ML20128E129
| ML20128E129 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Limerick |
| Issue date: | 07/17/1984 |
| From: | Helwig D AFFILIATION NOT ASSIGNED |
| To: | NRC |
| Shared Package | |
| ML19292B772 | List:
|
| References | |
| FOIA-84-624 NUDOCS 8505290305 | |
| Download: ML20128E129 (11) | |
Text
-
99 a
MECHANICAL ENGINEERING DIVISION N2-1 2301 Market Street MEMORANDUM
SUBJECT:
Limerick Generating Station Response to PSRB Questions a)
D. G. Eisenhut (NRC) letter to BWR Owners' Group dated 2/4/83 b)
Transcript of Limerick ACRS meeting, 10/13/83 c)
F. B. Litton (NRC) memo for K. Kniel (NRC) dated 5/10/84 d)
J. S. Kemper (PECo) letter to A. Schwencer (NRC) dated 6/15/84 e)
R. J. Stipcevich (PEco) telecon with W. Kennedy (NRC) dated 7/11/84 f)
PECo/NRC Conference Call dated 7/13/84 Questions regarding the reference d) submittal were discussed in references e) and f).
The following responses to these questions have been compiled in preparation for discussions with the NRC staff at a meeting scheduled for'7/20/84.
Question 1 -
Since LGS didn't use 2.0 times design pressure as the Primary Containment Pressure Limit:
a)
Provide calculations used.
b)
Describe what new plant calculations or data necessitated the departure from 2.0 times design.
c)
Describe what role the LCS PRA played in this.
Response
The BWR Owners' Group (BWROG) Emergency Procedure Guidelines (EPGs) call for venting of the primary containment at the " primary containment pressure limit".
In all situations where this step is taken, the plant will be significantly beyond its design basis.
This procedural step has been described to the NRC and approved on a generic basis per reference a).
The BVROG has continually asserted that the primary containment pressure limit is a value which must be selected. based on plant specific review and analysis.
The NRC staff _
has suggested that an appropriate value could be established basdd on generic analyses of the structural capability of the various BVR containment types and has recommended in reference a) that an interim value of 2.0 times containment design pressure be utilized
'provided containment integrity can be demonstrated for each plant at that pressure.
h52g5041015 SHOLLyg4-624 ppg
y
. t d
The staff has subsequently been advised by the BWROG that such generic structural analyses are not practical and that plant specific analysis of containment ultimate capability would not necessarily determine the proper limit for procedural use because of other potential negative safety impacts. Such competing considerations as pneumatic supply pressure required for safety relief valve (SRV) actuation and differential pressure limits for containment vent valve operability must also be considered (see references b) and c)).
The BWROG has advised the NRC staff that the bases of the EPGs will be revised to include specific discussion of these considerations in EPC Revision 4.
As des:ribed in reference d), plant unique evaluations'have been performed for LCS and a primary containment pressure limit of 1.3 times design pressure (70 psig) has been selected. Considerations in the selection of this value are described in reference d).
=
Specifically, the differential pressure limit of the large purge valves was found to be 76 psid.
In addition, it was determined that a pneumatic supply pressure 25 psig greater than containment pressure was required to facilitate opening of the LCS SRVs.
As discussed in reference b), a review of the published BWR PRAs will reveal the sensitivity of core damage frequency to postulated containment over-pressurization events.
In the sense that the EPGs include venting of containment in order to minimize the potential for indeterminant containment failure modes, the EPGs will reduce
~
the potential for consequential loss of RPV makeup and uncontrolled release of containment atmosphere which have previously been assumed in PRA's.
Therefore, containment venting may be considered a lesson learned from industry risk studies.
In fact, WASH-1400 recognized the potential benefit to be gained by the relatively simple action of containment venting but did not take credit for such action because procedures for its use did not exist. Although preliminary studies of containment venting were performed in support of the LCS PRA, no credit is currently taken for venting in the LGS PRA and the LCS PRA was not utilized explicitly to select the LCS specific pressure limit.
Question 2 -
Discuss operability of purge valves under accident conditions.
a)
How can valve operability be assured at specific pressures?
b)
Can the valves be reclosed after they are opened?
c)
Discuss the competitive risk of failed valves vs. not venting.
Responnes, As stated in reference c), the valves in each identified vent path for LCS have been determined to be operable for both closing and opening at differential pressures ranging from 76 paid to over 150 paid. The evaluation of operability has included consideration of all relevant loads, including flow induced dynamic effects. Those val'ves with the lowest differential pressure capability, the 18" and 24" valves, are all oriented such that containment pressure resists valve opening, or aids valve closure.
