ML20118A140

Staff Exhibit S-76,consisting of Paper Entitled, Westinghouse Setpoint Methodology for Control & Protection Sys

ML20118A140
Person / Time
Site:	Farley
Issue date:	05/18/1992
From:	Miller R, Tuley C WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
References
CIVP-S-076, CIVP-S-76, NUDOCS 9208030129
Download: ML20118A140 (5)

Text

W#

STAFF EXH' D6 hi do-a ys m /-dJUA i

w l

O /S ED IILE Troteactions on fiuvear Sciet;ce. Vol R No. t, retfuary 4 m

A.

[s wmmut un0m v.timwcv ioR Comet. m rmrcuos uses i4 p4 c, e, kley

+ %u mF 4

i

g. o mil,,

s t

I s.J Wst ita hvuse Elec tric Corporat ion Qy

r. o. m 3ss v2 JIS -2 All :29:

Fittsburgh, Fennsylvania isDO eI b ty i 18:2 c --

j w %,

SLhs0R DR!f r (SD) - the change '15 'intkt-putt >o APS W CT j g *m ct *j

, i b D:l West inyhouse t as due leped an ap;rcach f c>r de t e r er.i r. int c on d i t 1 tins.

M h D

[+ D c-han e l 2nst r ument ation urd e r taint ies whic h uses the 9.i I; *** h sqv.are roet cf t he sum cf the quares, wit h proper ENTIRONMINTA1. A1.1.0WA.NCE (EA) - t he thange in sensor /

k account ity f or stat istically dependert paranctess.

transmitter output due to adverse radiatiot) and

[ f/ w$

Ttta paper develtqs t he equat ion and disc s.rse n t he tempera ture ef f ects f rom a lia.iting ac cident conditien.'s E$

inderer.de nc e of t he vari m pararrters.

It mise tro-Typically this value is determitied f rom a conservative "

P G:

vide, a meth s et restructarirt t he plant't. Teibnical set et enveloping condit tens.

yC Ssecifications.

3 RACK cal.IFRATION ACCLRACY (RCA) - the reference INTRDUCTION (carit ration) accuracy, as defined by $AMA Standard

.D i p aa IMC 20.1-19 D [1], f or the process rack modules (cards)$,'

1ri the early I M r s, u stinphouse utilleet the methW in an i n s t r urne nt loop.

It is assumed tt.at whil( tach y l

of arit' m tir s unr.a t t e r fer Itx determination et the mod it or card is individually calibrated tn a

• ?

r i

D; uncertainties tur a Imac te r T r ip e r infim e r ed Sa f et y specific tolerance, that all rack medules are String I

featute* N tuattan H stix [retecti<m functarn. Tbi6 calit r at ed t o within an overall tolerance value.

methocolen w r ade wate at that time ' e c a u rn t he-uncertainti e and al1N m es us N wttc relatively easy EACF CCeUALATOR SET 71% ACCl%tCY (RCSA) - the ref erence

to coset and still alla more tr an adequate to r for (c a l i t'r a t i er. ) accutacy, a s d e f ined tw S AbtA Standard h

oper a tint ea t Ein and f le a t h li t.

hewe e r, in the PMC 20,1,19 M [1], of the instrument loop cotrparator p;

l i 1U 7/M t ine frau, with incr<ased avateness en the (bistarle).

M bl part of te t t Lntirmhmse and the bEC t hat additional A

uncertaintier, and allevances should be i nc lude d.

FACK U.MTEFAT1'RE Ef7FCTS (FTE) - change in input-output D l

Westanghaust-init ia te d a p rog r a!' t o r e-e w.l ua t e the selationship tor the process rack madule strirg da to [*

y !,

traditicnal approc h ar.d r e ficc t the nn or incre a"J a change in the ambient tempetature from the ref erence {[

f paramete r value s.

