ML20058Q237
| ML20058Q237 | |
| Person / Time | |
|---|---|
| Site: | MIT Nuclear Research Reactor |
| Issue date: | 12/17/1993 |
| From: | Bernard J, Lau E, Newton T MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| 50-20-1993-1, NUDOCS 9312280132 | |
| Download: ML20058Q237 (6) | |
Text
_ _ . .
4 !
I NUCLEAR REACTOR LABORATORY
' AN INTERDEPARTMENTAL CENTER OF MASSACHUSETTS INSTITUTE OF TECHNOLOGY
@W' !
O. K. HARLING 138 Albany Street, Cambridge, Mass. D2139-4296 J. A. BERNARD, JR. l Director Telefax No. (617) 253-7300 - Director of Reactor Operations Telex No. 92-1473-MiT-CAM Tel No. (617) 253-4211/4202 December 17,1993 i
U.S. Nuclear Regulatory Commission Washington, D.C. 20555 ,
Attn: Document Control Desk
Subject:
Reponable Occurrence 50-20/1993-1, Operation with Fewer than the Required -
Number of <100 kW Nuclear Safety System Level Channels Gentlemen:
The Massachusetts Institute of Technology hereby submits this ten-day repon of an occurrence at the MIT Research Reactor (MITR) in accordance with paragraph 7.13.2(d) ,
of the Technical Specifications. An initial report was made by telephone to NRC Headquaners, Mr. Alexander Adams, at about 0945 on December 8,1993. This was followed at 1015 by a more detailed repon that was teleconferenced to include both the
~ NRC Headquarters and NRC Region I (Mr. James H. Joyner and Stephen W. Holmes). '
NRC requested that a repon be also made to the Headquaners Operations Center. This call was placed ~at 1102 to Mr. John B. MacKinnon, the Duty Officer at the Headquarters Operations Center. :
r The format and content of this repon are based on Regulatory Guide 1.16, ;
Revision 1.
- 1. Report No.: 50-20/1993-1 ;
2a. Renon Date: 17 December 1993 I 2b. Date of Occurn nce: December 1993
]
- 3. Facility: MIT Nuclear ReactorLaboratory 138 Albany Street ;
Cambridge,MA 02139 ,
230049 !
i 9312280132 931317 /6 .j l PDR ADOCK 05000020 S PDR 4 /pv j
I Reportable Occunrence 50-20/1993-1
- s Page 2
- 4. Identification of Occurrence:
l Operation of the reactorin natural-circulation mode without the required two -
operable nuclear safety level channels set to scram at or less than 100 kW, MITR Technical Specification No. 3.7. This occurred at 1246 on December 7,1993. ,
- 5. Conditions Prior to Occurrence: l The reactor had been prepared for operation in the natural-circulation mode -
in order to perform an approved experiment. This mode of operation is permitted ;
under MITR Technical Specification No. 2.2. The maximum allowed power level +
is 100 kW and there is no forced flow of coolant thmugh the core. MITR Technical ,
Specification No. 3.7 also pertains to operation in the natural-circulation mode and 1 one ofits requirements is that there be two neutron flux level channels set to scram ,
at or below 100 kW. This is normally accomplished by replacing two of the three high-range safety channel amplifiers with low-range amplifiers that are capable of scramming the reactor at 100 kW or less. This had been done and the reactor instrumentation had been tested in accordance with the "Startup for Less than 100 kW Operation" checklist, PM 3.1.3.
- 6. Descrintion of Occurrence:
Reactor startup began at 1044. To accommodate the expament, power i was raised in steps, reaching 1 kW at 1109,10 kW at 1211, and 50 kW at 1245.
At 1246 it was noticed that power level safety channel #5 was not showing a reading, and the Assistant Superintendent of Reactor Operations and the Electronics -
Technician were notified. The experiment was halted and directions were issued !
to shut the reactor down as soon as the Electronics Technician finished examining -
the channel. While investigating the condition of the channel, the Electronics Technician caused the reactor to scram. This occurred at 1252. (Note: The signal from the scram amplifiers is displayed visually. - However, that signal is not-discernible at 10 kW and, by procedure, the operator monitors channels #7-#9 while maneuvering reactor power. Hence, the detection of the failed amplifier at 1246 was timely.)
The reactor power level was monitored during this startup on three independent channels (channels #7,8, and 9) that are not part of the safety system. 1 The readings from these channels were consistent and showed that reactor power :
never exceeded 50 kW. l
- 7. Descrintion of Anparent Cause of Occ.2rrence:
The MIT Research Reactor's nuclear safety system consists of six independent channels that are designated as channels #1 through #6. Channels
- 1-#3 provide period scrams. These were fully functional during the occurrence reported here. Channels #4-#6 provide power level scrams. The minimum- i required complement of nuclear instrumentation is two period and two level l channels. Monitoring of the reactor power by the console operator is done, as i mentioned above, using three other channels that are not part of the safety system. !
