ML20034F531
| ML20034F531 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 02/19/1993 |
| From: | Kiper K, Rau L, Vargas J NORTH ATLANTIC ENERGY SERVICE CORP. (NAESCO) |
| To: | |
| Shared Package | |
| ML20034F522 | List: |
| References | |
| 92-42, 92-42-R01, 92-42-R1, NUDOCS 9303030358 | |
| Download: ML20034F531 (8) | |
Text
. _ _ _ _ _ _ _ _ _ _ _
r PRA EVALUATION:
CHANGE IN PRIMARY COMPONENT COOLING TECH SPEC 3.7.7 4
J' Engineering Evaluation 92-42 Rev. I g-February 1993 Prepared by
_ld r A 1 7
. L Kiper
()
/ '/
(
Reviewed by
/S Z/,9 f 3
. W. Rau Approved by
~f/[ b Mr[rJ J. %
as
/
9303030358 930226 PDR ADOCK 05000443 PDR p
i l
l PRA EVALUATION: CHANGE IN PRIMARY COMPONENT COOLING TECH SPEC 3.7.3 1.0 Introduction This evaluation documents the change in operational risk, at the system level (system availability) and at the plant level (core damage frequency), for a proposed change in the Allowed Outage Times (AOTs) for the Primary Component Cooling (PCC) System.
1
2.0 Background
The current Primary Component Cooling Tech Spec (TS 3.7.3) applies AOTs to all four PCC pumps. These pumps are each 100% capacity and provide redundancy for each train. Thus, to I
define operability, one train of PCC must contain one PCC pump and the associated flow paths to the PCC loads. On that basis, a new TS has been proposed, summarized in the table below:
l Allowed Outage Time l
Components Current TS Proposed TS Inoperable 3.7.3 3.7.3 1 PCC pump 7d N/A 2 PCC pumps, 72 hr N/A opposite loop 2 PCC pumps, 24 hr 72 hr same loop One loop not explicit 72 hr (See proposed Tech Spec 3.7.3, Attachment 1). The change for PCC pumps brings this Tech Spec in line with the Standard Tech Specs, which have a 72-hour AOT for single PCC train unavailability.
Page 2 CCV242 htK U2/lve93 4
3.0 Discussion This Tech Spec change impacts risk by increasing the likelihood that a PCC pump would be unavailable due to planned or unplanned maintenance. This change is evaluated by considering the impact on system unavailability (Section 3.1) and on the frequency of shutdown due to loss of one train of PCC (Section 3.2). These impacts are combined in the plant model to produce a delta core damage frequency (Section 3.3).
3.1 PCC System Model The PCC system is included in the current Seabrook PRA. SSPSS-1990. This model includes the PCC pumps, the flow path through the PCC loads, and the associated pump area ventilation. The maintenance contribution to the PCC system model is described below (the " current" model):
then the model with the change in Tech Spec is presented (the "new" model).
(1) Current Maintenance Model This model includes contributions from unplanned maintenance, based on the number of j
pumps, the maintenance frequency, and the maintenance duration, as follows:
PCC pump, for each loop (7-day LCO)
= MNT1 = MNT2 1
= 2 x ZMPOPF x ZMPLSD = 0.00907 l
l where the frequency and duration variables are based on generic data from PLG-0500, as follows:
ZMPOPF = 1.58E-4 (mean)
- Maint. Freq. - operating PCC pumps l
ZMPLSD = 28.7 hr (mean)
- Maint. Duration - pumps,7-day LCO Assumptions in the current model:
Maintenance frequencies and durations are based on generic industry data and not
+
on Seabrook specitic data due to the limited operational data. This data was collected by PLG from a number of nuclear plants for similar equipment and isjudged to be reasonably representative of expected Seabrook experience, No planned maintenance is done on the PCC system during power operation that e
makes a pump inoperable.
No contribution is given to 2 PCC pumps in unplanned maintenance at the same time because of the low likelihood of dual pump failure or failure of the second pump while the first was being repaired.
The maintenance contribution from pumps covers unplanned maintenance from l
other components. Thus, no explicit maintenance contribution is modeled for valves, instrumentation, etc., that would make a loop inoperable. The pump contribution is assumed to dominate maintenance unavailability.
1 Maintenance contribution from failures of pump ventilation is not included because it l
is assumed that remedial action would be taken to keep the PCC system operational.
Maintenance is unrecoverable. This assumption may be very conservative for some maintenance activities where the system can be made operable quickly.
Page 3 CC9242hta 0219<9n c-.
