ML22238A085

NRC-2022-000208 - Resp 1 - Final, Inspection Report SEFOR Partial Scram 1970

ML22238A085
Person / Time
Issue date:	08/24/2022
From:	NRC/OCIO
To:
Shared Package
ML22238A081	List: ML22238A082 ML22238A084 ML22238A085 ML22238A086 ML22238A087 ML22238A088
References
FOIA, NRC-2022-000208
Download: ML22238A085 (22)

Text

U. S. ATOMIC ENERGY CO MMISS ION REGION II DIV JS ION OF COMPLIANCE

Report of Inspection

CO Report No. 50-231/70-4

Li en see: General Electric Company and SAEA License No. DR-15 Category B

Da t e of Inspection: October 6-9, 1970

Date of Previous Inspection: August 10 - 12, 1970 _

//-I Date 7-70

Reviewed By:

Proprietary Information: None

SOOPE

A. routine, announced inspection was made at the fast spectrum, sodium

cooled SEFOR reactor located approximately 20 miles southwest of Fayetteville, Arkan s as. On October 6-7, 1970, V. D. Thomas, Compliance

_ Headquarters, reviewed the licensee ' s corrective act ions following a

• ttpartial" scram whi c h occurred on September 12, 1970. The nfeede~*

B. report covering Thomas' portion of the inspection is attached as Exhibit

SUMM~RY

Safety Items - None

!oncompliance Items - Automatic flow control valves were installed in the discharge lines fro~ the redundant argon vent vacuum pumps. These pumps

a ::*e used to refill the auxiliary primary system if there is a 1 ine break

~

CO Rpt, No, 50-231/70-l ; ;\\

i I

• n the main primary coolant loop. Contrary to 1.0 CfR 50.59, the liceasee ~ *1

~id *not have a written safety evaluation to determine that this change did not involve an unreviewed safety question. A Form AEC-592 will be issued for this item of noncompliance.. (See Section M.)

Unusual Occurrences - The "part ial 11 scram reported in an Inquiry l-~emorandum :1 dated September 18, 1970, was reviewed during this inspection. The licensee repor ted the malfunction to DRL in a lett er dated September 21, 1970, as required by Section 3,C.(1) of the operating license. (Sec Section F,3,)

Overinflation of an access suit caused a mechanic to loose his bal. ance and fall across the reactor vessel head during a refueling cell entry.

Use of the man access su:i.t.s has been suspended by the facility manager as recommended by the Site Safety Committee until corrective action.has been completed to prevent similar occurrences. (See Section U.)

Status of Previously Reported Problems - The licensee's response to the Form AEC-592, issued for two items of noncompliance reported in CO Report No. 50-231/70-~, is considered unsatisfactory by Region II. Region 11's comments on the response were forwarded to Compliance Headquarters on October 6, 1970. The noncompliance it ems,,,ere concerned with hir i r:6 a maintenance supervisor and instrument engineer who did not meet the experience requirements for these positions as specified in the Technical Specifications.

Modifications have been completed to prevent failure of the containment ventilation isolation valves to open on demand.l/ Valve operation was transferred from the instrument air system to the instrument nitrogen systan to prevent freezing of moist ure in outside lines during cold weather. (See Section K.2,)

Fabrication of the new argon purification system has been completed and has been shipped from California. The new bubb ler system will be installed during the next maintenance outage of the facility.

Other Significant Items -

1. Wide-Range Monitor (WRM) - Heat Balance Correlation

Data obtained during S Mw static tests revealed that agreement between WRM's and reactor heat balances was dependent upon core average temperature. At any given thennal power level, WRM indications are ~ -

proportional to core average temperature. Diagnostic tests recommended,~

by the G-E. engineering group in Sunnyval e are being performed to

• evaluate the problem. (See Section F.l.)

l/Report ed in Inquiry Memorandum dated January 9, 1970, and in CO Report No. 50-231/70-1.

CO r.pt, No. 50-22 1/70-.'..

2, cover Gas Monitor Sensitivity -

Data was reviewed which demonstrates that the cover gas gamma monitor will detect a fission gas release equal to 1% of that contained in one fuel rod. Paragraph 3. 3.L of the Technical Specifications requires such a demonstration prior to operating above 10 Mw. (See S e*~t ion E. 3.)

3, Fuel Surveillance Program

Examination of. guinea *pig rods has been completed at 5, 10, l.'.>, and 17.5 Mw during the power ascension program. No evidence of fuel cladding failure has been observed during any part of the fuel surveil lance program. Paragraph 3.10.C of the Technical Specifications states that power shall not be increased above 15 or 17. 5 Mw until guinea pig rod examinations are satisfactory. (See Section G.)

4, Core Physics Requirements

Shutdown margin and core excess reactivity were reviewed and found to be as required by applicable sections of the Technical Specifications. (See Sec t io n F

• 2 * )

5. Primary Coolant Samples

Primary system sodium samples have been taken as required by paragraph 4.4.Q of the Technical Specifications. (See Section E.2.)

