ML20107D992

Provides Summary of B&W Rept, Review of On-Line Test Intervals for Reactor Trip Sys, Per Generic Ltr 83-28,Item 4.5.3 Re Adequacy of Testing Intervals for Reactor Trip Sys. On-line Test Interval Consistent W/High Availability

ML20107D992
Person / Time
Site:	Davis Besse
Issue date:	02/15/1984
From:	Crouse R TOLEDO EDISON CO.
To:	Stolz J Office of Nuclear Reactor Regulation
References
1126, GL-83-28, NUDOCS 8502250276
Download: ML20107D992 (7)

Text

3 c

i Docket No. 50-346 TOLEDO License No. NPF-3 AcxAno P. CAOUSE

. Serial No. 1126 vo p,,,,

ww.c February 15, 1984 Director of Nuclear Reactor Regulation Attn.:

Mr. John F. Stolz Operating Reactor Branch No. 4 Division of Licensing

' United States Nuclear Regulatory Commission Washington, D. C.

20555

Dear Mr. Stolz:

By lettier dated November 7,1983 (Serial No.1000), Toledo Edison Company submitted the Davis-Besse Nuclear Power Station, Unit No. I response to Generic Letter 83-28 (Log No. 1322), entitled " Required Actions Based on Generic Implications of Salem ATWS Events". Within that submittal, Toledo Edison provided detailed responses to each specific item identified in the Generic Letter 83-28.

Item 4.5.3 of the Generic Letter concerns the adequacy of testing intervals for the Reactor Trip System (RTS). Attachment 1 to this letter provides a summary of the Babcock & Wilcox (B&W) report, " Review of On-Line Test Inter-vals for the Reactor Trip System".

Through participation with the B&W Owners Group (BWOG), Toledo Edison has demonstrated that the current on-line test interval for the RTS.is consistent with high RTS availability.

Very truly yours, ff w

RPC:JSH:nif encl.

-cc: _DB-1 NRC Resident Inspector 8502250276 B40215

[

PDR ADOCK 05000346 P

PDR b

i O

\\

I THE TOLEDO EDISON COMPANY EDISON PLAZA' 300 MADISON AVENUE TOLEDO, OHIO 43652

n.

Dockst No. 50-346

-License No. NPF-3.

Serial No. 1126' February 15, 1985

' Attachment 1

'Page 1; l

Summary

'The following reports the-findings of an investigation to respond to Question 4.5.3 of Enclosure A of Nuclear Regulatory Commission (NRC) letter to all Licensees,f

SUBJECT:

Required Actions Based on Generic Implications of Salem ATWS. Events.(Generic Letter 83-28). Question

- 4.5.3 ' requests verification of the existing one-month Reactor Trip System (RTS) test interval and reads:

"4.5 REACTOR TRIP SYSTEM RELIABILITY (SYSTEM FUNCTIONAL TESTING)

Position

[

On-line functional testing of the reactor trip system, including

^

independent. testing of the diverse trip features, shall be per-formed on all plants.

1.

2.

3.

Existing. intervals for on-line' functional testing required by-Technicalc Specifications shall be reviewed to determine that-the. intervals are. consistent with achieving high reactor trip-

. system' availability when accounting for considerations such as:

n

~1.

Uncertainties in component failure rates-2.

Uncertainty in common mode failure rates 3.'

Reduced redundancy during testing.

14. !

Operator errors during testing.

5.-. Component " wear-out" caused by'the' testing.

Question 4.5 also addresses changes tot the test ' interval:

Changes 'to - existing required intervals: for' on-line testing as well as'the intervals to_be determined'by_

. licensees currently"not performing on-line testing shall1be justified by;information on. the sensitivity of reactor trip system availability.to parameters such as the test. intervals, component failure rates, and common mode failure rates."

J

,At this time,'the purpose of1the investigationLis to' explicitly respond to thu request.of 4.5.3'and provide estimates of the reliability for the "1

existing one-month test interval..

.: /*

f 4

4

h;e

~

Dockst No. 50-346 J*'

License No. NPF Serial No. 1126 February 15, 1985

-Attachment 1

'Page 2 Configuration Features of Importance The' investigation performed is generic to all plants with B&W NSS equip-ment, including 177 and 205 fuel assembly plants. Only two significant design configurations exist for these plants that must be accounted for by the reliability. evaluation of the RTS at the one-month test interval (or for longer intervals).

Consequently, two separate models have been constructed. For this project, all 177 fuel assembly plants, except Davis-Besse, are represented by one configuration'(Oconee group); Davis-Besse and Bellefonte, a 205 Fuel As-sembly plant, are the other configuration (Davis-Besse group). The funda-mental difference between the two-configurations is:

10conee. group:

Safety rodh (groups 1-4) are tripped by the Control Rod Drive Control System (CRDCS) using the D.C. Shunt and A.C. undervoltage trip de-vices of each A.C. or D.C. breaker. Regulating rods (groups 5-7) are tripped by the electronic trip (SCR trip) portion of the CRDCS-and CRDCS A.C. trip breakers using D.C. shunt and A.C.

