STF38 Reliability Data For Control and Safety Systems 1998
STF38 Reliability Data For Control and Safety Systems 1998
STF38 Reliability Data For Control and Safety Systems 1998
@$t'LiEF
;'ifiV}f
,'l';-15
KEMIRA
KIRJASTO
SINTEF REPORT
)
@s[Nr,,im
SINTEF lndustrial Management
Safety and ReliabilitY
Systems'
Reliability Data for Control and Safety
L998
Edition.
Address:
N-7034Trondhem'
NORWAY
Strindveien 4
Tefephone: +47 73 59 27 56
+47 73 59 28 96
Latin;
fa:
Vatn
Geir Klingenberg Hansen and Jm
srGN.).
It.
Lk^1
t999-01-l I
in this report' D
control and.safety systems are provided
eliability data estimates for components of
(etectronic.ar"
n::"-T:l Data dossiers
.nuor rogi.
r both fietd devices (senso;;
and expert judgements' The level
BSTBACT
t#;f"rm;t
reliabilitydataestimatesareessentiallybasedonthepreviouslyrecommendeddataforusewith
IV data'
method, updated with OREDA Phe
Also,amethodforobtainingapplication^specificreliabilitYdataestimatesisgiven.Asacase'
*",irtJ t
"ppfied
iltrol
Systems'
Feliability Data for Control and Safety
1998
Editon
PREFACE
ThePDsForumisaforumofoilcomparries,vendorsandlesearcherswithaspecialintefestln
if"il*
ror inrormatiJi-'"J*a"e
ft tp://www'sintef 'no/sipaa/prosjekt/pds-forum'html
TheresultsinthePlesenlreportistoagreatextendtasedonworkSlNlEFcarriedoutonrequest
Fe40s6 - Reliabilitv Data for
ff;siEf ;"I;':'sinzs
from Norsk Hydro in 1ee5 ffi"]i,
Hydro ailowed using
We appreciate ttfttttt that Norsk
Control and Safety Systems" t13l'
report'
these '95 results in the present
t,,t-.
tslindman/sipaa/prosjektioreda'/
""tri.nloni
Trondheim, 1999-01-1 I
.
o
e
.
o
.
.
BP Norge
ElfPetroleumNorgeAJS
Norsk HYdro ASA
Norway
Phillips Petroleum Company
SagaPetroleumASA
A"/S
Norske Shell
(Statoil) a's'
Den norske stats oljeselskap
Control and Safety Systems Vendors
. ABB Industi
o
o
.
o
o
o
.
.
.
Auronica
BaileY Norge
Boo Instrument AS
HoneYwell
ICS GrouP
Kongsberg Sirnrad
Norfass (Yokogawa)
SAASASA
Siemens
Consultnts
Engineering ComPanies nd
o
.
.
o
.
.
Aker Engineertng
Det Norske Veritas
Dovre Safetec AS
Kvrner Oil and Gas A'S
NORSOC
Umoe Olje og Gass
Norway
Phillips euoleum ComPanY
r,uo oljeselskap (Statoil) a's'
bln t*.rc
Sas Petloleum
ASA
ii"""".
TOTAL S.A.
B V'
Exploration and Production
Edition.
TABLE OF CONTENTS
LIST OF TABLF,S
LIST OF FIGURF,S
t.
INTRoDUcrIoN......""'
Rrsul,rSutt1t14RY""""""""'
::::
Hil:H*ir*i:'ffi
"""""""
Z.
Data
Summury Table of PDS Input
I
' """' rr
r+
I
"""""""' 17
"""""""""'17
"""""" 18
"""""" 18
"""""""'23
Tprobabilities""
2.3.2 Cotterages """"""""""'
2.3.3 P-factors
2.4 FufherVork
:' :::
a^1
2.4.1 Variability of the ?IF probability"""'-':"""""""""""1"":"""'
2.3.1
'
|""'T3
a
',
2.4.2Distinguon*.*.*u".*i'*i'"*anellofsduringtesttng......'''
3.
NIETHoD
ANIETHoDFoROBTAININGAPP"'"o",o*,""orrcTIFrnosILITIES.......'.'..''............25
A
lll.trnlllntion......'.......''...........'.
3.i
Edition.
it
2. RnsulrSulrulnY
2.1 Parameter Definitions
each component:
The following parameters are quantified for
-n
",=Totalcriticalfailurerateofthecomponent.Rateoffailuresthatwillcauseeithertripor
causing such
(unless cletected and prevented from
unavailability
failure).
";*#.r, ".ii*
.=RateoffailurescausingFail-To.operate(,FTo)failures,ndetectablebyautomaticself-
test.The,FlofailurescontributetotheCriticalSafetyUnavailability(csu)ofthe
comPonenlsYstem'
* \,\,,.
li,=RateofSpuriousoperaon(So)failures,undetectablebyautomaticself-test.Therateof
Spuriousoperation(So)failuresofacomponentcontributestotheSlRofthesystem
1a.p"nO"ntofoptrtionpbllosophy)' l\+'"
ndet
* 2i10"
Total rate of rdetectable failures' i'e' /ffi?t
lFTO
/het
Rate
lso
'"er
of failures
test.
detectable
causing FaiJ-To'Operate (-FIO) failures'
by automatic self-
t\\
=RateofspuriousOperation(So)failures,detectablebyautomaticself-test'Theeffectof
the operation philosophy'
these failures on tne spuriou
trip
W+ ftf'
h",
Totalrateofdetectablefailures,i'e'
TFTO
/brit
function
the component' Causes loss of safety
Total rate of critical FTO failures of
y* =
regularity
component. causes loss of production
Total rate of critical so failures of the
failure)' i'"'
m''
(unlessdetectedandpreventedfromcausingcriticalfaitur,i.e.,i,fl+,{f0"..
,no--Lw|^F[ll=Coverageoftheautomaticself-test+controlloomoperatoronFTo
-
,So=1r.t^n=Coverageoftheautomaticself-test+controlroomoperatolonSofailures.
nF-Theprobabilitythatacomponentwhichhasjustbeenfunctionallytestedwillfailon
eman (applies for FTO failures only)'
xr
: ,\
...:
'. .
Phase 4
components'
r,^_^r^^1
Risk Reduction
T'e
,fd;;;;-"-;*i,
targetfailureforasaferyfunction.allocatedtoanEiPEsafery"-relateds]_Stem
10.000 to 100.000
1000 to 10.000
100 to 1000
t0 to
100
Phase 5
Safeqv Requirements Allocaon
of a EUC operator
w't
for
take into account the requirements
t'e
Le'el
as meeting the required Safetv Integritv
Tlre functioning of the sIS needs to be verified
forming the qYstem architecture'
after the external risk
are then based on the remaining risk
In this gride, the risk assessmentand sIL determination
in the figure
box
have been implemented' i'' ttre leftmost
reduction facilities and oter safetv related s-vstems
to the three safegv s-vstems'
concept of safetv requirement allocation
The fo'owing figure illustrates the generar
t1 of23
t2
STNTEF
Syslems'
Belability Data t^- Contro and Safety
1998
Edtion.
}so
lFTO
/tndr
Detectable
so
2FrO
triet
7so
tudt
2FTO
'nr
'"d
Sum
Thus,notethatifanimperfectsrlngprinciple^isadoptedforthefunctional.testing,thiswill
by introducing
a procss switch is nar tested
if
Fail to operate
l,o",
2
'"det
A^,
rlf
niun.",
conribure to rhe IIF prouuffi.-nis no perfect
i*prirg u "icated test signal, there
change in rhe pro""r. itr"tt u'oir,". "i""ty
"""t a blocking of the sensing line'
functional testing, ttre test wil'not
(cs are
to the cridcal safe{ unavailabiliw
phvsical
are
faIures'
rate
S"rt,rtil.* t"n*"q io tt" ftut" to an operational state' The
illustrated in Figure 2.
,o
r"p;
,o*. t ind.ot
ComDonents with physical
bv tunctional iesting' on the other
contfiburion to csu ao*
"li";"
No repair is required but
nrs.
-iJtir"tutry *o*ol
and
The contributions of the T/F probability
x.-o
I'
fJ;;; ;q**
pri*i"i;.il,.i ,u';d
;
hand, failures contributing
,"*
the
suchfailureswi]]occurrepeatedlyifthesamescenariorepeatsitself,unless.modificationsare
ir'utto*t constant' independent of
,iiffi"n;:Ji;;
FTO
h.
Detected by automatic self-test, or by
operator/maintenance personnel
(inespective of funcrional testing).
SO
hd"t
{ro
'!undet
"t
nSo
4undet
Coveragec=
10'2
Revealed n
functional lesl, lrl2
(physical failures)
Unrevealed in
funclonal test, TIF
(luncional lailures)
103
10{
TTT
lool*,
t
.
E}
softwae
degreeofdiscrimination
'Wrong
Location
.
'
'
.
test if
forget to test
wong calibration
damage detector
leave in by-pass
CoveraRe
Thecoverageisthefractionofthecritica]failures,whichisdetectedbytheautomaticself-testorby
t;ure that in s91e way is detected in
include as part of the ":Yiq:.;
rn operaror. Thus, we
r"nro, t..g. t *r*itt"rj ti,i
betwien functional tests. Analo!
'
t4
@s5|LiiulllF
2.2
Safety Systems
Relabitily Dala for Conlrol and
1998 Edition
Itr
database, which
in the
;;;;;
th;
1t
t
where
tlAoD +
Notethatintheg5etlition,thedatawerepresente-in.asliehtlydifferentway.Insteadofusinga
is in the present repofl
types of frurel tn coverage
nfCj
comrnon coverage for both iO
split into its FTO -a so purt ]rJ"i."iin
r"
useo repeateay
"r"oiUf"
*itf,
Itr
Itr
rv
io;pd;;;
= 0) tlri.
database
(r = 6.
,r"r"
reriabiriry data
still apply. (Note that if
updare the
i.;;;;;';J".,
,ir"pi
Th" covemge updates are taken as a weighted average between the previous estimates and the
observed coverage in the OREDA phase IV databe.