6. Nitrogen Cooling System Drver Condensate

Sampling of the nitrogen cooling system dryer condensate was in accord cations. (See Sect ion K. 1.) ance with the req~irements of paragraph 4.5.K of the Technical Specifi

7. Partial Plugg age of Cover Gas Supoly Line

Partial pluggage in the reactor cover gas ~rgon supply line resulted in the inability to mai ntain the desir ed cover gas pressure. Normal opera tion was restored after clearing the pluggage with a "rotary snake. 11 (See Section E.1.)

Followup Items - See ~xhibit A for the current status of outstanding items.

Management Interview - The following items were discussed with Arterburn at the conclusion of the inspection:

1. Partial Scram (See Section F.3.)

The inspector stated that the time response data indicates that the cause of the partial scram had been identified and that the system as presently

llllii1 '

co f.p *,v. -t l'** 50- : 31;10- 4

adjusted should correct che probl em. Since routine s urveillance tests failed to detect this Pfoblem, consideration should be given to revising the tests to verify tha t a scram cannot be reset either manually or by a change in the adjustment of the auxiliary lockout contacts. Arterburn

&tated that the method and frequency of performing such tests would be considered based on the recomrnendat ions of the G-E engineering group in Sunnyvale.

2. Argon Vent Vacuum Pumo Discharge Valves (See Section MJ.

Arterburn was informed that a written safety evaluation was not available for the installation of automatic flow control valves in the discharge lines from the redundant argon vent vacuum pumps. A written-saf *ety evaluation is required by 10 CFR 50.59 to demonstrate that the modifica tion did not involve an unreviewed safety quest:ion. Arterburn stated that members of the Site Safety Committee had reviewed the changes individually in the course of approving the modification request form.

The inspector informed Arterburn that Region II would issue a Form AEC-592 for failure to maintain a written safety evaluation of this modification.

3. Discrepancies Between WRM' s and Heat Balances (See Section F. l}

Arterburn stated that diagnostic tests would be conducted based on recommendations from the G-E engineering group in Sunnyvale. The hi&h flux level scram point will be lowered to 95% to prevent exceeding the Technical Specification limit at low core average temperatures. Tests will be conducted at 5 Mw to determine the relationship between WRM indications and reflector position, sodium flow rate, and c ore inlet temperature. Core average temperature will be varied from 400° to 760°F.

Arterburn stated that the problem will be reported to DRL in accordance with Section 3. C(2) of the operating license.

• 4. Power Ascension Program

Arterburn stated that all Group II power ascension tests have been com pleted through the 17. 5 t1w level. Several of the Group III tests have been completed at 15 Mw. Only a few t e sts remain to be completed at 19 and 20 Mw before starting the balanced oscillator tests. Electronic equipment for th e se tests is currently being checked out for proper operation. Out-of-cell and in-cell testing of the Fast Reactivity _Excursio" Device (FRED) is complete and the test data sent to Sunnyvale for analysis.

There was good comparison between actual and predicted results according to Arterburn.

CO Rpt. No. 50-231/70- :1

The reflector configur..::tior i at 17.26 l1w (September 5, 1970) and the associated worths were as follows:

Position Worth Total Mea sured Worth I Reflector No. I 3 61% 96c 132c 8 70% llOc 133c 206c 265c

The difference between the total available excess and the required reactivity at 17.26 Mw is 59c. The additional reactivity required to reach 20 Mw is (20 - 17.26) Mw x 6.5c/Mw = 18c. This results in 41c exce s s reactivity at 20 Mw. The excess reactivity can be expected to decrease at a rate of le per month due to decay of Pu241 and 0.75c per full power day.

Technical Specifications 4.2.A states that Technical Specifications 3.3.A and 3.3.B shall be demonstrated at least once every four months after achieving 10 Mw. Becker stated that this requirement is fulfilled each time the reactor is taken critical. The reactor operators are given the estimated critical reflector position prior to each startup and are instructed to notify the shift physicist if the actual critical position differs by more than 6 cm from the predicted value. With a reflector in its position of maximum worth, 6 cm is equal to lOc and, therefore, would detect a change in core or reflector reactivity status greater than llOc.

Technical Specification 3.3.C states, "The reactor power coefficient of reactivity at constant inlet temperature and constant coolant flow rate shall be n ega tiv e. " Values obtained from the power ascension data indicate that the power coefficient is equal to -6.5c/Mw.

Technical Specification 3.3.F states, "The reactor shall have a phase margin of at least 30 degrees at the point where the Nyquist plot cross es the unit circ le." The evaluation of the Nyquist plots using on-line Fourier data for flow of 30, 50 and 100% indicate phase margins ranging fr om 74-77° at the Nyquist unit circle intercept.

Technical Specifications 3.!0.A r.nd 3.10.B are satisfied by the preceding results.

3. Partial Scram

Arterburn notified Region 11 that the reflector system had failed to function a~ designed on September 12, 1970, following a short duration scram signal.l/ The licensee's evauation and corrective actions

YReported in Inquiry-Memoranch.::*!l c~ted S ep t ember 18, 1970.

p llpt, No, 50-231/70-4

following the occurrence were reviewed by V. D. Thoma s, Complicnce Headquarters. Thomas' review is contained in Exhibit B.

Swartz stated that he would write a letter to Arterburn recommending that routine surveillance tests be revised to preclude this type of failure in the future. This item was added to the Outstanding Items List.

G. Core and Internals

Fuel Surveillance Program

Results of the fuel surveillance program were discussed with Johnson.

