ML20235H274

From kanterella
Jump to navigation Jump to search
Mechanical Level Instrument Availability Analysis
ML20235H274
Person / Time
Site: Cooper Entergy icon.png
Issue date: 05/31/1987
From:
DEVONRUE, LTD.
To:
Shared Package
ML20235H271 List:
References
NUDOCS 8707150082
Download: ML20235H274 (40)
v • d • e


Text

_ ____ _. _ _ _ - _ _ . _

l l

l i

IEBRASKA PUBLEC POWER DISEUCT l MECIIANICAL LEVEL INSTRUMENT AVAILAllILITY ANALYSIS May 1987 1

l PREPARED BY DEVONAUE 108 LINO 3LN STREST B707150002 870706 PDR ADOCK O5000298 BOSTON MASSACHUSETTS 02111 P PDR

Nebraska Public Power District Mechanical level Instrument Reliability Sudy June 8,1987 1

TABLE OF CONTENTS l

SECTION TITLE PAGE SECION I INTRODUCTION 2 SECTION 2 METHODOLOGY 4 SECTION 3 ANALYSIS 6 SECTION 4 DATA and RESULTS 24 SECTION 5 CONCLUSION 38 SECTION 6 REFERENCES 39 APPENDIX A DEFINITIONS and ABBREVIATIONS 40 1

1 1

1 De mnrue

. Nebraska Public Powe7 District Mechanich! Ire! Instrument Reliability Study June 8,1987 l

i

1. INTRODUCTION i i

On October 26,1984, the Nuclear Regulatory Commission (NRC) issued Generic Letter 84-23  !

(GL 84-23) to all Boiling Water Reactor (BWR) licensees of operating reactors conceming Reactor Vessel Water Level Instrumentation in BWRs (ref 1). GL 84-23 reviewed the BWR Owners Group 1 report, " Review of BWR Reactor Vessel Water Level Measurement System" [ S. Levy Inc. Report SLI 8211]. The Owners Group report provided a generic evaluation of water level instrumentation l adequacy of BWR/2 through BWR/6 plants, and identified severalimprovements in BWR water level measu.cment and instrumentation.

The Staff concluded that the changes identified in the emergency procedure guidelines were adequate for the short term, but that permanent physical improvements should be made on a deliberate schedule to reduce the burden on the operator. The potential improvements were placed into three categories: (1) level indication errors, (2) instrument reliability, and (3) protection system logic.

The first category consisted ofimprovements to the plants that would reduce level indication errors caused by high drywell temperatures. These improvements would prevent reference leg overheating

- or reduction of the vertical drops in the drywell. The second group ofimprovements address the

. issue of analog versus mechanical instrument reliability. The third group involved changes to the protection system logic may be needed for those plants in which operator action may be required to mitigate the consequences of a break in the reference leg and a single failure in a protection system channel associated with an intact reference leg. This change is still under evaluation by the Staff and i is not required to be mqncmented by GL 84-23. I i

By letters dated December 13,1984 [ref 2] and May 31,1985 [ref 3], Nebraska Public Power District (NPPD) responded to GL 84-23 for Cooper Nuclear Station (CNS) by evaluating the various methods ofimprovement that are available to reduce the burden on the operator with respect to indication errors due to overheating of the reference leg caused by high drywell temperatures. I NPPD also made provisions to replace mechanical level indication with analog devices in order to i

2 Devonrue i

Nebrada Public Power District Mechanical Level Instrument Reliability Study June 8,1987 increase instrument reliability and availability. The response and schedule of the proposed improvements were found to be acceptable to the NRC by letter dated August 21,1985[ref 4].

I This report summarizes the results of a reliability study comparing CNS mechanical instrumentation with analog devices. The results indicate that the availability of the mechanicallevel instruments at CNS not only exceeds the industry average for mechanical indication, but also compares favorably with the industry average for analog devices.

3 Devonrue

I Nebraska Public Power District Mechanical Level Instrument Reliability Study June 8,1987 2 METHODOLOGY l

l 2.0 General Two basic approaches are taken in this study to compare the reliability of mechanical level indication instruments to that of analog devices.