As a re uit of this prcgram, t he c.alibr at ion cenditions.

p

" b' 3

methodolvy uwd t o c ecl i r.c the uncartaintics was

4

=

I r e v t r.e d t o the squarr roet of t he sw c f t he quares LM K DRIFT (RC) - the change in input-eutrut relation-.j

]

apprcach w;th apptcyriate treat unt of ststint ually ship over a per Cod of

• me at ref erence conditions, ej c

i dependent parnetetk, The above parameter dca lt iers are applicatl, f or all M1i (LP We s t ir gnouse supplied equipment (Fambero, barton and j

W rit t ak /Tobat transeittera, and roxboro, Hagan 7100 ar.d Westinghouse 7 L process racks). Some slight 6

h<finitia 4

l rodi f ic ations at e ricc usary to the det ermination c f the m

The pr am te rs t erm1& red in t his arrre n t are pra-transmitter values for kasemount transmitterb, since l3 phrased tel N:

their specifications are r.omewhat ditterent than the j

other th*ee, but the everall errors are essentially the

[ i Ffst$s MhASnLUT Mit AAM (HiA) - allawane ter saw,

j e re;ats ctfects u ith have a dire ct

)

non-f rait r ume r o

tearing en ttt a curcev c tra t um r t c harme l ' h Sen ur/ Transmitter Alle.ances t w ing, e.g.,

t erp roture sttoinficat M to a lar diamettt rip As noted f r ott the definitions, Lestinghouse uses five I

parameters directly related to a transmitter. 5 Fk1MAFY fl &BT EG F m U LA) - trrct due te the use STE, STE and SS will be discussed in this section and v

c' a metering dn itt, c.v.,

c1tew, orifice, or IA will be dibcussei later. Two of these parameters

venturi, are considered to be indepsadent (SPE and STE) and two are conservatively considered to t,e depeadent i%A SENS0k CAllfRM.* ANACY (ECA) - th( referer=e and SD).

The SF[ and SU tern are indeperdent I, de (c alibrat iur.) acc urac y ter a sens t er transmitter as they -re deviations of two separat e ef f ec t s f rom two defined ty SPA Standard FT & l-lV 3 [13 different unr elated ref erenc e condi tions.

The

)

sagnit udes of ther.e two t erms a re based on We st inghouse j l

SENSDP FRESSCFl FFfECTS (SPE) - the charge in input

• and vendor equipment specifications. The SCA and SD cutput relatienship due te a c hange in t.he static h ad terms are conservatively rioted as being dependent i

l pressure from the calittaticu conditions (if calibra-because at the time the taethodology was developed, very e

tien is perforund at line pressure) on the accuracy to f ew plants were recording "as lef t/as fo md" data to l

whith a correctirn factor is int roduced f or the assist in the determination of drif t, It the mid 70's

{

differente Letwet n colit rJ.f en arm cl erat ing c c.nd i t i on x

ra n y plant s usir p Wer,tinghouse prete

tten systetDs g

l fer a d/r trarrmitter sitTly reterded the fact that the calibraiien was j

vithin spect'ication and net the ac tual "a s le f t" Srner H*J WT F E f rf

• C " (sni - the thme in input-u lue.

When periedir surveillance was lat r perfor ud, output re la t it rM i t chs o a charp in the art ier -

t h e. cnly.etcrrinatien possible 6as the de iation fic:

f temyerature (f or s ap ; t e d nerm! ope rating condit ions) the ner;.inal set pcint and not the a c t ual deg iat ier. f t cC f r e:: t h e ref erence c.. trratier r cn n t iens abo st a the "as left" point. Kestingbouse then det?rn.ined that I

t transmitter.

the a ssu:Lp t ien that these tsc parauters weie inter-l active, er dependent, weuld rcnservatively tound the

(

9208030129 920518 Oc% 9m W0mm OM19% IEEE PDR ADOCK 05000348 G

. h, PDR

. i

~ ~ - - - - _ _. - _ _ _

[

NUClDi' P!LU,fsIIlh'f ( DJ','.2$lON

' h !b. _

_j bLvmah0'att b_e_ifrmhk

1. !0 DO:/tt fi;

,'u

~, _. _ C

~~2Mu2:t % m airk, e, u < ~ o.-

c:

w.w.

,c.-

e, cf,

(,

1_,

bW

'ri _...

A_

th.

U'4%f__...__.

5 (}..

I)[)_b

^ $ fN.

N,d0

'k

=

t

611

- actual uncertainties of calibratica plus drift for a If a plant perferibeJ a string calibraticn of the given transeitter. 1ne magnitude of SCA 16 based on iristrument loop, including the bistalIs, and recorded the combination of the vendor's equipment specification the "as left" data for the string, it would be for the reference accurar) and the plant's calibration possible to reduce the number of tetes in tne procedure for deviction f rom the neatnal calibration mathodology to RCA and L. Littlitation of the "as value. The value for $D is based on the vendor's lef t" data to confirm that the RCA term was satisfied 3quipa,ent specification and Vestinghouse review of a N *as found" data to determine the RD ter% it would drift data.

be possible to remove the conservative assertion that f

the two tetts were interactive and calculate the loop j

Tf a plant recorded "as lef t/as found" data during uncertainty based on the assumption that they were I

ellbtstion and periodic surveillance it would be statistically independent. In the authors' opinion yaa reasonable to treat SCA and SD se independent this would result in an approxis, ate 251 reduction in i

{..

eth u.

