ML20097A359

Forwards Response to NRC 910628 Request for Addl Info Re Turbine Valve Surveillance Testing.Based on Westinghouse Rept,Total Turbine Missile Generation Probability Meets Appropriate Turbine Sys Reliability Criteria

ML20097A359
Person / Time
Site:	Beaver Valley
Issue date:	05/28/1992
From:	Sieber J DUQUESNE LIGHT CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
TAC-M77640, NUDOCS 9206020297
Download: ML20097A359 (11)

Text

A t

Beaver Valiey Power Station

, SNppmgport, PA 15077-0004 (412) 393-525$

JOHN D StEBEft vse pres oent - wuciear oroup Flay 28,1992 U. S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, DC 20555 Subject. Beever Valley Power Station, Unit No. 2 Docket No.30-412, License No. NPF-73 Turbine Valve Surveillance Testing (TAC M77640)

Ref: 1) Letter from A. W. De Agazio (Nuclear Regulatory Commission) to J.D. Sieber (Duquesne Light Company), dated June 28, 1991. Subject -

Request for Additional Information Ref: 2) Letter from J. D. Sieber (Duquesne Light Company) to the NRC, dated October 1, 1990. Subject -

P oposed Operating License Change Request No. 16 This letter provides a response to the NRC Request for Additional Information (RAI) forwarded by Reference 1. Baaed on the attached Westinghouse report it has been determined that, with an eighteen (18) month reheat stop and intercept valve test interval, the total turbine missile generation probability for Beaver Valley Power Station Unit 2 meets the appropriate turbine system reliability criteria.

Our proposed technical specification revision, submitted by Reference 2, will be revised to propose a reheat stop and intercept valve -test interval of 18 months. The revieed submittal will be forwarded in the near future.

In response to the concerns raised by the Salem turbine overspeed event, as described in Information Notice 91-83, a surveillance program is being developed to functionally verify the operability of each turbine trip solenoid valve on a refueling frequency. This program along with turbine surveillance tests and inspections -will ensure the turbine overspeed protection system is maintained and operated consistent with the assumptions used in the attached evaluation.

..n>  ;

qVDA '

O k i s 7D 9206020297 920528 \

PDR -ADOCK 05000412 P PDR ,

- - - - . - _ - - _ _ _ _ . _ . . l

A

' Turbine Valve Surveillance Testing (TAC M77640)

Page 2 9

Should you have any questions regarding this response, please contact Mr. Ken McMullen at (412) 393-5214.

Sincerely, N D. Sieber my cc: Mr. L. W. Rossbach, Sr. Resident Inspector Mr. T. T. Martin, NRC Region I Administrator Mr. A. W. De Agazio, Project Manager Mr. M. L. Bowling (VEPCO) 1 v - - _ _ _ - _ - . - _ - - _

6 REQUESTfFOR ADDITIONAL-INFORMATION

. TURBINE REHEAT STOP AND: INTERCEPT-VALVE-TESTING ^

Reauest

1. Please submit information relating to_the probabilities 1for missile ejection over the entire spectrum of overspeed conditions.

Response

Based on the Westinghouse evaluation (documented in the attached report),- it has been determined that with an eighteen (18) month reheat stop- and intercept ~ valve _ test interval, the total _ turbine missile generation probability for Beaver Valley Power Station Unit-2 meets applicable _ acceptance criteria.

Three _overspeed events- were'- considered in . evaluating- the probability -for missile- ejection: design overspeedL(120 percent of-rated speed),  : intermediate;overspeed-(132 percent of rated speed),

and destructive overspeed -(speeds greater than 170 percent of rated-speed). The Westinghouse- evaluation focused

• cn1 the design and intermediate overspeed events < since theyz would-beLaffected by-the test' -intervals of the reheat stop valves and intercept valves. :The destructive overspeed _ event does not result fror failures of reheat stop and intercept valves and therefore was excluded from further consideration.

