Part 21 Report - 1997-090

 
 

ACCESSION #: 9702120384



North Carolina State University



Nuclear Reactor Program

Department of Nuclear Engineering



Box 7909

Raleigh, NC 27695-7909

(919) 515-2321

FAX (919) 515-5115                      10 February 1997



U. S. Nuclear Regulatory Commission

Attn.:    Mr. Alexander Adams, Jr.

          Senior Project Manger,

          Non-Power Reactors and

            Decommissioning Project Directorate

          Office of Nuclear Reactor Regulations

Washington, DC 20555



Subject:  Nuclear Instrumentation information which may be useful to Non-

          Power Reactor facilities



Dear Mr. Adams,



Executive Summary



The PULSTAR reactor at North Carolina State University has two Gamma

Metrics Wide Range Linear neutron flux channels.  One channel, designated

the "Linear Channel" was operated in the auto-ranging mode and used for

power monitoring and control with a maximum range of 1 MW.  The other

channel was designated as the "Safety Channel" and was operated in the

manual range-change mode on the 1 MW range.



In November 1995, the Linear Channel experienced a failure to

automatically up-range and the range control card was returned to Gamma

Metrics for repair.  It was noted that both the linear and safety

channels had been operating with the neutron flux scram smart bi-stable

set so that it was enabled on the 1 MW range.  The bi-stable mode was

changed at that time so that both units would produce a scram at 120% of

range on all ranges.  In December 1995, the unit was successfully bench-

tested with the repaired card, but again failed to up-range and exhibited

spurious automatic range-changing when it was returned to service.  The

unit was returned to Gamma Metrics at that time.



In January 1997, the Safety Channel spuriously down-ranged from the 1 MW

range to the 300 kW range when the reactor power was increasing through

150 kW.  While operator manually scrammed the reactor, the scram function

remained operable at 360 kW due to the having the scram function enabled

on all ranges.  This unit was also returned to Gamma Metrics for repair.



Facilities using the same type of equipment should be aware of these

range change problems in both automatic and manual range mode and the

fact that a spurious range





North Carolina State University is a land-grant university and a

constituent institution of The University of North Carolina





Mr.  Alexander Adams, Jr.          (2)                   10 February 1997

U. S. Nuclear Regulatory Commission



change can defeat the scram function if the smart bi-stable is set so

that the scram is active on one range only.



Part 21 Assessment



The licensee has reviewed 10CFR Part 21 and concluded the Gamma Metrics

Wide Range Linear monitor ranging malfunctions experienced do not result

in disabling the protective function of the monitor by themselves and do

not create a substantial safety hazard.  Nevertheless, the information

has been shared with the TRTR community, Gamma Metrics, and the NRC sine

our experiences and the findings may be useful to other licensees.



Discussion



NC State University Linear Channel



The first problem experienced with the Gamma Metrics Wide-Range Linear

Monitor occurred in November 1995 with the PULSTAR Reactor "Linear

Channel" that is ranged with reactor power, used for power control, and

provides redundant scrams for the PULSTAR Reactor "Safety Channel".  This

event was previously reported to the NRC and was summarized in the

February 1996 TRTR Newsletter.  The range control card was suspected and

returned to Gamma Metrics where several components were replaced.  The

monitor was reassembled and successfully tested on the bench top and the

channel was returned to service.  On December 20, 1995 the reactor was

restarted under testing authorization and multiple range changes were

deliberately introduced during a gradual power escalation.  The channel

again failed to automatically up-range and the reactor was shutdown.



The entire monitor was returned to Gamma Metrics where modifications were

made in the internal grounding.  During the subsequent receiving

inspection it was noted that several cable connectors were not seated

properly and that a pin on one of these connectors was burned.  Gamma

Metrics provided a replacement connector and associated harness that was

installed by the reactor staff.  During pre-start testing, it was found

that actuation of the rod reverse demand would cause the range to jump to

a different range when the range control was in the manual mode.  A Gamma

Metrics field engineer spent one week at the site but was unable to

diagnose the problem.  The monitor was then sent back to Gamma Metrics. 

