ML19269B831
| ML19269B831 | |
| Person / Time | |
|---|---|
| Site: | MIT Nuclear Research Reactor |
| Issue date: | 01/12/1979 |
| From: | Lisa Clark MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE |
| To: | Grier B NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| References | |
| NUDOCS 7901170244 | |
| Download: ML19269B831 (5) | |
Text
.
\\
\\
t y
f n' n,,
i,4 - p !
1 3
, / {p'$SL#-(*
frA>
NUCLEAR REACTOR LABORATORY l
I cg
'%,h g #-d?
(
l '@
MASSACHUSETTS INSTITUTE OF TECHNOLOGY oK H ARLIP4G 138 Albany Street C3mDhdg0. Mass. 02133 L CLARK.JR D rec tr.r (617)253-4202 Director cf Reactor operations 12 January 1979 Mr. Boyce H. Grier, Director U.S. Nuclear Regulatory Commission Region #1 631 Park Avenue King of Prussia, PA 19406 Attention:
Mr. E. Stetka, Reactor Inspector Re: Reportable Occurrence 50-20/79-1, License R-37 Non-Conservative Nuclear Safety Channel Scram Trip Settings Gentlemen:
Massachusetts Institute of Technology hereby submits this 10 day report of an occurrence at the MIT Research Reactor in accordance with Paragraphs 7.13.2d, 1.15.1 and 3 7.1 of the Technical Specifications.
An initial report was made by telephone to Region #1 on January h, 1979 The format and content of this report is based on Regulatory Guide 1.16.
1.
Report No. 50-20/1979-1 2.
a) Report Date: January 12, 1979 b) Date of Occurrence: December 26-30, 1978 3.
Facility: MIT Research Reactor 138 Albany Street C bridge, Massachusetts 02139 h.
Identification of Occurrence:
Operation of the reactor between December 26 and December 30, 1978 with nuclear safety system settings for at least one, and possibly two, of the three neutron flux level scrams in excess of the limiting safety system settings (LSSS).
5 Conditions Prior to Occurrence: The reactor had undergone a normal weekend shutdown, and the startup check sheets for the following week were completed on December 26, 1978.
0 0490 t
700137opfi/.e
.s Mr. Boyce H. Grier 12 January 1979 Page 2 Reportable Occurrence 50-20/79-1 6.
Description of the Occurrence: On January 3, 1979, during operation of the reactor at low power preceding a full power startup it was dis-covered that the neutron flux level scram trip points, although within the LSSS, had apparently been incorrectly set in a non-conservative manner.
Startup checklists had been completed at 0113 that morning and the reactor had been brought to a nominal power of 2.5 MW at 0201 for 24 ninutes followed by a scheduled reactor shutdown. Later that morning, a normal reactor startup was begun at 0904 with the intent of raising reactor power to 4.9 MW for a standard operating veck. As power reached 2.5 MW it was noticed by the supervisor and operator that the scram setting for neutron flux level channel #6 was set too low and, judging from its indicated output, the reactor would scram on channel #6 high flux at approximately 3.3 FN.
Reactor power was held at 2.5 MW and data from channel h, 5 and 6 outputs logged.
It was also noticed that the scram setting for neutron flux level channels 4 and 5 seemed to indicate they would scram at power levels somewhat above the desired set point (5.5 MW).
The reactor was then shut down at 1008.
Following the reactor shutdown, the calculation and settings of the scram amplifier trips on the startup checklist were reviewed and tested using a battery powered signal generator to simulate detector output. A problem was found in the operation of the signal generator and corrected.
The appropriate flux level trips were then tested, found to be set slightly high for channels 4 and 5 (but lower than the LSSS), reset and tested again at correct value in accordance with checklist requirements.
The reactor was then returned to power by a normal reactor startup commencing at 1037, followed by a slow gradual power increase designed to verify operability of the channels.
Careful analysis of the operating logs and startup checklists for previous weeks showed that the reactor had been operated unconservatively during the period December 26-30, in that one of the three channels was set higher than the LSSS (channel #5, 6.3 MW) and a second was borderline (channel #6, 6.0 MW).
Channel #4 was at 5.7 MW.
