ML20134A493

Forwards Evaluation of Util Rept Re Overspeed Trips on Auxiliary Feed Pump Turbines,Overspeed Trip Throttle Valve Problem & Main Feed Pump Failure.Response by 851104 Requested

ML20134A493
Person / Time
Site:	Davis Besse
Issue date:	10/30/1985
From:	Stolz J Office of Nuclear Reactor Regulation
To:	Williams J TOLEDO EDISON CO.
References
TAC-59702, NUDOCS 8511070128
Download: ML20134A493 (17)

Text

i October 30, 1985 g M Q Cab Docket No. 50-346 DISTRIBUTION G0ick QDcket F W ACRS-10 NRC PUR Ringram L PDR ADe Agazio Mr. Joe Williams, 'Jr. ORBf4 Rdg CMcCracken Vice President, Nuclear HThompson Gray file Toledo Edison Company OELD EBrach Edison Plaza - Stop 712 EJordan H0rnstein 300 Madison Avenue BGrimes WPaulson Toledo, Ohio 43652 JPartlow GEdison

Dear Mr. Williams:

l

SUBJECT:

EVALUATION OF ROOT CAUSE FINDINGS REPORTS We have completed our reviews of nine Toledo Edison Company root cause findings and ccrrective action reports for plans IA/IB/1C, 10, 8, 9A/9B, 10, 12, 18, 26, and 27. Based on our reviews of these plans, we have concl"ded that Toledo Edison Company likely has identified the fundamental cause(s) of the equipment failure or malfunction being investigated.

The evaluations relating to our reviews of the above mentioned plans are included as attachments 1 through 8. Please note that we have identified additional testing, investigation, surveillance requirements or corrective actions in Attachments 2, 4, 6, and 8. Please review these evaluations and provide us your comitment to perform the additional actions identified. You should provide your response not later than November 4,1985.

Three of remaining reports covering SFRCS, Nuclear Instrumentation, and Pain Steam Header Pressure are still being reviewed by the staff. The results of these evaluations will be provided to you when completed.

Sincerely,

?h, b ."'

John F. Stolz, Chief Operating Peactors Branch #4 Division of Licensing

Enclosure:

As Stated cc w/ enclosure:

See next page OR8#4:0L ORB. :0L

~

0 ORR& gDL ADe Agaz o CMcCracken JSto l 10 85 10pp85 10/ /8

[W 186i!! 8%

3 -. . . - . . .. ~

. . s .

Mr. J. Williams Davis-Besse Nuclear Power Station Ta ndo Edison Company Unit No. I cc:

Donald H. Hauser. Esq. Ohio Department of Health The Cleveland Electric ATTN: Radiological Health Illuminating Company Program Director P. O. Box 5000 P. O. Box 118 Cleveland, Ohio 44101 Columbus, Ohio 43216 Mr. Robert F. Peters Attorney General Manager, Nuclear Licensing Department of Attorney Toledo Edison Company General Edison Plaza 30 East Broad Street 300 Madison Avenue Columbut, Ohio 43215 Toledo, Ohio 43652 Mr. James W. Harris, Director Gerald Charnoff Esq. Division of Powtr Generation Shaw, Pittman, Potts Ohio Department of Industrial Relations ,

and Trowbridge 2323 West 5th Avenue 1800 M Street, N.W. P. O. Box 825 Washington, D.C. 20036 . Columbus, Ohio 4'216 Paul M. Smart, Esq. Mr. Harold Kohn, Staff Scientist Fuller & Henry -

Power Siting Comission

. 300 Madison Avr;nue 361 East Broad Street P. O. Box 2088 Colurr. bus, Ohio 43216 Toledo, Ohio 43603 Mr. Robert B. Borsum President, Board of Babcock & Wilcox Ottawa County Nuclear Power Generation Port Clinton, Ohio 43452 Division ..

