ML20085D813

Requests Waiver of Compliance from Surveillance Requirement 4.8.1.1.2.d.3 Re Emergency Diesel Generator Overshoot Limits Until Permanent Tech Spec Amend Can Be Submitted & Approved

ML20085D813
Person / Time
Site:	Sequoyah
Issue date:	10/10/1991
From:	Joshua Wilson TENNESSEE VALLEY AUTHORITY
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9110170180
Download: ML20085D813 (6)

Text

. .

4 b

. m cg. ,<

,- e. ,

mm a: ,-

Jt% L & JMt s_-6, .4 3 , , , , , ,

October 10, 1991 U.S. Nuclear Regulatory Commission ATTN: Document Centrol Desk Washington, D.C. 20555 Gentlemen:

In the Matter of ) Docket Nos. 50-327 Tennessee Valley Authority ) 50-328 SEQUOYAli NUCLEAR PLANT (SQN) - REQUEST FOR WAIVER OF COMPLI ANCE FOR " NITS 1 AND 2 SURVEILLANCE REQUIREMEFT (SP) 4.8.1.1.2.d.3 AS APPLIED TO EMERCENCY DIESEL GENERA 13R (D/G) VOLTAGE OVERSHOCT LIMITS This letter documents TVA's request for s waiver of compliance with Tecnnical Specif i cation SR 4.8.1.1. 2.d. 3 f or SQN Ur.its i and 2 as applied to emergency D/G ve tage overshoot limits. The waiver of the subject SR voltage overshoot limita until a permanent TS amendment can be submitted and approved was determined to have no safety significance.

NRC verbal approval of the subject waiver with an administrative overshoot limit of 115 percent of pretest voltage was provided by G. C. Lainas of NRC to R. J. Beecken of TVA in a talephone conference call at 0020 Eastern daylight time on Octeter 9, 1991. This letter roquests NRC approval of a waiver of compliane: fer the same TS SR with an administrative overshoot limit of 120 percent of pretest voltage. Details concerning the justification for the waiver are prosidtd in the enclosure.

l l Please direct any questions concerning this issue to M. A. Cooper at l

(615) S43-8924 Sincerely,

\

IO'b i

q. L. Wilson Enclosure See page 2 Q

"M ggc:*fabinshh8@g7 1 i

P

. _ _ ._ .~ . - _ _ ,._ .

A

..; '-i n i

P U.S. Nuclear Regulatory Conmission ,

October 10.-1991 cc (Enclosure):

American Nuclear Insurers +

Attention: Librarian The Exchange, Suite 245

'270 Farmington Avenue Farmington, Connecticut 06032 L

l-l Mr. D. E. LaBarge, Project Manager U.S. Nuclear Regulatory Commission One White Flint,_ North-11555 Rockville Pike Rockville, Maryland 20852  :

Mr. Michael H. Mcbley, Director (w/o Enclosures)

Division of Radiological-Health  ;

._T.E.R.R.A. Building 15G 9th Avenue, 5

. Nashville, Tennest ee 37203

-NRC Resident Inspector Sequoyah Nuclear P.' ant 2600 Igou Ferry. Road-o Soddy Laisy Tennessee 37379 Mr. B. A. Wilson, Project Chief U.S. Nuclear Regulatory Commission Region II

[ 101 Marietta Stre -t, h"d,- Suite 2900 L Atlanta, Ge:rgia 30323 l'

4 4

+

1 9

. INCLOSURE gndItlon -

During performance of Surveillance Instruction (SI) 1-SI-OPS-082-20.A. "I.ons of off ite Power With balety Injection-D/G 1A-A Gontainment li.nl a t ion Test,"

on Dichel Generator (D/G) lA-A on October 7, l '! 91, the resulting transient overvoltage !:11 lowing a ful1-iond rejection of :ated load (l.c.,

4,400 kilowntt (kW), 5,000 kilovolt ampete (kVAl) exceeded the limita an entablished in Technten1 Specitlea ion (TS) Surveillance Requirem.nt (SM) 4.8.1.1.2.d.). The TS limitn the trnnuient overvoltage to the lower value of 114 percent of the pretent voltage or 8,276 volts 'V). the pretransient voltage prior to the tull-load rejection wuw me u. u r e d to be 7.175 V. The resulting t annicut overviltage t01 lowing the ti 11-lond rejection was measuted ott the vluiconder tracen to be approxinately M,250 V (i.e., approximately 11$ percent of 7,115 V). The is limit wan M,180 V (i.e.,

114 percent 01 7,115 V ) . The duintioti nt the voltage trannient above the limit was approximately 1:1 cycles or 0.22 second. D/G 1A-A was declaimi inoperable at l915 Eastern daylight time (EDT) on Oc tober 7, 1991. The 72-hout netton under Limiting Condition of Operation ( LCO ) 1.8.1.1 was entered for Unit 2 and would expire at 1915 E90 on Oc aber 10 1991. At 0020 EDT on October 9, 1991 NBC provided verbal approval of a waiver of compliance with SR 4.8.1.1.2.d.1 an it pertainn to voltage overshoot limit.. An part ot the basis for the approval, TVA cammitted to administratively e.tablle.h voltage o overrhoot limits of 115 petrent of the pretransient voltagn until a permanent TS amendment was approved.