Table 2- of the Westinghouse report _ presents the_ total probability of -turbine missile ejection for the design and intermediate _overspeed1 events- given. that system separation occurs. The total probability is based on conditional probabi.lities of missile-ejection for Unit--1 given that design' or intermediate overspeed. occurs.-

-. Unit- .1 conditional probabilities Eare more conservative than values calculated-for. Unit 2.

The total- probabilities for a. missile ejection must_be-multiplied Eby_.the average annual frequency of system' separation for the: Unit so that they can be measured >against' acceptance: criteria. ' Based on a reviewu of Unit 2 plant. trips, -the average annual frequency 1of-system separation was: -calculated--to be 0.22 -(one occurrence in'four.and -

one-half -years). -To -provide additional 7 conservatism, the average annual frequency Lof system : separation was assumed to be one-half (0.5).

The -Westinghouse report Ldid -not consider _desbructive.overspeed progability. Therefore,, the " general"~ acceptance: criteria-'of 1 x-110~ per year for- turbine missile Lejection-- from1 an unfavorably _

oriented turbine was reduced. A ten (10)Lpercent. fraction of the'

" general"- acceptance criteria.was assumed asithe acceptance criteria for (the ' design- -and intermediate overspeed- missileE probabilities

-ev~luated. -Thisa11 eaves an adequate reserve-margin of-90 percent of

_the acceptance criteria for other significar.t overspeed events such as destructive _overspeed.

t

' Request For Additional Information Turbine Rehent Stop'and Intercept Valve Testing Page 2 The product of the average annual-frequency of turbine separation for Unit 2 (conservatively assumed to be 0.5) and e total '

probability of turbine missile ejection (6.79 x 10 y) is- less ,

than the acceptance criteria of 1 x 10 . mus, the turbine system reliability is acceptable with an eighteen (18) month test interval for reheat stop aM intercept valves.

Beauest

< 2. Provide a discussion of the availrble data relating .to testing-the subject valves that could justify DICs conclusion regarding the reliability of the overspeed protection system.

Response

During the operating life of Beaver Valley Power Station Unit 1 and Unit 2 there has not been a single incidence of unplanned turbine overspeed nor a single turbine reheat- stop cr intercept valve malfunction that could have lead to-a turbine overspeed condition.

We Unit 1 reheat stop and intercept valves were tested during Unit start-ups and monthly thereafter until 1987. 'Ihe monthly testirg at Unit 1 was discontinued in 1987 when it was found to be 4 causing moisture separator reheater damage. Since 1980 the Unit i reheat stop aM intercept valves have undergone 31 operational surveillance tests (OST 3.26.1) .

We Unit 2 reheat stop and intercept valves have been tested during Unit start-ups aM monthly thereafter. Since the inception of operation in November of 1987, the Unit 2 reheat stop aM intercept valves have been tested at least 43 times (per OST 2.26.1) and continue to be tested on a monthly frequency.

Problems were encountered during performance . of 12 of the 31 surveillance tests at Unit 1. In these instances a limit or permissive switch associaced with the test circuitry required adjustment or repair. 'Ihese adjustments and repairs only .affected the test circuitry and did not represent a pc sibility for turbine overspeed since subsequent tests or plant trips showed that the valves operated as designed. 'Ihe other nineteen (19) of the 31 surveillance tests were conpleted satisfactorily with no problers.

In addition, the valves operated properly on 10 plant trips (closing on the turbine trip aM opening on the subsequent plant start-up) .

Each of the 43 operational surveillance tests performed at Unit 2 were completed satisfactorily. %at is, full reheat stop and intercept valve strokes were verified.

Request For Additiortal Information Turbine Reheat Stop and Intercept Valve Testing Page 3 e Minor problems were encountered during performance of 16 of the 43 surveillance tests at Unit 2. These problems typically involved a test push-button or valve position indication. In each case the problems would not have preventcd the reheat stop and intercept valves from closing in the event of a turbine trip.