~~

The first method compares the operating history of CNS mechanical instruments against similar instruments used in the industry. This comparison also compares the failure history of the mechanical instruments currently in use at CNS against the experience with analog devices .

The second method of analysis utilized in this study examines the availability of the system functions required by operation of the various level instruments. The data collected in the first method will be used as input to this phase.

2.1 Comparison of Instrument Retability Ilistory The history of the mec!.sical levelindicators at CNS between 1981 and 1986 was researched in order to determine the failure history of each instrument. Also the Nuclear Plant Reliability Data System (NPRDS) provided industrs data for comparable mechanical instruments as well as any existing analog instruments in use. The reliability of the instruments based on industry operating history was compared to the reliability of the CNS instruments based on CNS operating history.

k i
a.  ;

4 Devonrue i)

- Nebrasta Publi Power District Mechanical LevelInstrument Reliability Study June 8,1937

)

2.2 System Availability Due to Instrument Reliability IIIstory i

The second form of analysis developed a reliability model that determines the availability of the plant systems associated with the mechanical level instruments. This was accomplished by reviewing the General Electric circuit diagrams for system initiation and trip logic for all systems associated with the appropriate level indicators. Logic block diagams were then developed for each system to be used as input to the model. These logic diagrams are shown in Figures 3.3 through 3.14.

Several factors such as equipment failure rates and mean time to repair are taken into consideration in order to devise an availability number for each instrument. These availabilities are then used to develop a MARKOV model which the instruments as redundant systems and then determines the overall unavailability of the initiation rnd trip logic associated with these instruments.

i 5 Devonrue l

_ _____ _______-______- _ Q

' Nebroska Public Power District Mechanical Level Instrument : Reliability Study June 3,1987

?

u .

3. ANALYSIS 3.0 General This section describes the analytical process used to determine the availability and reliability of the mechanical level indication instrumentation at CNS. The analysis was undertaken in two phases.

The first phase was to gather data on mechanical instrument reliability and repairability for both the industry as well as CNS plant specific data. The second. phase consisted of developing a model to represent the availability of the system trip and initiation logic actuated by the various components.

The following subsections describe each of the two phasedn detail. j I

l 3.1 Mechanical Instrument Operating IIistory Operating history was gathered for mechanical level instruments for CNS components by reviewing the LERs issued for the plant between 1981 and 1986. The same information was developed for instruments throughout the industry that are. comparable to CNS instruments. Rate of

,L failure as well as Mean Time Between Failure (MTBF) and Mean Restoration Time (MRT) were determined for CNS instruments and instruments throughaut the industry. The results of this survey are discussed in section 5.

a-3.2 Availability Model The process widely used to determine the unavailability of systems consisting of redundant components is based on Markov models. Markov models consider a system to be made up of a number of states which represent the various conditions of operation in which the system may i exist. For example, the components of a system may be (1) operational, (2) in standby, (3) in a test mode or (4) in a repair status. Other states may be designated depending on the application. The probability of the system transitioning from one state to another can be predicted by detennining the 6 Devoorue

Nebraska Public Po wer District Mcchonical Level Instrument Reliability Study June B,1987 e

4" o i

components' rate of failure and rate of repair. ,

An illustration of a typical Markov model is shown in figure 3.1. The system consists of two .

redundant components, each having a failure rate 1 and a constant repair rate .

SIMPLE MARKOV MODEL:

TWO. REDUNDANT COMPONENTS n n

~

0 1 2 JL 46 2A A STATE O - both cornponients are operational STATE 1 - orie cornponent is under repair and the other is operational STATE 2 - both cornponents are under repair A is the hazard rate (1/hr) p is the repair rate (1/hr)

FIGURE 3.1 A system can transition from one state to another as shown in Table 3.1 7 Devonrue

Nebraska Public Power District Mechanical Level Instruraent Reliability Study June 3,1937

