ML20248L593
| ML20248L593 | |
| Person / Time | |
|---|---|
| Site: | University of Michigan |
| Issue date: | 06/03/1998 |
| From: | Fleming R MICHIGAN, UNIV. OF, ANN ARBOR, MI |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20248L596 | List: |
| References | |
| NUDOCS 9806110258 | |
| Download: ML20248L593 (6) | |
Text
- _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ - _ - _ _ _ - _ _ _ .
, The University of Michigan Michigan Memorirl Phoenix Project Office of the Director 2301 Bonisteel Boulevard
, . Ann Arbor, Michigan 48109-2100 June 3,1998 Docket 50-2 License R-28 United States Nuclear Regulatory Commission Document Control Desk Washington, D.C. 20555 Re: Report of Non routine Occurrence Hot Demineralized Leak Causes 75 Gallon Loss of Pool Water to Radioactive Liquid Retention Tank System.
Summary of Events l
On May 10,1998 at 0250 reactor operators at the University of Michigan's Ford Nuclear i
Reactor discovered a leak in one of the two resin columns comprising the "B" Hot Demineralized (Hot DI) system. The leak resulted in the loss of approximately 75 gallons of l reactor pool water to the facility's radioactive liquid waste retention tank system (total reactor pool water volume is approximately 48,000 gallons). Hot DI "B" was isolated and 2 Mw reactor operation continued. Safe operation of the reactor was not threatened at any time during this occurrence. No radioactivity was released from the facility as a result of l
this occurrence. This report is being made because the leak resulted from a failure in a component of the primary coolant boundary.
The leak was discovered while the reactor operators were investigating the cause of an unusually large observed decrease in pool water level amounting to one-half (1/2) inch during the prior two hours. The nonnal pool water loss rate is about one and one-half (l-1/2) inches per day.
An automatic shutoff valve that was installed with the most recent Hot DI system upgrade
, was not in service because of past erratic behavior m its flow sensor.
(
l Hot DISystem The FNR Hot DI system consists to two sets of fiberglass reinforced resin tanks containing mixed cation and anion ion exchange resin beads. The purpose of the Hot DI system is to maintain the pH and conductivity the primary coolant water. The tanks are physically located in the basement of the reactor building. The Hot DI system taps off of the primary .
coolant retum piping diverting a small fraction (14 gallons / minute) of primary coolant water to the Hot DI's. Nominal primary coolant flow rate is 1050 gallons / minute. Valves are I installed that allow isolation of the Hot DI system from the primary coolant system. A
! check valve prevents back flow from the pool through the Hot DI retum line. Most of the f
Hot DI system has been switched to Schedule 80 chlorinated polyvinyl chloride (CPVC) >
pipe, fittings and valves. See the two enclosed figures. }
9006110258 990603 PDR ADOCK 05000002 ,
S PDR . \
. .s v v v .
.. 4 During normal operation one set of DI columns will be put on-line when conductivity of the pool water reaches 2.5 pmho/cm. After the conductivity has been brought down to 1.5 pmho/cm, the DI's will be bypassed until conductivity again reaches 2.5 mho/cm.
When the on-line set of columns (either "A" or "B") is no longer effective the other set is put into service. The out-of-service set of columns are restored after a suitable decay i period by retiring the upstream column, moving the downstream column to the upstream j position, and replacing the downstream column with a tank containing fresh ion exchange resms.
I Safety Analysis l Testing conducted on May 18 during the shutdown maintenance period immediately following the event showed that the maximum primary coolant loss rate through a guillotine break in the Hot DI piping is 15.9 gallons / minute. This loss rate would cause a decrease in the water height above the core at the rate of 6.4 inches per hour (one inch of pool height is I equal to 150 gallons of water).
l The water would flow into pits and floor drains in the reactor basement that drain to a pair of sumps. Sump pumps transfer the water to the 3)0 gallon radioactive liquid retention tanks in the Phoenix Memorial Laboratory building. 29 gallons is pumped each time the sump pumps actuate. An alarm also annunciated in the reactor control room when the reactor centrol circuits are energized and either sump pump actuates. The normal frequency l at which the sumps pump out and alarm is about once per eight hour shift.
Reactor Operating A worst case leak in the Hot DI system would cause repeated sump alarms and would quickly be investigated, discovered and isolated by the reactor operators when the reactor is operating. This operator action would take place well before the low pool level alami at ,
i minus five inches or the automatic reactor rundown at minus 12 inches would occur. (Note.
the minus 12 inch level corresponds to the pool water level LSSS of 19 feet above the core, one foot above the safety limit of 18 feet.)
Reactor Secured Two pool level clarms are active when the reactor is secured. A local alarm is actuated at a pool level of about minus seven inches. The second alarm, actuated by the same sensor that gives the pool level reactor mndown signal at minus 12 inches, is sent directly to the alarm panel at the UM Department of Public Safety which is staffed 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> per day.
The worst case Hot D1 leak situation when the reactor is secured would occur if the local alarm actuates just after the security guard has passed through on his rounds. In this situation, knowledge that a problem exists will probably not be known until the minus 12 inch alarm actuates. Given a one hour response time between Public Safety contacting the On-Call Supervisor (OCS) and the OCS isolating the Hot DI system, an additional 6.4 inches would be lost from the pool. The pool level would still be at least 18 feet 5 inches above the core when the leak is isolated.
l
l ..