Thus, valve reclosure would
, i i
be cc isted by contrinment pressurization.
Since operability has been evaluated, it is judged that the risk of potentially failed valves is quite small and that the benefits of venting for these degraded conditions far exceed any risk of subsequent failure of the valves to reclose.
Question 3 -
Demonstrate that the sequence of valve opening is adequate to exercise control of depressurization.
Identify the differential pressure operability for each proposed vent path.
Response
a Since the EPGs are purely symptom based, there is no specific design basis event to consider in the assessment of containment depressurization rate.
As discussed in references b), c), and d),
the identified containment vent paths have been ranked in order of preference for sequential use to minimize the rate of containment depressurization; limit the rate of release to that required to stabilize containment pressure and ensure maximum filtratien of containment atmosphere during the venting process.
BasedonIDCgR and other analyses, it is estimated that a vent area of as 0.1 f t willbea6aquatetocontrolpressurefora}1eventswhentheteactor is suberitical and a vent area of v 1.0 ft vill be adequate for ATWS type events.
It is apparent that,given the range of vent
~
paths available for use at LGS, appropriately sized venting capability exists for this broad spectrum of events. Additionally, as stated in references b) and c), even if the rate of depressurization was rapid, loads would not be significant and backup injection sources would be available if NPSH limits were encountered. The differential pressure rating and size of each vent path is indicated on the attached sketch.
Question 4 -
Discuss any considerations that were given to venting to secondary containment rather than directly to atmosphere.
Response
As discussed in references b), c), and d), each of the identified vent paths has differing levels of desirability with regard to fission product retention, potential for causing adverse reactor enclosure environmental conditions, and potential for equipment damage.
It has been judged to be preferabia to favor avoidance of potentially adverse reactor enclosure environmental conditions ove r the dose reduction benefit that would be received from plateout, dilution, and radionuclide holdup in the reactor buildfpg for almest all situations.
Besides the potential negative impact *aventing to the react,or enclosure on equipment reliability, venting to the atmosphere assures access to reactor enclosure equipment for repair, inspection, and/or manual operation.
Question 5 -
Provide results of evaluations to determine the effect of early purge valve activation.
The analysis should consider the merits of delay with regard to the plateout and other considerations.
Responne Although no specific analyses have been performed for LCS, it is not expected that there will be any notable differences in this regard between venting at 1.3 timen desinn and ventina at 2.0 Bimoo MooRnmo
4 Qu stim 6 -
Dic::s3 concidsrction given to c:nt0inment v:nting
. crit:ria r:10tiva to rcdictirn Icvels and core, conditions.
Response
As discussed in references b) and c), the EPCs presently do not contain containment venting limits based on radiation level or core conditions.
The radiation control portion of the CPCs is currently under development and will be submitted as part of EPC Revision 4 As stated in reference b), venting of containment under conditions when.high radiation levels existed would only be performed with the cognizance and approval of the appropriate regulatory and govenmental agencies.
For all cases in which containment' venting is envisioned, significant quantities of radionuclides will not be present for release and/or the venting path will maximize their filtration ani holdup. As stated in reference c), for all conditions when venting is accomplished from the suppression pool air space, radiation releases will be insignificant from a risk point of view due to fission product retention in the suppression pool.
Por all transients, releases would be expected to be a small fraction 10CFR100 limits and releases would be expected to be less than 10CPR100 limits even for core melt accidents.
Question 7 -
Provide information'on the frequency and magnitude of the dynamic loads on the suppression pool resulting from SRV actuation at.the proposed vent pressure.
L.
Response
As discussed in references b) and c), there are multiple EPC directives to ensure RPV depressurization before the primary
}
containment pressure limit would be reached.
SRV discharge loads at low RPV pressure have been demonstrated to be small for SRV T-quenchers such as at LCS. Condensation will be stable at least up to saturation conditions.
^-
Question 8 -
ou state that vent paths will be used in order from small to large; however, the listed valves are not consistent with this statement. Provide your
+
justification for venting with the larger lines before the smaller ones.
Response
Reference d) indicates that the identified vent paths have been ranked in order of preference considering a number of important factors including size and location of the vent. The indicated order of preference is from small to large diameter paths connected to the su'ppress(on pool air space or flitered through SCTS, and then from small to large for lines connected to the drywell. The one exception to this is the "6 inch ILRT line from drywell" which is not an independent line. This line is used last because its use would tend to reduce fJow rate from the similar connection to.the suppreasion pool air space.