Technical Specification 3.3.K states that fuel rods having defects defined below shall not be reinserted in the core:

1. Cladding rupture, perforation, or other observable defects which may cast reasonable doubt on the integrity of the rods.

2. Local swelling of the cladding in excess of 10 mils or bowing of the rod sufficient to prevent reinsertion into the core.

3. An increase of more than one-half inch in the column height of either fuel segment.

Johnson stated that Item l above was confirmed by visual inspection. Item 2 was checked by a diameter spiral trace and diameter trace at fixed

azimuth in the fuel inspection station with a measuring sensitivi ty of 0.0002 inch. Bowing and overall rod l ength are measured with indicators having a sensitivity of 0.004 inch. Item 3 was checked by both trans mission and emission gamma scans having a sensi tivity of 1/32 inch. Both scans produced traces of good quality and the emission scan showed sharp peaks at the pellet interfaces over almost the full length of the fuel column. The results of all guinea pig rod inspections conducted to date are listed below. Changes in pellet column le ngt h and cladding diameter are referenced to preirradiation data.

Rod Al6 S Mw 15 Mw 17.S Mw

None None None Observable Defects <1/16" <l /1611 1/16" Change in Pellet Column Length £0.001 11 <O.OOltt <:'O. 001 tt Cladding Diametral Change 0.036" 0.036" 0.036" Bowing

• i

  POllT Or' AS S 1$1' UIS PE CT TON

Licensee: G-E Company ~nd SAEA (SEFOR)

Docket No. 50-231

Dates of Inspection: October 6-7, 1970

/1/~*l?c:

Inspected By: Vincent D. Thomas, Instrumentation Engineer Date

//;/ ;~:

1 Date Reviewed

• By: H. R. De n ton, Chief, Technical Support Branch

Propri etary Information: None

Int rod uct ion

Ao announ ced inspection was performed at the SEFOR Nuclear Plant of the G*E Company and Southwest Atomic Energy Associates located in Cov e Cr eek Township, Arkansas. This report is limit ed to the events surro un ding the incomplete scram that occurred on September 12, 1970.

\\

- The in spection was made to review the procedures, test data, and corrective action resulting from the subsequent investigation performed by G-E following the occurrence.

Summar y

1. Following the incomplete scram event on September 12 and during the investi gatio n that followed, th e actuator tabs on all 18 DC contactors were adjusted s o that the auxiliary contacts would operate with or before the main contacts. Paul Swar tz, G-E, stated that he will recom mend that the site per s onn e l estab li sh a testing program so the timin g sequence of contact operation is always known. He ai s o stated that the scram reset circuit will also be included as part of th e te s t program,

2. Examination of t es t data revealed that nin e of 12 safe ty s yst ems and all six cont ai~ment building isolation DC contactors were not ope ratin g

properly. Main contacts of the DC contactors we r e operating before the auxiliary contact s operated, Pr oper operation requir e s the opposite sequ e nce of cont3ct action. I

3. No at~empt was.. ~de to s imulate the conditions that caused the incid e nt prior to adju s tm e nt of the auxiliary contact operation. A short duratio n pulse test was compl e ted efter the fix.

Exhibit B Page 1 of 9

'- I 1

• . 2 -

.l

4

• Failur e in th e s cran re s et circuit, which could hav e caused the occurrence, was not con s ider e d during the invc 1, tigation. Examination of the electrical circuits and test data rev e aled that of the four DC contactors directly involved in the incident, more than one and poss\\~!y four could hav e failed to operate properly at the time of the ~ident, - *

5. Inspection of the main secondary flow controller set point station on September 14 revealed the adjustable slide wire-driven by the operator's thumbwhcel to be e rratic on the low end and had a short in the area of 3000-4000 gp~ *.,,

6, Mr. Swartz stated that EVESR nuclear plant in Vallecitos, California I II is the only other nuclear facility that utilized this type contactor in the safety system. He also stated that the EVESR facility is no longer operating.

7, G-E performed all testing requirements for the safety system, secondary flow monitoring system, and the reactor core outlet temperature moni~oring system set forth in the technical specifications.

DETAILS

I. PERSONS CONTACTED

The foilowing G-E and Compliance personnel accompanied and/or were contacted during the inspection:

General Electric Company - SEFOR site

Jesse Auterburn - Site Manager Mel Mathis - Instrumentation Engineer Dan Gellerman - Senior Instrumentation Technician

General Electric Company - Sunnyvale

Mel Muir - Manager, Control and Electrical System Unit Paul Swartz - Electrical Engineer

Division of Comoliance

Charles Upright - Reactor Inspector, CO:II Vincent Thomas - Instrumentation Engineer, CO:HQ

II, RESULTS OF INSPECTION

I A, Description of Events Surrounding Incomplete Scram ' l

On September 12, 1970, while attempting to increase the main secondary sodium flow, a short duration low flow trip signal caused the

Exhibit B Pnge 2 of 9 -,

J 1 I

~itiation of a reactor scram action; however, en incomi>lcte scram occurred. 1s:

he events surrounding the "incomplete scram" *,1erc ns fol10~ *

1. "D" scram bus was in a trip condition due to a secured freon unit which caused a "low freon pressure" trip condition to exist,

2. The reactor was operating at 5 NW with an es tablish ed flow rate in the main se condary sodium system of 1400 gpm.

3. The reactor operator was in the process of attempting to adjust to the secondary flow controller sc~ point station to increase the system flow rate to 4500 gpm.

