ML20029A754
| ML20029A754 | |
| Person / Time | |
|---|---|
| Site: | Oregon State University |
| Issue date: | 02/25/1991 |
| From: | Andrea Johnson Oregon State University, CORVALLIS, OR |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9103040213 | |
| Download: ML20029A754 (6) | |
Text
__ _-__ -__
- RAIdNiloN CINJ1 R Oni'cos Sirri: UNivi:nsity sw.am cento Aim n.m.n.. ong.n v7m wa Tekthsine SO 737 2VI Vu 3 o 737 WU)
Febo ary 25,1991 Document Control Desk Ur Nuclear Regulatory Commission Wasnington. D.C. 20555
Subject:
Oregon State University TRIGA Ronctor (OSTR), Licenso No. H 106, Docket No.
50-243; written report filed as a follow up to totophono converr.ations with USNRC staff fogarding en event which occtured on February 15,1991.
Gentlemon:
On February 15,1991, Dr. Brian Dodd, Reactor Administrator for the OSTR, contacted the USNRC's Region V of fico by tolophono to discuss a situation which occurred earlier on the samo dato. This written toport is bolng filed as a follow up to the above referenced telephone notiheution in keeping with the OSTR's policy of open communication with the NRC on matters which we believe to be of mutualinterest. In addition to notification of the Region V offico, the event was reported by tolophone to the Oregon Departmord of Energy and was reported verbally to the Chaitman of the OSTR Reactor Operations Cornmittoo and to the Oregon Stato University Hadiation Safety Officer Those latter notifications were also modo in a timely manner on February 15, 1991.
A description of the event itself, including our analysis of its causo; correctivo actions taken; measures implomonted or planned to provent or reduce the possibility of a rooccurrorico; lessons learnod; and specific conclusions ro0arding the ovent are included as part of this report, Tho information submitted in this report, including the conectivo and l proventativo actions, has boon reviewed and approved by the OSTR Reactor Operat. ions Committee.
D A_CKOILQl)RD INFORMATION in January of 1931, the OSTR staff decided to add a water rotontion and recirculation system to the reactor water purification loop. One major purpose of this now system was to provido a method for collecting water used during ion exchange resin changes so that this wotor could be retained and subsequently used as makeup water for the primary coolant. An additional featuro of the now system was its ability to allow the usa of distilled water for makeup water and thereby increase the lifetimo of the ion l axchan00 roains.
c.s 103040213 Y 102E' (-
USNRC 2 February 25,1991 Although the procoding now capabilities woro consider d desirable from on operational standpoint, the basic motivation boM the addition of the now systorn was the f ar,t that its use would virtually eliminato the dist,harge of any nater from the reactor f acility. This was viewod as a positivo action and very much in 1, coping with the OSTR's commitment to the ALARA concept, even though the amount of radioactivity rinnually discharged via water into the sanitary sewer is extremely omall.
The addition of the now water retention and recirculation system was reviewed and approved in a 10 CFR 50.59 safety evaluation dated January 14,1991. This evaluation dotormined that the now system did not creato an unroviewed safety question not did it require a chango to the OSTR license or technical specifications.
DESCRIPTlON OF_ OCCURRENCE AND_PSESIBLE CAllELE At approximately 8:10 a.m. on February 15,1991, the Reactor Supervisor noticot that the water lovelin the reactor tank was lower than normal (i.e., down about 6 to 10 inchos). Finding no obvious causo f or the low water lovel from his !! cation on the reactor top, the Reactor Supervisor proceeded to the first floor of the reactor bay and ontored the adjacent heat oxchanger room whcro the majority of the pumps and piping for the reactor cooling system are located. Upon entering the heat exchanger room, a small stream of wMor was observed to be running out from undet the concrete block shield which surrounds the domineralizer tank. The wrter was flowing to a floor drain in the center of the heat exchangor room. The Reactor Supervisor then looked down insido the concreto block shield and noticed that eno of the now PVC valves associated with the water fotontion and recirculation system had cracked near the point where the valve was scrowed into the dominoralizer tank. Water wrJ leaking through the crack onto the floor.
Upon making this observation, the Reactor Supervisor turned of f two nearby valvos in the water purification loop and immodTtely stopped the leak. Shortly thoroaf ter, the Reactor Supervisor notified the Reactor Administrator and the Senior Health Physicist, and a series of activities woro undertaken to assess the cause and impact of the event, to initiato required notifications, and to formulato appropriato correctivo and proventativo actions, in koopin0 with this action plan, a rcactor water sample was immediately taken and analyzed to determino the radioactiviy concontration of the water reioased down the drain.