Suite 200, 7910 Wood:nont Avenue

. Bethesda, Maryland 20814 Resident inspector U.S. Nuclear Regulatory Comission -

5503 N. State Route 2 Oak Harbor Ohio 43443 Regional Administrator, Region fil U.S. Nuclear Regulatory Comission 799 Roosevelt Road Glen Ellyn, Illinois 60137

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

6

, Attachnent .l' DAVIS-BESSE EVALUATION OF LICENSEE'S REPCRT REGARDING OVERSPEED TRIPS OF THE AUXILIARY FEED PUMP TURBINES TOLEDO EDISON PLANS NO. IA/1B and 1C We have reviewed the findings, corrective actions, and generic implications report entitled, "Overspeed Trips Of The Auxiliary Feed Pump Turbines on June 9,1985 at Toledo Edison's Davis-Besse Nuclear Power Station" concerning the problems associated with an overspeed trip of the auxiliary feed pump turbines (AFPT). The AFPT is a steam driven turbine which drives the auxiliary feedwater pump. Both of the auxiliary feedwater pumps.(AFP), including the turbine and overspeed trip mechanism (OTM), at Davis-Besse are identical except for the model of the governors.

EachAFPTisfedfromitsrespectivesteamgenerator(SG),thatisSG1 feeds AFPT 1 and SG2 feeds AFPT 2. In addition, there is a cross connection such that each SG can feed the redundant AFPT, i.e., SGI can feed AFPT 2 and SG2 can feed AFPT 1. These cross connected lines are normally closed. During the June 9th event, a low SGI level signal opened the steam line to AFPT 1. When the operator (five seconds later) trioped both channels on low SG pressure, the normal steam lines were isolated and the cross connected steam lines were opened. The liegnsee, by analysis, has determined that a large quantity of

, condensate could have been formed when the steam was admitted to the cold cross connect lines. The condensation in the steam lines formed a water slug at the ATPis and could have caused the overspeed of the AFPis. The li:ensee has ,

,aproposed three different scenarios where this water slug could cause the AFPT to trip on overspeed. The first scenario is where the water slug in the governor salve causes the valve to open too far in an atterpt to maintain turbine.1 peed. When the water clears the valve, the valve admits toomuch steam and the AFPT trips on overspeed. The second scenario is where the water flashes as it enters the turbine and thereby acc. 'erates the turbine due to the sudden expansion until it trips. The third scenario is similar to the first scenario except the water slows down the turbine and the gcVernor va'Ive opens i

to try to maintain speed. When the Watsr clears the turbire. the governor

2 l

valve is open too far and the turbine trips on overspeed. In order to support this hypothesis, the licensee also calculated the quantity of condensate which could be formed in the normal lines used to power the AFPTs. A comparison

< of the quantity of condensate formed in the line from SG1 to AFPT 1 is almost as much as from SG2 to AFPT 1. The licensee could not explain why the AFPT 1 3

had never tripped on overspeed whee fed from SG 1. In addition, the licensee has not determined how or why the condensate resulted in the overspeed tripping

%f the AFPTs. The identification of the root cause was done hypothetically and the licensee has not proposed to perform any verification tests.

The licensee has proposed maintaining all steam Ifnes from the SGs to the AFPTs at full pressure and temperature up to the turbine inlet isolation valves, which i are approximately 10 feet from the turbines. These turbine inlet isolation valves are to be replaced with pneumatically operated control valves.. Thus, on an initiation signal, only the new valves will be required to change position'. This valve lineup has been tested by the licensee, as indicated in a meeting on September 25, 1985. Each start of the AFPis resulted in accept-able perfomance, i.e_., no overspeed trips occurred.

Based on our review of the licensea's findings, corrective actions, and

. generic implications and the successful initiations of the AFPis with hot steam lires, we believe that the licensee has identified the root causes of the operators inability to reset the AFPis and that the licensee has proposed .

reasonable corrective actions. Therefore, we believe that the overspeed trip mechanism and associated linkage should be removed from the freeze list so that corrective action may begin.