With the teneration of a suffs.ler,tly large the total loop error.

1--

data base, a wre accusate, statistically r,ased j-pertination of t he SCA and SD terms would be nossible.

Other Allowances In the author s' opinion t his would result in an L

approxissate 5% reduction inthe overall uncertainty for The three remaining allowances tw be diatussed are IM,

.l s-instrument loop and would alot result in a bettet FEA and LA.

As noted in the defintrions. FttA is an determination of the adrquacy of a transad tter's allowance for non-instrument related efterts. These operation. The recording and utilization of the encomrass a large number of dif ferent uncet taintles -

as lef t /as f ound" data is strongly recommended by which have applicability to some, but not all, of the Westinghcose and is t.onsistent with the recomendationa Westinghouse protection functions. Examples of the J

of ISA Standard 567.0*,1982[23 PMA terms used aret temperature stratification in latte diameter piping for measurements with RTDs, the impact

'j The Westinghouse teference to SAMA standard PMC of a change in Tavt (which trapacts the water density 20.1-lV L) (1) t ri re gards t o $CA (and RCA and RCSA) has and thus the neutron leakage current) in the downtomer 1

quite an Imr.act on both tLe plant and the calculated region of the reactor vessel in f ront of neutron unc e r t a in t y for the various protection functions. The detectors, the impact of the accuracy of the rod SAMA standard not e s that when calibrating an inst rument control system (Tayg) on the density cf the primary i

(transmitter er rach module) the accuracy of the t ooh nt in the cold leg elbow where d/r tranmitters q

calibrAtton equipment must be taken into account if the are used to infer flow rate, the accuracy cf a l

uncertainty of the calibration of test equipe>cnt is precision flow calorimetric used to norr,altre the cold greater than 10; of the calibration value, e.g.,

if leg elbow tap d/p transmittera, ar.d the impact of the l

the calibration tolerance is 10.51, then the accuracy accuracy of the Pressurizer pressute cor trol system et the calibration or test equipment snu s t be less t han (which impact s the density of the water) on the 2.051 !! it is to be ignored.{3) If this accuracy I'ressuriser level transmitter. Some of these vatuo

+

Is not met, then the actual accuracy of the calibration involve detailed calculations with mult iple urget t atnt y j

equipment should be incluJed as part of the SCA, RCA, groupings which are reduced to a sini.le value. Other p

o-ECSA value, This is in contrast with IIII Standard values are simply allowances which are known to b, j

49bl980, which notes that the accuracy of the conservative and boundinr.

j calibration equipment should be less than + 25% of the calibrat tr'n t olerance.(4) Westinghouse has determined FEA in an allowance for use cf a metering device for

'V that ignoring the calibration equipment accuracy has the measurenent of flow. This term is med tsy little or rio tryac t on the Westinghouse methodology if Wertinghouse for elbows and veF urts, but w* l d d i so b

the SAMA otandard is met, however, this is not be used f or an orifice if one were unco in a necenarily truc if the IEEE standard i s n.e t and the protectien function. When an cibow is used for a calibration equipment ancer taint y is > IS? of t he protection function, the d/p transmitter used to mke f

cTlibration tolerance. The end tesult of this effect the measurement is normalized to a precision fin is that the plut must either purchase very accurate measurement to redoce trie uncertainty for the flev T

calibration and test equipw nt or lose some degree coefficient for the elbuw,

'lhe PEA term is t hen an

]

of operating margin, allowance f or the ncity signal that !s characteristic of use of an elbew. Vhen a ventuti is used, this term Rach Allow nces is the accuracy of the determinatibn of the fios coefficient for that device, This is usually pericued from the definitions it can be seen that Westinghouse at an independent laboratory tu a high confidence uses four parametern f or the procesh rack uncertainties,

level, kCA, ECSA, RTE and RD.