4. An annunciator alarm - "low Na flow main secondary loop" -

was received and all reflector upper limit lights were extinguish ed, carriage separatio n had occurred on reflectors No. 3 and No. 8, and all position indicators were some 5 CM below full-up position. The power level on the wide range monitor (WRM) also indicated a decrease from 5 MW to approxi mately 3 MW becau se of the reflector movement and bus "D" scram bus indicator was the only light lit. All scram solenoid position indicat o r s were extinghushed indicating J I no scram solenoids were deenergized. !

5. The operator immediately scrammed the reactor. i I

B. Background I I

Th e SEFOR safety system is a mul t iple bus (A thr ough F) type made up of r e dundant electric magnetic relays (DC contact or). Fast ac tin g primary r elays (four pole mercury-wetted contact type) are operated directly from trouble function trips, such as flow, pressure, and flu x.

Contact s of th ese primary r elays are wired in series or series-parall e l I arrangements, depending upon th e trip logic, to make up the safety chain that operates the DC contactors. Cont acts of th e DC contactors arc wired in redun dant one of three (A through C contactors) logic in series with ! I redundant two of three (D through F) loeic matrices. The two logic matrices l in turn operate two solenoid valves that are located in the hydraulic con I I

trol system of the reflector rod drives.

C. Licensee's Conclusions as of September 14 1 1970 l

Review of the report dated September 14 to the si te safety committee I covering the events and subsequent investigation r evealed the following possible problem areas were considered and investigated:

1. Intermittent loss of 125VDC power to the hydraulic scram solenoids - Ruled out because of indication on event recorder pen No. 24 low flow ma.in secondary and the associated low flow alarm. Contacts 12 and three of K3 in any one relay chassis Exhibit B Pase 3 of 9

\\.

---

=--

• ,~ =:::::i I.

(scram dr awe r s D, E, or F ) ha d to clos e to d e f l e ct th e p e n I on the event record e r. !

2. Intermittent los s of 26.5VDC power to the saf e ty control 1 center - DC cohtactors - Rul e d out pri n~rily be cau s e o r the,.

indication on th e ev e nt recor~er, Loss of 26,5 volt power ;

to the DC contactor coil circuitry would hot have pro d uc ed a pen deflection on the event recorder and loss of 26.5 volt power to the scram drawers A through F would have deflected all of the pens associated with the aff e cted chas s is plu s all '

associated annunciators would have been in the alarm condition.

• In either case, DC contactors AlKl/A2Kl throu g h C1Kl/C2Kl and ~

E1Kl/E2Kl through F1Kl/F2Kl should have rer..a.ined deenergized l (locked out). DC contactors DlKl and D2Kl wer e not considered ~

since this drawer (D)was in a scram (deenergized) condition..,

3. Temporary loss of 125VDC to any one of the scram draw e rs A i throu gh C or E through F was also considered as a possible cause ; however, it was ruled out due to the deflection of ~ (

pen No. 24 on the event recorder, The 125VDC power is not in the primary relay coil circuit it is in the relay contact circuit so that the pen deflection must have b e en an inter ~ I I ruption of the fast acting primary relay locat e d in the E j 11 and/or F scram drawers. This still do e s not explain the lack t

)

of a scram contactor's lockout feature. Fa ilcre of l o cko u t t I

~ r (auxiliary) contacts on Al/A2Kl through Fl/F2Kl to break was 1 considered and it was determined that it could cause the t

~

affected contactor to reenergize if a short duration trip ~

occurs. i I

4. The site personnel used two Simpson Voltohmists (VOH's) to I d etermine if the hydraulic scram solenoid logic circuitry I l could be interrupted (main contacts of the DC contactor) without the lockout (auxiliary contacts from some DC contactors) ever opening. I/

One VOM was plac e d acros s one set of main contacts of th e scram contact and the other was placed across the auxili a ry contacts. The relay actuation was manually mov e d against it s mechanical stop (fully energized) and slowly rciea s ed until e nt was then one or both VOM's indicated ope n circuit. Mov e m continued until the second contact (if only one opened)VOH indicated open circuit. The test was repeated with the s e cond set of main contacts of the scram contactor.

The results of the 125VDC. scram relay contact operation are listed below:

Exh ib i t B Page 4 of 9

I,

j t

.* 5 -

Hain Auxiliary

/,lKl X A2Kl X BlKl X B2Kl X ClKl X t,.

C2Kl X DlKl X ElKl X E2Kl X I FlKl X *~

F2Kl X !,

D2Kl X ~

Note: X indicates which opened first l

~

I It was felt from the information taken above that since one 1 I I trip was already in the system (D bus) and the event recorder ' 'I ! I indicated a short duration trip from pen No. 24 that buss e s j '

E and/or F received the trip causing interruption

• of 125VDCt the main contacts of the scram relay open deenergizing the i scram solenoids before the auxiliary contacts locked out the DC contactor coil. 1.

~ j The site safety committee then reviewed the findings above and agreed that the 12 contactors should be adjust e d so that ;,

the auxiliary contacts would release before its main contact s.