The previous stimates are given double
weight since they include expert judgements arid the dat
material is s"oc",
with the
OREDA
Phase
releai;;;;;;-t'n"
previous
il sources for the uario,rs components'-The
database'
OREDA
the
than
o'ht'
'o*t"t
estimates in the ss .auon *'ie;; ;;;;;;xt*bi9,:"
tutt dutu to*tts are given below'
failure
the
all
of
;;i;v.J;w
th"
data dossiers give informatirr;
Hll;:;;;'
';;;1.;r'r,
IV
dara.
"ven
installation'".i"
'i
l?e1
frqT"iiti"t
T8: '
!'"j''11-%:**r'
rt9ry-
(ref
l3t)' 2nd
there are
threeversionsoftheOREDAdatabase,ofwhichthelatestversion.isthemaindata
sourceinthisrepoft,denotedtheoneplpr'*"d"tab"s"(ref./15/).Thedatain
was collected in 1993-96'
te Onep pnle fV database
on Fire
Oseberg C 'Experience Dat
filters
the later versions' Thus new
totar
afier
the finishing of the 95 edition . lvhen this is-the ce
the previous estimates are updated
sequentially
with the complete OREDA Phase
approach described above.
reliabiLiry
WheretheoREDAPhelllorlVdatabasedoesnotcontaindata,ordataisscace,thefailurerate
in *'"t"'i*: *dl:lTl:*liduat
irn.".
SomefiltersusedinthepreviousstudywithearlierversionsoftheOREDAsoftweaenot
have to be set'
-l]i-
dossiers'
tisting - e included in the data
'nw
l)
anil
;;":ri
';:;u;rt
Publ.war:
";:::;:::"'
if:"tJ;i::ents
Gas
Detecton' ref'
/4/
Porsgnrnn' Norway
Norsk Hydro' Research Centre'
1994
IR name
data on catatvtic gas detectors'
Sea'
North
the
in
"-ry.-".:ifrom the Oseberg C patform
detectors an smoke detectors
ref' /5/
Process Safety Systems'
rerd
Methoil for
WLCAN - A Vulnerability Calculation
Lars Bodsbere
Author:
Norway
publisher: Nor*"giirirtituteofTechnology,Trondheim,
Publ.Year: 1993
detectors
railure data on fire and sas
';':r:;i::"?'' i#l;ffiT:serration incrudes experience
jl,;:;,gl*:m:,*:lJJff
rrom"J;,il;;;iglrlr:^.:
respect to ra
very comprehensive with
t" rt"i't
,nu,,n"
"iiit
in the oREDA
Phase
III
data'
l1"i:"1:
Systems
Reliability Data for Control and Salety
l6
1998
,@stltllllEm
NPRD-9L: Nonelectronic parts Reliability Data 1991, ref. /9/
Authors: william Denson, Greg chandler, william crowelr and
Rick wanner
Reliability Analysis Center, Rome, New york, USA
Publisher:
year:
PubI.
1991
"Reliability hediction
of
nor.o";.J;t '--HDBK
.^il^l:r:^**.es,
number of
chaacteristics.
LarsBodsberg
SINTEF Safety and Reliability, Tondheim, Norway
on:
Descriprton:
11'l'r"r'-
2.3.1
rrFprobabilities
/g/
Authos:
Publisher:
PubI.year:
Data based
;i;;dbelow,
.Process
1989
The report Presents field data and guide figures for prediction
of reliability of
computer-based process safety systems. Data is
based n eview of oil comiaay
data files, workshop with technical experts, interviews
with technical
questionnaires.
tffinrra
Descripton:
'
,".\-;1\, ",.,;..,,,
{.,.,:;r)
..,,;.-,,r."
lIF'
Fire detectors
Itisassumedthata.detectorwiththe,,right,'detectiorrP'il"'Pl:is.applied(Smoke
d: *-i::nt^where flame ftres
fires t"
detectors are applied where smoke
gives a very low
"*p"tt"J*a
s a.possibility tiat a fue may occur which
e expected') Even so' there
Butterworth-HeinemannLtd.,Oxford,Eneland
orobabilityofdetectionbythedetectornuro"".i*.bo"tothisfactanintervalis
the fire, essentially
";:Th; ir uu. *u1n ;dt,i"; to the size of
provided for
'
.n
-
evaluated
David J. Smith
Methods
Engineers" (ref. lZt) have a specific chaptr and an appendix on-iailue,rate for
data:
The data presented are mainly compiled from varius sources, such as
MILHDBK-217, NpRD-r985 (i.e. rhe 85 vrsion of MRD-91) an opGoe
Handbook
1984. The failure rate data presented in the textbook is an extract.from
the database
FARADIP.THREE.
'
NotethatanewexpertjudgemensessionlgasperformeddurngthelggSstudy,givingTIF
the
*itt r"sp""i to detectoitype S point or line)'
values for g* a.t"ior. dfferentiated
probability for IR
TIF
ihe
inflo"n."
size of the leakage, and other .onaition*p"ja
was not
a, 1at-probability for catalic gas detectors
detectors. s". cri"pto i, "t"1..
relevant'
less
and
old
* tfo' t"n"ology is considered to be
Data based
,:
'rt--tt-o''-t
Gas detectors
Author:
Publisher:
Publ. year:
tr-i:-1.1.:l),,:r,
"o*pl"t"'"tng,
;1 Y\r'rr'i--! ")\r.i
ilr';"'"'
i-\lo"-*'
/6/
Publisher:
year:
i;;{
ability
t''''-'""
Processtmdre"rs
;p"*;
Somecomments'basedontheexpertjudgementsessionperfolle]:nngthe^previousandpresent
and coverage'
in partiuhr onihe given values for l/F
u1., include
.rti*ut", of
Data
of the column
input data to pDS analysis. The definition
Table 24summaise the recommended
2'1
Chapter
given in
fr*aingr r.tut", to the parameter definitions
2r7,
part
failue
failures, rotal operaring.toun, an detailed part
11
Data based
Edtion.
*t li""t"t
ttt"-tJ*
"^.h
generally
depend on tne tocaor/envionmenr "r
detecto
19:t -pt:^l^"jtilt"ctors
quarter). n",
detectors are
Flame
grelter'
sigrrificantly
is
the value
serve as ,".onu iuri"., and
but oil fues in process
"
ir J;"n4_t""imalted ,IF = 3'104),
reliabte untess
as 0.5, could apply'
*"i"""' '*"t"
"t"
"f
*
d*"1;il;ir*"r.",
will
e
?Lprouuuiliry
as
high
systems
, - ^^ ^^+",'a .*^'q For dedic
^---"'T;;rIF
for the rogics is.essent4lt *:j.','J"::il""::rff:.t"#fiithlTH
Fo' standard
I
:*i,':"n::fff J l"ilii r'Jffi *md;;;,r,**" ""o's
PLC
/F
5{0-
appxes'
,,;*t},.-
lo
Safety Systems'
Reiability Data for Conlrol and
@)stlNTEF
18
1998
Edirion.
murtipricitv,gt-:'b:i:.^1":li:i'liltih::IJJJ;5':;:
0
H+
Valves
The zIF probabiliry for ESVs witl depend on the type
of functional resring. If the ESV is
shut in completely and pressure teste, iryF
ir al*"* because of rhe
= 10-6'ithis
possibility of human elrors' e'g. related to bypass and """
improper testing). If the ,,functional
testing"just involves a check that the valve moves
lstarts closng on dman, the value 10
r
is suggested. This.?IF val,re also applies ioi
ol valves. AII these values include the
"ont
pilot valve. The major contibution to the llF probabiJity
for psVs is wrong set point due
to enor of the maintenance crew, and the same TIF vaJue
used for switches is suggested
(sensing line nor included).
;;;
probabilitY
failed is 0.10'
single SimultanousY
failure lalure ol A and B
Unit A
2.3.2 Coverages
Feliability
btk diagrm ot
B single
lailure
Senson
Line testing gives a coverage of 20vo for switches, conventional
transmjtters and ESD push
In addition operato detect a significant
of p.o"".r-tanimitter failures
(transmitter being stuck), giving a total coverage
foi transrnitters which is significantly
higher. For gas detectors also drift are detected (low alarm)
an trris *-uy
trips to be
prevented. The given covefage for smoke detecrors
"uur"
applies for analog
buttons'
p*
sensors.
Control logic
For bus coupler and communication unit 1007o of rip tailures
actually gives trip. Further, it
is estimated that 957o of loss of safety failures e detected,
and a Fr iailure is prevented.
Valves
23.3
p-factors
the
_r.1,r,rn flq\a
In PDS dependent failures ae accounted for by introdu cing a multiplicity ttisibution. The
m-ultiplicity distribution specifes the probability that - given that a failure has
ccurred - exactly ft
of the n redundanr modules fail. Here, & equals r,2, ... , n. The probability of k
modures failing
simultaneously is denoted
p.
components
distribution for iluplicated
Figure 3 Example of multiplicity
Table6plesentsrecommendedp.factordistributionsadoptedfrom/11/.Thedistributionsare
pr"il"i
r
r
r
dependency
,tte following degrees of
Low
Medium
High
ComPlete
Table5pfesentsguidelinesforselectingappropriatedegreeofdependency(adoptedfrom/11.
sullilem
20
)@
1ee8
Gomponent
Co verage
-i
;Pf{ 106
hs
cFrQ
.
Process Switch,
FlQ
"ndd;:'
':
.t .: 'i,
'I-.r per
:..
.so
1SO
Ln
Inpffice
,,
Edtlon.
