ML20010C824
| ML20010C824 | |
| Person / Time | |
|---|---|
| Issue date: | 02/19/1981 |
| From: | Manning F NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Murphy F NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| NUDOCS 8108210053 | |
| Download: ML20010C824 (2) | |
Text
.
/
.s GLswmm~Gk q
sww s
}-.
y y
y gu7
., ~
1 I N R P.D R C,0 N et NEG-MREP j
~
February 19, 1981 g-KOLLtc M i
NOTE T0: Joe A. Murphy l
Division of Systems & Reliability Research, RES FROM:
Fred Manning Division of Systems & Reliability Research, RES
SUBJECT:
COMMENTS ON SYSTEM FAULT TREES SUBMITTED WITH IREP PHASE II SECOND STATUS REPORT FOR MILLSTONE POINT-1. DATED 01/15/81 1.
Core Spray Fault Tree
}
i (a) The basic event tabulation list table should be mylified to provide a'ddi-l tional information needed for fault tree evaluation and checking.
It i
should include columns for the failure rate used, the fault duration time, the calculated unavailability or unreliability, and the agiected j
range or variation of the calculation.
l (b) On system P&ID sketch, show which elements are common to both the A & B trains, e.g., does B train have separate piping all the way from the CST?
(c) On sheet CS-5 t;ie CS Pump 1401A failure to start logic is shown.
For j
completeness, show logic failures which give automatic or inadvertent stop signals.
(d) Should have more consistent use of symbols and teminology. For example, the P&ID shows valve CS-2A is normally open but FT shatt CS-9 has a gate i
showing CS-2A fatis to open on demand.
If the valve is' nomally closed, j
show it as such, othemise relabel the gate as a " confirmatory open" j
command.
)
(e) On FT sheet CS-10 and CS-47 there is nothing shown to indicate possible fault coupling between trains A & B and other systems.
For example, if i
Also, the water source (torus water) y might likely also clog on train B.
strainers are clogged on train A, the quantity and condition (e.g., its i
temperature) could have a comon effect on this and other trains or systems usin; the same source. Should include on sheets CS-10 and CS-47 1
(and other pertinent FTs) identical events labeled to indicate potential pump failures from cavitation because of condition of the common water 1
source.
l 8108210053 81';219 j
PDR REVGP NRCNREP P
"'c'>
i " "^*'>
OATEk..............;..............
- NRC FORM 318110<80l NRCM O240 OFFICIAL RECORD COPY
' " *2"2*
m j
f' ow. n.
no. r..,
- c.-
2.
Shutdown Cooling Fault Tree (a) Sheet SD-1 has inaccurate gate label. "No flow through SD Heat Exchanger B" does not result from "no cooling in SC Heat Exchanger B."
Also, this same gate should also be further developed to show loss of fluid in H.E.
through leakage or blockage.
(b) The P&ID sketch doesn't indicate that the Heat Exchansers can be separately fed from RBCCW A or B as sheet SD-1 indicates.
(c) The P&ID sketch shows that the pumps and Heat Exchangers are both cooled by RBCCW. It appears, therefore, that on sheet SD-1 the diamond event "no cooling in SD Heat Exchanger A" should be ti.e same as event on sheet SD-16, "RBCCW fails to cool H.O. SDL pump A."
Similarly for B loop components.
3.
Automatic Pressure Relief Fault Tree This tree is incomplete as submitted in that it has neither a P&ID sketch or electrical schematic to allow evaluating the fault tree and fault tree events--
hence it wasn't evaluated.
4.
Low Pressure Coolant Injection Fault Tree Not evaluated--xerox copies are not legible.
5.
Isolation Condenser Makeup Fault Tree I suggest the top fault tree event should be changed to read "IC Makeup System Inoperable." Interability would then be system operatien or parameters outside the tech spec requirements. The failure to maintain (or replenish) water level iould then be analyzed as a sub-mode en the fault tree.
6.
Genaral Comment for all Fault Trees Many of the ESFAS sensors (e.g., Reactor Water Level, Reactor Pressure, Drywell Pre:::,ure, etc.) are common to several safety systems. However, the fault trees do not indicate this by references or flag notes.
Even though these common sensors should eventually be tied together by common event names, we aren't assured of this by the fault trees submitted thus far. A similar contarn applies to the various AC and DC buses common to several systems.
/H Distribution :
Fred Hanning Subject /
Division of Systems & Reliability DSRR Reading Research CIRC Office of Nuclear Regulatory Research CHRON gNarningReading eme r ),,,RES/DSRR sua~4m) Manr.ing:psr carr) 02/19/81 nac rosu aisno om nacu om OFFICIAL RECORD COPY
"" ' * ** 3 2 * " 2 '