I
r r
-5i i
- Question 9 -
Provid] ju:tific: tion f r the conclusion that the air supply pressure will be sufficient for SRV operation at the selected pressure limit.
Response
The normal range of gas supply pressure for the Primary Containment Instrument Cas (PCIC) System is95-110 psig. Since a pneumatic system differential pressure (i.e. - above containment pressure) of 25 paid is required to open and hold open the LCS SRVs, supply pressure will be adequate to assure SRV operability at the primary containment pressure limit.
If somewhat higher pneumatic pressures were required, the PCIG operating pressure could be raised t6 the setpoint of its relief valves (120 psig) or the pressure regulators on the bottled backup supplies to the ADS valves could be raised, s
Question 10 - Presuming that the degraded condition is due to naavailability of ADS, would it be appropriate to have th* operator vent at 70 psig.
Response
The ADS SRVs are 5 of the total of 14 SRVs provided at LCS. The ADS function is provided by logic which automatically opens these selected SRVs to facilitate rapid depressurization under certain conditions.
The unavailability of this automatic function would not prevent RPV depressurization but may have some limited bearing on the selection of a vent pressure.
This conclusion is based on l
the evaluation of the ADS function which indicates that other failure modes dominate its failure probability, but that high 14 containment pressure could also induce ADS and other SRV failures.
Maintenance of a lower containment pressure will increase the likelihood of any SRV operation by maximizing the available pngunati,e supply differential pressure.
.~.
i Prepared byt. ""D. R. Helwig July 15, 1984 DRil/cmv/07188401
~~
Attachment
e p
p u
,.... m D
o e-0-
6 V
u
.g 4
h 1
W4 1
a a
M h j
_N
%g f
- p
.M a
a 9
j g
a x
x D
J -
3 J
i
~
i I
y et r
(
- E b
b~
o
=
=
=
W lb c*
-I 1 $ =,"
, M M.
~
t h4 NM 3
/
5
,d d
/70 o.
l General Comment
{
t Due to the very limited time available to review this document, it was only possible to carry out a cursory review of Sections 7.1 to 7.2.3.2, 7.9.1 and 7.9.2.
Unfortunately, the other sections were not reviewed.
It is our opinion that much more time is required in order to evaluate all of the items discussed in the report. Specific comments related to the Sections 7.1 to 7.2.3.2, 7.9.1 and 7.9.2 are given on the following page.
(
(
W
(
)-
s.
f 1
Comments on Seismic Ground Motion Hazard at Zion Nuclear Power Plant Site Section 7.9.1 1.
The validity of using peak ground acceleration as the only criterion of ground motion intensity is questionable.
2.
On page 3, basic assumption 3 mentioned that, "The peak accelerations at the site of interest can be represented as a function of the earthquake mapittude, the distance between the site and source of energy release, and the local soil condition." However, equation (2) expresses peak acceleration as a function of magnitude and distance only with no mention of soil conditions.
A I
2-
' (.
,, - -, - - - - -., ~,.,..
l r
!.{
Cursory Comments on g.
Conditional Probabilities of Seismic Induced Failures for Structures and Components for the Zion Nuclear Generating Station Section 7.9.2 (up to 5-1) l
~
- 1. ' Failure is based on peak ground acceleration only.
Load combinations L
are not considered. It appears that the results are not realistic.
2.
Statement on page 2-4, " Probability of failure estimated in this report are based on limited existing analysis and qualified engineering judgement and assumptions." This leads to the question as to whether such l
[
limited investigations can result in meaningful conclusions.
3.
Failure probability are based on the fragility curves. These fragility curves are estimated from median ground acceleration capacity 5 and
(
uncertainty variables 8g (inherent randonriess about the mean) and f u l
(uncertainty in the mean value). All the random variables are assumed to be logno rmal.
It appears that such crude models may not be-adequate.
L Furthennore the definitions of GR and tu are ambiguious l-4.
Statement on page 2-7, "Lognormal distributions can be justified as a
- reasonable distribution so long as one is not primarily concerned with the L
extreme tails of the distribution." On the contrary, failure probabilities l
are always dealing with the tail part of the distribution.
5.
The assumption of lognormal distribution for all the variables appears to be questionable.
In fact, published a.uments indicate many of these-variables are other than lognormal.
[ '.,. q-}
9
- ~ **
(
6.
The failure probability was not directly stated in the report.
It
{
appea'rs that a safety factor is used but, the relationship between the failure probability and the safety factor is not discussed.
' 7.
Specifically, in regard to the. strength of concrete, the decrease of strength of in-place concrete may be significant. (See Nureg/CR-1423, Vol.1).
