ML19209C866
| ML19209C866 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 03/10/1976 |
| From: | Bennett G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| Shared Package | |
| ML19209C834 | List: |
| References | |
| TASK-TF, TASK-TMR NUDOCS 7910180397 | |
| Download: ML19209C866 (20) | |
Text
.
MAR 101976 EJC CSURE HINAN RElABUZff S'IUDY Past Studies Survey i
' l f' il a u.n L,
-s Scrne work on hean reliability was dcne for the Feactor Safety Study (References 1 and 2). This scrk basically involved an extrap laticn of data cbtained fr:m other progrrs (e.g., aercscace, weapons) with sane en-site cbservatien of the h=an coerators. hhile the scrk that was done was undcubtedly adecuate for the pur:cses of the Feacter Safety Study (FSS) it is worth noting that FSS (Faference 1) states:
"An actuarial data base for he:mn error rates in nuclear pcwer plants does not exist."
and "In general, the desien of centrols and displays and their arrangerents en crerater panels in the nuc] ear plants s*udied in this analysis deviate frx. hcan engineering strda-ds specified for the design of ran-rachine systes a.d accepted as standard practice for military systcrs."
Specific proble s included possible errcrs in crerating similarly d::ig c.ted cen':ci fun:,.icn: er.:1 ::.
cf: :. : 2 cit-d :==
specific exrples of i.nger h=an engineering in nuclear plants:
. use of si-4' " appeardr.g centrol displays
. lack of practice in si--,ted emergencies
. poor fer at for cperating precedures
. tco rany auditcry/ visual pr.el displays cam 2 ting fer operator attentien
. side-by-side 1ccation of valves (which might cause both net to be cperated)
. occasicnal icw degree of hean redundr.6 in cer' *5
/
cperations.
'IM RSS also noted the abse ce of data en hcw the creraters right respond in a high-stress situatica (e.g., IrCA) r.d stated that "to cbtain acre accurate evaluaticns, each particular situation : ust be individually r.aly:ed to ascess the specific hu:an failure rate which is applicable."
(Eefere re 1).
N g-O [g, a t
()
j
)I 7910180 3 7 [
t
- ENCU3SUFE Beyond the cuestien of croratcr reliability, RSS stated that "h=an errors related to the testi.g and maintenance of c=renents can also be irgertant em en xde centributors."
(Faference 3). In additicn
" design defects" were noted as scurces of c2 men mde failures (Paference 3) and in a sense, Sis ie aleo a h=an er= r.
There are, of ccurse, a nu.ber of reference raterials dealing with the general subject of h=an er.gineering especially as it relates to aerospace and military rissicas (cf., e.g.,
Faferences 4-3).
These reference raterials highlight a ncher of general tcpics concerning human perferrance sucn as:
. wrking cenditiens (lighting, noise. levels, terperature, spacing, stress and fatigue).
. egai; ment desian (spacing of censoles; sice and lccatien of buttons, switches, etc.; need for standardication)
. definition of the appropriate man-machine interface gx
. selection and training of personnel h
\\gh
\\@h i
. raintainability QD 9 u
. application of decisien analysis
. What decisicns :=st the cperaar rake to acecr:=lish each fu.ction?
. What inferruticn is regaired for each decisien?
. isnnt action is regaired to execute the decisien?
Both EC and NSIC recularly collect and disse-i. ate i.foration en safety-related occurrences in nuc'ow pc. cr plants (cf., e.g.,
Faferences 9 and 10). Tcile not directly acplicable to esta.blishi.g the desired data base en hean er=r, dey de p= vide a useful reasure of the decree of hean in;civement = U2 safety-related cc==ences.
Tables 1 and 2 frca Re#-=~ a 9 and Tables 3 and 4 f=m Pafere.ce 10 show that desicn e ncr, crerator er = r, 4-*-=ce er=r and administrative e=cr head the list of causes of these safetv-related occurrences. The old adage alcut these cccuma.:eq beigc a eans cf the plants trying to tell us scretning is no dcubt t=e; derefore, it wuld behccve us to collate and study these occurr=ces fur.har.
NSIC has prcvided me with the results of a literature search en hu: an reliability in nuclear scwer plants. ' his is available in my office.
T. ~,' *l r,
\\
J ECICSUPS The Brrwns Ferrf nuclear plant fire provides a recent and p licly visible test of reacter cperater cor"crece (references 13 and 14).
