide the procedure for establishing cold overpressure protection when shutting down.
1 f
1 RESPONSE 45 Cold overpressure Protection at Seabrook does not require manually arming.
It is armed automatically on Reactor Coolant System temperature decreasing to less than 342 degree F.
The attached illustration below j
shows the logic employed in arming and operating the LTOP system (Low Temperature Overpressure Protection) during an RCS cooldown and depressurization for Protection Train ' A', Train 'B' logic is similar.
t
}
PORY (Power Operated Relief Valve) Block Valve auto-open (NOTE:
this is a NORMALLY OPEN valve.);
Power available to valve AND valve control in remote AND 4
I valve control in automatic AND l
LTOP train 'B' armed (auctioneered low wide range RCS cold leg i
temperature is less than 342 degrees F.)
i l
PORY Train 'A' will auto open if; l
Valve in automatic AND Wide Range Reactor Coolant pressure channel (PT-403) is greater
}
i than the LTOP setpoint (see attached graph, Protection Train
'A' l
LTOP setpoint generated by auctioneered low RCS hot leg temperature.)
3 l
7 5
' % g os.
h i
(
{
l i
X l
T a
m i
T m
x e
X C
W i
i g
f.
T g
1 t
- :t i
\\
a g
1 3
g
\\
E.
E
,l
,y-r g
h i
1 8i
'1I ~
~ ~~
E w
I S
s m
>11 4-
-lll s
t e
o e
o e
e 2
2'
~
~
(sisd) Amiodass Aged RARIEVR f
1 9
QUESTION 46 1s the primary system.made water-solid during shutdown?
i j
1
~.
RESPONSE 46 There is at present no plans to operate the RCS for long periods of time in the water solid condition.
If the plant is shutdown to refuel, or for other maintenance, the pressurtzer will be taken solid per operating procedure to collapse the steam bubble, then drained to the required level.
Another condition when the RCS is water solid would be when drawing a pressurizer steam bubble following filling and venting of the RCS. This would only be a transitory condition.. Another possible mode i
of operation is to have the RCS filled and vented to the PRT.
4 I
t 4
RAI 49
The FSAR gives RHR relief valve flow rate as 900 gpa with a set pressure of 450 psi.
The flow rate does not agree with the value used in Reference
-1,'section 3, page 6.
Please explain.
RESPONSE 49-The correct rated flow of each RHR relief valve is 900 GPM at 450 PSI.
Although the value of 990 GPM stated is incorrect, the V-sequence probability modeling used a total of 1800 GPM for both relief valves.
It was assumed that a valve rupture flow of greater than 1800 GPM would result in RHR pressurization to 2250 psia.
For the flow calculations in MAAP a pressure dependent flow capacity was used as described in RMEPS.
This is somewhat conservative since the initial driving pressure is over 2000 PSI.
t i
I i
1
)
't i
L _.
I~
.L
RAI 50
Please describe the mechanism for assuring that plant changes and new knowledge are promptly factored into the technical considerations which form a part of the foundation for staf f consideration of a reduced emergency planning zone radius.
RESPONSE 50 The PSA will be kept up-to-date with regard to plant changes and new knowledge.
Wherever. major plant modifications are performed which significantly affect the PSA, the report will be updated accordingly.
To ensure prompt and ef fective input to plant change decisions, the PSA will be considered in the design change review process. PSA considerations will include the impact on public health risk.
The RMEPS results already reflect some of the changes to plant Technical Specifications that have been made since the SSPSA was completed in 1983.
Recently, a review has been completed by the PSA team of changes made to the plant since 1983.
No changes were identified that would impact the current PSA results in any significant way.
i
RAI 51
Reference 1, page 3-7, paragraph 5) references both high and low level-sump alarms. What is s sump low level alarm?
RESPONSE 51 The CBS pump cubicle sump pump is tripped on low sump level and an alarm is annunciated at the. waste management system control papel (CP-38A) and remotely in the main control room via the VAS. The pump trip and low level alarm are set to protect the pump from cavitation.
1 i
RAI 53
Ref erence is made. on page 3-7 to the RHR system crosstie line and RHR system response due to flow in this line as well as in the miniflow bypass lines.