The Beaver Valley data regardirg reheat stop ard intercept valve testire dia' m M above along with the attached probability analysis (based on the Westirghouse program for tracking turbine valve failure rates) provides evidence that the reheat stop and intercept valves are reliable.

Recuest;

3. Timely detection ard correction of any problem (e.g.,

nechanical problem) that my arise in the main steam turbine system, includity the subject valves, is vital because of the potential for serious adverse consequences. Please address the plant-specific aspects of this concern as it relates to the propcsed deletion of surveillance tests for the subject valves.

Response

A proposed revision to the technical specifications will be sutrtitted to charge the requested reheat stop and intercept valve test interval to an 18 mmth frequency. The 18 raonth test interval is adequate to identiP/ and correct rechanical problems based on the favorable probability analysis described in response to Item 1 above, and the -favorable valve test experience described in response to Item 2 above. In addition, the turbine is inspected periodically as described in Updated Final Safety Analysis Report Section 10.2.3.5.

EeElest

4. Frequent testing of the subject valves at Unit 1 is rot required by the technical specifications, but DLC does perform testing prior to start-up. However, DIf asserts. that even these relatively infrequent tests have caused extensive damage. In view of this experience, please explain how DIE expects to avoid such damage to Unit 2 since DIC has committed to verify proper operation of the subject valves during each plant _ start up as is done at Unit 1.

J

Roquest For kklitional Inforriation

'Ibrbine Reheat stop and Intercept Valve 'Ibsting Page 4

Response

'Ibe damage _ described at Unit i resulted from testing conducted during power operation. During testing of an intercept /stop valve, closure of the _ valve results in the shut-off of cycle steam flw to one (1) noisture separator reheater (MSR) and diversion of flw mostly to the opposite side MSR.

When this testing in conducted at power, significantly -higher shell side (cycle steam) velocity (flw) rates and corresponding pressure drops occur. 'Ihese pressure drops can cause high stress levels in the shell- side closure plates. At the ssm time, the heat transfer across the three (3) active tube bundles is increased due to higher shell side flw arri this causes additional reheat steam to be cooled in the tubes resulting in larger tube side pressure drops. 'Ibe cumulative effects from a series of intercept /stop valve tests performed at power are considered the primary cause of the shell side closure plate damage.

Tests conducted during start-up demonstrate that valve stem sticking is not occurring, ard the .uts are performed at lower flw rates which result in lov * ;ressure drops and stress levels that are not as likely to cause MSh werage.

CALCULATION OF TURBINE MISSILE EJECTION PROBABILITY RESULTING FROM EXTENDING TIIE TEST INTERVALS OF INTERCEPTOR AND RElIEAT STOP VALVES AT BEAVER VALLEY UNITS 1 AND 2 prepared: April 11,1992 A calculation of the effects of extending the test intervals of the turbine interceptor and reheat stop valves at Beaver Valley Units 1 and 2 was perfonned using fault tree models and methodology from the Westirr, Souse report WCAP-ll525, "Probabilistic Evaluation of Reduction in Turbine Valve Test Frequency," dated June 1987. 'Ibe calculation considered the probabilities of turbine missile ejection due to overspeed. The evaluation focused on two of the three overspeed events defined in WCAP-ll525 that would be affected by the test intervals of the reheat stop valves (RSVs) and interceptor valves (IVs); design and intermediate overspeed. The third overspeed event in WCAP-Il525, destructive overspeed, de s not result from failures of RSVs and IVs and was not included in the calculation.

Design overspeed is considered to be an overspeed of 120 percent of rated speed and results from the following failure sequence:

1. System separation (total loss of load) occurs. (This event is not moceled or accounted for in the analysis. Its frequency should be determined by Duquesne Light Co. before applying acceptance cdteria.)

