0% found this document useful (0 votes)
96 views12 pages

A Collection of Formulas For GENG5507 (Version 9 5 14) : Acronyms

This document provides formulas that will be used in a course on reliability engineering. It includes formulas for probability, random variables, common probability distributions like the binomial, Poisson, normal, exponential, gamma, and Weibull distributions. The formulas cover topics like probability, mean, variance, probability mass functions, probability density functions, cumulative distribution functions, and reliability functions. The document is intended to help students with quizzes, exams, and understanding concepts covered in the course.

Uploaded by

Elena Todorovska
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
96 views12 pages

A Collection of Formulas For GENG5507 (Version 9 5 14) : Acronyms

This document provides formulas that will be used in a course on reliability engineering. It includes formulas for probability, random variables, common probability distributions like the binomial, Poisson, normal, exponential, gamma, and Weibull distributions. The formulas cover topics like probability, mean, variance, probability mass functions, probability density functions, cumulative distribution functions, and reliability functions. The document is intended to help students with quizzes, exams, and understanding concepts covered in the course.

Uploaded by

Elena Todorovska
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

1

A collection of formulas for GENG5507 (version 9514)


Manyoftheseformulaswillbeusedinthecourseofstudyingthisunit.Thiscollectionisintendedtoassistyou
indoingtheweeklyquizzes,midtermandfinalexam.Wewillcovertheuseofmany,butnotnecessarilyall,of
theseformulasduringthecourse.
Ifadditionalformulas(orcorrections)arerequired,pleaseusetheGENG5507discussionboardQuestionsfor
theunitcoordinatorabouttheorganisationofthecourseForumtopostyoursuggestion.
Acronyms
rv randomvariable
md mutuallydisjoint(theydonthaveanythingincommon,alsocalledmutuallyexclusive)
mde mutuallydisjointandexhaustive(exhaustiveunionistheentiresamplespace)
cdf cumulativedistributionfunction
pdf probabilitydensityfunction
pmf probabilitymassfunction
stddev standarddeviation
MTTF meantimetofailure
MTTR meantimetorepair

No. Formula Explanationofnotations Validity&


conditions
Application
1.Probability Rules
1
P(A) =
n
A
n
S

n
A
istheno.ofoutcomes
inA;n
S
isthetotalno.of
outcomesinthesample
space
u P(A) 1

n
S
= 1
uII

Probability
definition(for
equallylikely
outcomes)
2

P(A B) = P(A)
+P(B) - P(A B)
:union(OR). :jointor
intersection(AND)
Generaladdition
rule
3

P(A B) = P(A) +P(B)


A anu B aie
mu

4
P(A) = 1 - P(A)
A = not A, complement
of A
Complementrule
5

P(B|A) =
P(B A)
P(A)

Event (B|A): event B
given event A
P(A)>u Conditional
probability
6

P(A|B) = P(A)orP(A B) = P(A) P(B)


Independenceof
events
7 P(A B) = P(B|A) P(A)
OR= P(A|B) P(B)
WhenP(A)>0
ORwhen
P(B)>0
Multiplicative
rule
2

No. Formula Explanationofnotations Validity&


conditions
Application
8
P(B) = P(B|A) P(A) +P(B|A) P(A)

P(A
i
) = 1 -P(A)

0<P(A)<1
Lawoftotal
probability
9
P(A
k
|B) =
P(A
k
B)
P(B)
= P(A
k
)
P(B|A
k
)
P(B|A

) P(A

)
n
=1

k = 1, , n A
I
saremde.
P(B)>0and
P(A
i
)>0foralli

BayesTheorem
2.1RandomVariables
10
p = Hcon = _ t(t)

-
Jt
isthemeanorexpected
valueofthervtwithpdf
f.
Mean/expectatio
nofcontinuous
rvs
11
o
2
= Iorioncc = ] (t - p)
2
(t)

-
Jt
and o = o
2

o
2
=varianceofrvtwith
pdff.
o=stddev
Variance/stddev
ofcontinuous.
rvs
12
p = x(x)
uII x

isthemeanorexpected
valueofXwithpmff
Mean/expectatio
nofdiscretervs
13
o
2
= (x - p)
2
(x)
uII x
anu
o =

o
2

o
2
=varianceofXwith
pmff
o=stddevofX
Variance&std
devofdiscrete
rvs
14
F(x) = P(X x) = (y)
<x