The conclusion is drawn chat the RHR system pressure will tend to be uniform as a result.
Are flow conditions such that this is realistic? What is the impact of. this assumpation on conclusion pertinant to the discussion.
Response 53 Page 3-7 states that "...the entire. RHR system will tend to pressurize uniformly af ter valve f ailure (neglecting the time it takes for pressure valves to traverse the system)".
The figure included with this response provides a diagram of the RHR system. The configuration of the Residual Head Removal System when it is aligned as part of the emergency core. cooling system is for both RHR pumps to be aligned through separate suction paths f rom the ref ueling water storage tank.
The discharge side of the pumps are connected downstream of their respective heat exchangers by an 8-inch crosstie line containing two open motor operated valves.
Each pump has a minimum flow line that connects the outlet of the RHR heat exchanger to
~
the pump. suction piping. The minmum flow line is a 3-inch pipe containing a flow restricting orifice (750 GPM at 185 PSI) and a normally open motor operated valve.
In the RHR pump suction piping, branches from the refueling water storage tank, the containment sump and the reactor coolant loop join the RHR pump suction pipe.
It is the pump suction pipe from the reactor.
coolant loop that contains the 3x4 inch relief valve set at 450 psig with a rated flow of 900 gpm.
Table 4-6 and Figures 3-1 and 3-2 of PLG-0432 provides additional details of RHR flow paths.
Except for the RHR pipe f ailure case (which maps to the most severe release category, SI), the interfacing LOCA flow rates through to RHR system are expected to be moderate (a few thousand GPM) relative to the RHR piping equipment size (see, for example section 3.1.4.2 and Figure 3-7 of PLG-0432).
The pressure losses through the RHR system will not be large such that the pressure should be fairly uniform.
We see no significant " impact of this assumption on conclusions pertinent to the discussion" for the following reasons:
1)
As dismessed throughout PLG-0432, the RHR system overpressure failurew modeled conservatively envelop the spectrum of possible f ailure modes.
2)
The RHR pressure boundary failure (s), if any, are expected to be caused by the initial pressure wave or pulse, not f rom flow induced pressure drops.
3)
The precise pressure distribution through the RHR.is of little importance since the risk results are not extremely sensitive to the exact RCS blowdown /RHR pressurization rate.
t
i R)
)R O
O T
A TA L
t U
u M
M U
U C
C C
C A
i sb A
(
(
i s
A - B N - N I
I A
A R -
R T
T PC 3 R 4
yP 2
1 P
O O
L O
O L
x -
R 4
1
_ i P
W
=
PO g
s O
E y,
[a
- O O
P I
L L C V
R R
y E
VO 4
I I
M M
C E
C T
l C
S C
O Y
O S
R
)
)R p
R g __
R o
H O
r H
T u
R A
L
.o U
M U
o C
m C
A i
(
i i
t s
s X
I X
s r C
H H
O C
l l
?!
a l
d
=
RAI 55
What is the justification for the statement on page 3-10 that the first sign of trouble wil' be pressurizer low level or low pressure alarms?
We suspect a number of other indicators may be first, such as abnormal indication from the PRT or even a smoke alarm.
RESPONSE 55 The indications of an RHR Interf ace LOCA would primarily depend on the magnitude of the inleakage from the RCS as was stated on page 3-10.
If the RCS to RHR leak flow rate is less than the capacity of the RHR suction relief valves, the first indications the operator may receive are; PRT (pressurizer relief tank) high pressure RHR discharge pressure meters readings abnormally high,~ 450 PSIG (0-700 PSIG scale)
RHR heat exchanger outlet. temperature recorders pegged high (0-500 Degrees F)
Pressurizer Level low (-5% deviation f rom programed level)
Pressurizer pressure low (2210 PSIG)
For in-leakage less than 1800 GPM, the RHR system would pressurize to approximately 450 PSIG (the setpoint of the suction relief valve) and stay at this point.
The RHR suction relief valve (s) would cycle to maintain the RHR system at 450 PSIG.
It is important to note that this would entail no threat to'the RHR system integrity, the mechanical seal-package would remain INTACT and no leakage would be seen outside containment.