Each level channel consists of a neutron-sensitive ion chamber, a voltage supply, and a scram amplifier together with its power supply. Two types of scram
Reportable Occurrence 50-20/1993-1
, . Page 3 -
t amplifiers are used. For forced-circulation cooling of the core, 'high-range' amplifiers are installed in all three level channels (#4-#6).- These are set to scram at ;
5.5 MW - For natural-circulation cooling, two of the high-range amplifiers are ;
physically removed and, in their place, two ' low-range' amplifiers are inserted. l The result is that, when operating in the natural-circulation mode, channel #4 is a !
high-range amplifier and channels #5 and #6 are low-range amplifiers set to scram at or below 100 kW. .
3 The design of the low-range amplifiers differs fmm that of the high-range j ones. Specifically, the former are actually duat-range units that can be configured !
to either a high or low range application. However, they are only used for low- '
range amplification. The circuitry is similar to that of the high-range units except that each contains a low-range amplifier and a four-segment relay that is activated by a push button located on the amplifier chassis. Once activated, the relay performs several functions. First,it directs the signal from the neutron detector to the low-range amplifier. . Second, it cuts out the scrams on " Low Flow Primary 7 Coolant" and thereby allows operation in the natural-circulation mode. Third, it ,
activates a self-locking feature. The design is fail-safe in that should the push i button not be depressed and should primary flow not be established, the mactor's '
scram cimuitry can not be cleared. Hence, a mactor stanup in the natural-cimulation mode is not possible unless these low-range amplifiers are installed. Also, if these -
amplifiers am removed during reactor operation, a scram will result.
The immediate cause of this occurrence was found to be a faulty four-segment relay in the low-range amplifier of channel #5. It should be noted that the ;
high-range amplifiers, which are used during almost all critical operation of the ;
facility, do not contain this four-segment relay and hence are not subject to the j prookms identified in this occurrence.
An investigation has been conducted as to the root cause of this occurrence ,
and two factors have been identified. The first involves the content of the r preoperational checks. The second is related to the system's design. +
Preoperational checks are required to be performed on the nuclear instrumentation on a regular schedule. These checks, which include the setting of :
the scram trips as well as verification of the chamber and amplifier power supplies, entail transmitting a test signal through the amplifier circuitry. The input for this test signal is at the front of the amplifier at a specially-designed connection. It was found during the investigation of this occurrence that the test signal was not directed !
through one segment of the four-segment relay. It was also found that the entire i circuit could be checked by substituting the test signal for the chamber output at the !
rear connection to the amplifier. (Note A check of the high-range amplifiers ;
confirmed that all of their circuitry is checked upon insertion of a test signal through i the designated connection.) :
A second root cause of this occurrence was that, unlike the situation with !
the three high-range amplifiers, there are only two low-range units. Hence, if either fails during reactor operation, the technical specification on the minimum complement of nuclear instrumentation will be violated. The reason for there being ;
only two low-range amplifiers is that the reactor is operated infrequently in the ,
natural-circulation mode. Nevertheless,it appears prudent that a third low-range i amplifier be installed. His factor had been identified as desirable during a review ;
i
+ ;
' Reportable Occunence 50-20/1993-1 Page 4 q I
i 1'
- of the MITR nuclear safety system that was conducted in' January 1993. The recommendations of that review, which are in the process of being implemented, 1 would achieve this. This is further discussed under the section on " Corrective Action."
- 8. Analysis of Occunence:
This occurrence involved a failure to maintain the minimum complement of ;
safety channels during operation in the natural-circulation mode. No challenge o - ;
the safety system took place during this occurrence. The maximum allowed power :
level was 100 kW and power never exceeded 50 kW. Accurate indications of the .i reactor power were available to the reactor operator at all times via channels #7, #8, .; '
and #9 which were fully functional.
All of the other components of the reactor's safety system that are required for operation in the natural-circulation mode were fully operable. This included the ;
three period scrams channels (#1-#3) and channel #6 which was equipped with a ^i low-range amplifier. In addition, channel #4, which was set to scram at 5.5 MW, ,
was functional and would have caused a scram although it was not subject to the preoperational checks under PM 3.1.3. ;
9
- 9. Corrective Action: ,
The following immediate conective actions were taken:
(a) The reactor was shut down, the experiment was removed, the low-range safety channel amplifiers were replaced with the high-range ones, coolant flow was established, the full-power startup checklists were performed, and ;
the reactor was taken to full power.
(b) The four-segment relay in the channel #5 low-range amplifier was replaced.
At this point, it was felt that all necessary immediate corrective actions had .i been taken. However, as described in the addendun, to this report, such was not ;
the case. A further defective component of the low-range amplifier was found on December 14,1993. This was then replaced.
