ML20101J152

Provides Info Clarifying Position Re Operability Determinations Resulting from IST of Pumps & Valves,Per Generic Ltr 89-04,per NRC Request During 920428-29 Workshop on Generic Ltr in King of Prussia,Pa

ML20101J152
Person / Time
Site:	Three Mile Island
Issue date:	06/24/1992
From:	Broughton T GENERAL PUBLIC UTILITIES CORP.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
C311-92-2092, GL-89-04, GL-89-4, NUDOCS 9206300373
Download: ML20101J152 (5)

Text

- _ _ _ _ _ _ - _ _

GPU Nuclear Co:poration (uclear i, U =',,wsgreo Middletown, Pennsylvania 17057 0191 717 944 7621 TELEX 84 2386 Writer's Direct Dial Nurnber:

(717) 948-8005 June 24, 1992 C311-92-2092 U.S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, DC 20555

Subject:

Three Mile Island Nuclear Station, Unit 1 (TMI-1)

Operating License No. DPR-50 Docket No. 50-289 Inservice Testing (IST) of Pumps and Valves - Operability On December 3, 1990, GPU Nuclear responded to NRC Generic Letter (GL) 89-04 concerning the Inservice Testing (IST) and o)erability of pumps and valves at THI-1. GPU Nuclear intends to fulfill our o)li license, Technical Specifications (Tech Specs)and gations relative other NRC re to the However our response took exception to Position 8 of GL 89-04'quirements. which calls for declaring a component inoperable, without proper evaluation, when the test data for a component test falls within the ASME Section XI " required action range." GPU Nuclear has concluded that to do so could ultimately result in entering a 72-hour Tech Spec allowable outage time (A0T) and lead to an unnecessary plant shutdown.

Discussions with the NRC during a April 28-29, 1992 workshop on Generic letter 91-18 in King of Prussia, PA, as well as operating experience subsequent to our December 3, 1990 letter, have reinforced our desire to contic.ue our ongoing discussion of operability considerations regarding the pump and valve IST program. The purpose of this letter is to clarify our position regarding operability determinations resulting from IST, to demonstrate that our g, approach is consistent with the general view espoused by NRC management, and ono to attempt to achieve consensus with the NRC staff regarding this approach, gga. It is noteworthy that the history of component test operability detsrminations oc as a result of IST are a rare occurrence at TMI-1. Typically we have g8 experienced this condition one or two times each year and, in almost every instance, the final evaluation proved the component to be operable.

SU OQ f.o 1

This exception applies also to section 6.11, " Technical Specification Operability vs. ASME Code Section X7 Operative Criteria," of the NRC ,

om Inspection Manual Part 990o, which was transmitted to licensees as an 7

/ j k enclosure to NRC Generic Letter 91-18.

l Q Q O 'j 3 GPU Nuclear Corporanon is a subsidiary ct General Pubhc Ubbt<es Corporation

e C311-92-2092 Page 2 GPU Nuclear believes there are greater risks associated with unnecessarily shutting down the plant during an operating cycle than those associated with operating while a component's continued o>erability is being evaluated. We endorse the general guidance promulgated ay Dr. T. Murley in his memorandum of July 19, 1989. In his memorandum Dr. Murley stated that after careful consideration, NRR had concluded that guidance was not necessary concerning the length of time a licensee should be allowed to decide the question of operability when an equipment deficiency is discovered. He stated "There is no generally ap)ropriate timeframe in which operability determinations should be made." Furtier, Dr. Murley went on to note that equipment which is in fact inoperable should be immediately declared as such and the appropriate Tech Spec timeclock entered. However, when a question is raised regarding the operability of a component, operability determinations should be made using the best information available (e.g., analyses, a test or partial test, experience with operating events, engineering judgement, or a combination of these factors). We have and will continue to adhere to these guidelines.

Actions taken at TMI-; over the years support our position that compliance with this philosophy is prudent.