This adjustment was *made by the site personnel on September 13.

No attempt was made to duplicat e *the initial condition s, The adjustment consisted of repositioning the actuator tab so that l., I the desired contact sequence would occur. ~

-Tests on the reactor control sy s tem (Q0-5) and ~onthly t e ~t s ~ I.,1

~ I of the reactor s afe ty system (M0-9) we re run. Th e licen see ~ I test procedures we re compl e ted in order to sati s fy the r e quire

• I 11:,,

ments set forth in th e technical specifications and to d e termine l I i if the systems functioned properly following the adjustment. f i1:

On September 14 Dan Ge ll e rman, Instrum e ntation Te chnician, \\,, l'i discovered the cause of th e short low ~low trip was due to I I the erratic behavior on the low end of the set point adju s t I station of the main secondary flow controller. It was found j I that a short in the ar ~a of 3000-4000 g pm had e xi s t e d. - The ! i nlldc wire wan rcpnir c <l and placed bnck in ncrvice ofter functional testo (W0-5) were perform e d on the set point I r stnt ion.

I I' I.

Exhib it B I Page 5 of 9

'-.... '\\.

- G -

The report states that ::he <!Vents and investigation were discus sed with the site personnel (Mathis) and Jim Haar of Sunnyvale.

D. Licensee's Cone lus ion s as of September 21. 1970

Review of another report to the Site Safety Committee dat e d September 21,relating to the investigation of the incomplete scram event revealed the following:

l. A series of tests were run to obtain drop-out t imcs of the scrams solenoid, the DC contactor and finally, the combination of both as a typical scram unit.

2. In the first series of tests 125VDC was applied to a spare scram solenoid via a spare primary relay. The relay was I

• I pulsed by a signal generator to determine the solenoid drop out and pick-up times. A small voltage (1.5VDC) sign~l indi cating contacts of the relay opening and closing and also the position of the scram solenoid were recorded on a multi ch~nnel high speed oscillograph. The results of approximately 50 record ed pulses show that the scram solenoid r equires a lo s s of power for at least 12 -15 milliseconds before it will drop out and 8-20 milliseconds to pull back in.

' 3. The second series of tests were run to obtain information of the tim e relationship be tw een the operation of the DC con tactor (whose auxiliary contacts would op en after th e main contact s) and the solenoid scram valve. Using th e same spa re components and t est setup used in the solenoid scram valv e drop-out a nd pick-up test, the results revealed the following:

(a) Applying a series of varied pulse widths the DC contactor coi l circuit could be made to hold in while o~in contacts would open._ Pulses of 65 -7 0 milliseconds in duration to the coi l circuit, both main contacts opened be twee n 10-60 milliseconds. The DC contactor on several occasions dropp ed out.

(b) Applyin g a series (11 recorded) of the 65-70 millisecond pu lse widths show that the scram so lenoid opened for 20-60 Qilliseconds ten times. One pulse width of 80 milliseconds deenergizcd the scram contactor and scram sol e noid.

(c) A 70-75 millinccond pu~,w would opr.n the " 11 p,1rc tr:;t" scram relay 100'1. of the tim e while n 60-65 millisecond pulse width, will open the scram solenoid a.bout 40"1. of the time. No attempt to further destort the actU.'.ltion tab on the DC contactor under test was made.

Exhibit B Page 6 of 9 l

The report sta t es that the eve n ts a nd inv es tigation were discussed with the site personnel (Mathis ) and Jim Haar of Sun nyva le.

D. Licen s ee's Conc lusio~ s as of S eptember 21, 1970

September 21,relating to th e investi ga tion of th e incomplete scram event Review of another report to th e Site Safe ty Committee dated

revealed the following:

1. A series of te s ts were run to obtain drop-out tim es of th e scrams solenoid, the DC contactor and finally, the combin a tion of both as a typical scram unit.

2. In the first series of tests 12SVDC was applied to a spare scram solenoid via a spare primary r e l ay. The relay was pulsed by a signal generator to d e t e rmin e the so l eno id drop out and pick-up tim es. A small voltag e (l.SVDC) s ign~l ind i cating contacts of the relay opening and closing and also the posit i on of the scram s ol e noid were recor de d on a multi channe l high speed oscillogr a ph. The r e sult s of approximately 50 record ed pulses show th a t the sc ram so l eno id require s a loss of power for at l eas t 12-15 milliseconds before it will drop out and 8-20 milliseconds to pull back in.

3. The second series of tests were run to obtain informati on of the tim e relation sh ip be twee n the ope ration of t he DC con tactor (whose auxiliary contacts would open afte r the main contacts) and the solenoid scram v a lv e. Using t he same spa r e compon e nt s and test se tup u se d in the so l e noid scram v alve drop-out and pick-up test, the result s re v e aled the followi n g :

(a) Applyin g a series of varied pulse widths the DC cont~ctor coil circuit could be made to hold in while main contacts would open._ Pulses of 65-70 milliseco nd s in duracion t o the coil circuit, both main contacts ope n ed be twe en 10 - 60 milliseconds. The DC contactor on se v eral occa s ion s dropped out *

(b) Applyin g a series (11 recorded) of the 65-70 milliseco nd pulse widths show that the sc r am soleno id opened for 20 -

60 cilliseconds t en times. One pulse width of 80 millisecond s deenergized the scram contactor and scram solenoid.