21
10
.i lrs
So
Iff" || ^'nr
t
lL'*
Control logic units
3.4
90Vo
20Vo
2.1
0.2
0.9
l.lo3 - 5.10r
2)
l.J
9Vo
20Vo
1.6
0.1
0.4
3'104 - 5.104
3)
Level (displace)
Tansmitter
3.1
90Vo
50Vo
0.9
0.t
0.8
3.104
5.104
3)
Temperatue
Transmitter
I .8
60Vo
60Vo
0.6
0.3
0.4
3.104 - 5.104
3)
60Vo
5jVo
0.7
0.6
1.1
3.i0" - 5.104 3)
2.3
60Vo
4OVo
0.6
0.4
J .6
80Vo
7Vo
0.7
0.1
Conventional l)
5.10-s - 5.104
2)
Pressure
Tansmitte
Field
_
Flow
Transmitte
Gas
detector,
catalytic
Gas detector IR
point
Gas detector IR
line
t1 .0
4)
3.104 - 0.1
6.10-3 _
bus
couPler
l)
t)
4,8)
l.l0_3
,E
80Vo
7jVo
11.0
0;l
0.1
40Vo
507o
0.5
0.8
1.2
6.10-2 _ 7.70-2
j IilO,.,
"ndr
so'-
per 106
Component
3.6
COYeraBe
'hrs
4.8)
crro..l
,,ffi'
cso
--l
hrs
TU'
rff., I rf...
Outpul devices
Smoke
detector
lo-3 - o.o5
5)
ESV
Heat
detecto
2.4
50Vo
5OVo
0.6
0.5
1.3
0.05 - 0.5
6)
8.2
detector
5OVo
5OVo
1.0
2.1
2.1
3.10* - 0.5
7)
Push
button
1.0
20Vo
2OVo
0.3
0.2
0.6
l0-5
30To
1.1
0.8
0.5
r)
1O6 _ 10-s
1.6
OVo
Vo
+-3
1.3
0.3
lo{-105r)
20Vo
3O7o
0.7
4 .2
t.8
7.6
604o
'107o
17.8
2.8
0.1
10-s
,R
6O1o
'7j%o
3.0
0-8
u-
t0-
0Vo
5 .0
1.0
o.z2)
10-3
valve+actuator)
Pilot valve
ESD
OVo
X-Mas
Other ESV lmain
Flame
I .6
Control
valve,
small
)
2)
3)
4)
6)
1)
8)
Control
val-ve,
lge
Pressure relief
valve, PSV
1.2
07o
respectively
testing'
For complete and incomPlete functional
lead to system [aP
ttote tna tnp of fSV does not necessarily
Safety Syslems
Reliability Data for Conlrol and
,@ SINTEF
22
1998
Edition.
23
Fire/gas
detector
p-factol
te'rm
Component'.
tlistributions
Table 6 Recommended p-factor
:disfribution
=hl
Comment
r.'t.r.,..
d"pendence
-"er..
"f I Irigh
ruium
ut devices
mo
2: Medium
.so
dependence
contribute to CCFs
Ttr
<0.2
3: High
dependence
TIF
4: Complete
>0.2
dependence
CCFs
of
0.9800
0.0180
0.0015
signed to handle
Pressure switch
Pressure
atl
all
hansmitter
2: Medium
dependence
1:
[w
dependence
for
transmitters
as
compared to srilitches
Field bus
transmitters
all
1:
Low
dependence
Boththeg5editionandthepresentstudyi]lustates,thatfurtherworkshouldbecarriedoutonfailufe
validity of reliabiliry analyses:
io inir".rJ tn" cr"iility and
data definitions/cf*rifr"ution
probability
2.4.1 Variability of the TIF
PLC
all
"iO
2: Medium
dependence
Forseveralcomponents(e.g.sensors)thereisobviouslyawiderarrgeofTlFvaluesthatmayapply'
such as
depending on various factors
conibute.
Ouut devices/Valves
Pilot valves on
aIl
2: Medium
dependence
all
1:
same valve
Pilot valves on
different valves
ESV
Low
dependence
all
1:
Low
dependence
Lower fraction
of
all
l: Low
dependence
r)
;;;;s"(e'!'anaiogue/diqil4'Pginqn'].-,^^,,-line)
impulse
svstem boundary it'g' *ittt/*itttout
u*ount of self{esVmonitoring
different ESVs
Same design, medium ard maintenance conhibute
Couplers
Anefforthasbeenmadetomeetthischallenge,b.ytyfaronlyforgasdetectofs.However,itisan
*":"t'+;;"':"t:::t:i*l'r":*;mt"?ii:ttr#t'
testing
errors and human errors during
2.42 Distinguish between design
by
ItissuggestedthattheTlFprobabiityshouldberestrictedtoaccountforfac.*:'ll,arepresentfrom
These are failures caused
uuto*utl"f"".,1"J
in-ly
and which are
th-i|1{ errors introduced bv
d".:t:'.t:-t-t';;i;-suggested
be defined as
design enors, e.g' including
"f
iniquate testing) should
(e.g. by;pals ruilu,",
testing
should
upoi
models
crew
u"Jprov"d
the maintenance
U inctue i" ili'-p't"ility'
a separate category of f"ifor"s,--ar;d'no't
testing'
6r fitures inuouced during tunctional
day
l,
;;t.a
".""i';;#
*'" r""r*
"tt'
-J
\g
24
tlNULqf
1998 Edition'
)
The above suggestions will make analyses more credible and
accurate (ptant specifrc), and it will
facilitate the communication.between analysts and
maintenance/operational personnel.
"rr""
ri"
It wili
also
3.
25
pnosnnIr.rrIps
PPLIcMIoN sPEcIFIc TIF
3.1 Introduction
parameters in quantitative dependability
data are used as input
In most RAMS analyses generic
;uu"'ug"
it is theiefore desired to establish
;;;;;;i
"*i;unJ
assessments. These generic
into account' In this report
conditions
to tut'
a method for adjusting th"'"-;;;;g;;alues
'pt"int
future repofts we aim at
In
t^git-iirryrrs.
f",
present
a
merhod
vr'e
"a-unut
"urrJt;r;;
parameters and equipment classes'
i oter
;.:";ffi;;iit""l"gv
by a step by
Firstthemethodisestab]ishedandcalibatedbasedontheresultsfromanexpertseminar.The
N.*t tt" orJoi ift *ttito is described
ir S".,..
*.
:.S.
*urn**i."
main resulrs
is given' see Sections 3'7-3'8'
step procedure, and an example
3.2 ConcePtual aPProach
conceptual model
A.conceptualhierarchicalmodelhasbeenestablishedrelatinginfluencin.gconditionstodirect
u, if*rt ui" irifig*" 4' This
"rJ;-Tf;;"U,liry
zJr.r'* r"tutiu"
direct failure
i,npo,iult t*tig"1 of the various
causes.
Generic baselne
TIF values from
expert
Tminar
\
High
High
DC,,
-V
(
(S)
APplication specific scores
structure
Figure 4 Conceptual hierarchical
-
Thetotall/FprobabilityisthesumofTlF-contributionsfromthefollowingcontributingclasses
GA:
r
.
.
TIF'
Design enors (CCr) giving
glvingTlFz
(CC
Wroig Iocation
Insufficient functional
't po""ao'"
giving
or human errors (CC)
?lF:'
..Behind,,eachcontributingclassasetofdirectfailurecauses(DC)are.defined,forexample
each direct failure cause
t-"sign" The impottun"" of
are
"forset to test" and "*'o'l' ""t"
(wn' ninty the direct failure causes
within a contributing
"r"""i'#"y
"v'eight
26
,@srNTEF
19eB
-l
Edition.
27
Modifications
1,r.;ff::,:.:1t li;flft,l;
l;;
i:th*I
iltillu*;;
a direct
inJluenc.ing on
Notethatadirectfailurecausedoesnotdirecdycorrespondtotheconditionsthatafecontrollable
focuses i.,r'""i,i"ns
by a designer. Therefore *;;;Jt*ically
l"::* 1;Lj;l=*il.:T::"*:,t:?:tl';odi'":;
score w'r be
application
Thereisnostraightforwdmannertoestablishafe]ationbetweenthescore.sandThreTlF.values'
on tt'" following principles:
rt r"iu,ioo p.";*"u *
"
equal TIF,on\f all S=
tti"i t"d;;;;"
3.3 Definitions
The following definitions will be used throughout this presentation:
o A contributing
t TIFshould
1'T1,
1/F,,n3r' if all 'fu = 1,lurthll'---.n,
-' ir' Ji""ia equal tne
*.* f the low ardhighrlF-vaiues'
the
;.11;;'; tqt o flF strould equal Seometr
o
class (CO
to the TIF
probability.
A direct failure cause (DQ is a specific and clearly defined cause within one contributing
class, influencing the IIF probability.
'
c
10
of this principle (TIFnign=
Figure 5 i'ustrates the implications
:-+-
3.4 Method
o
The main idea is to establish rheTIF contribution from each of the contributing classes,
and then
next evaluate the diect causes within each contributing class. The following cntributing
classes
have been defined for gas detectors:
.
.
.
In the expert seminar baseline numerical T/F-values were established for each contributing
class,
CC, i = l;,'.,3. These baseline numercal /F-values represent the anticipated range
for TIF
values for vious conditions on an offshore installation. Notational we
leT TlFto*conesponds to
the "best case" and rlF,s cofiesponds to the "wost ce" for contributing
clasi.
.A set of direct failure causes are defined for each contributing class. For example
for the
contributing classwrong location the following diect failure
u.e,
"ous"i
0.5
Sco
.l+S, /
T, =iwDc, (TIF,,,"
and the total
rrn
IF for contributing
TIF
for
all
)T
(TIF,,
(l)
al-S,
J'
given by:
contnbuting classes is
irq ='oc,fr",""
h*''.'
gives:
all influence conditions
Note that average scores on
(z)
psnmrnm
28
rj--
TIF, = ) JTF,
r-'
Edition.
'i
T.o,ro
29
'
TIF consiilerat"
t"t *
That is, 71Fa is the sum of geometric means for each of the contributing classes.
ouiP.u*"t"rsettings
ioo"i."tioi
6lith
heavy/li ght
"n"itonment2,
weather
@of
r
o
lnsriion
o
o
.
