ML19306B416
| ML19306B416 | |
| Person / Time | |
|---|---|
| Issue date: | 12/03/1985 |
| From: | Harold Denton, Heltemes C NRC Office for Analysis & Evaluation of Operational Data (AEOD), Office of Nuclear Reactor Regulation |
| To: | James Keppler NRC Office of Inspection & Enforcement (IE Region III) |
| References | |
| NUDOCS 8512230117 | |
| Download: ML19306B416 (4) | |
Text
~
h CA,r w w.%
ym,%
y s<
, -,. m
.,y.
"M?
~'[
'O
'~~t S T
~di f
- - =
ytc 0 3 1585
!!EMORMIDUM FOR: Ja es G. reppler, Regional Administrator Region III FROM:
Harold R. Denton, Director Office of Puclear Reactor Regulation C. J. Heltenes, Jr., Director Office for Analysis and Evaluation of Operational Data SUP. JECT:
FAILURES IN UFPER HEAD If1JECTION SYSTEli This is in response to your nenorandum dated Novenber 28, 1984 (sane subj ect). The enclosure to your nenorandun outlined two concerns by an anonynous individual regardino the upper head injection systen (UHI):
first, failure of the UHI isolation valves to close could lead to injec-tion of the contents of the nitronen accunulator into the reactor coolant system (RCS); and second, the degassing of nitrogen-saturated water from the UHI water accumulator after UHI injection into the reactor vessel.
Both of these concerns were postulated to increase the likelihood of core nelt due to the non-condensible cas.
We have evaluated both of these concerns and concluded that failures of the UHI systen do not represent an " ultra hich risk" as postulated by the anonynous individual. The enclosure provides our evaluation which
)
shows that the risk due to failures of the UHI system is sufficiently low that no innediate action is needed, and that l!HI plants can continue to operate without undue risk to the health and safety of the public.
You should also be aware that licensees with l!H1 plants have inforned the staff that they may nake appilcation to renove the LMI systen fran sone or all of their UHI plants, in particular the two M:Guire Units.
Ve will keep you infomed of the staff activities in this regard.
.&, g' <3 6
c3512230117 951203 SUB J CF l
ep3_a-3-egs
-m
"" 5 >
our>
NRC FORM 3tB t10:80l NRCM C240 OFFICIAL RECORD COPY
- "5 " """ *"'
,--.~--]
w
Yih fj*%
tM "iti'*"'
y
~
V
'TQ T
GW W
t;
)
i James G. Keppler This memorandum should assist you in being responsive to the individual's concerns. If you should need additional information or if we can provide further assistance, please contact Nonnan Lauben at FTS 492-7579 or Wayne Lanning at FTS 492-4433.
6 H. R. Denton, Director Office of Nuclear Reactor Regulation
/5/
l C. J. Hei temes, Jr., Di rector Off:ce for And sis and Evaluation
/
Enclosure:
of Operational Data As Stated Distribution DCS AEDD CF AE00 SF l
ROAB CF l
VLanning KSeyfrit i
TIppolito CHeltemes i
WLauben BSheron WHouston j
RBernero DEisenhut HDenton JSneizek
]
l GHolahan J
, y r
R 11
/
B:DL p/\\ '.
GHo ahan PS.artin, RU 10 /d f/85 JMartin, R V
,j N#
,)
y y
p 4
- MP /d I
R0g&OD C:ROA :AEOD RSB:DSI
- r DSI AD /1,[M'D:DSI/
DD;A v'
WHouston[$
WLa Cing:jz KSeyfrit NLauben 3Sh ron F
RBernero TIp t
02/6 /85 02/.5 /85. G 02/j7./85 02/l3/85 L /j3 /85 2 //J/85
-)- /g/85 MAO
" ' " >.D'.$.......DDj).......
.t h 5"'"""4.C
.te9AS..
.0 a d u u t.....
' Eaton.......
- * " >}. /.N.8,5...
,,h.7,/.S.5......,.d/(.g.5.......
.ae ronu :ne novn wacu o2 o OFFICIAL RECORD COPY c u.s. c >o au-4oom ~
,,,.m
,,.wm wr mm
_ 4 m _.m.m.
-.m~+-~
4 EN?cOSURE Evaluation of UHI Concerns Two concerns were identified by an anonymous individual regarding the upper head injection system (UHl): first, a failure of the UHI isolation valves to close could lead te injection of the contents of the nitrogen accumula-tor into the reactor coolant system (RCS); and second, the degassing of nitrogen-saturated water from the UHI water accumulator af ter UHI injection into the reactor vessel.
Both of these concerns were postulated to increase the likelihood of core melt due to the non-condensible gas.
Each of these concerns is evaluated in the following discussion.
