ML20212E059
| ML20212E059 | |
| Person / Time | |
|---|---|
| Site: | 05000113 |
| Issue date: | 10/22/1997 |
| From: | Williams J ARIZONA, UNIV. OF, TUCSON, AZ |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9711030041 | |
| Download: ML20212E059 (6) | |
Text
.
i'.
- Nuclear Reactor Lateretwy
%E UNIVIUL5rTYOr i n = r * (' )
@(E EO*-*
Tucson Arinma 857210020 m%
}*** 520 l-October 22,1997 U.S. Nuclear Regulatory Commission Document Control Desk Mail Stop PI-37 One White Flint North 10CFR50.71(a) 11555 Rockville Pike Rockville, MD 20852
Subject:
Facility License No R-52, Docket 50-113 Infonnation report conceming reactor power calibration.
The occurrence documented on the following pages was reported to the NRC by telephone and fax on 1 October 1997. This report supplements that information and includes analysis, conclusions and remedial actions that have been reviewed by the Reactor Committe for this facility, ff John G. Williams, Director Director, Nuclear Reactor Laboiatory g-JGW/dg
/ O90 enclosure
/
. copies:
Non Power Reactors and Decommissioning Project Directorate, USNRC Mrc Marvin Mendonca, USNRC Dr. M. Cusanovich, Vice Pres. for Research, University of Arizona Reactor Committee Members p u v ~ ()
-n@O d
9711030041 971022 PDR ADOCK 05000113 8
pop
Ia
- .5 1
i 1
- FACILITY LICENSE R 52 l
P~ tor power calibratiaa shows safety chmaads r** dine lower than actual oower On September 30,1997, a power calibtation was performed by_ running the reactor for 60 minutes in steady state mode at an indicated reactor power of 90 kW, as indicated by the multirange llacar channel. During the run, the left safety channel was observed to indicate 84 kW and the right safety channel was observed to read 86 kW. De scram seuings on each safety channel had been
' set and tested to actuate at indicated powers of 100 kW on each channel, ne Limiting Safety.
System Setting for the power level scram in steady state operation in spect' led to be no greater than 110 kW. Limiting Conditions for Operation specify that in steady state mode two reactor power
-level scrams shall be ope [able, with seapoints not above 110 kW -
~]
s Analysis of the temperature data after completion of the run showed an average power during the run of 92 +/ 1.5 kW. Derefore,it was concluded that the left safety channel had been reading low by 8 kW and the right safety channel had been reading low by 6 kW. Based on these resu'ts
+
the actual power setting of each safety channel during the run was less than 110 kW, as required.
Nevertheless, the discrepancy found between the actual and indicated powers was larger than expected, based on historical operating experience for this reactor. Furthermore, the discrepancy was in such a direction as to raise tbc possibility that non-conservative settings could have arisen.
Oncratinn Records and Relevant Procedures i
ne calibration of September 30 was perfonned before its duc date because of an observation during the previous reactor run, performed at 100 kW for 30 minutes on September 24, of inconsistency between the readings of the two safety channels and the multirange linear channel.
On that occasion, it was observed that the linear channel read 100 kW while the left safety channel read 92 kW and the right safety channel read % kW. Prior to that run the scram 'et points had been f.et and tested at 106 kW. No calorimetric power determination was made during this run.
As a result of the discrepancy observed on September 24, and prior to any further reactor run, the Reactor Laboratory Director gave instructions for the scram set points be lowered to 100 kW and for a calibration to be performed at 90 kW, as indicated on the linear channel. This was done on September 30,1997.
Inspection of previous operating records provides the following additional information.
De annual power calibration was carried out on 31 July,1997. De calibration was timely: in compliance with the surveillance requirement for annual calibration of the safety channels. As a result of that calibration all channels were found to be within an expected range, both safety channels were found to read slightly high, and both were adjusted downwards by 2 kW, at an indicated reading of 100 kW.
The operating record showed that the calibration was done on the day after a 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> run at 100 kW the previous afternoon, July 30,1997. That run was terminated about twenty minutes short of the intended 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> by a trip on the Left Safety Channel (LSC), set at 106 kW. At the time the linear channel was reading 100 kW. The trip was attributed to the normal tendency for the flux to tilt c
1 g
m i,
- during an extended power run towards the East side of the core, in the direction of the LSC. His occurs because of withdrawal of the regulating (reg) control rod to compensate for reactivity loss
~
caused by fision products. ne reg rod is on the East side of the core..
On the day the power calibration was done, the core still contained xenon fission product poison.