.
r
.
o
r
where
Smail gas leakage, release ate <1
Large gas leakage, release rate 2 | kgls
kgls
air
<tu*ings, taglists'
@odification
@uuitiry
Gie-mandqualitatitelY/
rue.
ouantittively different from
plasuc oag'
emand (e'g., covered by
gas
wfong gas tyPe nd/of
Note that such a scenario consideation is only necessary for contributing class cc2 .,wrong
=
location".
On the expert semina focus was on the qualitative identification of direct failure causes and
influencing conditions. In addition, Z/F-values were stablished for each contributing class for
different detector types and scenarios. Based on the discussion on the expert semina SINTEF has
proposed numerical values fo the "weights" of each diect failure cause, and performed a
grouping of influencing conditions. The members of the "PDS-forum" have had this results for
comments. Table 8 summarises cci, DCs, ICs, wDCs and r/F-values established during the
expert seminar and the post processing of results.
'"i.dEf C"'u"t"or
ffidtit(ti*t
tesrcd'
i@e'
mis-understandings)
(wron g- derecro
6Tvoussnot t"mo"ed
i'p"r*' forgel to remove bypass)
Wpassed
modes ae made
I
No consideration of failure
t T"moerature, pressure, flaring etc'
:i:m::ti;;!lation
gasses
P*ssure' working
accessibilitv
componens
{(P
30
1998
Edtion'
Ventilation
type
Open
Naturally
ventilated aea
Mechanically
ventilated area
Ventilation
intake
Best
0.5
0.9
0.01
se\ffipaiJffi;:i,H:iir'iil::,:'1'i":r',ii';ilYl;
'Worst
{c,
It
0.1
0.1
0.3
5.10-3
5.102
5.10-3
0.1
l0-3
10-2
104
lo'2
104
10r
During the.expert
surr
and "global" effects'
lo"' eff ect, and'l 57o "global" effect
Best
3l
st]l,lulsF
on-the density of
does not depend
f/F-contribution
"local"
the
that
is reasonable to assume
..
derectors. How
ever,,n"
"
;r"1tr
berow
procedure suggested
TIF
Naturally
ventilated area
Mechanically
ventilted area
Ventilation
ntake
small ss leal(se
Best
Wrst
Best
Worst
0.05
0.09
0.002
0.02
0.01
0.03
1.10-3
l.1o-2
5.10-4
0.01
2.lf
2.10-3
104
7o'2
104
'Local"
detector density
Figure 6 TIF versus
ro
simp,irv
try
number per detector'
,yi*Uk*
:ffffi":lJ$
follows:
i::i"" *tr, o:t:t"^ot
i:,p::::.hr'ciu
pragmatic, ano is as
v^..'----Te I ro..uure is
uev formurus.
n-3
r.
asked:
3.
"Assume that there is only one detector installed to detect a gas leakage. What s the TIFprobability of not detecting such a leakage related to contributing class 'wrong location'?"
o. Denote this
scenario,,ro:i",ff"j:,",:,:,*iiyjfffif:tm;:it'*ratreastone
-=
lfi;; -nly on" d.t."tot.
*,1ii::;#''_-,,'_
means
0
/(
- o ;1 5k)
detecror. =
= TIF r^,"tn"(t
"'" ":ri::li:; :,{}:
For a given
number
*j,p:'f-::iiJii,:lfr
/<,
where
ro
This is rePeatedboth
I{'*;;,i[]Xi.'
r
r
AstepbystepprocedureisproposedtoestablishTlF-probabilitiesforaspecificapplication.
"local" effects related to a detector in an area containing gas
"global" effects related to the fact that there might not be gas at all in the area where the
detector is placed.
For a specific analysis where only one detector is considered, the TIF values may be used
Step 1: Identificationofdetection
stated in Table 9 and Table 10. However, in the situations whee several detectors ae used, it is
not straight forward to use these results. When the total CSU is calculated, the "T1F-contribution"
from each detector depends on the dependency, or so-called '-factors", and it is reasonable to
assign different dependency factors for the "local" and the "global" l/F-contribution.
will determine
--:-r-^red line detector. This choice
i'ti"t,.*g"^"t"::lio',t#:'o1"l,'J"';i:i';
aole
whether Table 9
as
system
or'l
Syslems'
Reliability Data for Control and Safely
9suNTEF
1998
Edilon.
33
(CC)
TIF for each contributine class
Step 8: Calculation of adjusted
contribiution is calculated
the
Foieach contributing
t
r
r
tl^t
.,-
=-l'"''l
formula:
'l+S' /
OPen
T,
=iw
DC u(Tr,.,,"
following
,l-S"
(Tr'0,
J'
3 in Table 12'
(S';) are ead from column 2 and
where the weights (wDC)and scores
TIF
Step 9: Calculation oftotal adjusted
class are sumnied up:
contributing
The TIF contributlons
"ut
TIF=TIFr +TIFz+TIF
will
the
''F
by
"o*
3.8 CalculationexamPle
highlight the content of each step'
A calculation example is given to
area
pas.
il1J;l*lrr3:iJ.i':ilii.':"
a inrrared
used in
point detector' hence rabre e is
Step 4.
TIF2,bn = TI Fz nn(1 - 03 5k)
TIF2s= TIF2,g(7 - 0.75k)
These numbers ae then to be inserted
e
using rhe "rert" part or rabre
$i,3iJi:Xt'Iii:,"[tflT.t:"tiT,u," . lksls
of area
Step 3: Identifcation of tvoe
a mechanically ventilated area
We assume that the gas'"utug" is in
,hr"ld
In the
identified
Yz
- Good state
l.
B ased on the
specification;
;;"
il;
:"d#;;;;;;:ti'
'"z'
3
TIF z ton = TIF 2.e*(1 - 0.7 5k) = ] 1']y-'
o'075
liF
ri,
1'
of influencing conditions
Step 6: Identification of state
Th scores are shown in Table I
I'
cause
scores for each direct failure
Step 7: Calculation of average
of avetage scores
See Tabe 1 I for calculation
(CC)
TIF.for.each^contributinB class
Step 8: Calculation of adjusted
on the formula:
based
is
contributing class inTable
The TIF contribution from-each
Il
@srNTEF
34
lL
.l+s,/,
T, =\wDCr(rm,.,,")' 1rm,,*
Edition.
35
,l-sr
TablellExamplecalculation;adjustingtheTlFprobability
lO-3
rj
r@srNTEF
36
Ediion.
1998
31
4. DemDossrnns
components'
control *d
the data dossiers of the
Y -sy-stem input data to
The following pages presents
"recomended" generic
the
4,
summarising
2-Table
Tab;
to
These are the input
PDS-II anaiYses'
e given in /13/ and
Thedatadossiersarebasedonthoseintheg5edition/13/,whichcontainsfailuremode
of these abbreviations
abbreviations no longer
l1'7 |
FollowingthedefinitionusedinoREDA,severaiseverityclassrypesarereferredtointhedata
are defined as follows:
dossiers. The various types
Critical failure
Afailurewhichcausesimmediateandcompletelossofasystem,scapabilityofprovidingits
outPut.
Degradedfailure i-:^^r
:"';li:l;l*:ii:J'i:::i'T;l'ili'ili";^,;"'n''
failure in time'
dru"lop into a critical
,;,tfo"'
t].""*
svstem's:'t*tl:tl1::viding
ts output'
Unknown
deduced'
recorded or could not be
Failure severiry was not
and quaitatively/
vely different
Notethatonlyfailuresclassifiedascritica]arepresentedandincluderltheestimatesofthe93
demand
edition.
1R
TIF3 r"-
0.001;
"'",
0.02
TIF = TIFI +
*)
38
:Retiability:DuhDjI!4 :
snmunr
39
1998 Edition.
PPQ&
Component:
Dte of Revion
DescrtPfion
1999-01-1
TheTlF-probabilityisentirelybasedonexpertjudgements.Detailsontheexpertjudgementare
sensor and
Pressure switch including
foundintheappendix.AsummaryofsomeofthemainargumentsisprovidedinSection2'3.
pneumatic switch
Overall
failure rate
Recommenileil Vlues
for Calcultion
Total rate
FTO
SO
2.3 Per
Overall
3.4 Per
106 hrs
1.39
SO:
0.00
Phase
r)
103 - 5 . 103
106 hrs
IV Softwe /15/.
Filter:
Observed:
cfro
100
Vo
(nv.
Previously Recomtneniled'
Values
r)
ailur e Rate
il
(95 edition)
for Calculntion
FTO:
hrs
No. of inventories = 12
No. of critical FTO failures = 1
No. of critical SO failures = 0
Cal. time ='l19 I
Coverage
ag-pobability
FTO:
SO:
Other:
0.61
"ui*"*
oREDAphaseIIIdata(phaserVcontainsnodataonprocessswitches).Theestimatedcoverage
(1007o in
the observecl coverage
judgement lassuming ZOVo coverage)and
is based on expert
oREDAphaseIII).TherateofFTofailuresisestimatedassumingacoverageol90vo
III was IOO 7o)' The rate of SO
T-boken
/6/:
Pressure switch
1.15
032
Pressure differential switch
For FTO: e=0'149 Per 10' demands
As s ess ment
Thegivenfailurerateessentiallyappliestopressure_switches.Thefailurerateestimateisan
and PDS I - with the complete
- *uinfy Uu'"a on OREDA-84
OR
ND
FTO:
SO:
Other:
2.28
T-boken
/6/:
T-boken
i6l:
T-boken
Flow switch
0.32
0.37
0.61
0.15
2.O4
40
Module:
uNUBLT
ano aIety
Reliabiily Data lor Control
1e98
Edition.
y5tErr1'
PS.data
RetiabilitYDaDo*t* t
Input Devices
Co*poo.nt,
Overall
DescriPtion
'
Dte
Failure mode
failure rate
Lo Med.
440
I Med.
320
5.6
Data source/comment
distributon
er 1 hrs)
Lo Me
1540
In Med.
