Você está na página 1de 42

Chapter 2

Basic Reliability Models


The Failure Distribution

C. Ebeling, Intro to Reliability & Maintainability Engineering,


Chapter 2 2nd ed. Waveland Press, Inc. Copyright © 2010 1
The Reliability Function

Let T = a random variable, the time to failure of a component

R(t)= Pr{T ≥ t}
where R(t ) ≥ 0 , R(0) = 1,and
limt →∞ R(t ) = 0

Often called the SURVIVAL FUNCTION

Chapter 2 2
Graph of a Reliability Function

R(t) Probability of surviving


1.2

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100 120

Chapter 2 3
The Cumulative Distribution
Function (CDF)

F(t)= 1- R(t)= Pr {T <t}


where F (0) = 0 and limt →∞ F (t ) = 1

Chapter 2 4
Graph of a CDF
F(t) The probability of a failure
1.2

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100 120

Chapter 2 5
The Density Function (PDF)

d F(t) d R(t)
f(t) = = -
dt dt

f (t ) ≥ 0 and ∫0
f (t )dt = 1

Chapter 2 6
Graph of a Density Function (PDF)
f(t)

0.02

0.018

0.016

0.014

0.012

0.01

0.008

0.006

0.004

0.002

0
0 20 40 60 80 100 120

Chapter 2 7
Relationship between PDF and CDF

F(t)= ∫ f(t ′) dt '


t
0


R(t)= ∫ f(t ′) dt '
t

Chapter 2 8
Finding Failure Probabilities

Pr{a ≤ T ≤ b} = R(a) - R(b)


= F(b) - F(a)

a b

Chapter 2 9
Example

The passive components of a distribution system for


natural gas has the following reliability function:

2
t
R(t) =1− ; 0 ≤ t ≤10 yrs
100
Find: a. R(3 yrs)
b. The CDF, F(t)
c. Pr{1<T<3}
d. The density function, f(t)

Chapter 2 10
Example - solution

32 t2
a . R ( 3) = 1 − = .91 R(t ) = 1 − ; 0 ≤ t ≤ 10 yrs
100 100

b. F(t) = 1 - [1 - t2 / 100] = t2 / 100

c. Pr{1<T<3} = F(3) - F(1) =. 09-.01 =. 08


f(t)
dF ( t ) 2t t
d . f (t ) = = = ; 0 ≤ t ≤ 10
dt 100 50
10 t

Chapter 2 11
Mean Time to Failure (MTTF)

∞ ∞
MTTF = ∫ t f(t) dt = ∫ R(t) dt
0 0

Note alternate notation: MTTF = E[T]

Chapter 2 12
Derivation of MTTF (Appendix 2A)

∞ dR(t )
MTTF = ∫ tf(t) dt = ∫

0
−t dt
0 dt
Integration by parts:

∞ ∞
MTTF = −tR(t ) 0 + ∫ R(t )dt = ∫ R(t ) dt
0
0
Since

−tR(t ) 0 = lim t R (t ) − 0 R(0) = lim t exp − ∫ λ (t ')dt '⎤ = 0



∞ t

t →∞ t →∞ ⎢⎣ 0 ⎥⎦

Chapter 2 13
Example - MTTF

For the distribution system, find the MTTF.

∞ 10 t
MTTF = ∫ t f (t )dt = ∫ t dt
0 0 50
3 10 3
t 10 100 2
= = = = 6 yr
150 0 150 15 3

Chapter 2 14
Example - MTTF revisited

For the distribution system, find the MTTF.

2
∞ 10 t
MTTF = ∫ R (t )dt = ∫ (1 − )dt
0 0 100
3 10
t 1000 2
=t− = 10 − = 6 yr
300 0 300 3

Chapter 2 15
Median Time to Failure and Mode

R( t med )= .5 = P{T ≥ t med }


f( t mode )= max f(t)
0≤t<∞

Chapter 2 16
Example - median and mode

R(t) = 1 - t2 / 100 = .5

t2 /100 = .5

t2 = 50 or tmed = 7.07 yrs

f(t)
by inspection
tmode = 10 yrs
10 t

Chapter 2 17
Comparison of MTTF, Median, & Mode

Chapter 2 18
Design Life

Find tR such that R(tR) = R

For example:
Find that time, t.99 such that R(t.99) = .99
Then t.99 is the 99 percent design life.
One percent will fail before time t.99