(c) A 70-75 milli~c-cond p ul_uc wou l d opr:n the 11 :; p,,rc t c-:;t "

scrnr.1 rclo.y 100'7. of the time while a 60-65 mil 1 isccond pul se width, wil 1 open the scram so l enoid ilbo ut 40'7. o (

the tim e. No at t empt to further destort the actuation tab on the DC contactor u nder test was made.

Exhibit B

- 7 -

A second insp e ction of th e DC contnctors (12) Al/A2Kl through Fl/F2Kl along with the DC con t nctors (six) used in the containment isol~tion sys t em was made o n Septemb e r 20. Using th e sam::: inspec ti on techniques which employed th e two VOM ' s pr e viou s ly discussed, it was found that all six c on tain me nt i solatio n DC cont~ctor s contact timing relation shi p was faulty, Th e site pe r ~onne l then proceeded to adjust the DC contactors i.n the s ame ma nn e r as previously descri be d until the auxiliary contact s r e leased before the main contacts, E. Licensee's Safety System Response Te s t S e pt embe r 24, 1970.

After the above adjustments and series of t ests we re completed, the site perso nn e l th e n performed a test to determine the safe ty systems '

(A through F) responses to simulated short durntion trips.

A set of contacts from the spa re primary relay was placed b e twP.e n each relay scram draw er and a selected trip unit. The primary r e lay was then pulsed with a series of varied puls e widths until one or both sc ram busses (solenoid dcen e rgized) opened. Recordings we re mad e of each sho rt duration pulse showing its effect on the safety system. The test wa s repeated for all 12 DC contactors that make up the six safety s y stems of A throu gh F.

deenergized are tabulated as follows :

• The pulse widths in milliseconds at which point the DC contactor I l AlKl-70 A2Kl-65 I',

BlKl-45 B2Kl -55 ClKl-67 C2Kl-45 DlKl-75 D2Kl-75 ElKl-65 E2Kl -60 I FlKl - 55 F2Kl.:6o I

signals sent to the six s af et y sys t em s cr am drawers, it was found that i n In a series of 72 r e corded 45-75 mi lli seconds s hort duration tr ip

all case s where the pulse was of suffi c i e nt duration to b reak the voltage to the sc ram solenoids, that scram went to completion, that i s, the auxiliary contacts on th e DC con tacto r ope n ed and the coi l circuit deenergizcd.

tests and two lic ense te s t procedures were performed as required by the Upon completion of the safety sys t em respon se, short duration trip

technical spec ifi ca tions. They were as fo l lows:

W-0-5 - Test of th e Main Seconda ry F low Monitor W-0 -15 - Test of the Reactor Core Outl e t Tempe ratur e Monit or

F. Compliance Examination of License e Inv estigation

CO:II and CO:HQ that one relay had failed, Initial r e view o f th e electrical G-E had state d in an ear l ier telecon between the s it e personnel,

Exhibit B circuits r eve aled the poss i bility Lhat two. an d po ss ibly four, relays had failed. After d i scuss ing thi s matter with the s it e pe r s onn el, th r.y concurr e d with our observation. During th e vi sit to th e o it e and aft e r reviewing th e data tak e n relatin g to the timing rclation s hlp be twee n the auxil i.ary and ma in contact s, th e r.e lay which initi ated the s cram :i long i: I with it s scram drawer was tentat ive ly id e ntified as FlKl.

Since one scram drawer ("D" bu s ) w.:is in a scram condition, a ny trip signal coming from E and/or F draw e r s would ~a ve caused a sc r am to occur. Since each drawe r has two DC contactors connect ed in para l lc 1, !

both will respond to the same trip signal; therefore, two and po ssib ly,*

four had to malfunction. G-E stated that the failure of th e cont ac ts within the same relay to operate in the proper sequence as being t he cause of the event. Review of th e settin gs taken before th e initi a l co~

tact adjustments were made revealed that of the three " good" contactor s,

scram draw e rs E and F had one each. Thi s means that even if a short dura tion trip occurred then the scram would have gon e to compl e tion since th e auxiliary contacts would have released on the "good" scram relay be fore the main contacts. G-E s tated that the contactors do not have the sam e release a nd pick-up characteristics, even thou gh they are id e ntical.

However, further exam ination of th e ir te s t data showed that E2Kl had a. I shorter tim ~ response characteri s tic than its redundant count e rpart ElKl. I G-E then s tated that the response time data tak e n was an av e rag e and not I ab so lut e accurate measurement. This insp e ctor th e n ob se rv e d t ha t "F" drawer had one "good" and one faulty r e lay where the faulty r elay had a shorter release ti me characteristic than its counterpart F2Kl. This indicate s that FlKl was the relay that "failed." G-E did not disagree with thi s observation.

During th e meeting held in Auter b urn's office thi s insp e ctor had I asked if the scram reset circuit had been considered susp e ct at any ti me ' t l during the inve s ti gation and G-E stated that it had not. G-E agreed tha t a failure in th e reset circuit would pro d uce the same re s ult s th a t occurr e d.

Addition al discussion on this matter r evea led that the last time th e r ese t circuit has been te s ted was durin g the initial pl a nt preops (two years ago ).