2520
FTO:
0.25
SO:
0.15
T-boken
Hi
Hi
Hi
Hi
FTOhys.
FTOlrorru
Undetected
Tol rate
FTO
SO
Overall
= 5. 104
(95 eiliton)
Values for Calculation
Previously Recommendeil
0.1
2.1
ho =
o =
F
Iso =
SOhys.
1.5
SOunct.
2.0
SO/roret
3.5
hrs
Coverage
0'60
L, =
hrs
--^L^Lilit\'
TlF-probability
= 5'10'
-smartansm.= 3'104
6.8
1999-01-11
;;i"t
included.
5.2
of Revson
includes the
The pressure transmitter
and the
electronics
element, local
FTOunct. 2.0
5;
Conu entional
Pressure Transmitter'
ailur e Rate
Ass es
sment
previous estimate
is an update of the
The failure rate estimate
- mainly
based on
nn^e Iv'
oREDA iII
The rate of
u^tJni"
with .REDA phase lV
* ^"
"t
to
FTo
.'
Qsnmuur
RetiabiiitY Data
Pressure
Component:
judgement are
rherlF-probabilitv
O
found in the appendix'
'o'o**
43
1998 Edtion.
Dossigl!!$e
Transnitteyy
is entireivbasedon
,and
Module:
o.porr.nt,
C'
is provided
of some of the main arguments
''''
InPut Devices
Pressure Transmitter, Conventonal
Overall
failure rate
@er
hrs)
ffi
total
OREDA IV- /13/: Pressure switch'
Phase-Ivs"ftwae lr5l'
Filter:
inil"equip*"'" cls:
SENsoRs AND
T:cEss
k"ttY
lr
AND
Processing *"
= c's
Oil Drocesslng,
Fail. SeveritY Class = Crtical
ftn". sy.t"t
inventories = 205
No.
^r. of
.i"ti i. frO
o. of
"ti"
SO:
"r
ters.
Obsertted:
fto
100
ifl,
.t"rlu'
TAxcoD=sPR''Al'{D'
Vo
(Calculated'
including
tansmitters having
some kind of self'
rc$ arranEement
onlY,)
failures = o
SO failures = 0
PS3l-'
OREDA Phe III /1/ Database
pressure transmit'
conventional
i" ,"n"*,
FTO:
FuNcrN='oP'
inclwletl
transmitter
snmrur
-.
Module:
Reliabilitf,Data'Dossier
- PDSdata
Input Devices
Date of Revision
1
Remarlts
isolation valves.
in
Only displacement level transmitters are included
the OREDA Phase
III
and
[V data
Undetected
Coverage
0.90
106 hrs
0.50
1.5 Per
TIF-probabItY
Overall
falure rate
(per 106 hrs)
C alculation
Total rate
Ass essment
Overall
1998 Edtion.
Module:
1999-01 -1
FTO
SO
Jr
InPut Devices
Description
P.'DS-91!
Conventional
Component: l*vel (Disptacement) Transmitter'
Re
= 5'
1.89
Failure mode
Data source/commenl
distribution
FTO:
0.00
SO:
1.89
transmitters.
FIter:
Observed:
104
,so =
t00
Vo
Level
Coverage =
L,
hrs
TlF-probability
smarttransm'
o'is
= : l:1
- 3'10-
AND
Transmitter AND
princ. = Displacement AND
Inv.Phase=4
(Inv. System = Gas processing
Oilprocessing)
SENsoRs AND
AND
OR
AND
No. of inventories = l7
No. of critical FTO failures = 0
No. of critical SO failues = I
Cal. time = 530 208
6.17
FTO:
SO:
4.94
1.23
Data relevant
PS31-.
transmitters.
Falure Rate Assessment
Observed:
l
Thefailurerateestimateisanupdateofthepreviousestimate-mainlybasedonoREDAIII.
withoREDAphaselVoata.TherateofFTofailuresisestimatedassumingacoverageof9ovo
(observedinOREDAPhaseIIIwasl00To).Therateofsofailuresisestimatedassumrnga
coverageof50To(previouslyassumedtobe2}Vo'observedinOREDAPhaselVwasl00T)'
cno =
100 7o
.OR,,GP'
including
No. of inventories = 65
Total no. of failures = 50
transmitters having
Cal. time
(CaIcuIated
some
kind of selfiest
arrangement only,)
FTO:
0.21
T-boken
the
failure rdte
esftmates'
SilMTEF
Reliability Data f or
Safetv Systems'
)and
1998 Edition.
tRetiabifitvDallPcrssier'
Transmitter'
o*porr"rrtt l*vet (Displncement)
PDS<!!
Rliability Dta Dossier
Module:
Conuentional
InPut Devices
Description
1999-01-1
er l hrg
L,o Med.
10
- PDS-data "
transmitter
Hi
Remarks
Note that the data material for temperature
estimate
ftansmitters is scarce, i e', the failure rate
20
total
OREDA IV- /13/: Pressure switch'
Recommendeil Values
for C alculntion
Coverage
0.60
0.60
Total rate
FTO
SO
OveraII
trs
TlF-probabilitY
=
=
=
Lr,,
= 5' lOa
smaftansm'
h*
Fro
trso
IJndetected
- 3'10-
hrs
Coverage
hrs
TlF-probability
- smart tfansm'
=
=
5'104
3'104
Thefailurerateestimateisanupdateofthepreviousestimate-basedonoREDAPhaseIII
data' The
andflowtransmitters.Theoverallcovelagegivenaboveisestimatedmainlybasedonexpert
Qsumunr
TIF -Prob
ab
ilitY
As
s es
stne
SafetV Systems'
,nd
Reliability Eat'Dossier
Transmtter'
Component: Temperature
"1998 Edition.
Con'
- PDS'qala
lconveily
nt
judgement is
judgements' Details on the expert
entirely based on expert
is
TlF-probability
The
foundintheappendix.asunlmaryofsomeofthemainargumentsareprovidedinSection2.3.
T-boken
/6/:
Temperarure transrru$er
ffiFh*"
ui"l"u-t
Iv software /15/'
ror conventional temperature
Filter:
inu. equip**,
il;.
u'
Inv. Phase = 4
(Inv. SYstem = Gas Processrng
Oil processing)
No. of inventoriss = 19
FTO failures = 0
| o. of critic
0
I No. of critical SO failures =
FTO:
5'06
transmitter.
Obsented:
cfro
(
C alc ulate
ffansmitter
PS31-'
OREDA Phase III /l/ Database
temperature
conventional
for
Data relevant
100 7o
il includin g
s hav in g s ome
kind of self-test
arrangement onlY,)
No. of inventories = 8
Total no. of failures = 7
Cal. time = 197 808 hrs
the
as
"critical"
)sumrun
50
Module:
'
Co,
1998 Editon.
ReliabilityData'Dossier,'
PDS:ilat
Module:
InPut Devices
-,,
PDS-.data
Input Devices
Date of Revision
Descrption
1999-01-l I
The flow transmitter includes the sensing
element, local electronics and the process
The TlF-probability is entirely based on expert judgements. Details on the expert judgement is
found in the appendix. A summary of some of the main arguments are provided in Secton 2.3.
Remarks
isolation valves.
OveraII
Recommeniled Values
fot Calculttion
Undetected
Coverage
Total rate
FTO
0.60
so
0.50
Overall
TIF-probability
- smaft transm
for Calculation
L",
},FTO
0.1
per
106 hrs
l.so
1.4 per
106 hrs
L,
3.0
106 hrs
per
failure rate
er 1 hrs)
Failure mode
distribution
5.70
FTO:
2.85
SO:
2.85
5.104
3.104
TIF-probability
- smart transm.
Filter:
Obsemed:
cfro =
"so =
(95 edition)
Coverage
Data source/comment
7Vo
100
Vo
0.50
No. ofinventories = 10
No. of critical FTO failures = I
No. of critical SO failures = 1
Cal. time = 350 640
5.104
3 . l0-4
2.89
FTO:
1.24
SO:
1.5
Failure Rate
Ass es srnent
on oREDA
The failure rate estimate is an update of the previous estimate based
Obsertted:
III - with
oREDAphaselVdata.TherateofFTofailuresisestimatedassumingacovelageof60vo
(observedinoREDAPhaseIIIandIVwas 10070 ando4o,respectively).TherateofFTO
(previouslyassumedtobe}}vo,observedinOREDAPhaselVwasl00To).lheSofailure
rate includes 'Erratic output' failures.
cno = 100 lo
(Calculated including
transmitters having
No. of inventories = 72
Total no. of failues = 92
Cal- time
arrangement only,)
Note! Onlyfailures
=2422200hs
included in
the
rsrNTEF
52
Con
,iO
S"t"ty Systems.
1998 Edition.
53
PDS.data
.:il
Reliability.:Data Dossier
Input Devices
Module:
PDS.data
na
Component: Catalytic
Refere nc g s
Description
Overall
Failure mode
distribution
failure rate
(per 106 hrs)
FTO:
Lo Med.
Date of Revision
1999-01-1
Data source/comment
0.25
T-boken
/6i:
Flow transmitte
FARADIP.THREE
/7 | :
Flow transmitter
l5zu
Coverage
Total rate
0.7 per 106 fus
0.60
0.40
TlF-probability
1.6
per
106 hrs
for Cahalation
Llndetected
0.6 per 106 hrs
0.4 per 106 hrs
see secrion
...
(95 edition)
1.5
per 106hrs
I., =
5.5 pe 106
hs
TlF-probability
3 . lO4 - 0.1
r)
III
OREDA phase IV data. The rate of FTo failures is estimated assuming a coverage of 60 To
(previously assumed to be 90 7, observed in OREDA phase III was 38 vo). The rate of so
failures is estimated assuming a coverage
of.