The NFC Special Review Grcuo stated that "de TVA coerating staff....
1 d-aved in exc nlary fadien." (Paference 14). Ecwcvar, the 3:30 Special Review Grcup gces en to note:
- 1. There were lapses in cuality assurance at Erowns Ferry.
- 2. There is a need to develcp standards and recuirerents for instru entation recuired for operator inferatica and acticn.
- 3. There is a need for ceriedic drills. inmlving all ensite and offsite orcanizatiens which ray be excected to res :end to a fire. Paference 14 cites the FRC inspecticn repcrt of the Browns Ferry fire (see Faference 13, pp. 213-6S5) to cbserve that there were "a n'ci:er of exarples where the acticns taken by the plant cperating staff during de fire are stated not to be irdicative of a high state of trai-dng of plant perscnnel in fire fightinc cperaticns."
Present Stufies NRR has recently lat a $100K/vear centract with Parcstace Corceration to ar21y o reacter centrol recr. disolays and cperater per#c=mce.
Cn Februar/ 19, 1975, I briefly dis =ssed t"is stufy.de Oc ric 7t:ndi, who ic 'M '?R project crager for the study. Ee said Aercscaca Cor;cration will study abcun five nuclear pir.ts, specifically plants which have Pad "ser.i-rajer. nds of events."* Ee said he is interested d
in raking decisiens en hcw cne c n reduce cperator errcrs particularl*/
during pes'"' W accident ccrditicns (Attac!: ent 1 is de secpe of w rk for this effort.)
The RES Prchabilistic Aralysis Branch is also startinc a study at Sandia en h= =, reliability.
"he secpe of is stuti is yet to be defined and FIS/ PAS u interested in any input EES/*.GSR ray have.
EPRI is funding'scre wrk en relidility and diacncstics (S3,576K in 1975) and cperatien a-d rainte 2nce (590CK in 1975) but these studies appear to be related rcre to ecuirnr.t a-d sofr. care (Peference 11).
Discussien with Jchn A. Ycder Per ycur recenrerdaticn I discussed the hran relicbility ratter with John A. Ycder befcre he left ::RC. A s=rary cf Yoder's points as derived frcnt his ::avy and If!" e.gerience fcllcws:
- J. A. ::crberg has cbserved t'at plants with gocd roccrds sPculd be studied also to determine where the differences are.
"! G s
- (
g-
\\l L s
s
D*OD
^D ~9',
fi fd o'
4 vvu_
m t
_4_
NM 1.
The optiens personnel shculd revica the design.*
2.
The systens desicners shculd write the precedures.** ?.e procedures shculd not he cluttered with superflucus raterial, i.e., assee the cperators are intelligent, well-trained perscnnel.
3.
Hean reliability is a functicn of the design ard each desien is different.
~4.
Early in the desien process, the project shculd genera *w a doct:nent that sets forth de cuidelines en hcw to censider the operaticnal asecces in the desien. Such a doc =ent wculd have to integrate all systers, e.g.,
. Ecw does the overall systen werk?
. Faa do the parts fit tecether fran an options point of viea?
Designs which rake cperatiens so reutine that an cperater can err shculd be avoided.
5.
L. frag:ent cperaticns are the rest likely cause of an acciC_-nt.
Dr" m.s sh~.:ld be er#cred pric te act-"'ly cenducti.g an infrequent cperaticn.
6.
Certain critical fenctions shculd be designed such that it takes a t.c-step precess to acco=lish the function.
7.
Fca an cperatcr responds to a situatien (e.g., an alarn) is a function of the status of the plant. Startue causes a lot of alar :'s and the ccerator ray overleck scre because they are not critical but cerely an i.dication of a change in the plant ecndition.
'3.us there is a need to catalcq what conditiens cause ala.ms and what to leck for when an alarn cccurs'and what to do about alars. Cne idea to help tra cperatcr sculd be to have a ca =uterized eniter which scans all sic.als, identifies which sicn l(s) caused the n' m and crevides recG TnGnced Ccursas c'. aC*OCC.
~
- J. A. Norberc reco re.ds "that cperators, and escecially the key su p isers, shculd be involved as early as pessible in the desian and constructica of nuclear pcwer plants."
- J. A. Norberti and W. S. Far:rr have reccnrunded that the cperators writo the procccures.