Chapter 2 19
Example – design life

R(t) = 1 - t2 / 100 = .96

t2 /100 = .04

t2 = 4 or t.96 = 2 yrs

f(t) .96

2 10 t

Chapter 2 20
Variance & Standard Deviation
definitional form:


σ = ∫ (t - MTTF ) f(t)dt
2 2
0

computational form:

∞ 2
σ = ∫ t f(t)dt - (MTTF )
2 2
0

Chapter 2 21
Example - standard deviation
2
⎛ 2⎞
3
10 t
σ =∫ 2
dt − ⎜ 6 ⎟
0 50 ⎝ 3⎠
10 2
⎛ 2 ⎞ 10, 000
4
t
= −⎜6 ⎟ = − 44.444 = 5.55
200 0 ⎝ 3 ⎠ 200

or σ = 5.55 = 2.36 yr .

Chapter 2 22
Hazard Rate Function
P( A ∩ B)
P( A | B) =
Pr {t ≤ T ≤ t+ Δt}= R(t)- R(t+ Δt)
P( B)

R(t)- R(t + Δt)


Pr {t ≤ T ≤ t + Δt | T ≥ t} =
R(t)

-[R(t+Δt)-R(t)] 1 -dR(t)
1 f(t)
λ(t) = lim Δt R(t) = =
Δt →0 dt R(t) R(t)

Chapter 2 23
Example - hazard rate function

t / 50 t / 50 2t
λ (t ) = = =
1 − t / 100 100 − t
2 2
100 − t 2

100

λ (t)
IFR

t
10 yr

Chapter 2 24
Derivation of R(t) from the
Hazard Rate Function
d R (t) 1
λ (t) = −
dt R (t)

t R (t ) d R(t ')
∫ λ (t ')dt ' = − ∫
d R(t )
λ (t )dt = −
R(t ) 0 1 R (t ')
t
− ∫ λ (t ')dt ' = ln R(t )
0

R (t ) = exp ⎡ − ∫ λ (t ')dt '⎤


t

⎢⎣ 0 ⎥⎦
Chapter 2 25
Hazard Rate Function & R(t)
t

R(t ) = e 0
λ ( t ') dt '

Example:

λ (t ) = .02t , then
t

R(t ) = e 0
− .02 t ' dt '
=e −.01t 2

Chapter 2 26
The Bathtub Curve

Chapter 2 27
Human Mortality Curve

Probability Male Death

0.024

0.019

0.014

0.009

0.004

-0.001
0 10 20 30 40 50 60 70
Age

Chapter 2 28
More on the Bathtub Curve
Burn-in Useful Life Wearout
Characterized by DFR CFR IFR

Caused by Manufacturing defects Environment Fatigue


Welding flaws, Cracks, Random loads Corrosion
Defective parts, Poor Human error Aging
quality control, "Acts of God" Friction
Contamination, Poor Chance events Cyclical loading
workmanship

Reduced by Burn-in testing Redundancy Derating


Screening Excess strength Preventive Maint.
Quality control Parts replacement
Acceptance testing Technology

Chapter 2 29
Average Failure Rate (AFR)
t2
1 ln R( t 1 ) - ln R( t 2 )
AFR( t 1,t 2 ) = ∫
t 2 - t 1 t1
λ (t ′) dt ' =
t 2 - t1

since:
- ∫ 0λ (t ′)dt '
t
R(t) = e
Note: AFR(0,t) = AFR(t) = -ln R(t) / t

Chapter 2 30
Example - Average Failure Rate

L
− ln M1 −
t O
P
2

AFR (t ) = N 100 Q
t

AFR(5 yr) = - ln [1 - .25] / 5 = .0575 failures / yr

Chapter 2 31
Conditional Reliability

R(t|T0 ) = P {T > T0 +t | T > T0 }


Event A Event B

P {T > T 0 + t} R( T 0 + t)
= =
P {T > T 0 } R( T 0 )