OREDA phase III was 1007o). The FTO failure rate includes ,No output' and .Very low
output' failures.
in
SINTEF
54
Reliability:Data Dossier
998 Edtion.
PDS-data
Reliability:Da Dossier
Module:
Component: Cafalytic
PDS-data
Input Devices
Component: Catalytic
TI F -probabil As s e s s me nt
The
Overall
failure rate
F ailure Rat e Refere nc e s
es
Failure mode
distribution
Frod"t:
Data source/comment
0.5
t
Irl'Oundet; 1.4 i"
SOo"t: 0.2
S6und"t: 0.4 e"t
catalytic gas
c
detectors.
r
.4, lt
i"
5Fs '.'-:r
lg
Fher:
Inv. Eq. Class = FIRE& CAs DETECToRS
Inv. Att. Sensing principle = Catalytic
Inv. Phase = 4
Fail. Severity Class = Critical
5.09
No. of inventories = 24
No. of critical FTO failures = 0
No. of critical SO failues = 0
OsebergC 14/.
Data elevant fo conventional
catalytic gas
detectors.
No. of inventories = 431
No. of failues = 85 (25 critical)
Time = 10 215 888 hrs
OnIy failures classified as "critical" are
Note!
FTOA{at.aging 3.83
VI.LCAN /5/:
FTO/Stress
0.06
FlOntervent.
0.1'7
FTOh)TAL
4.06
SO/lrlat.aging 0.74
NOO:
SHH:
Sum
3.62
0.79
FTO:
4.41
64
?o
(Calculated including
detectors having some
kind of self+est
arrangement only)
Note!
0.06
SOllnput
Solrort
0.17
FTOunct,
FTO/T}TAL
cno
0.06
SOllntervent.
"critical" are
classiJed. as
"critical" are
2
3
SOhys.
SOunct.
SO/roTAL
Note! Onlyfailures
1.03
FTOlPhys. I
SENSPRI=TATALYTIC'
Observed:
SO/Stress
Note!
snmrnr
56
),1
5l
1998 Edtion.
- PDS.data
Reliabilify,ata Dossier
Module:
Input Devices
Module:
- PDS.data
Input Devices
Description
1999-01-
1 1
The TlF-probability is entirely based on expert judgements. Details on the expert judgement is
found in the appendix. A summary of some of the main arguments are provided in Section 2.3.
Remarks
interface unit.
'F
Overall
Recotnmended Values
for C alculation
Total rate
failure rate
Coverage
Undetected
FTO
0.80
0.7 per
106 hrs
so
0.70
0.1 per
106 hrs
Overall
TlF-probablity
Failure mode
@er 1 hrs)
distribution
Data source/comment
3.49
FTO:
3.49
SO:
0.00
Observed:
seesection
,no
cso
Filter:
=
=
I00Vo
}Vo
Inv.Eq.Class =FrRE&GAsDETEsroRs
AND
Inv.Phase=3
AND
(Inv.Att. Sensingprinciple=IR OR
Inv.Att. Sensingprinciple=lR/W) AND
Fail. Severity Class = Critical
14",
l)
2rFTO
so
L, =
Coverage
0.70
TIF-probability
3.lo4-o.lr)
Failure Rate
Ass essment
The failure ate estimate is an updte of the previous estimate - essentially based the Oseberg C
data j with OREDA phase fV data. The rate of FTO failures is estimated assuming a coverage
of 8O 7o (previously assumed tobe70Vo, observed in OREDA Phase IV was 100 Vo).The rate
of S O failures is estimated assuming a coverage of 70 Vo (previous estimate). The FTO failure
rate includes 'No output' failures.
No. of inventories = 54
No. of critical FTO failures = 4
No. of critical SO failures = 0
Cal. time = 147 176
4.1
FIOdd:
FIOUn&r:
SO"'':
soono.r:
2.9
Oseberg C /4/.
1.2
0
0
tectors.
IR
gas de-
No. ofinventories = 4l
Total no. of failures = 26 (4 critical)
Time=977 472lus
Note!
Only
included in
Qsnmrum
'' ':|:
Reliability Dat.Dos5ier.
Modufe:
1998 Edition.
- PDSdata
Input Devices
Module:
InPut Devices
da
Description
1999-01-1
The TlF-probability is entirely based on expert judgements. Details on the expert judgement is
found in the appendix. A summary of some of the main arguments are provided in Section 2.3.
as the
address/interface
unit.
,F alur,Rte Referenc
es
Overall
Recommended Values
for Calculation
FTO
SO
rate
1.3 per 106 hrs
2.4 per 106 hrs
overall
failure rate
Coverage
0.40
0.50
Total
hrs
TlF-probability
@er
lJndetected
10-3 -
hrs)
3.70
10'hrs
0'05
r)
1.5 per
hrs
=
=
=
L,
106
1.31
SO:
2.39
Data source/comment
Filter:
=
"no
,to =
L*
Fro
fso
FTO:
Obsemed:
Failure mode
distribution
50
Vo
98
7o
o-5 Perlo6hrs
2.0 Per 106 hrs
r)
3.73
FTO:
1.01
SPO:
2.72
Observed:
Failure Rate
Asses sment
a coverage
29Vo and50 Vo,respectively). The rate of SO failures is estimated assuming
was
98 7o)'
III
(complete)
Phase
in
OREDA
observed
robe2\Vo,
of 60 7o (previously assumed
lllwas
Inv.Phase=4
Coverage
cno = 29 Vo
Vo
have a
have
(Calculated including
of loop and
no self-test feature.
kind of self-test
arrangement only)
7o
combination
Note!
OnIy
included in
.QsrNTEF
60
Reliability Data
for'
o_t
1998 Edton.
Reliability,Data,Dossier
Module:
PDS.data
Input Devices
t..,
..., :::..
F ailuie,Rate Rlpr enc e s,
'
1999-01-1
failure rate
er
Date of Revision
Description
Overall
hrs)
as the address/-
interface unit.
No. of inventories = 53
No. of failures = 4 (l critical)
Recommended Values
Time= 12'l8528hus
Note!
OnIy
rate
hrs
1.5 per 106 hrs
VULCAN/5/:
FTO/Stress
0.13
FTO/Intervent.0.03
0.97
FTO/ror,t
Covrage
0.50
0.50
Total
included in
FTO/1.{at.aging 0.8i
for Calculntion
0.9 per 10
Overall
t)
"PDS-model".
hrs
TlF-probabitity
0.87
SO/Stress
0.43
SOllntervent.
0.03
SO/Input
SOlrorAL
4.39
included in
the
"critical" are
5.72
PDS.I /8/: Smoke detector
FTO/Phys. 0.4
FTOunct.
0.4
FTOlrorAL
0.8
SO/Phys.
Note!
l)
106 hrs
106 hrs
0-05 - 0.5
r)
SO{at.aging
Undetected
0.5 Per
1.3 per
L.,
=
IFro =
?rso =
L,
2.5
for Calcalation
(95 edition)
Coverage =
0.40
106
hrs
TlF-probability
0.05 - 0'5
r)
SOlFunct.
are included.
SOlror,r
The failure rate estimate is an update of the previous estimate - based on OREDA Phase III
is
data - with complete OREDA trI data (no inventories in phase IV). The late of FTO failures
estimated assuming a coverage of 50 Vo (observed in OREDA incomplete and complete Phase
7o, respectively). The rate of SO failures is estimated assuming a
(previously assumed to be 2OVo, obsewed in OREDA (complete) Phase III
was 98
Vo).
Vo
snmrer
Reliability Data
for
1998 Editon.
OJ
PDS-data
Reliability,Data Dossier -,PDS.data
Input Devices
As s es s me
nt
es
Overall
failure rate
@er ld
hrs)
2.35
FTO/Irlat.aging
1.28
VULCAN /5/:
FTO/Stress
0.14
FTOllntervent.0.05
Failure mode
distibution
Data source/comment
FTo/rorer
FTO:
0.88
SO:
1.47
SO/Stress
Observed:
36
98
Vo
Vo
Inv.Phase=4
DETEcroRs AND
Hear
AND
AND
FTO:
SPO:
0.82
1.39
Observed:
: cno=50Vo
(Calculated including
Further, 77
kind of self+est
arrangement only)
Vo
Note!
1.47
SO/l.lat.aging 0.49
0.32
SO/ftrtervent. 0.14
Filter:
"fro =
cso
=
a t
SO/Input
SOh'orAL
0.51
1.46
included.
FTOhys. 0.1
FTOlFunct. 0.2
FTO/1rAL 0.i
SO/Phys.
PDS I
SOlFunct.
are included.
SO/rort
@snmunm
o+
Reliability:Data Dossier
Module:
Reliabrlity Data
Input Devices
Date of Revion
1999-01-1
The TlF-probability is entirely based on expef judgements. Details on the expert judgement is
found in the appendix. A summary of some of the main arguments are provided in Section 2.3.
Remarks
'
''. :
_:ir :
Coverage
0.50
0.50
rate
hrs
4.1 per 106 hrs
Total
Overall
l)
hrs
TlF-probabitity
Undetectd
2.1 per 106 hrs
r)
Filter:
,oo =
cso =
l)
2.5
per l0 hrs
Fro
7"so
Lr,
hrs
50
7o
100
Vo
'
104 - 0'5
FTO:
SPO:
r)
3.20
3.98
Failure Rate
Observed:
Ass es sment
The failurp rate estimate is an update oi the previous estimate - based on OREDA Phase III
data - with complete OREDA III data (no inventories in phase IV). The rate of FTO failures is
estimated suming a coverage of 40 7o (observed in OREDA incomplet and.complete Phase
III was 48 Vo and 50 Vo, respectvely). The rate of SO failures is estimated assuming a
coverage of50 Vo (previously assumed tobe2OVo, observed in OREDA (complete) Phase
was 100
7o).
