ML19309F124

RO 80-02:on 800305,low Flow Scram of Flow Recorder FRA-D1 Resulted in Reactor Trip & Subsequent Shutdown.Caused by Broken Teflon Gasket Pieces Restricting Flow.Gasket Repaired & Sys Refilled

ML19309F124
Person / Time
Site:	Neely Research Reactor
Issue date:	04/14/1980
From:	Kirkland R Neely Research Reactor, ATLANTA, GA
To:
References
NUDOCS 8004280512
Download: ML19309F124 (2)

Text

• % ?O q 4

GEORGI A INSTITUTE OF TECHNOLOGY SCHOOL OF NUCLEAR ENGINEERING ATLANTA, GEORGI A 30332

' / ." 16 r.!0 : 5 3 FR AN M H. N EE LY NUCLE AR RESE ARCH CENTER TELEPHONE: (404) 894 3400 April 14, 1980 U.S. Nuclear Regulatory Comission, Fagion II 101 Marietta Street, N.W. Suite 3100 Atlanta, Georgia 30303

Subject:

Licensee Event Report LER 80-2 Gentlemen:

Reference:

Docket 50-160 License R-97 At 1157 hours0.0134 days <br />0.321 hours <br />0.00191 weeks <br />4.402385e-4 months <br /> on March 5,1980 the Georgia Tech Research Reactor (GTRR) was shutdown frcm 1107 by a " Low2 D 0 Flow" scram. his trip came from flow recorder FRA-Dl. Observation of the recorder trace indicated a step decrease in flow from 1760 gallons per minute (gpm) to 1720 Epm.

The trip point setting is specified to be 1625 gpm.

An extensive investigation was begun to detemine the cause of the flow decrease. On March 6-7, 1980 boroscopic and indirect visual examinations were made inside the reactor vessel, particularily in the vicinity of the reactor outlet nozzle. No problems were noted. On March 10, 1980, the main coolant pump MD2A was operated several times while observers stationed themselves in the process equipment room. It was detemined that system pressures were generally greater than nomal, particularly the pressure between heat exchangers HX-D1 and HX-D2 which was 3-4 psi above normal.

During each startup of the punp, metallic-sounding noises were heard from the channel head of HX-D2. Additionally, the system m ssure gages were

, observed to fluctuate full scale during each startup, indication a pressure e pulse somewhere in the system.

Based upon these observations and with the concurrence of the Nuclear Safe-guards Comittee, the primry system was drained and the channel cover on HX-D2 was re::oved. Bree pieces of Teflon gasket material were found in the channel. he area of these pieces was estimated to be 36 sq. inches.

Be gasket mterial was thought to be from HX-D1 so the channel cover of this heat exchanger was removed. me gasket for Lhis cover was a full-face type and it aas founf to be damaged with missing pieces in the outlet quadrant.

Additionally, the horizontal channel separater was seen to be deflected downward, ie. toward the inlet line, appmximately 11/4 inch. ()1

//o 0FFICLE COPY h_ . . =

'h00418g

a .

s .w U.S. Nuclcar Regulatory Cormtission, Region II Page 2 April 14, 1980 Two additional pieces of Teflon gasket material were removed from the pipe ccnnecting HX-D1 and HX-D2. 'Ihese accounted for all the missing gasket material. Compression marks on the gasket for HX-D1 suggested that the deflection of the partition had occurred some time ago, perhaps in 1974 when the quick closing isolation valves were first installed and tested.

We believe that the pieces of Teflon Epsket broken off from HX-D1 were carried into HX-D2. During hydraulic startup, the pieces would be randomly impacted on the tubes of HX-D2, causing a pressure pulse in the D 20 system.

'Ihe Teflon pieces would act to restrict the flow somewhat through the system.

When the trip point from the flow recorder FRA-D1 was checked, it was found that the sealed mercur/ contact switch that actuated the trip had loosened in its holder causing the trip to occur at approximately 1690 gpm rather than the 1625 gpm set point. The loose switch was corrected.

Repair was effected by using two hydraulic jacks to nove the partition plate back to approximately its original position. A new Teflon gasket was made with cutouts and ribs for the partition plate. On March 21, 1980, the system was refilled and on March 26, 1980, the prirnry coolant pump was started.

Flow and system pressure readings were nomal.

On April 4,1980, during hydraulic startup, a very loud " buzzing" noise was heard frcm the channel section of HX-Dl. 'Ihe channel cover was removed and it was seen that a portion of the new gasket had extruded between the par-tition plate and the channel cover. 'Ihe noise was caused by vibration of the extruded section of gasket. We believe this occurred during hydraulic start-up as the pressure " lifted" the cover slightly from the channel. On April 8, 1980, a new, 1/8" thick full face Tef1cn gasket was installed on HX-D1. The main cool-ant pump was started and the system throughly checked. Flow was nonml, pressures nomal and no unusual sounds were noted.

We believe that the problem has now been corrected. If you have any questions about this matter please advise me.

Sincerely rours,

-1 .

Robert S. Kir d Associate Dimetor RSK:lm cc: Members, Nuclear Safeguards Conmittee L.D. McDowell (2)

L.E. Weaver 1

i

. . . . . _ . _ . _ . _ _ . _ _