Dated: October 30,1985 L The following NRR personnel contributed to this evaluation: John Ridgely

Attachment 2 -

l DAVIS-BESSE

(

t i

EVALUATION 0F LICENSEE'S REPORT REGARDING l

AFPT OVERSPEED TRIP THROTTLE VALVE PROBLEM

. TOLED0 EDIS0N PLAN NO. 10

\

We have reviewed the corrective actions and generic tmplementations report i entitled "AFpT Overspeed Trip Throttle Valve Probler,," Revision 1, concerning i

[ theproblemsassociatedwithresettingthetripth/ottle(T&T)valveduring

[

the June 9, 1985, event at Davis-lesse. The T&T valve is a steam admission

} valve to the terry turbine which drives the AFM pump. Both of the AFW pumps, j includingtheturbine.T&Tvalves,andoverspeedtripmechanism(OTM),at '

Davis-lesse are identical.

} . .

l  !

The OTM consists of a spring loaded poppet in the turbine casing. The poppet l

is str'uck by spring loaded weights when the weights are pulled sufficiently j

away from the turbine shaft by centrifugal force. Once the poppet is struck, i j

i it moves away from the turbine shaft and releases the spring loaded trip i i

linkage. The linkage releases the latch on the TAT valve, thereby allowing the spring in the TAT valve to close the valve. Resetting the AFpi overspeed  !

trip involves manually moving the linkage, resetting the OTM, resetting the  :

latch on the T&T valve and re-engaging the valve operator to the valve inter-

) nals. If the linkage is not moved far enough, the OTM will not reset and if l the TAT valve IEtches, the latch will only be due to friction between the parts l of the linkage.  !

i e I

' h The problem, as identified in the licensee's report involves three areast

1) improper procedures, 2) inadequate training and 3) insufficient trip status

• l l indication at the AFW pumps. Based on our review of the licensee's submittal.

l j it appears that the licensee has adequately identified the root causes of the l

l equipment operator's inability to reset the AFW pump after being tripped on

overspeed., In general, the NUREG-1154 report entitled " Loss of Main and

[

j Auxiliary Feedwater Event at the Davis Besse plant on June 9,1985," indicates l that the equipment operators performed their tasks associated with ysetting i I .

the AFW pump trips as well as possible with the infomation and training ,

I available. It is our opinion that if any one of the above areas had not been I j deficient, the equipment operators probably would have been successful in t j resetting the AFW pumps.

t l' I i-

j s -

l -

$ 2 t

The ifcensee has proposed the following corrective actions-i

\

1.

To modify the appropriate procedures to reflect th9 proper re:et sequence i for the OTM; l 1

l 2. To modify the testing procedures to ensure that the T&T valve and OTM are i reset after testing; 4

3. To provide operator training on the theory of operation for the OTM and

, T&T valve; . .

l

. 4. To-provide operator " hands-on" trairing in the proper reset of the OTM and opening of the T&T valve with a minimum steam pressure of 800 psi; 5.

To design and install local position indication of the OTMs and position indication of the T8T valves; i

6.

To paint the yoke of the T&T valve, the latch-up lever, trip yoke and connecting rod (for both AFpT's) yellow to distinguish this equipment as important in the cperation of the overspeed trip. In addition, the manual

, trip level will be painted red; and

, 7. To provide enhanced comunication fcr the equipment operators between both -

, pump rooms and with the operators in the control rocm.

I All of these corrective actions are to be cbpleted prior to restart, except for Item 4 which will be complete before leavir.g Mede 3. The licensee identi-

' fled additional planned actions in order to correct discrepancies noted during the course of their investigations. These actions in:19de additional surveillance tests, preventive maintenance and replacement of some  !

components. Some actions were proposed as one-tfre only actions. These include performing NDE on the poppet leaf spring; vertffcation cf proper i adjustment of the poppet; dimensional inspection of the poppet rJt ?dd head lever, connecting rod to trip hook lever and trip hook to latch sp lever mating

3 surfaces.