F(x)andf(x)aretheCDF
andpmfofrvX,
respectively
u F(x) 1,
(x) =
uII x
1
u 1
Discrete
probability
distribution
15
F(t) = ] (x)Jx
t
-
or (t) =
dP(t)
dt

F(t)andf(t)arethecdf
andpdfofrvX
u F(t) 1
] (t)

-
Jt =
1
Continuousprob.
distrib.
16
R(t) = P(I > t) = _ (x)Jx

t

R(t) = 1 - F(t)
R(t)isthereliability
functionwithrvTofpdff
u R(t) 1
u t <
Reliability
functionof
continuousrv
17
r(t) =
](t)
1-P(t)
=
](t)
R(t)
=
](t)
] ](t)dt

t
or
r(t) = -
1
R(t)
dR(t)
dt

r(t)isthehazardrateof
rvT(alsowrittenash(t)
andZ(t))
r(t) > u Hazardrateofa
rv
18
(t) = -
dR(t)
dt
=
dP(t)
dt

f(t)isthepdfoftherv pdfofarv
19
B(t) = _ i(x)ux
t
0

H(t)=cumulativehazard
rateofacontinuousrv

3

No. Formula Explanationofnotations Validity&


conditions
Application
20
NTTF = ] t(t)Jt =

0
] R(t)Jt

0

MTTF
2.2.1DiscreteModels
BinomialDistribution
21

(x) =
n!
(n-x)!x!
p
x
(1 - p)
n-x

f(x)=pmfofX;p=prob.
ofsuccessinatrial;n=
totalno.oftrials;n!isthe
factorialofn,i.e.,
n! = n(n - 1) 1

x=0,1,,n
Binomialpmf
22
F(y) =
n!
(n-x)!x!
p
x
(1 -p)
n-x

x=0


y=0,1,,n
BinomialCDF
23 p = np Meanof
binomialdistrib.
24
= npq = np(1 - p)or

2
= npq = np(1 - p)
q=1p Stddev&
varianceof
binomialdistrib.
PoissonDistribution
25
(x) =

x
x!
c
-

isthemeanrateof
occurrence

x = u,1, ,n
Poissonpmf
26
F(y) =

x
x!
c
-

x=0

y=0,1,,n
PoissonCDF
27 p = whenconsideringPoisson
distributions
MeanofPoisson
distribut.
28
o = oro
2
=

Stddev&
varianceof
Poissondistrib.
4

No. Formula Explanationofnotations Validity&


conditions
Application
29
(x) =
(x)
n
n!
c
-x

x=expectednumberof
eventsin(0,x)
n = u,1, , Poissonprocess
2.2.2ContinuousModels
NormalDistribution
30
(t) =
1
o2n
c
-
(t-)
2
2c
2

- < t <

Normalpdf
31
F(t) = P(I t) = _
1
o2n
c
-
(x-)
2
2c
2
Jx
t
-

- < t <

NormalCDF
32
R(t) = 1 - u_
t -p
o
]

- < t <

Reliability
function(for
Normal
distributions)
33
Z =
I -p
o

T=normalrvwithmean
andstddevo;Zis
standardnormal
Transformfrom
normalto
standardnormal
34 u(z)=1u(z) u(z)istheCDFofthe
standardnormalatz.Itis
foundinastatistictable.
ValueofCDFofa
standardnormal
atz
35 P(Z z
u
) = Z=standardnormalrv;z
u

isthe100(1o)
th

percentileofthestandard
normal
Percentileofthe
standardnormal
distribution
LogNormalDistribution
36
(t) =
1
ot2n
c
-
(Int-)
2
2c
2

andoarethemeanand
stddevofX=ln(T)
T = e
X
> u
foi all X
pdfoflognormal
distribution
37
F(t) = u_
ln t - p
o
_
u(*)isthecdfofthe
standardnormal
CDFoflognormal
distribution.
5

No. Formula Explanationofnotations Validity&


conditions
Application
38
R(t) = 1 - u_
ln t - p
o
_ = u_
p - lnt
o
_
Note: 1u(z)
= u(z)
Reliability
functionof
lognormal
distribution
39
Mean=e
_+
o
2
2
_

Meanof
lognormal
distribution
40
Stddev=

e
|2+2o
2
]
- e
|2+o
2
]