7 Description of Apparent Cause of Occurrence:
All neutron level channels are given a relative calibration check each time the startup checklist is performed.' A trip point is determined (corresponding to 5.5 FM thermal power) from the chamber output current of the previous operating period, in order to compensate for gradual changes due to burnup and compo-nent aging. The trip is adjusted and confirmed operational by inducing a simulated chamber output signal using the above signal generator. The simulated signal current is increased until a scram trip is achieved; the
bb. Boyce H. Grier 12 January 1979 Page 3 Reportable Occurrence 50-20/79-1 scram set is adjusted slightly (if necessary) to trip at the calibrated level as determined above, it is rechecked and the results legged.
Prior to start up on December 19, 1978 the scram trips were being verified when it was found, after checking channels 5 and 6, that the battery power was too lov to correctly trip the #h scram amp (highest input currer ').
An electronics technician increased the maximum output current of the signal generator. Channel h was then tested and noted to trip slightly early (input current 4 - 6 microamps below trip) but, as this was conservative, the trip was left as it was.
It was later discovered that increasing the output of the test signal generator caused a loading effect that led to an "early scram" on each channel when the input was 4 - 6 microamps below the trip point setting.
In the startup checklist of December 26, 1978 and January 2, 1979 credit was taken by the operator for this new offset when he adjusted the trip points up scale 4 - 6 microamps in order to achieve a trip, with simulated chamber output at the calculated 5 5 LM value. Since, during normal operation, the generator is not connected to the scram amps, the actual power level for the scram would correspond to the detector input equal to (not 6 microamps less than) the new, somewhat higher trip settings corresponding to actual power levels at scram point in excess of 5.5 FN.
Following the completion of the start up checklists on January 2, 1979, a resistor shunt on channel 6 input developed an open circuit. This open shunt increased the amount of detector output current seen by the scram amp at any power level, thus lowering the power at which it would scram with the same scram setting, and was responsible for the problem noticed on January 3, 1979, when the channel #6 scram point was found to be too lov (3.3 MW).
8.
Analysis of Occurrence: The check sheet calls for a set point of 5.5 MW to insure that the setpoint is less than 6 FM as required by the Technical Specifications. During the week in question, the set points, as stated above, were:
channel #4 - 5 7 tr. channel #5 - 6.3 MW, channel #6 -
6.0 MW.
A reasonable uncertainty for these set points is 1 0.1 MW, so it cannot be elab.ed that two of the three were set at or below 6 MW,as required by the table in Technical Specification 3.7 The safety level channels use a one-out-of-three logic; and so, if power had increased, the reactor would have been scrammed by channel #h at 5.7 FM.
Even if the most conservative channel had not scrammed the reactor at its set point, and reactor power had reached 6 MW, no damage to the reactor
\\
Mr. Boyce H. Grier 12 January 1979 Page 4 Reportable Occurrence 50-20/79-1 would have occurred. Conservative calculations have indicated that if the reactor were operating at the LSSS trip setting for flow and outlet temperature, the maximum fuel plate wall temperature at 6 MW would be less than that required to achieve incipient nucleate boiling.
In addition, the trip settings for flow and outlet temperature are conservatively set (1850 GPM and 55 C) in relation to the limiting settings for these parameters (1800 GPM and 60 C).
0 9
Corrective Actions: The immediate corrective actions consisted of a.
Lowering reactor power (shutdown).
b.
Correcting the signal generator loading effect.
c.
Resetting the scram trips.
d.
Trip testing the new setpoints as per startup checklist.
A return to power in incremental steps to verify correct c.
scram settings.
The steps taken to prevent a recurrence of this or similar problems are as follows:
The weekly startup procedures will be modified to add a step a.
for confirming that the maximum difference between the test input and the setpoint at the time of scram is less than 2 microamps.
b.
The occurrence and potential corrective actions were thoroughly reviewed at a meeting of reactor supervisors on January 3,1979 This report will be read and initialed by all operations personnel c.
to emphasize that they must be more aware of abnormal conditions in plant and should suspect and investigate a problem when something such as a scram setpoint shows a marked variation from normal.
d.
A better test signal generator is being investigated.
\\
Mr. Boyce H. Grier 12 January 1979 Page 5 Reportable Occurrence 50-20/79-1 10.
Failure Data: A similar problem of this nature occurred at MITR-I on February 2, 1973; however, at that time the error was made in the method of calculating the trip point, not in setting it, and has no direct impact on this occurrence.
Very truly yours, L -ul a 4 :.
Lincoln Clark, Jr.
LC:mos cc: MITRSC USNRC-0MIPC