ML20235S852

Forwards Matl & Cover Ltr Mailed to 24 Triga Facilities in Us & Foreign Facilities Re Maint of Control of CRD Switches, Per Recent Telcon

ML20235S852
Person / Time
Site:	Reed College
Issue date:	02/17/1989
From:	Whittemore W GENERAL ATOMICS (FORMERLY GA TECHNOLOGIES, INC./GENER
To:	Michaels T Office of Nuclear Reactor Regulation
Shared Package
ML20235S842	List: ML20235S841 ML20235S852
References
NUDOCS 8903080043
Download: ML20235S852 (5)

Text

_ - _ - _ _ _ . _ _ _ _ _ _ _ . ._ .

l' i e

$ sana - romics-t o

l February 17, 1989 Mr. T. S. Michaels Standardization and Non-Power Reactor ProjectDirector U.S. Nuclear Regulatory Comnission Washington,'D.C. 20555

Subject:

Notice to- All 'IRIGA Facilities Regarding Maintenance of Control Rod Drive Switches'.

Dear Mr. Michaels,

As per our recent telecon, I have iravigorated all the suggestions you made in our proposed information letter. Enclosed herewith is a copy of the material and cover letter we have mailed to 24 TRIGA facilities in the U.S. and to all foreign and overseas 'IRIGA facilities..

We trust-that the above actions will close the present matter con-cerning stuck control rod switches. We appreciate your'yt- vi and gener-ous help in this matter.

Sincerely, h,30 k , I,M N

• W. L. Whittemore, Ph.D.

Director, Irradiation Services

'IRIGA Group wtw/bb Enclosures 8

AI C O+ ,[

P PO. BOX 85608, SAN DfEGO. CA 92f 38 5608 (619)455 3000 10955 JOHN JAY HOPKINS DRIVE. SAN DIEGO, CA 921211194

,. , \;'. '

SENE N 'AF M ES.- 1 ,,,: E M yo Tu .

Fax: GJ19) 455-4169 lb February 15, 1989 I

[

Subject:

Maintenance of Control Rod Drive Switches

Dear In December 1988,

the Reed College Reactor Facility experienced an incident where a TRIGA reactor control rod switch malfunctioned by sticking in the-UP mode. Although this incident did not present a safety issue, it did involve a safety-related system and, as such, should receive operator attention. For your. information we have prepared the attached memorandum which describes maintenance procedures for control rod drive switches. It is recommended that . these procedures be followed every four or five years to reduce the likelihood that a sticking switch will occur.

~ If you have any questions on these procedures or their implementation, please i contact Bill Hyde at (619) 455-3593.

Sincerely yours,,

h William L. Whittemore, Ph.D.

Director, Irradiation Services i TRIGA Group JTG/lvg Enclosure P0. DOX 85608, SAN DIEGO. CA 92138-5608 (619) 455-3000 10955 JOHN JAY HOPKINS DRIVE, SAN DIEGO. CA 921211194

h. N..

u.. . .

}g ,

j February 7, 1989 MEMORANDUM Frm:: General . Atmics' 'IRIGA Reactor Division >

'Ib:: . All 'IRIGA Research Reactor Far 111 ties

Subject:

Maintenance of Control Rod Drive Switches The unual operation of the 'IRIGA reactor control rods is acccuplished through use of the UP, DCW, and SCRAM switches on the console. The switches are heavy duty, single action units with t apability for internal

,: illumination manufactured by MASTER SPECIALTIES CCEPANY. Fr m the perfor-mance history of these switches in nore than 70 'IRIGA consoles it can be expected that sticking (failure to respond correctly to manual manipula-tion) of one or nore of the switches can be expected during the lifetime'of the. console (10-20 years of use). In the event that sluggish or sticking-behavior is experienced by an operator, a switch should ur& y the mainte-:

nance givcMures set forth below. Preventive maintenance .using' the same  ;

givcMares can successfully be performed every four or five years to re-duce the likalih~wi that-s*4 eking switches will occur.

Ihe s*4 eking of the console control switches in no way, affects : the safety circuitry of the console. Power and fuel _ temperature scrams are un- 1 affected by the switch and will perform the safety functions intended.