Chapter 2 32
Residual MTTF

∞ ∞
R(t +T 0 )
MTTF ( T 0 )= ∫R( t | T 0 ) dt = ∫ dt
0 0
R( T 0 )

1
= ∫
R( T 0 ) T 0
R( t′) dt '

where t’ = t + T0

Chapter 2 33
Example - conditional reliability
R ( t + 1) 1 − ( t + 1) / 100
2 t2
R ( t |1) = = R(t) = 1−
R (1) 1 − (1 / 100) 100

100 − ( t + 1) 2
= = 1.01 − ( t + 1) 2 / 99
100 (.99 )

Therefore: R(5|1) = 1.01-36/99=.646


where R(5) = 1 - .25 = .75
10
1 10t 2
1 ⎡ t3 ⎤
MTTF (T0 ) = ∫
1 − T0 /100 0
2 T
1−
100
dt = t−
.99 ⎣ 300 ⎥⎦T =1

0

1 ⎡ 103 1 ⎤ 5.67
or MTTF (1) = ⎢10 − −1 + ⎥ = = 5.72 yr.
.99 ⎣ 300 300 ⎦ .99

Chapter 2 34
Residual MTTF(T0)

MTTF(T0)

7.000
6.000
5.000
4.000
3.000
2.000
1.000
0.000
0 2 4 6 8 10 12

Chapter 2 35
Student Exercise #1

A panel consisting of analog displays has a


reliability function given by
R(t) = (200-t)/200 for 0 < t < 200
where t is measured in 1000’s of hr. Find:
a. R(50,000) and R(12,000)
b. R(50,000 | 12,000)
c. MTTF
d. MTTF(12,000)

Chapter 2 36
Student #1- solution

a. R(50) = (200 - 50)/200 = .75


R(12) = (200 - 12)/200 = .94
b. R(50|12) = R(62) / R(12) = (200 - 62) /200 / .94=
.69 / .94 = .734
2 200
200 t 40000
c. MTTF = ∫ 1 − t / 200 dt = t − = 200 − = 100
0 400 0 400
200
1 200 1 ⎡ t ⎤ 2
d . MTTF (12) = ∫
.94 12
(1 − t / 200) dt = ⎢ t− ⎥
.94 ⎣ 400 ⎦12
= 94

Chapter 2 37
Student Exercise #1 (continued)

e. What is the shape of the density function?


f. Is the hazard rate function increasing or
decreasing?
g. Compute the average failure rate over the
first 100,000 miles.

Chapter 2 38
Student Exercise #1-solution

e. f(t) = -dR(t) /dt


= - d {(200-t)/200}/dt =1/200 f(t)
t
200

1 / 200 1
f . λ(t ) = =
(200 − t ) / 200 200 − t
t
200

g. AFR(100) = - ln {(200-100)/200} / 100


= .007 failures/ 1000 mi.

Chapter 2 39
Student Exercise #2
Insulators on a power distribution system have
a reliability function with t measured in yr.
R(t) = 1 / (1 + .05t) where t >= 0
Find:
a. F(1 yr) and R(2)
b. R(2|1)
c. The hazard rate function
(optional)
d. AFR(3)

Chapter 2 40
Student Exercise #2 - solution

a. F(1) = Pr{T<1} = 1 - 1/[1+.05(1)] = 1 - .9524 =


.0476
R(2) = Pr{T>2} = 1/[1+.05(2)] = .9091
b. R(2|1) = Pr{T>3 | T>1} = R(3) / R(1)
= [1+.05(3)]-1 / .9524 = .913
c. f(t) = -d [1+.05t]-1 /dt = .05[1+.05t]-2
λ(t) = f(t)/R(t) = .05/[1+.05t] which is DFR
d. AFR(3) = {-ln [1+.05(3)]-1 } / 3
= -ln .8696 /3 = .0466 failures per year.

Chapter 2 41
Summary - The Four Functions

f(t), the Probability Density Function (PDF)


F(t), the Cumulative Distribution Function
(CDF)
R(t), the Reliability Function
λ(t), the Hazard Rate Function

Chapter 2 42

Você também pode gostar