These items should be made part of the 18 month surveillance requirements as well as verification that the OTM hold down screws are p tightened and that the trip hook crank is in perpendicular alignment between the :onnecting rod and the trip hook pivot shaft.

Based on our review of the Itcensee's submittal and tr, creased 18 month surveillance requirements, we believe that the licensee has identified the root causes of the operators inability to reset the AFPT* and that the licensee has proposed reasonable corrective actions. Therefore, we believe that the overspeed trip mechanism and associated ifnkage should be rernoved from the freeze' list so that corrective action may begin.

The following NRR personnel contributed to the preparation'ofuation: this eval J. Ridgely S

e a

. . Attachlient 3 ON/IS-BESSE i EVALUATION OF LICENSEE"5 REPORT RELATING TO MATN FEED PUMP TURBINE AND CONTROL FAILURE i TOLECO E0! SON PLAN NO. 8

.. t We hive reviewed the findings report entitled, "i:afn Feed Pump Turbine"and i Contrcl System Failuie" concerning the problems asscciated with an everspeed trip of the main feed pump turbine (MFPT) No I which was the initiating failure of the hne 9,1955, event at ravis-Besse. The NFPT is a steam driven t:Jrbh e which drives the feedwater pump. Both of the main feedwater pumps (NTP), including the turbine, speed control sy: tem and overspeed trip me:hanism (OTM), at ravis-Besse are identical.

The hTPT speed is contro11:0 by an electronic hydraulic control system con- '

sisting of tha fellow!ng :.uhysters: _

1. Sigt.a1 con.erter cirryltry;

2. . peed pickup feedback circuitry; '

3. Speed sumstien and valve lift reference circuitry; and 4

Operator /pilet valve pcsition f. edback and r.ervo ampitfier circuitry.

The signal con'orter circui ryt accepts a speed setpoint signal and produces a referen:e sign'al which corresponds to the dtmanded feedwater flow requirement. Tna speed pickup feedback circuitry provides the signal which corresponds to tee speed of the HFFT. This signa *1 is determined by automatic selecticn of one of two redundtnt signals. Each signal is generated by a rickup which ronitors the passing of a toothed wheel which is mounted on the shaf t of the MFPT, The reference speed signal and the actual MFFT speed are surred and cobared by the speed sumstion and valve lif t reference circuitry. This circuitry produce: a speed error signal and a valve lift reference signal. This reference signal is sumed with the valve position feedback signals'from the pilot valve and the operating cylinder by the operitor/ pilot valve position fesdback and servo arrplifier circuitry which produces a salve position error 9

2 signal. This err 6r signal drives the servo valve to change the pesition of the pilot valve and operating cylinder. Thus the steam admission valve oper* I or closes to develcp a zero error signal and thereby maintain the t'.rbtra speed at its predetermined value.

The problem, as identified in the licensee's report is the result of the failu*e of the frequency to voltage converter in the speed summation circuitry. This . failure, which resulted in a fixed output cf 0.0 vulte, las been attributed by the licensee to a failed open capacitor.

Based en our review of the licensee's findings report, we talieve that the licenses nas identified the root causes of the overspeed tripping of tr.e MFPT. The ifcensee has concluded that thecapaciter failure on the

-afe en e circuit board was a randem failure of an electronic component.

Cer?ective action indicated is to replace the reference board, recalibrate,  !

and return the control system to operational condition.

ir,e following Nr.R personnel contributed to this avaluation: John Ridcoly.

4 l

l l

Attachment 4

' - DAVIS-BESSE EVALUATION OF LICENSEE'S REPORT REGARDING r

TURGidE BYPASS VALVE, SP 13A2, ACTUATOR FAILURE TOLEDO EDISON PLAN NO. 9A/98 The turbine bypass valves are part of the turbine bypass system and are used to control the flow of steam entering the condenser from the bypass header.