While a sticking control rod switch (UP, DCW ) does affect a reactor con-trol element and is important for givyer operation of the reactor, it does not dimin4=h the safety of the system.

In the experience of the many 'IRIGA facilities, a. sticking switch can be caused- by two irtr Ent effects. . By far the most usual is caused by the plastic lens which may stick in its frame. The sinple, effective cor-'

rective measure in this case is to remove the lens,- file its edges to re-nove any excess, polish the filed edges, and replace. - Usually this eliminates the sticking problem. .A far less frequent cause of sticking is the am=1ation of foreign material in the switch linkage. We have re-cently received a 1 yvit. fra a TRIGA owner of such a sticking switch; - The

_UP_ switch stuck during the manual withdrawal of a control rod at a power of about 100 watts. The upward movement of this control rod could- only be stopped when the DOR 4 switch was pressed, as designed, but when released, the upward rod movement continued. The power increased from 100 watts to 150 watts at which tine the operator scranned the reactor manually.

The purpose of this information memorandum is to (i) alert all 'IRIGA reactor owners to the possibility of malfunction of a control rod UP or DCM switch due either to an enlarged lens or to the result of arm = lated ,

debris, di.rt, or other foreign material in the switch and (ii) roccamend  !

simple, periodic preventive maintenance to keep the switches in proper working order.

.1-._ . . . _ .. . . . . . . . .. .

t..

O -

Figure 1 is a cutaway drawing of a typical cvukel rod switch shown

" exploded" for clarification. As designed, access to the linkage ==-haniam in these switches requires that the rod control panel on the reactor con-trol console be removed by removing the screws on the console top that hold the panel in place. Removal of the lens portion of the switch, Sections (3) and (4) in Figure 1, frtun the console top does not allow access to the linkage for the recemnended maintenance but does make it possible to file away excessive growth of the plastic lens (surfaces (a) and (b) on Item 4 in Figure 1) should this be necessary. 'Ib reach the linkage, it will be necessary to lift this control rod panel.

Use extreme care when renoving the rod control panel and associated wiring. Do not disassemble the switch assembly (See Note 2 in caption to Figure 1). Each switch assembly should be visually inspected for dirt or other foreign material, espe 4 ally around the switch linkage (Section (1) in Figure 1). Also note any unusual wear of the switch, and visually in-spect the switch wiring. Use u viassed air to gently blow away dust and .

debris between sections (1) and (2), i.e., linkage, and other areas as nec-essary. If cwyrossed air does not cenpletely remove the debris, contact the 'IRIGA Reactor Division at General Atomics for further instructions.

An additional note of caution is offered. Be esp eially careful not to interfere with the wiring to the switches. Many of the interlock circuits have wiring that are attached to these switches. If any soldering is needed, be sure this is performed by an expert tehnir-ian since many wires enter a close region and loose or cold solder joints nust be avoided in this wiring which is essential to the license reqaired interlocks.

Replace the rod control panel after the above inspection and mainte-nance. Inspect the top lens of each switch assembly (sections (3) and (4) in Figure 1) for dirt, debris, unusual wear, misaliemant or binding.

Clean as necessary. Contact the TRIGA Reactor Division at General Atcrnics for further instructions if needed.

After reassembly is emplete carefully check all rod control functions and interlocks as after empletion of any maintenance operation.

Artachment

5,. .

.l l'

1 I

W 0 ---

(I) ~

,r gk e t i

l (0 ,

cy m.ss i- -_

0 (4)

Figure 1. EXPIDDED VIEW OF A TYPICAL RCD CCN1ROL SWI'ICH.

Note 1: The surfaces (a) and (b) of the lens (4) scmetines stick in the switch opening.

Note 2 The switch section (1) nust not be removed during the cleaning process. That is, section (1) nnst not be pulled away frczn section 2.

~-

',c ,.3 ...,. , 1 c R E E D 'C O L L E G E i 7'ordend, Oregon 9-coe l

<g)j.

I f.'e.