~

undervoltage trip devices.

Davis-Besse group:

All rods are tripped by either the CRDCS using A.C. breakers or the electronic trip (SCR' trip).

Each breaker is tripped by D.C. - shunt and ' A.C.

s o

undervoltage' trip devices..

Both configurations. offer similar diversity since the mechanical breakers are backed by the electronic trip-(SCR trip). Thefelectronic trip reduces d

the possibility of common mode failures.

l Trip actuation features of~these two groups of plants are very similar.

t

.The Reactor Protection: System (RPS) and the sensor' inputs are virtually-

- th'e 'same.

The! evaluation also included the Anticipatory:Re' actor Trip LSystem' (ARTS) ~ that senses ' loss of feedwater and signals reactor 'and tur-lbine trip..Finor' differences"of the ARTS arrangement 1 exist between the

!.two-groups of plants.-

In addition to the diversity that ha existed in:the.B&W RTS due:to the

electronic. trip, ' two important equipment upgrades were -implemented to re-

'du'ce the. potential for common mode failure of the breakers'. Since'these-changes will-be.in effect for the remainder of plant life, they.were in-cluded.for. investigation. The changes'are:

zl.n Change of trip shaf t bearing lubricant to Mobil 28 from Lubriko, Lthe original trip shaf t bearing lubricant.. Accelerated aging

' tests by General-Electric indicate that Mogil'28 has an expected

. lifetime in excess of'approximately.I x'10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> (approximately

y Dockst No. 50-346

?

- License No. NPF-3

-Serial No.'1126 iFebruary-15,21985 m

AttachmentLi~

.Page 3 5

100 years), as compared to the Lubriko lifetime of 1 x 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> (approximately 10-11 years), at an expected-operating temperature of 40*C.

Thus, the past aging problems associated with lubricant stiffening are expected to be diminished. considerably for the re-

.mainder of plant life.

In particular, the common mode failure 7

contribution to RTS unavailability that has largely been attri-

.buted to breaker shaf t bearing stiffness resulting in slow re-s l

sponse when the undervoltage device is actuated is expected to

.be corrected by the lubricant change. This change also includes installation of a new latch roller assembly and new trip shaft bearings.

2.

Addition of an RPS trip signal to the D.C. Shunt trip device.

This device operates on a different principle than the A.C.

undervoltage trip device. Whereas the A.C. device is released by removal of power to the coil and uses the power stored in its spring to' rotate the trip shaft, the D.C. shunt device uses the power obtained by energizing a coil to rotate the trip shaf t to the " breaker open" position. The undervoltage device can apply about 24 in, oz. of torque,-but the. shunt device can apply _ap-

, proximately 200 in. oz. of torque and is able to overcome the

~

resistance ~of " frozen" lubricant in the trip shaft bearings (estimated maximum resistance is approximately 110 in, oz.).

.Thus, 'the D.C. - shunt trip provides a diverse mechanism f rom the

- A.C.; undervoltage1 device and would be expected to cause shaft

. rotation even with ' frozen bearings.

Methods and Approach The modeling methods use Reliability Block Diagrams-(RBD's) for the.RTS

.and the CRDCS. breaker subsystems involved in a reactor' trip. The PACRAT-code was used to calculate the time dependent unavailability of} equipment for~the one-month on-line itest intervals. The PACRAT code was developed

. tor B&W. ' It is similar to FRANTIC'11~and.has been'used.for other evalua-tions of.the RPS'previously submitted-to-the NRC-(see Topical. Report' i

BAW-10085P, ;" Reactor Protection System", Vol. 2, Rev. 6, : April, 1979).

The RBD models constructed for the Oconee and Davis-Besse groups' include

' sensors'and process equipment providing signals to the. Reactor _ Protection.

System (RPS),'the RPS_ trip module outputs, theLsensors and. processing tequipment~ associated with the Anticipatory Reactor Trip System (ARTS), the

%c ARTS outputs, the CRDCS breakers, and'the CRDCS. Electronic Trip.(SCR trip).

The analysis addressed automatic RTS operation and omitted credit.forfthe additional ' advantages of operator action for, manual' trips and operator :

action to " power-drive" the rods in'using the CRDCS in manual.

1

.3 Dockat No. 50-346 2*'

License No. NPF-3 Serial No.~1126 February-15, 1985 Attachment :1 Page 4 All of the five issues raised by Question 4.5.3 were addressed by the evaluation.. Random and common mode equipment failure rates were ac-counted for and. operating experience was used to support the evaluation

~'wherever possible.

Data from Licensee Event Reports (LER's) was used for sensors and instru-ment strings and B&W' operating experience data was relied upon to provide

. random and common mode breaker failure rates and was updated to show the effects of the lubricant changes.

s Historic operating experience for B&W and Combustion Engineering (CE) plants was evaluated to obtain failure modes and frequencies for the in-dividual breaker components. CE operating experience was used to provide failure rates for the shunt trip device and both B&W and CE experience provided data to support the undervoltage device failure rates.