In the Westinghouse approach, RCA, RCSA and kD are conservatively noted to be The EA term accounts for the impact of the hiph I

depender,t patameters and RTE is considered te be temperatures and radiation exposure due to a 11rtting rdependent, for much of the saw reasons as for the accident in the vicinity of the transmit ter, for tronardtters, Westinghouse assumes that the rack Reactor Trips and ESF function actuation, the r tu ry r

calibration and drift terms are dependent. The inclu-contributor is the hig,h temperature that may be sion of t he RCSA team is a specific allowance f or the em erienced at the beginning of a Nin Steamlire treak,

plants that calibrate the trip bistable separately from a Nin Feedwater Pipe Break, or a Loss of roclant the rew inder of the racks.

• his is not enattly Accident. The second principal contributor is the consistent with the string calibration assum tion ef fect of high radf ation exposure f rom a l~ss et noted in the definition but it is an acceptable Coolant Accident. These two contributors are procedure within the methodology. The magnitude of the arithmetically summed to conservatively bound the kTE rere is based on Westingbuse qualification test total impact of the event. Because of the methods in data for the racks. The RCA and ROSA values are based which the transmitters are specified and envitors-en the combination of the eeuipment's reference mentally qualified. Westinghouse treats EA as if it scuracy ar d the plant 's precedute (c r allowable were a bias t e ra.

As a result of recent testing of a

aviation fra the nominal calitration valu. The value mort than SC trant.mittars performed by ew of the for RD is based on equipment spccificaticas and the principle supphers to Westinghouse, it was determined resu?ts of a Westing %use review of rack drift data.

that the historir,al value used for EA for those transmitters was non-conservative by as much as 40%.

'!his demonstrates the need f or a corspreheraive 4

e i1.

j i

46 6tg i

T3 - {(kCAp kCSA +RD ) +(kCApRCSA +D ) 3" f$)

1 qualifiestion test program that involves mora than just y

y y

f three er four s q les to represent a large population if one expects to determine the magnitude and areas of and the subscripts indicate input chant.el 1 or 2.

The l

random vari tion of an uncertainty with a hifh Allowsble Value is then the Nominal Trip Setpoint plus confidence levtl. As a teault of the same testing the most limiting of the T values (11 or 12 for single p r o gr as., at ny be possibla to detero.ine random and input f unctions and 12 or T3 f or multiple input bias components of the temperature cortribatien.

functions), kh n the process rads plus Listable have tiewever, s. ore testing is necessary bef ore it is drifted past the Allowable Value then the IK staf f I

possible to otilite thie approach on all We6tinghouse has three optiunst (1) check the transmitter for the supplied transmitters. As can be seen, Westinghouse channel to detertnine if it has drif ted in a sianner that i

esercises great effort to determine uncettainties in a would ccupensate for the sack drif t, (2) determine if conservative asartner (usually by setting uppet bour:d the function has sufficient sargin to accept the l

values) to establish high cenfidence limits.

etnitude of the drift, and (3) do nothing to determine ustgin, just recalibrate the rack and note the 1:ncertainty Combination deviaticn foi the plant's reports, As noted in the previous sections, the Total Channel Westinphouse calculates the Maximum Value to be used in Allowanc e (TCA) is ahe square root cf the sum of the tLe plant's surveillance procedures tc assist the I&C squares af ter f orming statistically inoependent groups.

staif in the determination of whether snere is Troy the above discussion it ha s be en no t c 6 wha r e sufficient margin in the function (TA n 1CA) to Westinghouse u.akes conservative assumptichs soncerning accept the rack drift. The calculation for the britu.,

the treatment of the various paracters. The rasulting Value assumes that the transmitter has drifted to it's eauation as as follows:

worst condition allowed ir Equatior 1, i.e.,

the transmitten drift is SCA + SD and in the non-conserva-

+

TCA; * ((fha) +(PEA)'+(SCA+SD) +(SPE) +

tive direction. The equation f or this value 1s t 0)

- [ (i rA)*+ ( FLA)' t SU)2+ (ST E )2 a

(STE)* * (RCA+RCSA+D)*+ (RTE)y]if', ' fA.

9

+

T re

• TA (6)

)3 p' - (:TA* SD) - LA.

If the plant recorded "as lef t/as f ound" data aM (k1F) utilized it in the determination of drift and pe fctmed a string calibration on the racks that included the if the histab!e's tri-setroint is greater than Tmax, bistable, then Equation I could be revised to as then the I&C staff must either check the transmitter fellows:

for cog ensating drift or simply determine that the channel is out f specif; cation and recalibrate, y'

k 6

[(PMA)y (PEA)'+(SCA)2 (SFE)2 (STE)2

+

+

+

trgires*ing prudence indicates that any time the TCA;

=

(2) channel is fcund to be outside of the hominal Trip (SD)'+(RCA)'+(RTEf+(RD)y3 p* + EA.