A PULSTAR facility design change was approved for re-installing the

linear channel picoammeter that had been used prior to the Gamma Metrics

Wide-Range Linear Channel.



In October 1996, the linear channel was returned to NC State University. 

Changes that had been made included replacement of the printed circuit

board backplane.  During the initial bench-top check-out of the returned

unit, new spurious range changes were observed including oscillatory

behavior.  This included increasing the range to greater than 1 MW even

though this is inhibited by design in the NC State University version of

the monitor.  The unit was again returned to Gamma Metrics where it

currently remains.





Mr.  Alexander Adams, Jr.          (3)                   10 February 1997

U. S. Nuclear Regulatory Commission



NC State University Safety Channel



The same type of unit is operated on the 1 MW range in the manual range-

change mode.  In January 1997 this unit spuriously down-ranged from the 1

MW range to the 300 kW range as the reactor power was increasing through

150 KW.  At 150 KW, an internal trip circuit is enabled to protect

against loss of flow or the lower plenum flapper valve was not closed. 

While the operator manually scrammed the reactor, the scram function

remained operable at 360 kW due to having the scram function enabled on

all ranges.  This was verified during a post-shutdown check-out.  This

unit has also been returned to Gamma Metrics for repairs.



Results



Our bench-top testing with the Linear Channel and consideration of the

circuit design indicated in 1995 that:



o    In either the automatic or manual range change mode, spurious

     ranging did not affect the protective features of the monitors

     unless:



     -    There is a spurious down range and the smart bi-stable is set

          so that the scram is active on one range only



     -    The unit is on the 1 MW range, a spurious uprange occurs, and

          there is a concurrent or common mode failure of the maximum

          range limit function



By setting the smart bi-stable so that it is operable on all ranges, the

protective function will be maintained for single-failure spurious range

changes.  A spurious range change to a range greater than 1 MW requires

both a spurious range demand and failure of the maximum range limit

function.



It should be noted that the individual Gamma Metrics channels are not

single failure proof by design.  Many other single failures could cause a

channel failure.  Redundancy is provided through the use of two separate

channels for high neutron flux scram and low flow scram enable.



The first problem experienced with the Gamma Metrics Wide-Range Linear

Monitor was in November 1995.  We performed a series of bench-top tests

at that time in an attempt to determine the root cause of the auto-

ranging problem.



The bench-top tests and review of Gamma Metrics documentation for the

unit demonstrated the following:



o    Auto or Manual Mode ranging malfunctions are not likely to affect

     the protective function(s) of the unit,





Mr.  Alexander Adams, Jr.          (4)                   10 February 1997

U. S. Nuclear Regulatory Commission



o    A spurious down-range can defeat the scram function IF the smart bi-

     stable is set so that the scram is active on one range only,



o    The smart bi-stable comparator should be set on the tenth rotary

     position (dumb position) which causes the bi-stable to be operative

     on ALL decades.



It should be noted that Gamma Metrics explicitly explains the smart bi-

stable options in their technical manual.



Conclusion



The Wide-Range Linear Monitor configured to provide protection only on

the full power range with a coincident spurious down-range would disable

the bi-stable and not cause a trip (over-range for the lower scale). 

Such a scenario may lead to a Technical Specification violation for not

having a required protective action.  This specific possibility is

eliminated by placing the rotary switch on the smart bi-stable card in

position "10" (dumb setting) on the NCSU units.



We have advised the TRTR community via the Internet recommending that

facilities with identical equipment should review the smart bi-stable

description and configuration.



Please feel free to contact us at (919) 515-2321 with any questions or

comments.



                         Sincerely yours,



     Charles W.  Mayo, Ph.D             Pedro B. Perez

     Director,                          Associated Director,

          Nuclear Reactor Program            Nuclear Reactor Program



cc:  Mr. Craig Bassett, USNRC Region II

     Senior Project Manager



     Gamma-Metrics, Inc





*** END OF DOCUMENT ***