Flame
Phase=4
AND
AND
AND
0.40
Coverage
TlF-probability
Inv.Eq.Class =FIRE&GAsDETEcroRs
Obsened:
L",
for Calculation
Recomtnended Vlues
FTO
SO
- PDS-data
65
Module:
/I
PDS:iIata
Input Devices
Description
fr
1998 Edtion.
III
cfro = 48 Vo
(Calculated including
detectors having some
kind of self-test
Lrrangemenr only)
Note!
included in
the
"critcal" are
@snmrnr
66
Reliability'Data Dossier
'
{rol
o/
1998 Edition.
PDS'data
Reiability Data DO$liei
Module:
PDSdata
InPut Devices
Description
1999-01-l
Pushbutton including wiring
Remarks
@er 1 hrs)
Oseberg C /4/.
Reconmended Values
1.77
VI.JLCAN/5/:
FTO/Stress
O.l2
FTO/Intervent.0.12
FTOftort
"PDS-model".
2.01
SO/Stress
O.l2
SO/Intervent.
0.12
SO/Input
SO/rorAL
2.9'7
FTolrorer
SO/PhYs.
SO/Funct
SO/ror't
0.20
0.20
TIF-probabilitY
10-5
Coverage
0.20
TlF-probabilitY
lOs
rSO
I
I
I
L,
Note!
3.37
are
OnIy failures classified as "critical"
included.
I
I
1.1
FTOunct.
FTO
F ailur e Rt e As s es sment
FTO/PhYs.
OveraII
h.,
r
SO{at.aging
0.16
FTO
SO
lJndetected
Coverage
Total rate
rate estimates.
FTO/t{at.aging
for Calculaion
0.2
thexpert
sources, taking into account
The failure rate is estimated based on all listed data
I
I
I
I
I
I
1.3
N ot
ar e
B oth physic aI and functional failures
judgements.Theoverallcoveragegivenaboveisestimatedasiheaverageforbothfaiiure
judgement'
modes, also taken into account the expef
I
I
included'
O
nLy c
I
I
TI F -prob abilitY
As s es sm
ent
i
I
I
I
I
I
I
I
I
tu*^ury
of
to*"
of th"
-dn *g
@snmunm
68
Reliability Data
fc
)rtrot
1998 Edition.
Module:
Input Devices
PDS-data
Date of Revion
Overall
1999-01-1
Failure mode
dstribution
failure rate
er I hrs)
In Med.
Hi
0.
10
r 0.5
Data source/comment
5.8
Recommended Values
0.13
FTO
SO
OveraII
l)
for Calculation
Total rate
16 per 106 hrs
l6per
Coverage
106hrs
32 per 106
hrs
Undetected
0.90
0.90
TlF-probablity
5.lo-s-5.lo4r)
106
for Calculation
(95 edition)
hrs
L,i,
r)
III
III
data
discussed below: It is assumed that some of the observed FTO-failures in OREDA III is
included in the TlF-probabiiity. Further, for FTO-failures, only the current loop (i.e. one I-card,
etc.), not the entire PLC System, is required for a shut-down to be initiated. Thus, the estimated
rate of FTO-failures is reduced by approx. 7O Vo comparcd to the OREDA
III
@snmuen
'10
'
1cl
1l
PDS-data
Module:
F dilur
tRate,
Refeie nc es
er
Id
hrs)
OveraII
failure rate
(per
tion
106
hrs)
75.0
PDS I /8/:
FTO/unct.
channel
are incluiled.
0.05
FTO/Phys.
0.09
FTOunct.
0.05
channel
FTOIT1TAL
0.14
SOlPhys.
0.12
are included.
SOunct.
0.05
SO/rorAL
0.17
No. of inventories = 7 I
FTO/Phys.
FTOunct.
I
I
OREDA Phase
SO:
15.6
QO Vo) and
Inv.Phase=4
16.3
Obseried:
cno = 91
7o
Pe ch. 0.31
7o).
OREDA Phase
Loclc
UNITS
AND
AND
(Calculated including
No. of inventories = 52
Total no. of failures = 214
arrangement onlY)
Cal. time
Only
with
failure
PDS
I/8/: CPUMemorY
FTOITOTAL
Data elevant
Note!
0.14
SO:
units
Filter:
'14:7
0.09
59.4
FTO:
Note!
SO/Phys.
SOlFunct.
SOnorAL
IV Software i l5/.
FTO:
,fro = 9i 7o
,so = 88 7o
FTO/T)TAL
Data sourcelcbmment
Observed:
91.0
FTO/Phys.
SO/Phys.
SO/Funct.
SO/TqTAL
FTO/Phys. 0.02
FTOunct. 0.01
FTo/rorAL 0.03
PDS I /8/:
and
functional failures
are included.
OnIy crtical faIures are included'
@smunr
Reliability Data Dossier
Module:
Reliability Data
)ntrol
1998 Edition.
PSdata
Reliabilily Data Dossier
Component: PLC
Module:
SYstem
- PDS.dt
Date of Revision
1999-01-1
Overall
failure rate
@er 1 hrs)
Per ch. 0.21
Remarks
No data available in OREDA Phase IV
Data source/comment
FTO/Phys.
FTO/Funct.
0.17
0.01
FTO/TOTAL
O.]8
Recommended Values
0.02
SOunct.
0.01
SO/|OTAL
0.03
are included.
Total rate
0.01 per
0.2 per
Overall
for Cqlculatian
106
tus
0.18 per
0-001 per
0.02
0.90
0.001 per
0.90
0.02 per
TIF-probabIity
0.2
Unetected
Coverage
106 hrs
106 hrs
106 hns
10-s
106 hrs
106 hrs
per l0 hrs
per
106 hrs
TlF-probabilitY
10-5
T IF
-probability
Ass es s ment
@snmunm
'74
Reliability Data f
1998 Edition.
Module:
Reliability Data;Dossiei
Output Devices
1999-01-1
FTO
SO
Overall
0.90
0.001 per
0.90
TIF-probability
h., =
IFro =
lso -
0.18
L, -
0.2
per
0.001 per
for Calculntion
IV'
Undetected
Coverage
Recommended Values
for Calculation
10 hrs
10-5
(95 edfon)
10 hrs
FTO
SO
Total rate
0.8 per 106 hrs
0.7 per 106,hrs
Overall
1)
Coverage
hrs
106 hrs
106 hrs
IJndetected
0.00
0.30
0.5 per
TlF-probability
10-6 _
106
l0-s
hrs
r)
Date of Revision
Description
Remarks
.Total rate
- PDSidta
Valves
Date of Revision
1999-01-1
t)
;ntrol
=
h",
)"Fro =
Iso
=
o,
3.0 per
hrs
Coverage
106 hrs
t)
based on expert
No sourcs of failure rate data are identified. The failure rates are estimated
system'
judgement and the failure rate data found for PLC
hrs
TlF-probability
10-6
10-s
r)
T I F -probabilitY As
es s
ment
judgement rs
judgements. Details on the expert
The TlF-probability is entirely based on expert
Section 2'3
in
provided
aguments ae
found in the appendix. A summary of some of the maln
Qsnmrnr
76
1998 Edition.
PDS-data
:
Module:
-,
PDS-dat
OuQtut Devices
Valves
Overall
failure rate
(per 106 hrs)
1.1
Overall
F ailur e mo de di s t rib
FTO:
0.00
SO: l.l1
Observed:
,so =
failure rate
100
er 1 hrs)
9 .17
Vo
Fiher:
Inv. Eq. Class = \ilElIIADs AND X-MAS TREES ND
(Inv. System = Gas
OR
AND
Inv. System = Oil
production
Production)
Inv.Phase=4
Fail. Severity Class = Critical
op.
op.
AND
AI\'D
OR
OR
No. ofinventories = 18
No. of critical FIO failures = 0
No. of critical SO failures = I
Cal. time = 902 544
7.36
DOP:
EXL:
FTC:
FTOpen:
INL:
LCP:
PLU:
1.84
037
0.46
2.30
1.69
OR
0.15
0.15
SEAIJ'OR ACTUATOR'.
14
EXL:
FTC:
FTOpen:
INL:
OVH:
SEL:
SEP:
SIL:
SPO:
UNK:
0.28
OREDA Phase Il
3.81
2.1,2
0.14
/21 ,
Data relevant
0.28
No of inventories =322
No. of failures = 151
0.14
O.l4
1.12
Note!
Only
included in
0.43
0.14
FTOhys.
PDS
FTOunct.
valve etc.
FTO/ror,qt
I
N ote
SO/Phys.
SOunct.
SOlrorr
are
snmrur
't8
,R.U"lil!.itv'P4tq Po*l",
Module:
OutPut Devices
'
PDS'dat-
Date of Revision
1999-01 -1
TheTlF-probabilityisentirelybasedonexpertjudgements.DetailsontheexPertjudgementls
urgum"nts ar" p@
found in the appendix. A summary of some of th'e main
Remarks
F ailure Rate,References
and monitoring.
Recommended Values
for Clculation
Undetected
rate
Coverage
0'00
0'00
Total
hrs
0.3 Per 106hrs
1.6per
106hrs
TlF-probability
10-6 _ 10-s
FTO:
1.06
SO:
0.26
Filter:
Inv. Eq. Class = VALvES
(Inv. Syslem = Gas exPort.
Inv. System = Gas Processlng
Inv. System = Oil exPort .
Inv. System = Oil Processlng)
Inv. Phe = 4
Inv. Att, PPtication = ESD/PSD
Fail. SeveritY Class = Critical
(Fail. Item Failed <> Pilot valve
r)
,*"-r,
^t--***tlues
= 0.0 Per 106 hs
L",
IFro = 3.0 per 106 hrs
Xso = 0.5 Per lo hrs
Li,
t)
19
1998 Edition.
l)
Valves
Description
Overall
FTO
SO
hrs
Coverage
TlF-probability
uil. Suuunit
0.00
10-6.10sr)
FTOpen:
LCP:
1.12
1.12
ot
'cp"
OR
APPLIC=tsSD/PSD" MAffEM= tsODY'
Failure Rate
Ass essment
prevtous
data the failure rate estimte is an iterative updated' The
the
using
update
finally
is
estimate
esrimate is updared with the final phase III data, and this
of
coverage
a
assuming
estimated
oREDA phase IV data. The rate of FTO and so failures is
clefrrciency''
and
'structural
'Fail
on
demand'
to closc
0 vo .TheFTO failure rate incudes
Due to additional phase
III
Note!
are included
OnIyfailures classfied as "crtical"
in
the
falure rate
eslimt*
@snmunr
80
Module:
Output Devices
'
ol
8i
1998 Editon.
PDS-data
Retibility:Data Dossier - PDS'data
Valves
Module:
Output Devices
Valves
e R ete r e n c e s
Date
Description
Overall
failare rate
er I hrs)
9.17
t4
FaIure mode
EXL:
FTC:
FTOpen:
INL:
OVH:
SEL:
SEP:
SIL:
SPO:
UNK:
FTO/Phys.