< Their purpose is to minimize loss of condensate to the atmosphere by directing steam flow to the condenser. These valves themselves are not important-to-safety or safety-related in terms of fulfilling their function in the plant.

The safety-related or important-to-safety implications of this failure are as follows: , , ,

i ,

' 1. The valve disk and stem were separated prior to the incident for an

' unknown period of time. This indicates that planned maintenance and/or inspection was deficient. A water hanner occurred in the s riping upstream of the valve and coupled with impacting of the loose disk on the valve stem served to crack the valve casing and further damage the valve. The valve was not operational prior to the event and the licensee's maintenance plan did not discover it.

2. The connon drain / isolation valve was closed although it should have been open. This valve and its. associated header serves the turbine bypass valves and is intended to drain condensate from the lines in

, order to help prevent a water hamer event. This indicates that the operating procedure or that implementation of the operating procedure was deficient for this system.

3. Steam traps are p*ovided in the lines f rom the steart, ge,erators for the purpose of draining condensate frca the 'ines in order to minimize a potential water hanrer. The steam traps were blocked with debris and thus improperly raintained. This ir<ficates tact planned caintenance and/or inspection prccedures were deficient for these items.

4. There are missing loose parts in the system. Their potential effects on safety-related or im be assessed in detail. portant-to-safety equiptrer.. or systems should -

5. The other valves of this syste.1 were visually inspected. It wou'd appear that HDE of the castings might be justified. The licensee should assess the nect ssity to inspect the system and cor.:ponants further. The question of whether catastrophic failure of these valves could affect safety related or inportant to-safety systems or

, equipeent should be addressed in detail.

6. The cause of separation of the valve seat from the sten has not been identified and should be esaluated. This would have ivpact on the potential generation of louse parts. Further, such a failure might occh* in safety related equipment of similar design.

The following NRR per:.onnel cuatriluted to this evaluatinn: Owen Rothber;

. . . Attachment 5

) CAVIS-DESSE EVALUATION OF LICENSEE'S REPORT REGARDING PORV MALFUNCTION DURING THE EVENT OF JUNE 9, 1985 TOLEDO E0! SON PLAN NO. 10 During the Davis-Besse loss of feedwater transient of June 9,1985, the pressurizer power operated relief valve (PORV) opened to relieve pressure three(3) times. The third time the P0r,V opened, it did not reseat as it ~

should have when power was automatically removed from the actuating solenoid at the low pressure setpoint. Upon closure of the block valve the pressure hao dropped approximately 300 psi below tnis setpoint. When the block valve was subsequently reopened, the PORY was found closed.

The Davis-Besse PORY is a Crosby style HPV-SN pilot operated valve with a solenoid actuator. The solenoid moves to open the pilot valve when electrically energized and returns to close the pilot valve when electrical power is removed; The pilot valve, when open, provides a vent path to the -

main valve disk which is then opened by the inlet system pressure. The main valve disk should reseat once the pliot valve recloses to seal off this vent path which allows pressure to rebuild on the back side of the main disk.

The licersee conducted an investigation to determine the causes of the PORY failure. The PORV has been removed from the pressurizer, dismantled and inspected. The PORY vendor, Crosby , also participatad in the valve inspection and several abnomalities were found:

a) Three (3) of eight (8) inlet flange nuts were found loose.

b) The adjusting bolt locking nut in the pilot valve Ifnkage was found loose and a cotter pin only was in place to operate the adjusting bolt.

There was minor steam cutting on the pilot seat and utsk.

c) d) A brown substance, speculated by the licensee to be boric acid, was fcund on' the valve body in the vicinity of the pilot valve.

, e) A silver of metal frem the bellows housing flexitallic gasket and small gouge in the outside edge of the gasket surface was fcund.