[,"-
,,,)
- I S'
hM
D D
o' D qP
-[
oo 6
i eHL1Jr2 oe
- ncIosupp Discussien with W.
F'. Vesely (2/23/76)
A str:rary of this discussica is as folicws:
1.
The utilities tend to desien reacter cent:cl recrs to lock n:uch like fossil plant centrol recrs.
("his pecits then to use fossil plant cperatcrs at their nuclear plants.) The centrol h:ard layouts can be groceed approxirately by the utility.
2.
The layout of a centrol recm is deternined to scnu e:< tent by the manufacturers a.d the sucpliers. The utility has the final say en labeling, al2cuch each plant within a given utility ray have its cwn labelina. This difference in control board layouts inlies different si::ulaters wculd be needed and that the htr an reliability evaluaticn alrest recuires a plant-by-plant evaluation.
3.
The location of valves is set according to dr:raings and net for the ccnvenience of the operations persennel. Also there is often no posicive indication steder a valve is cpen or closed. The colcr coding has been noted to be i.censistent within a plant (e.g., red can nuan "cpen" and it can also
- r. san "cicsad").
4.
There are two basic areas to be stadied:
. existing plants - hcw to irpreve ths:n
. standard plants - hcw to stard'9 2e the centrol beards 5.
In studying hu:an reliability with the viea of ir preving htr an perforance, one trust censider the critical areas (cf., e.g.,
U SH-1400). since it rakes se.se to ec.cen. rate en these operatiens where the hran ele ent is critical.
(ca-dia is friliar with these cri.tical areas.)
It shculd also be noted that c.anges in existing laycuts or prcccdures can lead to accidents because of cperator unfa-iliarity with the nea syste:n.
6.
Typically a plant will have e.so or thrh crerators ter shift.
At ti:nes there ray cnly be ene crerator in the centrol rectn with the other cceratcrs perfe ri.c walk-thrcuch inspections (again pcssibly en an individual basis). Check lists are not always used.
7.
At Zicn there were an average of three to five alar s cer hcur.
Usually the crerators snut the:n eff.
(As an aside, 75s of de SAC 163 errer rate for air creas was caused by a failure to believe instrtncnts.)
80 ap c
D*9D W T
] I@r b J AL J]
o-J
~ DEIDST. RE 8.
Sandia has a h=an r'eliability creanicatien of at least 27 people who run a laboratcry where actual hean reliability tests can be perforn2d and 'd.o :aintain a data bank on hran errors. The laboratory is not like a reacter c cetml roem.
W.e labcratory is used to reasure hcw 1cnc it takes to perfern an cperation as well as hcw well it is cer#cred.
(If an operator takes too lcng to perform it ray be of no use.)
9.
'Itere is a need to establish an NPC-wide ccordination of hu: an reliability studies.
- 10. The FES/P.4B study at Sandia will:
. exterd the present Sardia study of ::icn (Faference 12) to other plants
. study candard board laycuts
. establish a data bank en the functicns h=ans actually perfor.n by en-site investiga icns.
Faccrrended Future Courses of Action
- erences, casec cn ene :.orecoing discussion u.d my cursory review or t.ye rea
.t. taink :Nm snculd institute p.vTs to adcress tre following 1:rs:
1.
A systcri of cericdically testi.g cre ators under pre-and ecst-accident conditiens shculd be es-d lished by de Office of Inspection and.*._.fercr.ent.
O.is ray be done d-cugh actual tests (i.strrenza _ien cnly) perfered en de cperators in de plant and withcut their advance :e.raledge.*
Scme srecific ite-s which need to be adhessed d ~-1 da restense tire, type of respense, a.d wheder de operators are evericaded by the types of =b-s which can cccur. Si ce perfo ra.cc effectiveness is a function of stre'ss levels (see Ficure 3), it might be sorth establishi c to the cperators' satisfaction that they are " safe" no ratter 'h.at the accident.
An cutccme of dis werk shculd be cptirized rath-ds cf instr = tion, warning a.d ccr=nica icns not to mentien irn:cved precedures.
In addition, recular traininc shculd i.-creve the crerator ras cnse time (Faferences 1 and 2 note that disbelief and inactica <.culd be the initial rescense to a rajor accident.)
- Cbvicusly dere is no uncmirity abcut using the plant itself as a testing grcu..d.