<p l

BEGINING STATE TRANSITION ACTION ENDING STATE

~

O 1 OF 2 COMPONENTS FAIL 1 1 REMAIN lNG COMPONENT FAILS 2 2 REPAIR 1 OF 2 COMPONENTS 1 i _

1 REPAIR REMAINING COMPONENT 0 l

TABLE 3.1 l

3.2.1 Formulae l

_ Determining the unavailability of the system involves calculating the probability that the system will be in any given state at any given time. For the purposes of this study the probability of a system being in state two is of particular interest. It is in this state that the system under examination is unable to perform its intended function. The transition probabilities can be mathematically i represented as:

dPo (t)/dt = -2AP 0 (t) + P i (t) dP3 (t)/dt = 2AP0 (t)- pPi (t)- AP3 (t) + pF2 (t) dP2(t)/dt = AP3 (t)- P 2(t) where:

P0 si the probability that the system is in state O Pi s ithe probability that the system is in state 1 P2 si the probability that the system is in state 2 t is the time time of concern A is the failure rate is the repair rate j

and the initial conditions are defined as:

P0(0) -1, Pi(0) - 0, P2(0) - 0 l

8 Devoorue

Nebraska Public Power D! strict Mechanical level Instrument Reliability Study June 8,1937 1

l This system of three first order ordinary differential equations can be solved for P2 (t) by eliminating P 3(t) and P o(t) to obtain one third order differential equation in P2(t). Although the full solution to this equation etn be found in reference 5, only the steady-state unavailability (i.e. t tends to infinity) is presented hee. The steady state unavailability for the system is given by:

1 l

P2(~) =, A2 j(p + gy2 3.2.2 System Modeling For each CNS system in the Technical Specifications, whose initiation, trip or interlock logic was affected by the mechanical level instruments, a Markov reliability model was developed. A complete listing of these systems is shown on Table 3.2. I For the purpose of this study two families of components will be analyzed. Those thr.t operate in I series and those that operate in parallel as shown in Figure 3.2 As is obvious from the block l diagrams, each of these systems can be broken into a combination of these families of components.

As shown below the unavailability for parallel components is the product of the individual components and the unavailability of the series components is the sum of the individual components unavailability.

Parallelwtavailability = (NA+p)2 Series unavailability = 2(NA+p)

' Each system shown in the block diagrams has been analyzed and the results are discussed in section 4.

i 9 Devonrue

Nebraska Public Power Dittrict Mecharical Level Instrument Reliability Study hoe 8,1987 l

l SEIIIES COMI*ONENTS CONIPONENT A -

COM PONENT B

I'AIIALLEL COMI'ONENTS l

1 COM PONENT COMPONENT ,

A 13 i

i

)

FIGURE 3.2 10 Devonme i

s

Nebraska Public Power District MechanicallevelInstrument ReliabiEty Study June 8,1937

. ,a q

~ j

,. i i

SYSTEMS / FUNCTIONS ANALYZED FOR LEVEL INSTRUMENT RELIABILITY  !

1 systern/ function figure Automatic Depressurization System 3.3

- Residual Heat Removal System 3.4 (Containment Spray Mode)

Core Spray System 3.5 EDG Initiation 3.6 HPCI System 3.7

' Primary Containment Isolation (Group I) 3.8 Primary Containment Isolation (Groups II, III, VI) 3.9 Primary Containment 1 solation (Group VII) 3.10

. RCIC System 3.11 Residual Heat Removal System . 3.12

- (14w Pressure Coolant Injection Mode)

Reactor Protection System 3.13 Reactor Recire Pump MG Set Trip 3.14 TABLE 3.2 11 Devourue

l Nebraska Public Power District Mechanica11evel Instrument Reliability Study June 8,1987 '

... . )

l 1

l AUTOMATIC DEPRESSURIZATION SYSTEM 1

l i

i NBI LIS 72A NBI LIS 72B I

NBI LIS-72C NBI LIS-72D NBI LIS 83A NBI LIS-83B l i

f

\

l l

FIGURE 33 l

l 12 Devoorue

. Nebraska Public Power District Mechanical level Instrument Reliability Study June 8,1987 1

( .i' l RESIDUAL HEAT REMOVAL SYSTEM CONTAINMENT SPRAY MODE 1

l l

l

.)