During jerformance of SI 1 -S I -Ol'S-08 2 - 26. h , " Lou s ot Ulfhite l'ower Ul th Sat r t y injectlon-D/C ik-B Contniament !s,olation Tent," on D/G 1B-B on October 9, 1991, tha resulting transient overvoltage tollowing a t u; 1-lunt rejection of rated load wan nreuuced at 8,220 V, which was approximately

[19 percent of the pretest voltage ot 6,9.10 V. D/G lb- h was declat e t inoperable e' 1252 EDT on October 9, 1991. The 72-hour action undet LCO 3.8.1.1 was entered for Unit 2 and will expire at 12S2 i<.DT on October 12, 1991.

The D/Cu are commen f or both unit s at equoyah Nuclear Plant (SQN). Unit 1 is currently in finde 5 tor the Cycle 5 retueling outage. Unit 2 in currently in Flode 1 and ope W ng at appr oxirantc ly 100 pei cent power. ,

Backgroun1:

TVA processed a TS change reipte n t i n Janua r y 1"89 to mod i f y Sk 4.8.1.1. 2.c .1 to the acceptance voltage Naluer listed above. The valtage iimit of 8,276 was based on 114 percent of the expected ma41 mum plant operat ing vol tage of 7,260 V. The 114 percent at pretest voltage valo" was baned on pant test results that indicated approximately 112 percent voltage overshoot and was n it haned on any <riteria an+.ociated with equipment damage.

Since the approval and hplementation at amondments t u r tln i t s 1 and .' that incorporated these valuca, modifications have been imnlemented to the I

transient voltage response of the D/G. The mediticationa have been to the generator' exciter control system that altered the excitation rurrent t ran4 t or me r (CT) tap set. tings to ptovide a boost or the tield current for a given D/G kVA loan and addition et the voltage overshoot redo ( tion device (VORD) that ahunta the fleld current to zero it the trnonient overnhoot during

s.

the D/G loading sequence exceeds 103 percent of nominal during ischronous mode-of operation.: The VORD-is-not active in1the Did circuit during performance of the load rejection test that is performed in the' droop mode. This'is because the shutdown board's normal voltage' range during plant operation exceeds the 103 percent operating setpoint of the VORD.

In addition, procedural changes have been recently incorporated that require generator loading to the design rating of 4,400 kW and 5,000 kVA. This change i in' testing was incorporated as a result of industry operating experience information (NRC-Information Notice 91-13) and is mor conservative than previous test conditions. This change results in an increase in the internal generator voltage =setpoint, which is required to overexcite the generator to be able to obtain a rated kilovoltampere reactive (kVAR) output of approximately 2,375. This change also increaset generator-field current.

Both the higher-internal voltage and the additional field current tend to ,

. create a higher voltage overshoot when the rated load is suddenly removed.

An additional. factor that contributed to. identification of the overshoot was y the use-of highspeed visicorder data to determine voltage overshoot.

L Previously, voltage overshoot had typically been determined from installed L panel gauge data. This panel instrumentation had significantly reduced capability for identifying high-speed, short-duration transients.

Technical-Evaluation of Deviations:

The-described conditions that identified-voltage overshoots in excess of TS limits and previously measured values are believed to have been caused by the s combination of the revised procedural requirements to test at rated generator load-(resulting from review of IEN 91-13), the CT modification, and the use of more accurate instrumentation capable of identifying short-duration transientst The combined effect of these changes increases the voltage overshoot of the_D/G during load rejections. The individual contributions of each: change are not known and would require additional D/G teFting that is not

recommended for reliability considerations. _ However, it is considered that J the changes in test kVA' loading are the driving contributor to this

! condition. The individual changes, as well as the combined effect of these

}; changet, were'not-applicable or were not expected at the time of the TS change l

reque.ted in 1989. Additionally the effect of these changes could not be quantified at the time of initial implementation and therefere was not-identified until performance of the subject testing.

There is' negligible safety significance associated with the subject voltage

. overshoot levels. The test deficiencies were of very brief durations (13 and 23 cycles for lA-A and 1B-B respectively) and of small magnitude relative to

-eqaipment considerations. The briet transient overvoltages (8,250 and 8,220 V for 1A-A andLIB-B respectively) would not overstress the generators' insulation system. Insulation life is inversely proportional <to voltage stress, but the duration of these transients is so small that the effect.on insulation' life would be negligible. The generator manufacturer has recommended to SQN that.an acceptance criteria of'S,712 V for a maximum duration of-30 cycles'would not overstress the generator insulation. The D/G l

l 1

l

l

~3- '

-vendor has determined that the engine and/cr generator controls would not experience detrimental effects for transient voltages up to 9,000 V.