(2) New Maintenance Model 1
A "new" PCC model was developed to account for the proposed changes in Tech Specs. The new model results in longer duration of unplanned maintenance and allows planned maintenance at power. These changes impact the modeling of maintenance as follows:
Unplanned Maintenance:
standby PCC pump in each train (no LCO) e
= MNT1' = MNT2'
= 2 x ZMPOPF x ZMPCCD = 0.0308 where the variables are based on generic data from PLG-0500, as follows:
ZMPCCD = 97.4 hr (mean) - Maint. Duration - PCC pumps, no LCO Other variable
- see current model Assumptions:
The standby PCC pump is repaired in unplanned maintenance with no special priority
- consistent with other pumps with no LCO. This is believed to be conservative; a PCC pump failure would still receive high priority. The variable ZMPCCD was developed from the data variable ZMPNSD in PLG-0500, using generic data for PCC and SW pumps, judged to be more representative of the SW and PCC pumps at Seabrook.
i Planned Maintenancefor the standby PCC pump in each train:
= MNT1(PLANNED) = MNT2(PLANNED)
= 2 x (1/4 yr) x (1 yr/ 8760 hr) x (336 hr) = 0.0192 Assumptions:
Each PCC pump is unavailable due to planned maintenance once every four years for 14 days.
Planned maintenance is done on one pump at a time - no MNT1(PLANNED) X MNT2(PLANNED) terms.
The quantification for the "new" PCC model is in general as follows l
PCC Unavail. = PCCpumps(hardware failure + unplanned maint. + planned maint.) +
common components failure where the terms in italic are the ones affected by the proposed TS change.
(3) Ouantitative Results - Systems Analysis The PCC system configuration is quantified for a several of different boundary conditions.
Boundary conditions are the signals and support systems, external to the PCC system, that impact the system configuration. For example, with loss of offsite power (LOSP), the PCC pumps must restart, presenting a different failure mode - pump fails to start - that is not present when offsite power is available. The important boundary conditions for the PCC Page 4 CC%42Jo st' C4'19/93 m
system are the number of support systems (e.g. AC power ) available, and LOSP.
The PCC system is quantified for four boundary conditions, as follows System Unavailability Maintenance Contribution (Current /New TS)
(Percent of TOTAL)
System Configuration TOTAL (Percent Unplanned Planned i
Change)
Maint.
Maint.
PCC1 Both trains, 2.37E-6 3.2 %
transient [ Normal 2.54E-6 (7.5 %)
10.6 %
6.8 %
system abgnment]
l PCC2 Both trains, LOSP 4.51E-5 2.0 %
[PCC pumps must 4.55E-5 (0.8 %)
6.7 %
4.2 %
restart]
PA1 Single train, 6.13E-4 2.0 %
transient 6.47E-4 (5.5 %)
6.7 %
4.3 %
PA2 Single train, LOSP 1.64E-3 2.6 %
1.77E-3 (8.0 %)
8.4 %
5.2 %
l The results at the system level indicate that the change in Tech Specs is not significant to system unavailability. Maintenance is a relatively small contribution to total system l
unavailability. Thus, increasing this small contribution results in small changes in system unavailability. This is due to the importance of non-maintenance related failure modes such as common cause pump failure, isolation valves failure to close on demand, etc. The impact of these changes in system unavailability are integrated into the plant model to estimate the change in risk.
3.2 initiating Event Frequency Loss of either train of PCC would affect the plant power generation through loss of cooling to the RCPs. The frequency of loss of one PCC train is given by the frequency ofloss of a PCC l
pump over one year of operation and failure of the other pump while the first is being repaired.
This also includes failure of the operating pump while the standby pump is out for maintenance
- either planned or unplanned. (There are also other combinations of valves, heat exchangers, l
etc. that could fail and contribute to loss of the train: however, they are not affected by this TS change.)
The equation for loss of one PCC train can be written as follows:
l L1PCC = F(PA)*T(yr)
- F(PC)*T(repair) + F(PA)*T(yr)
- M(PC) where:
l F(PA) = failure rate for operating PCC pump (A) to continue to run, F(PC) = failure rate for standby PCC pump (C) to start and run while pump A is being repaired.
Page 5 enwnx arm
T(yr) = duration the operating PCC pump must run = 8760 hr per yr
- 0.70, plant availability factor, T(repair) = duration of unplanned maintenance on failed pump A, 1
M(PC) = pump C unavailability due to planned and unplanned maintenance.
The last two term are the ones that change due to the new TS AOT as follows:
l l
Current TS Model New TS Model T(repair)
ZMPLSD ZMPCCD M(PC)
Planned Maint.
none 2*(1/4)*(1/8760)*336 Unplanned Maint.