3 i
Radiation levels at the surface of the pool would not be significantly altered by a drop in l pool level from 20 feet above the core to 18 feet 5 inches. The intensity of direct gamma l radiation from the core would increase by a factor of approximately 60 at the average fission gamma energy of 700 kev. However, the contribution direct core gammas make to the total dose rate at the pool surface is insignificant as compared to the dose due to activated impurities (primarily Na-24) in the pool water. The reactor has operated with the
- pool level lowered by as _much as five inches under special circumstances. No discemible increase in surface radiation level was detected at the time. Normal pool surface radiation l levels are 10 to 15 mR/hr.
Reactor pool water leaking from the Hot DI system would be captured by the facility's radioactive liquid waste system. There are no direct pathways between the basement of the FNR building, where the Hot DI's are located, and unrestricted areas. Therefore, no radioactivity would be released to unrestricted areas.
The conclusion is that a leak from the Hot DI system will have no reactor safety significance.
Possible Root Cause In the course of testing the Hot DI system following this event, it was observed that a partial vacuum could be drawn on the resin columns depending on how the system's valves were operated. This caused an inward flexing of the resin tank. Such flexing may have initiated the crack in the failed resin column.
hnmediate Actions Taken Immediate actions taken were: 1) to isolate and secure Hot DI "B", and 2) to isolate and secure the entire Hot DI system during periods when the reactor was not operating until further testing could be completed.
Long Tenn Actions to Prevent a Recurrence The vendor who supplied the resin columns was contacted. The burst pressure was quoted to be 250 psi. The maximum recommended working pressure was said to be 100 psi. A 100 psi pressure test will be performed on each DI resin column before it is put into service.
A number of tests were performed on the entire Hot DI system on May 18,26 and June 2. l l The entire Hot DI system was isolated and then pressurized and hydrostatically leak tested at 85 psi for more than one hour on May 18. No leaks were found. The system was hydrostatically tested a second time on June 2. One fitting containing a conductivity cell l
cracked at a pressure between 100 and 120 psi. The fitting was replaced by a CPVC pipe nipple and the system was tested again at 120 psi. No further leaks were detected. The maximum operating pressure in the Hot DI system occurs in the piping immediately after the return pump. The Hot DI return pump is normally operated with a back pressure of 50 -
60 psi to prevent cavitation. Maximum pressure will occur when the pumps are running
4 and the Hot DI return isolation valve is shut. Under this condition a pressure of 80 psi was observed on May 18. The June 2 leak test showed the Hot DI system piping is capable of withstanding the highest anticipated operating pressure without failure.
A new auto-shutoff valve and flow sensor using a different design have been procured and installed, but are not yet wired electrically. This will make the Hot DI system self-isolating l in case of a major leak when the installation is complete.
1
- i. Operating Procedure No. 211, titled Ooerating the Hot Demineralized System has been l modified to secure the inlet and outlet pumps prior to changing valve line-ups in the system. This should eliminate unnecessary flexing of the resin tanks.
= -
Ronald F. Fleming, Director Michigan Memorial - Phoenix Project [
ene.: as stated l
l cc: Theodore Michaels, USNRC Project Manager Thomas Burdick, USNRC Region III l
FILTER
- BYPASS a
- RTDI FILTERS POOL CHECK '
POOL- .
_a CONDUCTIVITY BYPASS VALVE ,
\
FILL '
f ' '
i -
- l ,4 . .n ,- INLET v
~
, : ~vq s .
j \
AUTO C 1 Z -2 SHUTOFF 3
,. , - - , VALVE 7 ,
i i i. --
l h 1 -. ;
Bl B -ilNLET'-
f j PUMPL 1 1
, ,, 9 1 p, ,
3%g * ,s m
i ._
h .,1
A
, "A" and "B" A
g CONDUCTIVITY CHECK 'j _._
7 N
VALVE
" I ' 7 1
LOCAL % ' ,, t-1 DI MET LOW CONDUCTIVITY M; -
i
,3 MANIFOLD THROTTLING ' '
PRESSURE INDICATOR VALVE TRIP
\ i 1
,A!! A
% 1 l' 0-100# y I ;
9 (n> '
RETURN a "g-TO - !' " ~
~
~~~
'l
-' I POOL V --
- 7 ,. lll OUTLET PUMP
._ ._/.
01 A l TANK :
FILL i N
,_. gg P3 PING CE $L ARRANGEMENT
. - .c - oo e m o- o ey e eve seuesay saucas Revision 17: 091594 l , Chapter 7 Hot Domineralizer System l
Figure 7.1 1 Hot Demineralized System 1
l l
l
-- 1
\
M l
w RX l
LJ l Q n A - Cod .......
, r ,,"""1.
J6 d L 1
$ t e en o u -
_ o ._ . n , m
'M m44," n
,o ,
W "q) M
....., .e
.aens .,.b ., ,>h
.nn ......
... > nn ' ,
n.a >en,..
nn n.n Hautan an
- ~
J.,.
'>, Hes os i
n t99 nn an ',.2 A avoi m n,.,,
i Page 7-3