2. One or more govemor (control) valves or two or more IVs stick open, or the initial protective action of dumping the govemor and interceptor valve emergency trip fluid header fails upon loss of load or upon reaching the setpoint of the overspeed protection control system.

3. Stop and reheat stop valves close successfully after reaching the overspeed trip setpoint.

IntermMiate overspeed is considered to be an overspeed of 132 percent of rated speed and results from the following failure sequence:

1. System separation occurs.

2. One or more combinations of RSVs and IVs (in the same steam path) stick open or fail to close.

3. Stop valves close successfully after reaching the overspeed trip setpoint.

Because the stop valves or governor valves must close to limit the overspeed and prevent destructive overspeed, the mechanical overspeed trip function a d dump of the autostop oil 1

system is assumed to be successful. Herefore, failures of the components of mechanical overspeed trip, including the interface valve and 20/AST solenoid valve which dump the emergency trip fluid, ws.w not modeled in the fault trees. Referring to item 2 of the design overspeed event, failure of the overspeed protection control (OPC) system and related solenoid valves, and failure of the 20/AST solenoid valve which dumps the autcstop on upon loss of load were considered in the analysis of design overspeed.

Three test intervals of the RSVs and IVs were considered; 3 months,12 months, and 18 months. Test intervals of the OPC solenoid valves and 20/AST solenoid valve were assumed to be 18 months in order to provide a bounding estimate of the failure probabuity of these valves. The failure rates of key components of the analysis are given in Table 1.

Turbine overspeed probability results are given in Table 1. Design overspeed is minimally affected by the extension of the IV and RSV test intervals because the overspeed is dominated by the failure of the control valves. His dominance is shown in Table 3.

Intermediate overspeed probabuity increases significantly with the extension of the IV and RSV test intetval. However, the products of intermediate overspeed probability and conditional probability of missile ejection, using the data from Table 2, are all small numbers, ne probabilities of turbine missile ejection are calculated in Table 2. Rese probabuities can be multiplied by the average annual frequency of system separation for the Beaver Valley Units so that they can be measured against acceptance criteria. He frequency of system separation typically ranges from 0.1 to 1.0 per year for plants that are equipped with reverse power relay systems. Rese systems delay the trip of the generator following a trip of the turbine for approximately 30 seconds or until there is adequate assurance that overspeed will not occur following a turbine trip which occurs prior to generator trip.

He acceptance criteria in WCAP-11525 was applied to a " total" probabuity of turbine missile ejection which summed the turbine missile p:obabilities for all known everspeed events. The analysis for Beaver Valley did not consider destructive overspeed probability or other possible causes of over md such as reverse steam flow through failed extraction non-return valves.

Therefore the " general" acceptance criteria of 1.0E-05 per year for turbine missile ejection from an unfavorably oriented turbine would not be applicable. Hewever, it would be reasonable to allocate a fraction of the " general" acceptance criteria (5 to 10 percent, nr example, but no more than 25 percent) for the design and intermediate overspeed missile probabilities evaluated herein. His would leave an adequate reserve msrgin of 75 to 95 percent of the acceptance criteria for other significant overspeed events such as destructiv-overspeed.

2 m .

. , ,a TABLE 1 Turbine Overspeed Probabilities as Function of Interceptor Valve (IV) and Reheat Stop Valve (RSV) Test Interval IV & RSV P(A) P(B)

Test Interval 3MO 1.12E-02 1.24E4 12 MO 1.14E-02 9.81E4 18 MO 1.15E-02 1.96E-05 Notesi P(A) Probability of design overspeed (120% overspeed) given that system separation occurs.

P(B) Probability of intermediate overspeed (132% overspeed) given that system sep~ation occurs.

m KEY DATA Component Failure Mode Failure Rate Basis (1/hr) (See Notes Below)

Interceptor valve sticks open 2.915-07 A Reheat stop valve sticks open 2.91E-07 A Control valve sticks open 7.27E-06 B Emergency trip fluid line (common to 6.97E-08 A govertw and interceptor valve) is clogged 20/OPC solenoid valve failure 1.67E-06 C Notes:

- A Original failure data from WCAP 11525 Rev.0.