If the RCS to RHR in-leakage was greater than 1800 GPM, che indications the operator would receive would be somewhat different.
Based on the worst case RHR Interf ace LOCA sequence run on the MAAP code by Westinghouse, the following alarms / indications would be received by the operator. (please ref er to plot of RCS Pressure vs Time f or basis of alarm sequence) These alarms / indications are listed in CHRONOLOGICAL order and DO NOT take credit for any radiation monitoring systems alarms, although there are both. area and ventilation process monitors that monitor the RHR equipment vault ' area; RHR DISCHARGE PRESSURE HIGH alarm train 'A' ('B') (560 PSIG)
MIR discharge pressure meters pegged high (0-700 PSIG)
RHR heat exchange outlet temperature recorders pegged high (0-500 Degrees F)
PRT Pressure High alarm (4 PSIG)
PRT Pressure and level increasing with no indication that the pressurizer PORV(s) or Safety Valves are discharging.
Pressurizer PORV(s) and Safeties have the following devices to monitor' their position and process flow thru them; Valve position for PORV(s)
Tailpipe temperature for PORV(s) and Safety Valves Acoustic flow monitoring for PORV(s) and Safeties PRESSURIZER PRESSURE LOW alarm (2210 PSIG)
PRESSURIZER LEVEL LOW alarm (-5 deviation from program)
RHR EQUIPMENT VAULT HIGH TEMPERATURE alarm Train ' A' ('B')
RHR EQUIPMENT V AULT HIGH SUMP LEVEL alarm Train ' A' (' B')
RAl 57 The statement on page 3 - 11 that "As soon as the pumps begin to produce flow to the RCS, valves in the miniflow lines close and all RHR pump flow is injected into the reactor vessel via the RHR cold leg injection lines" is not correct. The sensors are not located at the RCS to detect flow at that location. Further,'one is postulating a break in the RHR system, and a significant portion of the pump flow may never reach'the RCS (as it stated in a later paragraph).
RESPONSE 57 The referenced paragraph states that " As soon as the RHR pumps begin to produce flow to the RCS, valves in the miniflow lines close and all RHR flow is injected into the vessel via the RHR cold leg injection lines".
This paragraph describes the response of the Emergency Core Cooling Systems to a standard LOCA, and is correct. The RHR flow transmitters are not located on the RCS proper, but are indicative of RHR injection flow via the cold leg injection lines.
I m
RAI 59
An item under consideration for advanced nuclear power plants is the ability to monitor pressure on the low pressure side of check valves.
This could provide early warning of check valve leaks and would pro-vide monitoring capability to. help assure check valves were operating properly. The same monitoring capability with respect to RHR suction line valves could identif y if individual valves were mispositioned or malfunctioning. Would such a system for Seabrook be of significant benefit in reducing risk in a reduced size emergency planning zone?.
RESPONSE 59 The failure model and quantification for leaks greater than 150 gpm is summarized below for the four cold leg injection paths (two check valves in series) and the two hot leg suction paths (two MOVs in series).
4 COLD LEG INJECTION PATHS CHECK VALVE NO. OF FAILURE CONTINUOUS PATHS MODEL QUANTIFICATION (ANNUAL TESTING)
MONITORING 4
2 ( T/2) 4 [( 4X10-4 )2+Va riances 3.5 X 10-6 o
=
4 2
d 8X(4 X10-4 ) X( 2.7X10-4 )
0,9 X 10-6 0.9 X 10-6 2 HOT LEG SUCTION PATHS NO. OF FAILURE PATHS MODEL QUANTIFICATION (18 MONTH TESTING) 2 2 ( T/2) 2 [(4X10-4 )2+VarianceuXI.5 =
2.7 X 10-6 2.7 X 10-6 0.4 X 10-6 0.4 X 10-6 2
2 d
4X(4X10-4 )X(2.7X10-4 )
=
0.1 X 10-6 0.1 X 10-6 2
g 2X(4X10-4 )X( 1.1X10-4 )
=
7.6 X 10-6 4.1 X 10-6 The last column above assumes perfect continuous testing (monitoring) of the two series check valves.