The central issue related to an nperability determination is the component's ability to perform its intended function. A determination of operability based solely on generic limits, narrowly defined by ASME Section XI tables or formulae, may force a judgement to declare a-component inoperable when, in fact, the component will perform its intended function acceptably and predictably. Attachment 1 provides examples where TMI-l components have been determined to be in the " required action range" and remained fully functional.

Consequently, GPU Nuclear consiaers it imprudent to start the Tech Spec timeclock when a component has not been evaluated and may be fully able to perform its intended function.

GPU Nuclear intends to assure ccmpliance with NRC requirements, including the TMI-l Tecn Specs. However, GPU Nuclear believes it is appropriate to adequately evaluate test data before making a determination of liioperability which could potentially force a shutdown of the plant based on acceptance criteria which ray prove to be excessively conservative. We remain committed to the premise that obviously inoperable components (i.e., broken shaft, pump / valve will not move, etc.),- will be declared inoperable immeciately per Administrative Procedure (AP) 1041, and the appropriate action will be taken as' directed by Tech Specs. In regard to pump tests, AP 1041 states: " If confirmed test data falls within the ' Required Action Range' and the pun.p is clearly inoperable in the N gement of the Shift Foreman, the pump shall be declared inoperable and the acclicable Tech Spec requirements observed." In regard to valve tests, AP 1041 states: " If confirmed-test data exceeds the-

' maximum allowable stroke time for a MOV (or if a check valve fails to move to its safety position

the SF/SS, the Tech) Spec time clock shall be started per the applicable Techand the valv Spec."

t Therefore, based on the above rationale, GPU Nuclear has concluded thai. it is not appropriate to declare a component inoperable basu solely on IST results which fall within the " Required Action Range" as proposed in Position 8 of

, GL 89-04. When a component's IST results fall within that range, the preliminary determination will be made that s component is operable unless its inoperability is apparent. Operabili+y determinations are based upon everything we know about that comporent. GPU Nuclear will apply the resources t

r e *-,,..rv3-y,,s,., -

,a.- ,4- vv ----+m-,..e -,+-~m ~m.- ..--,.--,,vv . = ,----w-- .- c 4 e .r =~ " - - - - -

C311-92-2092 Page 3 .

1 necessary to complete a timely evaluation to support the initial determination '

of operability when the results from IST fall in the " required action range."

We agree with the statement made by the NRC during the GL 91-18 workshop that the time for evaluation will depend on safety significance. When a judgement based upon evaluation determines that a component is in fact inoperable, adherence to the Tech Spec A0T, retroactive to discovery of the discrepant data, will be effected.

Sincerely, l

T.G.Broub[tondNbww t/<%

Vice Presi ent and Director, TMI-l i

i MRK Attachment cc: Region I Administrator TMI-l Senior Project Manager TMI Senior Resident inspector l

i-l l

l

~

l l

ATTACHMENT l MS-V13B HS-V138 is an air operated valve which supplies steam to the turbine driven Emergency Feedwater (EFW) Pump, EF-P1, one of three safety related EFW pumps.

Since 1988, the stroke time for MS-V13B had been 19 to 20 seconds with a maximum allowable stroke time of 23 seconds. On 12/29/91 at ~0200 hours, its stroke time was 24.7 seconds which would have rendered MS-V138 immediately inoperable under strict GL 89-04 guidance. Tech Spec 3.4.1.1 requires the operability of all three EFW pumps. If one pump is inoperable, it must be restored in 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> or the plant is required to be in cold shutdown within the next 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />.

In an initial telephone call on 12/29/91 at -0800 hours, MS-V138 was evaluated to be operable based upon the satisfactory EF-P1 startup time data. Extensive analysis by the GPU Nuclear's technical support organization was completed on 2/11/92 whi .h showed that the maximum stroke time of MS-V13B could be increased to a conservative value of 27 seconds. It is noteworthy that the solenoid valve for MS-V138 was replaced on 1/25/92 leaving the "as lef t" stroke time for MS-V13B of 17.5 seconds.