Accordingly, it ~is considered that overshoot limits far in excess of those -

provided in citiiet 'cutrent SQN or standaro TSs (i.e., in excess of 120 percent) are technically adequate to prevent equipment damage or degradation following a full-load rejection.

In summary, while the testing results exceed the SQN TS overshoot value, the ,

subject voltage levels have not and will not adversely affect D/G equipment or capability to perform the intended function.

The' difference in transient voltage overshoot of 115 percent of 7,175 V for D/G 1A-A and 119 percent of 6,900 V for D/G 1B-B is primarily due to the manually adjusted internal generatar voltage setpoint required to overexcite

'the generator to be able to obtain a rated kVAR output of approximately 2,375. From a comparison-'of the test results for the two D/Gs, it was determined that the required internal sencrator voltage setpoint was approximately equal on both machines (i.e., 7,530 V) even though the pretransient voltage on the shutdown boards was different. This means that field current will flow for a longer duration on D/G 1B-B because of the -

greater voltage _differeuct between pretransient voltage and she internal voltage. The increased duration was documented in the test results (i.e.,

13 cycles for D/G 1A-A and 23 cycles for D/G 1B-B).

The potential effect of this' condition on the Unit 2 D/Gs was evaluated.

Although the absolute value of overshoot in the Unit 2 D/Gs is unknown, it would be expected that an increase in voltage overshoot would also be observed

~

v.nder- the revised testing conditions. However, as previously ci cussed, the possible voltage overshoot levels do not result in any technical or safety concerns. If further scheduled testing identifies potential for higher overshoots, generator- droop can be manually adjusted.

Based on the low safety significanco of the test defi lencies, it ir not considered prudent to subject the 6/Gs to additional testing in further

, attempts-to satisfy the existing TS value. Retesting requiring _further switchyard or power alignment configuration changes la additionally not considered prudent from an outage riuk minimization standpoint.

' Request:

TVA requests a waiver of compliance with TS SR k.8.1.1.2.d.3 for _ both SQN Units 1 and 2. TS SR 4.8.1.1.2.d.3 requires verification of diesel capability to reject a load of 4.400 kW without tripping; the generator voltage sitall not exceed 114 percent of tne initial pretest voltage or 8,276 V,=whichever is-less, during and following the load . ejection. Testir-g conducted en October 7 and 9, 1991, verified the capability of the Unit 1 D/Gs to accept a full-load rejecticn without tripping. The voitage overshaot values-that were

experienced exceeded the current SQN TS values, but were well below voltage-values, which would be required to effect damage or degradation to the D/G hardware.

I __

f

,, , _ . _. - - ~ . . _ . _ - ~ _ _ , . , _ _ _ _. .-~. _ _ . . . . .- ._

l-

-n ,

-The cause of the overshoot is believed to have resulted from several factors

-as previously discussed ard was not anticipated. The overshoot, combined v ths collective testing on the D/G,-does not indicate reduced capability to sustain 4r full-load rejection or to properly operate the voltage regulator. The Th value_of 114 percent is rot based upon criteria associated with equipment da nage. The D/G vendor I.as confirmed that voltage f ar in excess of this ':alue would not overstress genurator equipment. Standard TSs currently utilize ta value of 115 percent of nominal voltage, and improved TSs are proposing i ialue of 120 percent of nominal voltage. In summary, the test results do not indicate any impairment to D/G functionality or capability.

TVA will, propose to modify the SQN Units 1 and 2 TSs to revise the subject acceptan:e' criteria, e.g., raising the allowable voltage overshoot limit to

- 120 percent of pretest voltage,' by submittal an.i approval of an emergeacy TS change. In the interim, until the amendment is approved, TVA requests a

, waiver of compliance:with the requirements of TS SR 4.8.1.1.2.d.3 regarding L voltar,e overshoot limits. TVA will administratively establish voltage overshoot limits of 130 percent of the pretest voltage until -the emer.dment is approved. Prompt ac* ion by NRC to grant this waiver is needed to prevent unnecessary additional testing of the D/Cs and to prevent an unneces sa ry shutdown of Unit 2.

The requested action is viewed to have no safety significance becatse the l j subject deviation does not indicate any impairment of ability of the DGs to i

iully-perform its intended function. Accordingly, the requested waiver cannot

-result in=an increase ~in the probability or consequences of a previously evaluated accident, cannot create the possibility of a new accident, and cannot reduce the margin of safety. Therefore, the action does rot, 'nvolve a significant ha-ards consideration. Similarly, the waiver does ntt involve an L unreviewed env +onmental question because it does not increase ai) adverse L environmental imptets, change ef fluents or power levels, or result in j unreviewed enviroamental matters.

1

.PLO90205/109 l-L-

- ~ , _ - , - - _ _ . . _ - . . - - . . - . - - - - - . - , - , _ _ . . , - , , . _ - - , , - - - - . , , .-r. __,