As shown the total f requency changes very little (less than factor of 2).
Hot.ever, quantification of the other failure modes and treatment of MOV discs as check valve discs are believed to be conservative.
" Perfect continuous" leak testing of the RCS series check valves would require significant modification and is not practical.
Frequent, periodic testing can be performed to verif y that excessive leakage is not occurring through the inboard RCS isolation valves. Excessive leakage through the outboard check will possibly be detected by increased pressure in the RHR system and reduced accumulator level depending on the magnitude of the leakage.
Currently the normal RHR suction motor-operated valves cannot be tested as described above because permanent test lines are not connected to the process lines which the valves isolate.
In addition, as discussed in the response to RAI 48b, the interf acing LOCA event contributes approximately 12% to the total early release. frequency.
Therefore, if this was reduced to zero it would be a minor reduction to tctal release frequency. There is no significant benefit to reducing risk.
i i
i
7
RAI 60
Please elaborate on the page 3-23 list of actions an operator can take to mitigate the accident. This list appears to be short. Include iden-tification of what has been incorporated into operator training and proce-dures at Seabrook.
RESPONSE 60 At present, the following isolation sequence is being incorporated into training:
1.
Check proper valve alignment RC-V22 RHR pump "A" Th suction CLOSED RC-V23 RHR pump "A" Th suction CLOSED
- RC-V87 RHR pump "B" Th suction CLOSED RC-V88 RHR pump "B"- Th suction. CLOSED 2.
Identify and isolate leak CLOSE THE FOLLOWING VALVES
- RH-V21
- RH-V22 OBSERVE THE RESFONSE OF THE RRR SYSTEM Faulted train follows RCS pressure / temperature
.Non-faulted train depressurizes CLOSE FAULTED TRAINS TO DISCHARGE VALVE
- Energize MCC 522/622
- Close RH-V14/RH-V26 STOP FAULTED TRAIN RHR AND CBS PUMPS CLOSE FAULTED TRAINS RWST SUCTICN VALVE
- CBS-V2
- CBS-V5 3.
Check if break is isolated:
RCS Pressure - increasing or Reactor Vessel level - increasing and Faulted RHR loop pressure - decreasing
A 4.
If break is not isolated:
- Energize MCC 522 and 622
- Close RH-V14 and RH-V26 RETURN TO STEP 3 This strategy allows:
a quick reduction in leakage flow by' isolating the faulted train a.
from the non-faulted train b.
prevents cycling the RH injection valve's if not necessary 1:.
allows better diagnosis by separation of trains.
t a
a l
l
RAI 61
What is tk fragoency of failuras in the pipe tunne.1 that is mentioned on page 3-23, and which led the authora to conclude they are very lovt
~
Rf,SPONSE 61 As noted on page 3-24, top event FI in the VI and VS event trees.
represents any failure of piping or the heat schanger doe to the RRR systes high pressure challenge. Any piping f ailures are assigned to plant damage state 1 FV.
As Table 4-17 shows, damage state IFY always maps to release category 51-the moet severe release. In other words, soy RHR pipe f ailura is asaused to result in the most severs release, regardless of failure location; therefore results are f aaenaltive to the pipe failure location.
e l
g W D
--a. m
--__r-
~
o
RAI 62
Page 3-27 references situations where the combloed stusp pump espacity is sufficient to remove leaks and keep the vaalts from flooding. In t hes e cases, the RHR, 51, and CS pumps are assumed not to be impacted by flood-ing. What conalderation was given to failure of one (or both) snap pumpe?
RESPONSE
As shown in Figure 3-4 of PLG-6432, the sequences referred to in this question have frequencies on the order of IE-8 or less; these freqosucies, when multiplied by the chance of snap pump failure clearly make such sequences unimportant contributors.
l
\\-
e9 g t
--)
1
RAI 63
is the maximum flow rate that can be injected into the RCP pump What seals? -(Of potential interest since it may be an alternate path for
. injection into the RCS.)
RESPONSE 63 Flow to the reactor coolant pumps seals is significantly less than that required for core cooling.
l l
c -.
RAI 63
What is the maximum flow rate that can be injected into the RCP pump seals?