I'cwever, as Faference 13 brcught cut, scmetrl.g must be dcne abcut training reacter perscnnel en a regular basis to handle e ergencies.
3, 7 I) 8 ; r*
.(
k
0 D
- NW 2.
RSR should evaluate "the recevery facters to deternine the opti;nra degree of ccm=ensatien for hiran errers, e.g.,
. should irrediate feed 1:eck be provided en the results of each cperation?
. how nuch personnel redundancy is recuired?
. shculd there be i:r.Minte testing by a second party?
In general an everall ran-rachine interface study needs to be perfor:ned for nuclear plants (see Tables 5 and 6).
3.
The htran aspect must be integrated into the desien frun the very begi= ing. Apprcoriate guidelines for bran encineerinc should be established by PSR to chtain ecuircent characteristics which facilitate error-free and safe perforance of cceration, centrol and " Mtenance tasks. Ficures 1 and 2 illustrate dis process. The designers rust censider *.b.at tasks are to be per-ferred because this daterines the ::aining. Frem dis will ccrne training pir.s, pmcedures and the apprcpriate testing and evaluatien activities.
ucnsiceratien ray nave to ce given to ne use a
,cc= au..ws of centrol roces and cther critical plant features to deterine hcw best hirans can cperate in de nucle plant.
4.
Supplcrentary to the 57R and PIS/PA3 prcgrrs there should be an everall prec'In to establish de guideli.es for centrcl recn and plant desiens and inferation systrs with de view cf minimizing hiran er=rs thrcugh i proved design practices a.d procedures.
5.
CMIPC shetild establish a ccr=:rehensive, unified reference data base en hiran reliability as it refates to nuclear cc.cer plants.
NRR and IE shculd be ir/olved in de evaluatien and i:iterpretaticn of the data. This data base shculd include:
. cperator reliability
. raintenance persc=el reliability
. test persc=el reliability
. ad:inistrative centrol perrc=el reliability t
U. 6 m
. EMURE The data base may be develeped threugh a ccrnbination of:
. data frcra safety-related occurrences
. data frcra related progra:rs (only as directiv applicable)
. experinents with ooeraters under pre-and pcst-accident conditiens (including alle'.cnce for psycholcgical aspects and the " successive errer" p:chlea).
. evaluation of the differences in plant desicns, cpera* dens ard accident situations.
Frcra the foregoing five ite s, it shculd be possible to set up three study progrars as folicws:
Study 1: Testing Procedures (itea 1 above) (IE lead)
Sctdy 2: Design, interfaces and cuidelines (ites 2-4 abcve)
(FSR lead). This study cculd ccnceivably be a tPree-step process leading to the guidelines.
Stufy 3: Establish a data base (itea 5 above) (CME:C lead)
It is reccrr.e-ded that I: cst or all of these studies be done bv cer=ecent aerospace er :nilitary firs, using nuclear censultsnts as needed.
'~he reasons for this re m e.dation are:
. These firs have : cre experience in the hran engi..eering field.
. These firs w:uld bri.g a fresh viec:cint to t'.e studies.
h ol@o S s-B 865 177
D LN2 g h!h"
~
Peferences:
(/
1.
Practo" c' #atv Stutr, An Assess.ent of.secident Risks in U.S.
Q Cc r.crcial '.ac!. car 32 tor i'lants, Artenc III: Fa12ure Data, U.S. Nuclear Faculatcr/ CcrTalssicn, scH-1400 (:;UREG-75/014),
Octcber 1975.
2.
A.D.SwainandH.E.Guttr. ann,"HurE.PaliabilityAnalysis Applied to Nuclear Pc.er", SAND-74-5379 (CCNF-750108-2),1974.
3.
Ptmeter Safety Stu.+. r, An Assessment of Accident Risks in U.S.
Cctrercial ':uclear Fc.rer Plants, Accence IV:
r rn:rde s
Failures, U.S. ::cclear F.ecuAatory Ccn c.ssicn, FN:3H-1400 (NUEEG-75/014), Octcher 1975.
/
4.
R. M. Gag.e (ed.), Psycholocical Princieles in Svsten Develectent (N.Y.: Holt, Painhart and Pilnston,1962).
5.
L. J. Fogel, Bictech.olecv: Cc= ects and.:nclicatiens (N.J.:
Prentice-Hall, 1963).
