ML20215L741

Technical Justification for Changing Rosemount Analog Trip Unit Sys Tech Spec Test Interval

ML20215L741
Person / Time
Site:	Brunswick
Issue date:	12/24/1986
From:	Tony Brown, Michael Jones CAROLINA POWER & LIGHT CO.
To:
Shared Package
ML20215L691	List: ML20215L687 ML20215L724 ML20215L741 ML20215L757 ML20215L780
References
NUDOCS 8705120351
Download: ML20215L741 (15)

Text

m u

TECHNICAL JUSTIFICATION i

FOR CHANGING THE ROSEMOUNT ANAIDG TRIP UNIT SYSTEM TECHNICAL SPECIFICATION TEST INTERVAL

~

~

By ON-SITE NUCLEAR SAFETY Principal Contributors:

'T. L. Brown M. A. Jones i December 24, 1986 87'.iS120351 870501 PLR ADOCK 05000324 P PDR i

~

l 1

I CONTENTS

\

Page

~

l i

EXECUTIVE SINMARY 1

BACKGROUND RASIS 2 CONCLUSION .

3 AFFENDICES

1. Surveillance Test Risk

2. Trip Unit Failures

3. Repeatability Data Attachments:

1. Licensing Topical Report (NEDO-21617-A) ,

" Analog Transmitter / Trip Unit System For

• Engineered Safeguard Sensor Trip Inputs"

2. Trip / calibration System Model 510DU (Instrument Manual 4247-1)

3. Consultant Review:

Energy Incorporated Report. December 10, 1986

Ex7cutiva Summary:

Because of concerns expressed by Plant Maintenance and Operations personnel regarding the numerous and seemingly unnecessary half scrams, half group isolations, and actual ECCS isolations due to Rosemount Analog Trip Unit surveillance, the Brunswick On-Site Nuclear Safety Unit undertook a study of these instruments (drif t and failure rates).

The conclusions of the study are that a ne t reduction in risk (gain in safety) can be realized by increasing the surveillance interval on the Rosemount Analog Trip Units from monthly to semi-annually. -

The basis of these ' conclusions is a study of the instrument calibration repeatability (over 6000 calibrations) and failurgs l for over five years of operation on both plants for 192 trip units.

The study found these trip units to be highly repeatable and to have much lower. actual failure rates than assumed .in the design document (NEDO-21617-A) which provided the basis for a monthly surveillance.

The study also qualitatively evaluated the iner, eased risk (half scrams, half group isolations and out of service " time for ECCS systems) due to the monthly surveillance test interval..

The study concludes that the surveillance interval can be increased without a loss of availability.

In order to ensure the validity of the methods and assumptions used, this study was reviewed by a consultant (G.L. Crellin, PhD of ENERGY INCORPORATED) for independent determination of conclusions (See Attachment 3). The consultant's conclusions confirm the validity of the conclusions of this study.

e 4

e i

--- - - - - , - - - - . . - - - _ , , , - , , - . - - - . - ,- _ _ . - - - - - - - _ _ .- ?_ . - .

- Bruhswick Steam Electric Plant December 1, 1986 Backaround

In the late 1970's, General Electric Company investigated improvea its in the design of the safety system instrumentation of BWR's. This investigation led to proposed modificati'ons for operating plants to some instrumentation of the, Reactor Protection System (RPS) Nuclear Steam Supply Shutoff System. (NSSSS) and Emergency Core Cooling System (ECCS). Licensing Topical Report NEDO-21617-A (See Attachment 1) is General Electric's submittal to the NRC for justification of the design changes. NEDO-21617-A was issued in late 1978 as an approved method to be used to allow improvement s in safety system ins'trumentation.

Based on past operating history and the benefits indicated in the design change , (continuous monitoring of parameters, improved testing procedures and capabilities, and improved operational characteristics), the Brunswick Plant modified safety system instrumentation in the RP, NSSS and ECC systems. During the period of 1981 to 1984, the direct pressure and differential pressure switches that provided the input intgiligence to safety system logic were replaced with the Rosemount transmitters and the Rosemount analog trip system (Model 510DU) .

During'the past five years of operations of the Rosemount Analog Trip Systen

• observations made by both Maintenance and Operating personnel indicated the following:

1. The system is reliable (very low failure rate).

2. The system is stable (requires little, if any calibration).

1

.. . - . . - -. ~-

-1 I

Besia:

1 1

Based on the demonstrated reliability and stability of the Rosemount Analog Trip System it becomes apparent that the current monthly test as required by Technical Specifications and established by the design justification (NEDO-21617-A) is unnecessary and contributes adversely to the following areas related to safety:

1. Large number of half scrams taken for testing (288/ year).

(See Appendix 1).