There was foreign raterial in the pilot sensing tube which caused the pilot .

disk to leak during leak testing perfomed after the transient. The licensee indicates the caterial is a liquid lubricant and would not affect the ability of the valve to open and close.

The licensee has concluded that none of these abnomalities could have caused the failure on June 9, 1985. Several other failure modes have been hypothesized by the licensee, including:

a) differential therra) expansion between the main disk and the valve body due to non-unifom heating upon actuation. (Calculations by the licensee show that clearances are more than adequate to preclude this type of binding action.)

b) other mechanical malfunctions such as loose or misaligned internal c) parts, broken solenoid coil linkage, d)controlsystemmalfunction. ,

s 4-

,w- _ _

4-x,

• The licensee has determined that none of these failure modes is very probabic,

[ and has detemined that a more probable failure mode is that of foreign.

( -s material lodging in the pilot disk and seat. .

s

\ The staff agrees that this could have been a probable cause of failure, especially considering the long period of time since the last PORY actuation during which foreign material could have collected. Prior to the June g event, the licensee had not stroked the PORY since September 1,1982. The staff has detemined that the valve is required to be stroked according to the plant inservice testing (IST) program for pumps and valves at each cold shutdown.

Therefom, the licensee has not met the plant IST requirenents for

- the PORY during the time interval since September 1, 1982. The long period of

' time without actuation of the PORY may have contributed to the degradation of

, the valve operability and the lack of knowledge thereof.

Before the next restart, the licensee proposes to stroke the salve eight (8) times at reduced pressure (nominally 700 psig) and three (3) times at full pressure (nominally 2155 psig) during the plant restart in order to ensure that the valve is operable. Additionally, the licensee has proposed to stroke test the futhe PORY plant at dach shutdown in order to ensure its reliability during operation.

The licensee's recosmitment to strcke test the PORY in accordance with the plant IST program requirements is acceptable to the staff. Further, the staff finds l the liceasee's. proposed Startup test procedure of stroking the PORV three (3) times acceptable for assuring initial operability. With routine periodic

.. testing during cold shutdown, it is more likel closing the PORY may be detected early, i.e., yprior thattoproblems with of a challenge opening or overpressure.

As reqvfred_by'3ection XI of the ASME Code, the PORY must be repaired and retested

, if tre valve fails a test.

The licensee is also investigating whether an c'ternative PORY design would be more appropriate for the Davis-Besse pisnt. -

codification should such a change be deemed necessary.This couldwhich Any PORY design involve a future pl has Itemnot already II.D.1 mustbeen be soqualified by full flow testing as required by NUREG 0737, qualified. In addition, any changes to the plant PORV inlet ano Ites !!.D.1. discharge piping configuration must also be analyzed as required by Although the licensee has not been able to positively identify the cause of the PORY failure, the staff has concluded that the post event evaluation was thorough. i This evaluation did identif deficiencies, degradation mechanisms,and y aInservice number of valve installation Testing deviations which together clearly are evidence of at least a pre-event lax attitude ,on the part of the licensee relative to PORY operability. We have concluded that the testing to be performed by the licensee, both during startup and inservice, complemented by the additional PORV investigative effort yet to be performed should provide increased assurance of PORV operability for the Davis-Besse plant.

' This evaluation was prepared by G. Hamer.

3 ,

i .

A's

. .. : Attachment 6 DAVIS-BESSE EVALUATION AND STAFF POSITION REGARDING OPERATORS FOR VALVES AF-599, AF-608 AND MS-106, PLANS 12 AND 27 AND OTHER MOTOR OPERATED SAFETY RELATED VALVES We have reviewed the information provided by the licensee regarding the failures of the referenced valves to perform their safety related function to open during the June 9,1985 event. Valves AF-599 and AF-608 are required to open on demand against high differential pressure to admit auxiliary feedwater to the steam generators. Yalve MS-106 is required to open against high

- differential pressure to admit steam to the steam driven Auxiliary feed pump turbine.

o As described in Action Plans 12 and 27, the licensee has concluded that all three failures to open resulted from incorrectly adjusted operator torque switches. For corrective action the licensee has described a methodology to be used for readjusting the torque switch settings on these and other safety related MOVs. Significant elements of this methodology include use of substantial portions of a procedure developed by Torrey Pines Technology, GA-CL6881 as revised in December,1982 as well as the use of the MOVATS (Motor Operated Valve Assembly Testing System) testing technique.