6.
W. E. Wcodscn a.d D. W. Ccnover, Hu.r r--~4-eerino Guida fer Ecuitrent Desicners, 2nd ed. (University of California Press, 1966).
c.
- i. u.>tucri, ec at wu. i,
..r.u.m r.m me=t _ ~ ume w w-..c ~
4 Desien, (N.Y.: McGraw-Hill Bcck Co., Inc., 1963).
8.
E. Grandjean, Fittine the Task te % v'-
An Er-encnic Arc-cach (Icnden: Taylcr anc Franc:.s, Ltd,1969).
9.
R. L. Scott and R. B. Gallaher, Annoated Biblicerach*r cf Safetv-Palated Oc=rre.ces in Nuclear Pcwer Plants as Faccr ec in 1973, OENL-NSIC-ll4, i;cve=er 1974.
10.
R. L. Sectt 'and R. B. Gallaher, Annctated Bibliccrachv cf Safet 7-Related Cc=r ences in Yuclear Pcwer Plr.:s as Pancr cd in 1974, OE:iL-:: SIC-i:2, chy 1974.
11.
"A Su:rarf of P ogrc:n E=hasis for 1975", Electric Pcwer Research Institute, 1975.
12.
A. D. Swain, "Prelininary Hu an Facters Analysis of ica Nuclear Pcwer Plant", Sandia repcrt (unpublished), February 1975.
~
13.
"Rrewns Ferry Puclear Plant Fire," Fearings before the Joint CcrTaittee en Atenic rnercy, Septerter 16, 1975, pp. 25-26.
- 14. ikem'endations Pelated to Prcms Fer: r rire, Recert by Scecial Pcvicw Gwuo, :70Fm-0050 (Eccruary 197ti).
865.
,,pHd r
Ah ACIOTTP 1 SCOPE OF T PX ANALYSES CF CCEL Mt DISPIAYS.200 CFEPA'ICR PERFCP?REE PACKGRCC"D Study and functicnal analyses are needed of the h=an engineering featres and related accassibility to centrol devices in centrol centers of n"Me-pcsw plcnts. "cdificatiens to current practices may be needed to adecuately acccrredate the needs u-der the rcre extrena incident and accident c~-d4-4 ~s which require aided hu an operater initiative and perferance. An cbjective and thercugh a craisal ust include the desired cerforance of coeraters under pestulated high stress ccnditiens within a limited ti efrre envirc. rant; atterdant featres of centrol centers that wculd esance hran cerforance under such ccnditiens should be established. Such an assess ent ray be expected to provida recc: rended ratheds of cent:Ol rocm desien and cperati.c procedures tra will enhance sicnificantly the effectiveness of the reac:cr operater during postulated accident conditions.
Studies and f=r-icnal aralyses shall be cenducted to synt'.esize recc endaticns for the desired centrol rec = laycut ard desicn to better ass =e the dested perforance of ".ran cperaters under ccstulated accident ccnditiens in.uclear ccer clants. ' nasc evel=tiene shall i-.clude facters such as entrl W layout, data accuisiticn, irderaticn display, cperating pr:cedures, and hean cperater perferance requirmts under corditions of high stress.
Specific iters to be included in these evaluatiens are indicated belcw.
- 1. An assesr=nt of cent:01 beard laycut of the centrols for redu-dan:
engineered safety syste s.
With cr withcut canel craphics, thase controls may be lccated as discrete side-by-side fully secarated display ar-ays, er as redu. dant suceri. osed arrays, or as widely separated discrete centrol panel arays.
- 2. An assessrert and trade-off of using single er dual p= ose displays s
ard centrols for plant equirment used for both neral cperatien a-d during pcstulated acc W ea.
- 3. An assess ent of the c= rent t end frc= dere ccnventicnal infc:-atica display a-d full size centrols ccrnared to c. ater centrol a-d Cathode Pay Tabe display a.d recall technicues fer ecstulated accident ccrditions.
- 4. An evaluation of the desirability cf using selected space, size, array, colcr, and other physical feat =cs cf centrols for use under ecstulated plant upsets and accidents.
\\. hm o@
gO
< 6.. ' U9.
(
AN En1 An assess,ent ot-the F.ca #ac *- "-
recuircr.ents under pcstulat$d $c$s '.M~ic' ~ cc ~1..cns: a.d cs*E -is" 5.