2. Large number of half group isolations taken for testing (912/ year). (See Appendix 1).

3. Increased out of service time for safety systems (e.g., HPCI and RCIC) . During' testing systems isolate (192/ year). (See Appendix 1).

4. Increased stroking of valves inside the -drywell which may result in increased drywell leakage and potential valve failures.

5. Higher potential for personnel error during testing.

- As indicated in NEDO-21617-A, only the failure rates within the trip units and trip relays need be considered in determining the. test interval. Specifically, the sum of the essential non-annunciated l

! failure rates of the trip units and the trip relays is required to determine the overall system test interval.

I

! Actual failures and failure rates '(See Appendix 2) of the Rosemount Analog Trip Units (Model 510DU) experienced during the past five years

! of operation, applied to Figure 3-12 (Plot Locus of Equation 3-2) of NEDO-21617-A, indicate that a test interval in excess of'six months could be use.d and still maintain an exceptional availability of 0.9999.

l I

Rosemount Analog Trip System stability over the past five years supports an increase in test interval. In a sample of 6.177 of approximately 7,400 monthly calibration surveillance test results- (see l Appendix 3), 96% fell within manufacturer's trip point repeatability l

specifications.

l i

2

. _ .._. _ , .. . ... . _ n.

Conclusion:

Increasing the present monthly surveillance test interval to a six month test interval would provide the following:

o Exceptional System Availability of 0.9999 with simultaneous testing.

o Reduction of number of half scrans taken for testing (reduced from 288/ year to 48/ year). (See Appendix 1).

o Reduction of half group isolations taken for testing (reduced from 912/ year to 152/ year). (See Appendix 1) .

o Reduction of out o,f service time for safety systems (e.g.,

HPCI and RCIC). (Isolations reduced from 192/ year to 32/ year.) (See Appendix 1).

o Decreased stroking of. valves inside the,drywell.

o Lower potential for personnel error during testing. -

Justification for this change in test interval frequency is demonstrated by the operating characteristics of the Rosemount Analog Trip System over the past five years and the benefits related to safety that will be gained.

9 e

e O

3

~

Appsndix 1 Surveillance Test Risk A review of the forty monthly surveillance tests that cover the 192 Rosemount Analog Trip Units involved in this study was conducted to identify testing occurrences that adversely affect safety. The following items were calculated with the assumption that both Brunswick Units would be in operational condition one for the period specified:

. 24 half scrams are generated each month.

288 half scrams are generated annually.

. 76 half group isolations are generated each month.

912 half group isolations are generated annually.

. 16 HPCI and RCIC system isolations are generated each month (these systems are out of service a total of approximately 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> a month for these tests).

192 HPCI and RCIC system isolations are generated annually (these systems are out of service a total of approximately 96 hours0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br /> a year for these tests).

. 82 initiation or permissive occurrences are generated each month.

984 initiation or permissive occurrences are generated annually.

Surveillance testing on a six month frequency would change the above occurrences as follows:

. Half scrams reduced from 288 per year to 48 per year.

. Half group isolations reduced from 912 per year to 152 per year.

. HPCI and RCIC system isolations reduced from 192 per year to 32 per year (system out of service time reduced from 96 hours0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br /> per year to approximately 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> per year for these tests).

. Initiation or permissive occurrences reduced from 984 per year to 164 per year.

1

~ _ _ _ _ _ __. _ _ _ _ . _ . _ . _ _ . ~_

Appandix 2 i

i Trip Unit Failures l o Sources of Failure Data Failure data for the Rosemount Analog Trip Units was compiled in the following manner:

A search of the Trouble Ticket / Work order system to identify failures of the Rosemount Analog Trip Units, and a review of those failures identified against Stores issue records was completed.

The WDTS (Work Order Tracking System), RECO (Records portion of AQUARIUS /SIAIRS) . AMS (Automated Maintenance Management System) and the 0915 (Corporate Materials Management System) computer data bases were queried for failures of the Rosemount Analog Trip Units.

. Searches were conducted on equipment identification numbers.

keywords related to equipment / failures, and by part numbers.

. Trouble Ticket / Work Orders identified were reviewed via the microfilm system and those tickets identified as failures of the Rosemount Analog Trip Units were compiled. (See Page 4 of this appendix.)

.' Stores issue records of Rosea'o unt Analog Trip Units was used

! to verify identified failures.