Setpoint plus kCA + RCSA, the channel should be i

adjusted to within the calibration tolerance.

The Total Channel Allowance is then co::. pared with the Tc tal Allowance (TA) which is the dif f erence t,etween CONCLISIONS i

the Nottinal Trip Setpoint (in the plant 's Technical f pecifications) and the Safety Analysis Limit (the In the determination of instrument channel value for the trip setpoint used in the Safety uw ert a ntiet;, Westinghouse atten.pt-to include all Ana!ysis). Obviou*1y TCA must be less thw. er equal knern and reasonable errors and allowances, te TA if the Safety Analysis assumption concernir; the method used to combine the. 'ccriainties 1* the ttip setroint it to be satisfied, con screa t ive a pproac h of the a:uare root of the sum o

the squares aftet accounting its p t a :1.t tic a lly TECHNICd. SPECIFICATIONS dependent effects. The methodology makes several conse rvative assumptions with regards to the determin-Westinghouse uses the components of Equation 1 te ation el dependency between several sensor / transmitter det ermine an ellowable Value and a Lximum Value for cr rack paramercrs. The importance of proper use in the plant's Technical Specifications and calltratien techniques and use of accurate (,.librathn surseillar.ce procedures. The Allowable talue is the equilment was explained and the effects of the use of value to which the process racks and t istable se dett t either the SAMA or IEEE standard for the determinatien before specific action must be taken by the plant 16C vi acceptable calit ration equipment accuracy was noted.

staff. As part of the overall approach, Westinghour.e The equation developed to c.lculate the instrument has determined that the Allowable Value s.ly be ore of uncertainty leads to several subsets that may be used several calculated values, which ever is the most to calculate the allowed rack deviation determined l uiting. The first value (Tl) is simply the arithme-during periodic surveillance which is usually larger tt: s a of the rock calibration and drift values, than the rack drift value alone. This methodology has teen in ute sinc e June, 1976 with explicit nC Il = RCA + RCS A + ftD.

(3) approval (with vegard to the calculation of Allowable Values for use in the plant specific Technical i

The second value is based on the asseption that the Specifications) since August, 1982.[5]

transmitter drifts in a random fashion for both magni-i tude and direction and that parameters riot checked are REFERENO :

also randenly distributed with riorr.a1 or near normal prchability distributions. This calculation is

[1] Scientific Apparat,.ukers Association, Standard con.arvatively approximated by, FR: 20.1-19 7 3, "Prt c e r..

Measurement and Control Terminolog y, ' g. ge 4.

l

[(PMA)*+(FEA)'+(SCA+SD)2+(SFE)2 (STE) (

i T2 = TA -

+

)

[^

Instrureut Sectety of Am rica. Standard 567.04, (M E)2 W* - EA 19E2, "htpoints for Nuclear dafety-helated 3

Instrazentation Csed in Nuclear Pow =: Plaats,"

!l For protection func71cns that have more than one input page 12

-I to the bistable, the NkC has determined that either T2 cr T3 steuld be used f or the Allowable Value. wnere-o,,I l

1

r '

,h,

• ?

s DJScientificArpsrstusMakersAssociation,Ste.ndstd'

3) *

- PMC 20.1-1973, " Process Measurement and Control

'feroit. ology," page -16.

[4]InstituteoftacetricalandElectronicEnpineers, 4

5tandard 498-1980, "IEEE Standard Requirements for

' i

--*~ '

. the Calib : tion and Control of Measuring and Test Equipment Used in the ",onstruction and Maintenance

• g of Nuclear Power Generating Stations," page 6 et 4

{3) hTRIG-0717 Supplement No. 4. "Saf ety Evaluation

}

Report Related to the Operation of Virgil C. Summer Nuclear Station, Unit ha.1. Docket No. 393,"

pp. 7 7-3.

i i n >-

s' _..

-Q

.ua'k

? 3

.n i.

':?

6) g(I

.? k.

W I,'

,1

_b.'

^k N

.. -:rg

.r a

D d

k, l 1 ~

'g is l' i

.t

.s c

13 s

j

'.:r

.