• . Paul Swartz then s tated that a letter would be sent to th e site r ecommending testing of t h i s circuit at a more frequ e nt rate. Swartz also s tated that.,,

in ad -~ ition to this recomm e ndation, periodic testin g of the timing rela tionship between all DC contactors us e d in both s afety and cont ai nme nt isolation sys tem s will be t es ted. Auterburn s t ated that he would wai t until a 11 t he in forma t ion r ela tin g to t he r e commended t ests was revi ewed,_

before he would make a statement of agreement with these recomm e ndations.

lized a s p:,rt of the safety and s afegu:ird s y s tem s crnm and actu.1tinr. ci r c uit~ The review and examination of t he DC contactor t ha t i s be ing uti

t:,.v,.nl,*<I to tld. 11 i11 1:pr,ctnr th:1L th,* lll<'thrn l nf: :itl_l1111t 111 ~~ th,, :imd I i11 ry c-011-t11c _l11 Jor II dlfl,-r,*111. Ll111l11~ 1*1*l11Ll1111,,lilp w:111 t : 1!1*0111-'.I* 11i:111111tl li,*11oll11i~ ol tnl>s. Al 110 the u:i,~ of the auxiliary contact :i a s 11 H<\\111-111 control LOL i.Ls own coil circuit when not mounted on the s ame actuator l>oard as th e main

Exhibit B Page 8 of 9

• • ", I

contacts is of qu estio nable pra ctic e. This was discu~sed,,ith G-E personnel without th eir Cvlt'ments.

in the technical specifications for the safety system (M0-9) secondary The s ite peL*sonnel follm,ed all testing r equirem ':! nts set forth

flow monitoring system (W0-15). Test No. M0-9 - S.:ifety and Safeguards S)6 terns was run aft e r the initi al DC contactor actuato r ctrm tab adjust ment on September 13, after th e contain me nt isolation building DC contactor tab adjust Qe nt performed on September 20, and finally a ft e r the s hort duration pulse te s ts. Test W0-15 - Reac tor Co r e Outlet Temperatur~

Monitors was r un following the short duration pulse test.

secondary flow controller is common to th e three low flow safe ty tri ps The flow set point adjustrrent signal ori&inating from the main

units. Since thi s i s a deviation from the separa tion (channel indepe nde n ce) requirements set forth in IBEE-279, the matter was brought to th e at te ntion of DRS, DRS stated that the item had been r e vi e wed durin g the d esig n r e vi e w stages and was accepted. No further examination of this item was made by the writer.

nuclear facility that utili~ed this type contactor in th e safety system Mr. Swartz stated that EVESR nuclear plant is the only other

He also stated that EVESR facility is no longer in operation.

Exhibit B Pa ge 9 of 9 U. S. ATOMIC ENERGY COMMISSION REGION II

DIVISION OF COMPLIANCE

Report of Inspection

CO Report No. 50-231/71-2

General Electric Company and Southwest Atomic Energy Associates (SEFOR) jcensee:

License No. DR-15 Category B

March 17-19, 1971

)ates of Inspection:

January 6-8, 1971 Jates of Previous Inspection: ~ -Z.0-7/

Inspected By: C. Inspector (Operations) Dat e C. M. Upri (In Charge o(?_Y. ~

R. )l:* C bitt, Re tor Inspector (Operations)

- ~~~ ~

Reviewed By* W. C. Seidle, Reactor Inspector

None Proprietary Information:

SCOPE

A r outine, announced inspection was conducted at the fast spectrum, sodium cooled SEFOR reactor located approximately 20 miles southwest of Fayetteville, Arkansas.

SW-iHARY

Safety Items - None

!_oncompliance Items - A Form AEC-592 was issued for the following items of noncompliance:

,p

• t No. 50-231/ 71-2

~eactivity Control and Core Physics

1, Failure of Core Outlet Temperature Scram.!/

Arterburn notified Region II by telephone on February 25, 1971, that the outlet temperature scram had failed during a routine surveillance test conducted on February 21, 1971. A scram signal was not produced when a test switch was actuated to simulate high core outlet temperature. Three temperature inputs are connected to 1/3 scram l ogic and the other two trips functioned properly.

G-E reported the failure to DRL in a letter dated March 2, 1971.

The failure was discussed with Mathis and found to be correct as reported.

The failure was caused by relay contacts which failed to open I and deenergize 125 VDC relays that initiate scram action. The contacts were believed to be damaged by a cu rr ent surge back 11 through scram chassis C when 125 VDC power to the scram relays tripped off during a checkout of system modifications. (See Section F.4.) Scram busses A and B were deenergized during the checkout and experienced no contact damage. All other relay contacts in scram chassis C were checked for proper operation. The chassis was removed and inspected for damage to components

or wiring. The faulty relay was replaced and satisfactorily tested.

To prevent a recurrence of this failure, Zener diodes were installed in parallel with the coils of all 18 Kl relays used for reactor scram and containment isolation. The Kl rel ays are shown on Drawing No. 197R234 in FDSAR Supplement 2. Proper system opera tion was checked by LTP M-0-9, " Reactor Safety System Test,"

after completing the modifications.

The failure and subsequent modifications were r eviewed and approved by the Site Safety ColIUllittee as required by the Technical Specifications.

2. WRM - Heat Balance Correlation~/

Administrative control of high flux trip points was reviewed as specified in Engelken's memorandum to Skovholt dated November 27,

Yrnquiry Memorandum dated March 2, 1971.

Yeo Report Nos. 50-231/70-4 and 71-1, Section F, and letter from G-E to Region II dated November 3, 1970.

U.S. ATOMIC ENERGY COMHISSION REGION II

DIVISION OF COMPLIANCE

Report of Inspection

CO Report No. 50-231/71-1

ljlllsee: General Electric Company and Southwest Atomic Energy Associates

(SEFOR)

License No. DR-15 Category B

J,.S of Inspection : January 6-8, 1971

of P r evious Inspection: December 16, 1970 z-6 -7/

ected By: Inspector (Operations) Date I

~ I

irkparick,Reactor Inspector (Operati o ns)

~ Date ii Specialist 11

~fo/'II 11 ewed By: Date ;1 Reactor Inspector 'I

ffietary Inform a tion: None

SCOPE

utine, announced inspection was conducted at the fast spectrum, um cooled SEFOR reactor located ap9roximately 20 miles southeast 1ayet teville, Ar kansas.

SUl1HARY

~ty I terns - None

~t. No. 50-231/71-1

~eactivity Control and Core Physics

1. Wide-Range Monitor - Heat Balance Correlation.!./ I.

Arterburn informed the inspector during the previous site inspe c 1:

tion that discrepancies between WRM indications and reactor heat balances would be reported to DRL within 30 days in accordance with paragraph 3. C. (2) of the license. A report was submitted on November 3, 1970, but was sent to the Region II Compliance office instead of DRL as originally stated. A copy of the report was forwarded to Compliance Headquarters on November 1q, 1970.

Administrative controls regarding adjustment of the high flux trip settings were reviewed as specified by Engelken's letter to Skovholt dated November 27, 1970. WRM trip points were at 95%

and have been left set down since the discrepancies were first observed according to Arterburn and Becker. Tr ip points will not be reset to 105% until reactor is increased above 17.5 Mw (87.5%) at which time reactor core average temperature would be maintained at 760°F. Only one series of tests remain to be completed with core temperature below 760°F and the method of adjusting WRM trip points will be specified by the test pro cedure. Arterburn and Becker stated that the decision had not been made on whether to recalibrate to the lower temperature or to set the trip points down based on existing data.

Becker stated that the No. l WRM has been generally indicating higher than No. 2 or No. 3 for a given power level. The difference appears to be drift in the No. l system but bench tests have not revealed any component changes that could cause drift. There is no reason to believe that the de tector gamma compensation is changing. Becker stated that a program is being formulated which will determine if there is drift in the detector or in the amplifier. Shift personnel have been instruc t ed to notify Becker anytime a WRM is not withi n 2% of the reactor heat balance.

WRM operation and high flux trip setting will be reviewed during the next inspection.

2. Partial S~ram'l-_/

Arterburn stated in the Management Interview following the pre vious site inspection that consideration would be given to

CO Report No. 50-231/70-4.

CO Report No. 50-231/70- 4, Section F.3.

p.pt. No. 50-23 1/71-1

revising existing surveillance tests to assure proper adjus t ment of the auxiliary lockout contacts. Discussion of proposed tests with Mathis indicates that satisfactory testing will be performed. An annual system test will require a complete inspec tion of the scram relay (K-1 ) cabinets as well as insertion of a short duration scram signal while measuring the relative opening time of the main and auxiliary contacts.

This type of test was conducted to assure p roper system operation following the partial scram discusse d by V. C. Thomas of Compliance Headquarters in Appendix B of CO Report No. 50-231/70-4. Mathis stated that the test would be used following modification of the scram relays to provide the 400 msec delay required by operation of the Fast Reactivity Excursion Device (FRED). The procedures to accomplish the testing are in preparation and will be reviewed during the next inspection.

Auxiliary Systems

1. Failure of Main Secondary Expansion Tank Level Trip

Arterburn notified the inspector by telephone on January 4, 1971, that one of the three low level trips on the main secondary expansion tank had failed to trip during monthly surveillance testing on January 3, 1971. The other two channels functioned properly and provided a scram signal from the two-out-of-three logic. The occurrence was reviewed during this inspection.

During the completion of monthly surveillance test M-0-3,

lowering the expansion tank level failed to cause a low level trip input from safety chassis F. Safety chassises D and E tripped and provided a scram signal through the two-out-of-three logic circuit shown in FDSAR Supplement No. 2, diagram No. 197R234 of the reactor safety system. When the instrument mechanic moved the trip unit chassis associated with L.S.

244-6, one of the relays apparently dropped out giving a trip input from safety chassis F.

Either one of two relays in the trip unit failing to drop out or having contacts that failed to open could have caused the malfunction. One of the relays uses mercury wetted contacts and also provides alarm functions as well as the low level trip signal. This relay is controlled by the other relay in the unit which operates on a 50 Mv signal from the level probe and is considered to be more susceptable to such a failure.

According to Hixson an d Mathis, this is the first failure involving either type of relay.

~*:,-* -- -,.-,-- *- '

I,* * -. '.. l C r - -