[ ,

. Failure s were classified as per NEDO-21617-A into the  !

following categories:

f '

Detected Nonessential Failures (DNF)

! Detected Essential Annunciated Failures (DEAF)

Undetected Essential Nonannunciated Failures (UENF)

(See Page 4 of this appendix.)

. A review to ensure that the actual number of failures of the Trip Relays used in the Rosemount Analog System did not exceed the failure rate value specified in NEDO-21617-A was conducted.

1

._.__.,_.e.__.- _ _ _ _ .

Seven Trip Relay failures would have had to occur to exceed the failure rate value as specified in the NEDO-21617-A.  !

Maintenance Trouble Ticket / Work Order reports indicate that '

one trip relay was installed for troubleshooting purposes.

(Work order 2-E-84-1758) . The relay removed from service was functional and had not failed.

Note: The consultant used this relay replacement as a failure in his calculations. (See Attachment 3)

Stores records ihdicate that two other relays were issued but not installed-If all relays i'ssued were considered as failures we would have an actual failure rate 50% less than the failure rate specified in NEDO-21617-A. The conservative failure rate. .

value for the Trip Relays used in the Rosemount Analog Trip System,as specified in NEDO-21617-A is used in this study.

o Sources of Hours of Operation The hours of operation for the Rosemount Analog Trip Units was compiled as follows:

. Review of Plant Modification information with the engineer responsible for installation provided specific dates as to when each Rosemount Analog Trip Unit was placed into continuous service.

f . The last calibration data recorded for each Analog Trip Unit ,

in the Calibration History Data Base was used as the end period for* this determination.

Calculation of the number of days between the initial in-service date and the last calibration date recorded was

! completed for each Analog Trip Unit in this study. These figures were then converted to hours of operation and j totaled for use in calculation of failure rates. (See Page 4 of this appendix.)

l 2

l s

~

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

o S*urens of Fnilurm Rrta Point estimates of the actual failure rates per million hours of operation ( A*) exhibited by the Rosemount Analog Trip Units as described in NEDO-21617-A were calculated using methods described in IEEE Std 352-1975 and the results of the Failure and Operating Hour reviews. (See Page 5 of this appendix.)

O e

G G

4 6

9 6

0 e

D 4

3

i ROSEMOUNT ANAIDG TRIP UNITS (HODEL 5100U)

FAILURES AT BSEP

~

Trip Unit Work. Order Date classification

1. 1-321-PTS-N023D-2 1-E-82-1645 04/19/82 DEAF *

2. 2-521-FTM-N015A-1 2-E-82-3010 07/15/82 DEAF * ,

3. 1-321-FDrM-N008A-1 1-E-83-2043 06/03/83 UINF*

1 4. 2-321-LTM-N031A-3 2-E-83-2542 07/04/83 DEAF *

5. 2-521-LTS-N0315-3 2"E-84-1758 03/19/84 ,

DEAF *

6. 1-321-LTM-N036-1 1-E-84-1948 04/17/84 DNF*

i

7. 2-321-LTM-N017D-1 2-E-84-6236 10/20/84 -

UENF*

UINF*

8. 2-321-FIS-N023A-2 2-E-85-2441 05/10/85

9. 2-321'-PTS-N023A-2 2-E-85-2466 05/11/85 DEAF

• I 10. 1-E51-FDIM-3018-1 1-E-85-4615 10/08/85 DEAF

• 1-E-85-4615 10/08/85 DEAF *

11. 1-E51-FDTS-N018-2 2

1-E-85-4687 10/13/85 DEAF *

12. 1-E51-FDTS-N018-2

13. 1-B21-FIM-N023D-1 1-E-85-5211 11/05/85 DEAF *

14. 1-521-LTS-N031A-3 85-AMPG1 12/29/85 UINF*

f i

• DNF = Detected Nonessential Failures .

• DEAF = Detected Essential Annunciated Failures

! *UENF = Undetected Essential Nonannunciated Failures

, TOTAL OPERATING TIME IN NOURS FOR THE ROSEMOUNT ANAIhc TRIP SYSTEM: 6,467.784 i

i i

i 4

~ _ _ _ . _ _ _ _ . _ _ _ . _ _ _ _ _ _ _ _ . _ _ _ . . _ _ _ _ _ _ _ _ . _ _ _ . . _ _ . _ . _ _ _ . _ _ _ _ _ _ . _

PLANT FAILURE RATES:

Estimate of mean failure rate

[ = ,n . n = failures T* = hours of operation g4 .