NBI LITS 73A NBI LITS 73B 1

j l

1 I

l l

l FIGURE 3.4 13 Devonrue

~

~ Nebraska P5blic Power District Mechanical level Instnament Reliability Study June 8,1987 e.

i l

CORE SPRAY SYSTEM l

i NBI.LIS.72A - NBI.LIS 72B l

l

I NDI-LIS.72C NBI-LIS 72D -

l l

FIGURE 3.5 l

14 Devonrue

t i

Nebraska Public Power Dis:rict Mechanical Ixvel Instrument Reliability Study June 8,1987 EDG INITIATION NBI-LIS 72A NBI LIS 72B )

i NBI LIS-72C NBI LIS-72D FIGURE 3.6 l

15 Devoorue

Nebraska Public Powe7 District Mechanical Level Instrument Reliability Study June 8,1987

.~ O HPCI SYSTEM

  1. ssssss ,,sssssssssssssss,, ssssssssssssssssssssssssssssssss,,ssssssssssssssssssssssss I N t t t t t 4 t

t t 2 t t i R 4 i

s t

s NBI LIS 72A NBI LIS 72B t t t t t t t

? E INITIATION

t t t t t t t t t t t t

R R t t t t i NDI LIS 72C NBI LIS-72D E t s

i
4 t

t i

i I

4 2- t t i t- t t t

.t4m,mmmumsmmmmmsm,mmmmmmumsmummsmumm t

%%%,4%%s%%%%%%4%,s%%%%%%%%,K9 2

s I-E NBI LIS 101B E E L  !

s t t, t TURBINE TRIP i

t% t

NBI LIS 101D  !

t 1 i t t

t i i t I tmmmmmmummms t 1

1 FIGURE 3.7 l

16 Devonrue i

Nebratka Public Power District Mechanical level Instrument Reliability Study ' June 8,1987 PRIMARY CONTAINMENT ISOLATION .i (GROUP I) u j

l NBI LIS 57A NBI LIS 58A i

NBI LIS 57B NBI LIS 58B .

I 1 i FIGURE 3.8 1'

17 Devonrue

___-_-__--______s

, Nebraska Public Power District Mechanical level Instrument Reliability Study June 8,1937 y .

l PRIMARY CONTAINMENT ISOLATION l

(GROUPS II,III,VI) I l

l

.1 i

i i

l NBI LIS 101A NBI LIS-101C '

NBI-LIS 101B NBI LIS-101D FIGURE 3.9 18 Devonrue i

1

Nebraska Public Power District Mechanica! 1evel Instrument Reliability Study ' June 8,1987 PRIMARY CONTAINMENT ISOLATION (GROUP VII)

{

1

.j.

f NBI LIS 57A . NBI LIS 58A l

I NBI LIS-57B NBI LIS 58B i

i i

FIGURE 3.10 1

19 Devonruc l l

Nebraska Public Power District Mechanice.1 Level Instrument Reliability Study - June 8,1937 1

RCIC SYSTEM s m , m m m m m ss m s m m s m m ss m s, m m m m m s m m m , m m , m m ;

I 0 t t  :

I I I

t t t

i 1 i ts i, t NBI LIS 72A NBI LIS 72D I  :

i i t i

! j INITIATION t

4 i i

i i

i NBI LIS 72C NBI LIS-72B t I i i  ;

i i  :

i i

, t i o

t t t

t t t

i 1 i s ssssssssssssssssssss s ssssssssssssssssss,s sssssssssssssssss, sssssssss ssssssssss

%sssssssssss%% s%,,sssssssss, is is i NBI LIS 101A R t i i

0 s

ts R TURBINE TRIP 0

ts ts.

t NBI LIS 101C t  ;

R R t t t t t t t t t

t s'

, wmmmmmmvm mws FIGURE 3.11 20 Devoarue

..____________-_-_A

debrasks Public Power District Mechanical Level Instmment Reliability Study June 8,1987 )

=

l  ?

l RESIDUAL HEAT REMOVAL SYSTEM l

LOW PRESSURE COOLANT INJECTION MODE i

NBI LIS 72A NBI LIS 72B i

NBI LIS 72C NBI LIS 72D I i

FIGURE 3.12 21 Devonrue i

Nebraska Public Power District Mechanical level Instruraent Reliability Study June 8,1987 l REACTOR PROTECTION SYSTEM l