The routing of the drain, which was generally believed to go to the f acility's liquid holdup tank, was double-chocked and it was determined that the drain apparently went directly to the sanitary sowar, Tho water levelin the tank was measured to be approximately 101/2 inches below the normal fulllovel and from this it was determined that approximately 200 to 215 gallons of water were released to the sewer.
As a result of the previously mentioned water samp!o,it was determined that the concentration of radincetivo materialin tho water released to the sower was very low (approximately 1.6% of the applicablo limit). This conclusion was based on values and regulations specified in 10 CFR 20, Appendix B, Tablo I, Column 2, which is appUcable to sewer disposel of liquids.
l
4 USNRC February 25,1991 As part of the follow-up assessment, it was recognized that the reactor tank low water level alarm did not function correctly, although this alarm is tested monthly and has always functioned ccrrectly during each test. The water level alarm involves a float mechanism designed to trigger a microswitch for low water and a separate microswitch for high water. The monthly test of those alarms involves manually pushing the float down untilit engages the low water niicroswitch and then raising the float manually until it engages the high water microswitch. No abnormalities have been observed durin0 this monthly testing procedure, and tne Reactor Supervisor volunteered that he remembered the low water level alarm functioning properly when the reactor tank water was intentionally lowered at some previous time.
With respect to why this entire event occurred, first of all it appears that the new valve may have cracked due to stresses induced by piping vibrations together with the weight of the water in the pipes near the valve. These f actors cre believed to have been ef fective in causing the valve to crack because the piping was inadequately supported.
Tests of the reactor tank low water level alarm conducted on February 15 (the day the leak occurred) indicated that the low water level microswitch required slightly more force than originally needed in order to activate the alarm signal. It is not clear why this
.cituation existed.
Af ter the above evaluations, and an evaluation of the OSTR license and technical specifica; ions, an initial telephone notification to the NRC's Region V of fice was made at 11:17 a.n. on February 15. The initial conversation was between Dr. Brian Dodd of OSU and Mr. Dennis Schaef fer of the t' ~.C, This telephone call was followed by several other calls primarily to Mr. Jim Roese at the NRC's Region V office. During the final co<,/ersation witt, tha NRC, Mr. Marvin Mendonca from NRC headquarters was also on the phone. After notifying the NRC, an initial telephone notification was also made to the Oregon Department of Ener0y at apr oximately 11:.1 a.m. During each of those telephone conversations, the details of the event were described and plans for corrective and preventative actions were reviewed to the extent that such actions had boon taken or were definitely planned. On February 19,1991, A. G. Johnson also reviewed the entiro event with Mr. Al Adams, who is the OSTR project manager at NRC headquarters.
CORRECTIVE ACTIONS As indicated previously, immediate corrective action was taken by the Reactor Supervisor who stopped the water leaking from the cracked valve by closing two nearby valves b the water pu..fication system, which in turn isolated the demineralizer tank.
Subsequent to this action, the cracked valve was removed and replaced with a new valve, but the orir ntation of the now valve was reversed so that the heavier portion of the valve body cacting was screwed into the domineralizer tank thus providing more strength at the point where the previous valve cracked.
USNRC February 25,1991 After it was recognized that the low water level alarm did not function correctly, the Scientific Instrument Technician examined the microswitches currently being used and replaced both the low water and the high water micrc. switches with switches having significantly greator sensitivity. The significance of this action was immediately evident in that th i force required to operate the now microswitches was much loss than that previously requ.*ed, and therefore the sensitivity of the low and high water level alarms was greatly incrocod. This was vividly indicated when it was demonstrated that the weight of the float rei alone, without the weight of the float, would easily activue the microswitch for the low water level alarm.
MEASURES TO PREVEl4T THE REOCCURRENCE OF SUCH AN EVENT Any event of this nature is taken very seriously by the OSTR operational staff and management. Therefore, wo have implemented immediato positive action to ensure, to the maximum extent possib!o, that an event auch as the one described in this report will not reoccur. We believe that the following actions, most of which have already been implemented, will be ef fective in achieving this objective.
- 1. As noted, the reactor tank low water level and high water level alarm device has been equipped with more sensitive microswitches.
- 2. Month!y testing of the low and high water level alarm for the reactor tank will continue as ;:resently scheduled using the current testing procedure. However, on an annual frequency (interval not to exceed 15 months) the low water lovel alarm will ba tested by actually lowering the water levelin the reactor tank to a point whero the alarm will activate. As per technical specification 5.7, the alarm will be required to activate before the drop in water level exceeds 6 inches. NOTE: The now water containment system will make it possible to perform this test without discharging any water to the sewer system.