- A few weeks previously (6/3-4/97), the five yearly inspection of the fuel had been carried out.
l his entails removal of each element from its core position, and replacement of each element in the
~
same location after inspectich. Small reactivity effects are expected during this operation, owing to rotation of elements about their axis. A loss in reactivity amounting to several ecms was noticed when the reactor was restarted after the impection.
1 Owing to the fission product poison and the loss of reactivity from fuel element repositioning, the
. calibration on July 31 vias done with the reg rod withdrawn by about nine cents relative to its "i
~,
- previous position for that power.
he power calibration data for July 31 revealed no anomalies, but it was discovered in interviewing the Reactor Supervisor that several departures from the written procedure for power calibration occurred: (a) contrary to the procedure, a different temperature measurement system was used from that employed during the last pool heat capacity determination; (b) contrary to procedure, the adjustments to the two safety channels were made at 100 kW after the calorimetric power determinations had been made at 95 kW; (c) the Reactor Supervisor added to the measured power an amount equal to the uncertainty (1.5 kW) in the measured power before resetting the power channel reading, thus setting them intentionally high by this margin. Subsequent analysis and tests showed that these departures from procedures had a net effect of less than 2 kW on all channels, and had no bearing on the anomalies noticed later. Nevertheless these departures from procedures were incorrect, and resulted in counseling of the Reactor Supervisor concerning his obligation to follow written procedures, and to request that needed changes be reviewed and approved.
On August 5,1997, because of the small loss in reactivity that had been noticed since the fuel inspection, an effort was made to locate and reposition fuel elements that may have been incorrectly seated. If the locating pin at the foot of an element is not positioned correctly, the element can stand high by about one centhneter, resulting in a loss of reactivity. Nine elements that were suspected of being mispositioned were removed and repositioned. As a result, when the core was made critical it was found that 4 cents of reactivity had been added. De locations of the repositioned ricmc".a.r"hown in the attached diagram. It is not known whether all, or only some, of the elements that were repositioned had in fact been seated incorrectly. He repositioned elements were clustered in the West side 'of the core. Their repositioning resulted in a lowering of '
the reg rod position needed to attain criticality, and therefore would have tended to '.ilt the power profile back towards the West, away from the LSC.
- After the annual power calibration, and before the operation on September 24, described
- previously, the reactor was operated only once at a power above 10 kW. On September 4,1997, the reactor operated for 10 minutes at 100 kW, as indicated on the linear channel, while the right safety channel was observed to read 97 kW and the left safety channel indicated 94 kW. No calorimetric power determination was made.
[-
L 2
l~
y i.
~ _
. ' E.
-,m L.
i Repeated Power Calibrat}
idAfdhionalTests 3
LOn September 30 and again on October 2,1997, new power calibrations were run. On October 1, 1997, the temperature measuring system used in these calibrations and also used on July 31,1997 -
was compared whh the system used for the pool heat capacity measurement, performed in 1978.'
- At that time water had been sampled in a one liter jar from various d9ths of the pool.
1 5
Temperatures had been recorded using a mercury in-glans thermometer and then averaged.
Comparlaon of this method with rcoults obtained using an array of thermocouples mounted on a rod extending from the bottom to the surface of the pool showed that the methods agreed within 0.2
'C in a measurement of the 10*C difference of the pool temperature from the ice-point. De results of the power calibration on October 2 agreed with those obtained on September 30 within 1 kW, substantirl.tg the estimated u icertainty (-1.5 kW) of the measurements.
i i
., On October 7, with the trip net point for both safety channels set at 100 kW, the reactor was run at" ~ '
95 kW in order to confirm the eW effect of fission product poisoning on the power channel
?
measurement. De linear channel reading was maintained at 95 kW. Initially, the two safety channel readings agreed with this value, but gradually increased during the run. After 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 45 l
- minutes, as expected, the reactor tripped on the LSC, reading 100 kW. At that time the right safety channel (RSC) was reading 97 kW. Dese trends show the expected tilt of the power profile towards the East, as the reg rod is withdrawn to compensate for reactivity loss. If the safety channela laitially agree with the linear channel, their readings tend to become more conservative (i.e. high) as the run progresses. His run confirmcd that the LSC was operable: capable of performing its intended function, j
Coaciusions concerning the cause of the observed discrepancy in calibration Based on the above observations, the calibration of the reactor power channels changed, between July 31 and September 30, by an unusual amount, up to 8 kW.
Required surveillances have been carried out, and the reactor has been operated in accordance with Technical Specifications. De reactor has not been operated above its license limit of 110 kW, and therefore the safety systems witings relating to maximum power level have not been challenged in actual operation. The possibility is suggested by the data obtained during calibration on September 30 that un safety channel scrams settings may have been non-conservative during previous runs.
his is speculative, however, and we do not assume that this was the case.