MU-V51 MU-V51 is a 1", 150 lb Grinnell air operated diaphragm valve which fails closed. The valve opens with air pressure on one side of a diaphragm and closes by spring force when a solenoid valve bleads the air off. On 7/10/91 the measured stroke time of MU-V51 was 3.8 seconds. This exceeded its 2 second maximum allowable stroke time. Restroking MU-V51 gave acceptable stroke times of 1.6 and 1.5 seconds. The valve's stroke time was -1 second in 1988, ~1.2 sec in 1989, and -l.8 sec in 1990 and 1991.

The safety _ function of MU-V51 is to open and supply concentrated boric acid from the Boric Acid Mix Tank (BAMT) or Reclaimed Boric Acid Tanks (RBATs) to the Makeup Tank (MUT) if the Borated Water Storage Tank (BWST) is lost. -The MUT-is pumped to the Reactor Coolant System (RCS) by the High Pressure Injection (HPI) Pumps. As provided in the Basis of Tech Spec 3.2.2, the RCS can be borated-to one percent (1% Ak/k) subcritical at the worst time in core life with a stuck control rod assembly in less than-13 hours using one of the-10 gpm BAMT Pumps. In_ addition to the MU-V51 line, there is an alternative flow path through WDL-V61 from the BAMT or RBATs to the MUT. Operator action is required to start the pump to begin addition of concentrated boric acid to the MUT. The above shows that it takes significant time to add a sufficient amount of boric acid such that a stroke time of 3.8 second for MU-V51 has no significant impact on pumping time.

On 11/1/91 maintenance was performed on MU-V51 when tests demonstrated an increasing stroke time. The solenoid valve operator was found rusty with interior scale deposits. MU-V51 stroke time following maintenance was

-1.0 second.

4

'[ SF-Pl A/B THI-l FSAR, Table 9.4-1 states that the capacity of SF-PIA /B is 1,000 gpm at 100 ft. A flow rate of 1,000 gpm per pump ensures that under the highest heat load (574 spent fuel assemblies in storage accumulated from 11 successive

. yearly refuelings and the addition of the entire core,177 assemblies during the twelfth refueling, completed in 150 hours0.00174 days <br />0.0417 hours <br />2.480159e-4 weeks <br />5.7075e-5 months <br />) the pool temperature will not exceed 147'F. Therefore, 1,000 gpm from SF-PIA /B delivered to the pool (s) satisfies the safety function of SF-PIA /B. SF-PIA /B is operable if 21,000 gpm can be delivered to the pool (s) and vibration is satisfactory.

In 1985 SF-PIA /B were in the AP required action range. At 1000 gpm SF-PIA had 85.6 ft of head and SF-P13 had 83.3 ft of head. For SF-PIA this was 14.4%

less than the design point (100 ft and 1000 gam) and for SF-PIB this was 16.7%

below the design point. Pump vibration was extremely low (0.21H for SF-PIA and 0.15H for SF-PIB). ASME Section XI, Table IWP 3100-2 states that the required action range is <.9 times the reference value. For SF-PIA:

85.6 ft X (<.9) - <77 ft. For SF-PIB: 83.3 ft X (<.9) - <75 ft. For SF-PIA and SF-PIB trains the head loss (both piping frictional loss and elevation head) has been estimated to be <50 ft by standard head loss calculations in accordance with the Crane Handbock and other text book references. This indicates that the head of SF-PIA /B could drop to 50 ft and 1000 gpm still could be delivered to the pool (s). Therefore, <75 ft as a required action range is clearly acceptable and conservetive. IST Testing in 1991 shows that SF-PIA has remained at -85.6 ft at 1000 gpm while SF-PIB has dropped to -80 ft ,

at 1000 gpm. 30 ft is greater than the <75 ft chosen as the required action range.

)

i i k

-~~-- _- . , ,,_,__..__..__,.m_ _.,,,.m__ . . . _ . _ , , , _ . . . . . . , - . _ _ , _ , . . . - . _ _ .