8

, . ,r .no,1a n m 3

9 6 January 1989 Af c. e Ross A. Scarano, Director Division of. Radiation Safety and Safeguards U. S. Nuclear Regulatory Commission 1450 Maria Lane, Suite 210 Walnut Creek,- California 94596 5368 Re: Docket 50 286, R ll2 ..

REPORT ON ERRAN1 CONTROL-ROD INCIDENT OF 12/16/88 INIT]AL PROBLEM On the afternoon of December 16, 1988, the reactor was being operated as part of .the requalification- requirements of an RO and for an intended irradiation. -

The reactor checklist was completed at about 13:35 with no indication of any -

problems. The first operation, measuring of the core excess at 5 watts, was '

completed at 13:54 under the observation of the SRO. An inadvertent period scram occurred during this start up but was attributed to the relative inexperience of the operator. At that point, the SRO went into the adjacent radiochemistry lab to finish preparing a sample for irradiation leaving the RO of record and another licensed RO in the control room to practice reactivity manipulations.

The RO had raised the reactor power to 150 W and was manipulating rod up and down buttons to stabilize the reactor at that level. At approximately 14:17, the operator experienced an uninterided power increase, taking the reactor from 150 watts to about 250 watts. This increase was halted and reversed, without the occurrence of a reactor scram, by the operator inserting rods into the core using all three rod down buttons. Then, as the operator began to stabilize the reactor power again, the RO noticed that the Safety Rod was at it's upper extreme despite the fact that the RO believed that all rods had been driven down to correct for the original power increase. At this point, the SRO was notified and the reactor was manually scrammed, pending a review of the operation. When initial evaluation suggested a possible equipment problem, the Director was notified.

Following a review involving the Director (SRO), Associate Director (SRO of record), the RO of record, and the additional RO who had been present at the original incident, the Director concluded that the unexplained power rise was probably operator instigated, but to allow for the possibility of equipment malfunction, he requested the operation to be repeated exactly as originally performed under the careful observation of all present. Beginning at 15:16, the RO took the reactor to 5 watts, measured the core excess, and then proceeded to 150 watts as before. The power was stabilized at that level at 15:40 ',

with no repeat of the power rise. It was noted that this resulted in the Safety Rod position being approxi nately 1/2 to 2/3 of it's fully withdrawn position, far below the level it had achieved during the earlier operation.

8901250122 800113 PDF. ADO T O'JOOO269 P F DC e

• • u  ? !. E & -- _ .

l* *.,.

e

?*

2 e'

In order to test the hypothesis that the problem was a sticking switch, it was decided to maintain a reactor power of 150 watte, (50% on the 300 watt linear power range) but to raise and lower rods (especially the Safety) alternately by small increments, essentially to continue the operation which had been in progress. After about 15 minutes of uneventful operation of this type, the Director left the Control Room, requesting that the Associate Director directly observe additional testing and that the reactor be immediately scrammed if any additional problems occurred. At 15:52, following a short activation of the Safety Rod "UP" button, the Safety rod continued to rise even though there was no finger on the button. The reactor was thereupon scrammed by the operator. The Director was immediately notified and returned to the control room. I e ordered that no further criticalities would be undertaken until the problem was resolved. From 15:54 until 16:13, the safety rod only was tested by the Associate Director with the other two rods fully inserted in the core to insure that the reactor remained subcritical. Once during this time, the rod continued to rise after a short activation of the up switch. This rise was halted by a quick tap on the "UP" button. The mechanical portions of the switch were removed disassembled, and inspected, as were the mechanical portions of seseral other rod drise switches. All were noted to be dirty; however, due to the lateness of the ' day, it was decided to suspend further testing.

At this point, the Director initiated a discussion on whether an immediately reportable incident had occurred. He concluded in the affirmative, and the NRC, the Provost, the Chairperson of the Reactor Review Committee, and the Oregon Department of Energy were thereupon notified.