(Of potential interest since it may be an alternate path for injection into the RCS.)
RESPONSE 63 Normal seal injection flow into the reactor coolant system through the reactor coolant pump seal injection flow path is 5 GPM per reactor coolant pump or 20 GPM total. This flow path is judged insuf ficient by itself to satisfy core cooling requirements.
Other flow paths such as the normal charging line; the charging / safety injection flow path to.the cold legs or the safety injection pumps to the cold legs (at reduced RCS pressures) will provide greater flow capability.
It should be noted that there are no risk significant core damage sequences in the PSA results in which the charging pumps are available.
RAI 64
Shutting an RHR spstem crosstie valve is identified on page 3-35 as an action to help l'solate-a LOCA outside containment involving the RHR/SI systems. Has a careful evaluation of these systems been performed to assess isolation strategy? ~ If so, are procedures in place at Seabrook Station which reflect the work?
RESPONSE 64 See Question 60.
5
.--y,-,-
ne--
,,.,.-,,,m.,
,e-.-----
,.,..,,.,e.-..
m--
RAI 67
Page 3-37 contains the wording "End state DLOC contains sequences in which the interf acing LOCA has been terminated, and the ECCS has been -
degraded (D) (RHR or SI pumps have f ailed)..'.The point estimate frequency The additional failures required to achieve of DLOC is 4.0 x 10-7 per year.
core melt would lower this frequency by at least one order of magnitude."
What is the justification for this conclusion? (We have already lost' a water into the RCS via the usual portion or all of the ability to inject paths).
RESPONSE 67 End state DLOC represents sequences in which the interf acing LOCA has been terminated (isolated) and ECCS has been degraded but some core make up capability remains available.
(see top of page 3-27).
"The additional f ailures required to achieve core nelt would lower this~
f requency by at least one order of magnitude" means that, given the i
degraded ECCS state, the unavailability of remaining mitigation equipment is better (lower) than 0.1.
l l
l l
""^s"'--
RAI 68
The: bottom of page 3-37 contains a statement to the effect that failure of one charging pump will lead to core melt. Why is this the case? Our perception is that sufficient flow might be provided by alternate means to keep the core covered, such as use of the remaining two charging pumps, and perhaps the ractor makeup water pumps.
RESPONSE 68 For the DILOC accident sequences discussed at the bottom of page 3-37, l
the f ailure of one charging pump would not be expected to lead to core melt. This is a conservative assumption to estimate -the frequency of failures with the rest of the plant that would be needed to produce a core melt.
i l
3 RAI 69, 70, 71, 72, 73
RAI 69
~
What is to be the status of the " temporary" 34.5 kV power lines which are identified on page 3-457
RAI 70
I What is to-be the status of the mobile power supplies which are identified on page 3-467
RAI 71
What capability has been provided to connect external pumps as identified in the second and third. paragraphs of page 3-467 (This was briefly mentionsed on page 3-48).
Use of a pump to simply inject water into containment via the sprays on a short term basis (no recirculation) does not appear to be identified. 'Has this been considered?
RAI 72
Page 3-46 'iden' ifies a number of possibilities for recovery of various t
saf ety f unctions.
Are there. specific plans?. If so, please provide them.
RAI 73
There have been several~ references to purchase of a mobile electric generator by pooled resources on the pages prior to page 3-49.
What is the likelihood that such a generator would be needed by several plants at the same_ time, and hence might not be available to Seabrook Station when needed? Similarly, where is the generator to be stored, and how is it to be transported to Seabrook?
Include consideration of post. seismic and post severe storm in the response.
RESPONSE 69, 70, 71, 72, 73 i
These question refer to section 3.2 which discusses Containment Recovery following an extended loss of all AC power. The various potential recovery measures discussed are related to recovery which could prevent late containment failures predicted to occur one to several days after the initiating event.
These late containment failures, and therefore, the recovery measures discussed have no effect on early health risk. These recovery measures are not related to emergency planning decisions.
Reducing the chance of late containment failure could impact latent health effects which are not sensitive to evacuation assumptions.
Alternate ways of recovering late containment failures are still being evaluated.
i 1