Recent operating experience shows that there have been three events at Westinghouse-designed UHI plants that could be considered precursors for I
introducing nitrogen into the RCS. Two separate events at Sequoyah Units 1 i
and 2 involved a ruptured membrane between the nitrogen and water accumula-to rs. Both units operated for about a year providing the potential i
for the UHI accumulator water to become saturated with nitrogen. The licensee was aware that the membranes were ruptured, and more frequent surveillances of the UHI water for nitrogen concentration were performed.
However, the nitrogen concentration was erroneously calculated to be within technical specifications limits. Thus, had the UH1 actuated i
during a LOCA, the nitrogen from the injected water could have been released into the RCS as a non-condensible gas. The third event at McGuire involved the incorrect replacanent of the differential pressure i
transmitters, which sense water level and provide the initiating signal for isolation values closure. Consequently, the isolation valves would not have automatic'.y closed on low level had a LOCA occurred, and without operator action, the contents of the nitrogen accumulator could have l
been injected into the RCS.
The consequences of injecting nitrogen fran the accumlator into the RCS during a LOCA were part of the staff's review of the UHI design and discussed with the ACRS on several occasions (most recently on November 29, 1984). At the time the generic ECCS evaluation model was approved, neither the staff nor the licensees had the analytical capabil'ty to perform a LOCA analysis which assumed the accumulator nitrogen entered the system and to quantify the effects of this non-condensible ras on the predicted peak cladding tenperature. The safety evaluation (NUREG-0297) found that the UH1 system conformed to the staf f's acceptance criteria (e.g., isolation valves met single failure criteria and multiple failures would have to be postulated in order to inject the nitrogen cove r gas into tne RCS). Technical specifications were imposed for each !HI plant j
to minimize the dissolved nitrogen concentration in the UHI ac(Jmulator water.
Very recently, RES has completed development of the advanced systems performance codes (RELAP5, TRAC-PF1) and these codes have the capability to treat non-condensible gas effects. In the NRR/DSI FY85 technical assistance l
program, Sandia National Laboratory has been contracted to perform LOCA calculations for UHI plants to study the affects of accumulator nitrogen inj ection.
t h
i 5
i
~
y,
,,,m--
r4--
e
I i
As a result of the operating errerience, and in the absence of definitive i
calculations, we have performe. a qualitative assessment to determine what could be the potential corseque.ces if a LOCA occurred and the UH1 accumu-lator nitrogen entered the RCS.
If the nitrogen were injecud following the UH1 accumulator water injec-tion, it would most likely sweep out wa.er residing in the lower plenum.
Thus, at the end of blowdown, less water would remain in the vessel, and a longer period of time would be necessary for the low pressure safety injection pumps to replenish the water in the vessel and reflood the core. Because of this longer time needed to refill the vessel and j
reflood the core, the fuel would be vpected to heat up to higher tempera-l tures before being ref;ooded. Using conservative Appendix K assumptions, a l
LOCA analysis of this scenario would most likely cause the predicted peak cladding temperature to exceed 2200 F.
A realistic calculation of UH1 perfonnance during a LOCA without isola-i tion valve failure is expected to result in peak cladding temperatures i
on the order of 1200 F.
Thus, isolation valve failure and gas injection could raise.the peak cladding temperature in excess of 1200'F. Conservative estimates based on best estimate computer modeling indicate that the cladding temperature could reach about 2300 F.
However gross fuel failure j
or core melt should not result.
{
Estimates of the risk impact were also performed.
In general, break sizes of 3 inches or greater in Westinghouse plaints could result in pres-sure decreases great enough to possibly allow nitrogen injection (nitrogen injection is not expected to occur until the system pressure drops to about 400-500 psi). The probability o greater is somewhere between 10~f breaks jn the range of 3 inches and and 10- per reactor year.
}
Perhaps the most uncertain aspect is the probability of the human errors i
that were necessary at McGuire to cause all of the pressure transmitters to be installed incorrectly and the incorrect installation then allowed togoundetecged.
A typical value for this' type of error used in PRAs is about 10 or less.
If the conservative bounds of the above two LOCA contributions are used, and it is further conservatively assumed that j
all LOCAs with UHI isolation valve failures' result in core melt, then tge core melt estimate for this event is conservatively estimated to be 10'.
The consequence of degassing the nitrogen-saturated water after injection j
is. bounded by the above estimates for isolation valve failures. To further i
minimize the potential for the accumulator water to become saturated, we are evaluating the need to increase the surveillance frequency now required by the technical specifications.
1
- Based on this analysis, we 'do not believe'that failures of the UH1 system
.l represent an " ultra high risk" as postulated by the_ anonymous individual.
l Furthermore, the failure of the UHI system is not a risk " outlier," and i
the risk from the event is sufficiently low that no immediate action i
is needed and.that the UHI plants can continue to be operated without i
undue risk to the health and safety of the public.
i I
I i
i s,~- -, n,
_n,,
..n.,. - -, - c
~
.-..n.-.
_.,,.-.,.- - _