ADDITIONAL INVESTIGATIONS 12/17/88 Prior to additional testing, the Associate Director and the Reactor Supervisor completed a review of all past problems with control rod drives by examination of the Mechanical Maintenance Log. Main Log entries for several of these were also examined. One previous occurrence of a similar nature was identified as having occurred on 1/4/74. Mechanical Maintenance Log and Main log entries of that occurrence were located and copied . Phone discussions were held between the Associate Director, who by coincidence had been the operator of record on that date, and the person who had been the Director at the time and who was the individual who recorded the log entry on the final resolution of the problem on 1/8/74. In that case, the problem was traced to mechanical binding in the upper portion of the Shim Rod "UP" {

switch. It was resolved by exchanging this portion of the switch with i identical parts from the unused " Transient Rod" "UP" switch. The log describes testing of the switch successfully approximately 30 times following the exchange, and no further problems with that switch were recorded in the logs reviewed. J A discussion was also held with the chief TRIGA scientist at General Atomic, who advised that similar problems had occurred at G.A. in which mechanical binding of the plastic portion of the switch was found to have been j responsible. The GA official added that such problems have always been easily  ;

solved by cleaning the mechanism, sanding off the plastic "UP" label, or )

exchanging the mechanical portion of the switch. GA reported that no problem had eser been documented with the actual switch itself.

l 1

_ _ _ _ _ - - _ - _ - - _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1

.g ' (. ' .

?w

o. v 3

? NOTE:. . GA warned. us that removal and replacement of. the switch itself, should we decide that such is necessary,' must be done- with extreme care since the soldering is very fine, and many of the required reactor: safety systems including SCRAMS : and interlocks are wired through this switch.

TESTING 12/22/88 On :this date, additional testing of the Safety Rod was conducted by the. Associate Director (SRO) and the Reactor Supervisor (SRO). The primary goal was to ascertain, if possible, whether the. problem was mechanical, in the ' upper portion. of the switch, or electrical,' in. the lower portion of the switch. The Safety Rod was tested by repeatedly activating the "UP" and "DOWN" switches.

Tests of the : Safety Rod "UP" switch were tallied. The rod ~ was tested throughout

the~ range of it's travel. It was tested by both- SRO's. Tests were conducted by' holding the botton' down for long . periods of time, by quickly tapping the ,

button, and everything in- between. The botton was . tested repeatedly by .

pushing it'in the middle as well as on.cach of it's ; edges. After a total of at least 660 activations; of the -button, .the previously demonstrated unintentional rod -

raising had .not been observed. One unusual thing,- however, was observed. At approximately Test #170 with the Safety Rod at a Position of approximately 1260 (2/3 of it's maximum withdrawal distance), the rod faiM to rise when the switch was - depressed.- ' Both SRO's observed. this. occurrence and verified that the source interlock was not activated at the time. This aberration also never repeated.

CONCLUSION AND RESOLUTION 1/6/89

' By this 'date, a letter was on hand from NRC Region ;V reiterating our obligation to notify NRC before attempting any further criticalities. On the basis of the tests reported above, and in consideration of the advice from GA, we concluded that the sticking of. the switch is a rare phenomenon, whose cause is most

. probably in the mechanical portion of the switch. Hence, we removed the UP and DOWN buttons from all 3 rods, and the mechanical linkages. which they-activate. The linkages were cleaned in a pressurized gas stream, which had the effect of removing some solid material which had deposited on their surfaces. Next, in each linkage, lubricant was sprayed between the actuating rod and the bushing through which'it passes. Finally, pieces of dirt and other debris were removed from the housings into which the linkages are inserted.

The switches were then reassembled, and each UP and DOWN button was tested 20 times, with no unusual behavior observed. At this juncture, the Director informed the NRC .by phone of our proposal to resume criticalities. As soon as possible, the incident will be reviewed with all operators, who will. be alerted to watch carefully for any aberrant control-rod behavior, and to scram the reactor promptly and notify the Director if such occurred. The Director will also put the cleaning and lubrication of the control-rod linkages on a regularly scheduled basis.

Sincerely, l

l Y L Lawrence Rubk Professor '

Reactor Director

- _ __ - -_ - -_____ - - __ - _ __ _ _ - _ _ _ _ _ _ _ _ . _ _ _ __ _ _ _ - - -