M~ OC#CS tP~ 14te14k. w.. of h= an ene.e
,,,gm..smo, g ccn iticrq censiderir:g selected ambigucus dish 5av~s~o5 in 2e re*M.s of systems analvses to be used in evaluating control.m E
design and a-e a,. 5.an -actors shall te selected Wl'+5 +"x.e acpreval of ccgnl:: ant r,C staff.
h O 0$ $$0dNA m f m
865 180 9
6 9
O[
i i;ABLE 1 (REF. 9)
~
Classification of 3RR problems by cause, i
deficiency, and tine of occurrence 6p l
f
.o Percent Nunber l
Cause Maintenance
.33.7 119 Design 23.8 84 Operator error 12.2 43 Ad=inistrative error 11.0 39 Installatica error 10.8 38 Fabrication error 5.9 21 Constructica error 2.5 9
Deficiency Switch set-point drift 31.8 88 Lea.
15.5 -
43 l
dures 13.7 38 Vloratica 5.1 14 Fish kill 4.7 13
- Radioactivity release 4.3 12 Welds 3.6 10 Dirt 2.9 5
License violatien 2.5 7
Debris 2.2 6
Corrosion 2.2 6
Lubrica: ion 2.2 6
Personnel expcsure 2.2 6
Instru=ent calibratica 1.8 5
Lightning
.1.4 4
Quality assurance 1.1 3
Co==unicacica 0.7 2
Fuel densification 0.7 2
Water ha==er 0.4 1
Loose fitting 0.4 1*
Stress corrosien 0.4 1
Storn 0.4 1
Ti=e of occurrence Testing 47.3 247 Operation 38.9 203 Inspection 10.5 55 Construction 1.9 10 Preoperational 1.4 7
4
d e
g TABLE 2 (Ref. 9) d o
f Classification of PWR probic=s by cause, 7 deficiency, and ti=e of occurrence g
v Per:ent.
- =b ar I
Cause 1
[
Design error 22.6 77 i
Ad=inistrative error 19.4 66 Operator error 17.1, 58 Fabrication error 15.6 53 Maintenant2 error 13.2 45 Installation error 7.6 26 Construction error 4.4 15 Deficiency Welds 13.5 28 Quality assurance I3.0 27 l
Procedures 13.0 27 Leak 10.6 22 i
l Personnel egosures 7.2 15 l
Radioactivity release 6.7 14 Fuel densificarica 6.3 13 Vibration 6.3 13 i
License violation 4.8 10 3
Corrosion 4.8 10 Weather 3.4 7
Fire 2.9 6
Stress corrosion crac'dag 1.4 3
Lubrication
- 1. 4.
3
?
Set-point drift 1.4 3
j Storage 1.0 2
3 Co==unication 0.5 1
f
. Geologic fault 0.5 1
}
Dirt 0.5 1
,}
. Water ha==er 0.5 1
E2rthquake 0.5 1
j
's Ti=e of occurrence
- 1 s
4 5
Operation 39.2 161 Testing 23.9 98 l}
Construction 17.6 72
,[
/
Inspection 12.9 53 Preoperational 6.3 26 1
'4
)
865 J82.
0A wg ff TABLE 3 (Ref.10) d)
Classificacica of 3*'R proble=s by cause.
deficiency, and rir, of eccurren::
[
Percent Nu=ber Cause Design error 23 Maintenance error 154 21 Equip =ene failure 143 16 Operator error 105 13 Installation error 85 10 Ad=inistrative error 65 8
Fabr,1 cation error 54 8
Conscruction error 53 1
Weather 8
<1 Lightning 6
<1 Tornado 2
<1 1
Deficiency Set-point drift 26 Leak 142 15 oree d---*
S1 9
Vibration 48 7
5 bris 36 5
Failure to :es 26 4
Radioac:1vi:y release 24 4
Calibra:ica err :
21 4
Welds 21 4
Loose screvs 21 3
16 Corrosion 3
16 Coolant ene=
'y Moisture 2,
13 2
i Lubricacion 10 2
Stress corrosien 10 1
Erosion 8
1
'{
Failure to follev procedures 8
t 1
Water ha==er 7
1 Quality assuranca 7
1 Detonacion 5
is 4
Fatigue
<1 Personnel exposure 3
<1 Ther=al pollutien 3
<1 Loose connec:1cn 3
<1 Radiacion da= age 2
<1 Fire 1
~<1 Ti=e of occurrence Startuo testing 31 284 Tescing af:er going cc=:crcial 30 Operation 279 26 Inspection 246 10
[, C
} }
Construc:Lon 97 2
Prooperational 17
- - ~
1 12
O O
uGLE 4 (Ref.10)
Classificatien of P'iR probic=s by cause.