As a result of the reviews and tests d_ escribed above, it is concluded that at least two small anomatics affected the pwer calibration in the relevant period, and that these together resulted in the disagreement of the power channel readings noted on September 24. %c first was the presence
. of xenon fission product in the core during the calibration on July 31. The second was the rescanng of the fuel on August 5. As a result of these effects, the core power profile during subsequent short runs at high power was tilted towards the West, relat8vc to its shape when the power instruments were calibrated, nis made the LSC read low relative to the linear channel,
- with a smrect effect in the same direction on the RSC. Dese effects were corrected by the
' calibration done on September 30 and repeated on October 2.
l At no time was the reactor operated in excess of its licensed power of 110 kW, or in 'an unanalyzed condition.= There was no safety significance to the events reported here because the bas:s for the L
3 u
. m-
~. -
E
. ; N !..
o relevant technical specifications have very wide safety margins. Similar reactor cores are routinely -
operated at licensed powers of 2f0 kW and higher. '
_l i
Corrective and Remedial Actions q
Several actions resulting fmm this event have been implemented. Our procedures for normal l
operation did not specify any action that should be taken in event of observation of inconsistency between power channels. Dis has been corrected. De power channels have been recalibrated.
Instructions were issued to seUhc scrams more conservatively than hitherto pending additional investigation. Subsequently, the procedures for selection and checking of the scram scuing have been revised.
r.'.orrective actions taken on September 30 and October 1 were described in the October 1 report to' i
the NRC: the set points for the actuation of scram on each safety channel were set at 100 kW, l
instead of their usual values of 106 kW, and a calorimetric power calibration was performed at 90 kW, after which all the power indicating channels were set to agree with the meuured power.
Correct settings were confirmed in the additional calibration run on October 2, and agreement -
~ between the three channels was again cunfirmed on October 7.
Based on these observations and the diagnosis of the cause of the anomalous readings observed on September 24, the scram set points can in future be raised again to 100 kW provided that satisfactory agreement between the two safety channels and the linear channel is maintained.
On September 4, values were logged of the power channel readings that showed the LSC reading lower than the linear channel by 6 kW. Given that the margin between the scram set point of 106 kW and the Tech. Spec. llmit of 110 kW was less than this discrepancy, the Reactor Operator
. should have recognized that the agreement between the channcis was not satisfactory, and should
- have reported this. On September 24, the discrepancy between the LSC reading was lower than the linear channel by 8 kW. On this occasion, the RO on duty initially recorded both safety channel readings as 100 kW, but was corrected by the SRO, who happened to notice that the RO had not taken care to record the values accurately. After the 30 minute run, the Reactor Director was advised of the discrepancy.
The correct action that should have been taken by the operators on September 4 and on Septmber 24 was to shut down the reactor after noticing unsatisfactory agreement between th power channels and to notify the Reactor Supervisor or the Reactor Laboratory Director. Failure to follow this course of action is attributed to a lack of awareness of the significance of the.-
4 discrepancy and of the margin between the set points and the license limits. Operators have been -
advlaed of this ptoblem. In addition changes have been made in written procedures which now_
. specify the actions that the Reactor Operator must take. The reactor must be shut down if either
. safety channel reads lower than the linear channel by a margin greatcr than the margin between the scram set point and 110 kW, or if the two safety channels disagree by more than that margin. The magnitude of any disagreement between these channels must be entered in the operating log on -
each day thM the reactor is operated above 50 kW. In additkm. the Reactor Laboratory Director is required to review the data following a power calibration and to post on the Reactor Status Board
-the selected margin for the set points.
4 a
2
~
.,l e,
, p.,
In 6ddition to changes in_ written procedures, reinforced training measures have been taken.1hc)
_- Reactor Supervlaor has been instructed to ensure that written procedures are correcdy followed
- dwing surveillance and maintenance operations, and to promptly report the need for properly i reviewed revisions when required. All Reactor Operators have been instructed on the importance
- of making an acewate log during operations of each required meter reading, especially the two safety channels, and of the need to take timely actiop in event of anomalous readings. =
1hc analysis, conclusions and remedial actions presented here were reviewed by the Reactor
. Committee in meetings held on October 22 and October 24,1997. Revisions to procedures mentioned above were approved in ti.ase meetings, and are now in effcct.
It is expected that these actions will prevent the reactor from being operated in a condition that might violate technical specifications, 1
's s
October 30,1997 L O. Williams, Director
- Nuclear Reactor Laboratory Univenity of Arizona a
s
. - _ _ _ _ _ _ _ -