NBI LIS 101A NBI-LIS 101C I

NBI LIS 101B NBI-LIS 101D FIGURE 3.13 22 Devonrue

l Nebraska Public Power District' Mechanical 1.evelInstrument Reliability Study _ June B,1987 I

i l l

REACTOR RECIRC PUMP MG SET TRIP SET A -

i NBI LIS 57A NBI.LIS-53A I

a SET B NBI LIS 57B NBI LIS 58B FIGURE 3.14 Devonnie

, ', Nebraska Pubtic Power District Mechanical tevel Instrument Reliability Study June 8,1937

4. DATA AND RESULTS 4.0 General This section begins by describing the failure and repair rate data of the mechanicallevelindication -

instrumentation at CNS, the mechanical level indication instrumentation throughout the industry, and the analog instrumentation throughout the industry. The data is then applied to CNS systems to -

estimate system unavailability.

4.1 Data A failure (or hazard) rate and repair rate is required to model each type oflevel indication instrument of concern. For both mechanical and analog instruments in the entire industry, failure rates were extracted from the NPRDS. The data also includes a mean failure rate for those instruments with 95% confidence intervals.

For the mechanical instruments at CNS, licensee event reports and nonconformance reports were also used in combination with plant operating history to obtain a site-specific mean failure rate for each component. These reports revealed that from the ;;riod covering 1981 to 1986,14 failures occurred among the 16 CNS instruments during nearly 40,000 operational hours. A CNS instrument failure rate of 2.19 x 10-5/hr was calculated from this data. A student t-test was used to calculate a 95% confidence interval for CNS failure rate Although a similar procedure was followed for estimating repair rates, the data were analyzed more closely to reveal events that were not consistent with the modeling. Repair rates were established L for (1) mechanical instruments throughout industry. (2) mechanical instruments at CNS, and: (3) l l

analog instruments throughout industry.

24 Devonrue L

l

, , Nebrasks Public Power District Mechanical Level Instrument Reliability Study June 8,1987 l

l 4.2 Failure Rates The mean failure rates developed in this study are summarized in Figure 4.1.

MEAN FAILURE RATES 2.50 E-05 2.00 E --

s i

1.50 E -

M 1/HR 9/

1.00 E-05 -= E'y /m, 5.00E-06 -= ihd

" ;, i 0.00E+00- i i -- i i I  ! i MECH MECH ANALOG IND CNS Figure 4.1 Mean failure rates for various instruments.

From Figure 4.1, it is apparent that the mechanical instruments at CNS have a failure rate (FR) that is comparable to the industry norm, while analog instruments have a failure rate that is much less than the normal failure rate for mechanical instruments.

To bound the uncertainties in the estimated failure rates, a 95% confidence level was established for each of the various instruments. The results for the lower confidence limits are presented in Figure 4.2, and the results for the upper confidence limits are presented in Figure 4.3. l s

1 1

25 Devonrue  !

l I

, Nebraska Public Power District MechanicallevelInstrument Reliability Study June 8,1987 1

I i

l 2.5% CONFIDENCE LEVEL FAILURE RATES l

2.00E 05 l 1.80 E 05 -- ggy i 1.60 E is i l 1.40 E 3]nA -

1.20 b 05 - i&T 1/HR 1.00 E 05 - $Es. i 8.00 E ET l 6.006 x 4.00E 2%  !

2.00E 3(I$

0.00E+00- i i-i i'

MECH- MECH- ANALOG IND CNS Figure 4.2 Lower confidence limits for the various failure rates.

97.5% CONFIDENCE LEVEL FAILURE RATES 3.00E-05 2.50 E -

2.00 E -

1/HR 1.50E 7? -

1.00 E 'J' 5.00 E ' :#; -

1 0.00E+00 - I ii i i -u MECH MECH ANAL.OG I

IND CNS Figure 4.3 Upper confidence limits for the various failure rates.