0.28
3.81
2.12
0.14
No of inventories.= 322
0.28
No. of failures
0.14
Note! Onlyfailures
included in
0.43
Recommended Values
the
PDS
for Calculnton
FTO
SO
0.20
Undetected
1.4 per 106 hrs
0.30
Overall
Coverage
Total rate
0.14
151
0.14
l.l2
Data source/comment
tion
of Revion
1999-01-1
dtribu'
TlF-probability =
FTOlFunct.
FTOftoTAL
SO/Phys.
included.
SOlFunct.
Softorn
valve etc.
for Calcalation
(95 edition)
and
Failure Rate
TlF-probabilitY =
Ass essnent
III
of 2O 7o
OREDA phase IV data. The ate of FTO failures is estimated assuming a coverage
III was
Phase
complete
and
(previously assumed tobe0 To,observed in OREDA incomplete
of 30
coverage
a
assuming
The rate of SO failures is estimated
40 Vo and 67 7o, rcspectively).
7o (previously assumed to be 0
Reliability Data
@snmrem
82
f'
)rtrol
1998 Edition.
Reliabiliw'Data Dossie
PDSrdata
Reliabitity DCta,DoSiCi;'
Moduf
e:
Output Devices
Psiilata
Valves
Module:
The TIF-probabiliry is entirely based on expert judgements. Details on the expert judgement is
found in the appendix. A summary of some of the main arguments are provided in Section 2.3.
Overall
failure rate
@er I
es
0.45
Overall
failure rate
@er
ld
hrs)
hrs)
4.52
Data source/comment
FTO:
T-boken
0.45
tion
FTO:
1.69
SO:
2.83
"fro =
"so =
FTO:
0.11
Vo
94
7o
VALvEs
ESD/PSD
Shut-of
Phase=4
Critical
valve
Lo Med.
0.4 14
AND
OR
VA3l-.
0.07
0.36
0.07
Note!
T-boken
preted as FTO.
ND
OR
AND
AND
FTC:
FTOpen:
SO:
0.11
Filter:
67
/6/:
FTO.
Observed:
0.51
failure classif.catiott
i
:
Hi
snmrnr
84
Reliabilty Data
for
Module:
85
1998 Edtion.
Module:
TI F -p ro b ability A
Date of Revon
Description
1999-01-1
including actuator, pilot valve and local controVmonitoring. Both large and small control
s s ess
ent
judgement is
The TlF-probability is entirely based on expert judgements. Details on the expert
found in the appendix. A summary of some of the main arguments tt" plgytd:g tn Jgttion3'3'
Remnrks
valves ae included.
Recommended Values
for Calculation
Total rate
Small
FTO
'1
so
0.4
Overall
.1
.6
Coverage
- Iarge Valves
2.1per 106 hrs
0.60
0.70
O.l -0.2per
L.,
Undetected
TIF-probability
for Calculation
- Largevalves
- 8.0 per l06hrs
9.0 - 4.0 per 106 hrs
0.1 - 2-0 per106hrs
FTO:
3.97
SO:
l.O2
con'
trol valves including pilot valYe etc' Note! All
Vo of the registered valves
inches. Thus, 53 7o are
<
10
i.e.,
size
ae small,
sizes are
Obsemed:
^FO _-
106 hrs
,so =
.r<
LJ
oj^
'V
100
Vo
10-s
r FTO
(95 edition)
lL=
SO
L,
27.0
Coverage
0.65
> l0
inches.
No. of inventories = 99
No. of critical FTO failures = 10'5
No. of critical SO failures = 1
TIF-probability
1o-5
nt
The failure rate estimate is an update of the previous estimate - based on OREDA Phase III with OREDA phase IV data. Total rate of FTO-failures estimated by including the OREDA
and
Small
18.0
includ ed. 47
in
OREDA Phase IV was 25 Vo). The rate of SO failures is estimated assuming a coverage of 80
7o (previously assumed to be 65 %, observed in OREDA Phase IV was 100 7o).
DOP:
EXL:
FID:
FIC
FTOpen:
LCP
oTH
ovH
PLU
SO:
0.72
OREDA Phase
III
0.36
Data relevant
for
1.79
4.29
included.
2.15
1.43
.oR.
3.22
2.50
0.07
Note!
0;72
'GP'.
Only
in
included
failures classified as "crtcal" are
failure rate eslimates
the
Qsnmrum
86
Reliahility,Data:Dossier
Module:
hrs)
27.0'1
Module:
OuQtut Devices
Valves
Date of Revion
DOP:
FID:
Frc
FTOpen:
LCP
oTH
ovH
PLU
1999-01-l I
Data source/comment
1.04
OREDA Phase
4.17
Data relevant
5.21
l.M
3.12
3.12
.oR.'cP', srzE<=5.000.
2.o8
No. of inventories = 33
Total no. of failures = 66
7.29
DOP:
0.54
EXL:
FID:
0.54
Data relevant
0.54
/l/
Database
VA3l_.
FTC
3.81
FTOpen:
2.72
LCP
0.54
.oR. 'cP"
OTH
3.n
No. of inventories = 67
.18
T-boken
Ffo.
rate
Coverage
hrs
t)
0.2 per 106 hrs
OveraII
1.2
l)
1.0
per
per
106
106
hrs
0.oo
0.2 per
TlF-probabitity
t)
Undetected
0.00
106 hrs
1o-3
h", ?lFTo =
l,so =
L,
1-0
hrs
Coverage
0.00
120
Note! Onlyfailures
FTO:
for Calculation
FTO
so
slz>5.000.
No. offailures
Cai. time
Recommendeil Values
Total
SO:
Daoqsier :'PDS'dt.
Note! Onlyfailures
8.6
8'l
Relib,ility
14.16
}rol
fo
1998 Edition.
PDS-data
(per
Reliablty Data
per
106
r)
hrs
TlF-probability
l0
Note that trip of PSV does not necessarily lead to system trip
III'
The failure rate estimate is an update of the previous estimate - based on OREDA Phase
'Fail
to
as
classified
failures
OREDA 84 and other souces - with OREDA phase IV data. Only
'
T I F -p ro
ba
bility As s e s s m e nl
judgement is
The TlF-probabiliry is entirely based on expert judgements. Details on the expert
foundintheappendix.Asummaryofsomeofthemainarcu@
snmrnr
88
Reliabilty Data
fo.
1998 Edition.
PDS.data
,
Module:
Output Devices
Reliability-:Date :Dossier
P-DS.iIta
Valves
Module:
Output Devices
Valves
Overall
failure rate
er Id hrs)
failure rate
Overall
L .27
Data source/comment
FlO:
SO:
2.14
0.13
@er ld
4.4
Filter;
,fto = |vo
,so = 07o
VALvES
Inv. Phase=4
Inv. Att. Application = Relief
Inv. Eq. Class =
AND
AND
ANI)
INL/Degr.
22.06
INI-/Degr.
1.58
Sum/Degr. 23.63
EXl-/lncip.
1.58
EXl/krcip.
1.58
No. of inventories = 34
Sumllncip.
3.15
Note!
Also "Degraded" and
" In c ipent"
fai lures ar e
includeed, since no
" C ritic al "
ACT'.OR. 3.e.U-Or'.
17
the
failure rate
estimates.
failur es ar e
observed.
Lo Med.
28
Hi
t.5i
Observed:
o .78
hrs)
Data sourcelcomment
)snmrun
Reliabilty Dala
/t6l
ril
t17
13/
l4l
l5l
Trondheim, 1993.
NI\-
16/
nl
David J. Sflit}^, Retiability, MaintainabIty and Risk - Practical Methods for Engineers,
tgl
for
lgt
ll}t
William Denson et
Ragnar
Systems,
Aar/ et aI,
lt
Lars Bodsberg
tl2l
K.
ll3l
per
91
Harry F. Maftz and Ray A. \ffaller, Bayesian Reliability Analysis, IGieger Publishing
Company,1982.
REFERENCES
llt
fo
1998 Edtion.
Systems.
The PDS-II
report
ien and P. R. Hokstad. Handbook for performing exPert iudgmenL. SINTEF
sTF38 498419, 1998.
Hoktad and Ragnar Aa, Retiability Data for Control and Safety Systems, Revision
l.
41
tlst
1REDA Handbook; Affshore Retnbility Data Handbook, 3rd edition, oREDA Pafiicipants
(multiclient project on collection ofoffshoe reliability data)' 1997.
The PDS Forum was initiated in 1995, and follows up the PDS projects.
The main objective of the PDS Forum is to maintain a professional forum
for exchange of experience between Norwegian vendors and users of
control and safety systems. The primary focus is on safety and reliabilty
aspects of such systems. Research results are transferred, and personal
contacts between those working with offshore control and safety systems
are encouraged. Topics of the forum are:
Use of new standards for control and safetv svstems
.
.
.
The main activity of the PDS Forum in 1998 was to update the so-called
"PDS-recommended data". The present report summarizes the results from
this activity. For information regarding the PDS Forum please visit the web
s ite http ://www.s i ntef . n o/s i paalp rosjekt/pds-foru m.
The OREDA project is also acknowledged for allowing OREDA phase lV
data to be used in preparation of the present report. For information
regarding OREDA please visit the web site www.oreda.com
For information regarding the PDS-Tool please visit the web site of
Sydvest Software at www.sydvest.com.