The corrective actions that will be taken to address the root causes of this problem are as follows:
(i) Modify the preoperational test of the low-range amplifiers so that the tesi !
signal is substituted for the input from the neutron detector. (Action implemented under temporary approval on December 7,' 1993; revised ;
checklists to belssued by January 14,1994.) ,
(ii) Modify PM 3.1.3, "Startup for Less than 100 kW Operation" to make l explicit the requirement to place the low-range amplifiers in the low-range !
mode prior to performing the preoperational checks. (See adds im to this !
report. Action implemented under temporary wroval on December 15,-
, 1993; revised checklists to be issued by Jarrry 14,1994.) >
Reportable Occunence 50i20/1993-1
,- Page 5 !
, j (iii) Consider modifying the existing low-range amplifiers so that they are no l longer dual-range. This would eliminate the need for conective action (ii) above. (Note: This action will be proposed to the MIT Reactor Safeguards Committee on December 21,1993.) .
In addition to the above, we are considering eliminating the need for low-range amplifiers by installing a new, upgraded safety system. Funds have been procured for this effort, candidate equipment has been identified, and safety evaluations of the proposed changes are in progress. However, a final decision on this is not expected for at least six months. !
- 10. Failure Data:
There have been no previous failures rf the low-range amplifiers.
However, there have been several involving the high-range amplifiers. These are ROR #50-20/1979-1 which involved non-conservative setting of the scram amplifier trip points, ROR # 50-20/1979-5 in which a startup was conducted with l fewer than the required number of high-range power level scrams because of a safety scram amplifier circuit failure, and ROR #50-20/1989-1 which occurred when the voltage supply to one amplifier was inadvertently secured.
Sine ely, 4 i p g Svto Thomas H. Newton, Jr.,PE Edward S. Lau, NE Asst. Superintendent for Operations Asst. Superintendent for Engineering MIT Research Reactor MIT Research Reactor
< A 4L b /
John A. Bemard, Ph.D. !
Director of Reactor Operations MITResearch Reactor JAB /gw
Attachment:
Addendum to Reportable Occurrence 50-20/1993-1 cc- MITRSC (with enclosures)
USNRC - Project Manager, NRR/ONDD USNRC - Region I- Chief, FRSSB/DRSS USNRC - Region I- Project Scientist, Effluents Radiation Protection Section (ERPS)
FRSSB/DRSS
~
Addendum to Reponable Occurrence 50-20/1993-1 Page1 of1 l Addendum to Reportable Occurrence 50-20/1993-1 On December 14, 1993, it was decided to repeat the expenment that had been -
suspended on December 7. The reactor was again prepared for operation in the natural- ,
circulation mode and PM 3.1.3, "Startup for Less than 100 kW Operation" was performed ,
with the exception that the low-range amplifiers were tested by substituting a test signal for that of the neutron detector at the rear of each amplifier. A reactor startup was begun at 1511 with power attaining 1 kW at 1539,10 kW at 1642,25 kW at 1651, and 50 kW at 1702. It was again recognized that channel #5 was not responding and the reactor was shut .
down at 1703. The experiment was removed. The full-power startup checklists were then performed and the reactor was restarted with two high-range amplifiers being used to satisfy the technical specification requirements. The channel #5 low-range amplifier was run in parallel with the high-range units. This allowed its observation under actual i operating conditions while not placing any reliance on it as part of the safety system.
Power was increased in steps to 1 MW and it was observed that the channel #5 low-range amplifier did not respond. This confirmed previous analysis that the problem resided in the circuitry and not with the neutron sensor, its power supply, or cable. The reactor was then shut down and funher repairs made to channel #5. The integrated circuit that forms the low-range amplifier was found to have failed and it was therefore replaced. The channel #5 low-range amplifier was again installed in parallel with the high-range amplifiers and the reactor restarted. Power was increased in stcps to 250 kW and channel #5 was observed to function properly. Subsequently, the reactor was shut down, the channel #5 low-range amplifier removed, and reactor restart performed with three high-range amplifiers. ;
The cause of this related occurrence appears to have been human error. ,
Specifically, it will be recalled that the low-range amplifiers are actually dual range and, as best as can be determined, the individual who performed the continuity check of the low-range amplifier's circuitry prior to the startup at 1511 did not place the channel #5 low-range amplifier in the low-range mode. Hence the failed integrated circuit, which is unique to the low-range circuitry, would not have been detected. (Note: Actual operation of the reactor without the low-range amplifiers being in the low-range mode is impossible because of the interlock described in this report.)
7 The information contained in this addendum was reported to NRC Region I (Mr. Thomas F. Dragoun and Mr. Stephen W. Holmes) at 1530 on December 15,1993.
l
,