26 Devonrue

Nebrasko Public Powe7 D: strict Mechanical Irfel Instrument Reliability Study June 0,1987

, 'O 4.3 Repair Rate 4.3.1 CNS Mechanical Instrument Repair Using licensee event and nonconformance reports, it was established that all CNS repairs of mechanicalinstruments were completed within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of failure. Specifically the NPRDS failure rate code gives a mean repair time for CNS failures of 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />. A repair rate of 0.0416 (1/24) /hr is assumed for calculations to provide additional conservatism.

4.3.2 Industry Mechanical Instrument Repair The mean repair time for mechanical failures in industry therefore was obtained from a NPRDS listing for mechanical instmmentation. From the listing, a date for repair initiation and completion could be used to estimate the time needed to repair the component. In fact Eighty-three failure events and corresponding repair times were found. Figure 4.4 presents a statistical characterization of this data.

mechanical repair times  ;

Mean: Std. Dev.: Std. Error: Variance: Coet. Var.: Count:

532.916 1417.243 155.563 2008578.859 265.941 83 ,

1 Minimum: Maximum: Ranae: Sum: Sum Souared: # Mirsino: )

24 8808 8784 44232 188275392 1 Median: Mode: Geo. Mean:

l 24 24 92.015 l f 1

Figure 4.4 Statistical description of the repair times (hours) gathered from the NPRDS for mechanical instruments. )

27 Devonrue

_________-__-__--________________--___a

Nebre.ska Public Power Dittrict Mechanical Level Instrument Reliability Study June 8,1987 1

A scattergram showing the actual repair time values is given by Figure 4.5:

Scattergram of mechar.ical repair times 9000 ,

e 8000-c h

7000-a n e 6000- ,

.c 5000-a ,

I 4000-f 3000- e e , o P

a 2000- +3 i

1000- + d =-

r ,, ,,,,,,,,,,,,,,, ,,,,,,,,,,,, ,,,,,,,,,...,.y

,e,,,...

,, ,,,...v.......................

,,,,,,,,,,,,,,,, , e ,,,,,,,,,,,,,,,,,

s --nR 0 -y ............. 2 2.... -b; Observations Figure 4.5 j A scattergram of the mechanical repair time data (hours).

i The scattergram shows that most of industry's mechanicalinstrument repair times are very short, (i.e., typically requiring one or two days). A few outliers (8%) exist that require thousands of hours to repair. Since these outliers account for such a small percentage of the overall population surveyed, and since they vary considerably from the the population mean, they are judged to be nonrepresentative of the actual time needed to repair instrumentation. The large repair times reported are likely due to procedural or administrative concerns, rather than reflecting actual repair time. For instance, a utility may have decided to delay repair indefinitely until some convenient time

-- an outage for example. The long duratien entries then reflect the time the utility waited to repair i the component, which may be several months or even a year.

These outliers were consequently removed from the data, and new statistics generated that more properly reflect mean repair times for mechanical instrumentation. The censored data is provided in figure 4.6 28 Devorne

, Nebaska Public Power District Mechanical LevelInstament Reliabinty Study June 8,1987 mechanical repair time- select data Mean: Std. Dev.: Std. Error: Variance: Coef. Var.: Count:

177.662 246.335 28.072 60680.99 138.654 77 Minimum: Maximum: Range: Sum: Sum Sauared: # Missing:

24 912 888 13680 7042176 7 Median: Mode: Geo. Mean:

24 24 67.745 Figure 4.6 Statistical description of mean repair times (hours) for mechanicalinstruments with selected data removed.

In comparing the description of the full data against the description of the selected data, the mean is reduced by a factor of three. The variance is greatly decreased by a factor of 33 indicating significantly greater confidence in the revised data.

4.3.3 Industry Analog Instrument Repair 4

A similar procedure was carried out for analog instruments that have been in use by industry. l Fifty-eight failure events were found in the NPRDS and are described statistically in Figure 4.7. I analog repair times Mean: Std. Dev.: Std. Error: Variance: Coef. Var.: Count:

l 1657.552 2251.361 295.618 5068626.392 135.824 58 l 1

Minimum: Maximum: Ranae: Sum: Sum Souared: # Missina: l 24 6528 6504 96138 448265412 l26 l Median: Mode: Geo. Mean:

624 24 460.476 i

l l

Figure 4.7 Statistical description of analog instrument repair times (hours).