Our review of the Torrey Pines procedure indicates that it does contain some useful informat' ion, but leads to the erroneous conclusion that torque switch adjustments arrived at solely following the procedure will assure valve function under all operating conditions. We find the procedure deficient in that it makes no allowance for degradation in valve performance over time.

Use of the MOVATS technique can be useful for confirming that valves are

' adjusted correctly, if the correct valve required stem thrust capabilities are first known.

From reviewing information provided to date by the licensee we have concluded that the licensee does not have an adequate justification that the torque switch settings derived using the Torrey Pines procedure'(only based on new valve design dimensions, operating friction, etc., with no provision for degradation) can assure that the required open or closing forces will, in fact, be attained. Accordingly, in order to confirm the adequacy of the chosen adjustments prior to restart, we require that in-situ confirmatory l

testing be performed.

l

2

)'

The licensee should establish theoretical torque and limit switch settings in accordance with its current analytical methodology. As a minimum, for the valves of the emergency core cooling systems and auxiliary feed systems, torque and limit switch settings are to be confirmed by in-situ testing.

An in-situ testing plan should be prepared by the licensee and submitted for staff review which will effectively demonstrate, using whatever conditions of pressure and flow rate are available in the plant for a particular valve, that the valve operator will function under both normal and transient and accident conditions of pressure and/or fluid velocity for which the applicable valve function is credited in accordance with FSAR analysis. Thermal overload bypass limit switches are to be operating during the above described tests.

The following NRR personnel contributed to this evaluation: Joel page

& 9 4

l 90 O

4 W

-w.

e Attachment 7 DAVIS-BESSE EVALUATION OF LICENSEE'S REPORT REGARDING STARTUP FEEDWATER VALVE, SPe7A, PROBLEM ANALYSIS, CORRECTIVE ACTIONS AND GENERIC IMPLICATIONS TOLED0 EDISON PLAN 18 We have reviewed the Davis-Besse licensee's report regarding the apparent failure of the startup feedwater valve, SP-7A, during the June 9,1985 event at Davis-Besse. The report provides a step-by-step description of what appears to be a very thorough and methodical analysis and test program that was performed to determine the root-cause and possible generic implications of the indicated malfunctions of SP-7A and the SP-7A controls during the June 9, 1985 transient. This program was implemented in accordance with written procedures by licensee personnel supported by a representative of the valve nenufacturer and a . consultant on flow instrumentation. The reports state that the results of the tests and analysis indicate that: (1) the failed SFRCS channel 4 indication for SP-7A was due to a random or normal end-of-service-life indicating light bulb failure, and not to a system anomaly; (2) SP-7A was capable of providing a tight shutoff and responded in accordance with design to the June 9, 1985 transient; (3) the indicated flow through SP-7A was due to out of calibration and ambient temperature effects on the flow transmitter; and (4) there were no significant findings regarding generic implications. The reports do not include the raw event data and test data cited and the detailed design information necessary to enable the staff 1

to independently verify the specific step-by-step results of the analysis and test program. However, based on our review of the methodology employed and on the reported results of the program, we conclude that there is reasonable assurance that the conclusions reached in regard to the root-cause and

~

generic implications of the indicated malfunctions of SP-7A and itt controls are valid, and that the report provides an acceptable basis for the corrective actions taken v:ith respect to SP-7A and its controls.