The report does not consider this factor as important.
8.
In regard.to the Tables 4-3 to 4-19, the listed data, such as mean FS, @R. hu and mean acceleration capacity are not fully supported and explained by caputation.
There is no way to check their validity with the pertinent information.
9.
No load cabination of earthquake and any other hydraulic transient loads (LOCA) are considered.
Because of this the approach may not be on the safe side.
(
- 10. Strength of' a limited number of structural types are given (such as shear wall). However, important structural types such as containment shell structure strength are not discussed.
11.
No summaries or conclusions were presented in the report.
. i
M/
TD( D j
i i
y r
r M
Chdb.kQP.A GG..
1D MQ hg%
/(6 CHA CTERISTICS OF GROUND SHAKING 3
L24 j
s, r 4o
,,,,,u.a.,
and seishde moment. It'is also has a corneY frequency P coima o==.
(the frequency where the high-and low frequency
- a.s2 l
- s. is*E. component trends interact) that is related to the radius r of the equivalent circular fault causing the earthquake.
9 3
C h
I h --
In engineering seismology the spectral characteris-5 l
[
tics of ground motion are normally displayed as re-8,, ;, s sECoNoS sponse spectra, a form preferred by structural en-y-as2 gineers for the study of building response. Structures
-i.24 respond as oscillating systems with fairly well defined sackmd duation n==ro periods of vibration, and their response to ground mo-tion is strongly dependent on its spectral composition and duration. Simple systems such as viscous-damped pendulum or mass-spring systems have been success-fully used to model structural' elements (Blume and h
h k,,, ountion20.ost 12.2:
others,1961). When such a model is excited by ground motion from an earthquake, it will respond by vibrat.
ing. The vibratory. motion is described by the well
-,[
FAULT RUPTURE LENGTH.lN KILOMETERS known differential equation
},
s to to to 70 i4o 2so 500 i + 2w,, hi + w,, 'x = -a(t) i i
where x is the relative displacement between the mass a!
0 0""'*">"05' and ground, i andi are the velocity and acceleration of s
+ Values proposed try Housner(is6s) the mass relative to the ground, h is the fraction of C 4a critical viscous damping, w,, is the undamped natural
- 1, eachme sunson. o >c.2s, frequency of vibration of the system, and atti is the
+
g
--.sactmo sursiion. o >o.i, ground acceleration.
The Fourier amplitude spectrum has a close relation 33 4
to the undamped velocity response spectrum iHudson, i
1962). These two spectral respresentations, although fundamentally different, are essentially interchange-3
+
able in jeismic data analyses.
E Arupt sevaa. ises The response spectrum technique, proposed by Be-t!o Aimperias viney. is40 nioff(1934) and Biot (1943), is a method for determin-8 saa Fern.ndo.gri ref.p ing the maximum amplitudes of response of an ensem.
g As FL5 ado.is7:(r.coim.o,
ble of simple damped, harmonic oscillators (a narrow-
+
band filter) when excited by a given ground-motion a; "
~
, / rupisouna.inss time history (fig. 31). The various response spectra are:
i A
/
A Km covary. issa (1) pseudo absolute acceleration (PSAA), (21 pseudo 3g
/
relative velocity (PSRV), (3) absolute acceleration i gsIn Funcace. iss7 (AA), (4) relative velocity (RV), and (5) relative dis-placement (RD). Each response characterization has a o
s s.s s
s.s 7
7.5 s
s.s physical meaning. PSAA is a measure of the maximum MAGNITUoE elastic spring force per unit of mass. RD represents the Facuar 27.-Bracketed duration values for the S.16* E. accelero-maximum value of the relative displacement of the gram recorded at Pacoima Dam from the 1971 San Fernando simple system during vibratory motion. PSRV gives an eerthquake (top; from Ilays,1975at, and bracketed duration as a approximate index of the greatest velocity, relative to functi of magnitude and fault rupture length (bottom; modified its base, of the center of mass of the resonant simple structure. PSRV can also be related to the maximum 1970 to deduce source parameters from the far-field energy absorbed in the spring. For low damping, the di placement spectrum has stimulated much impor-PSRV spectrum provides an upper bound to the ttnt research in seismology. The displacement Fourier amplitude spectrum (Jenschke,19701 spectrum, after being corrected for instrument re-Response spectra for four values of damping i0,2,5, sponse and path propagation, has a flat low-frequency and 10 percent of critical) derived from the horizontal 1; vel (fig. 30), which is used to define the stress drop component accelerogram recorded at El Centro from N