1 29 Devonruc l l

1

, . Nebraska Public Power District MechanicallevelInstrument Reliability Study June 8,1987

)

i A scattergram of this data shows the variance in this data more plainly:

Scattergram of analog repair times 7000-a eees e se e n

6000-a l

o 5000 *e 9

4000- +3

r. )

e

  • p 3000 e a e a i

2000- ...............e.........................................................................+b- =

4 f

, o

,....................................................................R..........F........R t- 1000 e e e ee -Os se e e en i e e8 se e e e se e 0 2- : S*:::

m 222  ? ?", M e

S "d 1000-Observations Figure 4 8 A scattergram of the analog instrument repair times (hours).

I Again, it is judged that many of these data points do not reflect the actual repair times. For example, most (12%) of the long duration repairs occurred at a single unit, and are not applicable for our purposes. The data was screened, much in the manner the mechanica1 repair data was screened, and all repair times of greater than 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> were ignored. Figure 4.9 describes the " censored" data set.

I-1 30 Devonrue

_ _ _ _ __ _ - _ _ _ _ _ _ _ _ _ -___________ - ____~

Nebraska Public Power District Mechank a! L.evel Instru ment Reliability Study

~

June 8,1937 -

l i

analog repair times Mean: Std. Dev.: Std. Error: Variance: Coef. Var.: Count:

309A33 297.273 49.545 88371.057 l 95.946 36 MinirOm: Maximum: Range: Sum: Sum Sauared: # Missino:

24 888 <1 864 11154 6548868 48 l 1

Median: Mode: Gen Mean:

276 24 ,143.798 Figure 4.9 Description of analog repair times (hours) for the selected data.

F

)

.s

,0 r

31 Devereue e

Nebraska Public Power District Mechanical level Instrument Reliability Study June 8,1987 h

4.4 Results For each of the systems shown in the block diagrams, a low, mean and high unavailability was

] calculated based on the following assumptions:

es I, nd acan repsit rates for the industry were taken from the " select" data which ignored outliers kii h

$ {I 2. exh component in the system had the same failure rate ve.

e 3. nine failure rates were eonsidered:

a. the CNS failure rete -- lower limit, mean , and upper limit
b. *e industry meel,anical failure rate - lower limit, mean, and upper limit
c. the industry analog failure rate -- lower limit, mean and upper limit These results are provided on Tables 4.1 through 4.5 I

32 Devonrue m' ~ -

7. ,

Ecbraska Public Power District Mechanical level Instrument Reliability Study June 8,1987 -

1 l

d

' AUTOMATIC DEPRESSURIZATION SYSTEM UNAVAILABILITY

, -J I

l

)

t l

Low Mean High CNS 1.6 x 10-5 2.2 x 10-5 2.8 x 10-5

' Industry 6.4 'x 104 1.0 x 10-3 1.5 x 10-3

- Analog 4.1 x 10-6 9.5 x 10-5 5.0 x 104

(

1 i

TABLE 4.1 33 Devonrue 1

, 3 i, .

Nebraska Public Power District Mechansal LevelInstrument Reliability Study June 8,1987 -

.c RECIRCULATION PUMP MG SET TRIP A & B, CO.NTAINMENT SPRAY Low Mean 'IIigh CNS 2.1 x 10-7 2.8 x 10-7 3.5 x 10-7 Mechanical 8.3 x 10-6 1.4 x 10-5 2.1 x 10-5 Analog 5.1 x 10-8 1.1 x 10-6 6.8 x 10-6 i

l '

l 4

l I

TABLE 4.2 34 Devonrue IE.