ZMPOPF*ZMPLSD ZMPOPF*ZMPCCD l
where the variables are defined earlier.
The results are given below:
L1PCC Current TS Model New TS Model TOTAL 2.94E-3 per yr 4.74E-3 per yr Planned Maint. Contrib.
0.0 %
1.5 %
Unplanned Maint. Cont.
0.8 %
2.6 %
Thus, the initiator frequency increases by about a factor of 1.6. This is a significant increase despite the fact that maintenance is a small contributor to the total. The increase is due l
primarily to the increase in the repair time, which does not show up directly as a maintenance contribution.
3.3 Quantitative Results - Core Damage Frequency As a result of the change in TS, the CDF (mean) changes by about 4.1E-6 per year, or 4.4 %
increase in the total (9.4E-5/yr). This change is due to the following impacts:
system unavailability (1.7 %)
e initiating event frequency (2.7 %)
e This is a change in the mean value from 9.4E-5 to 9.8E-5, compared to the range of the CDF distribution, which is approximately one order of magnitude (from 5th to 95th percentile).
Thus, this is an insignificant change within the uncertainty bounds on the CDF distribution.
4.0 Conclusion As a result of the quantitative evaluation above, the effect of the changes proposed for TS 3.7.3 is generally small for the PCC system unavailability and is significant for the PCC initiating event Page 6 CC924Jhut 02c a 9'93
i frequency. However, with these changes in the plant model, the overall result is insigniricant to the core damage frequency. This evaluation is based on a best estimate of planned and I
unplanned PCC pump maintenance.
The evaluation does not include the positive contributions due to removing the major PCC pump maintenance activities from outages. These contributions include reducing the unavailability of PCC pumps during outages and permitting more flexibility in outage planning. The outage effects are very sensitive to the configuration of the primary system, time after shutdown, other systems unavailable, etc. and thus are difficult to estimate. The evaluation includes a slight l
increased risk from fire in this area as a result of additional maintenance at power. Both the SSPSA and the IPEEE Report showed this area to be important because on the location of all four PCC pumps. It is important that, during any maintenance at power, the combustible control program be strictly adhered to in order to limit this risk.
As a result, the proposed Tech Spec change does not increase the core damage risk within the bounds of the uncertainty.
l l
i l
l l
l Page 7 CC9242.DUC 02/2%'93 i
i
.. r t --
c..
2 NMCSO TCca h% [
t I
PLANT SYSTEMS 3/4.7.3 PRIMARY COMPONENT COOLING WATER SYSTEM LIMITING CONDITION FOR OPERATION 1
3.7.3 At least two independent primary component cooling water loops shall be OPERABLE, including OPERABLE pumps in each loop.
cne t
APPLICABILITY:
MODES 1, 2, 3, and 4.
i
/$2.hcuh ACTION:
cep With one primary com onent cooling water (PCCW).
i -- rable, a.
L to OPERABLE restore the require primary component cooling water status within 7 d;f; or be in at least HOT STANDBY within the next l
6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> and in COLD SHUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
f
-b.
uith two pri==ry enmnnnent cc; ling -atic pump; 'n;pcr:ble, re; tere Gt ledai One vi luc inupci dule pi ouns y COEpGnent Cooling atii psTpS to CD{ Dant _F ctatuc wi+hin 72 h;;r; Or b; ir a,
}cabL MOI 5IAiiCE'i wi+ sin +ha novt A sneer nd i,, ccLg 3Hgiggg;,,y,,n y,r,;;,,3m; 30 DOurS.
,u,ut ;;;74,7 m,,or nnmns witnin gn,30g3 c
uit3 ;;;
p,m,,,,
incpcr:ble, eac+ cee :: 1::;t one af it.e invgcc501 emEp; tc OcED22LE i
in oc ieast nDi 5i AiiDEi w iin in t r. : ext status ui+hin-20 hours c. uc 6 heues and in COLD SudiDUwn witnin tne roiiuwioy 30 5:ur;.
i SURVEILLANCE REQUIREMENTS 4.7.3 At least two primary component cooling water loops shall be demonstrated OPERABLE:
At least once per 31 days by verifying that each valve (manual, a.
power-operated, or automatic) servicing safety-related equipment that is not locked, sealed, or otherwise secured in position is in its correct position; and b.
At least once per 18 months during shutdown, by verifying that:
1)
Each automatic valve servicing safety-related equipment actuates to its correct position on its associated Engineered Safety Fea-ture actuation signal, and l
2)
Each of the four primary Component Cooling Water System pumo starts automatically upon loss of or f ailure to start of its redundant pump within the loop.
-SEABROOK - UNIT 1 3/4 7-12
-