Stddevoflog
normaldistrib.
ExponentialDistributions
41
(t) = c
-t

t u, > u pdfof
exponential
distribution
42
F(t) = 1 -e
-t
andR(t) = e
-t

CDFand
reliabilityof
exponential
distribution
43
p = HIIF = _ R(t)Jt = _ c
-xt

0
ut
= 1 z
For
exponential
distribution
meanof
exponential
distribution
44
o = stonJorJ Jc: =
1

For
exponential
distribution
Stddevof
exponential
distribution
45
NTTF =
1

For
exponential
distribution
Inverserelation
forexponential
distribution
46 MLEofthehazardrateforrightcensoredat
t
r
,thetimeofther
th
failure
z
`
=
1
1
r
| t

+ (n - r)t

=1
]

Beta=1
(exponential
distribution)
only
Specialcaseof
MLE=Number
offailures/Total
operatingtime
forcomponents
with
independent
Exponential
failuretimes
6

No. Formula Explanationofnotations Validity&


conditions
Application
47
t = -
ln|R(t)]

= -HIIF ln|R(t)]
For
exponential
distribution
Findingtgiven
R(t)
48

r(t) = alsowrittenash(t)andZ(t)
For
exponential
distribution
Hazardrate(also
knownas
instantaneous
failurerate)for
exponential
distribution
GammaDistribution
49
(t) =
z
([)
(zt)
[-1
c
-xt

t0
pdfofGamma
distribution
50
p =
[
z
onJ o
2
=
[
z
2

Meanand
varianceof
gamma
distribution.
WeibullDistribution
51
(t) =
|
q
_
t
q
]
|-1
c
-[
t
q

t0 pdfofWeibull
distribution
52
F(t) = 1 - c
-[
t
q

|
andR(t) = c
-[
t
q

t0 CDF/reliabilityof
Weibulldistr.
53
r(t) =
|
q
[
t
q

|-1
alsowrittenash(t)andZ(t)
ForWeibull
distribution
Hazardrate(also
knownas
instantaneous
failurerate)for
Weibulldistrib.
54
t = q|-ln (R(t))]
1
|

ForWeibull
distribution
Findingtfora
givenR(t)
55
p = p(1 +
1
[
)
I(x)istheGamma
functionfoundinTables.
(x) = x(x - 1). For
x=integer,I(x)=(x1)!
MeanofWeibull
distribution
7

No. Formula Explanationofnotations Validity&


conditions
Application
56
o
2
= p
2
_ [1 +
2
[
- j [1 +
1
[
[
2
_,
o = o
2

Variance/stddev
ofWeibull
distribution
3.SystemReliabilityandAvailability
57
R
Ss
= _R

n
=1

R
i
=reliabilityofitemi
[meansproductof
Seriessystem
reliability
58
R
Ss
= 1 -_(1 - R

)
n
=1

Asabove Parallelsystem
reliability
59
R
Ss
=
n!
i! (n - i)!
n
=k
R

(1 - R)
n-

R=reliabilityofitem,
n=totalnumberofitems,
k=numberofitemsmust
working
koutofn
systemreliability
60
R
Ss
(t) = c
-t

(t)

i!
n-1
=0

perfectswitch
ing,
exponential&
identicalitems
Standbysystem
reliability
61

R
Ss
= P(S|C)P(C) + P(S|C

)P(C

P(S|C)=prob.ofsystem
workingwhenCworks;
C

=Cnotworking
decomposition
method
Complexsystem
reliability
62
A = _
p

+ p

n
=1

i
=repairrate,
i
=hazard
rate
Exponential
distributed
failuretimes
Seriessystem
availability
63
A = 1 - _

+ p

n
=1

i
=repairrate,
i
=hazard
rate
Exponential
distributed
failuretimes
Parallelsystem
availability
8

No. Formula Explanationofnotations Validity&


conditions
Application
64
A = 1 -
1
( + )
n
[
n
i

n-1
=0
p

n-

=repairrate,=hazard
rate
Itemsare
exponentially
distributed&
identical
koutofn
system,
availability
65
HIIF(sys) = __
1
p

]
[
2
=1
_
-1 [
I _1 +
1
[
]
Components
musthave
Weibulldist.
withcommon
shape
parameter
MTTF(Sys)of2
Weibull(,
i
)
componentsin
series
configuration