Since the lubricant change will improve the reliability of undervoltage device ' actuation, the expected future performance was accounted for by crediting the breaker data base operating experience _for those past events

-for which failure would not have occurred had the new lubricant been in-stalled.

Generic data sources _were used to provide failure rates for quantification for electrical components'of the RTS other'than the breakers, sensors,>and instrument strings.

Wearout caused by test cycling and aging was evaluated for all components

. and emphasis.was placed on the breakers since 1they are most affected by.

~

~

utesting. The breakers are designed for 12,500 cycles and the lubricant change will virtually eliminate aging concerns.

It was concluded that wearout is a relatively unimportant concern, however,Jpossible effects

< were simulated by a sensitivity analysis using the IUSD's.

The effects of testing on RTS availability included the influence of oper-ator errors during testing and considered test and maintenance' errors that E could _ contribute to breaker failure to trip. The effect of reduced. system

~

-redundancy due to channel bypass _ during testing was also evaluated. The tests'of importance'are the monthly single channel RPS and ARISfinstrument-string, the trip module and breaker tests.

. Time dependent point. estimate values for the RTS unavailability were_ deter-

't

. mined using best ' estimate data to establish;a base line and the sensitivity -

analysis was performed to indicate the influence of-uncertainties for the '

five; considerations of Question-4.5.3. The sensitivity analysis'was based on a.95% upper bound distribution and provided insights for the effects of uncertainties..

M

~

Dsckst No. 50-346.

LLicense-No. NPF-3 Serial No. 1126-

+

February 15, 1985 l-

.AttachmentL1 Page 5.

t Summary of Results and Conclusions 1..

~The results of the best estimate and sensitivity analyses are:

l

- Average System Unavailability Using Best-Estimate Data Davis-Besse Class' Oconee Class

' Base Case

-Best-estimate time-averaged 6 x 10~ / demand 6 x 10~ / demand system unavailability Average System Unavailability Using 95% Upper Bound Data Davis-Besse Class-Oconee Class Sensitivity to

-6 f'

Uncertainties in randon com-7 x 10~ / demand

.2 x 10 / demand l

ponent failure rates (slightly sensi-(moderately sen-tive)-

sitive) 6

-6

~ Uncertainties in common mode 9 x 10 / demand 6 x-10 / demand

-failure rates / operator' errors (highly sensitive)

(highly'sensi--

~

tive)

Reduced redundancy during test 7 x'10~ / demand 6 x 10~ / demand 7(channel bypass)

(slightly.sensi-(slightly sensi-tive) tive)

' Breaker wearout caused by testing _

(not sensitive)

(not sensitive) 2.

The RTS configuration of both the Oconee and Davis-Besse groups have several features that 'contrib'ute to the high reliability,'auch as:

a.

The Electronic Trip (SCR trip) provides a diverse method of trip actuation that is separate from the CRDCS' mechanical breakers.

Thus,~the. potential for. failure to trip due-to common mode.

l

' failure of the bre,akers is ~significantly _ reduced.

-4,

-m.

_-,,,._,,,-,,7.,.+

---, +,,

r w4-3-v

- - - - + - - - -

J.

t: '

[Dsckst.No.50-346 ib -

-License No. NPF-3

~

l Serial No. 1126 L'-

February. 15, 1985 L

-Attachment 1.

)"

Page 6-l:^

l-b.

.The common. mode failure potential of the breakers is consider-ably reduced by the addition of the shunt trip device which provides diversity from the undervoltage device. The reli-ability of trip actuation by the undervoltage device is im-proved by the lubricant change from Lubriko to Mobil 28.

I I

c.

The:RPS and ARTS'are configured with=four channels.

?.

3.

The wearout evaluation indicated that the'RTS components are not susceptible to wearout_ caused by testing. The breakers are the l

major components affected by' test cycling and the GE AK-2 breaker

[

Lhas a design cycle objective of 12,500 cycles. Aging of the trip shaf t bearing lubricant is virtually eliminated as a concern when Ethe Mobil 28 lubricant is installed. Therefore, for the breakers, common modelfailure due to wearout is not a significant source of

[,

-~ indicate that wearout is a concern.

RTS' unavailability. Other components do not exhibit histories that f'

.- 4.

Reduced redundancy caused by. testing does not significantly contribute

.to.RTS unavailability.1' Reduced redundancy is primarily due:to bypass testing of the RPS and ARTS sensor strings, which has the effectJof reducing the trip logicLfrom 2/4'to'2/3 for the' duration'of the tests.

~

Other.on-line tests (breakers, ' electronic Ltrip, trip modules) are.per-formed with the channel tripped and, therefore, in'a " fail-safe"J con-dition that-does not ' affect unavailability, r

!J L

s i

~

e k

M-N 4.

i

~

o

\\'

n t