deficiec.cy, and tacc of occurrence Percent Nu=b e r Cause Design error 21 112 Operator error 18, 96 h intenance error 17 87 Ac.cinis trative error 13 66 Equip =ent failure 10
- 53 Installation error 8
44 Fabrication error 7
37 Construction error 4
21 Weather 2
10 Deficiency Leak 15 69 Procedures 13 60 Radioactivity reJesse 10 45 Welds 8
38 j
Se t-point drif t 7
34 Coolant :.5 e=1s t ej 7
32 Vib ration 5
25 i
Corrosion Deb ris
?
19 i
a la
@c.lity ass ri.-.
3 16 Fish kill 2
13 Failure to tes t 2
11 Calibratien error 2
10 Personnel exposure 2
9 Loose screws 2
d Fire Lub rication 8
2 8
Water ha==er 2
7 Fatigue 1
6 Jitation 1
6 Erosion 1
5 Stress corrosion 1
- 4 I
Mose consecci;n 1
4 Security
<1 2
Ceological f ault.
<1 2
l Explosion
<1 1
Ic: proper storage
<1 1
ebis tu re
<1 1
Tim. of occurrence i
ope ration 41 330 Startup tasting 22 177 testing af ter going ce=ercial 20 157 Construction 8
67 Inspection 6
46 hvore rational 3
21
)
865 184-f
M5-
' N/An,
r.
fr c -
/,9
%g i//h
. ;.g TABLE 5 (Ref. 6) c'-
M AN VS. MACIGNE
//
9 MAN EIC(L$ IN MACHINE $ (IC{L IN De'ettien of s erves a forms of v er y f ear Men sering (bort men end mesh nes) energy levels s
Senseliesty to en estremely w de versery og Pe.fer.o.ng rewhee, repehtn g, or very pre.
shaw31 sise operariens Forceiviag pe'veras ead making generei.as.
Respene.ag very eviekty to conteel s. gaels e s ebews th.m l
Detecting signals ja high no se le,ess
> go,,,iag greet f orte, smeethly and with precessa Ability to sfere large emewats of soforme.
Storsag and reselling large emewats of in.
eien for long periees - ead reces8.ag rete.
f ormetesa ia short time.pereses vent facts es opprope,ere momenes a
Ability to esercise,vegmeae onore e,ea's Performe*g comples and rep.d compwtetion
~
eme.a. n.
a..,,,,,
improvising end edeoeeng Aes.ble preced.
Senuriedy te st.mw33 bevead 'he rea ye of ures h, men sea s.es,.,y (,a#,e,e d.
ree's = eves, etc.)
Abelity to react 'e wasspected lew.pras.
De. g meay eiMereat th ags et one time abil.ty evenes Appfying originelary in selv aq problemse Dedochee processes Le ellernese solve.eas Ability to pre ## from esperieate end efter laseas.tieery to estreneews festers teerse of str.en Ability to perf orm Aae =ea.swier;ea. e sse.
A b.3s ey 'o reseet esoter.eas very resedly, seesly where m.sa.i g a m ea s esseers va.
c ent.a vow sly, sad pretesety the some -ev especteely e,e, e sea, pe,,ed Abitify to gene.aw, se perform eveg on n Coereesag in ea. reamenes which ere host ie e
e'etiesdod se maa or hersad *w ea es erence s
4 t
Abel.sy se reesen indver..ety I
t f
s s-r e
4)
Oo t
TABLE 6 (Ref. 6)
</"
LEVELS OF ANALYSIS AND 50ME RELATED TECHNIQL'ES g
t.fvfL CF AN A'Y5tS l P'.'tPC S E l AP*:tCal;f f!Cw'4tC'.'t System to determine edect.eaess of e, stem in Ceeret eas researca performing e esec Aee m.ssien morn gi. t.
e Subsystem to determiae best -ey of meenag e 5,trem anelvs s specified regwirement ei the misslea ta'eg'etrea metr's Pwesteen to ee'ermeae best combiaatica of com-mea mechine s, stem poacats regwered to mese we seiosystem o n e, v i s s ?