4.DataSummary
66
x =
x

n
=1
n

x isthesamplemean Calculationof
samplemean
67
s
2
=
(x

- x )
2 n
=1
(n - 1)

s
2
isthesamplevariance Calculationof
samplevariance
68
HIIF =
t

n
1
n

t
i
isthetimetothefirst
failureofeachitemi,nis
thenumberofitems
failed
Pointestimate,
onlyusedwhen
allitemshave
failurewithinthe
testtime
69 NTTF = In T=totaltimeontestofall
items,n=no.offailures
Pointestimateof
MTTFusedwhen
thetestistime
terminatedand
notallitemsmay
havefailed.
70
i
n

i = 1, , n Cumulative%
(ranking)
71
i
n + 1

i = 1, , n Meanranking
72
i - u.S
n + u.4

i = 1, , n Bernards
formulafor
medianranking
9

No. Formula Explanationofnotations Validity&


conditions
Application
73
N
t
i
=
(n+1)-
t
i-1
1+(n-m)
and
i
t
i
= i
t
i
-1
+ N
t
i

r
t
i
=
i
t
i
-u.S
n +u.4

n=samplesize;
i
t
i
-1
=ordernumberof
thepreviousfailure;
m=no.ofpreceding
items;r
t
i
=rankingoft

Meanorder
numberand
rankingwith
censoreddata
74 Ordernumber=Previousordernumber+INC
INC =
(n+1)-Pcous ordcr numbcr
1+k

k=nm=no.ofitems
followingsuspendedset
Alternativeto
findingmean
ordernumberfor
censoreddata
75 Forcompletedatar=n:I = t

n
t=1

ForType1censoring:
I = t

+ (n - r)t

=1

ForType2censoring
I = t

+ (n - r)t

=1

Type1timeterminated
att
*.

Type2terminatedafter
afixednumberoffailures
r.t
r
isthetimeoftherth
failure
r=totalnumberof
failures,n=samplesize

5.ConfidenceIntervals
76
P _X

-z
o
n
p X

+z
o
n
] = 1 - o

=samplemean,o=
populationstddev,
=populationmean,
n=samplesize,o=risk,zis
suchu(z)u(z)=1o
ocanbe
replacedby
samplestddev
owhenn>30
100(1o)%two
sidedCIforthe
meanofnormal
distribution
77
NTTRNTTR

+z
o
n

MTTR

=samplemean
timetorepair,otherssee
above
Asabove 100(1o)%one
sidedCIforthe
meanofnormal
distrib.(repair
time)whenis
known
78
Lowerlimit:X

- to
2
,n-1
c
n

Upperlimit:X

+ to
2
,n-1
c
n

to
2
,n-1
=o/2criticalpoint
ofthetdistributionwith
degreeoffreedomn1,
seeabove
Whenn30or
distributionis
Normal
100(1o)%two
sidedCIforthe
meanofa
distrib.whenis
unknown
79
HIIF
2I
_
1-u,2
2

T=totaltimeontest,
_
1-u,2
2
=_
2
criticalvalue
withparameter1oand
2r,r=no.offailures
Failure
terminated
test:fortime
terminated
test,df=2r+2
100(1o)%one
sidedCIforthe
meanof
exponential
distribution
(MTTF)
80
HIIF
2I
_
u,2
2

As above

Asabove

Asabove
10

No. Formula Explanationofnotations Validity&


conditions
Application
81
2I
_
u2,2
2
HIIF
2I
_
1-u2,2
2

Asabove Asabove 100(1o)%two
sidedCIforthe
meanof
exponential
distrib.(MTTF)
6.HypothesisTesting
82
Z =
X

-
0
on

=populationmean,

0
=nullhypothesisvalue
of,o=populationstd
dev,n=samplesize,
X

=samplemean
Normalpopul.
orn30
Teststatisticfor
onesampleZ
testfor witho
known:
distributedas
N(0,1)underH
0

83
t =
X

-
0
Sn

S=samplestddev,
othersasabove
Asabove Teststatisticfor
onesamplet
testfor witho
unknown:
distributedast
n1

underH
0

84
_
2
=
21

p=populationmean
p
0
=nullhypothesisvalue
ofp
T=totalusagetime
r=numberoffailures
Failure
terminated
contextonly:
fortime
terminated,
use2(r+1)in
placeof2r
Teststatisticfor
a_
2
testforan
Exponentialdist.
mean:distrib.
as_
2
2
underH
0