The following NRR personnel contributed to this evaluation: Faust Rosa ~

4 4

e

• i

. Attachment 8 DAVIS-BESSE EVALUATION OF LICENSEE'S REPORT REGARDING SPURIOUS TRANSFER OF AFW SUCTION TO SERVICE WATER TOLEDO EDISON COMPANY PLAN NO. 26

/

We have reviewed the findings, corrective actions, and generic implications i report entitled, " Service Water Transfer" concerning the spurious transfer i

of the auxiliary feedwater (AFW) pump No.1 suction from the condensate storage

f. tank to the service water system (SWS). The condensate storage tank is the non-safety related primary source of water for the AFW system. When the AFW

! system is needed and either the condensate storage tank is not available or has i been emptied by the AFW system, a safety-related transfer system transfers the suction from the condensate storage tank to the SWS. The SWS is the safety-related secondary source of water. The transfer is initiated upon a low suction pressure signal and is designed to transfer the suction to the alternate source of water without damaging the AFW pumps.

During the June 9,1985 event, the suction for AFW pump No. I transferred to the SWS while there was ample water in the condensate storage tank. AFW pump No. 2 did not experience any transfer. The licensee indicated that the pres-sure drop across the suction strainers in conjunction with the piping losses and load changer on pump No. I resulted in the low suction pressure. While the pressure drop across the strainers in the suction line of pump No. 2 and

! the effects of load changes would be similar to that experienced by pump No.1 the piping losses would be less for pump No. 2 and thereby would not result -

in the transfer to the SWS. The licensee's proposed solution is to remove

the strainers imediately ahead of each pump and to increase the mesh size of

[, the strainer in the comon suction line from the condensate storage tanks.

{ In addition, the licensee has proposed decreasing the low suction pressure set-point and incorporating a time delay to reduce spurious transfers to the SWS.

L r In our SER input dated August 29, 1983, concerning the TMI Task Action Plan I

(TAP) Item II.E.1.1, we stated that the licensee met Recomendation ,GS-4 by having an automatic transfer of the AFW suction to the alternate source of

! water and by having an automatic isolation of the AFW turbine steam inlet lines I

e

__ _.m.. _ _ _ _ _ _ _ _ . . _ . _ _ . _ . . . _ ..

s .

2 at a' suction pressure of 1 psig. These two features provide protection of the pumps from cavitation. In response to the additional short-term reconnendation No.1 the licensee identified that the low level alarm setpoint on the conden-sate storage tank corresponds to approximately 200,000 gallons of water in the tank which is more than I hour's worth of water.

Because of the automatic transfer of the AFW pump No. I suction on June 9th, it is not clear whether the required technically specified volume in the conden-sate storage tanks could actually be pumped by the AFWS. With the reduced transfer setpoint, it is not clear that the pumps will be adequately protected i

from cavitation. The' licensee has not proposed any testing to verify 1) the ability to pump the contents of the condensate storage tank into the steam generators or 2) the ability of the new transfer setpoint to provide adequate protection against cavitation for the AFW pumps. Therefore, the licensee should perfom the aforementioned verification tests or propose modifications t to the technical specifications which identify that a train of the AFWS is

, INOPERABLE whenever the related transfer system, transfer valves, or SWS is inoperable, in addition to the current requirements.

, In the TAP Item II.E.1.1 review the staff considered the need to lock open single or multiple valves in series which could interrupt all flow from the water source (s) to the pumps and from the pumps to the steam generators.

Additionally, where we realized that strainers were present we recomended their removal. Strainers are usually installed during construction and -

utilized during system pre-operational testing where there is the possibility i of items entering the suction of the pumps. After pre-ops the strainers are normally removed. We reconnend its removal thus eliminating a possible. source i of a comon mode failure.

! Based on our review of the licensee's findings, corrective actions, and i

generic implications report, we believe that the licensee has identified the root causes of the spurious transfer of the AFW pump No. I suction to the SWS. We 'do not believe that the licensee has proposed adequate corrective I action and testing. -

The following NRR personnel contributed to this evaluation: John Ridgely i