Fwa< tion eae#,ses e.g f.
w........ m..
.es.
.e.
.a....
- v. 6 c.......
g
!,ca. hit:ee.. a.
I
....a
- o.c.
- .e .....'.t1 tiens Logic meee sO U laformation sneeryi
$wbfesk te determeae best metaed et w ilisrag Coers'er lose e
men's cosenst; ties to perform the es.
easeysis-signed tasks,
Coe*etes seaweace esegrams Deosien t%eecy talerme* enJew enefyses e
Element to determine best metMed of utilising f:me-eae.metion men's cepenslities to perform esmagaed e a siese s su b rask s Efemented task enelyses 8=N I
J 5
1, k
h
-y w'0 O['o[ }.
sv57t.u ecvrtcevcNr 7,
$!J!r***
- e ' 1. s*;* $.: ::..* s
/
L c:
- Adva*Ced C s.;** 0; *:
tag rets;9 I
Alisteh "t 8' 8.FCf. *s s
to eas ase a:- es d
@C!=**4 5 l
- !i CE110N STA0Z Task c sce..-.:s Tass a*a:,s.s e
Job ces.gm i
Joba.ds 5ete::.ca OEVELCP'.'ENT T,4.a. aq Tra.a.ag c4v.:es STACE Perfar aace eas.<es Ceestetec cs :saeat s - - - - - - - - - - -- - Te:
S na.a.ag TESTING STA0E System training I
System evaluaties CPtR A TicNAt. sTACE t'
Figure I.
The pr=:e ures used i9 *e cevere: eat of h.
aa ;.*-
ponents of syste
- s. and f.a e r :-cer of.nir:at.:9 in re:4.on 'e ' e v. s -
system devele: ment. Pr=:e:wres used f:r e: i:r emi cevei: me"t s'e "~'
shown in dated.
(Ret.
. -- )
I s
I t e
.'(
I e
i t
e 865 187
II I
I l
Il o
I I J
s 9
5
,f'o 1
=
Ir 5
s y
a e
o a
I
- l.
~
, d),
o
~
m o
t' a
,;/
20~e Ii o
or 1
ac V
~
99 Il y
l I '.
3 ii
-=
)
35 e
fl l
> r 4
l
~-
Ii 0-
=*
o-gIl
-o
=8
=2 il, s%
_e,=
a.
a t,
o.
,t g g,g W-2 O-o.,.
eg:
2-3g 0
l l 2
- o~
0=
i, 1 I
.~~
41 5
os w
ii x-
~
ec a
'll i
==
l E
I I I
%s
{M l
l I
t g
a l
ii s
Ig 8
I s
m i i 8 II pe r
.> # c 8
%+
8-I-
u 18 I
.=
26
=
23
- J l og
. ggt gg
,_J j,
553
-3 9w
-=
. z.
O 4,
T
-- =a,==
\\
, ut,
8-7..
s I
i
.a.,
-=9
- y= =9
0 C.9<.<-.
.I d "e
s' w
" e*,'. # '
=
a it ez 2
=Q s-
,:.y: =9=s0~:=
fa 6
5 =a -
z.
o og i
<q2
- =5 s
)" " ~ 2 9<e
=u-1
~=
~
Zs e=
wg t
os
-g<
s-m=
<_3 t-e.
==
= =s 2 o.~,.
s
_z-
=-
em I
l
.= s, -
v:
-2
_3
-=
l e=<
1 i
o,
.m T-
,=
o-e-
.i os I
N a-3a
=.. = <
- Q V
C l
- .a Z
I
=
9 3=
0 I
$ :2 I
o 21,e=-
-=
<v<
~
l c-v2 = =
s=rsi :-
.i i
- =
I s
i;
==
I Ii o1
!I s
I i i ro
,I o=
..f
}
4 86b 1.88
ni,,
yp>
bo c'
S 60 i
c1]g '
E b-C i
1 E
E
./
3 C
c.
Low VetY Low Moderate Stress Level very Hign U
"4 HYPethe:L:a1 7elant.gg$r. u-e:seen perfg=,ng,
~~
-~~
and Stress (ref. 1)
~
865 189
.