- B Curree' data base for BB 296 nuclear steam chest valves. %e current failure data consists of 24 incidents of stickhg in 754 valva operating years. This data is crrrently under further revi:w ard documentrtion in conjunction with the Westinghouse program for tracking mrbine valve failure rates.

- C Utilizes criginal WCAP 11525 valve operating years (4620 years) but assumes 50 incidents for the calculation of the failure rate for failure to close.

3

TABLE 2 Probabilities of Turbine Missile Generation Affected by 'nterceptor Valve (IV) and Reheat Stop Valve (RSV) Failure IV & RSV P(A) P(hCA) P(B) P(M/B) P(T) =

P(A)*P(M/A)+P(B)*P(M/B) 4 Test 120 % 132 %

Interval Overspeed Ovenpeed 3.60E-05 1.24E-06 1.35E42 4.20E-07 3MO 1.12E-02 ,

12 MO 1.14E-02 3.60E-05 9.81E-06 1.35E-02 5.43E-07 18 MO 1.15E-02 3.60E 1.96E-05 1.35E42 6.79E-07 Notes:

P(T) Probability of turbine missile ejection given that system separation occurs. These probnbilities are multiplied by the aserage annual frequency of system separation to obtain annual probabilities of turbinc missile ejection.

P(A) Probability of design overspeed.

P(B) Probability of intermediate overspeed.

P(M/A) Conditional probability of missile ejection given that design overspeed occurs.

. P(M/B, Conditional probability of missile eje: tion given that intermediate overspeed occurs.

i KEY DATA Conditloaal Probabilities of Missile Ejection ,

LP. Rotor Beaver Valley 1 Beaver Valley 2 l Operating Time l (years / hours) P(M/A) P(M/B) P(M/A) P(M/B)

)

3/26280 1.13E-06 1.72E-03 4.23E 08 1.90E-04 l

3.60E-05

• 1.35E 02

• 2,69EM 2.47E.03

- 5/43800

• 7/61320 - 2.52E45 3.85E-02 2.96E45 ~ 9.71E-03

(*) 'Ihe calenlation of P(T) above utilizes the Unit 1 conditional .

probability data. The operating time of the L.P. rotors is assumed to not exceed 5 years or 43,800 houn between inspections. 'Ihe 5 year values have been used in the calculation above.

4

+

-~9 .,m*....,.-m-. n _ w y---f . , . . . ~ . - - ...,..-..,..%~..., ,v ,.v..w .w,.m7 -y, , . ,,,,yw w,. e.,,v,,v.-w-1,.e.-e-am-,-- y m, -

\. ;

}

TABLE 3 Importance Factors of Dominant Faults Contributing to Daign and Intermediate Overspeed probabilities IV & RSV 1mportance Freton

' " ' latermediate Ovenpeed Design Overspeed 3 MO A(94.68) B(85.36)

D( 4.09) D(55.34)

E(0.65) C(38.40) l 12 M O A(93.28) B(94.56) '

D( 4.03) I'QO.68)

E( 2.04) .>(25.26) 18 M O A(92.32) B(95.68)

D( 3.99) C(77.84)

E( 2.94) D(18.ti6)

Notes: ,

- Importance factors are detennined by summing the importances of all cusets in which the fault appears.

- The letters A through O cornspond to the following basic faults:

A Control valve (CV) sticks open.

B Reheat stop valve (RSV) sticks open.

C Interceptor valve (TV) sticks open.

D Emergency trip Guld line (common to govemor and interceptor valve) is clogged.

l- -E Two interceptor valves stick open due ta common cause.

F Dump valve /RSV - sticks closed. ,

O Dump valve /IV - sticks closed.

i 5

L