4 = Nonessential failures + 1 = .155/million hrs.

A -

6,467,784

= Essential annunciated = 9 = 1.39/million hrs.

1 B failures 4,467,784

. \s A = Essential nonannunciated = 4 = .618/million hrs. -

d -

C failures 6,467,784 W D .

Trip Units = .618 +d Trip Relay = 1 = 1.618/million hrs.

h C C ,

Total .

i

)

1 O

J 4

i l

5 I .. _ _. - - - ..

Appsndix 3 Repeatability Data The previous appendix concentrated on failure data. In this appendix, the concentration is on the drif t or repeatability of the 'instruasnts.

Operating history of the Rosemount Analog Trip Unit System was.

reviewed to see if it tended to support or confirm the stated annufacturer's specification for Trip point Repeatsbility as per Reference 1 of NgD0-21617-A. (See Attachesnt 2) This was accomplished by reviewing past calibratio.n data.

o~ Sources of Data Calibration data for the Rosemount Analog Trip Units was compiled ,

in the following manners Approximately 7,400 calibrations have been conducted since the Rosemount Analog Trip Units were installed. Copies of these .are maintained in the microfilm system. Of the total number of calibrations conducted. 84% (or 6.177 data sets) were still available in hard copy form in Maintenance calibration folders (for convenience, the hard copies are purged as storage is needed, even though they have already been microfilmed). It was.

determined, therefore, that only the hard copies of the calibration sheets need be reviewed as they represented a sample I size larger than necessary to lend credence to a statistical analysis of the results. -

Calibration data from the hard copy sheets was entered for each of the Rosemount Analog Trip Units in ascending order by dates. ,

i i

The breakdown of data is as follows:

. 2.757 (71% of the approximately 3,900) calibrations were for

• j

• Unit One instruments.

1

~

. 3,420 (98% of tha cpproxicctoly 3,500) celibrctions wero for ,

t Unit Two instruments.

. Note that although Unit one Analog Trip Units have been in servihe the longest, hard copy file storage constraints for j these working calibration folders necessitated purging some i of the folders so as to maintain the past few years of data sheets for review. (Those purged are still retained on  ;

mierofila.)

o Data Gaps In assembling the data base of the Rosemount Analog Trip Unit calibrations, a review was conducted to determine what gaps or

' discontinuities existed in the information by date. The results of this review were as follows: -

. Of 512 gaps identified in the data 329 (642) were related to unit ' outages. These tests are not normally required during outages and/or when fuel has been removed from the vessel.

. 68 (13%) were related to- the removal of data from the Unit one es11bration folders for storage constraints.

. 115 (23%) were identified as data sheets apparently microfilmed but not retained in the maintenance files.

It was determined that the presence of these gaps is not consequential,to the results for two reasons: (1) the significance .

of the sample size, and (2) the analys,is discussed below shows that 96% of the sampled surveillances (6,177 monthly surveillance tests) tes t results fell within the manufacturer's specification for trip point repeatability. .

o Data Base The assembled calibration history data base was therefore determined to be adequate for review of the drif t and setpoint repeatability of the Rosemount Analog Trip Units , installed based on the number of calibration data sheets recovered (84%) and the review of the gaps in the data entered.

2

o Repvtrbility Note: Data sets for each Rosemount Analog Trip Unit were compared

-for setpoint/ trip point drift by computing the absolute value of drif t in mil 11 amperes from one calibration data se t to the next. This was completed in ascending order by comparing the as-left value of the previous data set to the as-found value of the next data set.

A review of the data base was conducted to verify that the Rosemount Analog Trip Units performed within the manuf acturer's specification for Trip Point Repeatability as per Reference 1 of NED0-21617-A (see Attaclinent 2). The most stringent tolerances for normal operating conditions and normal environment were applied to the data base (+ or .13% of span for Master Trip Units and + or

.20% of span for Slave Trip Units) with the following results:

. 965 of the 6.177 monthly surveillance test results fell within the manufacturer's specifications for trip point repeatability.

. Review of the distribution of data indicates a normal

^ ^

distribution around an average drifL value of -0.000959 mil 11 amperes. ,

. Over their operating history, the Rosemount Analog Trip Units have exhibited a slight drift in and around their "

setpoint, staying well within the manufacturer's specifications. In fact. approximately 2.400 calibrations showed zero milliaspere drif t in setpoint values.

o conclusion The results of the calibration data review supports the performance requirements for trip point repostability for the Master and Slave Trip Units specified in Reference 1 to NEDO-21617-A.

8 W