- 3. There are two floor drains in the heat exchanger room. The floor drain which was involved in this event has boon plugged so that it is no longer functioning as a serviceGble drain. This drain was plugged because it tumed out to be the only drain in the reactor bay complex which did no11ead to the liquid holdup tank. As a result,it is now impossiola for l
reactor water to drain directly into the sanitary sewer system.
- 4. A small concrete barrier (dam) is being constructed across an exterior door in the heat exchanger room so that should there be any further water spilled on the floor of this room it will be touted into the floor drain in this room which leads to the liquid holdup tank.
- 5. To reduce stress on the pipes and other components s,f the new water retention and recirculation system, significant additional pipe supports have been added throughout the s;., tem. These supports have eliminated stresses and vibrations due to water moveme' .. rough the system and from the weight of water in the pipes.
\ .,
USNRC February 25,1991
- 6. OSTR operating procedures (OSTROPS 2 and 3) addressing the Reactor Startup Checklist Proceduros and the Reactor Shutdown Chocklist Procedures will be modified to incorportis routino opening and closing of appropriato valvos in the water purification system in order to open and then isolate the domineralizor tank and the now water rotontion and recirculation system from the reactor tank. Isolation of those systems will occur as part of the formal shutdown chocklist and opening will occur as part of the fonnal startup checklist. This action in itself will provent significant water leaka00 duo to an event similar to the one described in this report.
- 7. Two now valvos will bo installed near the reactor tank which will be opened and closed as part of tno startup and shutdown procedures. Af ter installation, those valves will be used in lieu of the valves mentioned in the procedin0 i tem and will enabio isolation of the entire water purification, rotontion, and rocirculation system from the reactor tank.
- 8. The Reactor Administrator conducted an audit of the techniques currentig in use at the OSTR to check and/or test various reactor systems. The purpose of the audit was to assess the validity of proceduros being used to perform tests or checks on systems like the low water level alarm, it was concluded that existing pmcodutos seem to be appropriate for the tests and checks they address.
LEESHMS kiLARNfa
- 1. There appear to be two aspects of this event which might bo helpful to other research reactor facilities. The first of these involves the type of test being performed on water level alarms. Sinco it is probable that few facilities actually lower the water level in their reactor tank to test the low water level alarm,it would follow that most facilities are using some type of manual activation of the alarm circuit. In view of the experience at the OSTR, we would encourago all facilitics to examine the validity of their testin0 process for their water level alarms.
- 2. As a second item,it would soom prudent to recommend that drain paths for all reactor f acility drains be double checked.
CONCLUSIONS Although there woro no radiological or reactor safety factors associated with the event described in this report, the OSTR staff ro0 rots that such an ovent occurred. Our reactor program continues to operato under a policy where safety and compliance with regulatory requirements are of the utmost importanco, and support from the Un;versity's administration makes it clear that they fully underwrite this modo of operation. As a result, we wish to emphasize that we do not take such events lightly and that wo will implomont overy reasonable action to provent a reoccurrence. With respect to this commitment, we would like to summarize for your consideration key factors relatin0 to the previous event.
l
..-g ean a a s __
1 a- x {
., i
.. 4 .
USNRC. 6- February 25,1991 L1. The component which failed in this event was in a new system that had boon i iristalled to virtually eliminate the release of liquids from the OSTR consistent with our
! strong commitment to ALARA. ,
2, The concentration of radioactivo materialin the water released to the sanitary. -!
sewer system was very low and contained only about 1.6% of the applicable NRC concentration limit. Of this 1.6% value,97% of the radioactivity released was tritium. As
.a result of this situation, we believe it is clear that there was no radiological risk.
3 The low water level alarm in the reactor tank had boon successfully tested overy i single month and in fact this alarm appears to have operated correctly in at least one l previous instance when the water lovel in the reactor tank was intentionally loworod. ,
Furthirrnoro, the sensitivity of this alarm has now been significantly increased, j
- 4. - All drains now operabio within the reactor bay and heat exchanger room will carry l water directly to the liquid holdup tank (i.e., thero is no internal pathway for liquid discharge directly into the sanitary sewer),
a
- 5. As specified previously, additional actions ( such as numerous new pipe supports) have been or.will be implemented to prevent a reoccurrence of an event of this typo.
. Should there be questions regarding the information in this report or should you require more information, pleaso lot me know, it is our intent that this report be as- ,
complete and helpful as possible.
Yours' sincerely,.
" .m ,, r
,,-w A.' G. Jo ,,on >
- Director ,
agl'dt/ntc/2 26.91' cc: INRC Region V .
'NRC-Al Adams -
ODOE David Stewart-Smith '
SEB c BD' I JFH1 i
.TVA >
-