N2 braska Public Power District idechanical Level Instrument Reliability Study June 8,1987 PRIMARY CONTAINMENT GROUP VII Low Mean Iligh CNS 3.2 x 10-6 4.4 x 10-6 5.6 x 10-6 Mechanical 1.3 x 104 2.1 x 104 3.3 x 10-4 Analog 8.1 x 10-7 1.9 x 10-5 1.0 x 10-4 TABLE 4.3 35 Devonrue

l Nebruka Public Power District Mechanical Level Instnament Reliability Study June 8,1987 RCIC,IIPCISYSTEMS UNAVAILABILITY I

Low Mean High CNS 3.8 x 10-5 4.4 x 10-5 5.0 x 10-5 Industry 4.9 x 10-5 6.9 x 10-5 9.4 x 10-5 Analog 1.6 x 10-6 9.4 x 10-6 3.1 x 10-5 TABLE 4.4 36 Devonrue

Nebraska Public Power District Mcchtnical Level Instrument Reliability Study June 8,1987 i

i BASIC SAFETY SYSTEM GROUPS UNAVAILABILITY

  • Low Mean Iligh CNS 4.1 x 10-7 5.5 x 10-7 7.1 x 10-7 Industry 1.7 x 10-5 2.7 x 10-5 4.2 x 10-5 Analog 1.0 x 10-7 2.4 x 10-6 1.4 x 10-5

l TAULE 4.5 l 1

37 Devonrue

____-_____________J

Nebraska Public Power District Mechanicallevel Instrument Reliability Study June 8,1987 5 CONCLUSION The conclusions of this analysis are as follows:

(1) Instrumentation repair practices at CNS are sufficiently effective to keep the unavailability of systems dependent on level indicating instrumentation at an acceptably low level.

(2) CNS mechanical instrumentation repair times appear to be better than the repair times experienced by analog devices of the type considered as replacements for the current mechanical systems.

(3) The availability of CNS systems that are dependent on level instrumentation would not necessarily be enhanced by converting to analog devices. l The failure history on analog devices is not yet developed enough to be able to raise a conclusive argument in favor ofinstalling these instruments at CNS. The high availability of CNS mechanical components further suggest that analog instrumentation will not provide substantial additional j reliability. On this basis, the replacement of the mechanical devices with analog instrumentation will not produce a decernable enhancement in safety.

1 J 38 Devonrue

, , Nebr:sta Public Power District Mechanical LevelInstrument Reliability Study June 8,1987

6. REFERENCE
1) EISENHUT, D. [1984], REACTOR VESSEL WATER LEVEL INSTRUMENTATION IN BWRs, ( GENERIC LETTER 84-23) (1984), Director Division of Licensing, Office of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Commission, Washington, D.C (1984)
2) PILANT, M. [ December 13,1984], Reactor Vessel Water LevelInstrumentation in BWRs (Generic Letter 84-23), Technical Staff Manager, Nuclear Power Group, N ebraska Public Power District, Columbus NE, (1984)
3) PILANT, M [May 31,1985], Reactor Vessel Water Level Instrumentation in BWRs (Generic Letter 84-23) - Supplementary Response, Technical Statf Manager, Nuclear Power Group, Nebraska Public Power District, Columbus NE, (1985)
4) VASSALLO, D [l985], NUREG-0737, ITEM ll.F.2, INADEQUATE CORE COOLING INSTRUMENTATION, Chief Operating Reactors Branch #2, Division of Licensing, U.S. Nuclear Regulatory Commission, Washington, D.C (1985)
5) MEYER, P.L. lottroductory Probabilistic and Statistical Applications, Addison-Wesley, New York (1972)
6) Institute of Nuclear Plant Operations, Nuclear Plant Reliability Data System, Atlanta GA (1987) i I

39 oevonrae 1

i.

' Nebraska Public Power District Mechanical level Instrument Reliability Study June 8,1987 .

c APPENDIX A DEFINITIONS and ABBREVIATIONS'-

BWR Boiling Water Reactor CWS CooperNuclear Station IER. Licensee Event Report t

Markov Model A reliability model that determines the probability that a particular component will transition between operating states based on operational history NPPD Nebraska Public Power District NPRDS Nuclear Plant Reliability Data System i

NRC Nuclear Regulatory Commission l

States A predetermined operating condition that applies to a particular component j within a system MRT. Mean Restoration Time - The mean time necessary to repair a particular instrument based on operating history 1

l MTBF Mean Time Between Failures - Mean operating time between instrument failures i i

based on operating failures.

l 1

40 Devonruc l

l i

______________

Retrieved from "https://kanterella.com/w/index.php?title=ML20235H274&oldid=3226402"