7.RegressionAnalysis
85 = o

+ [

X + e

Simple
regressionform.
86
S
x
= x

- nx y
S
xx
= x

2
n

- nx
2

x=xvalues
y=yvalues
x =averageofxvalues
y=averageofyvalues
n=numberofpoints

87
b =
S
x
S
xx

Slopeparameter
estimate.
88 o = y - bx b=slopeparameter
estimate
y=averageofyvalues
x =averageofxvalues
Intercept
estimate.
89
r =
S
x

S
xx
S

Correlation
coefficient.
11

No. Formula Explanationofnotations Validity&


conditions
Application
90
r
2
=
(

-y)^2
n
=1
(y

n
=1
- y)^2

=yvalueestimated
fromregression.
y=averageyvalue
y=actualyvalue
Coefficientof
Determination.
91
:or(e

) =
(

-y

)^2
n
=1
n - 2

sc = :or(e

)
Seeprevious. Estimates
varianceofthe
errorterms
92
to
2
,n-2
sc _
1
n
+
n(x
0
-x
|
)^2
nx
i
2
-(x
i
)^2

x
0
=xvalueinquestion

Confidence
intervalforata
pointx
0

93
to
2
,n-2
sc _1 +
1
n
+
n(x
0
-x
|
)^2
nx
i
2
-(x
i
)^2

Seeprevious Prediction
intervalforata
pointx
0

94
Lowerlimit:B

- t
n-p
u 2
SE(B

)
Upperlimit:B

+ t
n-p
u 2
SE(B

)
B

=coefficientestimator
SE(B

)=standarderrorof
estimate
n=numberofdatapoints
p=numberof
parametersinregression
model
Requires
standard
assumptions
oflinear
regression
analysis
1uu(1 - o)%CI
forpartialslope
ofalinear
regressionmodel
95
I =
B

SE(B

Asabove Asabove Teststatisticfor


testing
significanceof
slopeinlinear
regression:
distrib.ast
n-p

underH
0

8.AdditionalFormulasfromWorkshops
96

PLL(perannum)=

. p
]
. n

ofPLL=Potentiallossoflife
(fatalities/year)

i
=frequencyincidenti(perannum)
p

=probabilityoffailurefromincidenti
n
i
=numberofpeopleaffectedbyincident
i
0
i
=fractionalexposuretimeofpersonnel
whereincidenticouldoccur

97
I
LSIR
(x, y) = I
R
i
(x, y
n
=1
)

LSIR=Locationspecificindividualrisk(also
knownaspeakrisk)
n=allpossibleeventsthatcanimpactofan
individualatlocation(x,y)andI
R
(x, y)=
riskvalueforaneventI(yr
1
)

12

No. Formula Explanationofnotations Validity&


conditions
Application
98

FAR =
No. o otolitics o:cr H ycors x 1u
8
Iotol no. o bours workcJ in H ycors

FAR=Fatalaccidentrate
99
CPR =
r(1 + r)
n
(1 + r)
n
-1

r=costofcapitalpertime
period(years)
n=numberoftime
periods
Capitalprice
reductionfactor
100 EAC=CPRxNPV CPR=seeabove
NPV=NetPresentValue
EAC=Estimatedannual
cost.
Estimatedannual
cost/benefitof
operatingagiven
asset.
101 J
max
= max
1<I<n
|max(|F(t

) -
F
`
(t

)|, |F(t

) -F
`
(t
-1
)|)]
n=samplesize,F(t
I
)=cdf
attimet
I
,F
`
(t
I
)=plotting
position(estimation)of
F(t
I
)attimet
I
,u
max
=
criticalvaluefoundinthe
tablefornanda
significanceo
Kolmogorov
Smirnovtest
102
SH =
p
S
-p
L
_o
S
2
+ o
L
2

S
=meanstrength
L
=
meanloado
S
=
strengthstddevo
L
=load
stddev
Safetymargin
103
IR =
o
L
_o
S
2
+o
L
2

Loading
roughness
104
R = u
`

p
S
-p
L
_o
S
2
+o
L
2
/

= u(SN)
Normal
probabilityof
strengthand
load
Reliabilityby
loadstrength,
singleload
105
NPI =
C
n
(1 + r)
n
,
n
0

NPV=Netpresentvalue
n=yearinwhichcost
incurred
r=discountrate
Businesscases,
LCC,NPV

You might also like